Untitled

Emerging Infectious Diseases

Emerging Infectious Diseases is published month-

ly by the National Center for Infectious Diseases,

Centers for Disease Control and Prevention (CDC),

1600 Clifton Road, Mailstop D61, Atlanta, GA 30333,

USA. Telephone 404-371-5329, fax 404-371-5449,

email [email protected].

The opinions expressed by authors contributing to

this journal do not necessarily reflect the opinions of

the Centers for Disease Control and Prevention or the

institutions with which the authors are affiliated.

All material published in Emerging Infectious

Diseases is in the public domain and may be used

and reprinted without special permission; proper cita-

tion, however, is required.

Use of trade names is for identification only and

does not imply endorsement by the Public Health

Service or by the U.S. Department of Health and

Human Services.

∞ Emerging Infectious Diseases is printed on acid-

free paper that meets the requirements of

ANSI/NISO 239.48-1992 (Permanence of Paper)

Electronic Access

Retrieve the journal electronically on the World Wide Web (WWW) at

http://www.cdc.gov/eid or from the CDC home page (http://www.cdc.gov).

Announcements of new table of contents can be automatically emailed to

you. To subscribe, send an email to [email protected] with the following

in the body of your message: subscribe EID-TOC.

Editors

D. Peter Drotman, Editor-in-Chief

Atlanta, Georgia, USA

David Bell, Associate Editor

Atlanta, Georgia, USA

Vincent Deubel, Associate Editor

Lyon, France

Stephanie James, Associate Editor

Bethesda, Maryland, USA

Brian W.J. Mahy, Associate Editor

Atlanta, Georgia, USA

Takeshi Kurata, Associate Editor

Tokyo, Japan, USA

Martin I. Meltzer, Associate Editor

Atlanta, Georgia, USA

David Morens, Associate Editor

Washington, DC, USA

Tanja Popovic, Associate Editor

Atlanta, Georgia, USA

Patricia M. Quinlisk, Associate Editor

Des Moines, Iowa, USA

Gabriel Rabinovich, Associate Editor

Buenos Aires, Argentina

Didier Raoult, Associate Editor

Marseilles, France

Pierre Rollin, Associate Editor

Atlanta, Georgia, USA

Mario Raviglione, Associate Editor

Geneva, Switzerland

Polyxeni Potter, Managing Editor

Atlanta, Georgia, USA

Joseph E. McDade, Founding Editor

Rome, Georgia, USA

Mary Anne Castranio

Carolyn P. Collins

Teresa M. Hood

Ann Kitchen

Anne D. Mather

Carol D. Snarey

Reginald S. Tucker

Cathy E. Young

Dennis Alexander, Addlestone Surrey, United Kingdom

Ban Allos, Nashville, Tennessee, USA

Michael Apicella, Iowa City, Iowa, USA

Ben Beard, Atlanta, Georgia, USA

Barry J. Beaty, Ft. Collins, Colorado, USA

Martin J. Blaser, New York, New York, USA

David Brandling-Bennet, Washington, D.C., USA

Donald S. Burke, Baltimore, Maryland, USA

Charles H. Calisher, Ft. Collins, Colorado, USA

Arturo Casadevall, New York, New York, USA

Thomas Cleary, Houston, Texas, USA

Anne DeGroot, Providence, Rhode Island, USA

Vincent Deubel, Providence, Rhode Island, USA

Ed Eitzen, Washington, D.C., USA

Duane J. Gubler, Fort Collins, Colorado, USA

Scott Halstead, Arlington, Virginia, USA

David L. Heymann, Geneva, Switzerland

Sakae Inouye, Tokyo, Japan

Charles King, Cleveland, Ohio, USA

Keith Klugman, Atlanta, Georgia, USA

S.K. Lam, Kuala Lumpur, Malaysia

Bruce R. Levin, Atlanta, Georgia, USA

Myron Levine, Baltimore, Maryland, USA

Stuart Levy, Boston, Massachusetts, USA

John S. MacKenzie, Brisbane, Australia

Tom Marrie, Edmonton, Alberta, Canada

John E. McGowan, Jr., Atlanta, Georgia, USA

Stephen S. Morse, New York, New York, USA

Philip P. Mortimer, London, United Kingdom

Fred A. Murphy, Davis, California, USA

Barbara E. Murray, Houston, Texas, USA

P. Keith Murray, Ames, Iowa, USA

Stephen Ostroff, Atlanta, Georgia, USA

Rosanna W. Peeling, Geneva, Switzerland

David H. Persing, Seattle, Washington, USA

Gianfranco Pezzino, Topeka, Kansas, USA

Richard Platt, Boston, Massachusetts, USA

Leslie Real, Atlanta, Georgia, USA

David Relman, Palo Alto, California, USA

Nancy Rosenstein, Atlanta, Georgia, USA

Connie Schmaljohn, Frederick, Maryland, USA

Tom Schwan, Hamilton, Montana, USA

Ira Schwartz, Valhalla, New York, USA

Tom Shinnick, Atlanta, Georgia, USA

Robert Shope, Galveston, Texas, USA

Bonnie Smoak, Bethesda, Maryland, USA

Rosemary Soave, New York, New York, USA

P. Frederick Sparling, Chapel Hill, North Carolina, USA

Jan Svoboda, Prague, Czech Republic

Bala Swaminathan, Atlanta, Georgia, USA

Robert Swanepoel, Johannesburg, South Africa

Phillip Tarr, Seattle, Washington, USA

Timothy Tucker, Cape Town, South Africa

Elaine Tuomanen, Memphis, Tennessee, USA

David Walker, Galveston, Texas, USA

Mary E. Wilson, Cambridge, Massachusetts, USA

Editorial Board

The print journal is available at no charge to public health professionals

YES, I would like to receive Emerging Infectious Diseases.

Please print your name and business

address in the box and return by fax to

404-371-5449 or mail to

EID Editor

CDC/NCID/MS D61

1600 Clifton Road, NE

Atlanta, GA 30333

Moving? Please give us your new address (in the box) and print the number of your old

mailing label here_________________________________________

Full text free online at

www.cdc.gov/eid

Production Editors

Perspectives

Mycobacterial Aerosols and Respiratory Disease . . . .763

J.O. Falkinham III

Global Screening for Human Viral Pathogens . . . . . . .768

N.G. Anderson et al.

Synopses

Salmonella Control Programs in Denmark . . . . . . . . . .774

H.C. Wegener et al.

Disease Surveillance and the Academic,

Clinical, and Public Health Communities . . . . . . . . . . .781

R.W. Pinner et al.

Research

Acute Flaccid Paralysis

and West Nile Virus Infection . . . . . . . . . . . . . . . . . . . .788

J.J. Sejvar et al.

West Nile Virus in Farmed Alligators . . . . . . . . . . . . . .794

D.L. Miller et al.

Emergence and Global Spread of a

Dengue Serotype 3, Subtype III Virus . . . . . . . . . . . . .800

W.B. Messer et al.

Molecular Epidemiology of O139

Vibrio cholerae: Mutation, Lateral

Gene Transfer, and Founder Flush . . . . . . . . . . . . . . .810

P. Garg et al.

Amoeba-Resisting Bacteria and

Ventilator-Associated Pneumonia . . . . . . . . . . . . . . . . .815

B. La Scola et al.

Antimicrobial Resistance Markers of Class 1

and Class 2 Integron-Bearing Escherichia coli

from Irrigation Water and Sediments . . . . . . . . . . . . . .822

M.T. Roe et al.

Hantavirus Prevalence in the IX Region of Chile . . . . .827

M. Täger Frey et al.

Antimicrobial Susceptibility Breakpoints

and First-Step parC Mutations in

Streptococcus pneumoniae: Redefining

Fluoroquinolone Resistance . . . . . . . . . . . . . . . . . . . . .833

S. Lim et al.

Mutations in Putative Mutator

Genes of Mycobacterium tuberculosis

Strains of the W-Beijing Family . . . . . . . . . . . . . . . . . .838

M.E. Rad et al.

Yellow Pigmy Rice Rat

(Oligoryzomys flavescens) and Hantavirus

Pulmonary Syndrome in Uruguay . . . . . . . . . . . . . . . .846

A. Delfraro et al.

Dispatches

Serologic Evidence of West Nile Virus

Infection in Horses, Coahuila State, Mexico . . . . . . . . .853

B.J. Blitvich et al.

Serologic Evidence of West Nile Virus

Infection in Horses, Yucatan State, Mexico . . . . . . . . .857

M.A. Loroño-Pino et al.

Serologic Evidence of West Nile

Virus Transmission, Jamaica, West Indies . . . . . . . . . .860

A.P. Dupuis II et al.

A Peer-Reviewed Journal Tracking and Analyzing Disease Trends Vol.9, No.7, July 2003

On the Cover:

Henri Rousseau—known as

Le Douanier Rousseau

(1844–1910).

The Snake Charmer (1907).

Oil on canvas,

169 cm x 189.5 cm.

Musée d’Orsay, Paris,

France. Credit: Réunion des

Musées Nationaux/Art

Resource, NY

About the Cover, pg. 901

All material published in Emerging Infectious Diseases is in the public

domain and may be used and reprinted without special permission; proper

citation, however, is appreciated.

Sulfa Resistance and Dihydropteroate

Synthase Mutants in Recurrent

Pneumocystis carinii Pneumonia . . . . . . . . . . . . . . . . .864

A. Nahimana et al.

VIM- and IMP-Type Metallo-β-lactamase–

Producing Pseudomonas spp. and

Acinetobacter spp. in Korean Hospitals . . . . . . . . . . . .868

K. Lee et al.

Leishmaniasis in Germany . . . . . . . . . . . . . . . . . . . . . .872

G. Harms et al.

Probable Dengue Virus Infection among

Italian Troops, East Timor, 1999–2000 . . . . . . . . . . . . .876

M.S. Peragallo et al.

HIV Epidemic among Young

Thai Men, 1991–2000 . . . . . . . . . . . . . . . . . . . . . . . . . .881

K. Torugsa et al.

Letters

Taenia solium Cysticercosis,

Irian Jaya, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . .884

T. Wandra et al.

Recombinant Vaccine–Derived

Poliovirus in Madagascar . . . . . . . . . . . . . . . . . . . . . . .885

D. Rousset et al.

West Nile Virus Infection in Crocodiles . . . . . . . . . . . .887

A. Steinman et al.

Rickettsia aeschlimannii in Spain:

Molecular Evidence in Hyalomma

marginatum and Five Other Tick

Species that Feed on Humans . . . . . . . . . . . . . . . . . . .889

P. Fernández-Soto et al.

Hantaviruses in São Paulo State, Brazil . . . . . . . . . . . .891

L.T.M. Figueiredo et al.

Israeli Spotted Fever Rickettsia

in Sicilian Rhipicephalus sanguineus Ticks . . . . . . . . .892

G.M. Giammanco et al.

Co-feeding Transmission and Its

Contribution to the Perpetuation of the

Lyme Disease Spirochete Borrelia afzelii

(In Reply) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .893

S. Randolph and L. Gern

Co-feeding Transmission and Its

Contribution to the Perpetuation of the

Lyme Disease Spirochete Borrelia afzelii

(In Reply to Randloph and Gern) . . . . . . . . . . . . . . . . .895

D. Richter et al.

News & Notes

Conference Summaries

West Nile Virus Southeast Conference . . . . . . . . . . . .897

D. Rimland et al.

West Nile Virus and Wildlife Health . . . . . . . . . . . . . . .898

P.P. Marra et al.

About the Cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .901

A Peer-Reviewed Journal Tracking and Analyzing Disease Trends Vol.9, No.7, July 2003

Environmental opportunistic mycobacteria, including

Mycobacterium avium, M. terrae, and the new species M.

immunogenum, have been implicated in outbreaks of

hypersensitivity pneumonitis or respiratory problems in a

wide variety of settings. One common feature of the out-

breaks has been exposure to aerosols. Aerosols have been

generated from metalworking fluid during machining and

grinding operations as well as from indoor swimming pools,

hot tubs, and water-damaged buildings. Environmental

opportunistic mycobacteria are present in drinking water,

resistant to disinfection, able to provoke inflammatory reac-

tions, and readily aerosolized. In all outbreaks, the water

sources of the aerosols were disinfected. Disinfection may

select for the predominance and growth of mycobacteria.

Therefore, mycobacteria may be responsible, in part, for

many outbreaks of hypersensitivity pneumonitis and other

respiratory problems in the workplace and home.

ypersensitivity pneumonitis is an occupational hazard

of workers in two different industries, automobile

manufacturing (e.g., metal working) and leisure (e.g.,

indoor swimming pools). Pulmonary illness and infection

have also been a consequence of exposure to aerosols gen-

erated by hot tubs, spas, and coolant baths. Respiratory

problems have also been associated with exposure to

water-damaged buildings during reconstruction, and

mycobacteria isolated from materials from such buildings

have been shown to provoke inflammatory reactions. The

outbreaks share the common feature of aerosol exposure

and respiratory illness. I propose that exposure to aerosols

containing mycobacteria is a common feature of the out-

breaks and that mycobacteria or their products could be

responsible for the respiratory symptoms.

Epidemiologic studies have established that the work-

ers in such outbreaks were exposed to aerosols generated

in the workplace from water that was either a work tool

(e.g., metalworking fluid) or an integral part of the work-

place or household (e.g., swimming pools and hot tubs)

(1–7). Outbreaks of respiratory disease occurred in spite of

disinfectant treatment of the waters or fluids to reduce the

number of microorganisms. Living or working in water-

damaged buildings or as a consequence of reconstruction

of water-damaged buildings has also been associated with

outbreaks of respiratory problems (8,9). Respiratory dis-

ease has been associated with mycobacteria in reservoirs,

aerosols, or structural material in a number of cases

(2,3,6,7,9).

Hypersensitivity Pneumonitis in Workers

Exposed to Metalworking Fluid

An estimated 1.2 million workers in the United States

are exposed to aerosols generated by metal grinding (10).

Metalworking fluids are widely used in a variety of com-

mon industrial metal-grinding operations to lubricate and

cool the tool and the working surface. Metalworking fluids

are oil-water emulsions that contain paraffins, pine oils,

polycyclic aromatic hydrocarbons, and heavy metals

(10,11). Exposure to metalworking fluid aerosols can lead

to hypersensitivity pneumonitis and chronic obstructive

pulmonary disease (1,6,12–14). Mycobacteria were recov-

ered significantly more frequently from metalworking

fluid samples collected from facilities where hypersensi-

tivity pneumonitis was found; compared to facilities that

did not have hypersensitivity pneumonitis (6). In one

study, exposure to metalworking fluid mist resulted in

hypersensitivity pneumonitis in 10 workers (7). Acid-fast

microorganisms identified as mycobacteria were present in

the reservoir at 10

CFU/mL (7). A mycobacteria in the

reservoir was considered to be a likely cause of the hyper-

sensitivity pneumonitis because one patient was infected

by a Mycobacterium sp. and had antibodies against the

reservoir fluid (7).

Hypersensitivity pneumonitis appeared in spite of dis-

infection of the metalworking fluid with morpholine,

formaldehyde, or quaternary ammonium-based disinfec-

tants (1,6,12,13), and mycobacteria were recovered from

the metal working fluid (6,14,15). Mycobacteria are resist-

ant to formaldehyde and quaternary ammonium disinfec-

tants (16) and the heavy metals in metalworking fluids

(17). Further, mycobacteria can grow on the organic com-

pounds in metalworking fluid, including the paraffins, pine

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 763

PERSPECTIVES

Mycobacterial Aerosols and

Respiratory Disease

Joseph O. Falkinham, III*

*Virginia Polytechnic Institute and State University, Blacksburg,

Virginia, USA

oils, and polycyclic aromatic hydrocarbons (18,19) and

can degrade the disinfectant morpholine (20).

Mycobacteria present in the water (21) can likely grow on

the organic compounds in metalworking fluids in the

absence of competitors after disinfection. Cleaning would

not be expected to eradicate mycobacteria because of their

ability to form biofilms (21,22). Adding disinfectant and

cleaning the reservoir in one facility did not prevent the

reappearance of mycobacteria (7 x 10

CFU/mL by 2

weeks [7]). Further, disinfectant treatment would likely

result in selection of mycobacteria remaining after the

cleaning.

Hypersensitivity Pneumonitis in

Swimming Pool Attendants

Granulomatous pneumonitis has been reported in life-

guards (“lifeguard lung”) who worked at an indoor swim-

ming pool that featured waterfalls and sprays (5). Affected

lifeguards with symptoms worked longer hours than unaf-

fected lifeguards (5), which demonstrated a dose-response

effect. The waterfalls and sprays increased the number of

respirable particles fivefold and the levels of endotoxin

eightfold (5). Based on the presence of endotoxin in the

aerosol samples, endotoxin exposure was suggested as the

cause of the pneumonitis in lifeguards (5). However, sub-

sequent data provided evidence of a possible second factor

resulting in hypersensitivity pneumonitis; aerosols con-

taining mycobacteria were shown to cause granulomatous

lung disease (4). Others have reported high numbers of

mycobacteria in swimming pools and whirlpools (23) and

in hot tubs (2,3,24). Further, amoebae were reported in the

indoor swimming pool where lifeguards reported pneu-

monitis (5). Mycobacteria, including M. avium and M.

intracellulare, can survive and grow in phagocytic amoe-

bae (25) and protozoa (26). In fact, M. avium grown in

amoebae or protozoa are more virulent (25; Falkinham JO,

unpub. data). Mycobacteria are resistant to chlorine (27)

and preferentially aerosolized from water (28).

Mycobacterial Disease after Exposure to

Aerosols Generated by Hot Tubs

Hypersensitivity pneumonitis and mycobacterial pul-

monary disease has been reported after exposure to hot

tubs (2,3,24). The mycobacteria isolated (e.g., M. avium)

were likely responsible for the infections based on the

identity of patient and hot tub mycobacterial isolates by

either restriction fragment length polymorphism analysis

(24) or multilocus enzyme electrophoresis (2,3). Further,

exposure was followed closely by the onset of symptoms,

and the extent of symptoms was related to the length of

exposure (i.e., time spent in the hot tub) (2,24). Although

these reports do not document the use of disinfectants in

the hot tubs, the waters had been heated. Mycobacteria are

relatively resistant to high temperature (29) and concen-

trated in hospital hot water systems (30).

Hypersensitivity Pneumonitis in

Occupants of Water-Damaged Buildings

Inflammatory reactions—including eye irritation, res-

piratory infections, wheeze, bronchitis, and asthma—in

workers in water-damaged or “moldy” buildings have been

associated with the presence of high numbers of microor-

ganisms (8). Mycobacteria were recovered from materials

collected from water-damaged buildings, as well as from

microorganisms normally associated with building materi-

als (9). During reconstruction, those mycobacteria could

be aerosolized in the dust. Although other microorganisms

could be responsible for the respiratory problems, both

saprophytic (e.g., M. terrae) and pathogenic (e.g., M.

avium) strains isolated from moldy buildings were capable

of inducing inflammatory responses in a mouse

macrophage cell line (31). The mycobacteria elicited dose-

dependent production of cytokines interleukin-6 and tumor

necrosis factor-α, nitric oxide, and reactive oxygen species

from the murine macrophage (31). Because whole

mycobacterial cells were used in the assays (31), whether

cell metabolites, which are likely easily aerosolized, were

responsible for the induction of inflammatory reactions is

not known. Heat-shock proteins from a number of

mycobacterial species have been shown to generate Th1-

type responses, airway inflammation, and airway hyperre-

sponsiveness (32). This evidence suggests that mycobacte-

ria or their metabolites are possible causes of respiratory

disease in persons exposed to water-damaged buildings.

Ecology of Mycobacteria

The unique combination of physiologic characteristics

that distinguish the environmental opportunistic mycobac-

teria make them likely agents for causing respiratory dis-

ease in these diverse settings. Mycobacteria are found in a

great variety of natural and human-influenced aquatic

environments, including treated drinking water (21) and

aerosols (33). Mycobacteria in drinking water are associat-

ed with the presence of particulates (21). Although these

microbes are grown in rich media in the laboratory, they

are oligotrophic and capable of substantial growth in low

concentrations of organic matter. For example, M. avium

and M. intracellulare can grow in natural and drinking

water over a temperature range of 10°C to 45°C (34).

Mycobacteria are relatively resistant to high temperatures.

For example, 10% of cells of a strain of M. avium survived

after 1 h at 55°C (29). Mycobacteria are slow growing as a

consequence of their fatty acid- and wax-rich impermeable

cell wall (35). The resulting cell surface hydrophobicity

permits adherence to solid substrates (e.g., pipes and

leaves) in aquatic environments, which results in

764 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

PERSPECTIVES

mycobacteria’s persistence and resistance to being washed

away at high flow rates (21,22). Further, hydrophobicity is

undoubtedly associated with the ability of these bacteria to

metabolize a wide variety of nonpolar organic compounds

(18–20) that are constituents of metal working fluids

(15,16).

Resistance of Mycobacteria to Disinfection

Mycobacteria are very resistant to the disinfectants

used in water treatment, including chlorine and ozone (27).

For example, M. avium is almost 500 times more resistant

to chlorine than is Escherichia coli (27). Mycobacteria are

also quite resistant to agents used for surface and instru-

ment disinfection, including quaternary ammonium com-

pounds, phenolics, iodophors, and glutaraldehyde

(16,22,23,36) and can degrade the disinfectant morpholine

(20). Hydrophobicity and impermeability are undoubtedly

factors contributing to the disinfection resistance of

mycobacteria (35). Chemical or enzymatic removal of sur-

face lipid, while not reducing viability, reduces surface

hydrophobicity and alters cell charge (37). Because of

their inherent impermeability, mycobacteria grow relative-

ly slowly compared to other bacteria. The slow growth is

not necessarily a disadvantage because it correlates with

increased resistance to antimicrobial agents (35), including

chlorine (Falkinham JO, unpub data).

Exposure of a mixed microbial population to disinfec-

tants results in selection of a disinfectant-resistant or toler-

ant population (38). The persistence and growth of

mycobacteria in drinking water systems (21) are due, in

part, to their disinfectant-resistance (27) and ability to

grow under oligotrophic conditions (21). Disinfection of

swimming pools, therapy pools, and spas or hot tubs with

chlorine is expected to kill nonmycobacterial flora and to

permit the growth of even the slowly growing mycobacte-

ria in the absence of competitors for nutrients. High tem-

perature would also be expected to result in enrichment of

mycobacteria (29,30). Resistance to disinfectants could

also lead to the proliferation of mycobacterial populations

in metal working fluid and coolants after disinfection

(6,12,13).

Aerosolization of Mycobacteria

Although M. tuberculosis is transmitted between

patients through aerosols, little information exists on

aerosolization of the environmental opportunistic

mycobacteria (e.g., M. avium and M. intracellulare).

Patient-to-patient transmission of environmental oppor-

tunistic mycobacteria does not occur (39). M. avium and M.

intracellulare are readily aerosolized from aqueous suspen-

sion (28,33). Transfer of mycobacteria occurs as a result of

binding of mycobacterial cells to air bubbles and ejection

of water droplets after the air bubbles reach the liquid sur-

face (28). Aerosolization can result in >1,000-fold increase

in numbers of viable mycobacterial cells per milliliter of

water droplets ejected from water (28). Mycobacteria in

natural aerosols are found in particles and droplets (i.e., <5

µm) that can enter the alveoli of the human lung (28,33).

Cell surface hydrophobicity, not surface charge, is a major

determinant of enrichment in ejected droplets (28).

Transfer of mycobacteria from water to air is subject to

prevailing physiochemical conditions and can be manipu-

lated. Salts (e.g., NaCl) or detergents reduce the rate of

transfer of mycobacteria from water to air by ejected

droplets (28). The influence of the components of metal-

working fluid or of chlorine or other disinfectants in water

upon aerosolization mycobacteria is unknown.

Mycobacteria and Immune Responses

and Airway Inflammation

Mycobacterial cells and cellular components provoke

inflammatory responses. Cells of mycobacterial strains

isolated from material collected from water-damaged

buildings provoke inflammatory responses in

macrophages (31). Mycobacterial heat-shock proteins gen-

erate Th1-type responses, airway inflammation, and hyper-

responsiveness (32). The mycolic acid-containing glycol-

ipids, mannose-containing phospholipids, glycopepti-

dolipid mycosides, phenolglycolipid mycosides, and sul-

fatides that are unique to mycobacteria have all been

reported to stimulate immune responses in animals (40).

Further, mycobacteria produce a variety of extracellular

primary and secondary metabolites (19) that could be

aerosolized and trigger immune responses, including

hypersensitivity pneumonitis. Some of these immunostim-

ulatory compounds are produced in response to growth on

polycyclic aromatic hydrocarbons (18). Unfortunately, the

studies of inflammatory responses provoked by mycobac-

teria have been limited to whole cells grown under a single

condition (31) or single proteins (32). The influence of

growth conditions (e.g., growth in metalworking fluid or

chlorinated water) or cell fractions (e.g., membranes) or

metabolites to stimulate inflammatory responses has not

been measured.

Conclusion

Contemporary reviews of airway dysfunction all

describe the need for information concerning microbial

agents of workplace and household exposure (41).

Although many more studies are needed, the evidence

points to a role of environmental opportunistic mycobacte-

ria in provoking hypersensitivity pneumonitis, respiratory

disease, and respiratory infection in both the workplace

and home. In addition to the recovery of identical species

and types of mycobacteria from reservoirs and patients,

physiologic characteristics of mycobacteria are consistent

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 765

PERSPECTIVES

with their presence in the sources, transmission by means

of aerosols, and illnesses. Identifying the factors that influ-

ence the presence of mycobacteria in aerosols in these

workplaces would have an impact on workers in a variety

of occupational settings.

On the basis of several physiologic and ecologic char-

acteristics of mycobacteria, several approaches to reduce

the impact of mycobacteria in these settings are possible.

Because mycobacteria are associated with particulates

(21), their numbers in reservoirs can be reduced by

removal of particular matter (e.g., filtration). UV light can

be used to reduce mycobacterial numbers. Disinfection of

mycobacteria at high temperatures (e.g., 40°C) is more

effective at reducing numbers, especially if cells were

grown at lower temperatures (e.g., 30°C). Agents or com-

binations with surfactant or detergent-like and disinfectant

activity would increase permeation in cells and biofilms

and kill more mycobacteria. Finally, aerosolized or water-

borne mycobacteria may be trapped in filters coated with

hydrophobic compounds (e.g., paraffin) and thereby inter-

cepted before inhalation or ingestion.

Dr. Falkinham is a professor of microbiology in the

Department of Biology at Virginia Polytechnic Institute and State

University. His research interests include identifying the genes

and physiologic characteristics of Mycobacterium avium that are

responsible for its ecology, transmission, and virulence.

References

1. Bernstein DI, Lummus ZL, Santilli G, Siskosky J, Bernstein IL.

Machine operator’s lung. A hypersensitivity pneumonitis disorder

associated with exposure to metalworking fluid aerosols. Chest

1995;108:636–41.

2. Embil J, Warren P, Yakrus M, Stark R, Corne S, Forrest D. Pulmonary

illness associated with exposure to Mycobacterium avium complex in

hot tub water: hypersensitivity pneumonitis or infection? Chest

1997;111:813–6.

3. Kahana LM, Kay JM, Yakrus M, Waserman S. Mycobacterium avium

complex infection in an immunocompetent young adult related to hot

tub exposure. Chest 1997;111:242–5.

4. Martyny JW, Rose CS. Nontuberculous mycobacterial bioaerosols

from indoor warm water sources cause granulomatous lung disease.

Indoor Air 1999;9:1–6.

5. Rose CS, Martyny JW, Newman LS, Milton DK, King TE Jr, Beebe

JL, et al. “Lifeguard lung”: endemic granulomatous pneumonitis in

an indoor swimming pool. Am J Public Health 1998;88:1795–800.

6. Shelton BG, Flanders WD, Morris GK. Mycobacterium sp. as a pos-

sible cause of hypersensitivity pneumonitis in machine workers.

Emerg Infect Dis 1999;5:270–3.

7. Muilenberg ML, Burge HT, Sweet T. Hypersensitivity pneumonitis

and exposure to acid-fast bacilli in coolant aerosols. J Allergy Clin

Immunol 1998;91:311.

8. Brunekreef B. Damp housing and adult respiratory symptoms.

Allergy 1992;47:498–502.

9. Andersson MA, Nikulin M, Koljalg U, Andersson MC, Rainey F,

Reijula K, et al. Bacteria, molds, and toxins in water-damaged build-

ing materials. Appl Environ Microbiol 1997;63:387–92.

10. Eisen EA, Tolbert PE, Smith TJ, Monson RR, Hallock M, Woskie SR,

et al. Mortality studies of machining fluids; an exposure-response

analysis of respiratory and digestive cancers. Proceedings of the 9th

International Symposium on Epidemiology in Occupational Health,

Sept 23-25, 1992, Cincinnati, Ohio. DHHS (NIOSH) pub no. 94-112.

Cincinnati (OH): National Institute for Occupational Safety and

Health; 1994. p. 113–7.

11. Howell JK, Lucke WE, Steigerwald JC. Metalworking fluids: compo-

sition and use. The Industrial Metalworking Environment:

Assessment and Control (Symposium). Nov. 13–16, 1995. Detroit:

Automobile Manufacturers Association; 1996. p. 13–20.

12. Centers for Disease Control and Prevention. Biopsy-confirmed

hypersensitivity pneumonitis in automobile production workers

exposed to metalworking fluids—Michigan, 1994–1995. MMWR

Morb Mortal Wkly Rep 1996;45:606–10.

13. Centers for Disease Control and Prevention. Respiratory illness in

workers exposed to metalworking fluid contaminated with nontuber-

culous mycobacteria—Ohio, 2001. MMWR Morb Mortal Wkly Rep

2002;51:349–52.

14. Wilson, RW, Steingrube VA, Bottger EC, Springer B, Brown-Elliot

BA, Vincent V, et al. Mycobacterium immunogenum sp. nov., a novel

species related to Mycobacterium abscessus and associated with clin-

ical disease, pseudo-outbreaks and contaminated metalworking flu-

ids: an international cooperative study on mycobacterial taxonomy.

Int J Syst Evol Microbiol 2001;51:1751–64.

15. Moore JS, Christensen M, Wilson RW, Wallace RJ Jr, Zhang, Y, Nash

DR, et al. Mycobacterial contamination of metal working fluids:

involvement of a possible new taxon of rapidly growing mycobacte-

ria. American Industrial Hygiene Association Journal

2000;61:205–13.

16. van Klingeren B, Pullen W. Comparative testing of disinfectants

against Mycobacterium tuberculosis and Mycobacterium terrae in a

quantitative suspension test. J Hosp Infect 1987;10:292–8.

17. Falkinham JO III, George KL, Parker BC, Gruft H. In vitro suscepti-

bility of human and environmental isolates of Mycobacterium avium,

M. intracellulare, and M. scrofulaceum to heavy-metal salts and

oxyanions. Antimicrob Agents Chemother 1984;25:137–9.

18. Guerin WF, Jones GE. Mineralization of phenanthrene by a

Mycobacterium sp. Appl Environ Microbiol 1988;54:937–44.

19. Krulwich TA, Pelliccione NJ. Catabolic pathways of coryneforms,

nocardias, and mycobacteria. Annu Rev Microbiol 1979;33:95–111.

20. Combourieu B, Besse P, Sancelme M, Veschambre H, Delort AM,

Poupin P, et al. Morpholine degradation pathway of Mycobacterium

aurum MO1: direct evidence of intermediates by in situ

H nuclear

magnetic resonance. Appl Environ Microbiol 1998;64:153–8.

21. Falkinham JO III, Norton CD, Le Chevallier MW. Factors influenc-

ing numbers of Mycobacterium avium, Mycobacterium intracellu-

lare, and other mycobacteria in drinking water distribution systems.

Appl Environ Microbiol 2001;67:1225–31.

22. Schulze-Röbbecke R, Fischeder R. Mycobacteria in biofilms.

Zentralblatt fur Hygiene und Unweltmedizin 1989;188:385–90.

23. Havelaar AH, Berwald LG, Groothuis DG, Baas JG. Mycobacteria in

semi-public swimming pools and whirlpools. Zentralblatt fur

Hygiene und Unweltmedizin 1985;180:505–14.

24. Mangione EJ, Huitt G, Lenaway D, Beebe J, Bailey A, Figoski M, et

al. Nontuberculous mycobacterial disease following hot tub expo-

sure. Emerg Infect Dis 2001;7:1039–42.

25. Cirillo JD, Falkow S, Tompkins LS, Bermudez LE. Interaction of

Mycobacterium avium with environmental amoebae enhances viru-

lence. Infect Immun 1997;65:3759–67.

26. Strahl ED, Gillaspy GE, Falkinham JO III. Fluorescent acid-fast

microscopy for measuring phagocytosis of Mycobacterium avium,

Mycobacterium intracellulare, and Mycobacterium scrofulaceum by

Tetrahymena pyriformis and their intracellular growth. Appl Environ

Microbiol 2001;67:4432–9.

766 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

PERSPECTIVES

27. Taylor RH, Falkinham JO III, Norton CD, LeChevallier MW.

Chlorine, chloramine, chlorine dioxide, and ozone susceptibility of

Mycobacterium avium. Appl Environ Microbiol 2000;66:1702–5.

28. Parker BC, Ford MA, Gruft H, Falkinham JO III. Epidemiology of

infection by nontuberculous mycobacteria. IV. Preferential

aerosolization of Mycobacterium intracellulare from natural water.

Am Rev Respir Dis 1983;128:652–6.

29. Schulze-Röbbecke R, Buchholtz K. Heat susceptibility of aquatic

mycobacteria. Appl Environ Microbiol 1992;58:1869–73.

30. duMoulin GC, Stottmeier KD, Pelletier PA, Tsang AY, Hedley-Whyte

J. Concentration of Mycobacterium avium by hospital hot water sys-

tems. JAMA 1988;260:1599–601.

31. Hüttunen K, Ruotsalainen M, Iivanainen E, Torkko P, Katila M-L,

Hirvonen M-R. Inflammatory responses in RAW264.7 macrophages

caused by mycobacteria isolated from moldy houses. Environ Toxicol

Pharmacol 2000;8:237–44.

32. Rha Y-H, Taube C, Haczku A, Joetham A, Takeda K, Duez C, et al.

Effect of microbial heat shock proteins on airway inflammation and

hyperresponsiveness. J Immunol 2002;169:5300–7.

33. Wendt SL, George KL, Parker BC, Gruft H, Falkinham JO III.

Epidemiology of infection by nontuberculous mycobacteria. III.

Isolation of potentially pathogenic mycobacteria in aerosols. Am Rev

Respir Dis 1980;122:259–63.

34. George KL, Parker BC, Gruft H, Falkinham JO III. Epidemiology of

infection by nontuberculous mycobacteria. II. Growth and survival in

natural waters. Am Rev Respir Dis 1980;122:89–94.

35. Brennan PJ, Nikaido H. The envelope of mycobacteria. Annu Rev

Biochem 1995;64:29–63.

36. Collins FM. Bactericidal activity of alkaline glutaraldehyde solution

against a number of atypical mycobacterial species. J Appl Bacteriol

1986;61:247–51.

37. George KL, Pringle AT, Falkinham JO III. The cell surface of

Mycobacterium avium – intracellulare and M. scrofulaceum: effect of

specific chemical modifications on cell surface charge. Microbios

1986;45:199–207.

38. Young H-K. Antimicrobial resistance spread in aquatic environments.

J Antimicrob Chemother 1993;31:627–35.

39. Wolinsky E. Nontuberculous mycobacteria and associated diseases.

Am Rev Resp Dis 1979;119:107–59.

40. Clark-Curtiss JE. Identification of virulence determinants in patho-

genic mycobacteria. Curr Top Microbiol Immunol 1998;225:57–121.

41. Speizer FE. Occupational and environmental lung diseases: an

overview. Environ Health Perspect 2000;108(Suppl 4):603–4.

Address for correspondence: J.O. Falkinham, Fralin Biotechnology

Center, West Campus Drive, Virginia Tech, Blacksburg, VA 24061-0346,

USA; fax: (540) 231-7126; email:[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 767

PERSPECTIVES

Search ppast iissues oof EEID aat wwww.cdc.gov/eid

We propose a system for continuing surveillance of

viral pathogens circulating in large human populations. We

base this system on the physical isolation of viruses from

large pooled samples of human serum and plasma (e.g.,

discarded specimens from diagnostic laboratories), fol-

lowed by shotgun sequencing of the resulting genomes.

The technology for concentrating virions from 100-L vol-

umes was developed previously at Oak Ridge National

Laboratory, and the means for purifying and concentrating

virions from volumes in microliters have been developed

recently. At the same time, marine virologists have devel-

oped efficient methods for concentrating, amplifying, and

sequencing complex viral mixtures obtained from the

ocean. Given this existing technology base, we believe an

integrated, automated, and contained system for surveil-

lance of the human “virome” can be implemented within 1

to 2 years. Such a system could monitor the levels of

known viruses in human populations, rapidly detect out-

breaks, and systematically discover novel or variant human

viruses.

he traditional process of discovering previously

unknown human viruses, or variants of known viruses,

is neither rapid nor thoroughly systematic. The time

between back-calculated initial infection and final identifi-

cation is often many weeks, months, or even years. For a

totally new agent, the estimated interval between initial

infection and detailed characterization is variable and

depends on the presence of unusual symptoms, the failure

to identify a virus after using all available specific tests,

the recognition of a unique problem, and, in the past, the

ability to grow the agent in culture.

The idiosyncratic nature of virus discovery contrasts

with the broad survey approaches characteristic of

genomics and proteomics. Only in the relatively small

field of ocean viruses has a more inclusive, cataloging

approach been tested. Facilitated by the relative ease with

which viruses can be isolated from seawater (using com-

mercial filters), investigators in this area have examined a

broad and essentially unbiased population of viral agents at

the genome sequence level (including phage) and estimat-

ed the number of different genomes present (~5,000)

(1–3). One would expect that a comprehensive survey of

human viruses, defining what we might term the human

“virome” would be, at least conceptually, even more

straightforward.

Our proposed approach (Figure), in which large popu-

lations are continually monitored for new human-infective

768 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

PERSPECTIVES

Global Screening for

Human Viral Pathogens

Norman G. Anderson,* John L. Gerin,† and N. Leigh Anderson‡

*Viral Defense Foundation, Kensington, Maryland, USA;

†Georgetown University Medical Center, Rockville, Maryland,

USA; and ‡Plasma Proteome Institute, Washington, D.C., USA

Figure. Schematic representation of a process for systematic dis-

covery of human viruses. The basic process (left vertical series of

steps) depends on physical isolation and shotgun sequencing to

obtain sequences of frequent and rare viruses. A series of addi-

tional steps (right box) can be added to deplete known viruses at

two levels, thereby enhancing sensitivity for novel agents.

viruses, has not been considered technically feasible or

medically necessary in the past. For purposes of broad sur-

veillance, we propose using pools of serum or plasma from

large numbers of persons, the most likely source of which

is excess material collected for routine clinical purposes.

These samples would be pooled and processed by using

available technology to isolate virus particles en masse,

recover viral nucleic acids, produce amplified shotgun

libraries, carry out shotgun sequencing of the mixture of

viral genomes, and reconstruct these genomes in silico

with the techniques originally developed to sequence the

entire human genome from random fragments. A central

objective is to continually repeat this monitoring process to

determine which agents change in abundance over time,

find undiscovered agents already present, and detect new

viruses when they appear. If successful, we will have for

the first time a comprehensive picture of “what is going

around.” Surprisingly, most of the systems and technology

to carry out this process exist in a basic form and have

been successfully employed to survey the extremely varied

DNA virus population of the oceans. What remains to be

done, to create a system applicable to humans, is primari-

ly its integration, optimization, and implementation in a

safely contained environment. We briefly explore the com-

ponents of this process here and suggest that it can be

made operational in less than a year.

Availability of Large Pooled Samples

The major commercial diagnostic laboratories in the

United States discard approximately 500 L of excess

human serum or plasma each week. This material repre-

sents a broad cross-section of patients and illnesses.

Plasma viral loads as a function of time after onset of ill-

ness are not known for most viral diseases, but they appear

to be highest in the initial febrile stages. Since one of the

first steps in treating a febrile illness of unknown origin is

obtaining a blood sample, we expect that current diagnos-

tic networks contain appreciable quantities of virus.

Samples from subpopulations enriched for potential viral

illness could also be selected. For those viral diseases in

which viremia precedes major illness, the inclusion of

large numbers of randomly acquired specimens in the pool

(i.e., an unselected pool) offers the best chance of detec-

tion. Analysis of pooled samples from a large number of

persons should raise minimal privacy concerns.

Virus Isolation, Sequencing, and Assembly

Methods are required for the routine isolation of all

classes of viruses from a pooled sample and for concentrat-

ing them by factors of over a million while ensuring that

all nonviral nucleic acids have been removed. The concen-

trates may be dangerously infectious, and sophisticated

containment systems will be needed.

The preferred technology for virus concentration from

large volumes was developed by the Joint NIH-AEC Zonal

Centrifuge Development Project at the Oak Ridge

National Laboratory, Oak Ridge, Tennessee, USA, and the

Oak Ridge Gaseous Diffusion Plant in the 1960s (4). At the

outset, researchers wanted to determine whether viruses as

a class differed in a systematic way from all other small

particles in nature. When the sedimentation coefficients of

then-known viruses were plotted against their isopycnic

banding densities, nearly all viruses fell into an otherwise

essentially vacant area in the center of the plot, surround-

ed at higher or lower density and higher or lower sedimen-

tation coefficients by various subcellular organelles and

macromolecules. This area was termed “the virus window”

(5). Thus, viruses exhibit a unique size and density range

and have banding densities that reflect their combined pro-

tein and nucleic acid contents. In addition, viral nucleic

acids are shielded from attack by nucleases so that contam-

inating nucleic acid-containing particles (primarily

genomic DNA from apoptotic or disrupted cells) can be

selectively destroyed by added nucleases (6). The rules of

virus isolation are the following: 1) the sedimentation rate

(based largely on particle size) falls in a specific range; 2)

the banding density in a gradient falls in a specific range;

3) the genome is protected from nuclease attack until the

protein (+/- lipid) coat is disrupted; and 4) the major pro-

teins present have sequences that agree with at least part of

the genome.

Exploiting the virus window required a two-dimension-

al separation based on sedimentation rate (S) in one dimen-

sion and banding density (rho) in the other, usually carried

out in the order: S-ρ. For large-scale isolation and purifica-

tion, the challenge was to perform these separations con-

tinuously and simultaneously in large continuous-flow

centrifuge rotors spinning at high speed in a vacuum.

In this scheme, a flowing stream passes inboard of a

thin, nonflowing density gradient, held in place against the

rotor wall by centrifugal force. The recovered virus forms

a narrow band in the gradient, which is recovered after

reorienting the gradient to rest at the end of the run. If the

flow through two centrifuges is cascaded, the first operat-

ing at lower speed than the second, particles having a high-

er S rate than viruses could be removed from the flowing

stream, and the viruses then concentrated and banded in

the second higher speed centrifuge, thus providing a large-

volume S-ρ separation.

The end result of this work included the design and

construction of the K-II large-scale ultracentrifuge (7–10).

This device was designed to recover virus in a high state of

purity from 100-L batches of crude influenza vaccine in an

8-hour day (11,12) and subsequently was used for the

large-scale purification of the hepatitis B (Australia) sur-

face antigen (13,14) from human serum for use as a vac-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 769

PERSPECTIVES

cine and for mass isolation of polyhedral inclusion bodies

(15). K centrifuges have come into worldwide use for

large-scale virus isolation and have been commercially

available with little rotor modification for the last 35 years

(16,17). Approximately 200 such systems have been con-

structed.

Methods are now available to further purify compo-

nents of complex viral mixtures, to sediment the viruses

through gradients containing nonsedimenting zones (e.g.,

of nucleases), and to concentrate them down to tens of

microliters (18,19) for sequencing or mass spectrometric

analysis.

Ocean Viruses

During the Oak Ridge centrifuge development project,

large volumes of test material were required, and seawater,

among other sources, was examined as a possible source of

virus. The ocean was found to contain examples of almost

every known viral form at high titers (20), initiating the

exploration of marine virology (1–3,21–30). Recent data

in this field suggest that the oceans of the world contains

approximately 10

phage particles or virions (27) (c. 22

million metric tons), much of it turning over once per day

and including some human pathogens (28). This vast

mutation engine, even if one assumes a minimal mutation

rate, generates the equivalent of hundreds of new complete

human genomes per day. That viruses are ubiquitous in the

ocean has been demonstrated by studies on samples recov-

ered in widely separated locations from filtration systems

installed on surface ships (29), nuclear submarines (3), and

remotely operated vehicles (30). Indeed, the entire ocean

has an average viral content in the lower range of the viral

loads reported for human plasma from viremic patients.

Marine virologists have in fact come closest to implement-

ing a surveillance system such as we propose for humans.

On limited budgets, these researchers have developed the

means of recovering marine viruses from large volumes by

filtration (especially well-suited to such a dilute sample),

and for producing shotgun libraries from them by random

amplification (1). Marine virologists have also begun to

estimate the diversity of marine viruses (1–3,27,30) and

are reconstructing large numbers of complete viral

genomes. In one study (2), a 200-L sample of surface sea-

water was concentrated; ~2 x 10

viral particles were

recovered; the DNA was randomly sheared and cloned;

and 1,934 fragments were sequenced. Data analysis

showed that most of the sequences were from previously

unknown viruses. Approximately 3.5% of the total

sequence samples overlapped, suggesting that the marine

viral community was highly diverse. A unique mathemati-

cal analysis (2) further suggested that less than 10

differ-

ent viral types were present and that shotgun sequencing of

the most abundant could be done, with existing facilities,

in 1 month. Although efforts to date have focused on virus-

es with DNA genomes, most human viral pathogens have

RNA genomes. Genomic sequencing libraries will there-

fore have to be prepared from mixtures of both single- and

double-stranded DNA and RNA viruses (the latter generat-

ed by reverse transcription).

The feasibility of the genomics assembly and annota-

tion components of this project derives from the demon-

stration that the entire human genome could be fragment-

ed, the fragments sequenced, and the original sequence

reconstructed from overlaps; that abundant sequencing

capacity now exists in search of high-value projects; and

that marine virologists have succeeded in parallel ventures.

The challenge is to shorten the time of the entire process so

useful epidemiologic and viroterrorism response data can

be rapidly obtained.

Remaining Challenges

Recovering viruses from large pools of human serum

and plasma and routinely cloning and sequencing the viral

nucleic acids (using established shotgun approaches)

appears technically feasible. Consequently, the titers of

many known human viral pathogens may be estimated rou-

tinely, and new viruses (both pathogenic and nonpathogen-

ic) may be discovered systematically.

The initial choice is between filtration and centrifuga-

tion. Seawater contains little contaminating material of the

size and density of the virus particles, filtration is simple

and efficient, and no free nucleic acids have been report-

ed. Plasma and serum present different problems, compli-

cated by the presence of large amounts of protein, some

nonviral particles in the virus window, and variable

amounts of soluble nonviral nucleic acids (31) that must

be eliminated. An advantage of centrifugal methods is that

all separations, down to banding in microliter gradients,

can be (and have been) done with the virions in suspen-

sion, a process that avoids aggregation that may occur on

filter surfaces.

Several key questions remain to be addressed in a prac-

tical project: 1) Can this process be carried out rapidly

enough to support a timely therapeutic or prophylactic

response to a new agent (natural or engineered)? 2) Will

the novel virions that originated from one or a very few

infected persons be recovered and detected? 3) Can the

affected persons be located? 4) Will the prescence of anti-

bodies in the starting samples against most known viruses

affect the separations?

Speed of Operation

Samples can be collected weekly, and materials rapidly

transported to one or more processing sites. Virus isola-

tion, library construction, and preparation of clones for

sequencing require <7 days but the time may be com-

770 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

PERSPECTIVES

pressed into <4 days with a 24-hour per day operation.

Sequencing time will be determined by available capacity,

but since capacity is abundant, with large increases in

sight, extensive library sequencing (e.g., 10–100 megabas-

es) could be carried out in 2 to 3 days. Less than 1 day

would be required to assemble viral genomes as contigs in

silico. By the conducting of each step in sequence, turn-

around (serum pool to raw sequence data) could be com-

pleted in approximately 10 operating days. Additional time

would be required for bioinformatics analysis of the data

and annotation, but prevalence and novelty conclusions

should be available almost immediately. These estimates

assume an integrated and fully developed system in contin-

uous operation, analogous in some respects to those moni-

toring computer viruses.

Sensitivity

The mass of virus ultimately recovered from pooled

human plasma is difficult to estimate in advance. If the

average virus has a mass of 1.0 x 10

-15

g; if the average titer

of an infected person is 10

virions/mL, and if 0.1% of the

samples are from viremic patients, then ~0.5 µg of virus

would be recovered from each 500-L pool, substantially

more than the few nanograms required to make a large

library with current technology (1,2). If the average sam-

ple contributed to the pool was 1 mL, and if the final con-

centrated virus were in 1 mL, the final concentration of a

totally new virus would be close to that in the original indi-

vidual sample. The possibility of detecting all viruses for

which polymerase chain reactions (PCR) primers are

available, down to contributions from single patients,

therefore exists.

Dynamic Range

The problem of dynamic range can be addressed in

three ways. First, given large sequencing capacity, one

could sequence deeply into the libraries (millions of clones

instead of a few thousand), thereby detecting parts-per-

million sequences. Second, one could apply antibody-

based affinity methods to deplete known viral particles

from the initial concentrated viral sample. Third, one could

use subtractive hybridization to remove known viral

genomic sequences to further enrich libraries in novel

genomes. The last two approaches can be progressively

extended as viruses are characterized to provide a continu-

ous increase in sensitivity to new agents (Figure).

Identification of Viral Sources

Two approaches can be used, if necessary, to link virus-

es to patients. In the first approach, viruses would be

tracked geographically, first in terms of large regions, and

then, sequentially, in terms of smaller areas. Detecting a

new agent in large pooled samples would thus be repeated

in smaller, localized pools that had been combined hierar-

chically to generate the larger pool (32).

A potentially more efficient approach involves overlap-

ping subpools designed such that a new viral sequence can

be assayed (e.g., by PCR) in the subpools and the affected

persons identified in one step. To achieve this result, each

sample is added to a series of different pools, the identity

of these subpools providing an “address” of the sample

(33). This process can be visualized by analogy to a 3-D

chessboard, where each position represents a sample, and

the subpools are the various planes parallel to the top,

front, and side: each sample would contribute to three sub-

pools. In practice, additional pools would be created to

provide a relatively unambiguous means of backtracking

from the pattern of subpools positive for a specific

sequence to one or a few persons.

Viral Pathogens That May Be Missed

Not all human viral pathogens will be detected easily

by analyzing plasma or serum samples. Neurotropic virus-

es such as rabies, for example, are found in cells and tis-

sues and do not appear free in serum or plasma in appre-

ciable amounts. Thus, these viruses would escape the

screening system described to this point. Although using

rabies for viroterrorism would be unlikely, such viruses are

of great public health interest, and efforts should ultimate-

ly be made to include them in any global screening system.

The rapid turnover of viruses found in plasma suggests

that they are removed into cells, and that appears to be gen-

erally true. Centrifugal S-ρ technology was originally

developed for cell fractionation with the aim of isolating

viruses from tumors, cells, and tissues (5). Trace quantities

of virus could be added to tissue homogenates and recov-

ered in a high state of purity. The basic technology there-

fore exists for isolating viruses from lymphocytes and a

variety of different tissues. At a later stage, the proposed

approach should be applied to whole blood (with cells

lysed before virus recovery), nasal washings, tissues, and

other potentially virus-laden samples.

Automation and Containment

To routinely detect new and potentially lethal viruses,

researchers may need to create completely automated and

contained laboratories that continually search for and

sequence viruses from a wide variety of sources to hone

skills; demonstrate efficiency; and develop improved sys-

tems, methods, and reagents.

Containment was of great concern in the original

Manhattan Project to deal with radiologic hazards, and in

the Oak Ridge centrifuge project (34) to contain infectious

agents. Containment systems have since evolved in two

directions. In biological sciences, interest has centered on

schemes to allow investigators to work in a safe environ-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 771

PERSPECTIVES

ment using essentially the same tools they would use on an

open bench. As a result, designs has evolved in which

human operators are contained in “space suits.” In nuclear

programs, in contrast, (where containment systems actual-

ly originated), operators are completely isolated from the

contents of “hot cells,” and operations in these cells are

done remotely with specially designed equipment. Given

the national urgency for automated systems for virologic

studies, completely robotic automated systems should be

developed, analogous to those used in nuclear research,

because 1) the concentrated samples to be analyzed are

potentially extremely dangerous, 2) work must be done

without interruption, and 3) speed and precision depend on

automation. Although the K-II centrifuge has not thus far

been automated for totally remote operation, including

cleaning between runs, this is not an overwhelming prob-

lem and should involve cascaded centrifuges, as was done

for the mass isolation of Tussock moth polyhedral inclu-

sion bodies (15).

False Positives, False Negatives,

and the Price of Errors

All current diagnostic tests have the potential for false-

negative and false-positive results. In the atmosphere of an

actual terrorist attack with a biological agent, however, the

consequences of these false outcomes place an enormous

strain on public services, as demonstrated by the recent

anthrax episodes. A false positive triggers highly disrup-

tive responses, whereas a false-negative result exposes the

population to the obvious health concern. The approach

described here reduces the possibility of such outcomes

and only assumes that the virus has the expected biophys-

ical properties of size and mass and an internalized

genome. If a particle with the appropriate biophysical

properties coincides with an internal genome that codes for

structural proteins that are also found in the same fraction,

the possible number of false-positive error is acceptably

small and few mechanisms exist by which false evidence

for a truly nonexistent viral sequence might emerge from

the process described. The level of false negatives depends

not only on the overall quality of the analysis but also on

its sensitivity to rare events, i.e., the dynamic range. For

sequence-based analyses, sensitivity depends on the fre-

quency with which a sequence appears in a fragment

library, the number of clones produced, the efficiency of

known virus subtraction (if applied), and the number of

different clones sequenced.

The number of intentional false positives (duping) is

another matter and one that has two aspects. First, inten-

tional introduction of unexpected pathogens or their genes

into a global analytical system is itself a terrorist act and

one that should be detected and known. To insert a substan-

tial amount of recoverable viral particles into the sample

collection system, a person trying to deceive the system

would have to engineer and grow these viruses—an act

sufficiently close to actual bioterrorist use that it requires

detection, whether the agent is a serious human pathogen

or not. The second aspect concerns the best response to the

suspicion that such duping has occurred. Complete

sequencing of the agent(s) involved would be important

since one cannot initially distinguish a genuine sample

from one used in duping. Only after extensive further stud-

ies and the demonstration that an outbreak has not occurred

may the sample be determined not to be of patient origin.

Forewarned is forearmed. Given advance notice, even

by weeks, of an impending viral outbreak, the hope exists

that the tools and imaginations of molecular biology will

find the means to prepare some effective biological

defense.

Medical Contributions of Global Surveillance

The problem of developing new antiviral agents, espe-

cially those specific for one or only a few viral diseases, is

circular. Without such treatments rapid agent identification

is not necessary, but without such identification no press-

ing commercial justification for developing specific antivi-

ral agents exists (except for HIV) because they will not be

widely used. To be successful, diagnosis and therapy must

be linked. This project would assist in forging that link.

Conclusion

Isolating and sequencing the genomes of a wide variety

of viruses from pools of the excess human serum and plas-

ma currently collected and discarded by large diagnostic

laboratories is now technically feasible. This collection

and analysis process could allow new or unknown

pathogens to be identified in the first, or at most second,

round of infection. Not all human viral pathogens will be

present in such mixtures, but they will include a large frac-

tion of all known highly infectious viral agents. Since the

core technologies, though varied, are highly developed, we

believe that the initial feasibility studies could be complet-

ed in 1 year.

Former Senator Sam Nunn and William H. Wulf, pres-

ident of the National Academy of Engineering, have both

proposed setting up a project concerned with bioterrorism,

modeled after the Manhattan Project. We believe that the

project described could form the nucleus of such an effort

and suggest that lessons learned in the Oak Ridge cen-

trifuge project may apply. As noted by Alvin Weinberg,

that project was the first (and hopefully not the last) large-

scale project in the biological sciences in which facile

access to a wide range of technologies was provided, on

the model of the original Manhattan Project (35).

In separate articles, we will discuss the possibility of

linking rapid detection to rapid responses, including vac-

772 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

PERSPECTIVES

cine and therapeutic antibody development, in an attempt

to abort epidemics caused by new viruses while they are in

progress.

Acknowledgments

We acknowledge with gratitude the continuing support and

encouragement of Alvin Weinberg, who was director of the Oak

Ridge National Laboratory when the original development was

done.

Dr. Anderson was director of the Molecular Anatomy

Program at the Oak Ridge National Laboratory, held a similar

position at the Argonne National Laboratory, was chief scientist

at Large Scale Biology Corporation, and is now president of the

Viral Defense Foundation. His chief interests include proteomics,

virology, and biophysical separations.

References

1. Rohwer F, Seguritan V, Choi DH, Segall AM, Azam F. Production of

shotgun libraries using random amplification. Biotechniques

2001;31:108–17.

2. Breitbart M, Salamon P, Andresen B, Mahaffy JM, Segall AM, Mead,

D, et al. Genomic analysis of uncultured marine viral communities.

Proc Natl Acad Sci U S A 2002;99:14250–55.

3. Steward GF, Montiel JL, Azam F Genome size distributions indicate

variability and similarities among marine viral assemblages from

diverse environments. Limnology and Oceanography

2000;45:1697–706.

4. Anderson NG, editor. The development of zonal centrifuges and

ancillary systems for tissue fractionation and analysis. Natl Cancer

Inst Monogr 1966;21:526.

5. Anderson NG, Harris WW, Barber AA, Rankin CT, Candler EL.

Separation of sub-cellular components and viruses by combined rate-

and isopycnic-zonal centrifugation. Natl Cancer Inst Monogr

1966;21:253–83.

6. Allander T, Emerson SU, Engle RE, Purcell RH, Bukh J. A virus dis-

covery method incorporating DNase treatment and its application to

the identification of two bovine parvovirus species. Proc Natl Acad

Sci U S A 2001;98:11609–14.

7. Anderson NG, Waters DA, Nunley CE, Gibson RF, Schilling RM,

Denny C, et al. K-Series centrifuges. I. Development of the K-II con-

tinuous-sample-flow-with-banding centrifuge system for vaccine

purification. Anal Biochem 1969:32:460–94.

8. Perardi TE, Leffler RA, Anderson, NG. K-Series centrifuges. II.

Performance of the K-II rotor. Anal Biochem 1969;32:495–511.

9. Perardi TE, Anderson NG. K-Series centrifuges. III. Effect of core

taper on particle capture efficiency. Anal Biochem 1970;34:112–22.

10. Brantley JN, Willis DD, Breillatt JP, Gibson RF, Patrick LC,

Anderson NG. K-Series centrifuges IV. Measurement and control of

temperature. Anal Biochem 1970;36:434–42.

11. Gerin JL, Anderson, NG. Purification of influenza virus in the K-II

zonal centrifuge. Nature 1969;221:1255–6.

12. Reimer CB, Baker RS, Van Frank RM, Newlin TE, Cline GB,

Anderson NG. Purification of large quantities of influenza virus by

density gradient centrifugation. J Virol 1967;1:1207–6.

13. Gerin JL, Holland PV, Purcell RH. Australia antigen: large-scale

purification from human serum and biochemical studies of its pro-

teins. J Virol 1971;7:569–76.

14. Hilleman MR, McAleer WJ, Buynak EB, McLean AA. The prepara-

tion and safety of hepatitis B vaccine. J Infect 1983(Suppl 1);7:3–8.

15. Breillatt JP, Brantley JN, Mazzone HM, Martignoni ME, Franklin JE,

Anderson, NG. Mass purification of nucleopolyhedrosis virus inclu-

sion bodies in the K-Series centrifuge. Appl Microbiol

1972;23:923–30.

16. Alfa Wassermann Inc. Available from: URL: www.smartcentrifuge.com

17. Kendro Laboratory Products. Available from: URL: www.kendro.com

18. Anderson NG, Anderson NL. Detection and characterization of

microorganisms. U.S. Patent Application publication no.

2002/0137026 A1, Sept. 26, 2002.

19. Anderson NG, Anderson NL, Della-Cioppa, G. Method for discover-

ing new infectious particles. U.S. Patent Application publication no.

2003/0044771 A1, March 6, 2003.

20. Anderson NG, Cline GB, Harris WW, Green JG. Isolation of viral par-

ticles from large fluid volumes. In: Berg G, editor. Transmission of

viruses by the water route. New York: Interscience Publishers; 1967.

p. 75–88.

21. Bergh O, Borsheim KY, Bratbak G, Heldal M. High abundance of

viruses found in aquatic environments. Nature 1989;340:467–8.

22. Wommack KE, Colwell RR. Virioplankton: viruses in aquatic ecosys-

tems. Microbiol Mol Biol Rev 2000:64:69–114.

23. Noble RT, Fuhrman JA. Rapid virus production and removal as meas-

ured with fluorescently labeled viruses as tracers. Appl Environ

Microbiol 2000;66:3790–7.

24. Brussaard CP, Marie D, Bratbak G. Flow cytometric detection of

viruses. J Virol Methods 2000;85:175–82.

25. Chen F, Suttle CA, Short SM. Genetic diversity in marine algal virus

communities as revealed by sequence analysis of DNA polymerase

genes. Appl Environ Microbiol 1996;62:2869–74.

26. Fuhrman JA. Marine viruses and their biogeochemical and ecological

effects. Nature 1999;399:541–8.

27. Rohwer F, Edwards R. The phage proteomic tree: a genome-based

taxonomy for phage. J Bacteriol 2002;184:4529–5.

28. Griffin DW, Donaldson KA, Paul JH. Pathogenic human viruses in

coastal waters. Clin Microbiol Rev 2003;16:129–43.

29. Hara S, Koike I, Terauchi K, Kamiya H, Tanoue E. Marine Ecology

Progress Series 1996;145:269–77.

30. Steward GF, DeLong EF, Preston CM Assessing viral diversity in the

sea. Eos Trans 2002(Suppl);83:OS19.

31. Lee TH, Montalvo L, Chrebtow V, Busch MP. Quantitation of genom-

ic DNA in plasma and serum samples: higher concentrations of

genomic DNA found in serum than in plasma. Transfusion

2001;41:276–82.

32. Quinn TC, Brookmeyer R, Kline R, Shepherd M, Paranjape R,

Mehendale, S, et al. Feasibility of pooling sera for HIV-1 viral RNA

to diagnose acute primary HIV-1 infection and estimate HIV inci-

dence. AIDS 2000;14:2751–7.

33. Lefkovits I, Kettman JR, Frey JR. Proteomic analysis of rare molec-

ular species of translated polypeptides from a mouse fetal thymus

cDNA library. Proteomics 2001;1:560–73.

34. Cho N, Barringer HP, Amburgey JW, Cline GB, Anderson NG,

McCauley LL, et al. Problems in biocontainment. Natl Cancer Inst

Monogr 1966;21:485–502.

35. Weinberg AM The birth of big biology. Nature 1999;401:738.

Address for correspondence: Norman G. Anderson, Viral Defense

Foundation, Box 2126, Kensington, MD 20891, USA; fax: 301-770-

9091; email: [email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 773

PERSPECTIVES

We describe Salmonella control programs of broiler

chickens, layer hens, and pigs in Denmark. Major reduc-

tions in the incidence of foodborne human salmonellosis

have occurred by integrated control of farms and food pro-

cessing plants. Disease control has been achieved by mon-

itoring the herds and flocks, eliminating infected animals,

and diversifying animals (animals and products are

processed differently depending on Salmonella status) and

animal food products according to the determined risk. In

2001, the Danish society saved U.S.$25.5 million by con-

trolling Salmonella. The total annual Salmonella control

costs in year 2001 were U.S.$14.1 million (U.S.$0.075/kg

of pork and U.S.$0.02/kg of broiler or egg). These costs are

paid almost exclusively by the industry. The control princi-

ples described are applicable to most industrialized coun-

tries with modern intensive farming systems.

almonellosis is one of the most common causes of

foodborne diarrheal disease worldwide. Most of these

infections are zoonotic and are transmitted from healthy

carrier animals to humans through contaminated food. The

main reservoir of zoonotic Salmonella is food animals, and

the main sources of infections in industrialized countries

are animal-derived products, notably fresh meat products

and eggs. In developing countries, contaminated vegeta-

bles, water, and human-to-human transmission are

believed to contribute to a comparatively larger proportion

of the human cases than those in industrialized countries

(1). However, the incidence of human salmonellosis

increased in most industrialized countries in the 1980s and

1990s. Rapid spread of a limited number of successful

Salmonella clones in different sectors of food animal pro-

duction (swine, broiler chickens, and particularly layer

hens) has been suggested as the most important cause of

this increase (2).

Despite much research and many national and interna-

tional attempts to implement control strategies, the inci-

dence of human salmonellosis in most countries remains

high. One notable exception is Sweden, which remains

essentially free from the Salmonella problems typical for

most other industrialized countries. The background for

the Swedish success has been described (3). Unfortunately,

other countries cannot apply the Swedish model of

Salmonella control, which requires near freedom from

Salmonella in domestic food animal production from the

onset. In the European Union, the Zoonosis Directive (4)

was an attempt to initiate a European Union–wide control

effort against foodborne zoonoses, particularly Salmonella

in broiler chickens and layer hens. Most European Union

countries found that they either could not or would not

implement the directive, which did not permit use of vac-

cines, antimicrobial drugs, or both as elements in the con-

trol program of Salmonella in broiler chickens or layer

hens. This constraint was seen as an obstacle by some

countries. Recently a new directive has been formulated,

which is awaiting final approval by the European Union

Parliament.

In Denmark, the incidence in human salmonellosis

increased rapidly in the second half of the 1980s because

of the spread of Salmonella in broiler chickens. This

increase led to the initiation of a targeted national control

program (5). Subsequent spread of Salmonella in swine

and layer hens has also led to increases in human disease

incidence and subsequently to the development and imple-

mentation of targeted control efforts (6–8). We review

Denmark’s Salmonella control programs and the effect on

Salmonella in food animals, food, and humans. We also

evaluate and discuss control costs and public health econ-

omy aspects.

Control of Salmonella in Broiler Chickens

Objectives, Program, and Effects

The initial aim of the program was that <5% of broiler

flocks would be infected with Salmonella. The program

was successful and was gradually revised towards assur-

ance of complete freedom from Salmonella in broiler pro-

duction.

The program is based on the principle of top-down

eradication, ensuring freedom from Salmonella from the

top of the broiler-breeding pyramid down. Infected flocks

of breeding animals are destroyed, and infected birds are

774 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

Salmonella Control Programs

in Denmark

Henrik C. Wegener,* Tine Hald,* Danilo Lo Fo Wong,* Mogens Madsen,* Helle Korsgaard,*

Flemming Bager,* Peter Gerner-Smidt,† and Kåre Mølbak†

*Danish Veterinary Institute, Copenhagen, Denmark; and †Statens

Serum Institut, Copenhagen, Denmark

processed for slaughter. The testing program has devel-

oped gradually to adjust to higher food safety objectives.

As progress stalled, more intensive serologic and bacterio-

logic testing was developed and applied (5,9–11). The cur-

rent testing scheme is shown in Table 1.

Birds from infected flocks are slaughtered on separate

slaughter lines or late in the day to avoid cross-contamina-

tion. Farmers get a better price for birds from Salmonella-

free flocks, and slaughterhouses can use the label

“Salmonella-free” for birds that meet criteria determined

by the authorities. No decontaminants, such as organic

acids or chlorine, are used during carcass processing.

The proportion of Salmonella-infected broiler flocks

has been markedly reduced since the initiation of the con-

trol program. Figure 1 shows that >65% of broiler flocks

tested positive for Salmonella during the first year of the

program, 1988–89, versus <5% in 2000. This decrease in

Salmonella has led to a concomitant reduction in the pro-

portion of infected broiler carcasses after slaughter and at

retail.

The Danish government and the European Union equal-

ly compensate owners of destroyed breeding stock for their

losses. In 1993, a major Danish retailer (COOP-Denmark)

stopped the marketing of broiler chicken, which exceeded

a 5% target. Danish chicken could not meet this target at

that time, so producers suffered severe losses because they

had to export their chicken to lower priced markets.

Salmonella can be effectively reduced (nearly eliminat-

ed) from broiler chickens by intensive flock-level testing

and top-down eradication. Essential to success is a suffi-

ciently sensitive testing program in the breeding and rear-

ing flocks as well as in the hatcheries, i.e., one that

involves intensive sampling and a combination of serolog-

ic and bacteriologic testing methods (Table 1).

Bacteriologic testing alone is not sufficiently sensitive to

achieve control, especially if S. Enteritidis infections are

present. Removal of all organic material, thorough clean-

ing and disinfection of the poultry house, and an empty

resting period of 10–14 days between flocks can effective-

ly eliminate residual infections. In Denmark, most infec-

tions appear to be vertically transmitted (nearly always

traceable to an infected hatchery or parent flock), whereas

horizontal transmission from the environment and wild

fauna appear to play a minor role. Competitive exclusion

cultures, vaccines, or antibiotics have not been used in the

Danish control program.

Control of Salmonella in Layer Hens

Objectives, Program, and Effects

All shell eggs from commercial layer flocks should be

free from S. Enteritidis and S. Typhimurium. Control of

layer breeders in Denmark is essentially identical to the

control program for broiler breeders (Table 1). Blood and

fecal samples of rearing flocks are tested (8,11), and

infected flocks are destroyed. All commercial flocks of

layer hens in production are tested routinely every 9 weeks

by a combination of serologic testing of egg yolk and bac-

teriologic testing of environmental samples (Table 1,

Figure 2).

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 775

SYNOPSES

Table 1. Salmonella surveillance of the broiler and egg production, Denmark, 2000

Stage of production

Age or frequency

Samples taken

Method

Day-old chickens

10 samples of crate material, 20 dead or destroyed chickens

Bacteriologic

1 wk

40 dead chickens

Bacteriologic

2 wks

2 pairs of sock samples

Bacteriologic

4 wks

60 fecal samples

Bacteriologic

8 wks

2 pairs of sock samples

Bacteriologic

Central rearing stations,

broiler and egg sector

2 weeks before moving

60 fecal samples and 60 blood samples

Bacteriologic, serologic

Every 2 wks

50 dead chickens or meconium from 250 chickens

taken from the hatchery

Bacteriologic

Breeders (hatching egg

production)-broiler and

egg sector

Every wk

2 pairs of sock samples

Bacteriologic

Hatchery

After each hatching

Wet dust

Bacteriologic

Day-old chickens

10 samples of crate material and 20 dead chickens

Bacteriologic

3 wks

5x2 sock samples in floor production units

or 300 fecal samples

Bacteriologic

Rearing egg production

12 weeks

5x2 sock samples in floor production units

or 300 fecal samples, and 60 blood samples

Bacteriologic, serologic

Every 9th wk for eggs sold to

authorized egg-packing centers

2 pairs of sock samples in floor production

units or fecal samples and egg samples

Bacteriologic, serologic

Egg production

Every 6 mo for eggs sold at

barnyard sale

2 pairs of sock samples or fecal samples and egg samples

Bacteriologic, serologic

Requirements of the European Union Zoonosis Directive (92/117/EEC).

Samples taken by the district veterinary officer.

Samples taken by the district veterinary officer every 8 weeks.

Samples taken by the district veterinary officer every 3 months.

All eggs from suspect or confirmed-positive layer

flocks are pasteurized. All shell eggs are distributed in a

cold chain (not exceeding 12°C) and kept refrigerated at

retail; eggs are generally refrigerated in private homes.

The government and the European Union equally com-

pensate owners of destroyed breeding stock for their loss-

es. The proportion of layer flocks infected with

Salmonella, notably S. Enteritidis, has been markedly

reduced since the initiation of the control program. Figure

3 shows that >7% of layer flocks tested positive for

Salmonella in the first year of the program, 1998, versus

<2% in 2001. The level of Salmonella-contaminated shell

eggs has not been measured from the initiation of the con-

trol program. However, a year before the program began, a

study of 13,000 eggs from different types of production

determined the level to be 1 per 1,000 eggs (20% of the

contaminated eggs harbored S. Enteritidis) (12).

Top-down eradication of S. Enteritidis has effectively

reduced the level of Salmonella, notably S. Enteritidis, in

Danish commercial layer flocks. The program has been

effective in free range, deep litter, organic, and caged birds.

Frequent testing by a combination of serologic and bacteri-

ologic testing methods is essential to achieve adequate sen-

sitivity in the monitoring program. Control of residual

infections in poultry houses can be conducted with a suc-

cess rate of nearly 70% by thorough cleansing and disinfec-

tion of the depopulated house (removal of all organic mate-

rial, disinfection of surfaces, and resting of the empty

house for 2 weeks). Day-old chicks for rearing must be

antibiotic free. Competitive exclusion cultures and vacci-

nation are not used in the Danish program. Vaccination

cannot, at present, be used in combination with serologic

testing because of problems of cross-reaction.

Control of Salmonella in Pork

Objectives, Program, and Effects

Denmark is the only country with a nationwide control

program of Salmonella in pork that is integrated from

“feed-to-food.” The program is based on routine testing

and classification of slaughter pig herds and subsequent

slaughter of pigs according to the inherent risk, as meas-

ured by the continual test program (Figure 2; Table 2). The

program has been described in detail elsewhere (7,8,13).

Pre-Harvest Control

Pigs from breeding and multiplying herds are tested

monthly by serologic testing of blood samples. If a specif-

ic cutoff level is reached, bacteriologic confirmatory test-

ing is carried out. Further, if the serologic reactions exceed

a specific high level, all movement of animals is restricted.

Slaughter pig herds are monitored continuously by sero-

logic testing of “meat juice” (drip fluid released from meat

after freezing and thawing) (14). Meat samples for testing

are collected at the slaughter line, and the number of sam-

ples and frequency of sampling are determined by the size

of the herd. Approximately 700,000 slaughter pigs are cur-

776 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

Figure 1. Salmonella in Danish broiler flocks as determined by

bacteriologic testing of every flock 2–3 weeks before slaughter

(N>4,000 flocks/year).

Figure 2. A) Receipt of 60 eggs per producer every 9 weeks (bar-

code indicating producer is shown). B) The “eggbreaker” punches

a hole in 30 eggs at a time. C) Withdrawal of egg yolk from 30

eggs and transfer to microtiter tray. D) Enzyme-linked immunosor-

bent assay analysis, reading, and transfer of results to central

database.

Figure 3. Salmonella in Danish layer flocks as determined by sero-

logic and bacteriologic testing of each commercial flock in week 9

of production.

rently tested each year (Figure 4). Herds sending <200 pigs

to slaughter each year are not tested, leaving 1.6% of the

slaughter pigs outside the monitoring scheme (13). The

herds are categorized in three levels based on the propor-

tion of seropositive meat juice samples during the last 3

months. Owners in level 2 and 3 are encouraged to seek

advice on how to reduce the Salmonella problem in the

herd (e.g., feeding, hygiene, and management).

Furthermore, payment from the slaughterhouse is reduced

by 2% and 4%, respectively.

The postharvest surveillance program has been

described (8,15,16). Pigs from herds in levels 1 and 2 are

slaughtered traditionally without any special precautions.

Pigs from level 3 herds can only be slaughtered in special

slaughterhouses under special hygienic precautions.

Carcasses from level 3 herds are tested for bacteria after

slaughter, and if the level of contamination exceeds a cer-

tain limit all carcasses from the particular herd have to

undergo heat treatment or other risk-reducing processing.

All slaughterhouses do routine bacteriologic testing of car-

casses according to a sampling plan, which ensures that

testing is random and representative of the national swine

production (>30,000 samples/year). Slaughterhouses that

exceed a certain predetermined level of Salmonella in the

routine monitoring of carcasses are obliged to investigate

and reduce the contamination problem to an acceptable

level.

The prevalence of swine herds in level 2 and 3 respec-

tively, has been steadily reduced since the program began

(Figure 5). Bacteriologic testing has indicated that the herd

infection level was reduced by 50% (from 14.7% to 7.2%

in small herds and 22.2% to 10.4% in large herds) from

1993 (when the program was implemented) to 1998 (17).

In the same period, the level of Salmonella contamination

in pork products, as determined by the routine monitoring

program, was reduced from 3% to <1% (Figure 6).

As Salmonella eradication in swine herds is difficult

because of the continual nature of the production system,

reducing the infection level should be the aim of a control

strategy. The low infection level in the herds and contami-

nation in the products can likely be reduced further in

Denmark. As the contamination level goes below 1%, test-

ing for contamination requires increasingly large numbers

of samples and consequently becomes very expensive,

which is one reason the sampling plan of herds and prod-

ucts has become more sensitive in recent years. This

change in testing sensitivity makes it difficult to compare

current and past levels of infection and contamination but

is nevertheless a necessity for the continued improvement

of the program.

A combination of serologic and bacteriologic testing is

essential for the success of the program. Nearly 1 million

serologic samples are tested each year. Testing this large

number of samples would not be possible because of finan-

cial and logistic constraints if the program were to rely on

bacteriologic testing alone.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 777

SYNOPSES

Table 2. Salmonella surveillance in pig and pork production, Denmark, 2001

Type of production

Sample

No. and frequency

Response

Breeding and multiplying

herds

Blood

10 times per mo

Confirmatory bacteriologic testing and restrictions on the movement of

animals if above predetermined level

Pig herds

Feces

100 in 20 pools of five

collected on indication

Salmonella reduction plan implemented

Slaughter pig producers

producing >200 pigs per year

Meat juice

Depending on herd

size (60–100 samples

per yr). Samples are

collected continuously

and semi-randomly

Confirmatory bacteriologic testing (20 pools of 5 fecal samples). Herds are

assigned to one of three levels depending on serology. Level 1: no sanctions;

level 2: implementation of Salmonella reducing actions in the herd; and level

3: same as level 2 and obligatory slaughter of pigs under special hygienic

precautions, including postslaughter microbial testing and potential heat

treatment of all meat products

Carcass after slaughter

Surface

swab

Swabs of five

carcasses are pooled

into one sample. One

sample per day in each

slaughterhouse.

Slaughterhouses exceeding a predetermined number of positive swabs in a 3-

months period are obliged to implement corrective actions

Figure 4. A) Receipt of pork samples from the slaughterhouse.

Each tube is labeled with a barcode, indicating herd of origin.

Samples are frozen overnight. B) The tube is entered in a rack with

the barcode facing outward. Meat juice is sieved into the tube from

the container during thawing. C) Withdrawal of meat juice from

tube and transfer to microtiter tray. D) Enzyme-linked immunosor-

bent assay analysis, reading, and transfer of results to central

database.

Feeding strategy (e.g., increased coarseness of feed and

wet feeding) and improved management (e.g., sectioning

and all-in all-out production) and hygiene standards are

important elements in the preharvest control efforts. Using

commercial prebiotic cultures is not necessary; natural

microflora in the feed, especially wet fermented feed,

appear to have a protective effect (18).

A reduction of Salmonella in slaughter pig herds has

been attained without Salmonella–free breeding herds.

However, to ensure the highest degree of consistency in

the program, the levels of Salmonella in breeding herds

should be kept as low as possible, and infected breeding

herds should not be sold to producers of herds of a superi-

or Salmonella status.

Determination of Public Health Impact

To better explain the mechanisms in the occurrence of

Salmonella infections in humans, the Danish Zoonosis

Centre has previously described a method that estimates

the number of human cases attributable to each of the

major animal-food sources (19,20). Using this method, we

compared Salmonella types isolated from animals and

foods with Salmonella types isolated from humans. In

brief, subtypes of Salmonella that are almost exclusively

found in a particular food animal reservoir or food type

(unique types) are used as anchor points for the distribu-

tion of subtypes occurring in several reservoirs and

sources. All human infections caused by the unique types

are associated with the indicated food type or derived from

the indicated food animal reservoir (e.g., pork, beef, chick-

ens, or eggs). Salmonella types, which occur in several

reservoirs, are distributed relative to the prevalence of

unique types in each reservoir or food type. Detailed

knowledge of the distribution of Salmonella types in all

relevant food animals and food types, generated through

intensive and continuous monitoring, is a prerequisite for

the analysis. Recently, a stochastic model based on the

principles of the previous method was developed and

applied. This model allows us to consider the uncertainty

around the estimated parameters (21).

Figure 7 shows the human salmonellosis incidence

associated with the three major sources of human salmo-

nellosis in Denmark from 1988 to 2001. The year that a

control program was launched for a specific food animal

production system is also indicated. The control programs

have been successful in achieving the main objective, a

reduction of the incidence of human salmonellosis. The

broiler-associated salmonellosis incidence

(cases/100,000) has been reduced by >95.0, from 30.8 in

1988 to 0.5 in 2001; the pork-associated salmonellosis

incidence has been reduced by >85, from 22.0 in 1993 to

3.0 in 2001; and the egg-associated salmonellosis inci-

dence has been reduced by nearly 75, from 57.7 in 1997 to

15.5 in 2001. Trends in the animal and food-specific dis-

ease incidence estimates show a high degree of agreement

with the trends in prevalence of Salmonella in specific

food animals and the corresponding animal-derived food

products. These trends serve as an indirect validation of

the estimates because these estimates do not rely on preva-

lence data.

Economy of Salmonella Control

Costs of Salmonella Control

The Audit Office of Denmark has evaluated the govern-

ment spending in relation to the national Salmonella con-

trol efforts (22). From 1994 to 1999, the control program

for broiler chickens and layer hens involved government

finances of a total of 188.1 million Danish kroner (DKK)

(U.S.$26.5 million) (U.S.$1.00 = 7.1 DKK). A total of

109.7 million DKK (U.S.$15.45 million) was paid to com-

pensate farmers for destroyed animals; most of the remain-

ing costs were associated with establishing and running the

surveillance program. These costs were highest in the ini-

tial phase of the layer hen control program in 1997 but

778 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

Figure 5. Prevalence of Salmonella in Danish pig herds as deter-

mined by continuous serologic testing of all commercial pig herds

(N >700,000 samples tested/year). Herds are categorized in three

levels based on the proportion of seropositive meat juice samples

during the last 3 months. Owners in level 2 and 3 are encouraged

to seek advice on how to reduce the Salmonella problem in the

herd (e.g., feeding, hygiene, and management). Furthermore, pigs

from level 3 herds can only be slaughtered in special slaughter-

houses under special hygienic precautions. Data from the Danish

Veterinary and Food Administration.

Figure 6. Salmonella detected in pork, as determined by continu-

ous randomized sampling of pork end-products from all major

national pig slaughterhouses (N >30,000 samples/year).

have been reduced considerably since then. In 2001, all

costs associated with running the program were assumed

by the poultry industry (with the exception of govern-

ments’ compensation for flocks destroyed according to the

European Union Zoonosis Directive). The current control

costs for layer hens and broiler chickens are estimated to

be in the range of US$4.2 million per year (for 344 broiler

producers, producing 135 million broiler chickens year,

and 392 shell-egg producers, producing 1 billion shell-

eggs per year).

In the initial phase, the Salmonella control program of

pigs and pork cost the industry and government a total of

U.S.$14 million per year. With the recent revision of the

program, the responsibility has been taken over solely by

the industry, and operational costs have been reduced to

approximately U.S.$8.5 million per year (for 21,000 pro-

ducers, producing 21–22 million slaughter pigs each year)

(B. Nielsen, pers. comm.).

Public Health Economy

Direct health costs (e.g., hospitalization, consulting a

physician, and laboratory testing) as well as the costs of

lost labor (e.g., loss of production per day away from

work) in relation to a case of salmonellosis in Denmark

were evaluated as part of a multidisciplinary task force

(Korsgaard and Wegener, pers. comm.). For 2001, food-

borne salmonellosis cost the Danish U.S.$15.5 million.

The estimate is based on an incidence of 54.6 cases per

100,000, and approximately 10% of cases are laboratory

confirmed. Assuming that 5% or 20% of cases are labora-

tory-confirmed changes the estimate to U.S.$25.5 million

and U.S.$10.4 million, respectively.

Costs and Benefits

Assuming that salmonellosis associated with each of

the major sources would have remained at the precontrol

program incidence (and not increased further) (i.e., if no

action had been taken to curb the problem), we calculated

a hypothetical “no-control” salmonellosis incidence. This

incidence would have been 137.5 (pork 22, broiler chick-

ens 30.8, eggs 57.7, and average residual, 27). The socie-

tal costs, in the absence of the existing control programs,

would thus have been U.S.$41 million per year (assuming

10% of cases are laboratory confirmed). Thus, in 2001,

Denmark saved U.S.$25.5 million by controlling

Salmonella. The estimated annual Salmonella control costs

from 2000 and onwards are approximately U.S.$14.1 mil-

lion. These costs are borne almost exclusively by the ani-

mal producers and the food industry, which suggests that

the costs are passed on to consumers through higher food

prices. Based on the figures above and data on annual pro-

duction (23), control costs amount to approximately

U.S.$0.075/kg of pork and U.S.$0.02/kg of broiler or egg.

Discussion

Danish Salmonella control efforts have been successful

in achieving their objective; reduction of human salmonel-

losis. These efforts illustrate that with a focused and inte-

grated programs, including a strong element of preharvest

control, and based on a public-private partnership,

Salmonella can be reduced. At the same time, the indus-

tries involved have remained profitable and internationally

competitive (approximately 75% of the chicken products

and 85% of the pork are exported).

Initially, the programs have received some government

funding, primarily for research, development, and com-

pensation for destroyed animals. After the initial imple-

mentation and clean-up phase of the programs, the respon-

sibility for running and funding the programs have been

nearly completely taken over by the industries involved.

The government, however, maintains access to all relevant

information and data through a central database managed

by the Danish Zoonosis Centre, and food safety objectives

continue to be determined by the Danish government.

A proactive and collaborative approach to food safety

by food industry and government ensures consumers’ con-

fidence in the domestic food production. For example,

when the bovine spongiform encephalopathy (BSE) crisis

hit Europe, the beef industry in most countries was

adversely affected by reduced consumer demand. In

Denmark, the sale of beef remained nearly unaffected by

the crisis even after the first positive BSE findings

occurred in Danish cattle. These steady beef sales are like-

ly attributable to a high degree of consumer confidence in

the public and private control systems.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 779

SYNOPSES

Figure 7. Effects of Salmonella control programs as indicated by

incidence of human infection attributable to the different major

sources of human salmonellosis in Denmark; 1) Salmonella con-

trol program for broiler chickens implemented, 2) Salmonella con-

trol program implemented for pigs and pork, 3) Salmonella control

program implemented for layer hens and eggs The three sources

account for approximately 50% to 75% of Salmonella each year.

Remaining cases are attributable to beef, imported food products,

infections acquired while traveling abroad, and unknown sources.

Incidence (cases/100.000 inhabitants).

The success of the programs supports the effectiveness

of a preharvest control approach to Salmonella.

Monitoring and intervention at the farm and in the food

animal breeding systems are feasible means to achieve

lasting control of the Salmonella problem. Development

and application of a two-tiered detection system based on

a combination of serologic testing and bacteriologic con-

firmation have been essential for the success of the pro-

grams. Serologic testing enables semi-automated mass

screening of animals and eggs at a low price and with

good, and in some cases superior, sensitivity. Bacteriologic

testing serves to compensate for the sub-optimal specifici-

ty of a serologic-based monitoring system. The programs

could not have been operated solely on the basis of bacte-

riologic testing because of the higher costs involved and

logistical problems (i.e., screening nearly 2 million sam-

ples per year by bacteriologic testing is unrealistic).

Preharvest control tools, such as vaccines, antibiotics, or

competitive exclusion, are not used to control Salmonella

in Denmark; these tools might be counterproductive, as

they mask the Salmonella problem rather than aid in its

reduction or eradication.

Evaluating the costs and benefits of the national

Salmonella control efforts is difficult; estimating the pub-

lic health and societal costs in the absence of the control

program is impossible. However, this conservative esti-

mate suggests that efforts have been cost beneficial and

those benefits are likely to increase with time.

Dr. Wegener is a professor of zoonoses epidemiology and

head of the Danish Zoonosis Center. His main research interests

are the epidemiology of foodborne zoonosis and antimicrobial

resistant bacterial in the food chain. He is involved in the coordi-

nation of the World Health Organization Global Salmonella

Surveillance Program.

References

1. Salmonella. In: Acha PN, Szyfres B, editor. Zoonoses and communi-

cable diseases common to man and animals. PAHO Scientific and

Technical publications no. 580; 2001.

2. Thorns CJ. Bacterial food-borne zoonoses. Rev Sci Tech

2000;19:226–39.

3. Eld K, Gunnarsson A, Holmberg T, Hurvell B, Wierup M. Salmonella

isolated from animals and feedstuffs in Sweden during 1983-1987.

Acta Vet Scand 1991;32:261–77.

4. Directive concerning measures for protection against specified

zoonoses and specified zoonotic agents in animals and in products of

animal origin in order to prevent outbreaks of food-borne infections

and intoxications (Zoonosis Directive). Council Directive

92/117/EEC, December 1992.

5. Bisgaard M. A voluntary Salmonella control program for the broiler

industry, implemented by the Danish Poultry Council. Int J Food

Microbiol 1992;15:219–24.

6. Mousing J, Jensen PT, Halgaard C, Bager F, Feld N, Nielsen B, et al.

Nation-wide Salmonella Enterica surveillance and control in Danish

slaughter swine herds. Prev Vet Med 1997;29:247–61.

7. Nielsen B, Alban L, Stege H, Sorensen LL, Mogelmose V, Bagger J,

et al. A new Salmonella surveillance and control program in Danish

pig herds and slaughterhouses. Berl Munch Tierarztl Wochenschr

2001;114:323–6.

8. Feld NC, Ekeroth L, Gradel KO, Kabell S, Madsen M. Evaluation of

a serological Salmonella mix-ELISA for poultry used in a national

surveillance program. Epidemiol Infect 2000;125:263–8.

9. Skov MN, Carstensen B, Tornoe N, Madsen M. Evaluation of sam-

pling methods for the detection of Salmonella in broiler flocks. J Appl

Microbiol 1999;86:695–700.

10. Gradel KO, Andersen J, Madsen M. Comparisons of sampling proce-

dures and time of sampling for the detection of Salmonella in Danish

infected chicken flocks raised in floor systems. Acta Vet Scand

2002;43:21–30.

11. Skov MN, Feld NC, Carstensen B, Madsen M. The serologic

response to Salmonella Enteritidis and Salmonella Typhimurium in

experimentally infected chickens, followed by an indirect

lipopolysaccharide enzyme-linked immunosorbent assay and bacteri-

ologic examinations through a one-year period. Avian Dis

2002;46:265–73.

12. Bager F. Investigation of Salmonella prevalence in shell-eggs

(Undersøgelse af salmonella-forekomst i konsumæg) [In Danish].

Zoonose-Nyt 1996;3:7–8.

13. Alban L, Stege H, Dahl J. The new classification system for

slaughter-pig herds in the Danish Salmonella surveillance-and-

control program. Prev Vet Med 2002;53:133–46.

14. Nielsen B, Ekeroth L, Bager F, Lind P. Use of muscle fluid as a source

of antibodies for serologic detection of Salmonella infection in

slaughter pig herds. J Vet Diagn Invest 1998;10:158–63.

15. Nielsen B, Wegener HC. Public health and pork and pork products:

regional perspectives of Denmark. Rev Sci Tech 1997;16:513–24.

16. Hald T, Andersen JS. Trends and seasonal variations in the occur-

rence of Salmonella in pigs, pork and humans in Denmark,

1995–2000. Berl Munch Tierarztl Wochenschr 2001;114:346–9.

17. Christensen J, Baggesen DL, Nielsen B, Stryhn H. Herd prevalence

of Salmonella spp. in Danish pig herds after implementation of the

Danish Salmonella Control Program with reference to a pre-

implementation study. Vet Microbiol 2002;88:175–88.

18. Stege H, Jensen TK, Moller K, Baekbo P, Jorsal SE. Risk factors for

intestinal pathogens in Danish finishing pig herds. Prev Vet Med

2001;50:153–64.

19. The Ministry of Food, Agriculture and Fisheries. Annual report on

zoonoses in Denmark 1998. Copenhagen, Denmark: The Ministry.

Available from: URL: http://www.vetinst.dk

20. Hald T, Wegener H C. Quantitative assessment of the sources of

human salmonellosis attributable to pork. In Proceedings of the 3rd

International Symposium on Epidemiology and Control of

Salmonella in Pork, 4–7 Aug 1999, Washington, D.C. Urbana-

Champaign (IL): University of Illinois. p. 200–5.

21. Hald T, Vose D, Wegener HC. Quantifying the contribution of ani-

mal-food sources to human salmonellosis by a Bayesian approach.

Risk Anal. In press 2003.

22. Report from the National Audit Office of Denmark: The Danish

Governments Control of Salmonella [In Danish]: Beretning om

statens bekæmpelse af Salmonella). Rigsrevisonen. Report No. 8/99

2000. Available from: URL: http://www.rigsrevisionen.dk

23. The Ministry of Food, Agriculture and Fisheries. Annual report on

zoonoses in Denmark 2000. Copenhagen: The Ministry. Available

from: URL: www.vetinst.dk

Address for correspondence: Henrik C. Wegener, Danish Zoonosis

Centre, Danish Veterinary Institute, 27 Bülowsvej, DK-1790 Copenhagen

V, Denmark; fax: +45 35 30 03 77; email: [email protected]

780 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

The Emerging Infections Programs (EIPs), a popula-

tion-based network involving 10 state health departments

and the Centers for Disease Control and Prevention, com-

plement and support local, regional, and national surveil-

lance and research efforts. EIPs depend on collaboration

between public health agencies and clinical and academic

institutions to perform active, population-based surveillance

for infectious diseases; conduct applied epidemiologic and

laboratory research; implement and evaluate pilot preven-

tion and intervention projects; and provide capacity for flex-

ible public health response. Recent EIP work has included

monitoring the impact of a new conjugate vaccine on the

epidemiology of invasive pneumococcal disease, providing

the evidence base used to derive new recommendations to

prevent neonatal group B streptococcal disease, measuring

the impact of foodborne diseases in the United States, and

developing a systematic, integrated laboratory and epi-

demiologic method for syndrome-based surveillance.

uring the 1980s, clinicians added newly recognized

infectious diseases, such as toxic shock syndrome and

AIDS, to their differential diagnoses when evaluating pre-

viously healthy young adults with severe illness. More

recently, clinicians in the United States found themselves

considering the possibility of inhalational anthrax among

patients with influenzalike illnesses and adding West Nile

virus infection to their workup of posttransfusion fevers

(1–3). The existence of these and dozens of other emerg-

ing and reemerging infectious diseases, naturally or inten-

tionally transmitted, has removed any doubt about the

interdependence of clinical medicine and public health.

Clinicians are sentinels for detection of new or reemerging

diseases and may benefit from information acquired

through public health surveillance and research projects,

which helps to place the quantitative risks of these new

diseases in perspective amidst the media attention that

often accompanies the latest medical mysteries.

In 1992, the Institute of Medicine (IOM) articulated the

concept of emerging infections, discarding the naive view

that infectious diseases were problems of the past and cau-

tioning against complacency about public health prepared-

ness for infectious diseases (4). By defining emerging

infectious diseases as “new, reemerging, or drug-resistant

infections whose incidence in humans has increased with-

in the past two decades or whose incidence threatens to

increase in the near future,” IOM recognized the broad

scope of these diseases. The IOM report also cited factors

that influence the emergence of infectious diseases:

changes in human demographics and behavior; advances

in technology and changes in industry practices; economic

development and change in land-use patterns; increased

volume and speed of international travel and commerce;

microbial adaptation and change; and breakdown of public

health capacity at the local, national, and global levels. The

intentional release of anthrax in the United States in 2001

emphasized the need to add intentionally inflicted harm to

the list of factors that influence the emergence of infec-

tious diseases and to suspect the unexpected.

In response to the IOM report, Addressing Emerging

Infectious Disease Threats to Health: A Prevention

Strategy for the United States was developed by the

Centers for Disease Control and Prevention (CDC) (5). A

key recommendation of the plan called for establishing

population-based centers to complement and support local,

regional, and national surveillance and research efforts.

This concept was realized through Emerging Infections

Programs (EIPs), a network of state health departments

(Figure 1) coordinated by CDC. EIPs are intended to be a

national resource for surveillance and epidemiologic

research by conducting work that goes beyond the routine

public health department functions; by fostering collabora-

tions between the public health, academic, and clinical

communities; and by maintaining an infrastructure flexible

enough to address new infectious diseases challenges as

they emerge. An updated plan released in 1998 described

the important role assumed by EIPs in addressing emerg-

ing infections and identified several high-priority target

areas (6), which include: antimicrobial drug resistance,

foodborne and waterborne diseases, vector-borne and

zoonotic diseases, chronic diseases caused by infectious

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 781

SYNOPSES

Disease Surveillance and the

Academic, Clinical, and Public

Health Communities

Robert W. Pinner,* Catherine A. Rebmann,* Anne Schuchat,* and James M. Hughes*

*Centers for Disease Control and Prevention, Atlanta, Georgia,

USA

agents, diseases transmitted through blood transfusions or

products, vaccine development and use, diseases of preg-

nant women and newborns, diseases of persons with

impaired host defenses, and diseases of travelers, immi-

grants, and refugees. We describe EIP accomplishments

and future directions.

EIP Methods

The principal functions of EIPs are to perform active,

population-based surveillance for infectious diseases; con-

duct applied epidemiologic and laboratory research;

implement and evaluate pilot prevention and intervention

projects; and provide capacity for flexible public health

response. EIPs also develop and evaluate public health

practice and transfer what is learned to the public health

and medical communities.

These programs are supported through cooperative

agreements between CDC and state health departments,

who engage collaborators in local health departments, hos-

pitals, and academic institutions. Additional funding for

certain EIP activities comes from other sources; for exam-

ple, the U.S. Department of Agriculture and the Food and

Drug Administration provide support for activities involv-

ing foodborne illnesses, and the National Vaccine Program

Office has provided support for postlicensure vaccine eval-

uations.

The population base for EIP activities is approximately

36 million persons, though the base varies by project. This

population represents an approximation of the U.S. popu-

lation with respect to demographic characteristics such as

age, gender, race, and urban residence, as well as health

indicators such as population density and percentage of

persons at or below the poverty level (7). EIPs are geo-

graphically dispersed throughout the country (Figure 1).

Active, laboratory-based surveillance is the foundation

of two core EIP projects conducted at all sites: Active

Bacterial Core Surveillance (ABCs) and Foodborne

Disease Active Surveillance (FoodNet) (Table 1). These

active surveillance projects generate reliable estimates of

the incidence of certain infections and provide the founda-

tion for a variety of epidemiologic studies to explore risk

factors, disease spectrum, and prevention strategies (8,9).

For example, the total impact of foodborne illnesses in the

United States has been estimated by combining FoodNet

active surveillance data with other data sources and results

from FoodNet surveys of the general population (to learn

about the frequency of diarrhea in the general population

and to determine what proportion of persons with diarrhea

seeks medical care), physicians (to determine the frequen-

cy of stool-culturing by physicians), and clinical laborato-

ries (to determine the frequency of culturing for selected

foodborne pathogens) (9–11). These data provide esti-

mates of the overall occurrence of diarrheal illness (0.7 ill-

nesses/person-year), as well as the likely degree of under-

reporting for specific infections under surveillance (10).

Other projects are conducted by EIPs, depending on

local priorities and expertise. The Unexplained Deaths and

Critical Illness (UNEX) project, a prospective study that

uses epidemiologic and laboratory methods to detect and

investigate unexplained illnesses with clinical features

suggesting infectious diseases, has been in place at four

states with EIPs since the inception of the program (12,13).

The Connecticut EIP conducts active surveillance for

emerging tick-borne diseases that are transmitted by a sin-

gle tick vector (Ixodes scapularis) in the state (14). EIPs

also strive to maintain the flexibility to meet new chal-

lenges effectively. For example, in 1996 four EIP sites con-

ducted active surveillance for variant Creutzfeldt-Jakob

782 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

Table 1. Surveillance and focus area for two core projects conducted at all Emerging Infections Program sites

Projects

Type of surveillance

Focus

Active Bacterial Core Surveillance

Active, laboratory-based

Invasive disease (isolated from a normally sterile site such as

blood or cerebrospinal fluid) caused by group A streptococcus,

group B streptococcus, Haemophilus influenzae, Neisseria

meningitidis, and Streptococcus pneumoniae

FoodNet/Foodborne Disease Active Surveillance

Active, laboratory-based

Disease (first isolation from a person) caused by

Campylobacter, Listeria, Salmonella, Shigella, Yersinia,

Vibrio, Shiga toxin–producing Escherichia coli, including

O157:H7, Cryptosporidium, and Cyclospora

Intended to generate reliable estimates of the incidence of certain infections and provide the foundation for a variety of epidemiologic studies to explore risk factors,

disease spectrum, and prevention strategies.

Figure 1. Distribution of Emerging Infections Programs (EIPs), a

network of 10 state health departments and their collaborators in

local health departments, academic institutions, and clinical set-

tings, coordinated by the Centers for Disease Control and

Prevention. *New Mexico was added as the 10th EIP site in late

2002 and will begin EIP activities during 2003.

Disease (CJD) and physician-diagnosed CJD cases. This

study contributed to surveillance methods by confirming

that death certificate reviews are a sensitive method for

detecting CJD deaths while providing some assurance that

variant CJD was not occurring in these states (15).

Impact of a New Pneumococcal Vaccine

Through ABCs, we are evaluating the effect of the

pneumococcal conjugate vaccine on the epidemiology of

invasive pneumococcal disease in the United States.

Streptococcus pneumoniae (pneumococcus), which is an

important cause of serious illness among young children,

is the leading cause of bacterial pneumonia and meningitis

in the United States. For many years, immunization against

pneumococcus with a 23-valent polysaccharide vaccine

was recommended for persons >2 years of age who are at

high risk and for all adults >65 years of age. Although dis-

ease incidence is highest in the first 2 years of life, the

polysaccharide vaccine was poorly immunogenic in this

group. In February 2000, a protein-polysaccharide pneu-

mococcal conjugate vaccine for seven pneumococcal

serotypes (Prevnar, Wyeth Pharmaceuticals, Pearl River,

NY) was licensed for use in infants and children (16). This

conjugate vaccine is now recommended in the United

States for all children <2 years of age, with catch-up vac-

cination schedules suggested for children 2 to 4 years of

age. In clinical trials, the vaccine was efficacious against

invasive disease in infancy and reduced nasopharyngeal

colonization by vaccine-type strains, an indication of

potential for herd immunity.

One method used by ABCs is to collect available iso-

lates from identified cases. Serotyping data were analyzed

to learn about the epidemiology of S. pneumoniae in the

pre-conjugate vaccine era and to predict the potential

impact of the conjugate vaccine (17). Of pneumococcal

cases identified by ABCs from 1995 to 1998, at least 82%

in children <2 years of age were caused by serotypes

included in the 7-valent pneumococcal conjugate vaccine.

These population-based ABCs data were used to formulate

the original pneumococcal conjugate vaccine schedules

and provide recommendations for administering the vac-

cine to infants and children. When a vaccine shortage

became evident in 2001, ABCs data were again used by

public health officials to weigh alternative strategies for

delivering available doses (18). Surveillance is now

focused on evaluating changes in disease impact after the

conjugate vaccine was introduced, including whether it

interrupts transmission of antibiotic-resistant pneumococ-

ci. Analysis of ABCs data shows a substantial decline in

disease caused by serotypes in the vaccine formulation

among children in the age group for whom the vaccine is

recommended. More modest declines also occur in select-

ed adult groups (19).

ABCs will continue to evaluate the impact of the

recently introduced pneumococcal conjugate vaccine,

including whether vaccine shortages have slowed the ini-

tial steep decline in disease occurrence. Other goals are

measurement of vaccine efficacy, assessment of whether

the vaccine is interrupting transmission, and evaluation of

the distribution of serotypes causing disease (to determine

if decline in disease because of serotypes included in the

vaccine has been counterbalanced by emergence of inva-

sive disease caused by nonvaccine serotypes). While this

“replacement disease” phenomenon was recognized for

otitis media and colonization in the prelicensure vaccine

trials, no evidence of replacement invasive disease has

thus far been recognized.

Clinicians were challenged by the emergence of mul-

tidrug-resistant pneumococci during the 1990s, when new

treatment guidelines were developed for meningitis, otitis

media, and pneumonia (20). Vaccines, in concert with

campaigns to promote appropriate use of antibiotics, pro-

vide opportunities to transform the problem of drug-resist-

ant pneumococci from a treatment dilemma to a prevention

success story (21).

Revised Recommendations for Preventing Perinatal

Group B Streptococcal Disease

Data developed through ABCs provided a basis for

revising recommendations for the prevention of perinatal

group B streptococcal (GBS) disease. Since its emergence

in the 1970s, GBS disease has been the leading invasive

bacterial infection associated with illness and death among

newborns in the United States. Surviving infants may have

long-term developmental disabilities, such as mental retar-

dation or hearing and vision loss. Newborns at increased

risk for GBS disease are those born to women who are col-

onized with GBS in the genital or rectal areas. Although

the use of intrapartum prophylaxis has led to a 70%

decline in the incidence of GBS disease during the 1990s

(Figure 2) (22,23), early-onset GBS disease (in infants <7

days old) remains a leading cause of illness and death

among newborns. Guidelines issued in 1996 recommend-

ed either screening pregnant women for GBS colonization

by means of prenatal cultures (screening approach) or

assessing obstetric risk factors intrapartum (risk-based

approach) to identify candidates for intrapartum antibiotic

prophylaxis.

An EIP population-based, retrospective cohort study

compared the effectiveness of prenatal screening for GBS

with the risk-based approach for preventing early-onset

GBS sepsis (24). The analysis, which combined ABCs

population-based active surveillance data on GBS cases

with a sample survey representing >600,000 deliveries,

showed that infants born to women who had been screened

for GBS before delivering had less than half the risk for

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 783

SYNOPSES

early-onset GBS compared to infants of women who had

not been screened, after adjustments were made for poten-

tial confounders. The protective effect of the screening

approach resulted mainly from broader coverage of the

population at risk because many early-onset GBS cases in

the preprevention era occurred in GBS-colonized women

without obstetric risk factors. The evidence for updated

prevention recommendations from key health organiza-

tions (i.e., American College of Obstetricians and

Gynecologists, American Academy of Pediatrics,

American College of Nurse-Midwives, and CDC) was

based on the finding that routine screening for GBS during

pregnancy more effectively prevents cases of early-onset

disease than the risk-based approach (25). Through ABCs,

CDC will continue to monitor GBS disease trends to

understand the impact of the new recommendations and

detect potential adverse consequences of intrapartum

antibiotic use such as emergence of sepsis caused by other

organisms or new patterns of antimicrobial resistance

(26,27).

Decrease in Bacterial Foodborne Diseases

FoodNet documented a decrease in bacterial foodborne

illnesses from 1996 to 2001. Many infections are transmit-

ted through food and can cause illness ranging from mild

gastroenteritis to severe illness requiring hospitalization.

Foodborne pathogens cause an estimated 76 million ill-

nesses, 325,000 hospitalizations, and 5,000 deaths in the

United States each year (11). Clinicians treating patients

with acute gastroenteritis are principally focused on

whether empiric antimicrobial agents are warranted and

the value of diagnostic evaluation. However, the task of

providing accurate information on trends in specific food-

borne pathogens capable of causing this syndrome, as well

as probable sources of infection, has historically fallen to

public health authorities.

Data from FoodNet documented recent declines in the

occurrence of several major bacterial foodborne illnesses

(9,28); preliminary surveillance data for 2001 were com-

pared with 1996–2000 data (28). Significant declines

occurred in major bacterial foodborne illnesses, including

infections caused by Yersinia (49%), Listeria (35%),

Campylobacter (27%), and Salmonella (15%) (Figure 3).

The combined estimated incidence of infections caused by

Listeria, Campylobacter, Salmonella, and E. coli O157 in

2001 was 21% lower than in 1996, on the basis of a multi-

variate regression model.

The factors influencing the occurrence of foodborne ill-

nesses are complex. However, the observed declines in

foodborne disease incidence did occur in the context of

several control measures, including the U.S. Department

of Agriculture’s Food Safety Inspection Service’s imple-

mentation of the Pathogen Reduction/Hazard Analysis and

Critical Control Point regulations in meat and poultry

slaughter and processing plants, egg-quality assurance pro-

grams for Salmonella Enteritidis, and increased consumer

education in food safety (28).

FoodNet will continue to monitor the occurrence of

foodborne diseases. In 2003, FoodNet will also conduct

studies of the consequences of and risk factors for illness

caused by S. Enteritidis, S. Newport, and illness in infants

caused by Campylobacter and Salmonella. Other activities

include a project to improve collection and transport of

specimens during outbreaks so that a cause is identified in

a higher percentage of outbreaks.

Rapid identification of a cause for cases of infectious

diarrhea and appropriate reporting of cases of foodborne

illnesses to state or local public health authorities are

important not only in identifying and controlling outbreaks

but also for more precise assessments of the local, region-

784 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

Figure 2. Incidence of early- and late-onset invasive group B strep-

tococcal disease in three active surveillance areas (California,

Georgia, and Tennessee), 1990–1998, and activities for the pre-

vention of group B streptococcal disease (22). CDC, Centers for

Disease Control and Prevention. ACOG, American College of

Obstetricians and Gynecologists; AAP, American Academy of

Pediatrics.

Figure 3. Relative rates compared with 1996, adjusted for sites, of

laboratory-diagnosed cases of Campylobacter, Salmonella, and

Shigella, by year, FoodNet, United States, 1996–2001 (28).

Bacterial pathogens with highest incidences of the 10 studied dis-

eases are shown.

al, and national trends in foodborne illnesses (29). In turn,

such estimates can inform clinicians of likely causes, prob-

able sources, and prognostic factors for episodes of illness

in persons under their care.

Unexplained Deaths and Critical Illnesses Project

Many clinicians have treated patients with puzzling sit-

uations, in which the acute onset of a critical illness sug-

gestive of an infectious origin occurred in otherwise

healthy young people for whom diagnostic tests failed to

identify an etiologic agent. Occasionally, such episodes are

retrospectively diagnosed many years later with the recog-

nition of a new infectious disease and testing of stored

clinical specimens. For example, hantavirus pulmonary

syndrome was first recognized and described in the United

States in 1993 by an alert clinician during an outbreak in

the Southwest (30); retrospective reviews of fatal illnesses

showed that unrecognized cases of hantavirus pulmonary

syndrome had preceded the 1993 outbreak by at least 15

years (31). Similarly, cases of legionellosis and AIDS were

recognized in hindsight years after they had occurred (13).

These observations, coupled with the new laboratory tech-

niques for pathogen identification, particularly methods

that do not rely on culture, suggested that an effort to

prospectively identify pathogens causing unexplained syn-

dromes might yield useful information (12,13); this was

the beginning of the UNEX project. Laboratory evaluation

of cases includes traditional serologic and in vitro culture

diagnostic methods as well as molecular techniques. This

combined epidemiologic and laboratory approach is a hall-

mark feature of other EIP projects that study hepatitis,

acute respiratory diseases, and encephalitis (32).

The UNEX project has developed methods for evaluat-

ing severe syndromes indicating infection, including non-

culture-based methods to identify etiologic agents. From

May 1, 1995, to December 31, 1998, 137 illnesses meeting

the UNEX case definition were reported to participating

EIPs. After adjustments for age and race were made, this

number translates to an estimated 920 U.S. cases per year;

the overall annual incidence rates did not change during

this time. No differences were observed in the seasonal

distribution of cases of unexplained illnesses, nor did cases

cluster by time or place. The largest proportion of cases

was treated as a neurologic syndrome (29%), followed by

respiratory (27%) and cardiac (21%) syndromes.

Diagnostic testing through UNEX identified a cause in 34

(28%) of 122 cases from which specimens were available

(Table 2).

Two recent outbreaks demonstrate the usefulness of the

approach developed for UNEX. During a 1999 outbreak of

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 785

SYNOPSES

Table 2. Infectious causes and explanations for unexplained deaths and critical illnesses cases, 1995–1998, California, Oregon,

Connecticut, and Minnesota (n=34)

a,b

Syndrome

Etiologic agent (n)

Tests (n)

Neisseria meningitidis (4)

16S rDNA PCR (2), PCR (1), EIA IgM (1)

Bartonella hensaelae (1)

PCR, IFA, IgG

Bartonella spp. (2)

IFA, IgG

Chlamydia pneumoniae (1)

MIF, IgG

Mycoplasma pneumoniae (1)

EIA, IgM/IgG

Cytomegalovirus (1)

EIA and IFA,IgG

Coxsackie B virus (1)

EIA, IgM, viral culture

Enterovirus (1)

EIA, IgM

Epstein-Barr virus (1)

IFA, IgG (VCA and EA)

Human herpesvirus 6 (1)

IFA and EIA (IgM and IgG)

Neurologic (n=15)

Mumps virus (1)

IFA IgM, IFA and EIA, IgG

Chlamydia pneumoniae (2)

MIF IgG (2), IFA, IgM

Mycoplasma pneumoniae (4)

PCR (blood), EIA, IgM/IgG

Streptococcus pneumoniae (2)

16S rDNA PCR (pleural fluid)

Legionella spp. (1)

PCR (from lung)

Adenovirus (1)

EIA and IFA, IgG

Influenza B virus (1)

EIA and IFA, IgG

Influenza A virus (1)

EIA and IFA, IgM, EIA (IgG)

Respiratory (n=13)

Human parainfluenza virus types 1 and 3 (1)

EIA and IFA, IgG

Borrelia burgdorferi/Ehrlichia chaffeensis (1)

EIA/IFA flagella, IgG, Western blot (IgG and IgM)

Enterovirus (1)

EIA IgM

Cardiac (n=3)

Legionella spp. (1)

PCR (heart)

Neisseria meningitidis (1)

PCR (cerebrospinal fluid)

Adenovirus (1)

PCR (blood)

Multisystem (n=3)

Enterovirus (1)

IgM, EIA

PCR, polymerase chain reaction; EIA, enzyme immunosorbent assay; IFA, indirect immunofluorescent assay; Ig, immunoglobulin; EA,

early antigens; VCA, viral capsid

antigens; MIF, microimmunofluorescence.

Reference 12.

West Nile encephalitis in the northeastern United States,

which was recognized by an alert clinician (33), and dur-

ing an outbreak of unexplained illness among injecting

drug users in Scotland and Ireland (34), initial reports of

illness were received and initial laboratory testing per-

formed through the laboratory infrastructure established

for the UNEX project.

The frequency and distribution of the syndromes identi-

fied through this project undoubtedly reflect both the dis-

tribution of their occurrence and gaps in our ability to diag-

nose causes of neurologic and respiratory syndromes in

particular. Although novel pathogens have not yet been dis-

covered through the UNEX project, this systematic

approach improves chances of recognizing infectious dis-

ease causes earlier than in the past and lays the groundwork

for the development of improved diagnostic tools.

Moreover, concerns about bioterrorism have put a premi-

um on the early detection of an intentional release or infec-

tious or chemical agents; this syndrome-based surveil-

lance, which seeks early identification and diagnosis, can

contribute to public health preparedness for such events.

Future Directions of EIPs

Since the release of the plan that launched the EIPs,

these programs have made substantial contributions to the

practice of U.S. public health. Using domestic EIPs as a

model, CDC has begun developing a network of interna-

tional EIPs (IEIPs) in collaboration with Ministries of

Health and other international partners. The first IEIP was

established in Thailand during 2001, and a second IEIP is

being established in Kenya. Collaborations between EIPs

and IEIPs will provide valuable opportunities for training.

In addition, the new U.S. EIP in New Mexico will feature

work along the U.S.-Mexico border and also promises to

enhance international collaborations.

Opportunities presented by new laboratory and infor-

mation technologies, as well as challenges posed by poten-

tial bioterrorism, will influence the evolution of the EIPs

over the next several years. EIP work will build on experi-

ence gained through the combined epidemiologic and lab-

oratory evaluation of syndromes to enhance bioterrorism

preparedness and develop the capacity for identifying pre-

viously unrecognized pathogens. However, even as new

technologies are found, knowledgeable and engaged clini-

cians will remain a vital element in efforts to detect,

respond to, and prevent emerging infectious diseases.

Acknowledgments

We thank Cynthia Whitney for providing information about

the pneumococcal conjugate vaccine and pneumococcal disease

and Matthew Moore for providing information about foodborne

disease.

Dr. Pinner is the director of the Office of Surveillance,

National Center for Infectious Diseases, Centers for Disease

Control and Prevention. His scientific interests are public health

surveillance, especially for infectious diseases; trends in deaths

attributed to infectious diseases; and the epidemiology of several

bacterial and fungal diseases.

References

1. Gerberding JL, Hughes JM, Koplan JP. Bioterrorism preparedness

and response: clinicians and public health agencies as essential part-

ners. JAMA 2002;287:898–90.

2. Petersen LR, Marfin AA. West Nile virus: a primer for the clinician.

Ann Intern Med 2002;137:173–9.

3. Centers for Disease Control and Prevention. West Nile virus activi-

ty—United States, September 26–October 2, 2002, and investigations

of West Nile virus infections in recipients of blood transfusion and

organ transplantation. MMWR Morb Mortal Wkly Rep

2002;51:884,895.

4. Institute of Medicine. Emerging infections: microbial threats to

health in the United States. Washington: National Academy Press;

1992.

5. Centers for Disease Control and Prevention. Addressing emerging

infectious disease threats: a prevention strategy for the United States.

Atlanta: U.S. Department of Health and Human Services, Public

Health Service; 1994.

6. Centers for Disease Control and Prevention. Preventing emerging

infectious diseases: a strategy for the 21st Century. Atlanta: U.S.

Department of Health and Human Services, Public Health Service;

1998.

7. Hardnett FP, Hoekstra RM, Kennedy MH, Angulo FJ, EIP Foodnet

Working Group. Comparability of FoodNet and United States popu-

lations. Poster presented at: International Conference on Emerging

Infectious Diseases; 24–27 March 2002; Atlanta, Ga.

8. Schuchat A, Hilger T, Zell E, Farley MM, Reingold A, Harrison L, et

al. Active bacterial core surveillance of the emerging infections pro-

gram network. Emerg Infect Dis 2001;7:92–9.

9. Centers for Disease Control and Prevention. Foodborne diseases

active surveillance network. MMWR Morb Mortal Wkly Rep

1997;46:258–61.

10. Herikstad H, Yang S, Van Gilder TJ, Vugia D, Hadler J, Blake P, et al.

A population-based estimate of the burden of diarrhoeal illness in the

United States: FoodNet, 1996–7. Epidemiol Infect 2002;129:9–17.

11. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, et

al. Food-related illness and death in the United States. Emerg Infect

Dis 1999;5:607–25.

12. Hajjeh RA, Relman D, Cieslak PR, Sofair AN, Passaro D, Flood J, et

al. Surveillance for unexplained deaths and critical illnesses due to

possibly infectious cases, United States, 1995–1998. Emerg Infect

Dis 2002;8:145–53.

13. Perkins BA, Flood JM, Danila R, Holman RC, Reingold AL, Klug

LA, et al. Unexplained deaths due to possibly infectious causes in the

United States: defining the problem and designing surveillance and

laboratory approaches. Emerg Infect Dis 1996;2:47–53.

14. Ijdo JW, Meek JI, Cartter ML, Magnarelli LA, Wu C, Tenuta SW, et

al. Emergence of another tickborne infection in the 12-town area

around Lyme, CT: human ehrlichiosis. J Infect Dis

2000;181:1388–93.

15. Centers for Disease Control and Prevention. Surveillance for

Creutzfeldt-Jakob disease—United States. MMWR Morb Mortal

Wkly Rep 1996;45:665–8.

16. Centers for Disease Control and Prevention. Preventing pneumococ-

cal disease among infants and young children: recommendations of

the Advisory Committee on Immunization Practices (ACIP). MMWR

Morb Mortal Wkly Rep 2000;49(RR-9):1–35.

786 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

SYNOPSES

17. Robinson KA, Baughman W, Rothrock G, Barrett NL, Pass M, Lexau

C, et al. Epidemiology of invasive Streptococcus pneumoniae infec-

tions in the United States, 1995–1998: opportunities for prevention in

the conjugate vaccine era. JAMA 2001;285:1729–35.

18. Centers for Disease Control and Prevention. Notice to readers:

decreasing availability of pneumococcal conjugate vaccine. MMWR

Morb Mortal Wkly Rep 2001;50:783.

19. Whitney CG, Farley MM, Hadler J, Harrison LH, Bennet NM,

Lynfield R, et al. Decline in invasive pneumococcal disease after the

introduction of protein-polysaccharide conjugate vaccine. N Engl J

Med 2003;348;1737–46.

20. Whitney CG, Farley MM, Hadler J, Harrison LH, Lexau C, Reingold

A, et al. Increasing prevalence of multidrug-resistant Streptococcus

pneumoniae in the United States. N Engl J Med 2000;343:1917–24.

21. Heffelfinger JD, Dowell SF, Jorgensen JH, Klugman KP, Mabry LR,

Musher DM, et al. Management of community-acquired pneumonia

in the era of pneumococcal resistance: a report from the drug-resist-

ant Pneumococcus pneumonia therapeutic working group. Arch

Intern Med 2000;160:1399–408.

22. Schrag SJ, Zywicki S, Farley M, Reingold A, Harrison L, Lefkowitz

L, et al. Group B streptococcal disease in the era of intrapartum

antibiotic prophylaxis: population-based surveillance in the United

States, 1993–1998. N Engl J Med 2000;342:15–20.

23. Centers for Disease Control and Prevention. Early-onset group B

streptococcal disease–United States, 1998–1999. MMWR Morb

Mortal Wkly Rep 2000;49:793–6.

24. Schrag SJ, Zell ER, Lynfield R, Roome A, Arnold KE, Craig AS, et

al. A population-based comparison of strategies to prevent early-onset

group B streptococcal disease in neonates. N Engl J Med

2002;347:233–9.

25. Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of peri-

natal group B streptococcal disease. Revised guidelines from the

CDC. MMWR Morb Mortal Wkly Rep 2002;51(RR-11):1–22.

26. Hyde TB, Hilger TM, Reingold A, Farley MM, O’Brien KL,

Schuchat A. Trends in the incidence and antimicrobial resistance of

early-onset sepsis: population-based surveillance in San Francisco

and Atlanta. Pediatrics 2002;110:690–5.

27. Baltimore RS, Huie SM, Meek JI, Schuchat A, O’Brien KL. Early-

onset neonatal sepsis in the era of group B streptococcal prevention.

Pediatrics 2001:108:1094–8.

28. Centers for Disease Control and Prevention. Preliminary FoodNet

data on the incidence of foodborne illnesses—select sites, United

States, 2001. MMWR Morb Mortal Wkly Rep 2002;51:325–9.

29. Centers for Disease Control and Prevention. Diagnosis and manage-

ment of foodborne illnesses. MMWR Morb Mortal Wkly Rep

2001;50(RR-2):1–69.

30. Duchin JS, Koster FT, Peters CJ, Simpson GL, Tempest B, Zaki SR,

et al. Hantavirus pulmonary syndrome: a clinical description of 17

patients with a newly recognized disease. N Engl J Med

1994;331:546–8.

31. Zaki SR, Khan AS, Goodman RA, Armstrong LR, Greer PW,

Coffield LM, et al. Retrospective diagnosis of hantavirus pulmonary

syndrome, 1978–1993: implications for emerging infectious diseases.

Arch Pathol Lab Med 1996;120:134–9.

32. Park SY, Glaser C, Murray WJ, Kazacos KR, Rowley HA, Frederick

DR, et al. Raccoon roundworm (Baylisascaris procyonis) encephali-

tis: case report and field investigation. Pediatrics 2000;106:e56.

33. Asnis DS, Conetta R, Teixeira AA, Waldman G, Sampson BA. The

West Nile virus outbreak of 1999 in New York: the Flushing Hospital

experience. Clin Infect Dis 2000;30:413–8.

34. Centers for Disease Control and Prevention. Update: Clostridium

novyi and unexplained illness among injecting drug users—Scotland,

Ireland, and England, April–June 2000. MMWR Morb Mortal Wkly

Rep 2000;49:543–5.

Address for correspondence: Robert W. Pinner, Centers for Disease

Control and Prevention, 1600 Clifton Rd., Mailstop D59, Atlanta, GA

30333, USA; fax: 404-371-5445; email: [email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 787

SYNOPSES

All material published in Emerging Infectious Diseases is in the

public domain and may be used and reprinted without special per-

mission; proper citation, however, is appreciated.

Acute weakness associated with West Nile virus

(WNV) infection has previously been attributed to a periph-

eral demyelinating process (Guillain-Barré syndrome);

however, the exact etiology of this acute flaccid paralysis

has not been systematically assessed. To thoroughly

describe the clinical, laboratory, and electrodiagnostic fea-

tures of this paralysis syndrome, we evaluated acute flac-

cid paralysis that developed in seven patients in the setting

of acute WNV infection, consecutively identified in four hos-

pitals in St. Tammany Parish and New Orleans, Louisiana,

and Jackson, Mississippi. All patients had acute onset of

asymmetric weakness and areflexia but no sensory abnor-

malities. Clinical and electrodiagnostic data suggested the

involvement of spinal anterior horn cells, resulting in a

poliomyelitis-like syndrome. In areas in which transmission

is occurring, WNV infection should be considered in

patients with acute flaccid paralysis. Recognition that such

weakness may be of spinal origin may prevent inappropri-

ate treatment and diagnostic testing.

ost human infections with West Nile virus (WNV), a

flavivirus within the Japanese encephalitis virus

antigenic complex, are clinically inapparent (1,2). Mild

febrile illness develops in approximately 1 in 5 infected

persons; more severe neurologic disease, mostly meningi-

tis or encephalitis, occurs in and 1 in 150 (1–4). Less fre-

quently, acute WNV infection has been associated with

acute flaccid paralysis, which has been attributed to

Guillain-Barré syndrome, motor axonopathy, or axonal

polyneuropathy (4–6). However, these reports describe

clinical and laboratory features that seem inconsistent with

such diagnoses, and the exact cause of acute flaccid paral-

ysis has not been thoroughly assessed with rigorous elec-

trophysiologic, laboratory, and neuroimaging data. Brief

descriptions of six patients have suggested that this flaccid

paralysis is due to anterior horn cell involvement with a

resultant poliomyelitis-like syndrome (7–9). Because

understanding the clinical characteristics and underlying

etiology of WNV-induced acute flaccid paralysis is critical

for therapeutic decisions as well as prognosis, we describe

the detailed clinical, laboratory, and electrophysiologic

findings from these six patients and from one additional

patient.

Patients and Methods

Seven patients were detected through WNV surveil-

lance conducted by the Mississippi Department of Health

and the Louisiana Office of Public Health. For each

patient, a standardized questionnaire, including demo-

graphics, medical history, initial signs and symptoms, risk

factors, and treatment, was completed; a standardized neu-

rologic examination was performed by a single neurologist

(JJS). Electrodiagnostic studies were performed by neurol-

ogists (AAL and JAVG) specializing in electrodiagnostic

medicine.

Cerebrospinal fluid (CSF) and acute- or convalescent-

phase serum specimens (or both) from each patient were

tested for antibody to WNV by immunoglobulin (Ig) M

antibody-capture enzyme immunoassay (10) or plaque

reduction neutralization assay (11). The initial specimen

for one patient (patient 5, Table 1) was tested with a slight-

ly modified IgM antibody assay at a commercial laborato-

ry (12). IgM assays were considered positive if the optical

density ratio of the patient and negative control samples

(P/N ratio) was greater than three. For patient samples, a

P/N ratio for WNV at least three times that for St. Louis

encephalitis virus indicated WNV infection (13). A plaque

reduction neutralization test result of at least 10 was con-

sidered positive.

All seven patients had serologic evidence of WNV infec-

tion (Table 1). On the basis of serologic data, three of the

patients were classified as confirmed case-patients (patients

4, 6, and 7) and four as probable case-patients (patients 1–3,

5), according to the national case definition (14).

Case 1

On July 1, 2002, a previously healthy, 56-year-old,

male Mississippi resident was hospitalized with a 1-week

788 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Acute Flaccid Paralysis and West

Nile Virus Infection

James J. Sejvar,* A. Arturo Leis,† Dobrivoje S. Stokic,† Jay A. Van Gerpen,‡

Anthony A. Marfin,* Risa Webb,§ Maryam B. Haddad,* Bruce C. Tierney,* Sally A. Slavinski,§

Jo Lynn Polk,† Victor Dostrow,† Michael Winkelmann,† and Lyle R. Petersen*

*Centers for Disease Control and Prevention, Atlanta, Georgia,

USA; †Methodist Rehabilitation Center, Jackson, Mississippi,

USA; ‡Ochsner Clinic, New Orleans, Louisiana, USA; and

§Mississippi State Department of Health, Jackson, Mississippi,

USA

history of fever, chills, night sweats, myalgias, and acute

encephalopathy. Neurologic examination showed pro-

found weakness in both arms, asymmetric weakness in the

legs with a right foot drop, and acute respiratory distress

(Table 2). Sensory test results were normal. Although com-

puted tomography (CT) and magnetic resonance imaging

(MRI) of the brain showed normal results, heparin was

administered for suspected evolving stroke. Admission

laboratory values (Table 3) showed serum leukocytosis,

and cerebrospinal fluid (CSF) obtained on day 3 showed

elevated protein. By day 8, Guillain-Barré syndrome was

suspected, and intravenous immunoglobulin (IVIG) was

administered. Electrodiagnostic studies performed that day

were interpreted as showing a proximal neuropathy or

myopathy. A deltoid muscle biopsy for suspected inflam-

matory myopathy showed mild type 2 fiber atrophy but no

obvious necrosis or marked inflammatory response. On

day 30, the patient was transferred for rehabilitation with

flaccid, areflexic paralysis in the right leg and variable

weakness and diminished reflexes in all other limbs. Neck

flexors were normal in strength, and sensation was pre-

served in all limbs; loss of bladder function was evident.

MRI of the cervical spine showed normal results.

Electrodiagnostic studies showed widespread but variable

denervation, reduced compound muscle action potentials

(CMAPs), and normal sensory nerve action potentials

(SNAPs), consistent with a severe, asymmetric process

affecting anterior horn cells or motor axons. Myopathy,

demyelinating polyneuropathy, and diffuse axonal

polyneuropathy were not evident.

Case 2

On July 15, 2002, a 57-year-old male Mississippi resi-

dent with a remote history of prostate cancer and glucose

intolerance was hospitalized with a 3-day history of fever,

chills, nausea, vomiting, and headache. Neurologic exam-

ination showed encephalopathy and asymmetric weakness

in all limbs (Table 2). Results of a brain MRI weree nor-

mal. Admission laboratory studies showed a CSF pleocy-

tosis with elevated protein (Table 3). On day 5, acute res-

piratory distress developed, and the patient required

mechanical ventilation. Upon extubation 2 weeks later, the

patient had continued extremity weakness and was aspirat-

ing fluids; a modified barium swallow study showed

oropharyngeal dysphagia. Upon transfer to a rehabilitation

center on day 30, the patient had asymmetric weakness in

the legs and right arm and moderate weakness in neck flex-

ors and facial muscles. Hypotonia and areflexia were noted

in all limbs. Sensation was slightly diminished to vibration

and proprioception in toes bilaterally but preserved to light

touch, pinprick, and temperature. Sensory testing was nor-

mal in the upper limbs. Urinary incontinence was noted.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 789

RESEARCH

Table 1. Serologic results for West Nile virus (WNV)–specific antibodies in patients with acute flaccid paralysis associated with acute

WNV infection

IgM-capture enzyme immunoassay

Plaque reduction neutralization assay

Case no.

Onset

Collection

Sample

SLEV

WNV

SLEV

WNV

6/24

7/12

Serum

3.5

22.3

320

5,120

7/12

7/16

Serum

8.0

22.7

1,280

7/26

8/1

Serum

2.79

24.9

<10

640

7/29

8/3

Serum

1.1

14.1

<10

7/29

8/3

CSF

3.3

39.2

7/29

8/13

Serum

4.4

23.5

2,560

8/11

8/15

Serum

2.02

8/11

8/29

Serum

3.4

25.7

8/13

8/16

CSF

6.1

23.8

8/13

8/16

Serum

1.0

5.7

<10

9/1

10/24

Serum

2.8

10.6

320

9/1

9/6

CSF

Not performed

7.4

IgM, immunoglobulin M; SLEV, Saint Louis encephalitis virus; CSF, cerebrospinal fluid.

Table 2. Initial clinical signs and symptoms in patients with acute flaccid paralysis associated with acute West Nile virus infection

Case no.

Fever (>38.5°C)

Headache

Nuchal rigidity

Altered mental status

Tremor

Distribution of weakness

Upper and lower limbs, R > L

Lower limbs, R > L

R upper limb

Lower limbs, R > L

Upper and lower limbs, L > R;

bulbar muscles

R, right; L, left.

Electrodiagnostic studies showed widespread denervation,

reduced CMAP amplitudes in all nerves of the lower limbs

and right upper limb, and normal SNAP responses, consis-

tent with a severe, asymmetric process affecting anterior

horn cells or motor axons. Myopathy, demyelinating

polyneuropathy, and diffuse axonal polyneuropathy were

not apparent.

Case 3

On July 24, 2002, a low-grade fever, nausea, and vom-

iting, followed by shaking chills and sweats, developed in

a 56-year-old male Louisiana resident with a history of

hypertension and coronary artery disease. The next day,

asymmetric weakness developed in the lower extremities,

with no pain or numbness. Upper extremities were normal.

No bowel or bladder dysfunction was present. The patient

was hospitalized on July 29, and neurologic examination

showed a flaccid, areflexic right lower extremity and a

weak left lower extremity with diminished reflexes.

Results of strength and reflex testing of the upper extrem-

ities were normal. Sensory examination results were nor-

mal except for a mild decrease in sensitivity to pinprick,

temperature, touch, and vibration in a stocking-and-glove

distribution (i.e., distal arms and legs). A coarse bilateral

upper extremity action tremor was noted. The patient had

no headache, neck stiffness, or alteration of mental status

(Table 2). Admission laboratory values showed leukocyto-

sis and CSF pleocytosis (Table 3). Results of other diag-

nostic tests were unremarkable. Postviral demyelination

syndrome and viral-induced polyradicultis were consid-

ered, and IVIG, dexamethasone, and antibacterial and

antiviral medications were administered without patient

improvement. On day 15, the patient was discharged to a

skilled nursing facility for rehabilitation.

MRI of the cervical, thoracic, and lumbosacral spine

obtained during rehabilitation was notable for showing

mild cervical and lumbosacral spinal stenosis and forami-

nal restrictions from C3 through C7 and homogeneous

enhancement of the nerve roots of the cauda equina consis-

tent with meningitis. Electrodiagnostic studies showed

denervation in thoracic and lumbosacral myotomes, with

no muscle activation in the right leg and reduced muscle

activation in the left leg. CMAPs in the right leg were

absent; SNAPs were normal. Electrodiagnostic findings

suggested a severe, asymmetric process affecting anterior

horn cells or motor axons. Diffuse axonal polyneuropathy

was not evident, despite a slight sensory loss in the distal

extremities.

Case 4

On August 2, 2002, fever, headache, and neck stiffness

developed in a 69-year-old female Louisiana resident with

a history of diabetes and degenerative disc disease; the

next day acute weakness occurred in the right arm without

pain, numbness, or paresthesias. She was hospitalized on

August 4. On admission, physical examination document-

ed fever, vomiting, encephalopathy, nuchal rigidity, and a

bilateral rash on the lower extremities. Neurologic exam-

ination displayed a flaccid and areflexic right arm. Her

legs and left arm exhibited normal strength, reflexes, and

coordination, with normal sensation in all limbs. A coarse

tremor was noted in the chin, left arm, and legs (Table 2).

Laboratory findings included CSF pleocytosis (Table 3).

Differential diagnoses included meningoencephalitis with

associated motor polyradiculopathy and monoplegia sec-

ondary to stroke. The patient was treated with antibacter-

ial and antiviral medications. Results of CT and MRI of

the brain were normal. MRI of the cervical spine showed

multilevel degenerative disc disease. The patient

remained lethargic until day 13, when mental status

abruptly improved; right arm weakness persisted. On day

19, she was transferred to a rehabilitation facility.

Electrodiagnostic studies showed absent CMAPs and pro-

found denervation with no voluntary activation in muscles

of the right arm. Scattered denervation was also seen in

the other three limbs. SNAPs had borderline amplitudes

and conduction velocities bilaterally. The results were

most consistent with a severe, asymmetric process affect-

ing anterior horn cells or motor axons. The patient was

subsequently transferred back to intensive care because

her respiratory function was deteriorating, but she was not

intubated. After Guillain-Barré syndrome was diagnosed,

she was started on IVIG but had no improvement in

weakness.

790 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 3. Initial laboratory findings in patients with acute flaccid paralysis associated with acute West Nile virus infection

Case no.

Leukocytes

(x10

/mm

)

Hematocrit (%)

CSF

WBC (/mm

)

CSF

RBC (/mm

)

CSF

protein (mg/dL)

CSF

glucose (mg/dL)

17.6

38.0

1,778

3.6

38.2

2,600

204

11.8

44.4

140

234

9.5

37.8

143

116

119

7.9

45.6

Not performed

13.0

45.4

329

10.3

Not performed

182

CSF, cerebrospinal fluid; WBC, leukocyte count; RBC, erythrocyte count.

Case 5

On August 11, 2002, severe nausea, vomiting,

headache, and diarrhea in the absence of fever developed

in a 50-year-old male Mississippi resident with a history of

alcohol abuse; the next day, progressive right arm weak-

ness developed. He was hospitalized on August 14.

Neurologic examination on admission showed flaccid

paralysis of the right arm and mild weakness of the right

leg, with normal sensation in all limbs (Table 2).

Laboratory values are shown in Table 3; a lumbar puncture

was not performed. Acute stroke was diagnosed, and the

patient was treated with heparin. Mental status changes,

dysarthria, and dysphagia subsequently developed but

resolved. Upon transfer to a rehabilitation center on day

12, the patient had paralysis and areflexia limited to the

right arm, with normal sensation and diffuse tremor in all

limbs. Brain MRI results were normal; cervical spine MRI

displayed mild multilevel foraminal stenosis on the left.

Electrodiagnostic studies showed markedly reduced motor

responses in the right arm with normal sensory responses,

consistent with a severe asymmetric process affecting

anterior horn cells.

Case 6

On August 16, 2002, a 46-year-old male Louisiana res-

ident with a history of coronary artery disease was hospi-

talized with fever, headache, fatigue, and leg weakness of

3 days’ duration. He reported no nuchal rigidity or mental

status changes, although family members described him as

intermittently confused. Neurologic examination showed a

plegic and areflexic right leg and mild left leg weakness;

sensation was intact throughout. A bilateral tremor of the

upper extremities and jaw was noted (Table 2). Laboratory

abnormalities included a CSF pleocytosis (Table 3). He

was diagnosed with Guillain-Barré syndrome and started

on IVIG. Brain CT and MRI results were normal. Results

of an enhanced MRI of the spine suggested meningitis

involving the conus medullaris and cauda equina.

Electrodiagnostic studies performed on day 4 demonstrat-

ed early denervation and absent activation in muscles of

the right leg and reduced activation of muscles in the right

arm. CMAPs and SNAPs in the right arm and leg were

normal. These findings were consistent with a severe,

asymmetric process affecting anterior horn cells or motor

axons. He was transferred to a rehabilitation facility on day

6 with no improvement of weakness.

Case 7

On September 1, 2002, a previously healthy, 39-year-

old male Louisiana resident had onset of fever, headache,

and nuchal rigidity followed the next day by dysphagia and

bilateral arm and leg weakness that was worse on the left.

He was hospitalized on September 6 for acute respiratory

failure and intubated. Neurologic examination showed

normal cognition, asymmetric flaccid paralysis of the left

arm and leg with absent reflexes, hyporeflexic weakness of

the right arm and leg, and weakness of bulbar muscles

(Table 2). A partial supranuclear gaze palsy, cogwheel

rigidity, and bilateral Babinski signs were also evident.

Admission laboratory findings showed peripheral leukocy-

tosis and CSF pleocytosis (Table 3). Brain MRI showed

increased T2 signal in the periaqueductal gray matter, sub-

stantia nigra, and trigeminal motor nuclei.

Electrodiagnostic studies performed on day 15 showed dif-

fuse denervation in all myotomes, reduced CMAPs (worse

on the left), and preserved SNAPs. On day 25, he was

transferred to a long-term care facility with no improve-

ment of limb weakness.

Discussion

The clinical and electrodiagnostic findings in these

patients with WNV infection suggest involvement of

spinal cord gray matter, specifically anterior horn cells,

and a resulting acute poliomyelitis-like syndrome. All

patients exhibited features typical for polio, including

acute flaccid paralysis without paresthesias or sensory

loss, marked asymmetric weakness, diminished or absent

deep tendon reflexes in the affected limbs, and weakness

that developed during an acute infectious process. Other

typical features of poliomyelitis included CSF pleocytosis

in five of six patients with CSF examination, acute respi-

ratory distress in four, and acute changes in bowel or blad-

der function in two. In addition, electrodiagnostic findings

showed asymmetric muscle denervation, reduced CMAPs,

and preserved SNAPs. No patients had evidence of

demyelinating polyneuropathy or myopathy. The absence

of new sensory abnormalities localizes the disease process

to the anterior horn cells or motor axons. Although muscle

denervation and reduced CMAP amplitudes do not distin-

guish loss of anterior horn cells from loss of motor axons

(15), these patients’ clinical features can be explained only

by anterior horn cell disease, since no known infectious

processes limited to motor axons produce widespread,

asymmetric paralysis without sensory involvement. While

MRI signal abnormalities in the anterior spinal cord have

been noted in patients with poliomyelitis (16,17), these

findings are inconsistent (18,19), and the absence of such

changes in our four patients in which imaging was per-

formed does not preclude a diagnosis of a poliomyelitis-

like syndrome.

Since immunization has eradicated wild-type

poliovirus from the developed world, most cases of para-

lytic polio-like conditions in the United States have been

linked to other RNA viruses, including echoviruses,

enteroviruses, and coxsackieviruses (20). Case reports

have documented a poliomyelitis-type syndrome associat-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 791

RESEARCH

ed with other flaviviruses (21–23), as well as anterior

myelitis associated with WNV infection (24).

The assertion that WNV infection involves anterior

horn cells and causes a polio-type syndrome has a patho-

logic basis. The neuropathology of experimental WNV

infection in monkeys was most pronounced in the cerebel-

lum, medulla, and the cervical and lumbar regions of the

spinal cord (25). Anterior horn cells showed degeneration

and neuronal cell death; conversely, no changes were seen

in the oligodendroglia or peripheral nerves. Similarly,

WNV-infected horses displayed multifocal polioen-

cephalomyelitis, with involvement of the ventral and later-

al horns of the thoracic and lumbar spinal cord (26,27).

WNV antigen was mainly localized within the gray matter

of the spinal cord, with no lesions apparent in peripheral

nerves or ganglia. In WNV-infected birds, lesions and viral

antigen were most prominent in the cerebellum and the

gray matter of the spinal cord (28).

Previous case studies have attributed WNV-associated

acute flaccid paralysis to Guillain-Barré syndrome, motor

axonopathy, or severe axonal polyneuropathy (4–6). The

clinical signs and symptoms and electrodiagnostic findings

reported in those cases, and those described here, are most

consistent with a polio-like condition, and would be atypi-

cal for Guillain-Barré syndrome or other peripheral nerve

disorders. Although acute poliomyelitis and polio-like con-

ditions may occasionally simulate Guillain-Barré syn-

drome (29), our cases had several clinical, laboratory, and

electrodiagnostic features that differed from typical

Guillain-Barré syndrome (30–32; Table 4).

In Guillain-Barré syndrome, electrodiagnostic findings

generally suggest peripheral nerve demyelination or, less

commonly, a combined demyelinating and axonal process

(30,31). The cases reported here displayed reduced or

absent CMAPs with preserved SNAPs, no evidence of

demyelination, a neurogenic pattern of recruitment, and

widespread denervation; combined with the clinical pic-

ture of an asymmetric paralysis, these findings are typical

for a polio-like condition and uncommon for Guillain-

Barré syndrome. A pure axonal variant of Guillain-Barré

syndrome has been described (33) and may be confused

with poliomyelitis and polio-like conditions; however,

such cases are generally characterized by distally promi-

nent weakness and show subclinical sensory nerve

involvement on electrodiagnostic testing. Thus, in the con-

text of WNV infection, electrodiagnostic studies previous-

ly interpreted as motor axonal polyneuropathy or motor

axonopathy without sensory nerve involvement (4–6) are

more suggestive of anterior horn cell loss than of Guillain-

Barré syndrome.

Three of the seven patients had acute flaccid paralysis

without other findings, suggestive of severe central nerv-

ous system involvement caused by WNV infection.

Physicians should suspect WNV infection in patients from

areas where WNV is being transmitted and who have

acute, painless, asymmetric weakness, even if unaccompa-

nied by fever or apparent meningoencephalitis. Diagnostic

studies should include testing for WNV-specific IgM anti-

body in CSF or acute- and convalescent-phase serum sam-

ples. In patients from such areas who have acute flaccid

paralysis, CSF analysis, thorough electrodiagnostic stud-

ies, and spinal imaging should be considered before initi-

ating diagnostic evaluations or therapies directed at

Guillain-Barré syndrome, stroke, inflammatory

myopathies, or other peripheral inflammatory processes.

These therapies are ineffective for polio-like syndromes

and can produce serious sequelae (34–37).

Continued surveillance and investigation of WNV-

infected patients are needed to fully define the scope of

clinical illness and determine the incidence of acute flaccid

paralysis. In addition to assessing clinical outcome, the

identification of risk factors and the pathologic confirma-

tion of anterior horn cell involvement in patients with

WNV-associated acute flaccid paralysis remain important

public health goals.

792 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 4. Clinical characteristics of patients with West Nile virus–associated acute flaccid paralysis compared with patients with typical

Guillain-Barré syndrome (25–27)

Characteristic

West Nile virus–associated flaccid paralysis

Guillain-Barré syndrome

Timing of onset

Acute phase of infection

1–8 weeks after acute infection

Fever and leukocytosis

Present

Absent

Weakness distribution

Asymmetric; occasional monoplegia

Generally symmetric; proximal and distal muscles

Sensory symptoms

Absence of numbness, paresthesias, or

sensory loss; occasional myalgias

Painful distal paresthesias and sensory loss

Bowel/bladder involvement

Often present

Rare

Concurrent encephalopathy

Often present

Absent

CSF profile

Pleocytosis and elevated protein

No pleocytosis; elevated protein

(albuminocytologic dissociation)

Electrodiagnostic features

Anterior horn cell/motor axon: reduced/absent

CMAPs, preserved SNAPs; asymmetric denervation

Demyelination: marked slowing of conduction velocity;

conduction block, temporal dispersion; reduced SNAPs

CSF, cerebrospinal fluid; CMAPs, compound muscle action potentials; SNAPSs, sensory nerve action potentials.

Acknowledgments

We are grateful to Stanley W. Chapman, the Wilson

Research Foundation, Raoult Ratard, Andrea Vicari, Grant L.

Campbell, Michael Bunning, and Susan P. Montgomery for their

important contributions to this investigation.

Dr. Sejvar is a neurologist and epidemiologist with the

Centers for Disease Control and Prevention’s Division of Viral and

Rickettsial Diseases, National Center for Infectious Diseases. His

current areas of research include the epidemiology of encephalitis,

prion diseases, and other infections of the nervous system.

References

1. Tsai T, Popovici F, Cernescu C, Campbell G, Nedelcu N. West Nile

encephalitis epidemic in southeastern Romania. Lancet 1998;

352:767–71.

2. Mostashari F, Bunning M, Kitsutani P, Singer D, Nash D, Cooper M,

et al. Epidemic West Nile encephalitis, New York, 1999: results of a

household-based seroepidemiological survey. Lancet 2001;

358:261–4.

3. Chowers M, Lang R, Nassar F, Giladi D, Rubinshtein E, Itzhaki A, et

al. Clinical characteristics of the West Nile fever outbreak, Israel,

2000. Emerg Infect Dis 2001;7:675–8.

4. Nash D, Mostashari F, Fine A, Miller J, O’Leary D, Murray K, et al.

The outbreak of West Nile virus infection in the New York City area

in 1999. N Engl J Med 2001;344:1807–14.

5. Sampson B, Ambrosi C, Charlot A, Reiber K, Veress J,

Armbrustmacher V, et al. The pathology of human West Nile virus

infection. Hum Pathol 2000;31:527–31.

6. Asnis D, Conetta R, Teixeira A, Waldman G, Sampson B. The West

Nile virus outbreak of 1999 in New York City: the Flushing Hospital

experience. Clin Infect Dis 2000;30:413–8.

7. Leis A, Stokic D, Polk J, Dostrow V, Winkelmann M. A

poliomyelitis-like syndrome from West Nile virus infection. N Engl J

Med 2002;347:1279–80

8. Glass J, Samuels O, Rich M. Poliomyelitis due to West Nile virus. N

Engl J Med 2002;347:1280–1.

9. Centers for Disease Control and Prevention. Acute flaccid paralysis

syndrome associated with West Nile virus infection—Mississippi and

Louisiana, July–August 2002. MMWR Morb Mortal Wkly Rep

2002;51:825–7.

10. Martin D, Muth D, Brown T, Johnson A, Karabatsos N, Roehrig J.

Standardization of immunoglobulin M capture enzyme-linked

immunosorbent assays for routine diagnosis of arboviral infections. J

Clin Microbiol 2000;38:1823–6.

11. Beaty B, Calisher C, Shope R. Arboviruses. In: Lennette E, Lennette

D, and Lennette E, editors. Diagnostic procedures for viral, rick-

ettsial, and chlamydial infections. Washington: American Public

Health Association; 1995. p. 89–212.

12. Prince H, Hogrefe W. Performance characteristics of an in-house

assay system used to detect West Nile virus (WNV)-specific

immunoglobulin M during the 2001 WNV season in the United

States. Clin Diagn Lab Immunol 2003;10:66–9.

13. Martin D, Biggerstaff B, Allen B, Johnson A, Lanciotti R, Roehrig J.

Use of immunoglobulin M cross-reactions in differential diagnosis of

human flaviviral encephalitis infections in the United States. Clin

Diagn Lab Immunol 2002;9:544–9.

14. Petersen L, Marfin A. West Nile virus: a primer for the clinician. Ann

Intern Med 2002;137:173–9.

15. Kimura J. Electrodiagnosis in diseases of nerve and muscle: princi-

ples and practice. 2nd ed. Philadelphia: F.A. Davis Co.; 1989. p.

249–74.

16. Kornreich L, Dagan O, Grunebaum M. MRI in acute poliomyelitis.

Neuroradiology 1996;38:371–2.

17. Rao D, Bateman D. Hyperintensities of the anterior horn cells on

MRI due to poliomyelitis. Journal of Neurology, Neurosurgery, and

Psychiatry 1997;63:720.

18. Huang C, Liu C, Chang Y, Chen C, Wang S, Yeh T, et al. Neurologic

complications in children with enterovirus 71 infection. N Engl J

Med 1999;341:936–42.

19. Ohry A, Karpin H, Yoeli D, Lazari A, Lerman Y. West Nile virus

myelitis. Spinal Cord 2001;39:662–3.

20. Rotbart H. Viral meningitis and the aseptic meningitis syndrome. In:

Scheld W, Whitley R, Durack D, editors. Infections of the central

nervous system. Philadelphia: Lippincot-Raven Publishers; 1997. p.

239–63.

21. Solomon T, Kneen R, Dung N, Khan V, Thuy T, Ha D, et al.

Poliomyelitis-like illness due to Japanese encephalitis virus. Lancet

1998;351:1094–7.

22. Kho L, Sumarmo W, Jahja E, Gubler D. Dengue hemorrhagic fever

accompanied by encephalopathy in Jakarta. Southeast Asian J Trop

Med Public Health 1981;12:83–6.

23. Sumarmo W, Wulur H, Jahja E, Gubler D, Suharyono W, Sorensen K.

Clinical observations on virologically confirmed fatal dengue infec-

tions in Jakarta, Indonesia. Bull World Health Organ

1983;61:693–701.

24. Gadoth N, Weitzman S, Lehmann E. Acute anterior myelitis compli-

cating West Nile fever. Arch Neurol 1979;36:172–3.

25. Manuelidis EE. Neuropathology of experimental West Nile virus

infection in monkeys. J Neuropathol Exp Neurol 1956;15:448–60.

26. Cantile C, Di Guardo G, Eleni C, Arispici M. Clinical and neu-

ropathological features of West Nile virus equine encephalomyelitis

in Italy. Equine Vet J 2000;32:31–5.

27. Cantile C, Del Piero F, Di Guardo G, Arispici M. Pathologic and

immunohistochemical findings in naturally occurring West Nile virus

infection in horses. Vet Pathol 2001;38:414–21.

28. Steele K, Linn M, Schoepp R, Komar N, Geisber T, Manduca R, et al.

Pathology of fatal West Nile virus infection in native and exotic birds

during the 1999 outbreak in New York City, New York. Vet Pathol

2000;37:208–24.

29. Gorson K, Ropper A. Nonpoliovirus poliomyelitis simulating

Guillain-Barré syndrome. Arch Neurol 2001;58:1460–4.

30. Asbury A, Arnason B, Karp H. Criteria for diagnosis of Guillain-

Barré syndrome. Ann Neurol 1978;3:565–6.

31. Weinberg D. AAEM case report #4: Guillain-Barré syndrome.

Muscle Nerve 1999;22:271–81.

32. Hurwitz E, Holman R, Nelson D, Schonberger L, Breman D, Kaslow

R, et al. National surveillance for Guillain-Barré syndrome. January

1978–March 1979. Neurology 1983;33:150–7.

33. Visser L, Van der Meche F, Van Doorn P, Meulstee J, Jacobs B,

Oomes P, et al. Guillain-Barré syndrome without sensory loss (acute

motor neuronopathy): a subgroup with specific clinical, electrodiag-

nostic, and laboratory features. Brain 1995;118:841–7.

34. Norda R, Berseus O, Stemayr B. Adverse events and problems in

therapeutic hemapheresis. A report from the Swedish registry.

Transfusion and Apheresis Sciences 2001;25:33–41.

35. Stangel M, Muller M, Marx P. Adverse events during treatment with

high-dose intravenous immunoglobulins for neurological disorders.

Eur Neurol 1998;40:173–4.

36. Gottlieb S. Intravenous immunoglobulin increases the risk of throm-

botic events. BMJ 2002;324:1056.

37. Struble E, Dice Y. Intravenous immune globulin (IVIG) precipitating

acute myocardial infarction. J Miss State Med Assoc 2002;43:115.

Address for correspondence: James J. Sejvar, Medical Epidemiologist,

Division of Viral and Rickettsial Diseases, National Center for Infectious

Diseases, Centers for Disease Control and Prevention, 1600 Clifton Road,

Mailstop A39, Atlanta, GA 30333, USA; fax: 404-639-3163; email:

[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 793

RESEARCH

Seven alligators were submitted to the Tifton

Veterinary Diagnostic and Investigational Laboratory for

necropsy during two epizootics in the fall of 2001 and 2002.

The alligators were raised in temperature-controlled build-

ings and fed a diet of horsemeat supplemented with vita-

mins and minerals. Histologic findings in the juvenile alliga-

tors were multiorgan necrosis, heterophilic granulomas,

and heterophilic perivasculitis and were most indicative of

septicemia or bacteremia. Histologic findings in a hatchling

alligator were random foci of necrosis in multiple organs

and mononuclear perivascular encephalitis, indicative of a

viral cause. West Nile virus was isolated from submissions

in 2002. Reverse transcription-polymerase chain reaction

(RT-PCR) results on all submitted case samples were pos-

itive for West Nile virus for one of four cases associated

with the 2001 epizootic and three of three cases associat-

ed with the 2002 epizootic. RT-PCR analysis was positive

for West Nile virus in the horsemeat collected during the

2002 outbreak but negative in the horsemeat collected after

the outbreak.

est Nile virus (WNV) has been reported in a variety

of species but primarily endotherms. Arboviruses

have been reported to affect ectotherms, and in some cases

ectotherms are thought to serve as a reservoir (1–4). The

mode of transmission of the arbovirus to ectotherms has

often been presumed to be through ingestion or a bite from

the insect carrier (5).

During the fall of 2001 and 2002, two epizootics

occurred among captive alligators on a south Georgia alli-

gator farm that houses over 10,000 animals.

Approximately 250 alligators died between November and

December 2001, and >1,000 alligators died in 2002. These

epizootics tended to occur approximately 2 weeks after the

first abrupt drop in ambient temperature, which occurred

both years in mid-October and was characterized by mini-

mum temperatures between 0°C and 8°C and maximum

temperatures between 10°C and 18°C for a period of 1 to

3 days.

Methods

Animals and Housing

Animals were housed in six barns that were divided

into 10 pens; each pen contained approximately 100–200

alligators. The nursery animals are obtained either as eggs

from Florida or as hatchlings from onsite breeders.

All pens are cleaned in the morning starting at 6 a.m.

An automatic flushing system is used to drain the pens,

flush them, and fill them with clean water. Well water is

chlorinated with an automated system that injects chloride

gas into the water. The water is then piped into a central

collecting area and heated. The water temperature is main-

tained at 32.2°C year-round, and the buildings are kept

dark to reduce environmental stress on the animals. The

reduced stress and warm environment allow continued

growth (i.e., growth of >

1 m per year rather than 0.30 m

per year).

Alligators are fed in the mid- to late afternoon. The diet

consists of 95% ground raw horsemeat (obtained frozen

from a source in Pennsylvania) to which vitamins and min-

erals are added in a pelleted alligator diet carrier. The

ingredients are thoroughly mixed in a large commercial

mixer. The source of the horsemeat has remained constant

since 1985. The source of the vitamins and minerals has

varied, based upon availability.

The breeding population is maintained in a separate

fenced enclosure on the premises. This enclosure is a

native swampland and therefore subjected to ambient

weather conditions. A rookery was recently established in

the breeding area by native birds. Attempts to depopulate

the rookery (using U.S. Department of Agriculture–

approved methods) have been unsuccessful. The alligators

eat fledglings and older birds that fall from the nests and

branches or otherwise get within reach. Alligators do not

nest under the rookery. No mosquito control is practiced

on the farm.

Tissue Collection

Animals were seen moribund or dead upon arrival at

the laboratory. Blood was collected from the occipital

794 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

West Nile Virus in

Farmed Alligators

Debra L. Miller,* Michael J. Mauel,* Charles Baldwin,* Gary Burtle,* Dallas Ingram,*

Murray E. Hines II,* and Kendal S. Frazier*

*University of Georgia, Tifton, Georgia, USA

sinus or caudal vein of live animals. Gross observations

were made, and the animals were humanely euthanized.

Tissues were collected from the eye, thyroid gland, lymph

node, lung, heart, brain, spinal cord, kidney, liver, spleen,

pancreas, adrenal gland, gallbladder, tonsil, trachea, stom-

ach, intestines, and reproductive tract. Fresh tissue speci-

mens were submitted for virus isolation, reverse transcrip-

tion-polymerase chain reaction (RT-PCR), and bacterial

culture. Tissues were also collected in 10% buffered for-

malin, processed, and embedded in paraffin. Five-microm-

eter-thick sections were stained with hematoxylin and

eosin and viewed by light microscopy. Tissues opportunis-

tically collected from an adult clinically normal, free-rang-

ing alligator served as a control.

Multiple aliquots (totaling 1 g) of the ground raw horse-

meat (without additives) that was being fed during the

2002 epizootic (October and November) were collected

and processed for RT-PCR. Subsequent aliquots from

postepizootic horsemeat shipments (in December and

January) were similarly processed.

Virus Isolation

A 10% homogenate in Earle’s minimal essential media

(MEM) containing gentamicin was made of each speci-

men. The homogenate was centrifuged for 10 min at 2,000

RPM and 4°C. The supernatent was filtered and spread

onto a preformed monolayer of Vero cells. In 2001, fathead

minnow (FHM), white sturgeon skin (WSS), epithelioma

papillosum caprini (EPC), and channel catfish ovary

(CCO) cells were used instead of the Vero cells. Inoculated

cells were incubated in a 5% CO

atmosphere at 37°C.

Cells were examined each day for viral cytopathic effect

(CPE). If no CPE was observed, aliquots of the first pas-

sage were transferred to a second preformed monolayer of

Vero cells (FHM, WSS, EPC, and CCO cells in 2001) on

day 7. If no CPE was observed after a second 7 days of

passage, the culture was considered negative. Monolayers

demonstrating viral CPE were passaged to chambered

slides. The slides were fixed in cold methanol, and a West

Nile fluorescent-antibody test was conducted to confirm

the isolate.

Fluorescent-Antibody Testing

Mouse anti–WNV-specific polyclonal antibody

(Centers for Disease Control and Prevention [CDC],

Division of Vector-Borne Infectious Diseases, Fort

Collins, CO) was applied to the chamber and the slide

incubated in 5% CO

at 37°C for 30 min. The slide was

rinsed two times for 5 min in a sodium carbonate/bicarbon-

ate buffer (pH 9.3). The slide was then air-dried, followed

by an anti-mouse fluorescein-conjugated antibody, and

incubated as before for 30 min. The slide was washed

twice in carbonate buffer, followed by 5 min in 0.5%

Evans blue counter stain. Slides were dipped in distilled

water, and a glycerin/water mounting media and coverslip

was added. Slides were examined with a fluorescent

microscope. All isolates were tested for WNV. All isolates

were also tested for Eastern equine encephalomyelitis

virus (EEEV) by using a similar protocol. We tested for

EEEV because of its known prevalence within the geo-

graphic area. The EEEV-specific monoclonal antibody

(CDC, Atlanta, GA) was prepared against the New Jersey

1960 strain of EEEV.

RNA Extraction

RNA was extracted from various specimens (fresh tis-

sue, virus isolation homogenate or cell culture lysate, and

formalin-fixed paraffin-embedded tissue). For extraction

from fresh specimens, approximately 1 g of tissue was

placed in a whirlpack bag and homogenized by using a

Stomacher Lab Blender 80 (Tekmar Co., Cincinnati, OH)

with three times the tissue volume of phosphate-buffered

saline (PBS). Three milliliters of the tissue homogenate

was processed with a Rneasy Midi kit (QIAGEN, Inc.,

Valencia, CA) per manufacturer’s directions. If a virus iso-

lation homogenate or cell culture lysate in Earle’s MEM

was used, approximately 4 mL of the homogenate or lysate

was washed with 5 mL of PBS, the supernatant removed,

and the pellet processed with the Rneasy Midi kit. For

paraffin sections, several 5-µm sections from paraffin

blocks were cut and deparaffinized with xylene. The

xylene was removed, and samples were washed two times

with 100% ethanol for 10 min, once with 95% and once

with 70% ethanol. Samples were incubated overnight at

56°C in 80 µL of proteinase K with 5 mL of Buffer RLT

from the Rneasy Midi kit and then processed per manufac-

turer’s directions.

RT-PCR

RT-PCR for WNV was performed on the tissues

according to the procedure described by Kuno (6) and

using the RT-nested primer sets described by Johnson et al.

(7). In brief, a RT-PCR mixture was prepared by using the

outside primer set (P1401 – ACCAACTACTGTGGAGTC

and P1845 – TTCCATCTTCACTCTACACT) to amplify a

445-bp product. Forty microliters of the RT-PCR mixture

and 10 µL of sample were dispensed into a 0.2-mL thin

wall PCR tube, and 10 µL of Rnase-free water was added

for a final volume of 50 µL. With the use of a model PTC-

200 thermal cycler (MJ Research, Inc., Waltham,

Massachusetts), cycling conditions for the RT-PCR were

as follows: 53°C for 30 min, followed by 40 cycles of

94°C for 1 min, 53°C for 1 min, 72°C for 1 min, and then

held at 4°C. Ten microliters of RT-PCR first-round product

was used for the nested PCR (nPCR). The nPCR mixture

was prepared by using 40 µL of PCR mixture (now with

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 795

RESEARCH

the inside primer set [P1485 – GCCTTCATACACAC-

TAAAG and P1732 – CCAATGCTATCACAGACT]) to

amplify a 248-bp product. The cycling conditions for the

nPCR were as described above, but the first ramp was

omitted (53°C for 30 min). A 10-µL aliquot of each reac-

tion with 1 µL of loading buffer added was loaded onto a

1.5 % agarose gel in Tris-borate-EDTA (TBE) buffer and

run at 70 V for approximately 1.5 h.

This protocol was repeated on all samples with primer

sets for EEEV and St. Louis encephalitis virus (SLEV).

For the 262-bp EEEV genomic fragment, an outer set of

forward (P4 (EEE-4) - CTAGTTGAGCACAAACACCG-

CA) and reverse (P7 (cEEE-7) - CACTTGCAAGGT-

GTCGTCTGCCCTC) primers, followed by a nested set of

forward (P5 (EEE-5) - AAGTGATGCAAATCCAACTC-

GAC) and reverse (P6 (cEEE-6) - GGAGCCACACG-

GATGTGACACAA) primers, was used (8). The RT-PCR

mixture was similar to that described by Kuno (6). The

thermal cycling parameters varied from those of WNV as

follows: 94°C for 90 s followed by 30 cycles of 94°C for

20 s, 65°C for 35 s, 72°C for 17 s, and then a final elonga-

tion step of 72°C for 4 min. A single RT-PCR procedure

was used for SLEV. The 393-bp genomic fragment was

generated by using forward (SLE727 – GTAGCCGACG-

GTCAATCTCTGTGC) and reverse (SLE119c - ACTCG-

GTAGCCTCCATCTTCATCA) primers and using param-

eters as for WNV (9).

Bacterial Culture

Swabs of individual tissues were streaked onto 5%

bovine blood agar (BBA), Wilkins-Chalgren anaerobe

agar, mycoplasma agar, Lowenstein-Jensen agar slant, and

Hektoen Enteric agar (HE) agar (intestines only). Blood

was inoculated into thioglycolate broth and streaked onto

BBA. Inoculated media were incubated at 30°C with

duplicate blood agar plates incubated in the presence or

absence of 5% CO

, with the exception of the anaerobic

cultures, which were incubated at 37°C. The thioglycolate

broth was subcultured onto BBA after 24 h. Plates were

examined each day for growth and subcultured onto BBA

as needed. Bacterial colonies selected from pure cultures

were Gram stained. Cultures were injected into Sensititre

(Trek Diagnostic Systems, Westlake, OH) gram-negative

AP80 or gram-positive AP90 autoidentification plates and

the antibiotic sensitivity plate CMVIECOF and allowed to

incubate for 18 h at 37°C before automated reading of the

reactions per the manufacturer’s directions. Any isolates

that failed to be identified by the Sensititre system were

identified by using the RapID NF Plus System (Remel,

Norcross, GA) or the API20E system (API Analytab

Products, Plainview, NY).

Results

Clinical Findings

The affected alligators appeared to “star gaze” in the

water just before death, suggesting neurologic lesions (10).

Alligators sometimes became stranded in the dry part of

the pen with loss of leg control and neck spasms. No long-

term signs of stress were noted, and most animals were

eating well until a few days before death. The hatchlings

(approximately 30-cm long at the time of the epizootic)

and juveniles (1–2 m long) seemed to be more severely

affected.

A specific pattern of transmission was not noted in

2001. However, in 2002, the alligator deaths initially

occurred in one building and spread throughout the build-

ing in the opposite direction from that taken to feed and

clean the animals. At least one interruption of chlorine

addition to flush water occurred before the 2002 epizootic.

Deaths were not incurred in the breeding colony, and no

deaths were reported in birds that inhabited the rookery.

Gross Findings

2001

Both Florida and Georgia stock animals were affected,

but, in general, the Florida stock was affected first.

Initially, three juvenile alligators were sent for necropsy

during the 2001 epizootic. In general, the alligators were in

good to excellent body condition. One alligator had

approximately 25 mL of serosanguinous fluid in the peri-

cardial sac and 50 mL yellow serous fluid in the peritoneal

cavity. Two of the three had yellow-tan, caseous necrosis

of the palatine tonsils and multiple caseous yellow-tan

plaques, 2- to 10-mm in diameter, on the mucosal surfaces

of the esophagus, corpus, and pars pylorica. Only scant

ingesta were noted throughout the gastrointestinal (GI)

tract, and the intestinal mucosa was hemorrhagic in rare

instances. The liver and spleen of one alligator had multi-

ple 1- to 3-mm tan foci scattered throughout the parenchy-

ma. One alligator was in poor to moderate body condition

and had scattered bronchiectasis, no ingesta throughout the

GI tract, and mild multifocal serous atrophy of fat. No

other gross lesions were noted.

Approximately 2 months after the 2001 epizootic

began, another juvenile, live alligator was submitted to our

laboratory. The gross lesions were similar to those

described above but with numerous 1- to 3-mm tan foci in

the parenchyma of the liver, spleen, and kidneys.

2002

Three alligators were examined from the fall 2002 epi-

zootic, two juveniles and a hatchling. The two juveniles

had lesions similar to those described in the previous year.

796 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

The liver and kidneys of the hatchling were pale and mot-

tled tan/brown. Ingesta were scant throughout the GI tract.

The free-ranging alligator was in excellent body condition.

No significant gross changes were noted in its tissues.

Light Microscopic Findings

2001

Tissues of the alligators from the 2001 epizootic were

examined and were similar in two of the three alligators. In

the brain, rare glial nodules that contained occasional het-

erophils were present (Figure 1). The spleen was congest-

ed with moderate diffuse reticuloendothelial hyperplasia

and moderate numbers of heterophils. The tonsil had

severe multifocal coalescing areas of caseous necrosis and

heterophilic inflammation with reactive lymphoid follicu-

lar hyperplasia. In the esophagus, a focally extensive,

mixed ulcerative, and proliferative lesion was present; it

had a marked mixed but predominantly mononuclear

inflammation, colonies of bacteria, and extensive fibrin

deposition. In the liver, multifocal lymphoplasmacytic

aggregates and heterophilic granulomas were present, con-

sisting of caseous necrotic foci with degenerate heterophils

surrounded by an outer layer of macrophages, lympho-

cytes, and heterophils. The lungs were congested with mild

diffuse or patchy lymphoplasmacytic and heterophilic

interstitial infiltrates. The kidney had multifocal het-

erophilic granulomas. The pars pylorica region of the

stomach had multifocal mucosal abscesses and moderate

diffuse lymphoplasmacytic and heterophilic infiltrates of

the lamina propria. The small intestine had moderate, dif-

fuse mucosal and submucosal infiltrates of lymphocytes,

heterophils, and plasma cells and multifocal areas of acute

necrosis associated with bacteria. The remaining tissues

appeared within normal limits. Special stains for fungi and

acid-fast bacteria were negative. A population of primarily

gram-negative and fewer gram-positive bacteria was

observed in the heterophilic granulomas.

The third alligator had primarily pulmonary changes.

The airways contained moderate numbers of heterophils,

occasional mucous plugs with degenerate inflammatory

cells, and scattered bacterial colonies. The remaining tis-

sues were as described for the first two alligators.

Tissues from the alligator seen 2 months after the epi-

zootic had similar findings to those of the first two alliga-

tors with the addition of rare, small caseating granulomas

within the lungs. The granulomas contained numerous

large macrophages and multinucleated cells. Acid-fast

stains demonstrated low numbers of slender, beaded, acid-

fast positive bacilli consistent with mycobacteria.

2002

Multiple tissues from the two juvenile alligators from

the 2002 epizootic were examined. The tissue changes

were similar to those described for the 2001 epizootic

except that the inflammatory component was primarily

heterophils. The meninges within the brain and all spinal

cord sections except those from the sacral spinal cord had

stasis of heterophils within the blood vessels and perivas-

cular infiltration of mild numbers of heterophils (Figure 1).

One alligator had a small focus of macrophages and het-

erophils noted within the endocardium.

Multiple tissues were examined from the hatchling alli-

gator, and lesions differed from the previous submissions

on the basis of cellular composition of the inflammatory

cell infiltrates. Lymphoplasmacytic perivascular cuffs were

present throughout the brain and meninges (Figure 1).

Rarely, heterophils were admixed within the cuffs. Similar

changes were not seen within the spinal cord. Random foci

of necrosis were seen within the liver, pancreas, and tonsil.

Mild to moderate perivascular infiltrates of lymphocytes,

plasma cells, and heterophils were seen within the kidney

and heart, and similar but fewer numbers of infiltrates were

seen within the pulmonary interstitium. The heart had mul-

tiple, random foci of patchy vacuolar degeneration of the

myocytes and random aggregates of lymphocytes, plasma

cells and heterophils. Mild numbers of mixed inflammato-

ry cells were seen within the intestinal lamina propria. The

remaining tissues were unremarkable. Major pathologic

changes were not observed by light microscopy in the tis-

sues from the free-ranging alligator.

Virus Isolation/RT-PCR

Virus isolation was negative for all animals from the

2001 epizootic. WNV was isolated from tissues from all

animals in the 2002 epizootic. Additionally, all animals

from the 2002 epizootic and one animal from the 2001 epi-

zootic were positive for WNV by RT-PCR from fresh or

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 797

RESEARCH

Figure 1. Perivascular changes observed within the brain of alliga-

tors infected with West Nile virus (400x). A. Perivascular infiltrates

were composed of primarily lymphocytes, plasma cells, and

macrophages in the hatchling alligator. B. Perivascular infiltrates

were composed of primarily heterophils (arrows) in juvenile alliga-

tors.

formalin-fixed, paraffin-embedded tissues (Figure 2). In

general, liver was the most likely tissue to yield positive

results. Positive results were not obtained from any of the

tissues from the free-ranging alligator. All tissues tested

negative by RT-PCR for EEEV and SLEV. Retrospective

attempts to culture WNV at both 37°C and room tempera-

ture on FHM, CCO, EPC, and WWS cells were negative.

Aliquots from the horsemeat that was being fed during

the 2002 epizootic tested positive for WNV by RT-PCR

(Figure 2). Aliquots of the horsemeat from two postepi-

zootic shipments were negative for WNV by RT-PCR.

Bacterial Culture

Aeromonas sobria and Edwardsiella tarda were consis-

tently cultured from the intestines. These organisms and

occasionally others (Escherichia coli, Pseudomonas fluo-

rescens, α- and β-hemolytic Streptococcus) were isolated

from various tissues (liver, lung, and kidney) from the alli-

gators dying during the 2001 epizootics and the juveniles

from the 2002 epizootics. Alcaligenes spp. were isolated

from a tonsil swab in one of the animals in 2001.

Salmonella Group D was isolated from the intestines of the

hatchling alligator submitted in 2002.

Discussion

The histologic findings from the hatchling alligator

were most suggestive of a viral etiology, whereas those of

the older alligators were most suggestive of a primary bac-

terial cause. Given that both the RT-PCR and virus isola-

tion were positive for WNV, that virus is suspected to be

the underlying cause of both epizootics. Contaminated

horsemeat is the presumed source of the outbreak. We

speculate that the WNV infection led to the alligators’

immune systems’ becoming immunocompromised, which

resulted in the animals being more susceptible to various

environmental stressors and subsequent invasion by

opportunistic pathogens. Failure to isolate virus from the

alligators in 2001 may have been due to the inability of the

virus to propagate in the four cell lines used (FHM, CCO,

EPC, and WWS cells), as determined by retrospective cul-

ture attempts, rather than absence of virus.

Two important points to examine further are time of

year and age of affected animals. Both epizootics occurred

in the late fall to early winter. Although the epizootics

appeared to be correlated with the first abrupt drop in envi-

ronmental temperature, this finding was likely coinciden-

tal, especially given that the animals were housed in envi-

ronmentally controlled barns. The most likely factor in the

time of year is correlation with the occurrence of WNV

infection in horses. Historically, horses become infected

with WNV during the mosquito season (summer through

early fall). Undiagnosed WNV-infected animals sold for

food would most likely end up in the food supply during

the late summer and early fall months. As was found in this

study, deaths traced to consumption of contaminated food

would taper off in late fall or early winter as the food sup-

ply was less likely to contain virus. Furthermore, all ani-

mals have equal potential for viral exposure through con-

sumption because individual packages of horsemeat are

combined before mixing with the vitamin supplements and

being divided between all barns. In general, reptiles

achieve immunocompetence at an early age (often in a

matter of days), but this immunocompetence may be tem-

perature dependent until the animals are several months of

age (11). This fact may partially explain why the hatchling

alligators tended to die from the viral infection, whereas

the juveniles tended to die from infections caused by sec-

ondary invaders.

Extrinsic stressors may have increased certain animals’

susceptibility to the virus or opportunistic pathogens. For

example, the pens where the epizootics originated tended

to be the first to be washed out at 6 a.m., the coolest time

of the day. During the first abrupt drop in environmental

temperature, the first wash water was possibly cooler

because of colder water in the line between the boiler and

the pens. This cold stressor would serve as a shock to the

animals’ systems. During the 2002 outbreak, an additional

stress was internal construction, undertaken 2 weeks before

the epizootic within the initially affected building. The

environmental (temperature and darkness) control of the

building was maintained during this time, but silence was

not maintained. Additionally, sanitation-related stress may

have occurred during periods of intermittent flushing, such

as over weekends and during pen renovation activities.

Whether brood stock source had an affect on the sus-

ceptibility of the animals is not clear. Although Florida

stock animals were those initially affected, this finding

was likely coincidental because of their location in the

pens. The pens that were more exposed to external stres-

798 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure 2. West Nile virus (WNV) reverse transcription-polymerase

chain reaction results from epizootic die-offs in farm-raised alliga-

tors. The expected amplicon is 248 bp. Lane 1, a 100-bp molecu-

lar weight ladder. Lane 2, the positive WNV control. Lane 3, fresh

tissue samples from a juvenile alligator in the 2002 epizootic. Lane

4, virus isolation cell homogenate from a juvenile alligator in the

2002 epizootic. Lane 5, horsemeat that was being fed to alligators

during the 2002 epizootic. Lane 6, initial postepizootic horsemeat

shipment. Lanes 7, 8, and 9, formalin-fixed, paraffin-embedded tis-

sues of juvenile alligators in 2001 and 2002. Lane 10, fresh tissue

from a wild alligator. Lane 11, negative WNV control.

sors contained Florida animals. Additionally, most animals

in the production unit are from Florida brood stock.

Several management recommendations were suggested

to the producer. The primary recommendation was to stop

feeding horsemeat and switch to another food source such

as beef or fish. We also recommended that the water tem-

perature be reduced to 29.4°C in an attempt to reduce the

stress of rapid growth and perhaps produce an environment

less conducive for viremia. To date, neither of these rec-

ommendations has been implemented, but subsequent

horsemeat shipments have tested negative. Future investi-

gation will include the testing of the eggs from the brood

stock, clinically healthy animals, rookery birds, and free-

ranging alligators to explore the epidemiology of this virus

in ectotherms.

Acknowledgments

We thank the staff of The University of Georgia Tifton

Veterinary Diagnostic and Investigational Laboratory for help in

processing the samples.

Dr. Miller is an assistant professor in the Department of

Pathology at the University of Georgia (UGA) College of

Veterinary Medicine. She works as a veterinary pathologist at the

UGA Tifton Veterinary Diagnostic and Investigational

Laboratory. Her research interests are in wildlife disease and

reproduction.

References

1. Doi R, Oya A, Shirasaka A, Yabe S, Sasa M. Studies on Japanese

encephalitis virus infection of reptiles. II. Role of lizards on hiberna-

tion of Japanese encephalitis virus. Jpn J Exp Med 1983;53:125–34.

2. McLean RG, Ubico SR, Bourne D, Komar N. West Nile virus in live-

stock and wildlife. Curr Top Microbiol Immunol 2002;267:271–308.

3. Nir Y, Lasowski Y, Avivi A, Cgoldwasser R. Survey for antibodies to

arboviruses in the serum of various animals in Israel during

1965–1966. Am J Trop Med Hyg 1969;18:416–22.

4. Thomas LA, Patzer ER, Cory JC, Coe JE. Antibody development in

garter snakes (Thamnophis spp.) experimentally infected with west-

ern equine encephalitis virus. Am J Trop Med Hyg 1980;29:112–7.

5. Oya A, Doi R, Shirasaka A, Yabe S, Sasa M. Studies on Japanese

encephalitis virus infection of reptiles. I. Experimental infection of

snakes and lizards. Jpn J Exp Med 1983;53:117–23.

6. Kuno G. Universal diagnostic RT-PCR protocal for arboviruses. J

Virol Methods 1998;72:27–41.

7. Johnson DJ, Ostlund EN, Pedersen DD, Schmitt BJ. Detection of

North American West Nile virus in animal tissue by a reverse tran-

scription-nested polymersase chain reaction assay. Emerg Infect Dis

2001;7:739–41.

8. Linssen B, Kinney RM, Aquilar P, Russell KL, Watts DM, Kaaden

OR, et al. Development of reverse transcription-PCR assays specific

for detection of equine encephalitis viruses. J Clin Microbiol

2000;38:1527–35.

9. Lanciotti RS, Kerst AJ. Nucleic acid sequence-based amplification

assays for rapid detection of West Nile and St. Louis encephalitis

viruses. J Clin Microbiol 2001;39:4506–13.

10. Done LB. Postural Abnormalities. In: Mader DR, editor. Reptile med-

icine and surgery. Philadelphia: W.B. Saunders Co; 1996. p. 406–11.

11. Frye FL. Biomedical and surgical aspects of captive reptile hus-

bandry, Vol 1. 2nd edition. Malabar (FL): Krieger Publishing Co;

1991.

Address for correspondence: Debra L. Miller, University of Georgia,

Veterinary Diagnostic and Investigational Laboratory, Tifton, GA 31793,

USA; fax: 229-386-7128; email: [email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 799

RESEARCH

The opinions expressed by authors contributing to this journal do

not necessarily reflect the opinions of the Centers for Disease

Control and Prevention or the institutions with which the authors

are affiliated.

Over the past two decades, dengue virus serotype 3

(DENV-3) has caused unexpected epidemics of dengue

hemorrhagic fever (DHF) in Sri Lanka, East Africa, and

Latin America. We used a phylogenetic approach to evalu-

ate the roles of virus evolution and transport in the emer-

gence of these outbreaks. Isolates from these geographi-

cally distant epidemics are closely related and belong to

DENV-3, subtype III, which originated in the Indian subcon-

tinent. The emergence of DHF in Sri Lanka in 1989 corre-

lated with the appearance there of a new DENV-3, subtype

III variant. This variant likely spread from the Indian sub-

continent into Africa in the 1980s and from Africa into Latin

America in the mid-1990s. DENV-3, subtype III isolates

from mild and severe disease outbreaks formed genetical-

ly distinct groups, which suggests a role for viral genetics in

DHF.

rthropod-borne viruses are responsible for the emer-

gence of unexpected diseases in humans, as illustrat-

ed by the identification of West Nile virus encephalitis in

the American hemisphere in 1999 (1). The emergence of a

new disease is often attributable to the transport of a

pathogen (as in the case of West Nile virus) or changes in

the evolution or ecology of a native pathogen that hitherto

caused mild or no disease in humans (2,3). We studied

unexpected outbreaks of dengue hemorrhagic fever (DHF)

in Sri Lanka, East Africa, and Latin America caused by

dengue serotype 3 (DENV-3) virus.

Most persons infected with dengue viruses are asymp-

tomatic or develop dengue fever (DF). DHF and dengue

shock syndrome (DSS), which can be fatal, develop in a

minority of infected persons. The pathogenesis of DHF is

poorly understood, although factors such as age and previ-

ous exposure to dengue infections increase the risk for

severe disease (4). Epidemiologic studies point to particu-

lar DENV strains being more virulent than others (5–8).

For example, the dengue genotypes endemic to Central

and South America have caused mild disease, while the

Asian genotypes introduced to the region have led to DHF

epidemics (9–16). Similarly, outbreaks of DHF in some

Pacific islands have been traced to the introduction of

Southeast Asian dengue strains (17). DENV-2 subtypes

associated with mild and severe disease epidemics have

distinct mutations in the E gene and 5′ and 3′ untranslated

segments of the viral genome, although whether these

mutations directly contribute to pathogenesis is unproven

(18).

The distribution of DHF and DSS in Asia has been par-

ticularly puzzling. Before 1989, DHF was common in

Southeast Asia but rare in the Indian subcontinent despite

the circulation of all four serotypes in both regions. After

1989, this pattern of disease changed and regular epi-

demics of DHF were reported from several countries in the

Indian subcontinent (19). Sri Lanka, in particular, experi-

enced a dramatic and persistent increase in DHF cases

(20). Epidemiologic studies of dengue in Sri Lanka have

demonstrated that the intensity of virus transmission, as

well as the relative abundance of each serotype, remained

constant before and after the emergence of DHF (21).

Thus, DHF did not emerge in Sri Lanka because of an

overall increase in virus transmission or shift in serotype.

Although all four serotypes of dengue circulate in Sri

Lanka, persons who have the severe form of the disease

are most frequently infected with DENV-3 (20,22).

Lanciotti et al. characterized the genetic relatedness of

DENV-3 isolates from regions throughout the tropics and

subtropics and identified four geographically distinct sub-

types (23). All Sri Lankan isolates were classified as sub-

type III, which also includes isolates from East Africa and

India, as well as recent isolates from Latin America.

Because DENV-3 isolates from Sri Lanka isolated before

and after 1989 (when DHF emerged) formed separate

groups within subtype III, Lanciotti and colleagues postu-

lated that a genetic shift in DENV-3 may have been

responsible for the emergence of DHF (23).

In the current study, using phylogenetic methods, we

analyzed DENV-3 viruses isolated from Sri Lanka for up

to 10 years after the emergence of DHF to confirm the

800 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Emergence and Global Spread

of a Dengue Serotype 3,

Subtype III Virus

William B. Messer,* Duane J. Gubler,† Eva Harris,‡ Kamalanayani Sivananthan,§

and Aravinda M. de Silva*

*University of North Carolina, Chapel Hill, North Carolina, USA;

†Centers for Disease Control and Prevention, Fort Collins,

Colorado, USA; ‡University of California, Berkeley, California,

USA; and §Medical Research Institute, Colombo, Sri Lanka

establishment of a new genotype and evaluate the roles of

virus evolution and transport in establishing a new geno-

type. DENV-3, subtype III was introduced into Latin

America in 1994 (11), and the virus has subsequently been

isolated from DF and DHF outbreaks throughout Central

and South America (12–16). We also examined the genet-

ic relationships between DENV-3, subtype III isolates

from Latin America, East Africa, and the Indian subconti-

nent. On the basis of our results, we describe the most

likely scenario of events that led to the emergence of

DENV-3 –associated DHF in the Indian subcontinent and

the Americas.

Materials and Methods

Virus Strains

The dengue virus strains sequenced for this study as

well as sequences obtained from GenBank for this study

are listed in Table 1. The virus isolates were obtained

from the Centers for Disease Control and Prevention,

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 801

RESEARCH

Table 1. Dengue virus type 3 sequences used

Strain

Location

Name

Subtype

Sequence source

GenBank accession no.

D1266

1983

Sri Lanka

83SriLan1

III

This study

AF547225

D1306

1983

Sri Lanka

83SriLan2

III

This study

AF547226

D1307

1983

Sri Lanka

83SriLan3

III

This study

AF547227

D1336

1983

Sri Lanka

83SriLan4

III

This study

AF547228

D1440

1984

Sri Lanka

84SriLan1

III

This study

AF547229

073

1985

Sri Lanka

85SriLan

III

This study

AF547241

D2783

1989

Sri Lanka

89SriLan1

III

This study

AF547230

D2863

1989

Sri Lanka

89SriLan2

III

This study

AF547231

D2803

1989

Sri Lanka

89SriLan3

III

This study

AF547232

D3197

1990

Sri Lanka

90SriLan1

III

This study

AF547233

D5231

1993

Sri Lanka

93SriLan1

III

This study

AF547234

D9397

1994

Sri Lanka

94SriLan1

III

This study

AF547235

L57

1997

Sri Lanka

97SriLan1

III

This study

AF547242

1998

Sri Lanka

98SriLan

III

This study

AF547243

1557

1985

Mozambique

85Mozamb1

III

This study

AF547236

1558

1985

Mozambique

85Mozamb2

III

This study

AF547237

1559

1985

Mozambique

85Mozamb3

III

This study

AF547238

251991

1991

Kenya

91Kenya

III

This study

AF547239

SOM079

1993

Somalia

93Somalia

III

This study

AF547240

32267

1994

Nicaragua

94Nicara1

III

This study

AF547244

6845

1998

Nicaragua

98Nicara1

III

This study

AF547245

7431

1998

Nicaragua

98Nicara2

III

This study

AF547246

7071

1998

Nicaragua

98Nicara3

III

This study

AF547262

BC 96/94

1994

Panama

94Panama1

III

This study

AF547247

032231

1994

Panama

94Panama2

III

This study

AF547248

BC 13/96

1994

Panama

94Panama3

III

This study

AF547249

BC 20/97

1996

Mexico

96Mexico1

III

This study

AF547250

BC 172/97

1996

Mexico

96Mexico2

III

This study

AF547251

BC 184/97

1996

Mexico

96Mexico3

III

This study

AF547252

BC173/97

1996

Mexico

96Mexico4

III

This study

AF547253

17605

1995

Costa Rica

95CostaR1

III

This study

AF547254

17608

1995

Costa Rica

95CostaR2

III

This study

AF547255

322473

1995

Costa Rica

95CostaR3

III

This study

AF547256

322488

1995

Costa Rica

95CostaR4

III

This study

AF547257

20/8

1997

Guatemala

97Guatem1

III

This study

AF547263

366-781

1998

Puerto Rico

98PuertoR1

III

This study

AF547258

400-996

2000

Puerto Rico

00PuertoR1

III

This study

AF547264

1998

El Salvador

98ElSalv1

III

This study

AF547259

612210

2001

Venezuela

01Venezue1

III

This study

AF547260

VEN03

2001

Venezuela

01VEN03

III

This study

AF547261

Ref. 18

1981–91

Sri Lanka

81,85,89,91 SriLanA

III

GenBank

L11431,L11436–L11438

Ref. 18

1984

India

84IndiaA

III

GenBank

L11424

Ref. 18

1986

Samoa

86Samoa

III

GenBank

L11435

Ref. 18

1962–86

Thailand

62,73,86,86 Thailand

GenBank

L11440–L11442,L11620

Ref. 18

1983

Philippines

83Philipp

GenBank

L11432

Ref. 18

1989

Tahiti

89Tahiti

GenBank

L111619

Ref. 18

1992

Fiji

92Fiji

GenBank

L11422

Ref. 18

1973–85

Indonesia

73,78,85 Indones

GenBank

L11425,L11426,L11428

Ref. 18

1974–81

Malaysia

74,81 Malaysi

GenBank

L11429,L11427

Ref. 18

1956

Philippines

D3H-87

GenBank

L11423

Ref. 18

1963–77

Puerto Rico

63,77 PuertoR

GenBank

L11433,L11434

Ref. 18

1965

Tahiti

65 Tahiti

GenBank

L11439

Includes original identifier for strain, year of isolation, taxa name used in this paper, dengue virus 3 subtype, source of viral sequence, an

d GenBank accession numbers.

Dengue Branch, Puerto Rico, and Division of Vector-

Borne Infectious Diseases, Ft. Collins, Colorado;

Medical Research Institute, Colombo, Sri Lanka; School

of Public Health, Berkeley, California; Walter Reed

Army Institute for Research, Washington, D.C.; and

University of Massachusetts Medical Center, Worcester,

Massachusetts.

RNA Extraction

QiaAmp Viral RNA Mini Kit (QIAGEN, Valencia, CA)

was used to extract viral RNA from both the mosquito

grind supernatants and infected tissue culture media fol-

lowing the manufacturer’s protocol. Extracted RNA was

stored at –70°C or immediately subjected to reverse tran-

scription–polymerase chain reaction (RT-PCR).

RT-PCR

DENV-3 RT-PCR was carried out as described by

Lanciotti (23). Primers were designed to amplify and

sequence a 966-bp fragment from positions 179–1,144,

encompassing part of Capsid, all of PreM, and part of the

E gene sequences. The reverse primer (DEN3/735)

hybridized to positions 1,189–1,171 (5′-ctcctcaggcaaaac-

cgct-3′) and the forward primer (D1 consensus) hybridized

to positions 132–159 (5′-tcaatatgctgaaacgcgcgagaaaccg-

3′). The reverse primer DEN3/735 was added to extracted

RNA, incubated at 85°C for 90 s, and allowed to cool to

room temperature. RT was carried out for 45–60 min in 20

µL of reaction mix containing 25 U avian myeloblastosis

virus reverse transcriptase (Roche, Nutley, NJ), deoxynu-

cleoside tripophosphate, MgCl

, and RT buffer. PCR was

performed by adding a 30-µL cocktail containing D1 con-

sensus primer, PCR buffer, and EXPAND polymerase

(Roche) to the 20-µL RT reaction. PCR conditions were 4

min at 94°C, 30–35 cycles of 94°C for 30 s, 54°C for 30 s,

and 72°C for 90 s with 5 s/cycle added to elongation step

after the first 10 cycles. We separated 5 µL of the reaction

products on 2% agarose gels and visualized it by ethidium

bromide staining. When necessary, target bands were

excised and purified by using the Qiagen QIAquick Gel

Extraction kit (QIAGEN) following manufacturer’s

instructions. All remaining PCR reaction products were

purified by using the Qiagen PCR Purification kit follow-

ing the manufacturer’s protocol.

DNA Sequencing

Purified PCR products were sent to the automated DNA

sequencing facility at the University of North Carolina,

Chapel Hill, NC. The DENV-3 sequences used in this man-

uscript included 40 newly determined sequences, which

have been submitted to GenBank (accession nos.

AF547225–AF547264).

Viral Sequence Analysis

Overlapping individual nucleic acid sequences were

assembled with the aid of VECTOR NTI ContigExpress

(InforMax, Inc., Bethesda, MD). Sequences were aligned

and analyzed by using the following software: Clustal X

(available from: URL: http://inn-prot.weizmann.ac.il/soft-

ware/ClustalX.html), PAUP* (available from: URL:

http://www.sinauer.com), PHYLIP (available from: URL:

http://evolution.genetics.washington.edu/phylip.html), and

MEGA II (available from: URL: http://www.megasoft-

ware.net). Genetic distances were calculated by using

Tamura-Nei distance algorithm with 1,000 bootstrap repli-

cates; the trees were generated by using the Minimum

Evolution method. The phylogenetic tree in Figure 1 is

based on a 708-base segment, positions 437–1,145, span-

ning pre-M/M and a portion of the E gene. The phyloge-

netic tree presented in Figure 2 is based on 966-base region

spanning positions 179–1,145 on the viral genome, captur-

ing a portion of the C gene, all of pre-M/M gene, and a

portion of the E gene.

Results and Discussion

Many investigators have used viral nucleotide sequence

data and phylogenetic methods to understand genetic rela-

tionships between viruses, as well as the epidemiology of

viral disease. Phylogenetic studies have shown that dengue

viruses can move long distances between continents (24)

as well as short distances between neighboring countries

(25). Our goal was to use a phylogenetic approach to

understand recent DHF outbreaks caused by DENV-3

infections in the Indian subcontinent and Latin America.

Previous phylogenetic analysis of DENV-3 has princi-

pally relied on complete or partial sequences of the pre-

M/M and E genes (13,16,17,23). Our analysis used a 708-

base segment, positions 437–1,145, spanning pre-M/M

and a portion of the E gene, coding for 236 amino acids.

This region was selected because it both conserved the

original phylogenetic relationship identified by Lanciotti

et al. and, in preliminary analysis with previously estab-

lished sequences, captured 44% of the variable sites with-

in DENV-3, subtype III Sri Lankan sequences. No inser-

tion/deletion mutations and no hypervariable regions were

detected in this span.

A total of 40 DENV-3 sequences, including 21

sequences available from GenBank and 19 newly deter-

mined Indian subcontinent and African sequences (Table

1) were compared. Dates of isolation ranged from 1963 to

1998. With the exception of 63PuertoR, all sequences

were from low-passage (<4) virus cultures. Several

approaches to phylogenetic analysis, including maximum

likelihood, parsimony, and distance methods, were com-

pared. All approaches yielded identical or nearly identical

802 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

topologies. Results presented here used the Tamura-Nei

algorithm to calculate genetic distances and the minimum

evolution method to create the trees (Figure 1). The tree

identifies four distinct lineages that correspond to the

region of isolation, reproducing the same evolutionary

relationship first described by Lanciotti, et al. (23).

Subtype I includes isolates from Southeast Asia and the

South Pacific islands; subtype II consists of isolates from

Thailand; subtype III is comprised of isolates from the

Indian subcontinent, East Africa, and a single isolate from

Samoa; and subtype IV includes Puerto Rico and Tahiti.

Similarity within subtypes was high, with subtype III

showing the greatest mean similarity (98.4%), followed

by subtypes I, II, and IV (Table 4).

All 24 Sri Lankan, Indian, and East African strains fell

into subtype III (Figure 1). The circulating virus genotypes

within this region have remained closely related over the

relatively long period of 18 years (1981–1998), indicating

that countries bordering the western Indian Ocean form a

geographically distinct region with regard to DENV-3

viruses. DENV-2 viruses in the regions also form a subtype

with a similar geographic distribution (26,27). Frequent

trade between East Africa, Western Indian Ocean islands,

and the Indian subcontinent may have been responsible for

the movement of dengue viruses throughout the region

(26,28). Rico-Hesse, for example, demonstrated the intro-

duction of DENV-2 to Africa from islands in the Indian

Ocean (26). The earliest subtype III virus on record is an

isolate from India in 1966; this virus occupies a node that

is ancestral to all the subsequent Asian and African isolates

(R.S. Lanciotti, pers. comm.), suggesting that the DENV-

3, subtype III viruses have their origin in the Indian sub-

continent and have subsequently spread out of the region.

In Sri Lanka, regular epidemics of DHF have been

observed only after 1988. DENV-3 is responsible for many

of the infections that progress to DHF (20,22). DENV-3

isolates obtained before and after the emergence of DHF

are very closely related and belong to subtype III, indicat-

ing that the emergence of DHF on the island is not due to

the introduction of a new subtype from outside the region.

However, within subtype III, most Sri Lankan isolates

(except for 93SriLan1) from before and after the emer-

gence of DHF segregated into two distinct clades, desig-

nated groups A and B (Figure 1). Group A, with nine iso-

lates from 1981 to 1989, consists of viruses collected up to

the year epidemic DHF emerged in Sri Lanka but contains

no isolates from later than 1989. Group B includes eight

isolates from 1989 to 1998 but none from before 1989.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 803

RESEARCH

Figure 1. Phylogenetic tree of estab-

lished dengue virus 3 (DENV-3) sub-

types (23) and the relationship of Sri

Lanka pre– and post–dengue hemor-

rhagic fever DENV-3 isolates to the

established subtypes. This tree is based

on a 708-base segment, positions 437

to 1145, spanning pre-M/M and a por-

tion of the E gene. Scale bar shows

number of substitutions per bases

weighted by Tamura-Nei algorithm.

Horizontal distances are equivalent to

the distances between isolates.

Numbers at nodes indicate bootstrap

support values for the branch of the tree

inferred at that node. The origin of the

viruses and sequences used are listed

in Table 1. The amino acid substitutions

conserved within each DENV-3 subtype

are listed in Table 2. DHF, dengue hem-

orrhagic fever.

Temporally, the two groups are continuous, by virtue of

sharing isolates in 1989. Group A includes isolate

89SriLan2, while group B contains 89SriLan1, 89SriLan3,

and 89SriLanA. However, the groups do not form a contin-

uous lineage; they share a common ancestor only at the

node for subtype III (Figure 1). Group B shares ancestral

nodes with isolates from India and East Africa. Because

the Indian and East African isolates overlap temporally

with group A (all isolates are from the 1980s), group A and

group B lineages likely diverged sometime before 1981

and followed distinct evolutionary pathways.

We propose two likely scenarios that led to the emer-

gence of group B viruses in Sri Lanka. One possibility is

that the group B viruses were introduced from India or

East Africa into Sri Lanka because Indian and East African

isolates from the mid-1980s are closely related to Sri

Lankan group B viruses (Figure 1). Of the two regions,

India is the more likely source because of geographic prox-

imity to Sri Lanka, although the East African viruses could

be the direct ancestors of the group B viruses. Another pos-

sibility is that both groups co-circulated in Sri Lanka in the

early 1980s, with group B being a minor population. Some

804 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 2. Amino acid substitutions conserved within each dengue virus 3 subtype for the isolates used to create the phylogenetic tree in

Figure 1

Position

Subtype

Name

128

135

148

188

234

Outgroup

56Philipp

73Indones

–

74Malaysi

–

78Indones

–

81Malaysi

–

83Philipp

–

85Indones

–

89Tahiti

–

92Fiji

–

62Thailan

–

73Thailan

–

86Thailan

–

87Thailan

–

III

85Mozamb1

–

III

85Mozamb2

–

III

85Mozamb3

–

III

84IndiaA

–

III

91Kenya

–

III

93Somolio

–

III

81SriLanA

–

III

83SriLan1

–

III

83SriLan2

–

III

83SriLan3

–

III

83SriLan4

–

III

84SriLan1

–

III

85SriLanA

–

III

85SriLan

–

III

89SriLan2

–

III

89SriLanA

–

III

89SriLan1

–

III

89SriLan3

–

III

90SriLan1

–

III

91SriLanA

–

III

93SriLan1

–

III

94SriLan1

–

III

97SriLan1

–

III

98SriLan1

–

III

86Samoa

–

63PuertoR

–

65Tahiti

–

77PuertoR

–

Reference strain 56Philipp is the highly passaged laboratory strain H87. Positions are numbered sequentially from the first position in the pre-M/M protein

selective force operating in the late 1980s may have shift-

ed the balance in favor of group B viruses. In either case,

group B viruses emerged in Sri Lanka because a subtype

III variant already established in the greater region became

more common in Sri Lanka and not because a novel virus

evolved and emerged de novo on the island.

DENV-3, subtype III was detected in the Americas dur-

ing DF and DHF outbreaks in Nicaragua and Panama in

1994 (11). Subsequently, the virus has spread to many

countries in Latin America, and DENV-3–associated DHF

was confirmed in several countries (13,14,16,29–31)

(Figure 3). To establish the relationship of recent Latin

American DENV-3 isolates to each other and to the previ-

ously identified Indian subcontinent and East African sub-

type III isolates, we sequenced and analyzed a 966-base

region spanning positions 179–1,145 on the viral genome,

capturing a portion of the C gene, all of pre-M/M gene, and

a portion of the E gene. This region adds 288 positions to

the 5′ end of the sequences initially presented in this study.

Forty-three isolates were sequenced (21 from Mexico

and Central and South America, 16 from Sri Lanka, 1 from

India, and 5 from East Africa) (Table 1). The D3H-87

belonging to DENV-3, subtype 1 was used as an outgroup.

Years of isolation ranged from 1983 to 2001, an 18-year

span. Except for the D3H-87 outgroup, all other isolates

were low-passage clinical isolates. Most of the nucleotide

mutations were silent: only 12 amino acid positions

showed any variability and only 2 positions showed vari-

ability in more than one isolate. Consequently, the evolu-

tionary relationships observed in this analysis likely reflect

the results of genetic drift and are unlikely to have been

influenced by host-specific selection events on this portion

of the genome (26).

All sequences included in this analysis fell within sub-

type III (data not shown). Several approaches to phyloge-

netic analysis were compared, and all approaches yielded

identical or nearly identical topologies. We used the

Tamura-Nei algorithm to calculate genetic distances and

the minimum evolution method to create the trees (Figure

2). Bootstrap values are shown at critical nodes. Despite

the high overall similarity of the isolates in this analysis,

geographically and temporally distinct groups formed sep-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 805

RESEARCH

Figure 2. Phylogenetic tree of dengue

virus 3, subtype III group A, group B, East

Africa, and Latin America. Tree is based

on 966-base region spanning positions

179–1,145 on the viral genome, capturing

a portion of the C gene, all of pre-M/M

gene and a portion of the E gene.

Nucleotide substitutions conserved within

each dengue virus 3, subtype III group

(group A, B, East Africa, and Latin

America) are listed in Table 3. DHF,

dengue hemorrhagic fever.

arate lineages. Generally, two separate lineages formed

within subtype III. The first consists of group A viruses

isolated from 1981 to 1989 in Sri Lanka. These viruses

have been associated only with DF. The second is com-

posed of Sri Lankan group B, Indian, East African, and all

of the isolates from Mexico and Central and South

America.

Within group A, members are closely related, with a

nucleotide mean similarity of 99.4% (Table 5). Within the

expanded group B and related viruses, three distinct clades

exist: a group of closely related Sri Lankan isolates from

1989 to 1998, 5 East African isolates from 1985 to 1993,

and 21 isolates from 1994 to 2001 from Latin America.

Isolates 84India and 93SriLan1 are less closely related to

the other geographically distinct isolates in the larger sec-

ond lineage.

The isolates from Latin America all emerge from a

common node on the tree, suggesting a single introduction

806 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 3. Nucleotide substitutions conserved within dengue virus 3, subtype III groups

Position

Group

Strain

338

429

503

566

653

686

695

707

728

734

749

791

821

866

896

902

978

1010

1019

1056

Reference

H87

Group A

83SriLan1

Group A

83SriLan2

Group A

83SriLan3

Group A

83SriLan4

Group A

84SriLan1

Group A

85SriLan1

Group A

89SriLan2

unclassified

84India

unclassified

93SriLan1

East Africa

85Mozamb1

East Africa

85Mozamb2

East Africa

85Mozamb3

East Africa

91Kenya

East Africa

93Somalia

Group B

89SriLan1

Group B

89SriLan3

Group B

90SriLan1

Group B

94SriLan1

Group B

97SriLan1

Group B

98SriLan1

L. America

94Nicara1

L. America

94Panama1

L. America

94Panama2

L. America

94Panama3

L. America

CostaRica1

L. America

CostaRica2

L. America

CostaRica3

L. America

CostaRica4

L. America

96Mexico1

L. America

96Mexico2

L. America

96Mexico3

L. America

96Mexico4

L. America

97Guatem1

L. America

98Nicara1

L. America

98Nicara2

L. America

98Nicara3

L. America

98ElSalv1

L. America

01Venezue1

L. America

01VEN03

L. America

98PuertoR1

L. America

00PuertoR1

Reference strain is the highly passaged laboratory strain H87. Positions are numbered sequentially from the first nucleotide position at the 5′ end of the genome.

of a virus and the subsequent diversification of the virus

population from the founding strain. The DENV-3, sub-

type III isolates from Nicaragua, Panama, and Costa Rica

are closest to the Latin American group’s originating node,

with the more recent isolates found farther from that node,

reflecting the viral population’s ongoing evolution after

the point source introduction.

The internal branch from the Latin American group

shares a common node with the isolates from East Africa.

The common hypothetical ancestor for Latin America and

East Africa then shares a common node with the Sri

Lankan group B virus isolates. Both on the phylogenetic

tree and in pair-wise comparisons (Table 5), the Latin

American group was more closely related to the isolates

from East Africa than to the group B Sri Lankan isolates.

Furthermore, the East African isolates pre-date the earliest

Latin American isolates by 9 years, while the less closely

related Sri Lankan group B isolates are nearly contempo-

raneous with the Latin American isolates. Therefore, the

point source DENV-3 introduction into Latin America is

most likely to have its origins in East Africa and not the

Indian subcontinent (Figure 4).

Little is known about dengue activity in Africa, partic-

ularly DENV-3 (32). DENV-3 was first detected on the

African continent in 1984 to 1985 during an outbreak in

Mozambique (32). Later studies of U.S. troops in Africa

and the Persian Gulf suggested that DENV-3 is endemic in

those regions but largely undetected (33). Our results show

that all East African DENV-3 isolates belong to subtype

III. The fact that DENV-3 was only first isolated from East

Africa in 1985, whereas the viruses were present in the

Indian subcontinent at least as far back as 1966 (R.S.

Lanciotti, pers. comm.), suggests that DENV-3, subtype III

was introduced from the Indian subcontinent into East

Africa in or before 1984 (Figure 4). This introduction led

to the establishment of a stable East African group of

DENV-3, subtype III because all the isolates from

Mozambique, Kenya, and Somalia isolated from 1985 to

1993 form a distinct clade within subtype III (Figure 3).

The DENV-3, subtype III viruses introduced into Latin

America are most closely related to subtype III viruses in

East Africa (Figure 3). Although we can only speculate

about the exact mode of transport of DENV-3 into Latin

America, we propose that Panama, with its canal that

attracts goods as well as civilians and military personnel

from other parts of the world, may have been the point of

introduction of subtype III into the Americas. Similarly,

the introduction of DENV-2 in 1981 into Cuba may be

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 807

RESEARCH

Table 4. Summary of within- and between-subtype nucleotide

mean similarity for dengue virus 3 isolates shown in Figure 1

Between subtype similarity (%)

Subtype

Within subtype

similarity (%)

III

98.1

97.7

94.9

III

98.4

95.6

96.3

97.6

92.3

92.5

92.7

Mean similarities were calculated with the Tamura-Nei distance algorithm.

Figure 3. Map of the spread of dengue virus 3 (DENV-3), subtype

III through Latin America and the Caribbean. The introduction of

DENV-3, subtype III was first reported in November 1994 in

Nicaragua and Panama. This virus strain has been isolated, iden-

tified, and reported in at least 16 other countries in the region.

*Represents countries with dengue hemorrhagic fever (DHF)

caused by DENV-3. These countries are Nicaragua in 1994 and

1998, Brazil and Venezuela in 2001 (Pan American Health

Organization, unpub. data).

Table 5. Summary of within- and between-group nucleotide mean similarity for the dengue virus 3, subtype III virus isolates shown in

Figure 2

Between group similarity

Subgroup

Within-group similarity

Subgroup A

East Africa

Subgroup B

Subgroup A

99.4%

East Africa

99.5%

98.2%

Subgroup B

98.8%

97.9%

98.7%

Latin America

99.5%

98.0%

99.0%

98.5%

Mean similarities were calculated with the Tamura-Nei distance algorithm.

attributable to Cuban military personnel traveling between

Southeast Asia and Cuba (24,34).

Epidemiologic and clinical studies on dengue in

Indonesia in the 1970s pointed to strain differences

between DENV-3 viruses contributing to transmission and

disease severity (35,36). Despite their overall similarity at

the nucleotide level, the DENV-3, subtype III isolates

examined in this study have been associated with severe or

mild disease outbreaks (Figure 3). Sri Lankan group A

viruses were isolated during a time of little to no DHF,

while group B viruses were isolated after the emergence of

DHF in Sri Lanka. The emergence of DHF in Sri Lanka

was not accompanied by a change in dengue transmission

or the abundance of any particular serotype (21).

Implicating DENV-3 directly as the cause of DHF in Sri

Lanka has been difficult because few virus isolates are

available from DHF patients in Sri Lanka. However, dur-

ing dengue surveillance studies in 1997, only DENV-3 was

isolated from hospitalized dengue cases, whereas DENV-

1, DENV-2, and DENV-3 were isolated from patients vis-

iting outpatient clinics (22). These observations suggest

that DENV-3 is responsible for severe dengue disease in

Sri Lanka. Further studies are required to better establish

the relative contribution of DENV-3 to severe disease in

Sri Lanka.

The current studies support a viral genetic basis for

severe and mild disease outbreaks. We found that the popu-

lation of DENV-3 viruses associated with DHF in Sri Lanka

did not appear to be direct descendants of the group A virus-

es that were circulating before DHF emerged in that coun-

try. The Sri Lankan 1989–1997 isolates are more closely

related to the isolates from East Africa and the isolates from

the Americas than they are to the isolates from 1981 to 1989

in Sri Lanka (Figure 3). All three groups of subtype III

viruses (Sri Lankan group B, East African group, and Latin

American group) associated with DHF are more closely

related to each other than they are to the pre–DHF group A

viruses from Sri Lanka (Figure 3). Thus, all the viruses

within subtype III are closely related (mean 98.4% identity

at the nucleotide level), yet they form distinct phylogenetic

groups associated with mild or severe disease.

The Sri Lankan group B viruses may be associated with

severe disease unlike group A viruses because the group B

viruses are inherently more virulent. Alternatively, the

ability of preexisting dengue antibody to neutralize group

A viruses and enhance group B viruses may account for the

observed associations with severe and mild disease. In a

recent study, antibodies against American DENV-1 viruses

neutralized the Native American DENV-2 genotype better

than the Southeast Asian DENV-2 genotype that is current-

ly circulating in the Americas and causing DHF (37). This

study lends support to the idea that Asian DENV-2 may

produce a more severe disease not because of inherent vir-

ulence properties but because persons with previous pri-

mary DENV-1 infections may enhance infection with this

genotype and neutralize infections with the Native

American DENV-2 genotype. Similarly, DENV-2 and -3

are the common serotypes in Sri Lanka, and persons with

previous primary DENV-2 infections could neutralize the

DENV-3 group A viruses better than the group B viruses.

This difference may explain the unexpected emergence of

DHF associated with group B. Further comparative studies

with group A and B viruses are needed to understand their

association with mild and severe disease, respectively.

Acknowledgments

We thank Irene Bosch, Vance Vorndam, Niranjan Kanesa-

thasan, and Eric Wagar for virus isolates and Tissa Vitarana, Gaya

Colombage, Nalini Withana, and staff members in the Virology

Department at the Medical Research Institute in Colombo for

their help.

This work was supported by a Junior Faculty Development

Award and other funds from the University of North Carolina at

Chapel Hill (A.M.S.)

Dr. Messer earned his Ph.D. in ecology from the University

of North Carolina at Chapel Hill. His broad research interests

include the emergence, movement, and evolution of human

pathogens. For his doctoral research, he evaluated host, viral, and

environmental factors that may have contributed to the sudden

emergence of dengue hemorrhagic fever in Sri Lanka. He attends

medical school at the University of North Carolina.

808 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure 4. Global spread of dengue virus 3 (DENV-3), subtype III,

which has been continuously circulating in the Indian subcontinent

from the 1960s to the present. The virus was first isolated from

East Africa in 1985 in Mozambique and subsequently from Kenya

(1991) and Somalia (1993) (32,33). DENV-3 subtype III was first

detected in the American continent in 1994 (Nicaragua and

Panama) and the virus has subsequently spread through most of

Latin America (13,14, 16,29,30). The arrows depict the most likely

directions of spread based on the phylogenetic relationships

between the viruses (see text for details). The map also displays

countries in which dengue is known to occur.

References

1. Briese T, Jia XY, Huang C, Grady LJ, Lipkin WI. Identification of a

Kunjin/West Nile–like flavivirus in brains of patients with New York

encephalitis. Lancet 1999;354:1261–2.

2. Lanciotti RS, Roehrig JT, Deubel V, Smith J, Parker M, Steele K, et

al. Origin of the West Nile virus responsible for an outbreak of

encephalitis in the northeastern United States. Science

1999;286:2333–7.

3. Morse SS. Emerging viruses. New York: Oxford University Press;

1993.

4. Halstead SB. Epidemiology of dengue and dengue hemorrhagic

fever. In: Kuno DJGaG, editor. Dengue and dengue hemorrhagic

fever. New York: CAB International; 1997. p. 23–44.

5. Gubler DJ. Dengue and dengue hemorrhagic fever. Clin Microbiol

Rev 1998;11:480–96.

6. Rosen L. The emperor’s new clothes revisited, or reflections on the

pathogenesis of dengue hemorrhagic fever. Am J Trop Med Hyg

1977;26:337–43.

7. Gubler DJ, Reed D, Rosen L, Hitchcock JR Jr. Epidemiologic, clini-

cal, and virologic observations on dengue in the Kingdom of Tonga.

Am J Trop Med Hyg 1978;27:581–9.

8. Gubler DJ, Suharyono W, Lubis I, Eram S, Gunarso S. Epidemic

dengue 3 in central Java, associated with low viremia in man. Am J

Trop Med Hyg 1981;30:1094–9.

9. Guzman MG, Kouri GP, Bravo J, Soler M, Velasquez S, Morier L.

Dengue haemorrhagic fever in Cuba. A retrospective seroepidemio-

logic study. Am J Trop Med Hyg 1990;42:179–84.

10. Rico-Hesse R, Harrison LM, Salas RA, Tovar D, Nisalak A, Ramos

C, et al. Origins of dengue type 2 viruses associated with increased

pathogenicity in the Americas. Virology 1997;230:244–51.

11. Dengue type 3 infection. Nicaragua and Panama, October–November

1994. Wkly Epidemiol Rec 1995;70:41–3.

12. Gubler DJ. The global pandemic of dengue/dengue haemorrhagic

fever: current status and prospects for the future. Ann Acad Med

Singapore 1998;27:227–34.

13. Guzman M, Huelva G, Saenz E, Quiroz E, De los Reyes J, Balmaseda

A. Reintroduccion del dengue 3 en las Americas: 1994–1996.

Archivos Venezolanos de Medicina Tropial 1998;2:8–19.

14. Balmaseda A, Sandoval E, Perez L, Gutierrez CM, Harris E.

Application of molecular typing techniques in the 1998 dengue epi-

demic in Nicaragua. Am J Trop Med Hyg 1999;61:893–7.

15. Nogueira RM, Miagostovich MP, de Filippis AM, Pereira MA,

Schatzmayr HG. Dengue virus type 3 in Rio de Janeiro, Brazil. Mem

Inst Oswaldo Cruz 2001;96:925–6.

16. Usuku S, Castillo L, Sugimoto C, Noguchi Y, Yogo Y, Kobayashi N.

Phylogenetic analysis of dengue-3 viruses prevalent in Guatemala

during 1996–1998. Arch Virol 2001;146:1381–90.

17. Chungue E, Deubel V, Cassar O, Laille M, Martin PM. Molecular

epidemiology of dengue 3 viruses and genetic relatedness among

dengue 3 strains isolated from patients with mild or severe form of

dengue fever in French Polynesia. J Gen Virol 1993;74:2765–70.

18. Leitmeyer KC, Vaughn DW, Watts DM, Salas RA, de Chacon IV,

Ramos C, et al. Dengue virus structural differences that correlate with

pathogenesis. J Virol 1999;73:4738–47.

19. Srivastava VK, Suri S, Bhasin A, Srivastava L, Bharadwaj M. An epi-

demic of dengue haemorrhagic fever and dengue shock syndrome in

Delhi: a clinical study. Ann Trop Paediatr 1990;10:329–34.

20. Vitarana UT, Jayasekera N, Withane N, Gubler DJ. Finding the cause

of dengue hemorrhagic fever outbreaks in Sri Lanka. Arbovirus

Research in Australia 1993;6:125–9.

21. Messer WB, Sivananthan K, Elvitigala J, Preethimala LD, Ramesh R,

Withana N, et al. Epidemiology of dengue in Sri Lanka before and

after the emergence of epidemic dengue hemorrhagic fever. Am J

Trop Med Hyg 2002;66:765–73.

22. de Silva AM, Sivananthan K, Withana N, Vorndam V, Gubler DJ.

Dengue 3 virus is responsible for recent epidemics of dengue hemor-

rhagic fever in Sri Lanka. In: Annual Meeting of the American

Society for Tropical Medicine and Hygiene. San Juan, Puerto Rico;

1998.

23. Lanciotti RS, Lewis JG, Gubler DJ, Trent DW. Molecular evolution

and epidemiology of dengue-3 viruses. J Gen Virol 1994;75:65–75.

24. Rico-Hesse R, Harrison LM, Salas RA, Tovar D, Nisalak A, Ramos

C, et al. Origins of dengue type 2 viruses associated with increased

pathogenicity in the Americas. Virology 1997;230:244–51.

25. Kobayashi N, Thayan R, Sugimoto C, Oda K, Saat Z, Vijayamalar B,

et al. Type-3 dengue viruses responsible for the dengue epidemic in

Malaysia during 1993–1994. Am J Trop Med Hyg 1999;60:904–9.

26. Rico-Hesse R. Molecular evolution and distribution of dengue virus-

es type 1 and 2 in nature. Virology 1990;174:479–93.

27. Lewis JA, Chang GJ, Lanciotti RS, Kinney RM, Mayer LW, Trent

DW. Phylogenetic relationships of dengue-2 viruses. Virology

1993;197:216–24.

28. Wang E, Ni H, Xu R, Barrett AD, Watowich SJ, Gubler DJ, et al.

Evolutionary relationships of endemic/epidemic and sylvatic dengue

viruses. J Virol 2000;74:3227–34.

29. Gubler D. Dengue and dengue hemorrhagic fever: its history and

resurgence as a global public health problem. In: Kuno DGaG, editor.

Dengue and dengue hemorrhagic fever. New York: CAB

International; 1997.

30. Isturiz RE, Gubler DJ, Brea del Castillo J. Dengue and dengue hem-

orrhagic fever in Latin America and the Caribbean. Infect Dis Clin

North Am 2000;14:121–40.

31. Briseno-Garcia B, Gomez-Dantes H, Argott-Ramirez E, Montesano

R, Vazquez-Martinez AL, Ibanez-Bernal S, et al. Potential risk for

dengue hemorrhagic fever: the isolation of serotype dengue-3 in

Mexico. Emerg Infect Dis 1996;2:133–5.

32. Gubler DJ, Sather GE, Kuno G, Cabral JR. Dengue 3 virus transmis-

sion in Africa. Am J Trop Med Hyg 1986;35:1280–4.

33. Kanesa-thasan N, Chang GJ, Smoak BL, Magill A, Burrous MJ,

Hoke CH Jr. Molecular and epidemiologic analysis of dengue virus

isolates from Somalia. Emerg Infect Dis 1998;4:299–303.

34. Gubler DJ, Trent DW. Emergence of epidemic dengue/dengue hem-

orrhagic fever as a public health problem in the Americas. Infect

Agents Dis 1993;2:383–93.

35. Gubler DJ, Suharyono W, Lubis I, Eram S, Sulianti Saroso J.

Epidemic dengue hemorrhagic fever in rural Indonesia. I. Virological

and epidemiological studies. Am J Trop Med Hyg 1979;28:701–10.

36. Lee E, Gubler DJ, Weir RC, Dalgarno L. Genetic and biological dif-

ferentiation of dengue 3 isolates obtained from clinical cases in Java,

Indonesia, 1976–1978. Arch Virol 1993;133:113–25.

37. Kochel TJ, Watts DM, Halstead SB, Hayes CG, Espinoza A, Felices

V. Effect of dengue-1 antibodies on American dengue-2 viral infec-

tion and dengue haemorrhagic fever. Lancet 2002;360:310–2.

Address for correspondence: Aravinda de Silva, Department of

Microbiology and Immunology, CB#7290, University of North Carolina,

Chapel Hill, NC 27599, USA; fax: (919) 962-8103; email:

[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 809

RESEARCH

Vibrio cholerae in O-group 139 was first isolated in

1992 and by 1993 had been found throughout the Indian

subcontinent. This epidemic expansion probably resulted

from a single source after a lateral gene transfer (LGT)

event that changed the serotype of an epidemic V. choler-

ae O1 El Tor strain to O139. However, some studies found

substantial genetic diversity, perhaps caused by multiple

origins. To further explore the relatedness of O139 strains,

we analyzed nine sequenced loci from 96 isolates from

patients at the Infectious Diseases Hospital, Calcutta, from

1992 to 2000. We found 64 novel alleles distributed among

51 sequence types. LGT events produced three times the

number of nucleotide changes compared to mutation. In

contrast to the traditional concept of epidemic spread of a

homogeneous clone, the establishment of variant alleles

generated by LGT during the rapid expansion of a clonal

bacterial population may be a paradigm in infections and

epidemics.

n epidemic of cholera began in Madras, India, in 1992

and within a year had spread across the Indian sub-

continent, with cases numbering in the millions (1,2).

Vibrio cholerae isolates from this epidemic had a previous-

ly unidentified serotype, subsequently designated as O139

Bengal (1,2). This new serotype appears to have resulted

when a lateral gene transfer (LGT) event occurred that

replaced the 22 kb of the wbf region (encoding the O1 anti-

gen) of a seventh pandemic V. cholerae O1 El Tor strain

with a 37-kb region encoding the O139 polysaccharide

(3–5). The epidemic spread rapidly through all age groups,

as persons with previous exposure to V. cholerae O1 were

not immune to O139 infection. Since 1992, O139 strains

have established endemicity in this geographic region and

account for a variable percentage of cholera cases every

year (6).

Genetic variation observed in O139 isolates has been

attributed to many causes. Variation in restriction fragment

length polymorphism (RFLP) analysis of rDNA genes (7)

and in recA sequence (8) has been interpreted as evidence

for multiple origins. Genetic variability in RFLP of the

CTX element (6) has been attributed to phage-mediated

recombination. Variation in antimicrobial susceptibility (9)

has been attributed to plasmid exchange in response to

selective pressure from drug use. The variation in pulsed-

field gel electrophoresis (PFGE) analysis of genomic

restriction fragments (6,10) has been attributed to point

mutations. Multilocus sequence typing (MLST), which has

been used in the evaluation of a number of other bacterial

species (11–14), provides an alternative method for meas-

uring genetic relatedness and has provided data for identi-

fying both point mutations and LGT events (14). MLST

has improved discriminatory power over PFGE in some

cases, e.g., Enterococcus (15) and Salmonella (16); how-

ever, in the case of Escherichia coli O157, it does not

because of an absence of sequence variation in the clonal-

ly derived isolates (17). A small MLST study of O139 iso-

lates of V. cholerae did not identify any LGT events (18).

To understand the evolutionary dynamics of V. choler-

ae O139, we sequenced segments from nine loci, including

seven that may be classified as traditional housekeeping

genes, one that carries the genes for cholera toxin, and

another that is next to the insertion sequence within the

O139 wbf region (3–5). Thus, the last two loci might be

expected to show LGT, because they are associated with

known mobile elements, but the other seven loci would not

be expected to show LGT. However, we found putative

LGT alleles at all nine loci in the 96 clonally related O139

isolates.

Materials and Methods

We evaluated nine loci—dnaE, lap, recA, pgm, gyrB,

cat, chi, rstR, and gmd—from 96 V. cholerae O139 isolat-

ed from patients seen at the Infectious Diseases Hospital,

Calcutta, from 1992 to 2000 (see Appendix, online only).

DNA was prepared from overnight cultures by using

PrepMan Ultra (Applied Biosystems Inc., Foster City, CA)

at the University of Maryland School of Medicine. Each

810 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Molecular Epidemiology of O139

Vibrio cholerae: Mutation, Lateral

Gene Transfer, and Founder Flush

Pallavi Garg,*† Antonia Aydanian,‡ David Smith,‡ J. Glenn Morris, Jr.,‡ G. Balakrish Nair,*†

and O. Colin Stine‡

*National Institute of Cholera and Enteric Diseases, Calcutta,

India; †International Centre for Diarrheal Diseases Research,

Dacca, Bangladesh; and ‡University of Maryland School of

Medicine, Baltimore, Maryland, USA

locus was amplified by using polymerase chain reaction

(PCR) with primers (Table 1) selected from a conserved

region of the locus, as determined by aligning sequences

from GenBank. Our primers selectively amplified the orig-

inal O139 rstR gene found in all isolates and not the addi-

tional one found in some recently inserted CTX elements

(19). The presence of amplified products was confirmed on

agarose gels. Purification of the products was performed by

using Millipore filters. The purified PCR products were

sequenced in both directions by using the same primers

used for amplification and Big Dye cycle sequencing kit

(ABI) in accordance with manufacturer’s instructions. The

fluorescently labeled products were separated and detected

by using either an ABI 377 or 3700 Automatic Sequencer

(ABI). The trace files were read by using Phred (20,21) and

Phrap (22). Low-quality sequence at the ends was trimmed,

and the contigs from each individual isolate were aligned

by using Clustal X (23). Variable nucleotides were identi-

fied manually. Isolates with identical alleles were identified

from a distance matrix obtained from PAUP (24). The alle-

les have been assigned GenBank accession numbers

AY297845 to AY297921.

The expected number of alleles that were a result of

point mutations was calculated. All point mutations were

assumed to occur independently; thus, the expected num-

ber of alleles with >2 nucleotides (nt) can be calculated,

and the excess number of observed alleles was attributed to

conspecific LGT of homologous genes. If one assumes

that p is the probability of seeing a single mutation in an

allele, the chance of seeing two mutations on the same

allele is p

; the probability of seeing three or more muta-

tions is p

. Probability can be calculated from the data by

dividing the number of alleles with a single nucleotide dif-

ference, 34, by 785, the number of alleles in which observ-

ing a point mutation is possible (the 6-bp deletion; the

recombinant gmd, recA alleles; and all duplicate novel

alleles were excluded). Thus, p equals 0.043, p

equals

0.0018, and p

equals 8 x 10

-5

. When these probabilities are

multiplied by the total number of alleles, 785, the expect-

ed number of alleles containing two independent point

mutations is 1.45, and the expected number containing

three or more is 0.06 (Figure 1).

Results

Each of the loci examined had a variable number of

observed alleles: 9 for dnaE, 20 for lap, 11 for rstR, 11 for

gmd, 2 for recA, 8 for pgm, 4 for gyrB, 7 for cat, and 5 for

chi. The most variable, lap with 20 alleles, was expected

because it is a highly variable locus when analyzed with

multilocus enzyme electrophoresis (25). The most com-

mon allele was present in 91% of isolates (n=87) for dnaE,

77% (n=86) for lap, 79% (n=90) for rstR, 82% (n=87) for

gmd, 99% (n=96) for recA, 90% (n=94) for pgm, 97%

(n=92) for gyrB, 93% (n=88) for cat, and 94% (n=89) for

chi. Thus, the pattern for each locus consists of a common

or ancestral allele and a series of rare alleles, as expected

for the expansion of a clone.

Of the 64 less frequent alleles, some result from LGT

and others from mutation. The three alleles with the largest

changes are unlikely to be due to point mutations. First, a

gmd allele that differed by 113 of the 360 bp sequenced,

when compared with sequences in GenBank using BLAST

(available from: URL: www.ncbi.nlm.nih.gov/BLAST/)

showed greater similarity to gmd from E. coli (AF061251)

than gmd from V. cholerae, consistent with LGT of a

homologous gene into the V. cholerae genome. Second, an

alternative recA allele that differs by 24 nt is likely to be

the result of LGT of a homologous gene. Although sub-

stantial, the number of nucleotide differences is not large

enough for the allele to be clustered with sequences from

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 811

RESEARCH

Table 1. Primers used for multilocus sequence typing

Locus

Primer 1

Primer 2

dnaE

CGRATMACCGCTTTCGCCG

GAKATGTGTGAGCTGTTTGC

lap

GAAGAGGTCGGTTTGCGAGG

GTTTGAATGGTGAGCGGTTTGCT

rstR

CGTGTTAGAGCACAC

GAGTGAATCGTCGTG

gmd

CCTTATGCKGTGGCRAA

CTWGGATCACCTAACA

recA

GAAACCATTTCGACCGGTTC

CCGTTATAGCTGTACCAAGCGCCC

pgm

CCKTCSCAYAACCCGCC

TCRACRAACCATTTGAADCC

gyrB

GAAGGBGGTATTCAAGC

GAGTCACCCTCCACWATGTA

cat

ATGGCTTATGAATCGATGGG

TCCCATTGCCATGCACC

chi

CAYGAYCCRTGGGCWGC

ACRTCTTCAATCTTGTC

Figure 1. Bar graph of the number of novel alleles (y-axis) with a

specific number of nucleotide differences from the ancestral allele.

Two alleles with 24-bp and 113-bp differences are excluded from

the graph.

V. mimicus, the closest sibling species to V. cholerae (8), a

finding that suggests that recombination occurred within V.

cholerae. Third, a lap allele had a 6-bp deletion and a sin-

gle nucleotide difference that may be the result of a dou-

ble-strand break repair.

We calculated that at least 26 putative conspecific LGT

events occurred in the 96 isolates studied. Figure 1 shows

the number of nucleotide differences between each novel

allele and the ancestral allele. If all point mutations are

assumed to occur independently, the expected number of

alleles with two or more variable nucleotides can be calcu-

lated and the excess number of observed alleles attributed

to conspecific LGT of homologous genes. The expected

number of alleles containing two independent point muta-

tions is 1.45, and the expected number containing three or

more is 0.06. Since 11 alleles were observed with 2 nt dif-

ferences, 9 more than expected, and 16 were observed with

3 differences, 16 more than expected, all of these alleles

probably did not occur through mutation; more likely,

these alleles are the result of LGT. Thus, we would esti-

mate that 26 alleles (9 + 16 + recA allele above) are puta-

tively due to conspecific LGT of homologous genes.

The putative conspecific LGT alleles, although fewer in

number (26 alleles) than the assumed number of mutation-

derived alleles (34 alleles), provide most of the nucleotide

differences between alleles. The 120 nt changes introduced

by conspecific LGT events are approximately three times

the 38 (34 single mutations + 4 2x2 double mutations)

introduced by mutation. This calculation is conservative:

The 26 conspecific LGT events may represent an underes-

timate of the number because some of the alleles differing

by <

1 nt may have resulted from LGT.

The analysis of all nine loci from each isolate was

based on the sequence type (ST). Each isolate was defined

by a 9-digit number composed of the assigned allele num-

ber at each of the nine loci in the following order: dnaE,

lap, rstR, gmd, recA, pgm, gyrB, cat, and chi. The most

common allele was arbitrarily assigned as number 1. Thus,

the ST of all the most common alleles is ST

1,1,1,1,1,1,1,1,1. Missing data were assigned the most

common allele. This assumption is conservative, mini-

mizes the observed amount of variation, and is consistent

with the preponderance of common alleles found at each

locus.

Fifty-one unique STs were found in the 96 isolates test-

ed, reflecting relatively extensive genetic diversity. The

overall average of 0.53 unique STs per isolate examined is

similar to that seen in every year including 1992 (Table 2).

Six STs occur more than once. As expected, the ancestral

ST:1,1,1,1,1,1,1,1,1, found in 40 isolates, occurred in all

years. Among the others, ST:1,1,2,1,1,1,1,1,1 was found

three times, once each in 1995, 1996, and 1997.

ST:1,2,1,1,1,1,1,1,1 and ST:1,1,7,1,1,1,1,1,1 were found

twice in 1992 and 1994, respectively. ST 1,1,1,6,1,1,1,1,1

was found once in 1998 and again in 1999.

ST:1,1,1,1,1,4,1,1,1 was found in 1995 and 1998. Since the

number of STs is large (51 types), and number of samples

in a collection period is small (8–13 samples; Table 2), STs

seen in multiple years must not only persist but also repre-

sent a substantial portion of the epidemic O139 V. choler-

ae population.

Five of the novel STs are related to other novel STs by

allelic change at another second or third locus. One

sequence type evolved into three related types found in

subsequent years (Figure 2a). The starred gmd allele is one

related to the E. coli sequence, and its presence in two dis-

tinct related STs in two different years demonstrates its

establishment in the population. That the pattern seen in

Figure 2b of ancestral alleles rstR 1 and chi 1 and two vari-

ant alleles, rstR 7 and chi 5, was found in all combinations

is indicative of an LGT event. Figure 2c-e shows three

additional groups of related sequences. In Figure 2a, b, and

d, the ST with the larger number of novel alleles occurred

in later years. In contrast, in Figure 2c and e, the ST with

the larger number of novel alleles occurred in the earlier

years. The lack of an overall temporal relationship may

result from the small sample size (8–13 isolates) in any

year.

One isolate, CRC5, is unusual because it has no

sequenced alleles of the ancestral type. Nevertheless, the

alleles from this isolate are closely related to those of the

ancestral type. Each CRC5 allele differs from the ancestral

allele by 7 nt for dnaE, 3 nt for lap, 4 nt for rstR, 24 nt for

recA, 6 nt for pgm, 10 nt for gyrB, and 4nt for chi. A com-

prehensive survey of the genetic distances for these loci

could determine the average distance between alleles for

each of these loci. The data would provide insight into

whether this isolate represents a second derivation of the

O139 clinical type from an environmental strain or if it is

a genetic outlier within the clonally related, but diversified,

O139 epidemic type.

Discussion

The emergence and pandemic spread of V. cholerae

O139 Bengal represented a chance to examine evolution of

a bacterial strain in the midst of a clonal expansion. Our

results are consistent with clonal expansion and subse-

quent divergence as described by Spratt and Maiden (26).

Putative recombinant alleles were found at all nine loci

among the 96 clonally related O139 isolates. One gmd

allele from V. cholerae was most similar to a gmd allele

from E. coli. The number of base-pair differences among

other alleles was higher than expected on the basis of a

simple computation for the accumulation of independent

mutations. This finding suggests that many of these events

were due to LGT. When we applied our criteria to the

812 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

novel alleles identified in a previous study (18), 11 of the

13 would be considered to have resulted from LGT, since

the number of nucleotide differences to the ancestral allele

varied from 4 to 19. Thus, for V. cholerae, like Neisseria,

Streptococcus, and other bacterial species (11–14), conspe-

cific recombination of homologous genes appears to be

common and responsible for most of the alleles with mul-

tiple nucleotide differences and the majority of the

nucleotide differences. The genetic variability at the nine

loci alters our understanding of evolution in bacteria,

showing that recombination in V. cholerae occurs fre-

quently and most nucleotide changes occur by means of a

recombination that can alter any gene.

The proportion of recombinants from conspecific

recombination, 3.5% (28/785) is greater than that from

transgeneric recombination (0.01% from the acquisition of

E. coli gmd by one isolate). One potential implication of a

greater rate of conspecific recombination may be that, over

time, it will maintain the species identity of each individ-

ual bacterium, despite the constant bombardment of

homologous genes from other genera. Although at first

glance the frequency of the novel sequence types appears

to conflict between our study and an earlier study (18), the

observations may be reconciled on the basis of both the

observed frequencies and the timing of the observations.

Both studies identified a common ancestral allele in from

77% to 99% of isolates in our study and a series of rare

alleles with 1–19 variant alleles for each locus. These stud-

ies reported 10% novel sequence types in 29 isolates that

were collected from “the first epidemic period,” from 1992

to 1993 (18). Our data from 1992 showed 33% novel

sequence types from a sample of nine. These data are not

statistically different (chi-square test=2.4, p=0.12).

However, the researchers’ estimate of frequency (18) is

more likely to be correct because of the larger sample size.

The dates of collection may also be important because our

collection of isolates from 1993 began in March, when the

number of O139 cases at the Infectious Diseases Hospital

rose from <10 to >80 per month, corresponding to a rapid

population expansion or flush. Thus, we can predict that

we would see substantial variation in our sample.

The genetic diversity was greater in the V. cholerae

O139 isolates than in other clinically associated clones. In

V. parahaemolyticus O3:K6, a pandemic strain, 94% of

strains were identical at four loci (N. Chowdhury et al.,

unpub. data). In E. coli O157, all 77 isolates were identical

at seven loci in spite of variation between isolates on

PFGE (17). Although V. parahaemolyticus and E. coli are

widespread pathogens, they differ from V. cholerae O139

because their population size has expanded much more

slowly.

Among O139 isolates, the substantial genetic diversity

found in the first year of the epidemic may reflect a

“founder flush” phenomenon. During times of population

expansion, i.e., a flush, any novel genotype with similar or

even slightly deleterious fitness compared to the founder

genotype will produce sufficient offspring to become

established in the population (27). A founder flush appears

to have occurred in the establishment of Helicobacter

pylori in a single person (28). Although other previous

descriptions of this phenomenon have been limited to

insects, specifically butterflies (29) and drosophilids (30),

we believe that the founder flush phenomenon may

become the paradigm for epidemic bacterial expansion in

individual patients and populations. This founder flush

phenomenon, in turn, has implications for our interpreta-

tion of “clonality” among epidemic isolates and for our

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 813

RESEARCH

Table 2. Number of isolates tested and distinct sequence types, by year

No.

1992

1993

1994

1995

1996

1997

1998

1999

2000

Total

Isolates examined

Novel sequence types

Novel sequence types per isolate examined

0.33

0.66

0.5

0.7

0.62

0.55

0.58

0.33

0.88

0.53

Figure 2. Five groups of related sequence types of Vibrio cholerae

O139.

understanding of factors that contribute to the emergence

of new pathogenic strains.

This work was supported in part by an American Society for

Microbiology travel fellowship to Dr. Garg.

Dr. Garg completed this work as part of her Ph.D. disserta-

tion at University of Calcutta. She is a postdoctoral fellow at the

University of Maryland School of Medicine, pursuing studies of

thrombospondin in human microvascular endothelial cells.

References

1. Bhattacharya SK, Bhattacharya MK, Nair GB, Dutta D, Deb A,

Ramamurthy T, et al. Clinical profile of acute diarrhoea cases infect-

ed with the new epidemic strain of Vibrio cholerae O139: designation

of the disease as cholera. J Infect 1993;27:11–8.

2. Nair GB, Ramamurthy SK, Mukhopadhyay T, Garg S, Bhattacharya

MK, Takeda T, et al. Spread of Vibrio cholerae O139 Bengal in India.

J Infect Dis 1994;169:1029–34.

3. Comstock LE, Johnson JA, Michalski JM, Morris JG, Kaper JB.

Cloning and sequence of a region encoding surface polysaccharide of

Vibrio cholerae O139 and characterization of the insertion site in the

chromosome of Vibrio cholerae O1. Mol Microbiol 1966;19:815–26.

4. Stroeher UH, Parasivam G, Dredge BK, Manning PA. Novel Vibrio

cholerae O139 genes involved in lipopolysaccharide biosynthesis. J

Bacteriol 1997;179:2740–7.

5. Bik EM, Bunschoten AE, Willems RJL, Chang ACY, Mooi FR.

Genetic organization and functional analysis of the otn DNA essen-

tial for cell-wall polysaccharide synthesis in Vibrio cholerae O139.

Mol Microbiol 1996;20:799–811.

6. Basu A, Garg P, Datta S, Chakraborty S, Bhattacharya T, Yamasaki S,

et al. Vibrio cholerae O139 in Calcutta, 1992–1998: incidence, antibi-

ograms, and genotypes. Emerg Infect Dis 2000;6:139–47.

7. Faruque SM, Saha MN, Asadulghani, Bag PK, Bhadra RK,

Bhattacharya SK, et al. Genomic diversity among Vibrio cholerae

O139 strains isolated in Bangladesh and India between 1992 and

1998. FEMS Microbiol Lett 2000;184:279–84.

8. Stine OC, Sozhamannan S, Gou Q, Zheng S, Morris JG, Johnson JA.

Phylogeny of Vibrio cholerae based on recA sequence. Infect Immun

2000;68:7180–5.

9. Garg P, Chakraborty S, Basu I, Datta S, Rajendran K, Bhattacharya T,

et al. Expanding multiple antibiotic resistance among clinical strains

of Vibrio cholerae isolated from 1992–7 in Calcutta, India. Epidemiol

Infect 2000;124:393–9.

10. Kurazono H, Yamasaki S, Ratchtrachenchai O, Nair GB, Takeda Y.

Analysis of Vibrio cholerae O139 Bengal isolated from different geo-

graphical areas using macrorestriction DNA analysis. Microbiol

Immunol 1996;40:303–5.

11. Maiden MCJ, Bygraves JA, Feil E, Morelli G, Russell JE, Urwin R,

et al. Multilocus sequence typing: a portable approach to the identifi-

cation of clones within populations of pathogenic microorganisms.

Proc Natl Acad Sci U S A 1998;95:3140–5.

12. Dingle KE, Colles FM, Wareing DR, Ure R, Fox AJ, Bolton FE, et al.

Multilocus sequence typing system for Campylobacter jejuni. J Clin

Microbiol 2001;39:14–23.

13. Enright MC, Day NP, Davies CE, Peacock SJ, Spratt BG. Multilocus

sequence typing for characterization of methicillin-resistant and

methicillin-susceptible clones of Staphylococcus aureus. J Clin

Microbiol 2000;38:1008–15.

14. Feil EJ, Smith JM, Enright MC, Spratt BG. Estimating recombina-

tional parameters in Streptococcus pneumoniae from multilocus

sequence typing data. Genetics 2000;154:1439–50.

15. Nallapareddy SR, Duh RW, Singh KV, Murray BE. Molecular typing

of selected Enterococcus faecalis isolates: pilot study using multilo-

cus sequence typing and pulsed-field gel electrophoresis. J Clin

Microbiol 2002;40:868–76.

16. Kotetishvili M, Stine OC, Kreger A, Morris JG Jr., Sulakvelidze A.

Multilocus sequence typing for characterization of clinical and envi-

ronmental Salmonella strains. J Clin Microbiol 2002;40:1626–35.

17. Noller AC, McEllistrem MC, Stine OC, Morris JG, Boxrud DJ, Dixon

B, et al. Lack of DNA sequence diversity among Escherichia coli

O157:H7 isolates that are distinct by pulsed-field gel electrophoresis.

J Clin Microbiol 2003;41:675–9.

18. Farfan M, Minana-Galbis D, Fuste MC, Loren JG. Allelic diversity

and population structure in Vibrio cholerae O139 Bengal based on

nucleotide sequence diversity. J Bacteriol 2002;184:1304–13.

19. Mukhopadhyay AK, Chakraborty S, Takeda Y, Nair GB, Berg DE.

Characterization of VPI pathogenicity island and CTXΦ prophage in

environmental strains of Vibrio cholerae. J Bacteriol

2001;183:4737–46.

20. Ewing B, Greene P. Base-calling of automated sequencer traces using

Phred. II. Error probabilities. Genome Res 1998;8:186–94.

21. Ewing B, Hillier L, Wendl M, Greene P. Base-calling of automated

sequencer traces using Phred. I. Accuracy assessment. Genome Res

1998;8:175–85.

22. Green P. 1998. Phrap, SWAT and CrossMatch. Available from: URL:

http://www.washington.edu

23. Jeanmougin F, Thompson JD, Gouy M, Higgins DG, Gibson TJ.

Multiple sequence alignment with Clustal X. Trends Biochem Sci

1998;23:403–5.

24. Swofford D. PAUP* beta version. Sunderland (MA): Sinauer Assoc;

2000.

25. Beltran P, Delgrado G, Navarro A, Trujillo F, Selander RK, Cravioto

A. Genetic diversity and population structure of Vibrio cholerae. J

Clin Microbiol 1999;37:581–90.

26. Spratt BG, Maiden MCJ. Bacterial population genetics, evolution and

epidemiology. Philos Trans R Soc Lond Biol Sci 1999;354:701–10.

27. Wallace B. Basic populations genetics. New York: Columbia

University Press; 1981. p. 429–32.

28. Israel DA, Salama N, Krishna U, Rieger UM, Atherton JC, Falkow S,

et al. Helicobacter pylori genetic diversity within the gastric niche of

a single host. Proc Natl Acad Sci U S A 2001;98:14625–30.

29. Ford HD, Ford EB. Fluctuation in numbers, and its influence on vari-

ation, in Melitaea aurina, Rott. (Lepidoptera). Transactions of the

Royal Entomological Society London 1930;78:345–51.

30. Carson HL. The population flush and its genetic consequences. In:

Lewontin RC, editor. Population biology and evolution. Syracuse

(NY): Syracuse University Press;1968. p. 123–37.

Address for correspondence: O. Colin Stine, Department of

Epidemiology and Preventive Medicine, University of Maryland School

of Medicine, Baltimore, MD 21201, USA; fax: 410-706-1644; email:

[email protected]

814 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

To evaluate the role of amoeba-associated bacteria as

agents of ventilator-associated pneumonia (VAP), we test-

ed the water from an intensive care unit (ICU) every week

for 6 months for such bacteria isolates; serum samples and

bronchoalveolar lavage samples (BAL) were also obtained

from 30 ICU patients. BAL samples were examined for

amoeba-associated bacteria DNA by suicide-polymerase

chain reaction, and serum samples were tested against

ICU amoeba-associated bacteria. A total of 310 amoeba-

associated bacteria from 10 species were isolated. Twelve

of 30 serum samples seroconverted to one amoeba-asso-

ciated bacterium isolated in the ICU, mainly Legionella

anisa and Bosea massiliensis, the most common isolates

from water (p=0.021). Amoeba-associated bacteria DNA

was detected in BAL samples from two patients whose

samples later seroconverted. Seroconversion was signifi-

cantly associated with VAP and systemic inflammatory

response syndrome, especially in patients for whom no eti-

ologic agent was found by usual microbiologic investiga-

tions. Amoeba-associated bacteria might be a cause of

VAP in ICUs, especially when microbiologic investigations

are negative.

ospital-acquired pneumonia occurs in 0.5% to 1% of

admitted patients admitted, representing 10% to 15%

of all nosocomial infections; pneumonia is the most com-

mon cause of nosocomial infection in intensive-care units

(ICUs) (1). This pneumonia is associated with high death

rates. As the etiologic agent of pneumonia remains

unknown in 20% to 50% of cases (2), identifying new lung

pathogens is a major public health goal. Aquatic bacteria

such as Legionella spp., Pseudomonas spp., Stenotro-

phomonas spp., Burkholderia spp., or Acinetobacter spp.

may colonize in hospital water supplies and have previous-

ly been shown to be causally associated with cases of noso-

comial infections (3). Free-living amoebae have been

shown to be a reservoir of pathogens such as Legionella sp.,

Burkholderia picketti, and Cryptococcus neoformans (4–7).

The most studied amoebae-associated bacterium is

Legionella pneumophila, the agent of Legionnaires’ disease

(8), which frequently results from exposure to contaminat-

ed aerosols. Additional amoeba-associated bacteria might

be implicated in community-acquired pneumonia, includ-

ing Legionella-like amoebal pathogens (9) and members of

the genus Parachlamydia (10). As part of research into the

diversity of bacterial agents associated with amoebae in

hospital water supplies, we have identified a new α-

Proteobacteria belonging to the Bradyrhizobiaceae family

(11–13). We demonstrated that patients with nosocomial

pneumonia hospitalized in the vicinity of the contaminated

water in a public hospital of our city have elevated antibody

titers to these bacteria (14). In this study, we performed the

same kind of analysis but focused our work on a single ICU

during a 6-month period. Amoeba-associated bacteria were

periodically evaluated in all ICU water taps. To evaluate

contact of patients hospitalized in this ICU and amoeba-

associated bacteria in the water, serum and bronchoalveolar

lavages (BAL) samples were periodically sampled. Serum

samples were tested in an immunofluorescence assay

against the isolated bacteria to detect seroconversions, and

DNA of these bacteria were detected in BAL samples by

suicide-polymerase chain reaction (PCR) (15,16), a PCR

technique without positive controls that incorporates “dis-

posable” primers to avoid false-positive results. The second

part of this work was to evaluate if exposure to the amoe-

ba-associated bacteria in the ICU could be associated with

disease. Thus, we specifically studied some clinical mark-

ers of infection, including fever, systemic inflammatory

response syndrome (SIRS), and pneumonia for patients

admitted to the ICU. As a definition of pneumonia based

only on clinical and roentgenographic criteria has been crit-

icized for low specificity (17–19), we used strict criteria in

the definition. These criteria were applied to cases in which

bacterial documentation was negative to determine if dis-

ease observed in patients hospitalized in an ICU may be

attributed to amoeba-associated bacteria.

Materials and Methods

All patients admitted to the ICU during a 26-week

period who needed intubation and mechanical ventilation

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 815

RESEARCH

Amoeba-Resisting Bacteria and

Ventilator-Associated Pneumonia

Bernard La Scola,* Ioanna Boyadjiev,*† Gilbert Greub,* Atieh Khamis,*

Claude Martin,† and Didier Raoult*†

*Unité des Rickettsies, Marseille, France; and †Hôpital Nord,

Marseille, France

were included. Patients were evaluated at admission by the

Acute Physiology and Chronic Health Evaluation II score

(20). At admission and every week thereafter, temperature,

leukocytes and platelet counts, hepatic enzymes, and pres-

ence of SIRS and ventilator-associated pneumonia (VAP)

were recorded. Serum samples were obtained from

patients at admission, every 7 days afterwards, and at dis-

charge. BAL was obtained at admission immediately after

intubation by using protected mini-bronchoalveolar lavage

(Combicath, Plastimed, Le-Plessy-Bouchard, France),

then performed when a lung infiltrate suggestive of pneu-

monia appeared, and repeated every week until pneumonia

resolved. BALs are part of the routine diagnosis and fol-

low-up of pneumonia in the ICU and were not performed

specifically for the study. Informed consent was obtained

from the patient’s family, according to French legislation.

Data about bacteria isolated from blood cultures, lung

secretions, and urine by conventional procedures were

recorded. When isolated from urine and lung secretions,

bacteria were only considered as pathogenic when concen-

tration in the specimen was >

and >

CFU/mL,

respectively (21).

Definitions of VAP and SIRS were based on previously

published criteria (22–25), but to increase specificity, we

limited our study to severe cases of VAP by adding strict

criteria (Table 1). Taps and ice machine water were sam-

pled every other week, as previously reported (11). The

procedure for isolating bacteria from water and lung secre-

tions by using cocultivation with Acanthamoeba polypha-

ga followed by subculture onto BCYE agar plates has been

detailed elsewhere (11,26). Bacteria were identified by

using 16S rRNA gene sequence comparisons as previous-

ly described (27). Legionella species were identified by

using mip gene amplification and sequencing (28).

Twenty bacterial antigens were tested by microim-

munofluorescence as previously reported (14,29).

Bacterial species isolated from the ICU water during the

studied period and 10 other species previously isolated in

the same conditions from other sites (Bosea eneae, B.

vestrisii, B. thiooxydans, Mesorhizobium amorphae,

Azorhizobium caulinodans, Afipia felis, A. felis

genospecies A, A. clevelandensis, A. birgiae, and A. mas-

siliae [11–14]) were tested. The serologic tests were per-

formed on the first serum samples from all patients admit-

ted to the ICU; these patients were available for a second

sample (30 patients). The control group comprised 10

patients in the same ICU. These patients had shorter stays

and had samples taken at admission but was not available

(100 patients) and 114 patients with other diseases, includ-

ing Q fever (10 samples), trench fever (5 samples),

tularemia (8 samples), Mediterranean spotted fever (10

samples), epidemic typhus (5 samples), syphilis (10 sam-

ples), cat scratch disease (5 samples), pneumonia caused

by Chlamydia pneumoniae (5 samples), C. psittaci (5 sam-

ples), Mycoplasma pneumoniae (10 samples), L. pneu-

mophila (10 samples), hepatitis C virus (5 samples), infec-

tions caused by cytomegalovirus (5 samples), Epstein Barr

virus (11 samples), and HIV (10 samples). Serum samples

were diluted at 1:25, 1:50, and 1:100 for immunoglobulin

(Ig) G and IgM determination. The cutoff titer for a posi-

tive detection was determined as the lowest titer for which

all first serum samples and control serum samples were

negative. Then, all patients’ serum samples were tested and

serial twofold dilutions from 1:50 to 1:1600 were made on

these samples with titers at least equal to the cutoff titer.

Tests by Western blot were performed as previously

described (29) for any patient with a seroconversion.

DNA was extracted from BAL samples by using

QIAMP Tissue kit (QIAGEN, Hilden, Germany), accord-

ing to the manufacturer’s instructions. PCR detection was

attempted on all BAL samples for bacteria against which at

least one seroconversion was generated. DNA extraction

from serum samples was performed in a laboratory other

than the one in which the isolates were identified to avoid

vertical contamination from previous amplifications, and

no positive control was used to avoid horizontal contamina-

tion from the same experiment (16). We used a nested PCR

that incorporated two primer pairs used only once (Table 2)

followed by sequencing and comparison to the targeted

sequence, as previously described for suicide-PCR (15,16).

All samples were tested the same day in the same assay.

Sample testing was blinded, and positive amplicons were

sequenced. Negative controls consisted of BAL samples

from 10 patients of the same ICU obtained at admission,

BAL samples from 200 patients with nosocomial pneumo-

nia hospitalized in other medical centers of the city, water

samples, and a suspension of A. polyphaga. At least one

negative control was used for every two serum samples.

816 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 1. Definition criteria for ventilator-associated pneumonia (VAP) and systemic inflammatory response syndrome (SIRS)

VAP

SIRS

Unexplained VAP

Unexplained SIRS

New and persistent roentgenographic lung

infiltrate and new onset of:

a) Increase in white blood cells >10 g/L

b) Fever or hypothermia (>38°C or <36°C)

c) Purulent sputum

d) Duration of at least 2 weeks

e) PaO

/FiO

ratio <100

At least two of:

a) T° >38°C or <36°C

b) Heart rate >90/min

c) Respiratory rate >20/min or PaCO

<32 mmHg

d) Leukocytes >12 or <4 g/L or immature

(band) forms

Lack of recovery of

bacteria from:

a) Lung secretions

b) Blood cultures

Lack of recovery of bacteria from:

a) Lung secretions

b) Blood cultures

c) Urine

Comparisons of demographic, clinical, and laboratory

data between patients with evidence of amoeba-associated

bacteria contact (seroconversion or positive detection in

BAL samples) were performed by using chi square and

Mann-Whitney tests, respectively. The tested variables

were age; sex; an underlying disease; severity score; intu-

bation duration; hospitalization duration; SIRS; death;

increase of hepatic enzymes, platelet count, and leuko-

cytes; VAP; and fever. We also compared demographic,

clinical, and laboratory data between patients with and

without unexplained VAP, fever, and SIRS. Multivariate

analysis adjusted for age, sex, prolonged intubation, and an

underlying disease was performed to confirm observed

associations. STATA software (v. 7.0, Stata Corporation,

College Station, TX) was used for analysis.

Results

Ten species (310 isolates) were identified from 864

water samples (Table 3). B. massiliensis and L. anisa were

the two most commonly isolated species (62.3% of the 310

isolates). In the ICU admission rooms , isolation of L.

anisa ranged from 75% to 100% of the tested taps during

the first 20 weeks; whereas all were negative during the

last 6 weeks, after taps were changed. No amoeba-associ-

ated bacteria were isolated from BAL.

Ninety serum samples from the 30 patients and the 214

control serum samples were tested by an immunofluores-

cence assay for IgG and IgM on the 20 amoeba-associated

bacteria antigens (12,160 tests). The 30 first serum sam-

ples and the samples from blood donors did not have an

IgG titer of >1:50 or IgM titer >1:25. Cutoff titers for pos-

itive serologic tests with 100% specificity are shown in

Table 4. Twelve (40%) patients seroconverted from 10 to

35 days after admission to at least one antigen; 10 showed

IgM antibodies (Table 5). Five patients seroconverted to L.

anisa, six to B. massiliensis, including one to both, one to

L. quinlivanii, and one to A. broomeae. Western blots con-

firmed the seroconversions, with the appearance of sever-

al reacting bands on convalescent-phase serum samples

(Figure). Patients also seroconverted to amoeba-associated

bacteria not detected in the water in this study: two to

A. clevelandensis (patient 2, IgG=1:100 and IgM=1:100;

patient 8, IgG=1:50 and IgM=1:200;) and two to A. felis

(patient 2, IgG=1:100 and IgM=1:100; patient 9,

IgG=1:1600 and IgM=1:25). Seroconversions were signif-

icantly more frequent against amoeba-associated bacteria

obtained in this ICU than against amoeba-associated bac-

teria isolated in previous studies: 13 of 300 tests versus 4

of 300 tests and 12 of 30 patients versus 4 of 30 patients

(p=0.046 and p=0.039, respectively). Patients also sero-

converted more frequently to the most commonly isolated

bacteria, L. anisa and B. massiliensis (>50 isolates,

p=0.021).

Table 6 shows the clinical characteristics of patients

with serologic evidence of exposure to amoeba-associated

bacteria isolated in the ICU. Analysis for risk and potential

confounding factors did not indicate differences between

patients with or without seroconversion. However, sero-

conversion was statistically associated with VAP

(p=0.026), unexplained VAP (p=0.034), SIRS (p=0.024),

and unexplained SIRS (p=0.045). Multivariable logistic

regression demonstrated that seroconversion was inde-

pendently associated with VAP even after we adjusted for

intubation duration, hospitalization duration, number of

serum samples, and underlying disease (p=0.014 to 0.030).

The DNA of L. anisa and B. massiliensis were each

detected once in the 66 BAL samples from 30 patients. For

two samples, seroconversion to the identified bacteria, L.

anisa and B. massiliensis, respectively, was observed 4

and 2 weeks after the PCR was positive in BAL. None of

the 210 control patients was positive for these bacteria in

BAL samples compared to 2 of 30 patients in the ICU

(p<0.01).

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 817

RESEARCH

Table 2. Polymerase chain reaction primers used for amplification and sequencing of bacterial DNA within human samples

Primers

Bosea-related strain

Legionella anisa

Afipia broomeae

L. quinlivanii

External forward

5′-TGCGAGTGTA

GAGGTGAAATT-3′

5′-TATTGGTGC

TGATTTAGGAA-3′

5′-TCTTTTGTGCG

GGAAGATAATG-3′

5′-TTGTTGATGTTT

GTTTTGAGACC-3′

External reverse

5′CGCTCGTTG

CGGGACTTAA-3′

5′-GCTAAGTCTGAA

GGTACA-3′

5′-TAAACTTTCCAA

CGGCTGGCAT-3′

5′-TTCAACACTTCT

TTCATCTGATC-3′

Internal forward

5′-GAGGTGAAA

TTCGTAGATATT-3′

5′-GCCCAATTG

ATTTTGACAG-3′

5′-GCTAACTTCGT

GCCAGCAG-3′

5′-TCCAAGAATAA

AAGGGGATTG-3′

Internal reverse

5′-GAGCTGACG

ACAGCCAT-3′

5′-GCATTAATTGT

AATGCTTCA-3′

5′-GTTTGCTCC

CCACGCTTTC-3′

5′-CCATACCAT

CCTGTAAGCCTT-3′

Table 3. Identification of the 310 bacterial strains isolated by

using amoebal co-culture procedure

Species

No. of isolates

Legionella anisa

126

Bosea massiliensis

Rasbo bacterium

Bradyrhizobium liaoningense

L. quinlivanii

L. pneumophila

L. rubrilucens

L. worsleiensis

B. japonicum

Afipia broomeae

The death rate was 33.3%; disease was mainly associ-

ated with fever (96.6%), VAP (56.6%), and SIRS (76.6%).

No microbial etiologic agent was found in 18 (62%) of 29

patients with fever, 10 (43%) of 23 patients with SIRS, and

in 8 (47%) of 17 patients with VAP. VAP was significantly

associated with duration of hospitalization (median hospi-

talization days/interquartile range of 23/18–41 with VAP

versus 14/10–21 without VAP, p=0.04), SIRS (16/17 with

VAP versus 7/13 without VAP, p=0.025) and seroconver-

sion to amoeba-associated bacteria (10/17 with VAP versus

2/13 without VAP, p=0.026). No other statistical difference

was observed between patients with VAP and without VAP

in terms of demographic, clinical, and paraclinical data,

risks, and potential confounding factors. Most patients

received cephalothin for the first 3 days of hospitalization

as an antibiotic prophylaxis. The most commonly used

antibiotics in patients with VAP were third-generation

cephalosporins. Patients with VAP received antibiotics

more frequently and for more than 1 week (16/17 with

VAP versus 7/13 without VAP, p=0.025).

Discussion

We confirmed that several amoeba-associated bacteria

are common in the water in ICU; we recovered 310 amoe-

ba-associated bacteria isolates from 10 species from the

water of the ICU. As exposure to a microorganism is a pre-

requisite to infectious disease, we first evaluated contact of

patients hospitalized in the ICU with amoeba-associated

bacteria in water. We found that 12 (40%) of 30 of patients

seroconverted to amoeba-associated bacteria and that these

seroconversions were significantly more common against

local isolates than to amoeba-associated bacteria from

other ICUs. Moreover, antibody response parallels that of

water contamination with B. massiliensis and L. anisa,

which cause 83% of seroconversions identified in 62% of

isolates (p=0.021). The cutoff titers, chosen to have 100%

specificity, and the detection of several reactive antigens in

the Western blots (Figure) suggest that these seroconver-

sions represent specific serologic reactions. Detection of

antibodies reacting to amoeba antigens would also be

important because infections could occur through inhala-

tion (perhaps after colonization) of infected amoebae act-

ing as a “Trojan horse” (6). In future studies, isolation of

amoebae in the aquatic environment as in the BAL sam-

ples of patients will be performed for further use as anti-

gens.

Patient exposure to amoeba-associated bacteria from

the ICU was also evident in results of amoeba-associated

bacteria DNA detection in BAL samples from 2 of 30

patients as compared with 0 of 210 control patients

(p<0.01). Moreover, for these patients, bacterial DNA was

detected in the BAL sample before the seroconversion to

the same bacteria 2 to 4 weeks after admission. This rate is

compatible with an acute infection occurring during hospi-

talization rather than colonization of the respiratory tracts

of patients. Contamination is unlikely when using a sui-

cide-PCR procedure (15,16). Isolation of these agents was

probably hampered by the antibiotic prophylaxis instituted

at admission in this ICU for trauma patients and likely

explains the lack of amoeba-associated bacteria isolation

in BAL samples.

Among the 30 patients, VAP occurred frequently

(56.6%) and was associated with hospitalization duration,

as previously reported (30,31). Of the patients with VAP,

58.8% seroconverted to amoeba-associated bacteria, as

compared to 15.4% of the remaining 13 patients

(p=0.026). In patients with seroconversion, SIRS was also

more prevalent (p=0.024). The percentages of unexplained

VAP and unexplained SIRS were four times more common

in amoeba-associated bacteria seroconverters than in non-

seroconverters and remained statistically significant in

spite of the small population studied (Table 6). Thus,

amoeba-associated bacteria may be a common cause of

unexplained VAP and SIRS.

Finally, we identified a cryptic outbreak in the ICU

caused by L. anisa, a pathogen commonly encountered in

the environment (32,33), only implicated in a few epi-

demics of Pontiac fever (34,35) and four cases of legionel-

losis (26,36–38). Five (16.7%) of 30 patients were infect-

ed with this bacterium, considered a relatively rare

pathogen. Serologic tests are not currently used for L.

anisa, and no urinary antigen test is available. Therefore,

818 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 4. Definition of cutoff titers for positive serologic tests and

100% specificity according to antigen tested by using the 224

control serum samples

Antigen

IgG

IgM

Legionella anisa

>1:50

>1:25

Bosea massiliensis

>1:50

>1:25

Rasbo bacterium

>1:100

>1:25

Bradyrhizobium liaoningense

>1:50

>1:25

L. quinlivanii

>1:100

>1:25

L. pneumophila

>1:50

>1:25

L. rubrilucens

>1:200

>1:25

L. worsleiensis

>1:200

>1:25

B. japonicum

>1:100

>1:25

Afipia broomeae

>1:200

>1:25

Bosea eneae

>1:800

>1:25

B. vestrisii

>1:100

>1:25

B. thiooxydans

>1:50

>1:25

Mesorhizobium amorphae

>1:100

>1:25

Azorhizobium caulinodans

>1:100

>1:25

Afipia felis

>1:100

>1:25

A. felis genospecies A

>1:100

>1:25

A. clevelandensis

>1:100

>1:25

A. birgiae

>1:50

>1:25

A. massiliae

>1:50

>1:25

Ig, immunoglobulin.

diagnosis based only on isolation may explain why L.

anisa is so rarely reported. However, its absence in the

BAL samples from the 200 patients from other ICUs

shows that local epidemiology plays a major role. This

study confirms that L. anisa is common in the environment

(32,33).

Members of the Bosea genus are gram-negative, oxi-

dase-positive, catalase-positive rods belonging to the α-2

subgroup of Proteobacteria. All are motile. They grow well

on BCYE agar from 25°C to 37°C but do not grow or grow

weakly on Columbia agar with 5% sheep blood. Colonies

are smooth, mucoid, round, and cream colored and are ure-

ase positive and α-hemolytic on Columbia agar with 5%

sheep blood and 0.2% yeast extract. B. massilensis are

negative in assays for arginine dihydrolase activity, esculin

and gelatin hydrolysis, β-galactosidase activity, maltose

assimilation, and acid production by fermentation or oxi-

dation of substrates tested in API 50 CH (Biomérieux,

Marcy l’étoile, France), especially D-glucose, D-fructose,

D-mannose, and sucrose. The species of the Bosea genus

have high MICs to penicillin and amoxicillin and low

MICs to doxycycline (13). In co-culture with A. polypha-

ga, B. vestrisii, B. eneae, and B. massiliensis are phagocyt-

ed and form progressively large vacuoles that lead to

amoebal lysis; however, they have never been reported as

pathogenic agents before. As B. massiliensis has not been

tentatively isolated elsewhere, whether our findings reflect

a local phenomenon or whether the bacterium is widely

encountered is not known. We think that our data support

the role of B. massiliensis in severe VAP, but confirmation

is needed to definitely establish a role.

Our study indicates that most patients with VAP

received β-lactam agents, mainly amoxicillin-clavulanic

acid and third-generation cephalosporin. These antibiotics

may have inhibited bacterial culture, which explains why

no amoeba-associated bacteria were isolated from BAL

samples. B. massiliensis has low MICs (<0.5 mg/L) to cef-

triaxone, doxycycline, rifampin, and erythromycin (13).

However, β-lactam agents are also active in vitro on

Legionella spp., but animal models and clinical studies

have demonstrated their inefficacy in the treatment of

legionellosis (39).

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 819

RESEARCH

Table 5. Antibody titers of 12 serum samples with seroconversion to at least one of the bacteria isolated in the intensive care unit

Bosea massiliensis

Legionella anisa

L. quinlivanii

Afipia broomeae

Case

Wk of sampling

IgG

IgM

IgG

IgM

IgG

IgM

IgG

IgM

<1:50

<1:25

1:50

1:100

1:800

1:50

<1:50

<1:25

1:400

1:100

<1:50

<1:25

<1:50

1:200

<1:50

<1:25

1:200

1:800

<1:50

<1:25

<1:50

<1:25

1:400

1:50

1:400

1:50

1:25

1:800

1:50

<1:50

<1:25

1:50

<1:50

<1:25

1:400

1:50

<1:50

<1:25

1:50

<1:25

1:100

<1:25

1:200

<1:25

<1:50

<1:25

1:400

<1:25

<1:50

<1:25

1:200

1:50

<1:50

<1:25

1:400

1:25

<1:50

<1:25

1:400

1:25

1:200

1:25

Patients with positive PCR results in bronchoalveolar lavage samples (patient 1, L. anisa; Patient 2, B. massiliensis); Ig, immunoglobulin. All patients were sampled at

admission (wk 1); titers in bold are those greater than or equal to defined cutoff titers (Table 4).

The results of our study confirm that the bacteriologic

tests of hospital water supplies is largely ignored. Our work

demonstrates that patients are exposed specifically to the

most common water amoeba–associated bacteria in their

environment, as evidenced by seroconversion against these

bacteria and DNA of these bacteria in BAL samples. The

route of infection, even if caused by aerosols generated in

the ICU, remains unclear. Patients of this ICU sometime

receive water through nasogastric tubes but only bottled,

sterile water. However, we cannot exclude mistakes caused

by not following recommended procedures. Patients for

whom exposure to these bacteria is supported by serocon-

version or DNA detection in BAL samples have unex-

plained VAP and SIRS more commonly. We speculate that

amoeba-associated bacteria in the environment of intubat-

ed patients may concurrently cause unexplained infections

and cryptic outbreaks. Research of new etiologic agents of

pneumonia in ICUs should be based on environmental

study of each ICU since ecologic findings of amoeba-asso-

ciated bacteria in water points in hospital vary.

Acknowledgments

The authors are indebted to Lina Barrassi for technical help

and Kelly Johnston and John Dumler for reviewing the manu-

script.

This work was supported by the Programme Hospitalier de

Recherche Clinique.

Dr. La Scola is associate professor at Marseilles Medicine

Faculty. He is a member of the Unité des Rickettsies (CNRS UMR

6020, World Health Organization reference center for rickettsiae

and rickettsial diseases). His fields of interest are isolation and

description of fastidious bacteria, including Coxiella, Rickettsia,

Bartonella, Tropheryma, and amoeba-associated bacteria.

820 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure. Western blot showing seroconversions in immunoglobulin

(Ig) G (Lanes 1, 3, 5, 7) and IgM (Lanes 2, 4, 6, 8) of patient 2

against Bosea massiliensis (Lanes 1 to 4) and patient 9 against

Legionella anisa (Lanes 5 to 8). Lanes 1, 2, 5, 6: acute-phase

sera; Lanes 3, 4, 7, 8: convalescent-phase sera.

Table 6. Clinical characteristics of patients with or without seroconversion to one of the amoeba-associated bacteria

Clinical characteristics

Seroconversion (N=12)

No seroconversion (N=18)

p value

Demographic data

Median age in y (IQR)

35 (25–43)

24 (21–52)

0.85

Male (%)

10 (83.3)

15 (83.3)

Risk and potential confounding factors

Underlying disease (%)

2 (16.6)

3 (16.6)

Circulation injury (%)

9 (75)

14 (77.8)

Median APACHE II

score (IQR)

21 (14–4)

23 (16–34)

0.12

Intubation in ICU (%)

8 (66.6)

7 (38.9)

0.26

Median hospitalization days (IQR)

25 (19–41)

17 (10–23)

0.094

Median intubation duration in days (IQR)

11 (7–20)

0.8

Median number of serum samples (IQR)

3 (3–5)

3 (2–4)

0.12

Clinical data

VAP (%)

10 (83.3)

7 (38.9)

0.026

Unexplained VAP (%)

6 (50)

2 (11.1)

0.034

Fever > 38.5 °C (%)

12 (100)

17 (84.4)

Unexplained fever (%)

7 (58.3)

11 (39.3)

0.27

SIRS (%)

12 (100)

11 (61.1)

0.024

Unexplained SIRS (%)

7 (58.3)

3 (14.3)

0.045

Death (%)

2 (16.7)

8 (44.4)

0.23

Paraclinical data

Leukocytes > 12 g/L (%)

12 (100)

14 (77.8)

0.13

Platelets > 500 g/L (%)

5 (41.7)

6 (33.3)

0.7

PCR detection of ARB in BAL samples(%)

2 (17)

0.15

IQR, interquartile range; VAP, ventilator-associated pneumonia; SIRS, Systemic Inflammatory Response Syndrome; APACHE II, Acute Physiology and Chronic Health

Evaluation II; PCR, polymerase chain reaction; ARB, angiotensin receptor blockers; BAL, bronchoalveolar lavage samples; bold p values are those that are significant

(<0.05).

References

1. Wenzel RP. Hospital-acquired pneumonia: overview of the current

state of the art for prevention and control. Eur J Clin Microbiol Infect

Dis 1989;8:56–60.

2. Marrie TJ, Durant H, Yates L. Community-acquired pneumonia

requiring hospitalization: 5-year prospective study. Rev Infect Dis

1989;11:586–99.

3. Rutala WA, Weber DJ. Water as a reservoir of nosocomial pathogens.

Infect Control Hosp Epidemiol 1997;18:609–16.

4. Rowbotham TJ. Preliminary report on the pathogenicity of

Legionella pneumophila for freshwater and soil amoebae. J Clin

Pathol 1980;33:1179–83.

5. Michel R, Hauröder B. Isolation of an Acanthamoeba strain with

intracellular Burkholderia pickettii infection. Zentralbl Bakteriol

1997;285:541–57.

6. Barker J, Brown MRW. Trojan horses of the microbial world: proto-

zoa and the survival of bacterial pathogens in the environment.

Microbiology 1994;140:1253–9.

7. Steenbergen JN, Shuman HA, Casadevall A. Cryptococcus neofor-

mans interactions with amoebae suggest an explanation for its viru-

lence and intracellular pathogenic strategy in macrophages. Proc Natl

Acad Sci U S A 2001;98:15245–50.

8. Stout JE, Yu VL. Legionellosis. N Engl J Med 1997;337:682–7.

9. Marrie TJ, Raoult D, La Scola B, Birtles RJ, de Carolis E. Legionella-

like and other amoebal pathogens as agents of community-acquired

pneumonia. Emerg Infect Dis 2001;7:1026–9.

10. Greub G, Raoult D. Parachlamydiaceae potential emerging

pathogens. Emerg Infect Dis 2002;8:625–30.

11. La Scola B, Barrassi L, Raoult D. Isolation of new fastidious α

Proteobacteria and Afipia felis from hospital water supplies by direct

plating and amoebal co-culture procedures. FEMS Microbiol Ecol

2000;34:129–37.

12. La Scola B, Mallet MN, Grimont PAD, Raoult D. Description of

Afipia birgiae sp. nov., Afipia massiliae sp. nov. and recognition of

Afipia felis genospecies A. Int J Syst Evol Microbiol

2002;52:1773–82.

13. La Scola B, Mallet MN, Grimont PAD, Raoult D. Bosea eneae sp.

nov., Bosea massiliensis sp. nov. and Bosea vestrisii sp. nov., isolat-

ed from water supplies, and emendation of the genus Bosea (Das

1996). Int J Syst Evol Microbiol. In press 2003.

14. La Scola B, Mezi L, Auffray JP, Berland Y, Raoult D. Intensive care

unit patients are exposed to amoeba associated pathogens. Infect

Control Hosp Epidemiol 2002;23:462–5.

15. Raoult D, Aboudharam G, Crubezy E, Larrouy G, Ludes B, Drancourt

M. Molecular identification by “suicide PCR” of Yersinia pestis as

the agent of Medieval Black Death. Proc Natl Acad Sci U S A

2000;97:12800–3.

16. Raoult D, Fournier PE, Fenollar F, Jensenius M, Prioe T, de Pina JJ,

et al. Rickettsia africae, a tick-borne pathogen in travelers to sub-

Saharan Africa. N Engl J Med 2001;344:1504–10.

17. A’Court CH, Garrard CS, Crook D, Bowler L, Conlon C, Peto T, et

al. Microbiologic lung surveillance in mechanically ventilated

patients using non-directed bronchial lavage and quantitative culture.

Q J Med 1993;86:635–48.

18. Kollef MH, Bock KR, Richards RD, Hearns ML. The safety and

diagnostic accuracy of minibronchoalveolar lavage in patients with

suspected ventilator-associated pneumonia. Ann Intern Med

1995;122:743–8.

19. Sanchez-Nieto JM, Torres A, Garcia-Cordoba F, El Ebiary M, Carillo

A, Ruiz J, et al. Impact of invasive and non-invasive quantitative cul-

ture sampling on outcome of ventilator-associated pneumonia: a pilot

study. Am J Respir Crit Care Med 1998;157:371–6.

20. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II : a

severity of disease classification system. Crit Care Med

1985;13:818–29.

21. Reisner BS, Woods GL, Thomson RB Jr, Larone DH, Garcia LS, Shimizu

RY. Specimen processing. In: Murray PR, Baron EJ, Pfaller MA, Tenover

FC, Yolken RH, editors. Manual of clinical microbiology. 7th ed.

Washington: American Society for Microbiology; 1999. p. 64–104.

22. Kollef MH. Ventilator-associated pneumonia, a multivariate analysis.

JAMA 1993;270:1965–70.

23. Salata RA, Lederman MM, Shlaes DM, Jacobs MR, Eckstein E,

Tweardy D, et al. Am Rev Respir Dis 1987;135:426–32.

24. American Thoracic Society. Hospital-acquired pneumonia in adults:

diagnosis, assessment of severity, initial antimicrobial therapy, and

preventive strategies. Am J Respir Crit Care Med 1996;153:1711–25.

25. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et

al. Definitions for sepsis and organ failure and guidelines for the use

of innovative therapies in sepsis. Chest 1992;101:1644–55.

26. La Scola B, Mezi L, Weiller PJ, Raoult D. Isolation of Legionella

anisa using an amoebal coculture procedure. J Clin Microbiol

2001;39:365–6.

27. Drancourt M, Bollet C, Carlioz A, Martelin R, Gayral JP, Raoult D.

16S ribosomal DNA sequence analysis of a large collection of envi-

ronmental and clinical unidentifiable bacterial isolates. J Clin

Microbiol 2000;38:3623–30.

28. Ratcliff RM, Lanser JA, Manning PA, Heuzenroeder MW. Sequence-

based classification scheme for the genus Legionella targeting the

mip gene. J Clin Microbiol 1998;36:1560–7.

29. La Scola B, Rydkina L, Ndihokubwayo JB, Vene S, Raoult D.

Serological differentiation of murine typhus and epidemic typhus

using cross-adsorption and Western blotting. Clin Diagn Lab

Immunol 2000;7:612–6.

30. George DL, Falk PS, Wunderink RG, Leeper KVJ, Meduri GU,

Steere EL, et al. Epidemiology of ventilator-acquired pneumonia

based on protected bronchosopic sampling. Am J Respir Crit Care

Med 1998;158:1839–47.

31. Vincent JL, Bihari DJ, Suter PM, Bruining HA, White J, Nicolas-

Chanoin MH, et al. The prevalence of nosocomial infection in inten-

sive care units in Europe. JAMA 1995;274:639–44.

32. Gorman GW, Feeley JC, Steigerwalt A, Edelstein PH, Moss CW,

Brenner DJ. Legionella anisa: a new species of Legionella isolated

from potable waters and a cooling tower. Appl Environ Microbiol

1985;49:305–9.

33. Bornstein N, Vieilly C, Marmet D, Surgot M, Fleurette J. Isolation of

Legionella anisa from a hospital hot water system. Eur J Clin

Microbiol 1985;4:327–30.

34. Fenstersheib MD, Miller M, Diggins C, Liska S, Detwiler L, Werner

SB, et al. Outbreak of Pontiac fever due to Legionella anisa. Lancet

1990;336:35–7.

35. Fields BS, Barbaree JM, Sanden GN, Morrill WE. Virulence of

Legionella anisa strain associated with Pontiac fever: an evaluation

using protozoan, cell culture, and guinea pig models. Infect Immun

1990;58:3139–42.

36. Bornstein N, Mercatello A, Marmet D, Surgot M, Deveaux Y,

Fleurette J. Pleural infection caused by Legionella anisa. J Clin

Microbiol 1989;27:2100–1.

37. Thacker WL, Benson RF, Hawes L, Mayberry WR, Brenner DJ.

Characterization of a Legionella anisa strain isolated from a patient

with pneumonia. J Clin Microbiol 1990;28:122–3.

38. Fallon RJ, Stack BH. Legionnaires disease due to Legionella anisa. J

Infect 1990;20:227–9.

39. Vergis EN, Yu VL. Legionella species. In: Yu VL, Merigan TC Jr,

Barriere SL, editors. Antimicrobial therapy and vaccines. Baltimore

(MD): Williams & Wilkins; 1999. p. 257–72.

Address for correspondence: Didier Raoult, Unité des Rickettsies, CNRS

UMR 6020, Faculté de Médecine, 27 Boulevard Jean Moulin, 13385

Marseille Cedex 05, France; fax: 33(0)4.91.83.03.90; email:

Didier[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 821

RESEARCH

Municipal and agricultural pollution affects the Rio

Grande, a river that separates the United States from

Mexico. Three hundred and twenty-two Escherichia coli

isolates were examined for multiple antibiotic resistance

phenotypes and the prevalence of class 1 and class 2 inte-

gron sequences. Thirty-two (10%) of the isolates were

resistant to multiple antibiotics. Four (13%) of these iso-

lates contained class 1–specific integron sequences; one

isolate contained class 2 integron–specific sequences.

Sequencing showed that the class 1 integron–bearing

strain contained two distinct gene cassettes, sat-1 and

aadA. Although three of the four class 1 integron–bearing

strains harbored the aadA sequence, none of the strains

was phenotypically resistant to streptomycin. These results

suggest that integron-bearing E. coli strains can be present

in contaminated irrigation canals and that these isolates

may not express these resistance markers.

ntegron gene sequences contribute to the spread of

antimicrobial resistance alleles by lateral gene transfer

of gene cassettes in a variety of enteric bacteria, including

Campylobacter spp., Escherichia coli, and Salmonella

enterica serotype Typhimurium (1–4). The gastrointestinal

environment is suspected of serving as a reservoir for inte-

gron-bearing strains; when antimicrobial exposure occurs,

gene transfer events—which spread cassettes between

commensal organisms that are expelled into the environ-

ment (2)—would also occur.

The Rio Grande, the river separating the United States

from Mexico along the Texas-Mexico region, serves as a

source for irrigation water in Texas and Mexico. Previous

studies in our laboratory and others have shown that the

transboundary region is subject to extensive microbial and

chemical contamination. This contamination has been

associated with agricultural, municipal, and industrial

wastes originating from both sides of the border (5,6).

Leaking septic tanks and wastewater effluent discharges

result in fecal contamination levels as high as 2,000

CFU/mL of fecal coliforms (7,8).

Because of the strategic importance of the Rio Grande

for U.S. agriculture and the potential transmission of

antimicrobial resistance determinants by means of food

crops, we investigated the prevalence and characteristics

of class 1 and class 2 integron–bearing E. coli strains.

These strains were previously isolated from a study inves-

tigating fecal contaminants in irrigation water and associ-

ated sediments at specific locations along the river (9).

Methods

Three hundred and twenty-two E. coli isolates were

previously isolated from irrigation water and associated

sediments at the El Paso, Presidio, and Weslaco regions of

the river (9). After being confirmed as E. coli by MUG (4-

methyl umbelliferyl-β-D-glucoronide)–based fluores-

cence, these isolates were screened for antimicrobial sus-

ceptibility by using the agar dilution method (10,11). The

isolates were tested against ampicillin, tetracycline, ceftri-

axone, cephalothin, gentamicin, kanamycin, streptomycin,

chloramphenicol, ciprofloxacin, and trimethoprim/sul-

famethoxazole. The antibiotics were tested at concentra-

tions established by the National Antimicrobial Resistance

System (12).

Isolates that were multidrug resistant (resistant to two

or more antimicrobial agents) were grown overnight in 5

mL of Mueller-Hinton broth (Accumedia, Baltimore, MD)

with the appropriate concentration of antimicrobial com-

pound. A 1-mL aliquot of the culture was centrifuged at

10,000 rpm for 2 min. The cell pellet was resuspended in

500 µL of sterile water and boiled for 10 min. The result-

ing DNA suspension was used as template DNA in poly-

merase chain reaction (PCR) amplification for the class 1

and class 2 integrase gene and variable regions using the

primer sequences shown in the Table (13–15).

The PCR reactions used 10 µL of template DNA, 5 µM

of primers, 25 mM MgCl, 10 mM deoxynucleotide

triphosphate, and 23 ng bovine serum albumin. Nuclease-

free water (Ambion, Austin, TX) was added to achieve a

volume of 50 µL. A “hot start” method was used, and 1.25

U of Taq DNA polymerase (Sigma, St. Louis, MO) was

822 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Antimicrobial Resistance Markers

of Class 1 and Class 2 Integron-

bearing Escherichia coli from

Irrigation Water and Sediments

Matthew T. Roe,* Everardo Vega,* and Suresh D. Pillai*

*Texas A&M University, College Station, Texas, USA

added after initial template denaturation. The PCR cycle

was as follows: initial denaturation for 12 min at 94°C, hot

start pause at 80°C followed by 35 cycles of denaturation

at 94°C for 1 min, primer annealing at 60°C for 1 min, and

extension at 72°C for 5 min at first cycle. An additional 5

s was progressively added to each cycle to reach a final of

7 min, 55 s. PCR products were analyzed on 1% agarose

gel.

Amplification products were extracted from the gels

with the QIAGEN QIAquick gel extraction kit (Valencia,

CA). The amplified products were sequenced at a commer-

cial facility (MWG Biotech Inc., High Point, NC) with the

class 1 and class 2 integron variable region primers (integ

and hep) (Table). Contiguous sequences were created from

single sequence reads by using the CAP3 sequence assem-

bly program (16). Contiguous sequences were analyzed by

using the GenBank database of the National Center for

Biotechnology Information and the BLASTX search

engine (17). Putative gene relationships and sequence data

were analyzed by using a multiple sequence alignment cre-

ated by using Clustal W version 1.82 (18).

Results

Of the 322 E. coli isolates from sediment and irrigation

water samples analyzed for antimicrobial resistance, 104

(32%) isolates showed resistance to at least one of the

antimicrobial compounds (Figure 1). Approximately 10%

(32/322) of all the isolates showed a multidrug resistance

phenotype. Eighteen percent of the isolates were resistant

to cephalothin; however, only 5 (2%) of 322 were resistant

to ceftriaxone, which also belongs to the cephalosporin

family. Resistance to ampicillin was prevalent in approxi-

mately 35 (11%) of the isolates. Resistance to tetracycline

(9%), kanamycin (2%), gentamicin (0.3%), and strepto-

mycin (4%) was also observed. Resistance to the fluoro-

quinolone ciprofloxacin was seen in one isolate. Three

(<1%) of the 322 isolates were resistant to sulfonamide

sulfamethoxazole. On the basis of analysis of variance,

antimicrobial resistance and the sampling location were

correlated. Isolates from the El Paso sampling region had

significantly higher (p<0.05) antimicrobial resistance as

compared with the Presidio and Weslaco sampling regions

(data not shown).

The 32 isolates identified as multiple antimicrobial

resistant were assayed by PCR amplification for class 1

and class 2 integrase genes. Four isolates (approximately

13%) had the class 1 integrase gene intI1 (Figure 2A), and

one isolate had the class 2 integrase gene intI2 (Figure 2B).

Isolates identified as having the class 1 or class 2 integrase

genes were further characterized through PCR amplifica-

tion of the class 1 and class 2 variable regions. Of the four

amplified class 1 integron variable regions, three isolates

(isolate 16, isolate 19, and isolate 21) were approximately

1 kb in size, but the fourth isolate (isolate I-6) harbored a

2-kb fragment (Figure 3A). The 1-kb amplification prod-

ucts were observed in isolates from the El Paso area.

Nucleotide sequencing showed that all of the 1-kb

sequences contained a conserved configuration of a 780-

bp gene cassette identified as the aadA gene (Figure 4).

The 2-kb amplification product was seen in an isolate from

the Presidio sampling region. Nucleotide sequencing

showed that the variable region contained a 498-bp gene

cassette, identified as the dhfrXII gene, which encodes

trimethoprim resistance. The gene cassette did not exhibit

perfect homology with the dhfrXII gene (Figure 4). Within

the identified 498-bp gene cassette, a 323-bp stretch

showed 97% sequence homology; in addition, 59-bp and

56-bp fragments showed 88% and 89% homology, respec-

tively. “Islands” of sequence within the variable region

showed no sequence homology to any known genes.

When the 32 multiply antimicrobial-resistant isolates

were screened for class 2 integrons, only 1 isolate was pos-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 823

RESEARCH

Table. Oligonucleotide primer sequences used for amplification of class 1 and class 2 integrase and variable regions

Primer

Primer sequence

Target

Reference

integ-1

5´-GGCATCCAAGCAAG-3´

5′-Class 1 integron variable region

Levesque et al. 1995 (13)

integ-2

5´-AAGCAGACTTGACCTGA-3´

3´-Class 1 integron variable region

Levesque et al. 1995 (13)

hep51

5´-GATGCCATCGCAAGTACGAG-3´

5´-Class 2 integron variable region

White et al. 2001 (14)

hep74

5´-CGGGATCCCGGACGGCATGCACGATTTGTA-3′

3´-Class 2 integron variable region

White et al. 2001 (14)

intI1.F

5´-GGGTCAAGGATCTGGATTTCG-3´

5´-intI1 gene

Mazel et al. 2000 (15)

intI1.R

5´-ACATGGGTGTAAATCATCGTC-3´

3´-intI1 gene

Mazel et.al. 2000 (15)

intI2.F

5´-CACGGATATGCGACAAAAAGGT-3´

5´-intI2 gene

Mazel et al. 2000 (15)

intI2.R

5´-GTAGCAAACGAGTGACGAAATG-3´

5´-intI2 gene

Mazel et al. 2000 (15)

Figure 1. Frequency of antimicrobial resistance observed in 322

Escherichia coli isolates from irrigation water along the Texas-

Mexico border.

itive (Figure 2B). This particular isolate (isolate 29) was

obtained in the Presidio region and had a 2,600-bp variable

region (Figure 3B). Nucleotide sequencing identified two

distinct gene cassettes, namely, the sat-1 and aadA genes,

which code for streptothricin acetyl transferase, and amino-

glycosede adenyltransferase, respectively (Figure 5).

Discussion

Antimicrobial resistance in human pathogens has

become a major public health issue. Resistant organisms

have been isolated from a number of natural and man-

made environments (6,19,20). In natural environments,

resistant organisms can be indigenous or introduced

through natural or anthropogenic causes (21,22). Integron

gene sequences have been identified as a primary source of

resistance genes and are suspected to serve as reservoirs of

antimicrobial resistance genes within microbial popula-

tions (1,2,23,24). Previous studies along the Texas-Mexico

border have shown that fecal contamination of the Rio

Grande does occur (7,25). The isolation of 322 E. coli iso-

lates from irrigation water and associated sediments fur-

ther confirms that fecal wastes are affecting this body of

water. Previous studies have reported that municipal and

animal wastes regularly harbor multidrug-resistant E. coli

strains (6,26,27). In this study, 18% of the isolates were

resistant to cephalothin. These results are similar to those

from a recent survey of U.S. rivers, which found cefo-

taxime (a third-generation, cephalosporin-resistant, gram-

negative bacterium) to range from 16% to 96% across 22

rivers (19). The higher frequency of isolation of resistant

strains from the El Paso region compared with the other,

less urbanized sampling locations is not surprising since

the effluent from a number of wastewater treatment plants

enters the river at that region (W. McElroy, unpub. data;

28). Previous studies with sludge and septic tank wastes

showed relatively high levels of antimicrobial resistance in

E. coli (6). The precise sources of the E. coli isolates used

in this study could not be identified because of technical

limitations in source tracking (29).

Class 1 and class 2 integron gene sequences were found

within these E. coli isolates. Together, they accounted for

5 (16%) of 32 multidrug-resistant isolates characterized in

this study. This prevalence was higher than that reported

by Rosser et al. (30), who showed that 3.6% of gram-neg-

ative bacteria in an estuarine environment contained the

class 1 integron. Three of the four class 1 integron-bearing

E. coli in this study contained the nucleotide sequence of

the spectinomycin-streptomycin resistance gene aadA1

(31). Resistance to streptomycin was not observed in these

isolates, but resistance to the closely related kanamycin

was seen. These results are similar to those reported by

Zhao et al. (3), who identified that the aadA gene trans-

ferred to a strain of Hafnia alvei but did not report resist-

ance to streptomycin or spectinomycin. These researchers

attributed their findings to the inefficient expression of the

inserted gene cassette by the integron promoter. Previous

studies have also shown that the antimicrobial resistance

phenotype can be modulated once these strains are

exposed to specific environmental conditions (32).

The aadA gene cassette is not novel in class 1 inte-

grons. Earlier work by Zhao et al. (3) and Bass et al. (24)

has shown that the aadA gene is highly conserved among

Shiga toxin–producing and avian clinical E. coli isolates,

respectively. The only class 2 integron-bearing strain iso-

lated in this study also contained the aadA gene in addition

to the sat-1 gene, which codes for resistance to kanamycin,

a finding in agreement with the phenotypic expression.

824 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure 2. Agarose gel electrophoresis of integrase gene poly-

merase chain reaction (PCR) amplification products. A: PCR prod-

ucts of class 1 integrase gene intI1. Lane M; molecular marker;

lanes 1 and 19: no template (negative) control; lanes 2 and 20:

positive control (In2); lanes 3–36: multiple drug–resistant isolates.

B: PCR products of class 2 integrase gene intI2. Lane M: molecu-

lar marker; lanes 1 and 19: no template control (negative) control;

lanes 2 and 20: positive control (Tn7); lanes 3–36: multidrug-

resistant isolates.

Figure 3. A: Polymerase chain reaction (PCR) amplification prod-

ucts with primers targeted against the class 1–specific conserved

sequences. Lane 1: no template control; lane 2: positive control

(In2); lane 3: Escherichia coli isolate 16; lane 4: E . coli isolate 19;

lane 5: E. coli isolate 21; lane 6: blank; lane 7: E. coli isolate I-6.

B: PCR amplification products with primers targeted against the

class 2–specific conserved sequences. Lane 1: no template con-

trol; lane 2: positive control (In2); lane 3: E. coli isolate 29.

The sat-1 gene, which codes for the streptothricin acetyl

transferase, was not detected in any other E. coli isolate.

The presence of the sat-1 gene cassette, in combination

with the aadA gene, suggests that this class 2 integron is

likely a derivative of the class 2 integron found on transpo-

son Tn7 (33,34).

The aadA gene was conserved among the class 1 and

class 2 integrons, which suggests a possible selective

mechanism for this cassette in enteric bacteria from natu-

ral waters. The 2-kb integron-specific variable region–con-

taining strain, which was isolated from the Presidio area,

harbored the dihydrofolate reductase gene (dhfrXII)

instead of the aadA gene (35).

Overall, these results suggest that the irrigation canals

and sediments associated with the Rio Grande are contam-

inated by bacteria of fecal origin that contain antimicrobial

resistance genes. Of 322 E. coli isolates, 32 (approximate-

ly 10%) were resistant to multiple antimicrobial drugs.

Five of these 32 E. coli isolates harbored class 1 and class

2 integron sequences. This study did not investigate the

possibility that other integron-bearing nonfecal bacteria

were present. The occurrence of integron-bearing E. coli in

irrigation water is important since these organisms are

known fecal contaminants, and the potential for lateral

gene transfer exists. The results also indicate that integron-

bearing strains may not always express the antimicrobial

phenotype; thus, phenotype-based isolation of resistant

organisms can underestimate the levels of resistant organ-

isms. Studies are needed to identify whether integron-

mediated antimicrobial resistance transfer does indeed

occur within the irrigation canal sediments and on veg-

etable surfaces, when they are irrigated with contaminated

irrigation water.

Acknowledgments

We thank Anne Summers and Cynthia Liebert for providing

the integron-positive strains and James Zhu for helpful discus-

sions concerning the sequencing data.

This work was supported by funds from the State of Texas

ATP project 00517-0361-1999, the U.S. Department of Agriculture

(USDA)/CSREES-IFAFS grant 00-52102-9637, the USDA

CSREES grant 2001-34461-10405 and Hatch grant H8708.

Mr. Roe conducted this study while an M.S. student in the

Food Safety and Environmental Microbiology Laboratory in the

Poultry Science Department at Texas A&M University. His

research interests are in environmental microbiology and food

safety.

References

1. Ochman H, Lawrence JG, Groisman EA. Lateral gene transfer and the

nature of bacterial innovation. Nature 2000;405:299–304.

2. Lucey B, Crowley D, Moloney P, Cryan B, Daly M, O’Halloran F, et

al. Integronlike structures in Campylobacter spp. of human and ani-

mal origin. Emerg Infect Dis 2000;6:50–5.

3. Zhao S, White DG, Ge B, Ayers S, Friedman S, English L, et al.

Identification and characterization of integron-mediated antibiotic

resistance among Shiga toxin–producing Escherichia coli isolates.

Appl Environ Microbiol 2001;67:1558–64.

4. Briggs CE, Fratamico PM. Molecular characterization of an antibiot-

ic resistance gene cluster of Salmonella typhimurium DT104.

Antimicrob Agents Chemother 1999;43:846–9.

5. Mroz RC Jr, Pillai SD. Bacterial populations in the groundwater on

the US-Mexico border in El Paso County, Texas. South Med J

1994;87:214–7.

6. Pillai SD, Widmer KW, Maciorowski KG, Ricke SC. Antibiotic

resistance profiles of Escherichia coli isolated from rural and urban

environments. J Environ Sci Health 1997;32:1665–75.

7. International Boundary and Water Commission. Water Bulletin #63:

Flow of the Rio Grande and related data. El Paso (TX): The

Commission; 1993. p. 132.

8. Scandura JE, Sobsey MD. Viral and bacterial contamination of

groundwater from on-site sewage treatment systems. Water Sci

Technol 1997;35:141–6.

9. Rayburn EL, Sternes KL, Weidenfeld RP, Pillai SD. Microbial

pathogens in irrigation canals and associated sediments. In:

Proceedings of the 101st General Meeting of the American Society

for Microbiology. 2001. Orlando (FL): American Society for

Microbiology; 2001.

10. Food and Drug Administration. Bacteriological analytical manual

online. 8th ed. 2001. Available from: URL: http://www.cfsan.fda.gov/

~ebam/bam-toc.html

11. National Committee for Clinical Laboratory Studies. Methods for

dilution antimicrobial susceptibility tests for bacteria that grow aero-

bically. Vol. 10. Villanova (PA): The Committee; 1999. p. 33.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 825

RESEARCH

Figure 4. Schematic representations of the four class 1 integrons

and one class 2 integron sequenced from multiple

antibiotic–resistant Escherichia coli isolates.

12. Centers for Disease Control and Prevention/Food and Drug

Administration; U.S. Department of Agriculture. National Antibiotic

Resistance Monitoring System. Atlanta: The Centers; 1999.

13. Levesque C, Piche L, Larose C, Roy PH. PCR mapping of integrons

reveals several novel combinations of resistance genes. Antimicrob

Agents Chemother 1995;39:185–91.

14. White PA, McIver CJ, Rawlinson WD. Integrons and gene cassettes

in the enterobacteriaceae. Antimicrob Agents Chemother

2001;45:2658–61.

15. Mazel D, Dychinco B, Webb VA, Davies J. Antibiotic resistance in

the ECOR collection: integrons and identification of a novel aad

gene. Antimicrob Agents Chemother 2000;44:1568–74.

16. Huang X, Madan A. CAP3: a DNA sequence assembly program.

Genome Res 1999;9:868–77.

17. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local

alignment search tool. J Mol Biol 1990;215:403–10.

18. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the

sensitivity of progressive multiple sequence alignment through

sequence weighting, position-specific gap penalties and weight

matrix choice. Nucleic Acids Res 1994;22:4673–80.

19. Ash RJ. Antibiotic resistance of gram-negative bacteria in rivers,

United States. Emerg Infect Dis 2002;8:713–6.

20. Goni-Urriza M, Capdepuy M, Arpin C, Raymond N, Caumette P,

Quentin C, et al. Impact of an urban effluent on antibiotic resistance

of riverine Enterobacteriaceae and Aeromonas spp. Appl Environ

Microbiol 2000;66:125–32.

21. American Academy of Microbiology. Antimicrobial resistance: an

ecological perspective. Washington: American Society for

Microbiology; 1999. p. 1–14.

22. Wegener HC, Aarestrup FM, Jensen LB, Hammerum AM, Bager F.

Use of antimicrobial growth promoters in food animals and

Enterococcus faecium resistance to therapeutic antimicrobial drugs in

Europe. Emerg Infect Dis 1999;5:329–35.

23. Hall RM, Collis CM. Antibiotic resistance in gram-negative bacteria:

the role of gene cassettes and integrons. Drug Resist Updat

1998;1:109–19.

24. Bass L, Liebert CA, Lee MD, Summers AO, White DG, Thayer SG,

et al. Incidence and characterization of integrons, genetic elements

mediating multiple-drug resistance, in avian Escherichia coli.

Antimicrob Agents Chemother 1999;43:2925–9.

25. Cech I, Essman A. Water sanitation practices on the Texas-Mexico

border: implications for physicians on both sides. South Med J

1992;85:1053–64.

26. McEwen SA, Fedorka-Cray PJ. Antimicrobial use and resistance in

animals. Clin Infect Dis 2002;34(Suppl 3):S93–106.

27. Schwarz S, Chaslus-Dancla E. Use of antimicrobials in veterinary

medicine and mechanisms of resistance. Vet Res 2001;32:201–25.

28. Texas Water Development Board, New Mexico Water Resources

Research Institute. Transboundary aquifers of the El Paso/Ciudad

Juarez/Las Cruces region. El Paso (TX): The Institute; 1997. p. 152.

29. Simpson JM, Santo Domingo JW, Reasoner DJ. Microbial source

tracking: state of the science. Environ Sci Technol 2002;36:5279–88.

30. Rosser SJ, Young HK. Identification and characterization of class 1

integrons in bacteria from an aquatic environment. J Antimicrob

Chemother 1999;44:11–8.

31. Sundstrom L, Radstrom P, Swedberg G, Skold O. Site-specific recom-

bination promotes linkage between trimethoprim- and sulfonamide-

resistance genes. Sequence characterization of dhfrV and sulI and a

recombination active locus of Tn21. Mol Gen Genet

1988;213:191–201.

32. Pillai SD, Pepper IL. Transposon Tn5 as an identifiable marker in

Rhizobia: survival and genetic stability of Tn5 mutant bean Rhizobia

under temperature stressed conditions in desert soils. Microb Ecol

1990;21:21–33.

33. Sundstrom L, Roy PH, Skold O. Site-specific insertion of three struc-

tural gene cassettes in transposon Tn7. J Bacteriol 1991;173:3025–8.

34. Hansson K, Sundstrom L, Pelletier A, Roy PH. IntI2 integron inte-

grase in Tn7. J Bacteriol 2002;184:1712–21.

35. Huovinen P, Sundstrom L, Swedberg G, Skold O. Trimethoprim and

sulfonamide resistance. Antimicrob Agents Chemother

1995;39:279–89.

Address for correspondence: Suresh D. Pillai, Department of Poultry

Science, 418D Kleberg Center, MS 2472 TAMUS, Texas A&M

University, College Station, Texas 77843-2472, USA; fax: 979-845-1921;

email: spillai@poultry.tamu.edu

826 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

OPPORTUNITIES FOR PEER REVIEWERS

The editors of Emerging Infectious Diseases seek to increase

the roster of reviewers for manuscripts submitted by authors all

over the world for publication in the journal. If you are interested

in reviewing articles on emerging infectious disease topics, please

e-mail your name, address, curriculum vitae, and areas of expert-

ise to [email protected]

At Emerging Infectious Diseases, we always request reviewers’

consent before sending manuscripts, limit review requests to three

or four per year, and allow 2-4 weeks for completion of reviews.

We consider reviewers invaluable in the process of selecting and

publishing high-quality scientific articles and acknowledge their

contributions in the journal once a year.

Even though it brings no financial compensation, participation

in the peer-review process is not without rewards. Manuscript

review provides scientists at all stages of their career opportunities

for professional growth by familiarizing them with research trends

and the latest work in the field of infectious diseases and by

improving their own skills for presenting scientific information

through constructive criticism of those of their peers. To view the

spectrum of articles we publish, information for authors, and our

extensive style guide, visit the journal web site at

www.cdc.gov/eid.

For more information on participating in the peer-review

process of Emerging Infectious Diseases, e-mail

[email protected] or call the journal office at 404-371-5329.

An epidemiologic and seroprevalence survey was con-

ducted (n=830) to assess the proportion of persons

exposed to hantavirus in IX Region Chile, which accounts

for 25% of reported cases of hantavirus cardiopulmonary

syndrome. This region has three geographic areas with dif-

ferent disease incidences and a high proportion of aborigi-

nals. Serum samples were tested for immunoglobulin (Ig) G

antibodies by enzyme-linked immunosorbent assay against

Sin Nombre virus N antigen by strip immunoblot assay

against Sin Nombre, Puumala, Río Mamoré, and Seoul N

antigens. Samples from six patients were positive for IgG

antibodies reactive with Andes virus; all patients lived in the

Andes Mountains. Foresting was also associated with

seropositivity; but not sex, age, race, rodent exposure, or

farming activities. Exposure to hantavirus varies in different

communities of IX Region. Absence of history of pneumo-

nia or hospital admission in persons with specific IgG anti-

bodies suggests that infection is clinically inapparent.

antaviruses, RNA-containing viruses, compose a

genus within the family Bunyaviridae. The natural

reservoirs of the pathogenic New World hantaviruses are

rodents of the family Muridae, subfamily Sigmodontinae,

in which a chronic and asymptomatic infection develops

(1). Hantavirus is a zoonosis transmitted from rodents to

humans by inhaling contaminated aerosols from feces,

urine, and saliva of infected mice (2).

Human infection with hantaviruses have been associat-

ed with two diseases. One is the hemorrhagic fever with a

renal syndrome (HFRS) caused by Hantaan, Puumala,

Seoul, and Dobrava/Belgrade viruses, first recognized dur-

ing the Korean War in 1950 to 1953. HFRS occurs mainly

in Asia and Europe; death rates range from 0.1% to 15%

(3,4). The other disease, a severe respiratory illness known

as hantavirus cardiopulmonary syndrome (HCPS), occurs

in the Americas and has a death rate of 40% (1,5–7). We

prefer the term hantavirus cardiopulmonary syndrome to

an alternative term, hantavirus pulmonary syndrome,

because most deaths associated with HCPS are related to

cardiac failure rather than pulmonary failure, and this

aspect of the syndrome remains underappreciated by prac-

titioners and others.

HCPS has been identified in several countries in North

and South America and is caused by different hantavirus-

es: Sin Nombre in North America, Juquitiba virus in Brazil

(8), Laguna Negra virus in Paraguay and Bolivia (9,10),

Andes in Argentina and Chile (11,12), Choclo virus in

Panama (13), and several subspecies or viral genotypes of

Andes virus in Argentina (e.g., Oran, Lechiguanas, et al.)

(14–16).

In 1995, Andes virus was first identified in the

Argentinean Patagonia and was recognized in central and

southern regions of Chile. The reservoir is the long-tailed

pygmy rice rat (Oligoryzomys longicaudatus), a species

that occurs primarily in temperate forest (17).

Since HCPS emerged in Chile, 287 cases have been

confirmed as of March 2003, causing a substantial impact

on the public health system; the death rate for HCPS in

Chile has exceeded 40%. In addition to the 287 cases of

HCPS, 17 cases of mild infection with no cardiopulmonary

involvement have been reported, demonstrating that han-

tavirus infection may have variability in its expression.

Serologic studies have established both clinically asymp-

tomatic infections and symptomatic infections not recog-

nized at the time as hantavirus infection (18–23).

The phenomenon of clinically nonapparent infections

varies in different areas and populations of the Americas.

In the United States, the proportion of infection versus dis-

ease is thought to be close to 1% (i.e., disease develops in

most of the infected patients, and human population sero-

prevalence varies between 0.2% to 1.7%) (4,24,25).

Seroprevalence studies in South America, however, have

shown that some populations have had much more fre-

quent exposure to hantaviruses in the absence of known

clinical manifestations, as seen in some populations native

to Paraguay and Northern Argentina (40% and 17%,

respectively) (26), the general population in Aysen,

Southern Chile (2.0% and 13.1% in urban and rural areas,

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 827

RESEARCH

Hantavirus Prevalence in the

IX Region of Chile

Marlis Täger Frey,*† Pablo C. Vial,*‡ Constanza H. Castillo,§ Paula M. Godoy,*

Brian Hjelle,¶ and Marcela G. Ferrés*

*Universidad Católica de Chile, Santiago, Chile; †Universidad

Austral de Chile, Valdivia, Chile; ‡Universidad del Desarrollo,

Santiago, Chile; §Universidad de La Frontera, Temuco, Chile; and

¶University of New Mexico School of Medicine, Albuquerque, New

Mexico, USA

respectively) (27), São Paulo and Bahia, Brazil (1.23 %

and 13.1%, respectively) (28,29), and recently in Panama

(30%) (13). The populations with the highest seropreva-

lence were indigenous persons rather than those of

European ancestry.

Two hypotheses have been proposed to explain these

phenomena. The first one presumes that strains in South

America are of lesser pathogenicity, though in many cases

nonpathogenic viruses have yet to be detected in local

rodent populations. The second hypothesis involves the

existence of at least two variables: the nature of the expo-

sure and genetic constitution of the host population. The

high seroprevalence seen in Paraguay (26) could be caused

by a higher resistance in the aboriginal population and

greater exposure to the virus in some regions. The preva-

lence and high case-fatality ratios seen in North America

might be the result of a lower exposure, a lesser genetic

resistance to disease, or both.

More than 20 strains of hantaviruses are known world-

wide; not all of them are associated with human disease.

The observations discussed in this article suggest the need

to study whether hantaviruses of lesser pathogenicity and

other clinical entities until now unrecognized may occur in

areas where the seroprevalence is high and disease inci-

dence is low (11,14–16,18). For example, hantaviruses

previously thought to be endemic only in Europe and Asia

have also been recognized in the North American continent

(30–32).

To study the seroprevalence of hantavirus in Chile, we

studied the general population in the IX Region for the rea-

sons that follow. This area is ranked third among those

affected by HCPS, accounting for 25% of the known cases

in Chile. Three different areas can be distinguished on the

basis of geographic features in relation to the two moun-

tain ranges that traverse the region (the Pacific Coast and

the coastal range [Coastal], the Central Valley [Central],

and the pre-Andean Region [Andean]). The Mapuche, the

main native ethnic population of Chile, account for 26.3%

of the total regional population. Members of the tribe can

readily be identified from surnames, which are derived

from both parents. Therefore, we were able to study the

seroprevalence of hantavirus in the same region but in

areas having different incidence rates of the disease and

evaluate the asymptomatic infection in the Mapuche abo-

riginal population.

Methods

Study Design and Area

We conducted a cross-sectional epidemiologic and

serologic survey to determine the prevalence of IgG anti-

bodies against hantavirus N antigen in the adults living in

nine communities of the IX Region of Chile. The IX

Region of Chile is located in the southern part of the coun-

try, between parallel 37° and 40° South and the meridians

70° and 74° East (Figure). The region has a surface area of

31,842.3 km

and 781,242 inhabitants in 31 communities;

of the total population, 38.71% is rural. The characteristics

of these populations can be distinguished according to geo-

graphic features. The Mountainside Andes range (Andean)

is rural area with small towns; the population there is pri-

marily sustained with farming, woodcutting, and tourist

services. The population composes 13.8% of the region’s

total population, yet accounts for 82% of HCPS cases. The

Central Valley intermediate depression (Central) is a pre-

dominantly urban area in which the main sources of

income are industry, farming, and ranching. Temuco, the

capital and main urban center, is located in this area. The

Central Valley contains 67.3% of the total population and

accounts for 5.9% of HCPS cases. Coastal inhabitants

account for 19% of the region’s population; the communi-

ty sustains itself with farming and forestry activities, fish-

ing, and tourism. Twelve percent of HCPS cases originate

in this area.

Sample Design

The sample included 847 persons, enough to give us

accurate point seroprevalence given estimated seropreva-

lence rates of 7%, 3%, and 5% for the Andean, Central,

and Coastal areas, respectively, as derived from previous

studies in the country (27) and the distribution of HCPS

cases in the region. The sample was designed to be repre-

sentative of each community according to information

available from the National Statistics Agency about popu-

lation sex, ethnicity, age distribution, and rural-urban pro-

portion. Persons were contacted in their homes with a pre-

determined plan that included the selection of the housing

blocks to be studied, the home at which to start enrollment,

and how to proceed thereafter. Only one person was

828 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure. Geographic IX Region, Chile, and hantavirus seropreva-

lence in the tested communities of the IX Region. HCPS,

Hantavirus cardiopulmonary syndrome.

HCPS

enrolled in each household contacted. That person was

informed of the study, signed an informed consent, and

agreed to answer a questionnaire and donate a sample of

blood. Only persons >

15 years of age who were free of any

current febrile respiratory illness were asked to enroll. The

epidemiologic questionnaire, previously validated in the

community, requested information about sex, age, race,

residency, labor activity, rodent exposure at home or work,

cardiorespiratory disease history, hospital admissions, and

direct contact with HCPS patients.

Nine communities were included, three per geographic

area. Curacautín, Melipeuco, and Pucón (Figure) were

chosen as representative of the Andean area (n=277);

Victoria, Temuco, and Loncoche from the Central area

(n=279); and Purén, Carahue, and Toltén from the Coastal

area (n=291). The Andean and Coastal areas were consid-

ered rural; the Central area was considered to be a compos-

ite of rural and urban. Persons were classified as Mapuches

when they had at least one surname of Mapuche deriva-

tion. The study was approved by the Catholic University

Ethics Committee.

Collection, Processing, and Analysis of Samples

Blood samples were collected, transported, and cen-

trifuged the same day. Serum samples were kept frozen at

–20°C and sent for analysis to the Laboratory of Virology

at the Pontificia Universidad Católica de Chile, Santiago,

Chile. An enzyme-linked immunosorbent assay (ELISA)

was performed to detect hantavirus-specific immunoglob-

ulin (Ig) G antibodies in all samples.

ELISA was performed as described (33,34) by using

recombinant antigen for the Sin Nombre strain (produced

in Escherichia coli and provided by Centers for Disease

Control and Prevention [CDC]), which crossreacts with all

American hantaviruses. Briefly, serial dilutions (1:100,

1:400, 1:1,600, and 1:6,400) of patient serum samples

were incubated for 1 h at 37°C in antigen-coated, 96-well

plates. Peroxidase-coupled anti-human IgG was used as

secondary antibody and incubated 1 h at 37°C. After sub-

strate reaction, plates were read at 414 nm.

The net absorbance values are results of the hemiplates

absorbance subtraction with and without antigen. To be

considered positive, the net absorbance values for a sam-

ple had to be >0.2 in the 1:100 and 1:400 dilutions, and the

sum of all net absorbance values had to be >0.95. Positive

serum samples were tested twice and verified to be posi-

tive at CDC.

A confirmatory strip immunoblot assay (SIA) was per-

formed an all positive samples (B.H.). Four recombinant N

antigens from Sin Nombre, Puumala, Rio Mamoré, and

Seoul hantaviruses were fixed onto a nitrocellulose mem-

brane by vacuum. Serum samples (1:200) were incubated

with the strips and alkaline phosphatase-coupled anti-

human IgG added. Reactivity was estimated visually, and

each band given an intensity value on a four-point scale, as

previously described (35,36). We regarded a sample as

positive for antibodies when reactive to both Sin Nombre

and Rio Mamoré on the basis of the presumption that anti-

bodies to Andes virus will always react to closely related

Rio Mamoré and generally will crossreact with the other

sigmodontine rodent-borne virus, Sin Nombre. A sample

was regarded as confirmed seropositive when the serum

was reactive in the ELISA and confirmed by the SIA.

Fisher exact test was used for the analysis of independent

variables. A p value <0.05 was considered significant.

Results

Sample Population Characteristics

A total of 829 persons were included in the study, with

271, 272, and 286 from the Andean, Central, and Coastal,

respectively. The age range was 15–88 years of age (mean

39.4 years of age); 47.6% were men, and 18.3% were of

Mapuche origin. In all, 73.3% of the sampled population

was considered rural. Men worked in farming, forestry, or

both in 61.8% of the cases; 62.2% of the women were

housewives. A history of exposure to rodents, their excre-

ta, or both at home was reported in 88.0% of enrolled per-

sons; 54.6% reported exposure at work.

Serology

Six (0.72%) samples were hantavirus antibody-positive

by both ELISA and SIA. The samples came from persons

who lived in the Andean area, giving a seroprevalence of

2.15% (6/271) for this area, significantly higher than the

other two regions studied (p=0.0001). The seropositive

cases belonged to each of the three counties studied; how-

ever, relative frequencies varied: Curacautín, 2.6%

(2/132); Melipeuco, 6.1% (2/33); and Pucón, 2.0%

(2/100). All three counties have previously had HCPS

cases reported (Figure). As shown in Table 1, four case-

patients were male, and two were Mapuche. The case-

patients ranged from 28 to 76 years of age, with a mean of

52.8 years of age. Seropositivity could not be associated

with sex (p=0.43), race (p=0.30), or age (p=0.18). Five of

the case-patients worked in farming or forestry, although

only forestry had a significant association with seroposi-

tivity (p=0.018), meaning the risk for infection was 10

times higher (relative risk 9.72; confidence interval 1.15 to

82.44). Five of the case-patients said they had been

exposed to rodents or their excreta either at home or work.

This exposure, however, did not reach statistical signifi-

cance (p=0.1 and p=1.0, respectively). The sixth case-

patient was not exposed to rodents; however, he had been

working in a large shed, the woods, and a sawmill; he also

had been weeding. Finally, none of the six antibody-posi-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 829

RESEARCH

tive persons had previous contact with HCPS patients, his-

tory of pneumonia, or hospital admission.

Discussion

This study shows serologic evidence of past infection

by hantavirus in the Chilean population. The hantavirus

associated with symptomatic infections in Chile is Andes

virus, identified in patient blood samples by reverse tran-

scription-polymerase chain reaction (RT-PCR) and recent-

ly isolated from human blood (12,37).

Our results show the global seroprevalence to han-

tavirus in the adult general population in the IX Region to

be 0.72% (ranging from 0 to 6%, depending on the com-

munity). Seroprevalence is significantly higher in the

Andes rural area, consistent with the observed elevated

incidence of HCPS disease in this area of Region IX

between 1997 and 2000 (Table 2). This geographic distri-

bution is probably related to the great tracts of native for-

est, where Chusquea quila, a bamboo-like shrub that pro-

tects and feeds the carrier rodent, is abundant. Moreover,

the increasing development of the forestry industry plus an

increase in the rodent population caused by the favorable

climatic condition because of the El Niño effect have

caused humans and mice to interact closely.

We did not find an association between seroprevalence

and reported exposure to rodents, probably because this

type of exposure is frequent in all groups in the region,

occurring both at work and at home. The exposure to

rodents or their excreta is necessary but apparently not suf-

ficient to acquire the infection. Presumably an exposure to

a specific reservoir mouse or virus carrier, human behav-

ior, and biologic factors are important.

Forestry is associated with a higher risk for infection.

This labor, whether in the woods or at sawmills, is per-

formed almost exclusively by men; both men and women

share farm work in most instances. This finding may

explain why a higher proportion of HCPS develops in men

(75% of reported cases). However, any apparent associa-

tion between hantavirus exposure and a particular occupa-

tion does not necessarily implicate these occupations as a

risk factor. In fact, occupation may be a marker for living,

sleeping, or housing conditions that constitute the proxi-

mate risk factor for exposure. Indoor exposure to rodents

is common in patients with HCPS (38).

On the basis of these results, we argue that infections by

hantavirus follow a gradient of exposure to the virus in the

830 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 1. Relationship between independent variables and

seropositivity to hantaviruses, IX Region, Chile

Variables

Seropositive

(total)

p value

RR (95% CI)

Sex

Male

4 (395)

0.43

Female

2 (435)

Age

15–44 y of age

2 (550)

0.18

>44 y of age

4 (280)

Race

Mapuche

2 (151)

0.30

Other

4 (679)

Area

Andean

6 (271)

0.0001

Central

0 (273)

Coastal

0 (286)

Risk activity or labor

Agrarian

No activity

1 (271)

Two or more

4 (341)

0.38

Foresting

No activity

1 (279)

With activity

5 (135)

0.018

9.72

(1.15 to 82.44)

Exposure to rodents

Peridomestic

5 (731)

0.10

Laboral

5 (453)

1.00

p <0.05; p values were determined by using 2-tailed Fisher exact test; RR,

relative risk; CI, confidence interval.

Not applicable.

Clearing, shrubbery cutting, working in pastures or cellars.

Foresting, working at a sawmill.

Table 2. Incidence of hantavirus cardiopulmonary syndrome (HCPS) and seroprevalence in Chile’s IX Region according to geographic area

Region

Global

Andean (A)

Central (B)

Coastal

p value

(A vs. B)

RR (95% CI)

Disease

Cases

Population

781,242

127,974

424,278

229,190

0.0001

29.6 (11.46 to 76.47)

Cumulative Incidence

(1997-2001) (x10

)

4.35

22.66

0.23

1.74

Infection

Seropositive

Sample

829

271

272

286

Prevalence (%)

0.72

2.15

0.00013

p values were determined using 2-tailed Fisher exact test; RR, relative risk; CI, confidence interval.

B is Central and Coastal regions combined.

Chile’s Department of Health, March 2001.

Not applicable.

IX Region. Our population has a different epidemiologic

profile to those of the aborigines of Paraguay (40%) and

North Argentina (17%) in South America, which have low

seroprevalence, similar to that described for North

American populations, as reported by Vincent (13) and

Ferrer (26). This lower seroprevalence could be because of

a greater pathogenicity and clinical severity in infections

by the prototypical (southern) form of Andes virus, similar

to Sin Nombre virus.

Less pathogenic hantaviruses may cause a greater

amount of asymptomatic infections, as seen for HFRS in

Europe and Asia (seroprevalence 7.9% to 10%, death rate,

0.1% to 15%) (39) and some American hantaviruses

(Laguna Negra in Bolivia and Paraguay, Choclo in

Panama, Oran and Lechiguanas in Argentina) for which

seroprevalence is high and case-fatality ratios are <30%

(9,13,15). This finding is in contrast to the findings of han-

taviruses with severe clinical syndromes and high death

rate (i.e., Sin Nombre and Andes viruses, both of which

have been associated with few subclinical infections). The

lack of subclinical infections can be caused by a variation

in virulence or by different genotypes of the hosts, which

give them greater resistance to infection and disease.

A slightly disproportionate fraction of seropositive

samples (33%) were from Mapuche (18.3% of those sam-

pled were Mapuche). Moreover, when both hantavirus

seroprevalence and the HCPS incidence rates in this

region are considered, more infection but less disease (not

significant) is found in the Mapuche population (Table 3).

Future epidemiologic studies should address this finding

and use a larger sample to evaluate possible associations

between racial origin and the incidence rates of infection

and disease.

Acknowledgments

We thank Thomas G. Ksiazek for thoughtful technical assis-

tance in hantavirus serologic analysis, Jovita Mardones and

Ligia Sanhueza for providing assistance in collection and pro-

cessing of samples, Gustavo Jimenez for helping to critically

revise the manuscript, and Julio Valdivia for providing statistical

analysis.

Funding was provided by Child Health Foundation,

Alabama, Birmingham and a grant from Universidad de la

Frontera, Temuco, Chile and was supported in part by U.S. Public

Health Service grants U19 AI45452 and D43 TW01133.

Dr. Täger Frey is an assistant professor in the Pediatrics

Department at Universidad Austral de Chile. Her interests

include infectious diseases.

References

1. Schmaljohn C, Hjelle B. Hantaviruses: a global disease problem.

Emerg Infect Dis 1997;3:95–104.

2. Wells R, Young J, Williams R, Armstrong L, Busico K, Khan A, et al.

Hantavirus transmission in the United States. Emerg Infect Dis

1997;3:361–5.

3. Smadel JE. Epidemic hemorrhagic fever. Am J Public Health

1953;43:1327–30.

4. McCaughey C, Hart C. Hantaviruses. J Med Microbiol

2000;49:587–99.

5. Duchin J, Koster F, Peters C, Simpson G, Tempest B, Zaki S, et al.

Hantavirus pulmonary syndrome: a clinical description of 17 patients

with a newly recognized disease. N Engl J Med 1994;330:949–55.

6. Khan A, Khabbaz R, Armstrong L, Holman R, Bauer S, Graber J, et

al. Hantavirus pulmonary syndrome: the first 100 U.S. cases. J Infect

Dis 1996;173:1297–303.

7. Doyle T, Bryan R, Peters C. Viral hemorrhagic fevers and hantavirus

infections in the Americas. Infect Dis Clin North Am

1998;12:95–110.

8. Iversson LB, Travassos da Rosa APA, Rosa MBD, Lomar AV, Saski

M, LeDuc JW. Infeccao humana por Hantavirus no sul e sudeste do

Brasil. Rev Assoc Med Bras 1994;40:85–92.

9. Williams R, Bryan R, Mills J, Palma R, Vera I, de Velásquez F, et al.

An outbreak of hantavirus pulmonary syndrome in western Paraguay.

Am J Trop Med Hyg 1997;57:274–82.

10. Johnson A, Bowen M, Ksiasek T, Williams R, Bryan R, Mills J, et al.

Laguna Negra virus associated with HPS in western Paraguay and

Bolivia. Virology 1997;238:115–27.

11. López N, Padula P, Rossi C, Miguel S, Edelstein A, Ramírez E, et al.

Genetic characterization and phylogeny of Andes virus and variants

from Argentina and Chile. Virus Res 1997;50:77–84.

12. Toro J, Vega J, Khan A, Mills J, Padula P, Terry W, et al. An outbreak

of hantavirus pulmonary syndrome, Chile, 1997. Emerg Infect Dis

1998;4:667–702.

13. Vincent M, Quiroz E, Gracia F, Sánchez A, Ksiazek T, Kitsutani P, et

al. Hantavirus pulmonary syndrome in Panama: identification of

novel hantaviruses and their likely reservoirs. Virology

2000;277:14–9.

14. López N, Padula P, Rossi C, Lazaro ME, France-Fernández MT.

Genetic identification of a new hantavirus causing severe pulmonary

syndrome in Argentina. Virology 1996;220:223–6.

15. Lewis S, Rowe J, Morzunov S, Henria D, St. Jeor S. New hantavirus-

es causing pulmonary syndrome in central Argentina. Lancet

1997;349:998–9.

16. López N, Padula P, Rossi C, Miguel S, Edelstein A, Ramirez E, et al.

Genetic characterization and phylogeny of Andes virus and variants

from Argentina and Chile. Virus Res 1997;50:77–84.

17. Murua R. Ecología de los reservorios silvestres de hantavirus en

Chile. Rev Chil Infect 1998;15:79–83.

18. Levis S, Morzunov S, Rowe J, Enria D, Pini N, Calderon G, et al.

Genetic diversity and epidemiology of hantaviruses in Argentina. J

Infect Dis 1998;177:529–38.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 831

RESEARCH

Table 3. Incidence of hantavirus cardiopulmonary syndrome (HCPS) and seroprevalence in Chile’s IX Region according to race

Seroprevalence

Incidence

HCPS cases/population

Race

Positive/total (%)

RR (95% CI)

(rate x 100.000)

RR (95% CI)

Mapuche

2/151 (1.3)

2.23 (0.41 to 12.06)

4/205,466 (1.9)

0.37 (0.13 to 1.06)

Other

4/678 (0.6)

–

30/575,776 (5.2)

Relative risk (95% confidence interval). p values were determined by using two-tailed Fisher exact test; RR, relative risk; CI, confidence interval.

19. Kitsutani P, Denton R, Fritz C, Murray R, Todd R, Pape J, et al. Acute

Sin Nombre hantavirus infection without pulmonary syndrome,

United States. Emerg Infect Dis 1999;5:701–5.

20. Zavasky D, Hjelle B, Peterson M, Denton R, Reimer R. Acute infec-

tion with Sin Nombre hantavirus without pulmonary edema. Clin

Infect Dis 1999;29:664–6.

21. Weissenbacher M, Cura E, Segura E, Hortal M, Baek L, Chu Y, et al.

Serological evidence of human hantavirus infection in Argentina,

Bolivia and Uruguay. Medicina 1996;56:17–22.

22. Baró M, Vergara J, Navarrete M. Hantavirus in Chile. A review and

case analysis since 1975. Rev Méd Chile 1999;127:1513–23.

23. Parisi M, Enria D, Pini N, Sabattini M. Detección retrospectiva de

infecciones clínicas por hantavirus en la Argentina [in Spanish].

Medicina 1996;56:1–13.

24. Vitek CH, Breiman RF, Ksiazek T, Rollin P, McLaughlin J, Umland

E, et al. Evidence against person-to-person transmission of hantavirus

to health care workers. Clin Infect Dis 1996;22:824–6.

25. Zeitz PS, Butler JC, Cheek JE, Samuel MC, Childs JE, Shands LA, et

al. A case-control study of hantavirus pulmonary syndrome during an

outbreak in the southwestern United States. J Infect Dis

1995;171:864–70.

26. Ferrer J, Jonsson C, Esteban E, Galligan D, Basombrio M, Peralta-

Ramos M, et al. High prevalence of hantavirus infection in Indian

communities of the Paraguayan and Argentinean Gran Chaco. Am J

Trop Med Hyg 1998;59:438–44.

27. Valderrama R, Vega J, Terry W, Aguilar G, Gallegos N, Escobar P, et

al. Community serological survey of infection by hantavirus in the XI

Region, Aysen, Chile. In: The Fourth International Conference on

HFRS and Hantaviruses. 5–7 Mar 1998. Atlanta, GA, USA.

28. Holmes R, Boccanera R, Figueiredo L, Mancano S, Pane C.

Seroprevalence of human hantavirus infection in the ribeirao preto

region of São Paulo State, Brazil. Emerg Infect Dis 2000;6:560–1.

29. Mascarenhas-Batista AV, da Rosa ES, Ksiazek TG, da Rosa AP, Leduc

JW, Pinheiro F, et al. Anti-hantavirus antibodies in school children in

Salvador, Bahía. Rev Soc Bras Med Trop 1998;31:433–40.

30. Childs JE, Korch GW, Glass GE, LeDuc JW, Shah KV. Epizootiology

of hantavirus infections in Baltimore: isolation of a virus from

Norway rats, and characteristics of infected rat populations. Am J

Epidemiol 1987;126:55–68.

31. Korch G, Glass J, Rossi C, Le Duc J. Serologic evidence of hantavirus

infections within small mammal communities of Baltimore,

Maryland: spatial and temporal patterns an host range. Am J Trop

Med Hyg 1989;41:230–40.

32. Diglisic G, Rossi C, Doti A, Walshe D. Seroprevalence study of han-

tavirus infection in the community based population. Md Med J

1999;48:303–6.

33. Glass GE, Watson AJ, LeDuc JW, Kelen GD, Quinn TC, Childs JE.

Infection with a ratborne hantavirus in U.S. residents is consistently

associated with hypertensive renal disease. J Infect Dis

1993;167:614–20.

34. Rossi C, Ksiazek T. Enzyme-linked immunosorbent assay (ELISA).

In: Lee HW, Calisher C, Schmaljohn C, editors. Manual of hemor-

rhagic fever with renal syndrome and hantavirus pulmonary syn-

drome. Seoul: WHO Collaborating Center for Virus Reference and

Research (Hantaviruses) Asian Institute for Life Sciences; 1999. p.

87–91.

35. Hjelle B, Jenison S, Torrez-Martínez N, Herring B, Quan S, Polito A,

et al. Rapid and specific detection of Sin Nombre virus antibodies in

patients with hantavirus pulmonary syndrome by a strip inmunoblot

assay suitable for field diagnosis. J Clin Microbiol 1997;35:600–8.

36. Hjelle B, Torrez-Martínez N, Bharadwaj M. Recombinant western

blot and strip immunoblot assays for hantaviruses. In: Lee HW,

Calisher C, Schmaljohn C, editors. Manual of hemorrhagic fever with

renal syndrome and hantavirus pulmonary syndrome. Seoul: WHO

Collaborating Center for Virus Reference and Research

(Hantaviruses) Asian Institute for Life Sciences; 1999. p. 122–30.

37. Galeno H, Villagra E, Fernández J, Ramírez E, Mora J. Infección por

hantavirus en humanos: experiencia del laboratorio de referencia para

enfermedades infecciosas emergentes [in Spanish]. Rev Chil Infect

2000;17:216–9.

38. Hjelle B, Glass GE. Outbreak of hantavirus infection in the Four

Corners region of the United States in the wake of the 1997–1998 El

Nino-southern oscillation. J Infect Dis 2000;181:1569–73.

39. Clas A, Juto P, Stegmayr B, Settergren B, Wadell G, Tärnvik A, et al.

Prevalence of serum antibodies to hantaviruses in Northern Sweden

as measured by recombinant nucleocapsid proteins. Scand J Infect

Dis 1997;29:349–54.

Address for correspondence: Marlis Täger Frey, Instituto de Pediatría,

Facultad de Medicina Universidad Austral de Chile, Valdivia, Chile; fax:

56-63-229570; email: [email protected]

832 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Use of trade names is for identification only and does not imply

endorsement by the Public Health Service or by the U.S.

Department of Health and Human Services.

Clinical antimicrobial susceptibility breakpoints are

used to predict the clinical outcome of antimicrobial treat-

ment. In contrast, microbiologic breakpoints are used to

identify isolates that may be categorized as susceptible

when applying clinical breakpoints but harbor resistance

mechanisms that result in their reduced susceptibility to the

agent being tested. Currently, the National Committee for

Clinical Laboratory Standards (NCCLS) guidelines utilize

clinical breakpoints to characterize the activity of the fluoro-

quinolones against Streptococcus pneumoniae. To deter-

mine whether levofloxacin breakpoints can identify isolates

that harbor recognized resistance mechanisms, we exam-

ined 115 S. pneumoniae isolates with a levofloxacin MIC of

>2 µg/mL for first-step parC mutations. A total of 48 (59%)

of 82 isolates with a levofloxacin MIC of 2 µg/mL, a level

considered susceptible by NCCLS criteria, had a first-step

mutation in parC. Whether surveillance programs that use

levofloxacin data can effectively detect emerging resist-

ance and whether fluoroquinolones can effectively treat

infections caused by such isolates should be evaluated.

he emergence of Streptococcus pneumoniae resistance

to β-lactam and macrolide antimicrobial agents has led

to recommendations that fluoroquinolones with increased

activity against S. pneumoniae, such as levofloxacin, mox-

ifloxacin, and gatifloxacin, be used to treat patients at risk

for infection caused by such multidrug-resistant strains

(1–6). Fluoroquinolone resistance in S. pneumoniae is pri-

marily due to mutations in the genes encoding the target

topoisomerase enzymes, namely parC, which encodes the

A subunit of DNA topoisomerase IV, and gyrA, which

encodes the A subunit of DNA gyrase (7). Mutations in

parE and gyrB have been reported, but to a lesser extent

(8–10). Most pneumococcal isolates with reduced suscep-

tibilities to fluoroquinolones have amino acid substitutions

in either ParC alone or ParC and GyrA (11–14). Resistance

can also be mediated by active efflux (15), although the

role of efflux in contributing to resistance by the newer

fluoroquinolones is unclear (16).

The MIC of an antimicrobial agent is a value that has

been used to determine breakpoints that predict the proba-

bility of clinical success, detect resistant populations, or

both (17). Clinical breakpoints are dependent on the

antimicrobial activity and pharmacology of the drug; such

breakpoints are ascertained with the goals of eradicating

the infection and ultimately achieving clinical success with

the antimicrobial agent. In contrast, microbiologic break-

points are established to identify isolates that may be cate-

gorized as susceptible when applying clinical breakpoints

but harbor resistance mechanisms that result in their

reduced susceptibility to the agent being tested. These

microbiologic breakpoints are therefore useful in monitor-

ing the emergence of resistance. The current National

Committee for Clinical Laboratory Standards (NCCLS)

guidelines make no distinction between these two interpre-

tations of MIC, with clinical breakpoints used to character-

ize most antimicrobial agents, including the fluoro-

quinolones.

Levofloxacin has been used as a surrogate marker to

predict fluoroquinolone susceptibility in clinical laborato-

ries and surveillance studies (18). To establish whether

current levofloxacin breakpoints are also able to function

as microbiologic breakpoints, we determined the percent-

age of S. pneumoniae isolates with first-step parC muta-

tions that would go undetected by using the current

NCCLS breakpoints for levofloxacin (19).

Materials and Methods

A total of 6,076 clinical isolates of S. pneumoniae were

collected as part of a 1993–1998 surveillance program

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 833

RESEARCH

Antimicrobial Susceptibility

Breakpoints and First-Step

parC Mutations in Streptococcus

pneumoniae: Redefining

Fluoroquinolone Resistance

Sue Lim,*† Darrin Bast,*† Allison McGeer,*† Joyce de Azavedo,*† and Donald E. Low*†

*Toronto Medical Laboratories/Mount Sinai Hospital, Toronto,

Ontario, Canada; and †University of Toronto, Toronto, Ontario,

Canada

throughout Canada. All isolates were identified as S. pneu-

moniae by standard methods. The isolates were frozen at

–70°C, thawed, subcultured onto blood agar, and incubated

at 37°C in 5% CO

for 24 h twice before testing. In vitro

susceptibility testing was performed by broth microdilu-

tion, according to NCCLS guidelines (20,21).

Susceptibility interpretive criteria used were those pub-

lished in the NCCLS M100-S12 document (19). The non-

susceptible category was defined as those isolates with

MICs of fluoroquinolines in the intermediate and resistant

category. The parC gene of 115 isolates with a levofloxacin

MIC >2 µg/mL (82 = MIC 2 µg/mL; 8 = MIC 4 µg/mL; 10

= MIC 8 µg/mL; and 15 = MIC ≥16 µg/mL) was amplified

by polymerase chain reaction (PCR), and the nucleotide

sequence determined as previously described (9). All iso-

lates (n=33) with a levofloxacin MIC of ≥4 µg/mL, and a

random sample of 29 isolates with a levofloxacin MIC of 2

µg/mL were examined for gyrA mutations. For compara-

tive purposes, the parC gene of 14 isolates with a

ciprofloxacin MIC of 2 µg/mL, regardless of their lev-

ofloxacin MIC, was amplified and sequenced. Although

numerous single mutational events occur in parC, the focus

of this investigation was on amino acid substitutions for

Ser-79 or Asp-83, because previous studies have consis-

tently demonstrated that mutations at either of these posi-

tions are associated with decreased susceptibility (9,14).

Crude cell lysates were used as DNA templates for

PCR. After overnight growth on Columbia nutrient agar

and supplemented with 5% sheep blood, a loopful of

growth was suspended in 100 µL of lysis buffer (100 mM

NaCl, 10 mM Tris-HCl [pH 8.3], 1 mM EDTA, 1% Triton

X-100) and boiled for 10 min. Ten microliters of the super-

natant was used as the template in a 50-µL reaction vol-

ume. The quinolone-resistance–determining regions of

parC and gyrA were amplified by PCR. Primers used were

based on published sequences (7,8), and amplification

products were purified with either the QIAquick PCR

purification kit (Qiagen Inc., Mississauga, Ontario,

Canada) or the Concert Rapid PCR purification kit (Life

Technologies, Burlington, Ontario, Canada).

DNA sequencing was performed by ABI prism Big Dye

terminator cycle sequencing with the ABI 377 automated

sequencer (PE Applied Biosystems, Mississauga, Ontario,

Canada). Nucleotide and amino acid sequence compar-

isons were performed by the multiple-alignment sequence

function of Vector NTI Suite software (InforMax Inc.,

Bethesda, MD). The GenBank accession numbers for the

wild-type sequences used for comparison purposes were

Z67739 for parC and parE (22), AB010387 for gyrA, and

Z67740 for gyrB (23).

Isolates were examined for active efflux by agar dilu-

tion on Mueller-Hinton agar containing 5% sheep blood in

the presence of ciprofloxacin with or without 10 mg/mL of

reserpine (Sigma Chemical Co., St. Louis, MO) (24).

Strains for which a fourfold or greater decrease in the MIC

of ciprofloxacin existed in the presence of reserpine were

considered in this study to be positive for reserpine-inhib-

ited efflux. S. pneumoniae strain P121/1N27 and clinical

isolate BSP 823 were used as quality control strains, the

latter of which demonstrated a 16-fold decrease in the

ciprofloxacin MIC in the presence of reserpine (9).

Results

Of the 115 S. pneumoniae isolates with a levofloxacin

MIC >2 µg/mL, 78 (69%) had an amino acid substitution

in ParC (Ser-79 or Asp-83) (Table 1). Mutations in gyrA

were not found in any of the randomly selected isolates

with a levofloxacin MIC of 2 µg/mL, but were present in

three (38%) of eight isolates with a levofloxacin MIC of 4

µg/mL and in all isolates with a levofloxacin MIC ≥ 8

µg/mL (Table 2). The specific ParC amino acid substitu-

tions of the isolates and their corresponding levofloxacin

MICs are shown in Table 1. The most common substitution

was Ser-79 to Phe, accounting for 60% of all observed

amino acid substitutions. The prevalence of first-step ParC

amino acid substitutions among all strains according to

their levofloxacin and ciprofloxacin MICs is shown in

Table 3. Using the current MIC interpretive standards for

levofloxacin, 48 (59%) of 82 of isolates with a first-step

mutation fall in the susceptibility category of levofloxacin

834 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 1. ParC amino acid substitutions found in 115 Streptococcus pneumoniae isolates with levofloxacin MICs >2 µg/mL and

corresponding levofloxacin MICs

No. isolates inhibited by levofloxacin MIC (µg/mL) of

ParC amino acid

substitution

≥32

Total no. of strains

Ser79→Phe

Ser79→Tyr

Ser79→Ala

Asp83→Asn

Asp83→Gly

Asp83→Tyr

Asp83→Val

Asp83→Ala

No. isolates/total with

amino acid substitutions

48/82 (59%)

5/8

(63%)

10/10

11/11

4/4

78/115 (69%)

One isolate with no ParC amino acid substitution found to have active efflux; two isolates had ParC amino acid substitutions at sites other than Ser79 or Asp83.

(MIC<4 µg/mL). In comparison, 4 (29%) of 14 randomly

chosen isolates with a ciprofloxacin MIC of 2 µg/mL har-

bored a first-step mutation.

Thirty-three isolates were nonsusceptible to lev-

ofloxacin (MIC>4 µg/mL); for 25, which harbored both

ParC and GyrA amino acid substitutions, the levofloxacin

MIC was ≥8 µg/mL (Table 3). For eight isolates, the lev-

ofloxacin MIC was 4 µg/mL; three (38%) of those isolates

had a substitution in GyrA (Ser-81-Phe) as well as a sub-

stitution in ParC (Ser-79-Phe, Asp-78-Asn and Ala-115-

Pro) (Table 4). In addition, two of the eight (25%) isolates

had no substitution in GyrA, but were considered positive

for reserpine-inhibited efflux, while three isolates had a

Ser-79-Phe amino acid substitution in ParC. No mutations

were found in either parE or gyrB in the isolates with a

levofloxacin MIC of 4 µg/mL.

Discussion

Before the development of fluoroquinolones such as

levofloxacin, ofloxacin was used to determine trends of

pneumococcal fluoroquinolone resistance in the United

States (25). By using this system, an increase of ofloxacin-

nonsusceptible isolates from 2.6% in 1995 to 3.8% in 1997

was reported. However, the significance of such an

increase was questioned, since ofloxacin-resistant strains

could be observed with only a single topoisomerase muta-

tion, whereas for fluoroquinolones such as levofloxacin,

multiple mutations are required for a strain to be classified

as resistant according to NCCLS breakpoints (25–27). As

a consequence, ofloxacin was replaced by levofloxacin in

1998 as a marker for fluoroquinolone nonsusceptibility,

and not surprisingly, given levofloxacin’s increased activi-

ty against S. pneumoniae, fluoroquinolone resistance rates

were only 0.2% in 1998 and 1999 (25).

Since effective surveillance depends upon the ability to

detect the emergence of resistance, the prevalence of pneu-

mococci that harbor resistance mechanisms to the fluoro-

quinolones may not be accurately represented if surveil-

lance systems that rely on levofloxacin MIC data are used

(25,28–34). We found that 59% of isolates with a lev-

ofloxacin MIC of 2 µg/mL, a level considered susceptible

according to NCCLS criteria, had a first-step mutation in

parC. Similarly, Davies et al. (12) found that of 14 strains

for which levofloxacin MICs were 2 µg/mL, 10 (71%) had

a parC mutation. Therefore, if the goal of surveillance is to

detect emerging problems, then by extension, the detection

of first-step mutations is also important and the use of cur-

rent NCCLS breakpoints to estimate fluoroquinolone

resistance is clearly inadequate. Apart from DNA sequenc-

ing, currently no accurate test can reliably identify isolates

with first-step mutations (35). Although decreasing lev-

ofloxacin breakpoints has been proposed as a solution to

this problem, we found that 8 (25%) of 32 strains for which

the levofloxacin MIC was 1 µg/mL already had first-step

mutations (data not shown). Similarly, the replacement of

levofloxacin as a surveillance indicator by another fluoro-

quinolone has also been suggested. However, use of

ciprofloxacin does not fare significantly better, with 4

(29%) of 14 isolates in the susceptible category (MIC of 4

µg/mL to define nonsusceptible isolates) harboring first-

step mutations.

In addition to causing an underestimation of the emer-

gence of fluoroquinolone resistance, the use of clinical

breakpoints has therapeutic implications, as supported by

recent reports of treatment failure when a fluoroquinolone

was used to treat an infection caused by a strain of pneu-

mococci with a first-step mutation (36,37). Clearly, a first-

step mutation is necessary for the development of subse-

quent mutations, the latter of which result in MICs that fall

within the nonsusceptible category. However, studies have

shown that upon acquisition of a first-step mutation, the

likelihood of developing a subsequent mutation is

enhanced in comparison to the development of the first-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 835

RESEARCH

Table 2. Number of isolates with ParC and GyrA amino acid

substitutions and their corresponding levofloxacin MICs

No. strains with amino acid substitutions in

MIC (µg/mL)

ParC (%)

ParC and GyrA (%)

48/82 (59)

0/29

(0)

5/8 (63)

3/8 (38)

0/10 (0)

10/10 (100)

≥16

0/15 (0)

15/15 (100)

29/82 isolates were randomly examined for GyrA mutations.

Table 3. The prevalence of ParC amino acid substitutions

among all strains according to their levofloxacin and

ciprofloxacin MICs

No. strains with ParC amino acid

substitution at 79 or 83

MIC (µg/mL)

Ciprofloxacin (%)

Levofloxacin (%)

4/14 (29)

48/82 (59)

24/37 (65)

5/8 (63)

11/12 (92)

10/10 (100)

22/22 (100)

15/15 (100)

Total

62/87 (71)

78/115 (69)

Table 4. Characterization of Streptococcus pneumoniae isolates

with levofloxacin MIC 4 µg/mL

Amino acid substitution

Isolate

no.

In ParC

In GyrA

Change in MIC

with inhibition of

efflux

Ser79→Phe

None

8-fold

Ser79→Phe

None

No effect

Asp78→Asn

Ser81→Phe

No effect

Ala115→Pro

Ser81→Phe

No effect

Ser79→Phe

None

No effect

Ser79→Phe

None

No effect

None

4-fold

Ser79→Phe

Ser81→Phe

No effect

parE and gyrB sequencing was performed on all isolates, but no mutations were

found in the quinolone-resistance–determining region.

step mutation itself (38–40). Studies are required to deter-

mine whether isolates with one or more mutations in genes

encoding ParC, GyrA, or both, can still be effectively treat-

ed with a fluoroquinolone when that fluoroquinolone is

found to be susceptible by using current clinical break-

points. Recognizing the presence of underlying mutations

may be especially important when using these agents to

treat patients with large biomass infections such as pneu-

mococcal pneumonia.

Lastly, the acquisition of a second-step mutation

appears more likely than not to raise the MIC to >8 µg/mL

and not to 4 µg/mL as would be expected. Isolates with a

levofloxacin MIC of 4 µg/mL represented 0.1% of the total

number of isolates in our study, which is notable, consid-

ering that a levofloxacin MIC of 4 µg/mL is currently used

to define nonsusceptibility. Furthermore, these isolates

were for the most part either genotypically or phenotypi-

cally distinct from other isolates characterized (Table 4);

two had efflux mechanisms, one singly and the other con-

current with a ParC amino acid substitution, and two had

unusual substitutions in ParC (Asp78→Asn and

Ala115→Pro). The importance of this latter finding

remains to be determined.

In summary, levofloxacin susceptibility testing that

uses current MIC clinical breakpoints does not identify

most S. pneumoniae isolates with only first-step parC

mutations. This finding may not only have implications for

the ability of surveillance programs to detect emerging

resistance, but therapeutic implications as well.

Dr. Lim is a physician trained in internal medicine and cur-

rently a fellow in infectious diseases and medical microbiology

at the University of Toronto. Her research interests include

infections in the immunocrompromised host and transplant

recipients.

This work was supported by grants from Hoffman-La Roche

and the Canadian Bacterial Diseases Network.

References

1. Bartlett JG, Dowell SF, Mandell LA, File TM Jr, Musher DM, Fine A.

Practice guidelines for the management of community-acquired

pneumonia in adults. Clin Infect Dis 2000;31:347–82.

2. Niederman MS, Mandell LA, Anzueto A, Bass JB, Broughton WA,

Campbell GD, et al. Guidelines for the management of adults with

community-acquired pneumonia. Diagnosis, assessment of severity,

antimicrobial therapy, and prevention. Am J Respir Crit Care Med

2001;163:1730–54.

3. Mandell LA, Marrie TJ, Grossman RF, Chow AW, Hyland RH.

Canadian guidelines for the initial management of community-

acquired pneumonia: an evidence-based update by the Canadian

Infectious Diseases Society and the Canadian Thoracic Society. Clin

Infect Dis 2000;31:383–421.

4. Heffelfinger JD, Dowell SF, Jorgensen JH, Klugman KP, Mabry LR,

Musher DM, et al. Management of community-acquired pneumonia

in the era of pneumococcal resistance: a report from the Drug-

Resistant Streptococcus pneumoniae Therapeutic Working Group.

Arch Intern Med 2000;160:1399–408.

5. File TM Jr, Segreti J, Dunbar L, Player R, Kohler R, Williams RR, et

al. A multicenter, randomized study comparing the efficacy and safe-

ty of intravenous and/or oral levofloxacin versus ceftriaxone and/or

cefuroxime axetil in treatment of adults with community-acquired

pneumonia. Antimicrob Agents Chemother 1997;41:1965–72.

6. Petitpretz P, Arvis P, Marel M, Moita J, Urueta J. Oral moxifloxacin

vs high-dosage amoxicillin in the treatment of mild-to-moderate,

community-acquired, suspected pneumococcal pneumonia in adults.

Chest 2001;119:185–95.

7. Pan XS, Ambler J, Mehtar S, Fisher LM. Involvement of topoiso-

merase IV and DNA gyrase as ciprofloxacin targets in Streptococcus

pneumoniae. Antimicrob Agents Chemother 1996;40:2321–6.

8. Perichon B, Tankovic J, Courvalin P. Characterization of a mutation

in the parE gene that confers fluoroquinolone resistance in

Streptococcus pneumoniae. Antimicrob Agents Chemother

1997;41:1166–7.

9. Bast DJ, Low DE, Duncan CL, Kilburn L, Mandell LA, Davidson RJ,

et al. Fluoroquinolone resistance in clinical isolates of Streptococcus

pneumoniae: contributions of type II topoisomerase mutations and

efflux to levels of resistance. Antimicrob Agents Chemother

2000;44:3049–54.

10. Weigel LM, Anderson GJ, Facklam RR, Tenover FC. Genetic analy-

ses of mutations contributing to fluoroquinolone resistance in clinical

isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother

2001;45:3517–23.

11. Ho PL, Yam WC, Que TL, Tsang DN, Seto WH, Ng TK, et al. Target

site modifications and efflux phenotype in clinical isolates of

Streptococcus pneumoniae from Hong Kong with reduced suscepti-

bility to fluoroquinolones. J Antimicrob Chemother. 2001;47:655–8.

12. Davies TA, Evangelista A, Pfleger S, Bush K, Sahm DF, Goldschmidt

R. Prevalence of single mutations in topoisomerase type II genes

among levofloxacin-susceptible clinical strains of Streptococcus

pneumoniae isolated in the United States in 1992 to 1996 and 1999 to

2000. Antimicrob Agents Chemother 2002;46:119–24.

13. Jones ME, Sahm DF, Martin N, Scheuring S, Heisig P, Thornsberry

C, et al. Prevalence of gyrA, gyrB, parC, and parE mutations in clin-

ical isolates of Streptococcus pneumoniae with decreased susceptibil-

ities to different fluoroquinolones and originating from worldwide

surveillance studies during the 1997–1998 respiratory season.

Antimicrob Agents Chemother 2000;44:462–6.

14. Brueggemann AB, Coffman SL, Rhomberg P, Huynh H, Almer L,

Nilius A, et al. Fluoroquinolone resistance in Streptococcus pneumo-

niae in United States since 1994–1995. Antimicrob Agents

Chemother 2002;46:680–8.

15. Gill MJ, Brenwald NP, Wise R. Identification of an efflux pump gene,

pmrA, associated with fluoroquinolone resistance in Streptococcus

pneumoniae. Antimicrob Agents Chemother 1999;43:187–9.

16. Piddock LJ, Johnson MM, Simjee S, Pumbwe L. Expression of efflux

pump gene pmrA in fluoroquinolone-resistant and -susceptible clini-

cal isolates of Streptococcus pneumoniae. Antimicrob Agents

Chemother 2002;46:808–12.

17. Mouton JW. Breakpoints: current practice and future perspectives. Int

J Antimicrob Agents 2002;19:323–31.

18. Jones RN, Pfaller MA. Can antimicrobial susceptibility testing results

for ciprofloxacin or levofloxacin predict susceptibility to a newer flu-

oroquinolone, gatifloxacin? Report from the SENTRY Antimicrobial

Surveillance Program (1997–99). Diagn Microbiol Infect Dis

2001;39:237–43.

19. National Committee for Clinical Laboratory Standards. Performance

standards for antimicrobial susceptibility testing. Twelfth informa-

tional supplement, M100-S12. Wayne (PA): The Committee; 2002.

836 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

20. National Committee for Clinical Laboratory Standards. Methods for

dilution antimicrobial susceptibility tests for bacteria that grow aero-

bically, 5th ed. Approved standard M7-A5. Wayne (PA): The

Committee; 2000.

21. National Committee for Clinical Laboratory Standards. Performance

standards for antimicrobial disk susceptibility tests, 7th ed. Approved

standard M2-A7. 7th ed. Wayne (PA): The Committee; 2000.

22. Pan XS, Fisher LM. Cloning and characterization of the parC and

parE genes of Streptococcus pneumoniae encoding DNA topoiso-

merase IV: role in fluoroquinolone resistance. J Bacteriol

1996:178;4060–9.

23. Munoz R, Bustamante M, De La Campa AG. Ser-127-to-Leu substi-

tution in the DNA gyrase B subunit of Streptococcus pneumoniae is

implicated in novobiocin resistance. J Bacteriol 1995:177;4166–70.

24. Brenwald NP, Gill MJ, Wise R. Prevalence of a putative efflux mech-

anism among fluoroquinolone-resistant clinical isolates of

Streptococcus pneumoniae. Antimicrob Agents Chemother

1998;42:2032–5.

25. Centers for Disease Control. Resistance of Streptococcus pneumoni-

ae to fluoroquinolones—United States, 1995–1999. MMWR Morb

Mortal Wkly Rep 2001;50:800–4.

26. Jorgensen JH, Weigel LM, Ferraro MJ, Swenson JM, Tenover FC.

Activities of newer fluoroquinolones against Streptococcus pneumo-

niae clinical isolates including those with mutations in the gyrA,

parC, and parE loci. Antimicrob Agents Chemother 1999;43:329–34.

27. Jorgensen JH, Weigel LM, Swenson JM, Whitney CG, Ferraro MJ,

Tenover FC. Activities of clinafloxacin, gatifloxacin, gemifloxacin,

and trovafloxacin against recent clinical isolates of levofloxacin-

resistant Streptococcus pneumoniae. Antimicrob Agents Chemother

2000;44:2962–8.

28. Ho PL, Yung RW, Tsang DN, Que TL, Ho M, Seto WH, et al.

Increasing resistance of Streptococcus pneumoniae to fluoro-

quinolones: results of a Hong Kong multicentre study in 2000. J

Antimicrob Chemother 2001;48:659–65.

29. Jones ME, Karlowsky JA, Blosser-Middleton R, Critchley IA,

Karginova E, Thornsberry C, et al. Longitudinal assessment of antip-

neumococcal susceptibility in the United States. Antimicrob Agents

Chemother 2002;46:2651–5.

30. McGee L, Klugman KP, Wasas A, Capper T, Brink A. Serotype 19f

multiresistant pneumococcal clone harboring two erythromycin

resistance determinants (erm(B) and mef(A)) in South Africa.

Antimicrob Agents Chemother 2001;45:1595–8.

31. Critchley IA, Blosser RS, Karlowsky JA, Yamakita J, Barth A, Sader

HS, et al. Antimicrobial resistance in respiratory pathogens isolated

in Brazil during 1999–2000. Braz J Infect Dis 2001;5:294–304.

32. Thornsberry C, Jones ME, Hickey ML, Mauriz Y, Kahn J, Sahm DF.

Resistance surveillance of Streptococcus pneumoniae, Haemophilus

influenzae and Moraxella catarrhalis isolated in the United States,

1997–1998. J Antimicrob Chemother 1999;44:749–59.

33. Schmitz FJ, Perdikouli M, Beeck A, Verhoef J, Fluit AC. Molecular

surveillance of macrolide, tetracycline and quinolone resistance

mechanisms in 1191 clinical European Streptococcus pneumoniae

isolates. Int J Antimicrob Agents 2001;18:433–6.

34. Reinert RR, Simic S, Al Lahham A, Reinert S, Lemperle M,

Lutticken R. Antimicrobial resistance of Streptococcus pneumoniae

recovered from outpatients with respiratory tract infections in

Germany from 1998 to 1999: results of a national surveillance study.

J Clin Microbiol 2001;39:1187–9.

35. Richardson DC, Bast D, McGeer A, Low DE. Evaluation of suscep-

tibility testing to detect fluoroquinolone resistance mechanisms in

Streptococcus pneumoniae. Antimicrob Agents Chemother

2001;45:1911–4.

36. Davidson R, Cavalcanti R, Brunton JL, Bast DJ, de Azavedo JC,

Kibsey P, et al. Resistance to levofloxacin and failure of treatment of

pneumococcal pneumonia. N Engl J Med 2002;346:747–50.

37. Weiss K, Restieri C, Gauthier R, Laverdiere M, McGeer A, Davidson

RJ, et al. A nosocomial outbreak of fluoroquinolone-resistant

Streptococcus pneumoniae. Clin Infect Dis 2001;33:517–22.

38. Pestova E, Millichap JJ, Noskin GA, Peterson LR. Intracellular tar-

gets of moxifloxacin: a comparison with other fluoroquinolones. J

Antimicrob Chemother 2000;45:583–90.

39. Li X, Zhao X, Drlica K. Selection of Streptococcus pneumoniae

mutants having reduced susceptibility to moxifloxacin and lev-

ofloxacin. Antimicrob Agents Chemother 2002;46:522–4.

40. Gillespie SH, Dickens A. Fluoroquinolone resistance in

Streptococcus pneumoniae: evidence that gyrA mutations arise at a

lower rate and that mutation in gyrA or parC predisposes to further

mutation. Microb Drug Resist 2003;9:17–24.

Address for correspondence: Donald E. Low, Department of

Microbiology, Rm. 1487, Mount Sinai Hospital, 600 University Ave.,

Toronto, Ontario, Canada, M5G 1X5; fax: 416-586-8746; email:

[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 837

RESEARCH

Alterations in genes involved in the repair of DNA

mutations (mut genes) result in an increased mutation fre-

quency and better adaptability of the bacterium to stressful

conditions. W-Beijing genotype strains displayed unique

missense alterations in three putative mut genes, including

two of the mutT type (Rv3908 and mutT2) and ogt. These

polymorphisms were found to be characteristic and unique

to W-Beijing phylogenetic lineage. Analysis of the mut

genes in 55 representative W-Beijing isolates suggests a

sequential acquisition of the mutations, elucidating a plau-

sible pathway of the molecular evolution of this clonal fam-

ily. The acquisition of mut genes may explain in part the

ability of the isolates of W-Beijing type to rapidly adapt to

their environment.

uberculosis (TB) and AIDS cause more deaths in

adults worldwide than any other infectious disease.

Globally, the number of TB cases is growing at a rate of

2% per year. Resistance, especially multidrug-resistance

(MDR), is an increasing problem (1) and a growing hazard

to human health. Many outbreaks of MDR-TB, defined as

resistance to at least rifampicin and isoniazid, have been

reported, with poor response to therapy and very high dis-

ease and death rates. Some TB outbreaks have involved

patients with HIV co-infection (2,3). Although in several

instances, MDR outbreaks associated with a particular

genotype, such as the W strain, have been identified (4,5),

drug-susceptible variants of the W strain account for most

of this group of isolates characterized to date.

In 1995, the largest proportion of the Mycobacterium

tuberculosis strains from Beijing, China, shared a high

degree of similarity in IS6110 restriction fragment length

polymorphism (RFLP) patterns and identical spoligo pat-

terns (6). Subsequent molecular analyses have indicated

that the W and Beijing isolates constitute a single group of

strains designated as the W-Beijing genotype (Figure 1).

The global distribution and success of M. tuberculosis iso-

lates of the W-Beijing genotype have led to the hypothesis

that these strains may have selective advantages over other

M. tuberculosis strains. In addition to the W-MDR strain

identified in New York City, and areas in Cuba, Estonia,

Vietnam, and Russia, the W-Beijing genotype has been

significantly associated with drug resistance (7 and unpub.

data). Several studies have suggested that the W-Beijing

genotype strains are disseminating throughout the world

(7). In Vietnam, the proportion of W-Beijing strains was

71% in patients <25 years of age and 41% for those >25

years of age (8). Furthermore, W-Beijing strains have been

implicated in several TB epidemics globally, including

ones in New York, Texas, California, South Carolina, and

New Jersey in the United States (9) and South Africa,

Russia, and Spain (10). A recent study showed that 82% of

MDR strains isolated in a prison in Azerbaijan, Eastern

Europe, are of the W-Beijing genotype (11).

Ongoing research is focused on identifying the factors

responsible for the worldwide spread of the W-Beijing

strains and their ability to adapt and enhance their patho-

genicity or virulence. Identifying a possible mechanism for

increased adaptation of these bacteria to the human

immunologic host defense system or human interventions

such as anti-TB treatment is of the utmost importance.

Such mechanisms may indicate how the bacterium adapts

to the host, a prerequisite for an enhanced accumulation of

genomic mutations associated with resistance. In M. tuber-

culosis, resistance to antibiotics occurs because of genom-

ic mutations in certain genes, such as the katG gene for iso-

niazid (INH) resistance and the rpoB gene for rifampicin

resistance (12). In contrast to several other pathogens with

MDR phenotypes, plasmid or transposon-mediated mech-

anisms of resistance have not been reported in M. tubercu-

losis (13–15). Since resistance to bacteriostatic in M.

838 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Mutations in Putative Mutator

Genes of Mycobacterium

tuberculosis Strains of the

W-Beijing Family

Mina Ebrahimi Rad,* Pablo Bifani,† Carlos Martin,‡ Kristin Kremer,§ Sofia Samper,‡ Jean Rauzier,*

Barry Kreiswirth,¶ Jesus Blazquez,# Marc Jouan,* Dick van Soolingen,§ and Brigitte Gicquel*

*Institut Pasteur, Paris, France; †Institut Pasteur, Lille, France;

‡Universidad de Zaragoza, Zaragoza, Spain; §National Institute of

Public Health and the Environment, Bilthoven, the Netherlands;

¶Public Health Research Institute, Newark, New Jersey, USA; and

#National Institute of Health, Madrid, Spain

tuberculosis is exclusively due to genomic mutations, the

bacterium would benefit from an increased mutation rate.

Recent studies provided evidence for a role of mutator

phenotypes in the emergence of MDR clinical

Pseudomonas isolates (16). Such phenotypes not only

enable the bacteria to acquire resistance to antibiotics more

easily but also facilitate their adaptation to a new niche.

Bacteria can escape immune surveillance by modulating

bacterial resistance to host defense mechanisms (16–18).

This finding prompted us to investigate whether a similar

situation exists in M. tuberculosis. We have undertaken a

comprehensive comparative sequence analysis of selected

target genes to evaluate and study the presence of muta-

tions in putative genes expected to play a role in the muta-

tion frequency in such strains.

Mutated phenotypes commonly result from defects in

DNA repair (19). An in silico analysis suggested that most

mismatch repair systems (e.g., mutS, mutL, or mutH) were

missing in the M. tuberculosis genome (20). However, the

frequency of spontaneous mutations in M. tuberculosis (in

vitro cultures) is similar to that found in other bacteria-car-

rying mismatch repair systems (21), which suggests that

other DNA repair mechanisms must be present.

Hypothetical open reading frames (ORF), similar to genes

known to be responsible for the avoidance or repair of

DNA lesions resulting from the alkylation or oxidation of

nucleotides, are present in the genome of M. tuberculosis.

We searched for variations in these genes in 139 clinical

isolates to detect possible mutations that could allow an

enhanced adaptability to the host and increased resistance

to anti-TB drugs.

Methods

We searched for mut genes variation in 139 M. tubercu-

losis complex strains originating from 35 different coun-

tries. Ninety-four of these strains were selected because

they were representative strains characterized with 13 dif-

ferent genetic markers in previous studies (6,22).

This set comprised 125 M. tuberculosis strains, 1 M.

africanum, 8 M. bovis, 3 M. bovis BCG, and 2 M. microti.

Fifty-five strains had a W-Beijing genotype; 12 had an

MDR phenotype. Strains representing different branches

of the W-Beijing genotype were studied. Eight MDR M.

tuberculosis strains with a genotype other than Beijing

were included. Five M. tuberculosis strains of the W-

Beijing genotype and three strains of unrelated genotype

were obtained from the national program for surveillance

of MDR tuberculosis in Spain. Four M. tuberculosis W-

Beijing genotype strains isolated in the Netherlands and

one from Vietnam were included because they showed

spoligo patterns with fewer than nine spacers. Five other

W-Beijing genotype strains showed hybridization to an

additional spacer, as demonstrated by using the extended

set of spacers, two of which lacked hybridization to spac-

er 37. Strain W4 is part of a drug-susceptible outbreak in

New Jersey (4); W147 is a drug-resistant isolate widely

spread in Russia (23). Eleven strains were representative

of ancestral W-Beijing strains, which diverged early in the

evolution of the W-Beijing phylogenetic lineage. Finally,

29 strains of another frequently observed genotype, the

Haarlem genotype (6), were investigated.

The collection consisted of 55 W-Beijing genotype iso-

lates, 29 Haarlem genotype isolates, 8 strains of the

African genotype, 1 M. bovis strain, and 46 representatives

of other genotypes. Principal genetic grouping (PGG),

according to the polymorphism in katG and gyrA, was

known (24) for most of the isolates in this study. Seventy-

four strains belong to PGG 1, 54 to PGG 2, and 3 to PGG

3. All isolates were subjected to at least IS6110 RFLP typ-

ing and spoligotyping (6). Drug susceptibility was deter-

mined for 41 of 139 strains (Table 1 and 2). Several puta-

tive mut genes were annotated as such in the released

genome sequence of M. tuberculosis (25). In addition,

using the BLAST program (26), we identified Rv3908 as

an ORF carrying a mutT domain (27) and have since

named it mutT4.

Primers were designed to amplify putative mut genes:

mutY (5'-CCGGCGACGAATCGCTCGTT-3', 5'-

AGCTGGGACAGTCGTCGCGG-3'), mutM (5'-CTG-

GTTCGATGGTGATGACC-5', 5'-GTGCGCTCGACC-

CACAG-3'), mutT2 (5'-TCCGGATGATGATTTACCTCC-

3', 5'-TCCGCCGGGTCGGGGAC-3'), mutT1 (5'-ATCG-

TCGGCGTGCCGTG-3', 5'-GTCAGCGTCCTGCCCGG-

3'), mutT4 (5'-TCGAAGGTGGGCAA- ATCGTG-3' 5'-

TGGGGTTCGCTGGAAGTGG–3'), ogt (5'-CAGCGC-

TCGCTGGCGCC-3', 5'-GACTCAG CCGCTCGCGA-3'),

and mut T3 (5'-GTCACGTCTGTTAGGACCTC-3', 5'-

CGCGCAACGGCTGCCGG-3'). Similar primers were

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 839

RESEARCH

Figure 1. Characteristic patterns

of Mycobacterium tuberculosis

Beijing genotype strains. RFLP,

restriction fragment length poly-

morphism.

designed to amplify the rpoB gene (5'-TACGGTCGGC-

GAGCTGATCC-3', 5'- TACGG- CGTTTCGATGAACC-

3').

DNA sequencing was performed directly on the ampli-

fied fragments by using the dideoxy chain-termination

method with the Big Dye Terminator Cycle sequencing Kit

(Perkin Elmer Applied Biosystems, Courtaboeuf, France)

on a GeneAmp polymerase chain reaction (PCR) system

9600 (Perkin Elmer) and run on a DNA analysis system

model 373 or 3100 (Applied Biosystems). Sequences of

mutY, mutT2, mutT4, rpoB, mutT1, mutT3, and ogt of the

M. tuberculosis strains H37Rv, CDC1551, and MT210

were obtained from published sequences or at the TIGR

Web site (available from: URL: http://www.tigr.org/).

Results

We searched for allele variation in putative genes cod-

ing for DNA repair enzymes: mutT (which hydrolyzes 8-

oxo-deoxyguanosine triphosphate) (28), ogt (which

removes methyl groups from O6-methylguanine in DNA)

(29), mutM (formamidopyrimidine-DNA glycosylate)

(30), and mutY (specific adenine glycosylate) (31) in 12

MDR M. tuberculosis strains. Subsequently, we genotyped

for the observed single nucleotide polymorphism (SNP)

variation in 124 strains, members of the M. tuberculosis

complex, and in the three published sequences of M. tuber-

culosis strains. In total, the sampling comprises 55 W-

Beijing genotype M. tuberculosis strains, including 11

ancestral W-Beijing isolates (unpub. data ), 84 M. tubercu-

losis strains of other genotypes, and 1 M. bovis strain.

Several putative mut genes were annotated as such in

the released genome sequence of M. tuberculosis. A

BLAST search using the E. coli mutT sequences as tem-

plate identified, in addition to mutT1, mutT2, mutT3, the

hypothetical ORF Rv3908, which we have designated as

mutT4. The best matches with E. coli mutT gene were

observed for mutT2 and mutT4. Figure 2 depicts sequence

alignment of the conserved region of the different genes of

the M. tuberculosis genome showing similarity with mutT

of E. coli. The search for sequences similar to ogt, mutM,

and mutY identified a single ORF in each case. Primers

were designed for PCR amplification of all the genes men-

tioned above.

We first determined the sequences of the different

genes mentioned above in 12 MDR M. tuberculosis strains

(ZA20, ZA65, ZA67, ZA68, ZA69, ZA11, ZA16, ZA12,

ZA13, ZAA14, ZA17, and ZA19), including 5 W-Beijing

strains (ZA20, ZA65, ZA67, ZA78, and ZA69). For the

mutY, mutM, mutT1, and mutT3 putative genes, PCR

amplification was obtained in all strains tested, but

sequence analysis did not indicate any nucleotide changes

at these loci except for the same silent SNP in mutT3 in

strains with a Haarlem genotype. We confirmed these find-

ings by sequencing mutT1, mutT3, mutM, and mutY in a

collection of 26 MDR strains from North Africa. No vari-

ation was observed in mutT1 or mutM. Only one strain had

a major variation in mutY. All Haarlem strains carried one

characteristic silent mutation in mutT3 and one character-

istic mutation in ogt (Ser 15 replaced by Thr). These defin-

ing SNPs were also observed for all Haarlem strains of this

study. No other variations were observed in mutT1, mutT3,

mutM, or mutY. However, comparative sequence analysis

of H37Rv, CDC1551, and the five MDR–W-Beijing iso-

lates indicated polymorphisms in mutT2, mutT4, and ogt.

These mutations in mutT4, mutT2, and ogt were also found

in the W-Beijing strain 210 (TIGR) but not in MDR strains

other than those belonging to the W-Beijing genotype. We

therefore decided to extend this investigation and look for

mutations in these three genes in a collection of M. tuber-

culosis complex isolates, including well-defined branches

of the W-Beijing phylogenetic lineage (Table 1 and 2).

In 43 of 55 strains with a W-Beijing genotype, either

susceptible to bacteriostat or MDR, we found a mutation in

mutT4. Codon 48 (CGG) of the annotated ORF had been

changed to GGG, resulting in the amino acid substitution

of Arg by Gly (Table 1 and 2). All 11 W-Beijing isolates

known to be closely related to the ancestral W-Beijing

strain (AM, HI, N16, DU2, DV, LB2, KY, IK, 122(C11),

113, and 107(LB)) were found to have the wild-type geno-

type as all other 84 isolates with a genotype other than W-

Beijing.

Thirty-nine of 43 W-Beijing strains with the mutation

in mutT4 carried an additional mutation in mutT2 and in

ogt. The mutT2 mutation constitutes a change in codon 58

(GGA to CGA), resulting in an amino acid substitution of

Gly by Arg. The active site of the E. coli MutT enzyme

comprises amino acids 53, 56, 57, and 98. Therefore, a

mutation Gly to Arg at position 58 may have a important

effect on enzyme activity and lead to a mutator phenotype.

All 39 W-Beijing isolates carrying the mutT2 polymor-

phism at codon 58 also displayed a concurrent silent muta-

840 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure 2. MutT proteins’ sequences alignment. Mycobacterium

tuberculosis Rv2985(MutT1), Rv1160(MutT2), Rv0413(MutT3),

and Rv3908(MutT4) were selected from the M. tuberculosis

genome because of their annotation or after a BLAST analysis.

These sequences were compared to Escherichia coli mutT by

using alignments available from: URL: http://www.biochem.uthsc-

sa.edu/~barnes/mutt.html. The detected region of similarity is

shown here. # absolutely conserved residues; * residues that are

strongly conserved and that define the mutT or nudix motif.

tion in codon 12 (Gly GGG to GGA Gly) of the ogt gene.

Of four possible codons encoding for glycine, GGG and

GGA had the lowest relative synonymous codon usage

(RSCU) in genes with high expression levels (0.20 and

0.17, respectively, compared to 1.32 and 2.31 for GGU and

GGC). For genes with low expression levels, the RSCU

values are 0.92, 0.37, 0.65, and 2.06 for GGG, GGA, GGU,

and GGC, respectively (32).

The five W-Beijing isolates with a mutation in mutT4

and a wild-type mutT2 gene did not contain the ogt silent

mutation on codon 12 either. Instead, they all shared a din-

ucleotide substitution in codon 37 (ACC to CTC) of ogt,

resulting in amino acid substitution of Arg to Leu. These

five W-Beijing isolates of 43 with the mutT4 mutations,

without the mutT2 (codon 58) or the ogt (codon 12) muta-

tions, differed molecularly from all other W-Beijing iso-

lates in their spoligotype pattern and accompanying dele-

tion flanking the DR locus. Four of five were isolated from

Dutch patients in the Netherlands; the fifth originated from

a patient in Vietnam. The Vietnamese isolate (no. 94)

shared >95% IS6110 pattern similarity with the Dutch iso-

late 115 when standard RFLP analysis was used. Overall,

the five isolates were closely related to each other accord-

ing to IS6110 profiling (>90% similarity). Spacer 37 in the

DR locus of Dutch isolates 114 and 139 was absent, while

sample 115 was missing spacers 37 and 38, and 111 had a

deletion of spacers 38 and 39 but not spacer 37, suggesting

that these isolates may belong to a different sublineage. A

tentative phylogeny of the W-Beijing strains analyzed in

this study is proposed in Figure 3. Seven of nine MDR W-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 841

RESEARCH

Table 1. Characteristics of Mycobacterium tuberculosis complex strains originating from 35 different countries

Strains

Genotype

No.

isolates

Country of

isolation

Group

mutT2

mutT4

ogt

ZA20/65

W-Beijing

Spain

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

ZA67-69

W-Beijing

Spain

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

ZA11/16

Haarlem

Spain

Thr ACC 15 Ser AGC

ZA12-14/17

other

Spain

ZA19

M. bovis

Spain

ZA15

other

Spain

ZA60-62

W-Beijing

Spain

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

CDC1551

USA

H37Rv

USA

MT210

W-Beijing

USA

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

Mongolia

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

South Africa

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

Malaysia

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

China

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

Thailand

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

Malaysia

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

91/102-6

W-Beijing

Vietnam

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

110/116/119/12

4-5/140-2

W-Beijing

the Netherlands

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

133

W-Beijing

South Africa

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W4/10/126/129

W-Beijing

USA

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W99

W-Beijing

Singapore

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W147

W-Beijing

Russia

Gly GGA 58 CGA Arg

Arg CGG 48 GGG Gly

Gly GGG 12 Gly GGA

W-Beijing

Vietnam

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

111

W-Beijing

South Korea

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

115

W-Beijing

the Netherlands

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

5107(HG1)

W-Beijing

USA

Arg CGG 48 GGG Gly

Arg CGC 37 Leu CTC

114, 139

W-Beijing

the Netherlands

Arg CGG 48 GGG Gly

166(HD6)

W-Beijing

USA

Arg CGG 48 GGG Gly

165(001)

W-Beijing

USA

Arg CGC 37 Leu CTC

107(LB)

W-Beijing

USA

113

W-Beijing

the Netherlands

122(CI1)

W-Beijing

USA

IK/KY/LB2/DV

/DU2/HI

W-Beijing

Russia

N16

W-Beijing

USA

W-Beijing

USA

wt, wild-type alleles (identical to H37Rv strain); nd, not determined

Beijing strains carried missense mutations in two muT

genes (mutT2 and mutT4), and two had a missense muta-

tion in both mutT4 and ogt (Table 1 and 2).

No mutations in mutT4 or in mutT2 were observed in

any of the 84 M. tuberculosis complex strains, including

19 strains of PGG1, 54 strains of PGG2, and 2 strains of

PGG3; the strains originated from 29 countries and were a

genotype other than W-Beijing. A Thr15Ser mutation was

observed in 24 of 29 strains of the Harlem family. No other

change was observed in ogt.

Resistance to rifampicin in MDR strains was correlated

with mutations in the rpoB gene. The three tested MDR W-

Beijing strains isolated in Spain, with the mutations at the

mutT2 and mutT4 loci, harbored a different mutation in the

rpoB gene. These strains were isolated from patients who

had emigrated from Eastern Europe to Spain (ZA67,

842 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 2. Characteristics of Mycobacterium tuberculosis complex strains originating from 35 different countries

Strains

Genotype

No. isolates

Country of isolation

Group

mutT2

mut T4

ogt

Haarlem

USA

Thr ACC 15 Ser AGC

3,5,22,32,39,48,50

,52-3,55

Haarlem

Argentina

Thr ACC 15 Ser AGC

Haarlem

Vietnam

Thr ACC 15 Ser AGC

13/28

Haarlem

Sri Lanka

Thr ACC 15 Ser AGC

Haarlem

the Netherlands

Thr ACC 15 Ser AGC

57/59

Haarlem

Czech republic

Haarlem

Czech Republic

Thr ACC 15 Ser AGC

86/143/145

Haarlem

Bolivia

Thr ACC 15 Ser AGC

Haarlem

USA

Thr ACC 15 Ser AGC

Haarlem

Italy

Thr ACC 15 Ser AGC

123

Haarlem

Czech Republic

Thr ACC 15 Ser AGC

144/146-7

Haarlem

Bolivia

Apr-35

Africa

Rwanda

Africa

Uganda

40/120

Africa

Burundi

Africa

Central African Republic

Africa

Uganda

121

Africa

Central African Republic

BCG

the Netherlands

6/47/73/130

M. bovis

the Netherlands

Other

Tunisia

15/31

Other

Iran

Other

Canada

Other

Greenland

Other

USA

19/36/74

Other

India

25/62

M. microti

Other

Zimbabwe

Other

Ethiopia

38/42

Other

Tahiti

41/46

Other

Chile

Other

Tanzania

Other

Curacao

Other

Honduras

65/112

Other

the Netherlands

BCG

Japan

76/101/126

M. bovis

Argentina

BCG

Russia

89/95

Other

Spain

Other

the Netherlands

Other

Ecuador

100

M. africanum

the Netherlands

108

Other

China

118

Other

Honduras

nd, not determined; wt, wild-type alleles (identical to H37Rv strain).

ZA68, and ZA69). Analysis of the IS6110 RFLP of the

respective isolates showed a difference of a single band.

These findings suggest that the three strains may be relat-

ed. The acquisition of the three different mutations in the

rpoB gene leading to rifampicin resistance must have

occurred after the acquisition of mutations in the putative

nucleotide repair enzyme genes mutT4 and mutT2.

Discussion

Our results show that M. tuberculosis strains of the W-

Beijing genotype acquired missense mutations in DNA

repair genes. These M. tuberculosis W-Beijing genotype

strains are genetically highly conserved and widespread.

DNA repair genes have been previously shown to be asso-

ciated with mutator phenotypes in other microorganisms.

The success of this group of strains may result in part from

mutations in DNA repair enzymes, which might provided

a true selective advantage for these bacteria to adapt and

persist, including through the acquisition of resistance to

anti-TB drugs. Mutations in the DNA repair genes might

be the evolutionary answer of the TB bacillus to increase

adaptation to hosts. This adaptation will lead to increasing

trends in the TB epidemic in the coming decades. The

World Health Organization considers MDR and resistance

as a problem of local rather than of global importance (1).

If the relative contribution of W-Beijing genotype strains

to the current worldwide TB epidemic is increasing as sug-

gested (7), this approach should be revised. In areas with

an increasing problem with MDR-TB, such as Estonia and

Russia, W-Beijing genotype strains are predominantly

associated with MDR cases (33). In Germany, the relative

proportion of W-Beijing strains among isolates from resist-

ant cases has increased from 12% in 1995 to 35% in 2000

(unpub. data). The latter observation may indicate an

increasing influence of W-Beijing strains on the worldwide

TB epidemic.

We identify polymorphisms in M. tuberculosis in genes

that might result in a mutator phenotype and therefore a

plausibly better adaptation of the bacilli to a hostile envi-

ronment (34). Forty-three of 55 W-Beijing isolates ana-

lyzed were found to have a unique mutation on the ORF

Rv3908. This ORF contains a MutT domain and is denot-

ed here as mutT4. Thirty-nine of 43 W-Beijing strains car-

ried an additional and identical mutation in a second puta-

tive gene of the mutT family, mutT2, and an identical silent

mutation in ogt.

The W-Beijing phylogenetic lineage probably acquired

the mutation on codon 48 of the mutT4 only once and

before other mutations associated with the mutator genes

we describe. This mutation clearly distinguishes ancestral

W-Beijing isolates from contemporary W-Beijing strains.

The 11 W-Beijing isolates that did not have the character-

istic mutT4 mutation on codon 48, consist of a collection

of isolates known to be ancestral within this phylogenetic

lineage, as determined by various other molecular tech-

niques (unpub. data).

Nine of W-Beijing strains with the wild-type mutT2

gene had a characteristic mutation on codon 37 of the ogt

gene, which suggests that these isolates constitute a branch

of the W-Beijing family that diverged after the acquisition

of the mutT4 mutation but before the development of the

nucleotide substitution on mutT2. One strain carries the

mutation 37 in ogt but no mutation in mutT4, a reversion

that might have occurred after a transient mutator pheno-

type.

A mutation in mutT2 was always associated with a

mutation in muT4. A first mutation may have occurred in

mutT4 and thereafter a second mutation either in mutT2 or

ogt was acquired. As observed for other bacterial popula-

tions, mutator phenotypes may be transient in many cases

to limit deleterious effects (35). Identifying these muta-

tions may aid in the identification of mut genes in M.

tuberculosis. These mutations associated with mutator

genes provide a reliable tool for the identification of W-

Beijing isolates and thus a useful marker for strains

endowed with capacity to yield epidemics. The biologic

consequences of these mutations and function of these

DNA repair genes are currently been investigated in the

laboratory.

Nine MDR strains with a W-Beijing genotype were

among strains carrying two missense mutations in putative

mutator genes. Phylogenetically unrelated M. tuberculosis

MDR isolates had no mutations within the DNA repair

genes investigated in this study. Our data support the idea

that M. tuberculosis strains of the W-Beijing genotype may

have adapted to hostile environments, including exposure

to anti-TB drugs, because of a succession of alterations of

DNA repair enzymes. Other genes involved in other DNA

repair mechanisms or in the fidelity of DNA replication

may also be involved and remain to be investigated.

The acquisition of mutator alleles was described as an

adaptive response of bacteria to a succession of different

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 843

RESEARCH

Figure 3. Schematic representation of a plausible pathway to

explain the accumulation of mutations in mut genes.

environments (18,35,36). After infecting a host, M. tuber-

culosis has to adapt to different environments such as alve-

olar macrophages and dendritic cells and subsequently to

granuloma containing inactivated macrophages or to acti-

vated macrophages after induction of the acquired immune

responses. In addition, the bacilli have to adapt to the

caseous media with low oxygen concentration in the cen-

ter of tubercles and to different types of tissues during dis-

semination of the disease. Such variable growth conditions

might select for mutations in M. tuberculosis strains, as

described in other bacterial populations exposed to differ-

ent environmental challenges. Mutations and selection

might occur with an increased frequency caused by the

toxic radicals produced in phagocytic cells.

However, a mutator phenotype is often transient.

Otherwise a continual accumulation of mutations would

lead to deleterious effects and loss of fitness. No difference

in the frequency of spontaneous mutations, resulting in a

rifampin resistance phenotype, was observed for W-

Beijing strains (37). We suggest that a transient mutator

phenotype allowed a better adaptation of W-Beijing

strains. Subsequent compensatory mutations occurred to

reverse the mutator phenotype. An alternative hypothesis

would be the existence of a higher mutation rate in specif-

ic conditions (i.e., in mutagenic radicals inside phago-

cytes). The accumulation of mutations leading to antibiot-

ic resistance in W-Beijing strains may be a consequence of

the appearance of strains with a better adaptation to the

hosts. MDR strains would be easily selected when patients

with strains that have adapted better received inadequate

anti-TB regimens.

Acknowledgments

We thank Maxime Schwartz and Lluis Quintana-Murci for

many helpful discussions and Xavier Nouvel for help with the

manuscript.

This work received support from the European Commission

(grant QLK2-CT-2000-00630), and from Louis D. French

Academy of Sciences award.

Dr. Rad is a molecular biologist. Her expertise includes

DNA amplification, sequencing, and cloning.

References

1. Raviglione MC, Gupta R, Dye CM, Espinal MA. The burden of drug-

resistant tuberculosis and mechanisms for its control. Ann N Y Acad

Sci 2001;953:88–97.

2. Farmer P, Kim JY. Community based approaches to the control of

multidrug resistant tuberculosis: introducing “DOTS-plus.” BMJ

1998;317:671–4.

3. Telzak EE, Sepkowitz K, Alpert P, Mannheimer S, Medard F, el-Sadr

W, et al. Multidrug-resistant tuberculosis in patients without HIV

infection. N Engl J Med 1995;333:907–11.

4. Bifani PJ, Mathema B, Liu Z, Moghazeh SL, Shopsin B, Tempalski

B, et al. Identification of a W variant outbreak of Mycobacterium

tuberculosis via population-based molecular epidemiology. JAMA

1999;282:2321–7.

5. Bifani PJ, Mathema B, Kurepina NE, Kreiswirth BN. Global dissem-

ination of the Mycobacterium tuberculosis W-Beijing family strains.

Trends Microbiol 2002;10:45–52.

6. Kremer K, van Soolingen D, Frothingham R, Haas WH, Hermans

PW, Martin C, et al. Comparison of methods based on different

molecular epidemiological markers for typing of Mycobacterium

tuberculosis complex strains: interlaboratory study of discriminatory

power and reproducibility. J Clin Microbiol 1999;37:2607–18.

7. Glynn JR, Whiteley J, Bifani PJ, Kremer K, van Soolingen D.

Worldwide occurrence of Beijing/W strains of Mycobacterium tuber-

culosis: a systematic review. Emerg Infect Dis 2002;8:843–9.

8. Anh DD, Borgdorff MW, Van LN, Lan NT, van Gorkom T, Kremer

K, et al. Mycobacterium tuberculosis Beijing genotype emerging in

Vietnam. Emerg Infect Dis 2000;6:302–5.

9. Bifani PJ, Plikaytis BB, Kapur V, Stockbauer K, Pan X, Lutfey ML,

et al. Origin and interstate spread of a New York City multidrug-

resistant Mycobacterium tuberculosis clone family. JAMA

1996;275:452–7.

10. Caminero JA, Pena MJ, Campos-Herrero MI, Rodriguez JC, Garcia

I, Cabrera P, et al. Epidemiological evidence of the spread of a

Mycobacterium tuberculosis strain of the Beijing genotype on Gran

Canaria Island. Am J Respir Crit Care Med 2001;164:1165–70.

11. Pfyffer GE, Strassle A, van Gorkum T, Portaels F, Rigouts L, Mathieu

C, et al. Multidrug-resistant tuberculosis in prison inmates,

Azerbaijan. Emerg Infect Dis 2001;7:855–61.

12. Ramaswamy S, Musser JM. Molecular genetic basis of antimicrobial

agent resistance in Mycobacterium tuberculosis: 1998 update. Tuber

Lung Dis 1998;79:3–29.

13. Van Rie A. Analysis for a limited number of gene codons can predict

drug resistance of Mycobacterium tuberculosis in a high incidence

community. J Clin Microbiol 2001;39:636–41.

14. Telenti A, Imboden P, Marchesi F, Lowrie D, Cole S, Colston MJ, et

al. Detection of rifampicin-resistance mutations in Mycobacterium

tuberculosis. Lancet 1993;341:647–50.

15. Finken M, Kirschner P, Meier A, Wrede A, Bottger EC. Molecular

basis of streptomycin resistance in Mycobacterium tuberculosis:

alterations of the ribosomal protein S12 gene and point mutations

within a functional 16S ribosomal RNA pseudoknot. Mol Microbiol

1993;9:1239–46.

16. Oliver A, Canton R, Campo P, Baquero F, Blazquez J. High frequen-

cy of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung

infection. Science 2000;288:1251–4.

17. Giraud A, Matic I, Tenaillon O, Clara A, Radman M, Fons M, et al.

Costs and benefits of high mutation rates: adaptive evolution of bac-

teria in the mouse gut. Science 2001;291:2606–8.

18. Richardson AR, Stojiljkovic I. Mismatch repair and the regulation of

phase variation in Neisseria meningitidis. Mol Microbiol

2001;40:645–55.

19. Horst JP, Wu TH, Marinus MG. Escherichia coli mutator genes.

Trends Microbiol 1999;7:29–36.

20. Mizrahi V, Andersen SJ. DNA repair in Mycobacterium tuberculosis.

What have we learnt from the genome sequence? Mol Microbiol

1998;29:1331–9.

21. David HL, Newman CM. Some observations on the genetics of iso-

niazid resistance in the tubercle bacilli. Am Rev Respir Dis

1971;104:508–15.

22. Supply P, Lesjean S, Savine E, Kremer K, van Soolingen D, Locht C.

Automated high-throughput genotyping for study of global epidemi-

ology of Mycobacterium tuberculosis based on mycobacterial inter-

spersed repetitive units. J Clin Microbiol 2001;39:3563–71.

844 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

23. Kurepina N, Ramaswamy S, Shashkina EF, Sloutsky AM, Blinova

LN, Mishustin SP, et al. The sequence analysis of the pncA gene

determining the PZA-resistance in the predominant M. tuberculosis

strains isolated in the Tomsk penitentiary system, Western Siberia,

Russia. In: 32nd World Conference on Lung Health of the

International Union against Tuberculosis and Lung Disease. Paris,

France; 2001.

24. Sreevatsan S, Pan X, Stockbauer KE, Connell ND, Kreiswirth BN,

Whittam TS, et al. Restricted structural gene polymorphism in the

Mycobacterium tuberculosis complex indicates evolutionarily recent

global dissemination. Proc Natl Acad Sci U S A 1997;94:9869–74.

25. Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D, et al.

Deciphering the biology of Mycobacterium tuberculosis from the

complete genome sequence. Nature 1998;393:537–44.

26. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W,

et al. Gapped BLAST and PSI-BLAST: a new generation of protein

database search programs. Nucleic Acids Res 1997;25:3389–402.

27. Fujii Y, Shimokawa H, Sekiguchi M, Nakabeppu Y. Functional sig-

nificance of the conserved residues for the 23-residue module among

MTH1 and MutT family proteins. J Biol Chem 1999;274:38251–9.

28. Maki H, Sekiguchi M. MutT protein specifically hydrolyses a potent

mutagenic substrate for DNA synthesis. Nature 1992;355:273–5.

29. Margison GP, Cooper DP, Potter PM. The E. coli ogt gene. Mutat Res

1990;233:15–21.

30. Boiteux S, O’Connor TR, Laval J. Formamidopyrimidine-DNA gly-

cosylase of Escherichia coli: cloning and sequencing of the fpg struc-

tural gene and overproduction of the protein. EMBO J

1987;6:3177–83.

31. Au KG, Clark S, Miller JH, Modrich P. Escherichia coli mutY gene

encodes an adenine glycosylase active on G-A mispairs. Proc Natl

Acad Sci U S A 1989;86:8877–81.

32. Andersson GE, Sharp PM. Codon usage in the Mycobacterium tuber-

culosis complex. Microbiology 1996;142:915–25.

33. Kruuner A, Hoffner SE, Sillastu H, Danilovits M, Levina K, Svenson

SB, et al. Spread of drug-resistant pulmonary tuberculosis in Estonia.

J Clin Microbiol 2001;39:3339–45.

34. Tenaillon O, Toupance B, Le Nagard H, Taddei F, Godelle B.

Mutators, population size, adaptive landscape and the adaptation of

asexual populations of bacteria. Genetics 1999;152:485–93.

35. Taddei F, Radman M, Maynard-Smith J, Toupance B, Gouyon PH,

Godelle B. Role of mutator alleles in adaptive evolution. Nature

1997;387:700–2.

36. Rainey BP, Moxon ER. Microbiology. When being hyper keeps you

fit. Science 2000;288:1186–7.

37. Werngren J, Hoffner SE. Drug-susceptible Mycobacterium tuberculo-

sis Beijing genotypes does not develop mutation-conferred resistance

to rifampin at an elevated rate. J Clin Microbiol. In press 2003.

Address for correspondence: Brigitte Gicquel, Unité de Génétique

Mycobactérienne, Institut Pasteur, 28 rue du Dr. Roux, 75015 Paris,

France; fax: 33145688843; email: bgicquel@pasteur.fr

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 845

RESEARCH

Search ppast iissues oof EEID aat wwww.cdc.gov/eid

During 5,230 trapping nights, 672 small mammals

were trapped in the areas where most hantavirus pul-

monary syndrome (HPS) cases occur in Uruguay. Yellow

pygmy rice rats (Oligoryzomys flavescens) were the only

rodents that showed evidence of antibodies to hantavirus,

with a seroprevalence of 2.6%. The rodents were trapped

in all the explored environments, and most of the seropos-

itive rodents were found in habitats frequented by humans.

Nucleotide sequences were obtained from four HPS case-

patients and four yellow pygmy rice rats of the M genome

segment. Sequence comparison and phylogenetic analysis

showed that rodent-borne viruses and viruses from three

HPS case-patients form a well-supported clade and share

a 96.4% identity with the previously characterized Central

Plata hantavirus. These results suggest that yellow pygmy

rice rat (O. flavescens) may be the host for Central Plata, a

hantavirus associated with HPS in the southern area of

Uruguay.

he family Bunyaviridae consists of five genera.

Viruses in the Hantavirus genus are unique among

them because all members (except Thottapalayam virus)

are rodent-borne. Viruses in the other four genera are

arthropod-borne. Hantavirus pulmonary syndrome (HPS)

was first identified in the United States in 1993. The dis-

covery of the outbreak was followed by the identification

of Sin Nombre virus (SNV) as the primary etiologic agent

of HPS (1). Since these findings, many countries in the

Americas have identified cases and outbreaks of this syn-

drome, and several other related viruses (New World han-

taviruses) have been recognized (2–9).

New World hantaviruses are carried by different species

of sigmodontine and arvicoline rodents (Muridae). Indeed,

genetic diversity and geographic distribution of these

viruses are related to the genetic diversity, geographic dis-

tribution, and phylogenetic history of their rodent hosts. In

South America, studies of the correlation between rodent

hosts and indigenous hantaviruses are complicated by the

great diversity of sigmodontine rodents in this area. Also,

the sympatric distributions between the different species of

sigmodontine rodents in South America provide opportu-

nities for spillover infections and host-switching events

(10).

In Uruguay, the first evidence of the circulation of these

viruses came from a study of serum specimens collected

from blood donors between 1985 and 1987 that showed a

seroprevalence of 1%, as measured by indirect fluorescent

antibody (IFA) test using Hantaan and Seoul antigens (11).

Since 1997, the Ministerio de Salud Pública, through the

Departamento de Laboratorios,

began the surveillance and

diagnosis of HPS. In 2000 Padula et al. (12) reported par-

tial sequences (G1 and G2 glycoprotein) derived from two

HPS cases that occurred in Uruguay in 1997 and 1999.

These viruses clustered within a previously reported line-

age named Central Plata.

Knowledge about small mammal communities and

habitat preferences is limited in Uruguay. However, some

studies about systematic distribution, reproduction, and

cytogenetic aspects have been published (13–17).

Research regarding the distribution and habitat preferences

of the Muridae family in Uruguay is currently being con-

ducted (18–20).

The purpose of this study was to identify the carrier

rodents of hantavirus in Uruguay and their potential associ-

ation with HPS cases, to determine the prevalence of infec-

tion in different habitats, and to begin to genetically charac-

terize the hantaviruses recovered from these rodents.

846 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Yellow Pygmy Rice Rat

(Oligoryzomys flavescens) and

Hantavirus Pulmonary Syndrome

in Uruguay

Adriana Delfraro,* Mario Clara,* Lorena Tomé,* Federico Achaval,* Silvana Levis,†

Gladys Calderón,† Delia Enria,† Mario Lozano,‡ José Russi,§ and Juan Arbiza*

*Facultad de Ciencias de la Universidad de la República,

Montevideo, Uruguay; †Instituto Nacional de Enfermedades

Virales Humanas “Dr. Julio I. Maiztegui,” Pergamino, Buenos

Aires, Argentina; ‡Universidad Nacional de Quilmes, Bernal,

Argentina; and §Ministerio de Salud Pública, Montevideo,

Uruguay

Material and Methods

HPS Case Identification

National surveillance for HPS was reinforced when a

case definition was established by the Ministerio de Salud

Pública in 1997. An HPS case was suspected in a previous

healthy person with an acute febrile illness (temperature

>38°C), associated with dyspnea, acute respiratory distress

syndrome with pulmonary noncardiogenic edema, or inter-

stitial bilateral infiltrates, hypotension or shock, elevated

leukocyte count, and thrombocytopenia (21). A case of

HPS was confirmed when, in addition to clinical illness,

circulating specific hantavirus immunoglobulin (Ig) M

was detected.

Human serum samples were tested for the presence of

IgM and IgG antibodies with an enzyme-linked

immunosorbent assay (ELISA) developed by MRL

(Hantavirus ELISA IgM and Hantavirus ELISA IgG, MRL

Diagnostics, Cypress, CA). The test was used to screen

patients, and in every case, the results were confirmed by

retesting the specimens by an in-house enzyme immunoas-

say with a recombinant nucleocapsid antigen specific to

Andes virus, according to the procedure developed by

Padula et al. (12).

Site Selection

Rodent sampling was conducted at the most likely sites

of infection for known HPS case-patients and included the

places where the person had lived or worked in the 6

weeks before onset of symptoms and nearby natural habi-

tats. The trapping sites were classified as 1) domestic and

peridomestic, including all sites in the immediate vicinity

of houses, sheds, gardens, road borders, and fence lines,

and 2) rural natural ecosystems and agro-ecosystems,

including representative habitats of each area such as open

fields, cultivated areas, wetlands, shrublands, brook bor-

ders, natural forests, and artificial woods (planted by

humans) (Table 1).

The trapping expeditions were performed in the follow-

ing areas: Puntas de Valdéz (34°32′S/56°36′W) and

Piedritas (34º20′S/55°39′W) (one expedition each);

Cerrillos (34°38′S/56°19′W), Melilla (34º44′S/56°16′W),

and Sauce (34°35′S/56°08′W) (two expeditions each)

(Figure 1). The geographic area covered by the trapping

expeditions corresponded to areas where 16 HPS cases

occurred in Montevideo and Canelones, two cases

occurred in San José, and one case occurred in Florida.

The other 19 cases were dispersed in the southern half of

Uruguay, and for some of them, the probable site of infec-

tion was not clearly identified.

The landscape of Canelones, rural Montevideo, Florida,

and San José shows cultivated areas, stubble areas, shrub-

lands in the territories abandoned by rural people, range

lands, natural and artificial woodlands, small wetlands,

and small brooks. In recent years, many rural inhabitants

have migrated to the cities; abandoned farmlands have thus

been transformed into shrublands.

Small-Mammal Trapping and Processing

Small mammals were trapped in six expeditions in the

above-mentioned areas from May 1997 until September

2001. Each trapping site was sampled with Sherman live-

capture traps (model LFATDG 23 cm x 8 cm x 8.5 cm)

(Sherman Traps Inc., Tallahassee, FL). The number of

traps depended on the available area for trap placement at

each trapping site. The traps were placed at 5-m intervals

in line transects, along the different environments at the

trapping site. The traps were set out in the late afternoon

and checked in the early morning for the next two

mornings. The animals were trapped and sampled

according to established biosafety guidelines (22). Each

animal was anesthetized, and blood was collected from the

retroorbital sinus. The animals were humanely killed, and

their size, mass, sex, and reproductive status were

recorded. Samples of liver, kidney, lung, and brain were

extracted and stored in liquid nitrogen for further

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 847

RESEARCH

Table 1. Trapping efficiency/environment/species in the total of captures

Env

288

9.4

1,010

172

17.0

1,420

3.1

265

5.2

680

8.4

190

26.3

1,198

268

22.4

117

179

22.3

5,230

672

12.8

194

198

157

Sp%

0.6

28.9

3.3

29.5

5.1

7.7

0.3

0.4

23.4

0.3

0.6

Env, environments; TN, trapping nights; C, captures; E, efficiency; NW, natural woods; RB, road borders; PD, peridomestic; WE, wetlands; BB, brook borders; AG,

agroecosystems; SH, shrublands; AW, artificial woods; T, totals; Sp%, species %; Md, Monodelphis dimidiate; Of, Oligoryzomys flavescens; Od, O. delticola; St,

Scapteromys tumidus; Aa, Akodon azarae; No, Necromys obscurus; Hb, Holochilus brasiliensis; Cl, Calomys laucha; Mm, Mus musculus; Rr, Rattus rattus; Ca, Cavia

aperea.

processing. The individual animals were tentatively

identified in the field by external characteristics, and the

carcasses were kept in 10% formalin. Identification was

confirmed by cranial measurements and dental

examination at the laboratory. All specimens were

deposited at the collection of the Sección Zoología

Vertebrados, Facultad de Ciencias, Universidad de la

República, Montevideo, Uruguay.

ELISA

Serologic testing of rodents was performed by IgG

ELISA. Briefly, the IgG ELISAs were performed by coat-

ing polyvinyl chloride microtiter plates (Dynex

Technologies, Chantilly, VA) overnight at 4°C with a

Lechiguanas virus (LECV) antigen (inactivated, 3 M rad

gamma-ray irradiation detergent-extracted lysate of Vero-

E6 infected cells, with a 100% infection index controlled

by indirect immunofluorescence). An uninfected Vero E6

cell culture antigen was used to determine the specificity

of mouse antibodies. Unbound antigen was removed by

washing three times with phosphate-buffered saline (PBS)-

Tween 20, 0.1% (Sigma-Aldrich, St. Louis, MO). After

blocking with PBS-Tween 20, 0.1%-dry milk 5% (37°C, 1

h), sera diluted fourfold, beginning with 1:100, were added

to react with the antigen-coated wells. Bound antigen was

measured by the use of a hyperimmune mouse ascitic fluid

and by using goat anti–Peromyscus leucopus IgG (H+L)

and goat-anti rat IgG (heavy- and light-chain–specific;

Kirkegaard & Perry Laboratories, Gaithersburg, MD) con-

jugated to horseradish peroxidase. Optical densities (ODs)

at 405 nm were recorded on a microplate spectrophotome-

ter (Labsystems Multiskan EX; Thermo Labsystems,

Finland, Vartaa, Finland), and the ODs of the uninfected

antigen-coated well were subtracted from that of its corre-

sponding viral antigen to yield the adjusted OD. A serum

dilution was considered positive if OD was >0.2 U after

adjustment. A serum titer >

400 was considered positive.

Total RNA Extraction and RT-PCR

Total RNA was extracted from lung tissue of seroposi-

tive rodents and from blood clots from HPS case-patients.

Approximately 100 mg of tissue was treated with 1 mL of

TRIzol reagent (GIBCO BRL, Life Technologies,

Rockville, MD), according to manufacturer’s instructions.

An M genome segment of the G2 glycoprotein encoding

region was amplified by using reverse transcription-poly-

merase chain reaction (RT-PCR) and specific oligonu-

cleotides as previously described by Levis et al. (4).

RT was carried out using MMLV reverse transcriptase

(GIBCO BRL) and the oligonucleotide 3348(-)

(5'CTGTCCAGATTTAGTGTTCCA 3'). cDNA was then

precipitated with NaAc (ICN Biomedicals, Costa Mesa,

CA) 3 M pH 5.6, ethanol (Merck Química Argentina,

Buenos Aires, Argentina), and lineal polyacrylamide 2.5

µg/µL (ICN Biomedicals), and resuspended in 20 µLof

double distilled water. Two microliters of first-strand

cDNA was used in the PCR reaction. Two rounds of PCR

were performed by using Taq DNA polymerase (GIBCO

BRL). The first round was performed with oligonu-

cleotides 3348 (-) and 2765 (+) (5'CTGTATGTGAGTAC-

CAAG 3'), and the second round (heminested) was per-

formed with 1 µL of first-round reaction and oligonu-

cleotides 3221 (-) (5'TCAGAAGAGCAGTCAGTGT-

CATG 3') and 2765 (+), giving a 456-nucleotide (nt) frag-

ment. PCR fragments were visualized on ethidium bro-

mide 1.5% agarose gels.

Sequencing and Phylogenetic Analysis

The PCR fragments obtained from rodent and HPS

case-patient samples were purified for further sequencing

by using the Concert rapid gel extraction system (GIBCO

BRL) or QIAquick gel extraction kit (QIAGEN Inc.,

Valencia, CA). Nucleotide sequencing was conducted by

using the oligonucleotide 2765(+) and an ABI 377 Genetic

Analyser (PE Applied Biosystems, Inc., Foster City, CA).

Alignment of sequences was done by using

CLUSTALX (1.5) (23). Phylogenetic analyses and

sequence comparison were carried out with PAUP* 4.0b10

(24) and MEGA version 2.1 (25). Maximum parsimony

848 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Figure 1. Locations of capture sites and enzyme-linked

immunosorbent assay–confirmed hantavirus pulmonary syndrome

case-patients in Uruguay. Capture sites: triangle = Puntas de

Valdéz, diamond = Piedritas, circle = Sauce, star = Cerrillos,

square = Melilla.

analysis was carried out by using the heuristic search

option. Maximum parsimony trees were searched by apply-

ing the tree bisection reconnection branch-swapping algo-

rithm. A consensus tree was obtained through 50% majori-

ty rule consensus. For the distance-based approach, MOD-

ELTEST 3.06 (26) was used to establish the most suitable

model of DNA substitution that best fitted our dataset, and

a phylogenetic tree was obtained by using the neighbor-

joining algorithm. Bootstrap analysis (27) was performed

to estimate topologic accuracy of the trees (500 replicates),

and only values >70% were considered significant.

For comparison, existing sequence data from GenBank

were used: hantavirus sequences from Argentina

(GenBank accession nos. AF028023 to AF028027,

AF028029 to AF028063), Central Plata genotype from

Uruguay (GenBank accession nos. AY101184 and

AY101185), and Sin Nombre virus (L37903, isolate

NMR11); the last one was used as outgroup.

Results

HPS Cases

From April 1997 to August 2002, 38 cases of HPS were

confirmed by ELISA, with a fatality rate of 21.0%.

Twenty-four (63.2%) of these cases occurred in rural or

suburban areas of Montevideo and Canelones, 6 (15.7%)

in Colonia, 3 in Soriano (7.9%), 2 in San José (5.3%), 2 in

Rocha (5.35%), and 1 in Florida (2.6%) (Figure 1). As of

August 2002, HPS cases in Uruguay had occurred in the

southern half of the country.

Distribution of Rodents by Species and Capture Site

During 5,230 trap-nights, 672 small mammals were

collected (trap success = 12.8%). The trapped small mam-

mals belonged to two families (Muridae and Caviidae)

within the order Rodentia and one family (Didelphidae) in

the order Didelphimorphia. The mammals belonged to 11

species, 75.1% of the captured animals corresponded to the

Sigmodontinae subfamily, 23.7 % to the Murinae subfam-

ily, 0.6% to the Caviidae family, and 0.6% to the

Didelphidae family (Table 1).

Captures and percentage of trap success by habitats

were as follows: natural woodlands, 27 (9.4%) of 288;

road borders, 172 (17.0%) of 1,010, peridomestic areas, 44

(3.1%) of 1,420; wetlands,14 (5.2%) of 65; brook borders,

57 (8.4%) of 680; agroecosystems, 50 (26.3% of 190;

shrublands, 268 (22.4%) of 1,198; and in artificial wood-

lands, 40 (22.3%) of 179 (Table 1). The most common cap-

tured small mammals were the following: swamp rats

(Scapteromys tumidus), 198 (29.5%); yellow pygmy rice

rats (Oligoryzomys flavescens), 194 (28.9%); and house

mice (Mus musculus), 157 (23.4%) (Table 1). No sigmod-

ontine rodents were found inside the houses, where only

house mice and black rats (Rattus rattus) were found.

Yellow pygmy rice rats were found in areas of human dis-

turbance such as peridomestic areas, agroecosystems, road

borders, and shrublands. We found that the trapping suc-

cess in these sites was higher than in natural areas. As

shown in Table 1, yellow pygmy rice rats were found in all

of the habitats where traps were set.

Screening for Hantavirus Infection of Rodents

Serum specimens collected from rodents were screened

for IgG antibodies to LECV by ELISA. As mentioned

above, 672 small mammals were trapped in six areas

where HPS cases occurred between 1997 and 2001. Anti-

LECV antibodies were detected in five rodents (O.

flavescens) from four different locations. Absorbances

with LECV antigen of positive samples screened at 1:400

dilution were at least fourfold the absorbance of the nega-

tive control. Further titration showed that three samples

had titers >1:1,600, and one had a titer >1:6,400 (Table 2).

The proportion of positive rodents in the different locali-

ties ranged from 2.1% to 2.9%. Piedritas was the only

locality where no antibody-positive rodents were recorded

(Table 3).

Total RNA Isolation, RT-PCR, and Sequence Analysis

Total viral RNA was extracted from the lungs of the

five seropositive yellow pygmy rice rats and blood clots

from four case-patients. RT of viral RNA and PCR ampli-

fication of a 456-nt fragment of the G2

glycoprotein–encoding region of the virus M genome seg-

ment (bases corresponding to LECV 2,805–3,215) and

nucleotide sequences were obtained from four rodent sam-

ples and four human blood clots. Amplified DNA was not

recovered from the rodent sample CE155, which had the

lower antibody titer (Table 3). A 292-nt segment (LECV

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 849

RESEARCH

Table 2. Hantavirus-seropositive rodents found in the different geographic areas where captures were performed

Sample

Rodent species

Geographic area

Habitat

Antibody titer (arbitrary units)

U89

Oligoryzomys flavescens

Puntas de Valdéz (San José)

Road border

1,600

SA63

O. flavescens

Sauce (Canelones)

Peridomestic

1,600

Ce20

O. flavescens

Melilla (Montevideo)

Peridomestic

>6,400

Ce22

O. flavescens

Melilla (Montevideo)

Shrublands

1,600

Ce155

O. flavescens

Cerrillos (Canelones)

Shrublands

400

Specimens deposited at the Specimen Collection of the Sección Zoología Vertebrados, Facultad de Ciencias, Universidad de la República, Montevideo, Uruguay with

the following numbers: U89=ZVC-M2154, SA63=ZVC-M2155, Ce20=ZVC-M2156, Ce22=ZVC-M2157, and Ce155=ZVC-M2158.

2,815–3,106 G2 glycoprotein–encoding region) was used

for further comparison and phylogenetic analysis.

Sequence Comparison and Phylogenetic Analysis

For phylogenetic comparisons, a 292-nt fragment of the

M gene from lung RNA of four yellow pygmy rice rats

(GenBank accession nos. AY204677 to AY204680), as

well as clot RNA from four HPS case-patients (GenBank

accession nos. AF283896 to AF283899) was used. These

nucleotide sequences were compared with the equivalent

region of published hantavirus sequences. Phylogenetic

analysis indicated that two previously known hantavirus

genotypes are circulating in Uruguay: Central Plata and

LEC (Figure 2).

By both maximum parsimony and distance-based

analysis, the four sequences recovered from Uruguayan

yellow pygmy rice rats were closely related to each other

and formed a monophyletic group with the hantavirus

sequences derived from three HPS case-patients from

Canelones and Montevideo and two HPS case-patients

from the same geographic area, previously characterized as

Central Plata. This clade was supported by high bootstrap

values (Figure 2A,B). Comparison of these sequences at

the nucleotide level showed 96.4% identity. The most

closely related genotype was LEC, with 87.9% identity,

followed by Bermejo (85.0%), Orán (83.1%), Andes

(82.5%), and Hu39694 (82.1%). The less-related geno-

types of the Argentinean hantaviruses were Maciel (79.3%)

and Pergamino (78.6%). One viral sequence from an HPS

case-patient in Soriano clustered into LEC genotype.

Conclusion

Most HPS cases were in rural and suburban areas of

Montevideo and Canelones (24 of 38 cases) (Figure 1) in

southern Uruguay. Rodent sampling was conducted at the

most likely sites of infection for known HPS case-patients.

The trapping success rate was higher in the environ-

ments influenced by humans (agroecosystems, road bor-

ders, shrublands, artificial woods, peridomestic areas) than

in the natural areas (natural woods, wetlands, brook bor-

ders): 574 (85.4%) of trapped individual animals were cap-

tured in environments influenced by humans, and 98

(14.6%) were captured in natural environments. Swamp

rats, yellow pygmy rice rats, and house mice were the most

frequently trapped species. Of 44 (3.1%) rodents trapped

in peridomestic environments, 19 (43.2%) were yellow

pygmy rice rats (Table 1). Five seropositive yellow pygmy

rice rats were captured in modified environments: one was

captured along a road border, two were captured in perido-

mestic environments, and two were captured in shrub-

lands, at <

150 m from homes. Three of five seropositive

rodents were therefore trapped in environments frequented

by humans (road borders and peridomestic environments).

These findings could indicate an increased risk for infec-

tion for human inhabitants.

Yellow pygmy rice rats were the only rodents that

showed evidence of antibodies to hantavirus, with a preva-

lence of 2.6%. Researchers have found that hantavirus

seroprevalence in rodents may vary widely, according to

the season, geographic area, altitude, and rodent species

analyzed (29–33). We found that the percentage of

seropositive rodents (2.6%) is the same as encountered in

the central zone of Argentina (2.6%) (32), although the

habitats are not similar to the southern area of Uruguay. In

Uruguay, we found that only yellow pygmy rice rats were

antibody positive, while in central Argentina seropositive

yellow pygmy rice rats, Azara’s field mice (Akodon

azarae), dark mice (Necromys benefactus), and small

water rats (Holochilus brasiliensis) were found (2–4). In

the different locations in Uruguay, seroprevalence was

similar, ranging from 2.1% to 2.9%. In Piedritas, where no

positive rodents were found, only four yellow pygmy rice

rats were trapped. All seropositive rodents in Uruguay

were adult males, which is consistent with horizontal

transmission and in accordance with the findings of sever-

al authors (30–32).

The phylogenetic analysis on a 292-nt region of the M

segment showed that these rodent sequences clustered

together with those from five Uruguayan HPS case-

patients from the same geographic area (Canelones and

Montevideo); these data suggest that the yellow pygmy

rice rat can be considered as the putative reservoir host for

Central Plata hantavirus in this region of Uruguay. This

study also showed the circulation of LEC genotype in the

western location of Soriano, 250 km from Montevideo,

separated from the Argentinean central HPS-endemic area

by the Uruguay River. This virus shared a 99% identity at

the nucleotide level with LEC genotype.

Phylogenetic analysis shows that the genotype Central

Plata recovered from rodents and HPS case-patient from

850 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Table 3. Small mammals trapped in the different sites, number and % of positives Monodelphis dimidiate

Area

+/%

Total

-/-

1/2.6

-/-

0/0

-/-

0/0

-/-

131

-/-

1/2.1

0/0

-/-

0/0

-/-

153

-/-

2/2.9

0/0

-/-

0/0

-/-

0/0

-/-

0/0

297

-/-

0/0

-/-

0/0

-/-

0/0

-/-

0/0

1/2.4

0/0

-/-

0/0

-/-

173

0/0

-/-

0/0

-/-

0/0

Total

0/0

194

5/2.6

0/0

198

0/0

157

0/0

672

Md, Monodelphis dimidiate; Of, Oligoryzomys flavescens; Od, O. delticola; St, Scapteromys tumidus, Aa, Akodon azarae, No, Necromys obscurus; Hb, Holochilus brasiliensis; Cl, Calomys laucha; Mm, Mus musculus; Rr

,Rattus rattus; Ca, Cavia aperea; PV, Puntas de Valdéz; Ce, Cerrillos; Me, Melilla; Pi, Piedritas, Sa, Sauce; Ca, Canelones. O. flavescens, S. tumidus, and M. musculus were the most frequently captured rodents.

Canelones, San José, and Melilla is phylogenetically dis-

tinct from (although related to) the previously described

LEC genotype, whose reservoir host in Argentina is also

the yellow pygmy rice rat. Hantaviruses have been associ-

ated with subspecies of closely related rodents: Sin

Nombre–like hantaviruses with mice from the genus

Peromyscus (34) and Andes virus recovered in southwest-

ern Argentina and Orán virus in northwestern Argentina,

both recovered from long tailed pygmy rice rats (O. longi-

caudatus) (4). Recent studies have shown that these two

rodent populations differ with respect to their mitochondr-

ial DNA (10). This fact raises the question of whether

rodents morphologically identified as O. flavescens in

Uruguay are indeed a different subspecies of O. flavescens

in Argentina. Further experiments will be needed to iden-

tify both the interspecific and intraspecific phylogenetic

relationships of O. flavescens in these regions.

Acknowledgments

We thank Noemí Pini and Juan Cristina for critical reading

and useful suggestions; Raúl Maneyro, Juan José Porta, and Ana

Liñares for rodent trapping; W. Slenczka for contributing biosafe-

ty equipment; Leandro Jones, Mónica Galiano, and Guillermo

D´Elía for help in Modeltest implementation.

This research was supported in part by grant no. 37/2000

from the Pan American Health Organization/Red

Latinoamericana de Ciencias Biológicas (Silvana Levis, Mario

Lozano, Juan Arbiza, and Mario Clara). Mario E. Lozano is a

research career member of Consejo Nacional de Investigaciones

Cientificas y Técnicas, Argentina. A. Delfraro received a predoc-

toral training fellowship from Programa de desarrollo de las

Ciencias Básicas. L. Tome was supported by a grant from

Laboratorio Santa Elena.

Adriana Delfraro is a doctoral student in microbiology and

assistant professor in the Sección Virología of the Facultad de

Ciencias, Universidad de la República. She also worked in the

Hantavirus Program of Uruguay in the Departamento de

Laboratorios of the Ministerio de Salud Pública, Montevideo,

Uruguay, from its beginning in 1997 until 2002.

References

1. Nichol ST, Spiropoulou C, Morzunov S, Rollin P, Ksiazek G,

Feldman H, et al. Genetic identification of a Hantavirus associated

with an outbreak of acute respiratory illness. Science

1993;262:914–7.

2. López N, Padula P, Rossi C, Lazaro ME, Franze-Fernandez MT.

Genetic identification of a new hantavirus causing severe pulmonary

syndrome in Argentina. Virology 1996;220:223–6.

3. López N, Padula P, Rossi C, Miguel S, Edelstein A, Ramirez E, et al.

Genetic characterization and phylogeny of Andes virus and variants

from Argentina and Chile. Virus Res 1997;50:77–84.

4. Levis S, Morzunov S, Rowe J, Enria D, Pini N, Calderon G, et al.

Genetic diversity and epidemiology of hantaviruses in Argentina. J

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 851

RESEARCH

Figure 2. A: Maximum parsimony phylogenetic tree B: Distance-

based phylogenetic tree. The tree was built under the Tamura-Nei

model of DNA substitution with estimation of the shape parameter

of the gamma distribution (28). This model and the associated

parameters resulted from testing our dataset with the program

MODELTEST 3.06 (26). Both trees include Argentinean and

Uruguayan hantaviruses from hantavirus pulmonary syndrome

(HPS) case-patients and rodents. Hantavirus sequences from

HPS case-patients in Uruguay: Ac1Hu, Ca2Hu, Bu3Hu, Pe4Hu,

Uru1/97, Uru1/99. Hantavirus sequences from yellow pygmy rice

rats: U89, Sa63, Ce20, Ce22. Sin Nombre virus was used as out-

group.*Specimens deposited at the Collection of the Sección

Zoología Vertebrados, Facultad de Ciencias, Universidad de la

República, Montevideo, Uruguay, with the following numbers:

U89=ZVC-M2154, SA63=ZVC-M2155, Ce20=ZVC-M2156,

Ce22=ZVC-M2157, and Ce155=ZVC-M2158.

Infect Dis 1998;177:529–38.

5. Vasconcelos M, Lima V, Iversson L, Rosa M, Travassos da Rosa E,

Travassos da Rosa A, et al. Hantavirus pulmonary syndrome in the

rural area of Juquitiba, Sao Paulo, Metropolitan Area, Brasil. Rev Inst

Med Trop (S. Paulo) 1997;39:237–8.

6. Johnson AM, Bowen MD, Ksiazek TG, Williams RJ, Bryan RT, Mills

J, et al. Laguna Negra virus associated with HPS in western Paraguay

and Bolivia. Virology 1997;238:115–27.

7. Johnson AM, de Souza LTM, Ferreira IB, Pereira LE, Ksiazek G,

Rollin P, et al. Genetic investigation of novel hantaviruses causing

fatal HPS in Brazil. J Med Virol 1999;59:527–35.

8. Vincent MJ, Quiroz E, Gracia F, Sanchez AJ, Ksiazek TG, Kitsutani

PT, et al. Hantavirus pulmonary syndrome in Panama: identification

of novel hantaviruses and their likely reservoirs. Virology

2000;277:14–9.

9. Padula P, Gonzalez Della Valle M, Garcia Alai M, Cortada P, Villagra

M, Gianella A. Andes virus and first case report of Bermejo virus

causing fatal pulmonary syndrome. Emerg Infect Dis 2002 8:437–9.

10. Plyusnin A, Morzunov SP. Virus evolution and genetic diversity of

hantaviruses and their rodent hosts. In: Schmaljohn CS, Nichol ST,

editors. Hantaviruses. Berlin: Springer-Verlag; 2001. p. 47–75.

11. Weissenbacher M, Cura E, Segura E, Hortal M, Baek L, Kyu Chu Y,

et al. Serological evidence of human hantavirus infection in

Argentina, Bolivia and Uruguay. Medicina (Buenos Aires)

1996;56:17–22.

12. Padula P, Colavecchia S, Martínez P, González Della Valle M,

Edelstein A, Miguel S, et al. Genetic diversity, distribution and

serological features of hantavirus infection in five countries in South

America. J Clin Microbiol 2000;38:3029–35.

13. Devicenzi GJ. Mamíferos del Uruguay. Anales del Museo de Historia

Natural de Montevideo 1935;4:1–96.

14. Langguth A. Contribución al conocimiento de los Cricetinae del

Uruguay (Especies halladas en los regurgitos de búho). Anais do

Segundo Congresso Latino-Americano de Zoología (Sao Paulo,

16–21/7/1962). 1965;2:327–35. San Pablo, Brasil: Sociedad

Latinoamericana de Zoología. Secretaria de Agricultura de San

Pablo.

15. Barlow JC. Observations on the biology of rodents in Uruguay. Life

Sciences Contributions of Royal Ontario Museum, 1969. Toronto:

The Museum, 1969. p. 75.

16. Brum-Zorrilla N. Investigaciones citogenéticas sobre algunas

especies de Cricetinae (Rodentia) del Uruguay. Anais do Segundo

Congresso Latino-Americano de Zoología (Sao Paulo, 16-

21/7/1962), 1965, 2:315–20. San Pablo, Brasil: Sociedad

Latinoamericana de Zoología. Secretaria de Agricultura de San

Pablo.

17. Brum-Zorrilla N, Degiovanangelo C, Di Tomaso MV. Estado actual

de los estudios cromosómicos en mamíferos del Uruguay. Actas de

las Jornadas de Zoología del Uruguay (Montevideo, 23-28/9/1985);

1985:77–8. Montevideo, Uruguay: Sociedad Zoologica del Uruguay;

1985.

18. Clara M, Achaval F. Datos preliminares sobre reservorios de

Hantavirus en Uruguay. Bol. Soc. Zool. Uruguay (2ª época). Act. V.

Jorn Zool Uruguay 1999;11:12.

19. Clara M, Achaval F. Preferencia de hábitat de algunas especies de

Muridae (Rodentia) del Uruguay. Act. VI . Jorn Zool Uruguay 2001;

33.

20. Clara M, Achaval F, Arbiza J, Delfraro A, Tomé L, Lozano M, Enría

D. Caracterización de Hantavirus y Arenavirus en sus reservorios

naturales en Uruguay. Act. VI . Jorn Zool Uruguay 2001;34.

21. Aguila J, Roucco G, Curto S, Willat G, Russi J, Gaye G, Salvatella R.

Guía para la vigilancia epidemiológica de los casos de Síndrome

Pulmonar por Hantavirus. Montevideo, Uruguay: Ministerio de Salud

Pública; 1998.

22. Mills J, Childs J, Thomas G, Ksiazek G, Peters CJ. Methods for trap-

ping and sampling small mammals for virologic testing. editors.

Atlanta: Centers for Disease Control and Prevention; 1995.

23. Thompson JD, GibsonTJ, Plewniak F, Jeanmougin F, Higgins DG.

The ClustalX windows interface: flexible strategies for multiple

sequence alignment aided by quality analysis tools. Nucleic Acids

Res 1997;24:4876–82.

24. Swofford, DL. PAUP*: Phylogenetic analysis using parsimony (and

other methods), v. 4.0b4a. 2000. Sunderland (MA): Sinauer

Associates, Inc. Publishers; 2000.

25. Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: Molecular evo-

lutionary genetics analysis software. Tempe (AZ): Arizona State

University; 2001.

26. Posada D, Crandall KA. MODELTEST: testing the model of DNA

substitution. Bioinformatics Application Note 1998;14:817–8.

27. Felsenstein J. Confidence limits on phylogenies: an approach using

the bootstrap. 1985. Evolution 1985;39:783–91.

28. Tamura K, Nei M. Estimation of the number of nucleotide substitu-

tions in the control region of mitochondrial DNA in humans and

chimpanzees. Mol Biol Evol 1993;10:512–26.

29. Childs JE, Ksiazek TG, Spiropoulou CF, Krebs JW, Morzunov S,

Maupin GO, et al. Serologic and genetic identification of Peromyscus

maniculatus as the primary rodent reservoir for a new hantavirus in

the southwestern United States. J Infect Dis 1994;169:1271–80.

30. Kuenzi AJ, Morrison ML, Swann DE, Hardy PC, Downard GT. A

longitudinal study of Sin Nombre virus prevalence in rodents, south-

eastern Arizona. Emerg Infect Dis 1999;5:113–7.

31. Mills JN, Ksiazek TG, Peters CJ, Childs JE. Long-term studies of

Hantavirus reservoir populations in the southwestern United States: a

synthesis. Emerg Infect Dis 1999;5:135–41.

32. Calderón G, Pini N, Bolpe J, Levis S, Mills, J, Segura E, et al.

Hantavirus reservoir hosts associated with peridomestic habitats in

Argentina. Emerg Infect Dis 1999;5:792–7.

33. Kuenzi AJ, Douglass RJ, White Jr D, Bond CW, Mills JN. Antibody

to Sin Nombre virus in rodents associated with peridomestic habitats

in west central Montana. Am J Trop Med Hyg 2001;64:137–46.

34. Morzunov SP, Rowe JE, Ksiazek TG, Peters CJ, St Jeor SC, Nichol,

ST. Genetic analysis of the diversity and origin of hantaviruses in

Peromyscus leucopus mice in North America. J Virol 1998;72:57–64.

Address for correspondence: Juan Arbiza, Sección Virología, Facultad de

Ciencias. Iguá 4225. CP 11400; fax: 5982-5258617; email:jarbiza@

fcien.edu.uy

852 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

RESEARCH

Serologic Evidence

of West Nile Virus

Infection in

Horses, Coahuila

State, Mexico

Bradley J. Blitvich,* Ildefonso Fernandez-Salas,†

Juan F. Contreras-Cordero,† Nicole L. Marlenee,*

Jose I. Gonzalez-Rojas,† Nicholas Komar,‡

Duane J. Gubler,‡ Charles H. Calisher,*

and Barry J. Beaty*

Serum samples were obtained from 24 horses in the

State of Coahuila, Mexico, in December 2002. Antibodies to

West Nile virus were detected by epitope-blocking enzyme-

linked immunosorbent assay and confirmed by plaque

reduction neutralization test in 15 (62.5%) horses. We

report the first West Nile virus activity in northern Mexico.

est Nile virus (WNV; family Flaviviridae, genus

Flavivirus) is a member of the Japanese encephalitis

virus complex, which also includes Japanese encephalitis

virus, Saint Louis encephalitis virus (SLEV), and Murray

Valley encephalitis virus (1). These viruses are maintained

in cycles between mosquitoes and birds (2). The principal

vectors for WNV are Culex species mosquitoes, and many

species of wild birds act as vertebrate hosts (3). Humans,

horses, and other mammals usually serve as dead-end

hosts. In humans and equines, WNV infection is usually

asymptomatic or characterized by a mild febrile illness,

although fatal meningoencephalitis or encephalitis may

occur (3–5). WNV has a broad geographic distribution,

recently including North America (4,5). The initial out-

break of WNV in North America was recognized in New

York City in August 1999. Since then, WNV’s geographic

range has increased. WNV activity has now been reported

in 44 states in the United States, the District of Columbia,

and 5 of the 10 Canadian provinces (6,7).

In anticipation of the possible emergence of WNV into

Mexico, we conducted equine infection surveillance in the

northeastern states of Mexico. Coahuila State is bordered

on the north by Texas. WNV activity has been detected in

204 (80%) of 254 Texas counties, including most counties

that border Coahuila State (8). Therefore, Coahuila State

was considered to be a likely point of incursion of WNV

into Mexico from the United States.

Case Study

We present data from a small equine serosurvey con-

ducted in Coahuila State in December 2002. A more exten-

sive equine serosurvey is currently under way and will be

described in detail elsewhere (B.J. Beaty, unpub. data). In

the present study, blood samples were taken from 24

domestic horses at study sites located in Ciudad Acuña,

Jiménez, and Saltillo (Figure). The Ciudad Acuña and

Jiménez sites are approximately 40 km apart, and both are

<15 km from the Texas border. Saltillo is located in the

southeast region of Coahuila State and is approximately

220 km from Texas. All study sites are privately owned

ranches.

The climatic conditions of the three study sites are sim-

ilar and can be described as hot, dry, and arid. The average

annual temperature ranges from 18°C to 22°C. The aver-

age rainfall is from 100 to 300 mm per year. The sites in

Ciudad Acuña and Jiménez are approximately 300 m

above sea level. The Saltillo site is situated at an elevation

of approximately 1500 m.

A horse from the Ciudad Acuña ranch died October 17,

3 days after being observed with neurologic signs. The

case was not reported immediately, and we were unable to

obtain a tissue specimen postmortem. On December 19,

with the assistance of a local veterinary practitioner, we

sampled 14 horses at this site, 5 of which had developed

neurologic disease in mid- to late October. Clinical symp-

toms included ataxia, weakness of limbs, trembling, and

anxiety. All five horses survived. The horses with clinical

signs were from 1 to 5 years of age; three were male, and

two were female. Ages and sexes of horses without clini-

cal symptoms were not documented. The veterinarian

reported a great abundance of mosquitoes in the area.

Another six horses were sampled in Jiménez and four in

Saltillo, none of which had signs of illness. According to

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 853

DISPATCHES

*Colorado State University, Fort Collins, Colorado, USA;

†Universidad Autonoma de Nuevo Leon, Nuevo Leon, Mexico;

and ‡Centers for Disease Control and Prevention, Fort Collins,

Colorado, USA

Figure. Geographic location of West Nile virus study sites in

Coahuila State, Mexico.

the owners, none of the horses had ever been outside the

State of Coahuila, and none of the horses had been vacci-

nated against WNV.

All serum samples were tested for antibodies to WNV

by epitope-blocking enzyme-linked immunosorbent assay

(ELISA). Blocking ELISAs were performed by using the

WNV-specific monoclonal antibody (MAb) 3.1112G, as

previously described (9). The ability of the Mexican horse

serum samples to block the binding of the MAb to WNV

antigen was compared to the blocking ability of horse

serum without antibody to WNV (Vector Laboratories,

Burlingame, CA). Data were expressed as relative percent-

ages by using the formula of Hall et al., (10). Previously,

we considered an inhibition value >

30% to indicate the

presence of viral antibodies (9). Recently, we have shown

that ELISAs performed with MAb 3.1112G detect WNV

antibodies in various vertebrate species, including horses

(9,11).

Fourteen serum samples were positive in blocking

ELISA that utilized MAb 3.1112G (Table). Serum from

another horse (H-16) inhibited the binding of MAb by

25%, which is close to the diagnostic criterion. Previously,

we observed that the nonspecific inhibition values for

serum samples from noninfected control birds ranged from

0% to 24.3% (9). Therefore, if we used a less stringent

threshold value of >25%, this serum could be considered

positive for WNV antibodies.

To validate the above assays, we tested serum samples

for neutralizing antibodies to WNV and SLEV by plaque

reduction neutralization assay (PRNT). Testing for neutral-

izing antibody to SLEV was important because this virus

is enzootic in the Americas and antibodies to WNV and

SLEV often cross-react. Furthermore, horses are suscepti-

ble to SLEV infection, although clinical manifestations

have not been reported (12). Viral isolates of WNV (strain

NY99-35261-11) and SLEV (strain TBH-28) were

obtained from the World Health Organization Center for

Arbovirus Reference and Research, maintained at the

Centers for Disease Control and Prevention, Division of

Vector-Borne Infectious Diseases, Fort Collins, CO.

PRNTs were performed by using Vero cells. Serum sam-

ples were tested by using a starting dilution of 1:20. Titers

were expressed as the reciprocal of serum dilutions reduc-

ing the number of plaques that were >90% (PRNT

854 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table. Summary of serologic data for horses in Coahuila State, Mexico

PRNT

titer

Horse

Study site

Clinical illness

% inhibition by

ELISA

WNV

SLEV

PRNT diagnosis

H-1

Ciudad Acuña

>320

—d

WNV

H-2

Ciudad Acuña

—

Negative

H-3

Ciudad Acuña

—

Negative

H-4

Ciudad Acuña

—

Negative

H-5

Ciudad Acuña

>320

WNV

H-6

Ciudad Acuña

>320

—

WNV

H-7

Ciudad Acuña

>320

—

WNV

H-8

Ciudad Acuña

Yes

>320

—

WNV

H-9

Ciudad Acuña

Yes

>320

—

WNV

H-10

Ciudad Acuña

Yes

>320

WNV

H-11

Ciudad Acuña

>320

—

WNV

H-12

Ciudad Acuña

>320

WNV

H-13

Ciudad Acuña

Yes

>320

—

WNV

H-14

Ciudad Acuña

Yes

>320

WNV

H-15

Jiménez

>320

—

WNV

H-16

Jiménez

—

WNV

H-17

Jiménez

—

Negative

H-18

Jiménez

>320

WNV

H-19

Jiménez

Flavivirus

H-20

Jiménez

—

WNV

H-21

Saltillo

—

Negative

H-22

Saltillo

—

Negative

H-23

Saltillo

—

Negative

H-24

Saltillo

—

Negative

ELISA, enzyme-linked immunosorbent assay; PRNT, plaque reduction neutralization test; WNV, West Nile virus; SLEV, Saint Louis encephalitis virus

Inhibition values >30% are considered significant.

Neutralizing antibodies to WNV in selected horse serum samples were confirmed at Centers for Disease Control and Prevention-Division of Viral and Bacterial

Infectious Diseases by PRNT.

—, <20.

Conclusions

Overall, PRNT and ELISA data were in concordance.

Fifteen (62.5%) horses were considered to be seropositive

for WNV by PRNT because the antibody titers for WNV

were greater than or equal to fourfold higher that the cor-

responding SLEV titer (Table). These 15 were the same

serum samples that had inhibition values of >25% by

ELISA. Of these, 11 horses were from Ciudad Acuña, and

4 were from Jiménez. Evidence for WNV infections was

detected in 5 (100%) of 5 horses with clinical symptoms,

and 10 (52.6%) of 19 horses without clinical symptoms.

Therefore, the rate of asymptomatic seropositivity was

high, with 10 (66.7%) of 15 WNV-infected horses showing

no signs of illness. Similarly, 21 (58.3%) of 36 WNV-

infected horses sampled during a serosurvey in New York

in 1999 showed no clinical signs (13). However, the sam-

ple population (n=24) in the present serosurvey was

notably small, and data from our large equine serosurvey

will provide a more reliable estimate of the asymptomatic

seropositivity rate.

We were unable to detect RNA in any horse serum by

reverse-transcription polymerase chain reaction with

WNV-specific primers (14). We plan to isolate and ampli-

fy WNV RNA sequences from tissue specimens obtained

from seropositive horses, as well as from birds, in future

studies.

We are currently conducting avian infection surveil-

lance in the State of Coahuila and the neighboring states of

Tamaulipas and Nuevo Leon. Preliminary evidence sug-

gests that several birds from a region in Nuevo Leon State

have antibodies to WNV (I. Fernandez-Salas, unpub. data).

The birds were trapped in February 2003, 2 months after

we obtained samples from the horses in Coahuila State.

However, equine cases often precede the detection of

seropositive birds. For example, an equine case was the

first indication of WNV activity in 29% (660/2,289) of the

United States counties to report virus activity in 2002 (6).

In summary, we have obtained serologic evidence for

antibodies to WNV in horses in the State of Coahuila,

Mexico. In the accompanying manuscript, we report the

detection of antibodies to WNV in horses in the State of

Yucatan (15). These two reports provide the first published

evidence of WNV activity in horses in Mexico. Antibodies

to WNV, or a closely related virus, were detected in a sin-

gle bovine during a serosurvey in Chiapas, Mexico, in

mid-2001 (16). WNV will probably become endemic in

Mexico, which is a major concern to public health author-

ities in the Americas. Our findings demonstrate the impor-

tance for continued WNV surveillance in Mexico.

Acknowledgments

We thank Nicolas Fernandez-Rangel, Oscar Lorenzana, and

Leopoldo Chavarria for collecting serum specimens.

This study was supported by grant U50 CCU820510 from

the Centers for Disease Control and Prevention and in part by

grant AI45430 from the National Institutes of Health.

Dr. Blitvich is a postdoctoral scientist in the Department of

Microbiology, Immunology and Pathology at Colorado State

University, Fort Collins, CO. His research interests include the

mechanisms of vector and host interactions in arbovirus trans-

mission cycles.

References

1. Heinz FX, Collett MS, Purcell RH, Gould EA, Howard CR,

Houghton M, et al. Family Flaviviridae. Virus taxonomy. In: Van

Regenmortel CM, Fauquet CM, Bishop DHL, Carstens E, Estes MK,

Lemon S, et al., editors. 7th Report of the International Committee for

the Taxonomy of Viruses. San Diego: Academic Press; 2000. p.

859–78.

2. Burke DS, Monath TP. Flaviviruses. In: Knipe DM, Howley PM, edi-

tors. Fields virology. 4th ed. Philadelphia (PA): Lippincott Williams

and Wilkins; 2001. p. 1043–126.

3. Hayes CG. West Nile fever. In: Monath TP, editor. Vol I. The

arboviruses: epidemiology and ecology. Boca Raton (FL): CRC Press

Inc; 1988. p. 59–88.

4. Campbell GL, Marfin AA, Lanciotti RS, Gubler DJ. West Nile virus.

Lancet Infect Dis 2002;2:519–29.

5. Roehrig JT, Layton M, Smith P, Campbell GL, Nasci R, Lanciotti RS.

The emergence of West Nile virus in North America: ecology, epi-

demiology, and surveillance. Curr Top Microbiol Immunol

2002;267:223–40.

6. Centers for Disease Control and Prevention. Provisional surveillance

summary of the West Nile Virus epidemic—United States,

January–November 2002. MMWR Morb Mortal Wkly Rep

2002;51:1129–33.

7. Health Canada. Population and Public Health Branch WNV surveil-

lance updates, December 4, 2002. Available from: URL:

http://www.hc-sc.gc.ca/pphb-dgspsp/wnv-vwn/mon_e.html#sitrep

8. Texas Health Department. West Nile virus in Texas. Available from:

URL: http://www.tdh.state.tx.us/zoonosis/diseases/Arboviral/west

Nile/default.asp

9. Blitvich BJ, Marlenee NL, Hall RA, Calisher CH, Bowen RA,

Roehrig JT, et al. Epitope-blocking enzyme-linked immunosorbent

assays for the detection of serum antibodies to West Nile virus in mul-

tiple avian species. J Clin Microbiol 2003;41:1041–7.

10. Hall RA, Broom AK, Hartnett AC, Howard MJ, Mackenzie JS.

Immunodominant epitopes on the NS1 protein of MVE and KUN

viruses serve as targets for a blocking ELISA to detect virus-specific

antibodies in sentinel animal serum. J Virol Methods

1995;51:201–10.

11. Blitvich BJ, Bowen RA, Marlenee NL, Hall RA, Bunning ML, Beaty

BJ. Epitope-Blocking enzyme-linked immunosorbent assays for the

detection of West Nile virus antibodies in domestic mammals. J Clin

Micorbiol. 2003;41:2676–9.

12. Tsai TF, Mitchell CJ. St. Louis encephalitis. In: Monath TP, editor.

Vol IV. The arboviruses: epidemiology and ecology. Boca Raton

(FL): CRC Press; 1988. p. 113–43.

13. United States Department of Agriculture Animal and Plant Health

Inspection Service (APHIS). Summary of West Nile virus in the

United States. Available from: URL: www.aphis.usda.gov/

vs/ep/WNV/summary.html. 2000.

14. Lanciotti RS, Kerst AJ, Nasci RS, Godsey MS, Mitchell CJ, Savage

HM, et al. Rapid detection of West Nile virus from human clinical

specimens, field-collected mosquitoes, and avian samples by a

TaqMan reverse transcriptase-PCR assay. J Clin Microbiol 2000;

38:4066–71.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 855

DISPATCHES

15. Loroño-Pino MA, Blitvich BJ, Farfán-Ale JA, Puerto FI, Blanco JM,

Marlenee NL, et al. Serologic evidence of West Nile virus infection

in horses, Yucatan State, Mexico. Emerg Infect Dis 2003;9:857–9.

16. Ulloa A, Langevin SA, Mendez-Sanchez JD, Arredondo-Jimenez JI,

Raetz JL, Powers AM, et al. Serologic survey of domestic animals for

zoonotic arbovirus infections in the Lacandón Forest Region of

Chiapas, Mexico. Vector Borne Zoonotic Dis. In press 2003.

Address for correspondence: Barry J. Beaty, Arthropod-borne and

Infectious Diseases Laboratory, Department of Microbiology,

Immunology and Pathology, College of Veterinary Medicine and

Biomedical Sciences, Colorado State University, Fort Collins, Colorado,

80523, USA; fax: 970-491-8323; email: [email protected]

856 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

The opinions expressed by authors contributing to this journal do

not necessarily reflect the opinions of the Centers for Disease

Control and Prevention or the institutions with which the authors

are affiliated.

Serologic Evidence

of West Nile Virus

Infection in

Horses, Yucatan

State, Mexico

María A. Loroño-Pino,*† Bradley J. Blitvich,*

José A. Farfán-Ale,† Fernando I. Puerto,†

José M. Blanco,‡ Nicole L. Marlenee,*

Elsy P. Rosado-Paredes,†

Julian E. García-Rejón,† Duane J. Gubler,§

Charles H. Calisher,* and Barry J. Beaty*

Serum samples were obtained from 252 horses in the

State of Yucatan, Mexico, from July to October 2002.

Antibodies to West Nile virus were detected by epitope-

blocking enzyme-linked immunosorbent assays in three

(1.2%) horses and confirmed by plaque reduction neutral-

ization test. We report the first West Nile virus activity in the

State of Yucatan.

est Nile virus (WNV) is a member of the Japanese

encephalitis virus complex within the genus

Flavivirus, family Flaviviridae (1). The virus is transmit-

ted in natural cycles mainly between mosquitoes and birds,

with humans and horses serving as incidental hosts (2).

WNV was first isolated in 1937 from the blood of a febrile

adult human in the West Nile District of Uganda (3). This

virus has since been reported in Africa, the Middle East,

Asia, southern Europe, Australia, and, more recently,

North America (4,5). The initial outbreak of WNV in

North America was recognized in New York City in

August 1999, with deaths reported in humans, horses, and

numerous species of birds. Since then, the geographic dis-

tribution of WNV in North America has greatly increased.

WNV activity has now been reported in 44 states and the

District of Columbia in the United States and in 5 of the 10

Canadian provinces (6,7).

In response to the incursion and rapid spread of WNV

in North America, we established equine and avian infec-

tion surveillance in Yucatan State, Mexico, in March 2000.

Yucatan State is a likely point of incursion of this virus into

Latin America because this area is a principal landfall for

many species of birds that migrate from the northeastern

and midwestern United States (8).

To determine whether WNV had already reached this

part of Mexico, we obtained blood samples from 252

domestic horses in 14 study sites from July to October

2002 (Table 1). The age distribution of the horses was 3

months to 25 years, and the mean age was 8.2 years. One

hundred and fifty-one horses were male, and 101 were

female. All study sites were on privately owned ranches,

where the horses were primarily used to perform heavy

labor and herd cattle. According to the owners, none of the

horses had ever been outside the State of Yucatan.

Furthermore, none of the horses had been vaccinated

against WNV.

The climate and topography of the study sites are simi-

lar. The climate can be described as tropical. The average

annual rainfall in each study site ranges from 600 to 1,100

mm, and the average annual temperature is 26°C. The

average elevation is approximately 17 m.

All serum samples were screened for antibodies to fla-

viviruses by hemagglutination inhibition (HI) assays and

epitope-blocking enzyme-linked immunosorbent assays

(ELISAs) at the Universidad Autonoma de Yucatan in

Merida. HI assays were performed by using Saint Louis

encephalitis virus (SLEV) antigen as previously described

(9). This antigen recognizes cross-reactive HI antibodies to

WNV and to other flaviviruses. To preclude nonspecific HI

reactions, samples were treated with kaolin, then adsorbed

with goose erythrocytes, according to standard methods

(9). Epitope-blocking ELISAs were performed by using

the flavivirus group-reactive monoclonal antibody (MAb),

6B6C-1, or the WNV-specific MAb, 3.1112G as previous-

ly described (10). The ability of the Mexican horse serum

samples to block the binding of MAbs to WNV antigen

was compared to the blocking ability of horse serum with-

out antibody to WNV (Vector Laboratories, Burlingame,

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 857

DISPATCHES

*Colorado State University, Fort Collins, Colorado, USA;

†Universidad Autonoma de Yucatan, Merida, Yucatan, Mexico;

‡Universidad Autonoma de Yucatan, Xmatkuil, Yucatan, Mexico;

and §Centers for Disease Control and Prevention, Fort Collins,

Colorado, USA

Table 1. Study sites and numbers of horses sampled per site,

State of Yucatan, Mexico

Study site

Global Positioning

System location

No. (%) of

horses bled

Acanceh

20° 48' 46" N, 89° 27' 14" W

8 (3.2)

Caucel

21° 00' 53" N, 89° 42' 25" W

1 (0.4)

Hobonil

20° 00' 54" N, 89° 01' 15" W

26 (10.3)

Hunucma

21° 00' 55" N, 89° 52' 28" W

7 (2.8)

Mani

20° 23' 11" N, 89° 23' 37" W

1 (0.4)

Merida

20° 58' 04" N, 89° 37' 18" W

63 (25.0)

Molas

20° 48' 57" N, 89° 37' 55" W

5 (2.0)

Progreso

21° 17' 04" N, 89° 39' 48" W

31 (12.3)

Sierra Papacal

21° 07' 16" N, 89° 43' 41" W

14 (5.6)

Timucuy

20° 48' 34" N, 89° 30' 51" W

5 (2.0)

Tixkokob

21° 00' 08" N, 89° 23' 37" W

15 (6.0)

Tizimin

21° 08' 32" N, 88° 09' 03" W

49 (19.4)

Uman

20° 49' 38" N, 89° 41' 08" W

26 (10.3)

Xbec

21° 14' 54" N, 88° 49' 29" W

1 (0.4)

Total

252 (100)

CA). Data were expressed as relative percentages, and

inhibition values >

30% were considered to indicate viral

antibodies. Recent studies in this laboratory have shown

that epitope-blocking ELISA provides a rapid and reliable

serologic technique for the detection of WNV antibodies in

various vertebrate species, including horses (10,11).

Six horses had evidence of flavivirus infection by HI

assay or ELISA (Table 2). Serum samples from three of

these horses (H-117, H-126, and H-252) were positive in

the ELISA that used the WNV-specific MAb. H-117 (7-

year-old stallion) and H-126 (2-year-old stallion) were

both sampled at the Tizimin study site. Neither horse

showed signs of illness at the time of serum collection or

during the 7 months that followed. Furthermore, neither

horse had a history of WNV-like illness. H-252 was a 3-

year-old stallion from Caucel that exhibited neurologic and

muscular symptoms at the time of sampling; it was eutha-

nized several hours later. We were not able to obtain tissue

specimens from this horse postmortem. Of the 252 horses

sampled, the only other horse to exhibit signs of clinical

illness was H-60, which had signs consistent with gastroin-

testinal illness.

Serum samples positive for flavivirus antibodies by HI

assay or ELISA were tested by plaque reduction neutral-

ization assay (PRNT) to identify the infecting virus.

PRNTs were conducted in the BSL-3 facilities at Colorado

State University. Serum sample results shown to be nega-

tive by HI assay and ELISA were not tested. PRNTs were

done by using WNV (strain NY99-35261-11), SLEV

(strain TBH-28), Ilhéus virus (ILHV, original strain), and

Bussuquara virus (BSQV, strain BeAn-4073). Virus stocks

were obtained from the World Health Organization Center

for Arbovirus Reference and Research, maintained at the

Centers for Disease Control and Prevention, Division of

Vector-Borne Infectious Diseases, Fort Collins, CO. We

tested serum samples for neutralizing antibodies to SLEV

because the virus is enzootic in the Americas and because

antibodies to WNV and SLEV often cross-react.

Furthermore, horses are susceptible to natural SLEV infec-

tions, although clinical manifestations have not been

reported (12). ILHV and BSQV are also present in the

Americas, although neither virus is known to naturally

infect horses (2). PRNTs were performed by using Vero

cells. Serum samples were tested by using a starting dilu-

tion of 1:20. Titers were expressed as the reciprocal of

serum dilutions yielding >90% reduction in the number of

plaques (PRNT

Neutralizing antibodies to WNV were detected in three

horses (Table 3). The PRNT-positive horses were H-117,

H-126, and H-252, which exhibited PRNT

antibody titers

of 320, >2,560, and 160, respectively. The SLEV, ILHV,

and BSQV antibody titers of the three horses were all <20.

Therefore, we considered H-117, H-126, and H-252 to be

seropositive for WNV because the PRNT

antibody titers

for WNV were more than fourfold higher than the other

flaviviruses tested. Overall, the PRNT and ELISA data

were in concordance; all serum samples that contained

neutralizing antibodies to WNV were positive in the assay

that used MAb 3.1112G (Tables 2 and 3). However, H-252

was negative in the assay that used MAb 6B6C-1, although

the percent inhibition value was close to the diagnostic cri-

terion. The three other horses (H-60, H-134, and H-141)

that were positive for flavivirus antibodies by HI assay or

ELISA did not have neutralizing antibodies to WNV. H-60

had a low SLEV PRNT

titer, suggesting it had been

infected with SLEV or a closely related virus. H-134

exhibited a HI titer of 10 but was negative by the other

serologic tests, suggesting that the HI antigen had reacted

nonspecifically. H-141 was positive by HI assay and

ELISA but had no neutralizing antibodies to any flavivirus

tested; thus, the identity of the infecting virus was not

determined.

We obtained serologic evidence for antibodies to WNV

in Yucatan State, Mexico. The mode of entry of this virus

into Yucatan State is not known; however, the virus may

have been brought in by birds migrating from the north.

858 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table 2. Summary of horses with HI assay or epitope-blocking ELISA antibodies to flaviviruses

% inhibition by

blocking ELISA

Horse

Sampling date

Study site

Age (y)

Sex

Clinical symptoms

Outcome

HI assay

titer

3.1112G

6B6C-1

H-60

July 2, 2002

Merida

Male

Gastrointestinal

(recurrent colic)

Survived

H-117

July 5, 2002

Tizimin

Male

None

Survived

H-126

July 5, 2002

Tizimin

Male

None

Survived

H-134

July 5, 2002

Tizimin

Female

None

Survived

H-141

July 5, 2002

Tizimin

Male

None

Survived

H-252

Oct. 15, 2002

Caucel

Male

Neurologic and

muscular symptoms

Euthanized

HI, hemagglutination-inhibition; ELISA, enzyme-linked immunosorbent assay.

Inhibition values >30% are considered significant.

MAb 3.1112G is WNV-specific.

MAb 6B6C-1 is flavivirus group–reactive.

We have also detected antibodies to WNV in certain

species of migratory birds, which supports this hypothesis.

Data from the avian surveillance studies conducted in

Yucatan State will be described separately (J.A. Farfán-

Ale, unpub. data). We plan to isolate and amplify viral

sequences from migratory and resident birds, as well as

from specimens from other seropositive animals, to deter-

mine the origin of the WNV strain in Yucatan State. We

also provide serologic evidence for WNV infection in

horses in Coahuila State (13). These two reports provide

the first published evidence of WNV activity in horses in

Mexico. Neutralizing antibodies to WNV have also been

detected in a bovine in Chiapas, Mexico, in mid-2001,

indicating that the animal had been infected with WNV or

a closely related virus (14). WNV may become endemic in

this country, which demonstrates the importance for con-

tinued WNV surveillance in Mexico, and elsewhere in the

south.

Acknowledgments

We thank Eduardo Sierra-Lira and Edwin J. Gutierrez-Ruiz

for collecting the serum specimens and Luis F. Flores-Flores and

Richard Bowen for technical assistance.

This study was supported by grant U50 CCU820510 from

the Centers for Disease Control and Prevention and in part by

grant AI45430 from the National Institutes of Health.

Ms. Loroño-Pino is a Ph.D. student in the Department of

Microbiology, Immunology and Pathology at Colorado State

University, Fort Collins, CO. Previously, she was the chief of the

Laboratorio de Arbovirologia, Centro de Investigaciones

Regionales Dr. Hideyo Noguchi, Universidad Autonoma de

Yucatan, Mexico. Her research interests include the mechanisms

of West Nile virus and dengue virus pathogenesis.

References

1.Heinz FX, Collett MS, Purcell RH, Gould EA, Howard CR,

Houghton M, et al. Family Flaviviridae. Virus taxonomy. In: Van

Regenmortel CM, Fauquet CM, Bishop DHL, Carstens E, Estes MK,

Lemon S, et al., editors. 7th report of the International Committee for

the Taxonomy of Viruses. San Diego (CA): Academic Press; 2000. p.

859–78.

2. Burke DS, Monath TP. Flaviviruses. In: Knipe DM, Howley PM, edi-

tors. Fields virology. 4th ed. Philadelphia (PA): Lippincott Williams

and Wilkins; 2001. p. 1043–126.

3. Smithburn KC, Hughes TP, Burke AW, Paul JH. A neurotropic virus

isolated from the blood of a native of Uganda. Am J Trop Med Hyg

1940;20:471–92.

4. Campbell GL, Marfin AA, Lanciotti RS, Gubler DJ. West Nile virus.

Lancet Infect Dis 2002;2:519–29.

5. Roehrig JT, Layton M, Smith P, Campbell GL, Nasci R, Lanciotti RS.

The emergence of West Nile virus in North America: ecology, epi-

demiology, and surveillance. Curr Top Microbiol Immunol

2002;267:223–40.

6. Centers for Disease Control and Prevention. Provisional surveillance

summary of the West Nile Virus epidemic—United States,

January–November 2002. MMWR Morb Mortal Wkly Rep

2002;51:1129–33.

7. Health Canada. Population and Public Health Branch WNV surveil-

lance updates, December 4, 2002. Available from: URL:

http://www.hc-sc.gc.ca/pphb-dgspsp/wnv-vwn/mon_e.html#sitrep

8. Howell SNG, Webb S. A guide to the birds of Mexico and northern

Central America. Oxford: Oxford University Press; 1995.

9. Calisher CH, Beaty BJ, Chandler LJ. Arboviruses. In: Lennette E,

editor. Laboratory diagnosis of viral infections. 3rd ed. New York:

Marcel Dekker Inc; 1999. p. 305–32.

10. Blitvich BJ, Marlenee NL, Hall RA, Calisher CH, Bowen RA,

Roehrig JT, et al. Epitope-blocking enzyme-linked immunosorbent

assays for the detection of serum antibodies to West Nile virus in mul-

tiple avian species. J Clin Microbiol 2003;41:1041–7.

11. Blitvich BJ, Bowen RA, Marlenee NL, Hall RA, Bunning ML, Beaty

BJ. Epitope-blocking enzyme-linked immunosorbent assays for the

detection of West Nile virus antibodies in domestic mammals. J Clin

Microbiol. 2003;41:2676–9.

12. Tsai TF, Mitchell CJ. St. Louis encephalitis. In: Monath TP, editor.

Vol. IV, The arboviruses: epidemiology and ecology. Boca Raton

(FL): CRC Press; 1988. p. 113–43.

13. Blitvich BJ, Fernandez-Salas I, Contreras-Cordero JF, Marlenee NL,

Gonzalez-Rojas JI, Komar N, et al. Serologic evidence of West Nile

virus infection in horses, Coahuila State, Mexico. Emerg Infect Dis

2003;9;853–6.

14. Ulloa A, Langevin SA, Mendez-Sanchez JD, Arredondo-Jimenez JI,

Raetz JL, Powers AM, et al. Serologic survey of domestic animals for

zoonotic arbovirus infections in the Lacandón Forest region of

Chiapas, Mexico. Vector Borne Zoonotic Dis. In press 2003.

Address for correspondence: Barry J. Beaty, Arthropod-borne and

Infectious Diseases Laboratory, Department of Microbiology,

Immunology and Pathology, College of Veterinary Medicine and

Biomedical Sciences, Colorado State University, Fort Collins, CO 80523,

USA; fax: 970-491-8323; email: [email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 859

DISPATCHES

Table 3. Neutralizing antibody titers to West Nile, Saint Louis

encephalitis, Ilhéus, and Bussuquara viruses in serum samples

from six horses

PRNT

titer

Horse

WNV

SLEV

ILHV

BSQV

H-60

–

H-117

320

–

H-126

>2,560

–

H-134

–

H-141

–

H-252

160

–

WNV, West Nile virus; SLEV, Saint Louis encephalitis virus; ILHV,

Ilhéus virus; BSQV, Bussuquara virus; PRNT, plaque reduction

neutralization test; –, <20.

PRNT

titer: 20.

PRNT

titer: 40.

All material published in Emerging Infectious Diseases is in the

public domain and may be used and reprinted without special

permission; proper citation, however, is appreciated.

Serologic Evidence

of West Nile Virus

Transmission,

Jamaica, West

Indies

Alan P. Dupuis II,* Peter P. Marra,†

and Laura D. Kramer*

In spring 2002, an intensive avian serosurvey was ini-

tiated in Jamaica, Puerto Rico, and Mexico. We collected

>1,600 specimens from resident and nonresident neotropi-

cal migratory birds before their northerly migrations. Plaque

reduction neutralization test results indicated specific neu-

tralizing antibodies to West Nile virus in 11 resident species

from Jamaica.

est Nile virus (WNV) is maintained in nature

between birds and Culex species mosquitoes (1,2).

Unlike other viruses maintained in bird and mosquito

transmission cycles (for example, St. Louis encephalitis,

western equine encephalomyelitis, and eastern equine

encephalomyelitis), the WNV strain responsible for the

current epizootic in the Western Hemisphere is associated

with a high number of avian deaths (3). Although crows

and other corvids appear to be especially susceptible to

disease (4), WNV has been documented in >190 bird

species, including neotropical migratory species and exot-

ic zoo specimens (5).

Birds have been implicated in spreading WNV during

migratory events in Europe, Asia, Africa, and the Middle

East (6–9). Because of the apparent ease of infecting a

multitude of avian hosts, WNV can potentially be intro-

duced during annual migratory events in the Western

Hemisphere. Predicting the incursion of WNV into the

tropics is complicated by our incomplete knowledge

regarding geographic connectivity in populations of

migratory birds between winter and summer. We do not

know where most species of migratory birds from North

America spend their winter, where wintering birds spend

their summer, or the routes they use while in transit.

Evidence from mark-and-recapture efforts suggests that

birds from the northeastern United States tend to winter in

the southeastern United States and Greater Antilles (e.g.,

Puerto Rico, Jamaica) whereas birds from the western

United States migrate to Mexico and Central America

(10–12). In response to the potential introduction of WNV

in tropical America during the fall migrations, we estab-

lished a network of monitoring sites on the overwintering

grounds of neotropical migratory birds in Jamaica, Puerto

Rico, and Mexico.

The Study

The primary goal of the monitoring system was to

obtain a large number of blood specimens from birds

belonging to as many species as possible across the

Caribbean. At seven study sites in Jamaica, Puerto Rico,

and the Yucatan Peninsula of Mexico, we erected 15 mist

nets (12-m) daily for 3 to 4 weeks during January through

March 2002. We collected 1,619 avian blood specimens,

which represented 98 species, 25 families, and eight

orders. In Jamaica, 542 samples were collected from

Westmoreland, Manchester, and St. Catherine Parishes;

649 samples were collected at Roosevelt Roads Naval

Station in Puerto Rico; and 430 samples were collected

from the states of Yucatan and Campeche in Mexico (Table

1). At the time of capture, all migratory birds were banded

with an aluminum U.S. Fish and Wildlife Service band and

3–5 breast feathers were removed for isotope analysis.

Resident birds had outer rectrices cut. All birds were eval-

uated for age and gender if possible, bled with microcapil-

lary tubes from the brachial vein, and released. Blood was

added immediately to BA-1 medium, consisting of M199

medium with Hank’s salts, 1% bovine albumin, TRIS base

(tris [hydroxymethyl] aminomethane), sodium bicarbon-

ate, 20% fetal bovine serum (FBS), and antibiotics. The

samples were placed in a cooler on ice packs until storage

in a –20°C freezer. Samples were sent on dry ice or hand-

carried on ice packs to the Arbovirus Laboratories,

Wadsworth Center, New York State Department of Health

for serologic analysis and virus isolation attempts.

Specimens were screened at 1:100 for antibodies

against flaviviruses by using an indirect enzyme-linked

immunosorbent assay (ELISA) (13). Samples with a P/N

ratio >2.0 were tested further by a plaque reduction neu-

tralization test (PRNT) for Ilhéus virus (ILHV), St. Louis

encephalitis virus (SLEV), and WNV, as described (14).

The particular virus strains used for the PRNTs were ILHV

(original), SLEV 59268 Parton, and WNV (3100365), an

isolate from a pool of Culex sp. mosquitoes collected in

Staten Island, New York. The indirect ELISA was chosen

to screen the samples in order to take advantage of its abil-

ity to detect antibodies against a wide range of flavivirus-

es. PRNT was used as a confirmatory assay to differentiate

among recognized flaviviruses (15,16). Briefly, serial dilu-

tions of test samples were mixed with an equal amount of

virus suspension containing 200 PFU/0.1 mL and incubat-

ed at 37°C for 1 h. We then added 0.05 mL of each virus-

860 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

*New York State Department of Health, Slingerlands, New York,

USA; and †Smithsonian Environmental Research Center,

Edgewater, Maryland, USA

diluted blood sample onto 1 well of a 12-well tissue culture

plate containing confluent monolayers of African green

monkey kidney cells (Vero). The plate was incubated for 1

h at 37°C, after which an agarose overlay was added and

incubation was continued. When virus plaques became

visible, we added a second overlay containing neutral red

and counted plaques. The antibody titer reported is the

dilution of serum that inhibited 90% of the test virus inocu-

lum. For virus isolation attempts using confluent Vero cell

monolayers, 0.1 mL of each serum specimen was added

onto one well of a six-well tissue culture plate, incubated

for 1 h, rinsed with phosphate-buffered saline, and then the

cells were refed with minimum essential medium contain-

ing 2% FBS. Cells were monitored once a day for 5 days

for cytopathic effect. Cell cultures showing any abnormal

cell morphology were then blind passaged after 5 days.

ELISA results indicated 34 of 1,413 serum specimens

tested from the three study sites contained immunoglobu-

lin G antibody against a flavivirus. Of the 34 reactive sam-

ples representing 20 bird species, 27 were collected in

Jamaica (26 residents, 1 migrant), 5 in Puerto Rico (3 res-

idents, 2 migrants), and 2 in Mexico (1 resident, 1

migrant). PRNTs on the Jamaican bird samples indicated

18 WNV infections, 3 SLEV infections, 5 undetermined

flavivirus infections (positive results for two or more

viruses without a fourfold difference in antibody titer); one

additional reactive serum sample was negative for the

three viruses tested. Results on the serum samples collect-

ed in Puerto Rico indicated one WNV infection in a migra-

tory bird and one SLEV infection in a resident bird; three

additional reactive serum samples were negative for all

three viruses tested. In Mexico, we found evidence of

WNV infection in one migrant bird and SLEV infection in

one resident bird (Table 2). Virus isolation attempts were

negative for all 1,603 specimens tested (16 samples

destroyed). Negative isolation results were not entirely

unexpected, considering that birds are viremic for a short

period of time (17) and maintaining a proper cold chain

(i.e., temperatures) to preserve virus is difficult when

working in the tropics.

Conclusions

We detected neutralizing antibodies to WNV in resident

birds from two parishes in Jamaica. This detection marks

the earliest evidence of WNV introduction into the

neotropics; WNV antibodies have been demonstrated in

birds and horses in Mexico (late 2002, spring 2003)

(18,19) and detected in resident birds from the Dominican

Republic (spring 2003) (20). This evidence of WNV in the

neotropics may be an important development in the spread

of the virus. The tropics provide all the necessary compo-

nents (i.e., high temperatures, dense avian populations, and

large numbers of Culex sp. mosquitoes) to maintain an

enzootic focus. Furthermore, the climates of Mexico and

the Caribbean are suitable for year-round transmission of

the virus. No dead birds have been reported in Jamaica, but

surveillance activity there is less intensive than in the

United States; the study sites, being rural in nature, are not

conducive to observing dead birds. Another contributing

factor to the lack of reports of dead birds may be the rapid

decomposition of dead birds as a result of the heat, humid-

ity, and detritivore foragers, such as ants.

Arbovirus activity, particularly of flaviviruses, is well

documented in the Caribbean and Mexico. Dengue and

yellow fever viruses are recurring public health threats in

these areas. SLEV is endemic to Mexico and has been iso-

lated from mosquitoes and one Northern mockingbird

(Mimus polyglottos) nestling in Jamaica (21). This disease

is still active in the region, and its known range may have

expanded into Puerto Rico, considering the one seroposi-

tive Caribbean elaenia (Elaenia martinica) sampled during

this study, although the antibody may have resulted from

infection with yet another flavivirus. Neutralizing antibod-

ies to WNV in migratory birds collected in Mexico and

Puerto Rico, coupled with the apparent absence of anti-

body to WNV in the resident bird population, indicate that

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 861

DISPATCHES

Table 1. Number of birds collected at seven field sites in the Caribbean

Field site

Migratory birds

Resident birds

Total

Flavivirus positives (%)

Westmoreland Parish, Jamaica

156

232

388

18 (4.6)

Manchester Parish (site 1), Jamaica

1 (4.4)

Manchester Parish (site 2), Jamaica

107

4 (3.7)

St. Catherine Parish, Jamaica

4 (16.7)

Yucatan, Mexico

102

172

2 (1.2)

Campeche, Mexico

114

144

258

0 (0.0)

Mexico totals

184

246

430

2 (0.5)

Puerto Rico (one collection site) totals

391

256

647

5 (1.1)

Jamaica totals

194

348

542

27 (5.0)

Totals

769

850

1,619

34 (2.4)

Based on screening enzyme-linked immunosorbent assay (ELISA) results.

Results based on 441 samples suitable for ELISA (206 samples were blood clots only).

Results based on 1,413 samples tested by ELISA.

infection likely occurred in an enzootic area of the United

States, but this observation shows that individual birds

from at least three species of neotropical migratory birds

have survived WNV infection and may serve as hosts for

spreading the virus. The results from this study suggest

that WNV now appears to be established in Jamaica, on the

basis of the neutralizing antibodies to WNV found in the

resident bird population.

Acknowledgments

We thank Brian Gibbons, Karin Roux, Javier Salgado-Ortiz,

Joe Smith, and Jenn Barg for collecting field samples and Robert

Reitsma, Kim Kent, and Donna Young for tremendous logistical

and laboratory support.

This work was supported by Grant Agreement Number

U50/CCU320544-01 from the Centers for Disease Control and

Prevention (CDC) and the National Science Foundation (NSF),

Grant Number DEB 0089565. Its contents are solely the respon-

sibility of the authors and do not represent the official views of

CDC or NSF.

Mr. Dupuis is an assistant research scientist in the Arbovirus

Laboratories, Wadsworth Center, New York State Department of

Health. His research focuses on the role of birds in the mainte-

nance of arbovirus transmission cycles and dispersal of the virus.

References

1. Hayes CG. West Nile fever. In: Monath TP, editor. The arboviruses:

epidemiology and ecology. Vol. V. Boca Raton (FL): CRC Press, Inc.;

1989. p. 59–88.

862 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table 2. Ninety percent plaque reduction neutralization titers to West Nile virus, St. Louis encephalitis virus, and Ilhéus virus in

enzyme-linked immunosorbent assay reactive bird specimens

Field site

Species

ILHV

SLEV

WNV

Interpretation

White-chinned thrush

(Turdus aurantius)

640

WNV

White-chinned thrush (T. aurantius)

<40

320

WNV

Jamaican elaenia (Myiopagis cotta)

<40

160

640

WNV

Bananaquit (Coereba flaveola)

<40

160

Flavivirus

Loggerhead kingbird (Tyrannus caudifasciatus)

<40

320

WNV

Bananaquit (C. flaveola)

<40

160

WNV

White-chinned thrush (Turdus aurantius)

<40

>1,280

WNV

Bananaquit (C. flaveola)

<40

WNV

Northern mockingbird (Mimus polyglottos)

<40

320

WNV

Caribbean dove (Leptotila jamaicensis)

<40

>1,280

SLEV

Jamaican elaenia (Myiopagis cotta)

<40

160

WNV

Bananaquit (C. flaveola)

<40

Flavivirus

White-chinned thrush (Turdus aurantius)

<40

640

WNV

Jamaican elaenia (Myiopagis cotta)

<40

SLEV

Common ground-dove (Columbina passerina)

<40

160

WNV

Caribbean dove (Leptotila jamaicensis)

<40

Negative

Black-faced grassquit (Tiaris bicolor)

<40

320

WNV

Westmoreland Parish, Jamaica

Jamaican vireo (Vireo modestus)

<40

WNV

Manchester Parish (site 1), Jamaica

Caribbean dove (Leptotila jamaicensis)

<40

WNV

Black-faced grassquit (Tiaris bicolor)

<40

160

WNV

Jamaican oriole (Icterus leucopteryx)

<40

160

>1,280

WNV

White-eyed thrush (Turdus jamaicensis)

<40

320

WNV

Manchester Parish (site 2), Jamaica

Orangequit (Euneornis campestris)

<40

Flavivirus

Greater Antillean grackle (Quiscalus niger)

160

Flavivirus

Loggerhead kingbird (Tyrannus caudifasciatus)

320

Flavivirus

Golden warbler

(Dendroica petechia sp. )

<40

SLEV

St. Catherine Parish, Jamaica

Northern waterthrush (Seiurus noveboracensis)

<40

320

WNV

Caribbean dove (Leptotila jamaicensis)

<40

>1,280

SLEV

Yucatan State, Mexico

Yellow warbler (Dendroica petechia)

<40

320

WNV

Prairie warbler (Dendroica discolor)

<40

Negative

Golden warbler

(Dendroica petechia sp. )

<40

Negative

Pearly-eyed thrasher (Margarops fuscatus)

<40

Negative

Caribbean elaenia (Elaenia martinica)

<40

SLEV

Roosevelt Roads Naval Station,

Puerto Rico

Black and white warbler (Mniotilta varia)

<40

>1,280

WNV

ILHV, Ilhéus virus; SLEV, St. Louis encephalitis virus; WNV, West Nile virus.

Resident bird.

Resident yellow warbler subspecies.

Migrant bird.

2. Kramer LD, Bernard KA. West Nile virus in the western hemisphere.

Curr Opin Infect Dis 2001;14:519–25.

3. Bernard KA, Maffei JG, Jones SA, Kauffman EB, Ebel GD, Dupuis

AP II, et al. West Nile virus infection in birds and mosquitoes, New

York State, 2000. Emerg Infect Dis 2001;7:679–85.

4. Bernard KA, Kramer LD. West Nile virus activity in the United

States, 2001. Viral Immunol 2001;14:319–38.

5. U.S. Department of the Interior, U.S. Geological Survey, National

Wildlife Health Center. Species found positive for WNV in surveil-

lance efforts. [Cited April 5, 2003.] Available from: URL:

http://www.nwhc.usgs.gov/research/west_nile/wnvaffected.html

6. Malkinson M, Banet C, Weisman Y, Pokamunski S, King R, Drouet

MT, et al. Introduction of West Nile virus in the Middle East by

migrating white storks. Emerg Infect Dis 2002;8:392–7.

7. Malkinson M, Banet C. The role of birds in the ecology of West Nile

virus in Europe and Africa. Curr Top Microbiol Immunol

2002;267:309–22.

8. Ernek E, Kozuch O, Nosek J, Teplan J, Folk C. Arboviruses in birds

captured in Slovakia. J Hyg Epidemiol Microbiol Immunol

1977;21:353–9.

9. Rappole JH, Derrickson SR, Hubalek Z. Migratory birds and spread

of West Nile virus in the Western Hemisphere. Emerg Infect Dis

2000;6:319–28.

10. Johnson LS, Wise J. Wintering grounds of North American house

wrens as revealed by band recoveries. J Field Ornithol

2000;71:501–5.

11. Webster MS, Marra PP, Haig SM, Bensch S, Holmes RT. Links

between worlds: unraveling migratory connectivity. Trends in

Ecology and Evolution 2002;17:76–82.

12. Elphick J. The atlas of bird migration. New York: Random House;

1995.

13. Ebel GD, Dupuis AP, Nicholas D, Young D, Maffei J, Kramer LD.

Detection by enzyme-linked immunosorbent assay of antibodies to

West Nile virus in birds. Emerg Infect Dis 2002;8:979–82.

14. Lindsey HS, Calisher CH, Matthews JH. Serum dilution neutraliza-

tion test for California group virus identification and serology. J Clin

Microbiol 1976;4:503–10.

15. DeMadrid AT, Porterfield JS. The flaviviruses (group B arboviruses):

a cross-neutralization study. J Gen Virol 1974;23:91–6.

16. Calisher CH, Karabatsos N, Dalrymple JM, Shope RE, Porterfield JS,

Westaway EG, et al. Antigenic relationships between flaviviruses as

determined by cross-neutralization tests with polyclonal antisera. J

Gen Virol 1989;70:37–43.

17. Komar N, Langevin S, Hinten S, Nemeth N, Edwards E, Hettler D, et

al. Experimental infection of North American birds with the New

York 1999 strain of West Nile virus. Emerg Infect Dis

2003;9:311–22.

18. West Nile virus, equine, 2002—Mexico (Tamaulipas). ProMed mail

archive no. 20030325.0739. [cited 25 March 2003]. Available from:

URL: http://www.promedmail.org

19. West Nile virus, birds—Mexico (North). ProMed mail archive no.:

20030315.0640. [Cited 15 March 2003]. Available from: URL:

http://www.promedmail.org

20. West Nile virus, birds—Dominican Republic. ProMed mail archive

no.: 20030315.0645. [Cited 15 March 2003]. Available from: URL:

http://www.promedmail.org

21. Belle EA, King SD, Griffiths BB, Grant LS. Epidemiological inves-

tigation for arboviruses in Jamaica, West Indies. Am J Trop Med Hyg

1980;29:667–75.

Address for correspondence: Alan P. Dupuis II, Griffin Laboratory, 5668

State Farm Rd., Slingerlands, NY 12159, USA; fax: (518) 869-4530;

email: [email protected].us

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 863

DISPATCHES

The print journal is available at no charge to public health professionals

YES, I would like to receive Emerging Infectious Diseases.

Please print your name and business

address in the box and return by fax to

404-371-5449 or mail to

EID Editor

CDC/NCID/MS D61

1600 Clifton Road, NE

Atlanta, GA 30333

Moving? Please give us your new address (in the box) and print the number of your old

mailing label here_________________________________________

Full text free online at

www.cdc.gov/eid

Sulfa Resistance

and Dihydropteroate

Synthase Mutants

in Recurrent

Pneumocystis

carinii Pneumonia

Aimable Nahimana,*

Meja Rabodonirina,†

Jannik Helweg-Larsen,‡ Isabelle Meneau,*

Patrick Francioli,* Jacques Bille,*

and Philippe M. Hauser*

Failure of sulfa or sulfone prophylaxis is associated

with mutations in Pneumocystis carinii gene coding for

dihydropteroate synthase (DHPS). The DHPS genotype

was analyzed in AIDS patients who had two separate

episodes of P. carinii pneumonia. The results suggest that

DHPS mutations can be selected de novo within patients

by the pressure of a sulfa or sulfone drug.

o-trimoxazole, the antifolate drug combination of

trimethoprim and sulfamethoxazole, is the drug of

choice for the prophylaxis and treatment of Pneumocystis

carinii pneumonia (PCP), a life-threatening disease in

immunosuppressed patients. Trimethoprim is an inhibitor

of dihydrofolate reductase, whereas sulfamethoxazole

inhibits dihydropteroate synthase (DHPS). The antipneu-

mocystis activity is believed to be due mainly to sul-

famethoxazole (1). Dapsone is a sulfone drug, also fre-

quently used, that targets DHPS. Widespread use of sulfa

and sulfone drugs to prevent and treat PCP in recent years

has correlated with an increase of the prevalence of muta-

tions in the gene coding for DHPS (2,3). The most frequent

DHPS mutations occur at nucleotide positions 165 and

171, which lead to an amino acid change at positions 55

(Thr to Ala) and 57 (Pro to Ser). These mutations are locat-

ed in the sulfa-binding site and may appear as either a sin-

gle or double mutation in the same isolate. Similar muta-

tions in other microbial pathogens confer sulfa resistance

(4,5). In P. carinii, DHPS mutations are associated with

failure of sulfa or sulfone prophylaxis (1,6) and decreased

survival of the patient at 3 months after PCP (2). However,

patients harboring P. carinii types with DHPS mutations

are most often successfully treated with high-dose co-tri-

moxazole (6). Because a standardized culture system for P.

carinii does not exist, the level of sulfa resistance con-

ferred by these mutations cannot be determined with in

vitro susceptibility tests. A key issue is whether the recent

emergence of DHPS mutations is a result of P. carinii

transmission between patients or arises from selection

within patients by the pressure of a sulfa or sulfone drug,

two possibilities that are not mutually exclusive. To inves-

tigate the latter possibility, we analyzed patients who had

had two separate episodes of PCP.

The Study

P. carinii DNA was extracted from bronchoalveolar

lavage specimens by using the Qiamp Blood Kit (QIA-

GEN GmbH, Hilden, Germany). Bronchoalveolar lavage

specimens from 13 patients with recurrent PCP episodes

were collected from four European hospitals (Lyon,

France; Copenhagen, Denmark; Lausanne, Switzerland;

and La Chaux-de-Fonds, Switzerland). To determine the

prevalence of the different P. carinii molecular types, we

analyzed bronchoalveolar lavage specimens from 310 PCP

patients from two Swiss hospitals (Lausanne, 111 patients;

Zurich, 64 patients) and Lyon’s hospital (135 patients).

Specific information on demographic and clinical charac-

teristics, chemoprophylaxis, and treatment regimens were

obtained from the medical charts. P. carinii infecting

humans (now named P. jiroveci [7]) was typed by using the

multilocus method developed in our laboratory as previ-

ously described (8–10). In this method, four variable

regions of the P. carinii genome are amplified by poly-

merase chain reaction (PCR), followed by the detection of

polymorphisms using single-strand conformation poly-

morphism (SSCP). A P. carinii type is defined by a combi-

nation of four alleles corresponding to the four genomic

regions. If a specimen harbored two alleles of one or more

of the four genomic regions, the patient was considered to

be co-infected with two or more P. carinii types (9). This

typing system has been validated and the stability of its

markers assessed; its index of discriminatory power has

been estimated to be 0.93 (10). The full length of the

DHPS gene was amplified by PCR as described previous-

ly (1). PCR products (765 bp) were cloned, and both

strands were sequenced (5 clones per sample). The five

clones had identical sequences for all samples, except for

those from patients 3 and 4, which contained a mixture of

DHPS sequences.

864 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

*Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland;

†Université Claude-Bernard, Lyon, France; and ‡Hvidovre

Hospital, Copenhagen, Denmark

Aimable Nahimana contributed to the design of the study, ana-

lyzed the samples by polymerase chain reaction and DNA

sequencing, and wrote the draft of the manuscript. Meja

Rabodonirina and Jannik Helweg-Larsen reviewed medical charts

and provided bronchoalveolar lavage specimens. All authors con-

tributed to the analysis of data and writing of the paper. Philippe M.

Hauser initiated and supervised all aspects of the study.

Thirteen patients with two separate PCP episodes

were analyzed (Table). All patients had recovered

between episodes. The intervals between the episodes

ranged from 4 to 25 months. All patients had AIDS and

all, except patients 8 and 9, were men, with a median age

of 35 (range 23–51) and median CD4 cell count of 9.5

cells/µL (range 0–98). Some patients were co-infected

with two different P. carinii types, as shown by PCR-

SSCP multilocus typing method (patients 4, 5, 8, 11, and

13) or DHPS genotyping (patients 3 and 4). In seven

(54%) patients (patients 1–7), the same PCR-SSCP type

was observed in both episodes of PCP; six (46%) patients

(patients 8–13) had different types in the first and second

episodes. This rate of genotype switch is similar to that

reported in previous studies, in which such a change was

observed in approximately half of recurrent episodes

(11–14). The importance of a genotype switch remains

uncertain. Indeed, the switch might be due to a de novo

infection or to the reactivation of a genotype not detect-

ed in the first episode because of the compartmentaliza-

tion of different co-infecting P. carinii types in the lungs

(15).

A second episode of PCP could result either from reac-

tivation of organisms that caused the first episode or from

de novo infection with a new P. carinii type acquired from

an exogenous source. In seven patients (patients 1–7),

reactivation was strongly suggested by the detection of

identical SSCP types in both episodes of PCP. An alterna-

tive explanation could be de novo infection in the second

episode by the same P. carinii PCR-SSCP type as that

which caused the first episode. However, the prevalence of

the types observed in the seven “reactivation” cases was

low in Lyon and Switzerland during the study period

(types no. 2, 5, and 7 represented 7%, 6%, and 10%,

respectively, of Lyon’s isolates; type 6 represented only

3.5% of the Swiss isolates [Figure]). Thus, reinfection with

these specific types was unlikely. All Danish patients (1, 3,

and 6) were infected with type 6. Although no prevalence

data for SSCP genotypes in Denmark are available, no

indication of possible contact between these patients, over-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 865

DISPATCHES

Table. Pneumocystis carinii DHPS and PCR-SSCP genotyping in AIDS patients with recurrent pneumonia

Patient no.

City

Age

Date of episode 1/

date of episode 2/

interval (mo)

CD4/mm

Prophylaxis at

PCP episode

Treatment

Outcome of

treatment

P. carinii PCR-

SSCP type

DHPS genotype

7/16/1993

CO → P

Success

6/8/1994 (11)

Success

1/31/1994

Success

5/18/1995 (16)

Success

8/19/1994

CO → C/P

Success

WT/M1

12/23/1994 (4)

Success

11/23/1994

Success

2, 5

WT/M3

3/23/1995 (4)

Death

2, 5

4/19/1995

Success

7, 8

3/1/1996 (11)

Success

11/16/1995

P → CO

Success

5/6/1996 (6)

Success

2/3/1998

Success

7/22/1998 (5)

C/P

Success

11/24/1990

Success

6, 10

7/29/1991 (8)

Success

12/8/1992

Success

11/5/1993 (11)

Success

3/22/1993

CO → P

Success

10/28/1994 (7)

CO → P

Death

3/30/1994

CO → A

Success

4, 7

3/28/1995 (12)

D+T → A

Success

9/21/1994

Success

10/21/1996 (25)

P+A

Success

10/12/1994

Success

1, 2

3/25/1996 (17)

PM/SD

P+A

Success

1, 3

PCP, Pneumocystis carinii pneumonia; DHPS, dihydropteroate synthase; PCR, polymerase chain reaction; SSCP, single-strand conformation polymorphism.

Co, Copenhagen (Denmark); CF, La Chaux-de-Fonds (Switzerland); La, Lausanne (Switzerland); Ly, Lyon (France).

A, atovaquone; CO, cotrimoxazole; C/P, clindamycin/primaquine; D,

dapsone; D+T, dapsone and trimethoprim; P, pentamidine; P+A,

pentamidin

e and atovaquone;

PM/SD (pyrimethamine/sulfadoxine inhibitors of dihydrofolate reductase (DHFR) and DHPS, respectively); T, trimetrexate (an inhibitor of DHFR).

WT, wild type (Thr55 Pro57); M1, mutant 1 (Ala

55 Pro57); M2, mutant 2 (Thr55 Ser57); M3, muta

nt 3 (Ala55 Ser57 double mutant).

Switch of molecules because of toxicity for patients 3, 6, and 11 and because of toxicity and treatment failure for patients 1

and 10.

Caused by PCP.

lap in hospitalization dates, or similar zip codes for home

address suggested transmission of type 6 between these

patients.

A change of P. carinii DHPS genotype between the two

episodes was observed in three reactivation cases, either

from wild type in the first episode to DHPS mutations in

the second one (patients 1 and 5) or from DHPS with a sin-

gle mutation (at position 57) in the first episode to a dou-

ble mutation in the second one (patient 2). In two patients

(3 and 4), the DHPS mutant strain was selected out of a

mixture of wild-type and DHPS mutant strains. Because

both episodes of each patient were caused by the same P.

carinii types and because all patients received co-trimoxa-

zole or dapsone as treatment, maintenance therapy, or

both, these results strongly suggest that selection of P.

carinii DHPS mutations occurred within the patients. The

results of tests on patients 3 and 4 isolates highlight the

fact that some patients may harbor genetically different

strains of P. carinii and that the mutant strain may be read-

ily selected when drug pressure is exerted. In the two

remaining patients (6 and 7), the P. carinii DHPS mutant

found in the bronchoalveolar lavage specimen from the

second episode was already present in the first episode.

The wild-type DHPS allele was more frequently

observed in the six reinfection cases than in the reactiva-

tion cases (8 wild-type alleles among 12 genotypes versus

4 among 16, Table). This finding is probably related to the

fact that, with the exception of the second episode of

patient 13, patients who were reinfected had no prophylax-

is or did not receive sulfa drugs for prophylaxis.

In all the second episodes caused by reactivation,

mutant DHPS strains were observed (7/7), compared to

only two of six second episodes caused by reinfection

(Table). This observation suggests an association between

mutant DHPS and second episodes attributable to reactiva-

tion (p<0.02, Fisher exact test).

Conclusions

Our study suggests that P. carinii DHPS mutants may

be selected in vivo (within a given patient) under the pres-

sure of co-trimoxazole or dapsone and that DHPS muta-

tions may be associated with reactivation of P. carinii.

Whether DHPS mutations are induced by the pressure of

the drug or preexisting and selected out by the pressure of

the drug remains to be determined. Physicians should be

alert to the increased risk for drug resistance during recur-

rence of PCP infection, although the impact of DHPS

mutations on retreatment with sulfa or sulfone drugs

remains to be determined. De novo selection of P. carinii

DHPS strongly favors the hypothesis that P. carinii is

developing sulfa and sulfone resistance.

Acknowledgments

We thank R. Lienhard for the samples he kindly provided

and S. Picot for storage of the specimens. We also thank A.

Telenti for critical reading of the manuscript, as well as J.L.

Touraine, D. Peyramond, and C. Trepo for access to patients’

charts in Lyon.

Financial support for this study was provided by the Swiss

National Program on AIDS Research grant 3345- 65407, Swiss

National Fund for Scientific Research grant 32-56715.99; Centre

de Coordination de la Lutte contre les Infections Nosocomiales

Sud-Est et Hospices Civils de Lyon; the Swiss Federal Office for

Education and Science for participation in EUROCARINII proj-

ect, Framework V Program, European Commission; and a North-

South fellowship from Lausanne University (supporting A.N.).

Dr. Nahimana obtained his bachelor’s degree and master’s

degree in microbiology at the University of Lausanne. The pres-

ent work was submitted by A. Nahimana as partial fulfillment for

a PhD degree at the University of Lausanne and was performed

under the supervision of P. Hauser.

References

1. Ma L, Borio L, Masur H, Kovacs H. Pneumocystis carinii dihy-

dropteroate synthase but not dihydrofolate reductase gene mutations

correlate with prior trimethoprim-sulfamethoxazole or dapsone use. J

Infect Dis 1999;180:1969–78.

2. Helweg-Larsen J, Benfield TL, Eugen-Olsen J, Lundgren JD,

Lundgren B. Effects of mutations in Pneumocystis carinii dihy-

dropteroate synthase gene on outcome of AIDS-associated P. carinii

pneumonia. Lancet 1999;354:1347–51.

3. Kazanjian P, Armstrong W, Hossler PA, Burman W, Richardson J,

Lee CH, et al. Pneumocystis carinii mutations are associated with

duration of sulfa or sulfone prophylaxis exposure in AIDS patients. J

Infect Dis 2000;182:551–7.

4. Olliaro P. Mode of action and mechanisms of resistance for anti-

malarial drugs. Pharmacol Ther 2001;89:207–19.

5. Sköld O. Sulfonamide resistance: mechanisms and trends. Drug

Resist Updat 2001;32:1608–14.

6. Navin TR, Beard CB, Huang L, del Rio C, Lee S, Pieniazek NJ, et al.

Effects of mutations in Pneumocystis carinii dihydropteroate syn-

thase gene on outcome of P carinii pneumonia in patients with HIV-

1: a prospective study. Lancet 2001;358:545–9.

866 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Figure. Frequency distribution of Pneumocystis carinii types

observed in different locations. Each type, co-infecting or not, was

considered as one isolate.

7. Stringer JR, Beard CB, Miller RF, Wakefield AE. A new name

(Pneumocystis jiroveci) for Pneumocystis from humans. Emerg Infect

Dis 2002;8:891–6.

8. Hauser PM, Francioli P, Bille J, Telenti A, Blanc DS. Typing of

Pneumocystis carinii f. sp. hominis by single-strand conformation

polymorphism of four genomic regions. J Clin Microbiol

1997;35:3086–91.

9. Nahimana A, Blanc DS, Francioli P, Bille J, Hauser PM. Typing of

Pneumocystis carinii f. sp. hominis by PCR-SSCP to indicate a high

frequency of co-infection. J Med Microbiol 2000;49:753–8.

10. Hauser PM, Blanc DS, Sudre P, Senggen Manoloff E, Nahimana A,

Bille J, et al. Genetic diversity of Pneumocystis carinii in HIV-posi-

tive and negative patients as revealed by PCR-SSCP typing. AIDS

2001;15:461–6.

11. Tsolaki AG, Miller RF, Underwood AP, Banerji S, Wakefield AE.

Genetic diversity at the internal transcribed spacer regions of the

rRNA operon among isolates of Pneumocystis carinii from AIDS

patients with recurrent pneumonia. J Infect Dis 1996;174:141–56.

12. Keely SP, Baughman RP, Smulian AG, Dohn MN, Stringer JR. Source

of Pneumocystis carinii in recurrent episodes of pneumonia in AIDS

patients. AIDS 1996;10:881–8.

13. Keely SP, Stringer JR. Sequences of Pneumocystis carinii f. sp.

hominis strains asssociated with recurrent pneumonia vary at multi-

ple loci. J Clin Microbiol 1997;35:2745–7.

14. Hughes WT. Current issues in the epidemiology, transmission, and

reactivation of Pneumocystis carinii. Semin Respir Infect

1998;13:283–8.

15. Helweg-Larsen J, Lundgren B, Lundgren JD. Heterogeneity and

compartmentalization of Pneumocystis carinii f. sp. hominis geno-

types in autopsy lungs. J Clin Microbiol 2001;39:3789–92.

Address for correspondence: Philippe M. Hauser, Centre Hospitalier

Universitaire Vaudois, 1011 Lausanne, Switzerland; fax: +41 21 314 40

60; email: Philippe.Hauser@chuv.hospvd.ch

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 867

DISPATCHES

Search ppast iissues oof EEID aat wwww.cdc.gov/eid

VIM- and IMP-Type

Metallo-β-lactamase–

Producing

Pseudomonas spp.

and Acinetobacter

spp. in Korean

Hospitals

Kyungwon Lee,* Wee Gyo Lee,† Young Uh,‡

Gyoung Yim Ha,§ Jihyun Cho,¶ Yunsop Chong,*

and the Korean Nationwide Surveillance of

Antimicrobial Resistance Group

We determined the occurrence of acquired metallo-β-

lactamase (MBL)–producing bacteria in Korean hospitals.

Among the isolates nonsusceptible to imipenem that were

collected from 28 hospitals from 2000 to 2001, 44 (11.4%)

of 387 Pseudomonas spp. and 38 (14.2%) of 267

Acinetobacter spp. produced MBL and had alleles of

bla

VIM-2

or bla

IMP-1

. MBL-producing isolates were detected

in 60.7% of the hospitals.

arbapenems are often used as a last resort for treating

serious infections attributable to multidrug-resistant

gram-negative bacilli because these drugs are stable even

to extended-spectrum and AmpC β-lactamases. However,

gram-negative bacilli with acquired metallo-β-lactamase

(MBL), IMP-1, emerged and spread during the early 1990s

in Japan (1). IMP-1 and its variants were then detected in

other countries (2).

Another type of acquired MBL, VIM-1, was first

reported in Pseudomonas aeruginosa in Italy (3), followed

by reports of VIM-2 in France and Greece. VIM-2 was

detected in P. aeruginosa in a Korean hospital isolated as

early as 1995 (4). The occurrence of the VIM enzyme has

continued to evolve: VIM-3 was reported in Taiwan (5),

and VIM-4 in the United States (6).

The bla

IMP

and bla

VIM

genes are horizontally transfer-

able because they are inserted in integrons, and some of

these integrons are located on conjugative plasmids (7).

Because of its ability to spread, carbapenem resistance

related to IMP and VIM β-lactamase production has

become a serious concern (8). Laboratory personnel and

physicians must consider the therapeutic and infection-

control implications of not detecting carbapenemase-pro-

ducing bacteria (9). A large number of VIM-2–producing

Pseudomonas spp. have been detected in a Korean hospi-

tal since 1995 (4), but the occurrence of MBL-producing

isolates has not been studied at other Korean hospitals,

despite the high prevalence of carbapenem-resistant P.

aeruginosa and Acinetobacter spp. (10). The aim of our

study was to determine the occurrence of acquired MBL-

producing P. aeruginosa and Acinetobacter spp. among

isolates collected by Korean Nationwide Surveillance of

Antimicrobial Resistance Group hospitals. The MBL types

produced and the sources of the MBL-positive isolates

were also investigated. In addition, pulsed-field gel elec-

trophoresis (PFGE) patterns were compared to determine

intra- and inter-hospital spread of resistant strains.

The Study

Nonduplicate, imipenem-resistant isolates of 387

Pseudomonas spp. and 267 Acinetobacter spp. were col-

lected from 2000 to 2001 from 28 hospitals in the Korean

Nationwide Surveillance of Antimicrobial Resistance

Group hospitals located in six cities or provinces. The

identification of the species and the imipenem susceptibil-

ity were confirmed at the coordinating laboratory by using

conventional tests (11) or ATB 32 GN system (bioMerieux,

Marcy-l’Etoile, France) and by using the disk diffusion

test (12), respectively.

MBL production was screened by using the Hodge test

and the imipenem-EDTA double disk synergy test (13).

The bla

IMP-1

and bla

VIM-2

alleles were detected by poly-

merase chain reaction (PCR), and three of the positive iso-

lates were confirmed by sequencing, as described previ-

ously (4). XbaI-digested genomic DNA of P. aeruginosa

isolates was separated by PFGE using the CHEF-DR-II

system (Bio-Rad Laboratories, Hercules, CA) (4). The pat-

tern was analyzed visually and by using UVIBand and

Map software (UVItec Ltd., Cambridge, UK).

Some of the Pseudomonas and Acinetobacter isolates

collected were not fully resistant to imipenem but showed

intermediate resistance when retested. Among the isolates

not susceptible to imipenem, 44 (11.4%) of 387

Pseudomonas spp. (42 P. aeruginosa and 2 P. putida) and

38 (14.2%) of 267 Acinetobacter spp. were considered

MBL producers on the basis of positive results by the

868 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

In addition to the listed authors, this group includes the following:

Jung Oak Kang, Moon Yeon Kim, Nam Yong Lee, Mi-Na Kim,

Myungshin Kim, Kyung Soon Song, Ki Sook Hong, In Ki Paik, Hye

Soo Lee, Sook-Jin Jang, Ae Ja Park, Sung Ha Kang, Won Keun

Song, Insoo Rheem, Eui-Chong Kim, Yeon Joon Park, Jong Hee

Shin, Myungseo Kang, Young-Kyu Sun, Hee Joo Lee, Hwan-Sub

Lim, Jong Wook Lee, and Bo-Moon Shin.

*Yonsei University College of Medicine, Seoul, Korea; †Ajou

University School of Medicine, Suwon, Korea; ‡Yonsei University

Wonju College of Medicine, Wonju, Korea; §College of Medicine of

Dongguk University, Kyongju, Korea; and Wonkwang University

College of Medicine, Iksan, Korea

Hodge test and imipenem-EDTA double disk synergy test

(Table 1). MBL-producing Pseudomonas spp. and

Acinetobacter spp. were detected in 11 (52.4%) of 21 and

10 (41.7%) of 24 hospitals that were located in four of five

and five of six cities or provinces, respectively. We detect-

ed the bla

VIM

allele by PCR from all 42 isolates of MBL-

producing P. aeruginosa and 2 isolates of P. putida. The

bla

VIM-2

and bla

IMP-1

alleles were detected in 27 (71.1%)

and 11 (28.9%) of 38 Acinetobacter isolates, respectively

(Table 2). Nucleotide sequencing for three representative

PCR-positive isolates confirmed the presence of the

bla

VIM-2

gene in one isolate each of P. aeruginosa and

Acinetobacter spp., and the bla

IMP-1

gene in one isolate of

Acinetobacter spp.

The MBL-producing strains were isolated mainly from

intensive-care unit patients (31.7%) and other inpatients

(50.0%); five (6.1%) were from emergency service and

other outpatients (Table 3). Overall, MBL-producing iso-

lates were mainly obtained from specimens of sputum

(50.0%) and urine (29.3%). However, the proportion of

MBL-producing isolates was relatively higher among

urine isolates: 17.3% for Pseudomonas spp. and 29.2% for

Acinetobacter spp. We obtained one MBL-producing

Acinetobacter isolate from each of the following specimen

types: blood, spinal fluid, pleural fluid, and venous

catheter tip (Table 4).

The PFGE of the XbaI-digested genomic DNA of 39

isolates of P. aeruginosa showed 22 patterns (data not

shown). Six isolates from one hospital had an identical pat-

tern. Thirteen isolates (33.3%) belonged to another identi-

cal pattern—six from one hospital, two from each of two

hospitals, and one from each of three hospitals, which were

located in a city and two provinces.

Conclusions

In this study, >10% of all imipenem-nonsusceptible iso-

lates of Pseudomonas spp. and Acinetobacter spp. were

attributable to MBL production (Table 1), and these MBL-

producing isolates were detected in 62.5% of the partici-

pating hospitals. Our finding indicates that MBL-produc-

ing P. aeruginosa is more prevalent in Korea than in other

countries (2) and that MBL-producing Acinetobacter spp.

is increasing. The percentage of hospitals with MBL-pro-

ducing isolates might have been higher if a larger number

of imipenem-nonsusceptible isolates had been collected

for this study.

VIM-2 was the only type of acquired MBL identified

initially in Korea. VIM-2–producing P. aeruginosa was

isolated at almost the same time in Europe (7) and Korea

(4). However, IMP-1–producing isolates were rare until

2000 in Korea. Only one and three IMP-1-positive P.

aeruginosa and Acinetobacter spp., respectively, have

been isolated at the coordinating laboratory (4, unpub.

data). In our study, 11 (28.9%) of 38 MBL-positive isolates

of Acinetobacter spp. were IMP producers (Table 2). This

increase suggests the possible introduction of IMP-produc-

ing strains of Acinetobacter spp. from Japan, where 28 iso-

lates of bla

IMP-1

-positive Acinetobacter baumannii were

reported in a hospital as early as 1994 to 1996 (14).

Rasmussen and Bush (15) predicted that an increase of

MBL-producing organisms was inevitable, given the more

frequent use of carbapenems. Imipenem has been used for

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 869

DISPATCHES

Table 1. Detection of metallo-β-lactamase–producing isolates among imipenem-nonsusceptible isolates of Pseudomonas spp. and

Acinetobacter spp.

No. hospitals (%)

No. isolates (%)

Organism

City/province

Tested

Positive

Tested

Positive

Seoul

4 (36.4)

144

12 (8.3)

Kyungki

2 (100)

6 (15.0)

Kangwon

1 (50.0)

2 (3.5)

Chulla

4 (100)

108

24 (22.2)

Kyungsang

0 (0)

Pseudomonas spp.

Tota l

11 (52.4)

387

44 (11.4)

Seoul

4 (36.4)

107

12 (11.2)

Kyungki

0 (0)

Kangwon

2 (25.0)

8 (19.5)

Chulla

1 (12.5)

13 (52.0)

Kyungsang

1 (50.0)

1 (1.9)

Chungchung

2 (100)

4 (33.3)

Acinetobacter spp.

Tota l

10 (41.7)

267

38 (14.2)

Four were tertiary-care hospitals.

Table 2. Detection of bla

VIM-2

and bla

IMP-1

allele from metallo-β-

lactamase–producing Pseudomonas spp. and Acinetobacter

spp. by polymerase chain reaction

No. isolates (%)

Organism

Tested

bla

VIM-2

positive

bla

IMP-1

positive

Pseudomonas aeruginosa

42 (100)

0 (0)

P. putida

2 (100)

0 (0)

Acinetobacter spp.

27 (71.1)

11 (28.9)

Tota l

71 (86.6)

11 (13.4)

only 9 years in Korea, but the imipenem-resistance rate of

P. aeruginosa has rapidly risen from 6% in 1996 to 19% in

2001. A study by the Korean Nationwide Surveillance of

Antimicrobial Resistance Group showed that the mean

imipenem-resistance rates of P. aeruginosa in 1997 did not

differ substantially depending on hospital size, (i.e., 17%

in medium hospitals [<1,000 beds] and 18% in large hos-

pitals [>1,000 beds]). The mean resistance rates to imipen-

em were not lower than those to ceftazidime in 2000, i.e.,

21% versus 18% in large hospitals and 20% versus 19% in

medium hospitals (data not shown).

Acinetobacter spp. are also common nosocomial

pathogens with multidrug resistance. The imipenem resist-

ance rate of this organism isolated in Korea was found to

be much lower than that of P. aeruginosa, but its resistance

rate rose from 4% in the first quarter to 20% in the third

quarter of 2002 at the coordinating laboratory (data not

shown).

In our study, MBL-producing Pseudomonas spp. and

Acinetobacter spp. were isolated mainly from sputum and

urine specimens, and most (81.7%) isolates were from

inpatients and intensive-care unit patients. Therefore,

proper treatment of respiratory secretions and urine from

intensive-care unit patients is considered an important

aspect of preventing the spread of MBL-producing organ-

isms. The presence of P. aeruginosa isolates with identical

PFGE patterns among those collected not only from cer-

tain hospitals but also from different hospitals suggests

that clonal spread is at least a part of the cause of intra- and

inter-hospital dissemination of MBL-producing isolates.

The presence of VIM-2-producing Serratia marcescens,

Enterobacter cloacae, and Achromobacter xylosoxidans

subsp. denitrificans (unpub. data) in other hospitals also

suggests horizontal transfer of the resistance determinants.

Cornaglia et al. reported that five of seven patients

infected with MBL-producing P. aeruginosa died,

although the cause of death was difficult to establish with

certainty (16). Clinical studies on the infection are rare

because isolation of MBL-producing gram-negative bacil-

li increased only recently. We anticipate difficulties in

treating patients infected with MBL-producing gram-neg-

ative bacilli, which can hydrolyze, in vitro, all available β-

lactams, except aztreonam for which clinical efficacy is

unknown.

Our study indicates the urgent need for action to pre-

vent further spread of MBL-producing organisms.

Previous experiences with penicillin-nonsusceptible pneu-

mococci, methicillin-resistant Staphylococcus aureus, van-

comycin-resistant Enterococcus faecium, and extended-

spectrum β-lactamase–producing Klebsiella pneumoniae

indicate that once resistant bacteria can become wide-

spread they cannot be controlled (10). Our first task is to

detect MBL producers among clinical isolates (9).

Although the National Committee for Clinical Laboratory

Standards document (12) does not contain procedures for

detection, simple procedures are available (13).

The prevalence of bla

VIM-2

allele-positive P. aeruginosa

and bla

IMP-1

allele-positive Acinetobacter spp. is increasing

possibly because of clonal and horizontal spread of the

resistance determinant in Korean hospitals. Sputum and

urine from inpatients and intensive-care unit patients were

found to be the main sources of MBL-producing isolates.

Laboratories not only in Korea but also in other countries

with carbapenem-resistant organisms must be prepared to

screen MBL-producing isolates to determine the clinical

impact and prevent further spread of MBL-producing

organisms.

870 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table 3. bla

VIM-2

and bla

IMP-1

allele-positive Pseudomonas spp. and Acinetobacter spp. isolated by service

No. isolates (%)

Organism

Outpatient

Inpatient

Intensive-care unit

Others

Tota l

Pseudomonas spp.

3 (6.8)

26 (59.1)

11 (25.0)

4 (9.1)

44 (100)

Acinetobacter spp.

2 (5.2)

15 (39.5)

6 (15.8)

38 (100)

Tota l

5 (6.1)

41 (50.0)

26 (31.7)

10 (12.2)

82 (100)

Two were emergency service patients, and one was a urology patient.

One was an emergency service patient, and one was a pediatric patient.

Table 4. bla

VIM-2

and bla

IMP-1

allele-positive Pseudomonas spp. and Acinetobacter spp. isolated by source

No. (%) of isolates with metallo-â-lactamase

Pseudomonas spp.

Acinetobacter spp.

Tota l

Source

Tested

Positive

Tested

Positive

Tested

Positive

% positive by

source

Sputum

200

22 (11.0)

143

19 (13.3)

343

41 (12.0)

50.0

Urine

17 (17.3)

7 (29.2)

122

24 (19.7)

29.3

Wound

2 (4.1)

7 (9.9)

120

9 (7.5)

10.9

Other

3 (16.7)

5 (17.2)

8 (17.0)

9.8

Tota l

387

44 (11.4)

267

38 (14.2)

654

82 (12.5)

100

Others included one Acinetobacter isolate of specimens from blood, spinal fluid, pleural fluid, and a venous catheter tip.

Acknowledgments

We thank Jong Hwa Yum and Dongeun Yong for detecting

the metallo-β-lactamase genes and Yonghee Suh for screening

the MBL producers.

Dr. Lee is director of the Research Institute of Bacterial

Resistance and a professor in the Department of Laboratory

Medicine, Yonsei University College of Medicine, Seoul, Korea.

He is the Korean coordinator in the World Health

Organization/Centers for Disease Control and Prevention

External Quality Assurance Scheme, and he is the organizer of

the Korean Nationwide Surveillance of Antimicrobial Resistance

Group. His research interests include antimicrobial resistance of

bacteria and its mechanisms.

References

1. Osano E, Arakawa Y, Wacharotayankun R, Ohta M, Horii T, Ito H, et

al. Molecular characterization of an enterobacterial metallo-β-lacta-

mases found in a clinical isolates of Serratia marcescens that shows

imipenem resistance. Antimicrob Agents Chemother 1994;38:71–8.

2. Gibb AP, Tribuddharat C, Moore RA, Louie TJ, Krulicki W,

Livermore DM, et al. Nosocomial outbreak of carbapenem-resistant

Pseudomonas aeruginosa with a new bla

IMP

allele, bla

IMP-7

Antimicrob Agents Chemother 2002;46:255–8.

3. Lauretti L, Riccio ML, Mazzariol A, Cornaglia G, Amicosante G,

Fontana R, et al. Cloning and characterization of bla

VIM

, a new inte-

gron-borne metallo-β-lactamase gene from a Pseudomonas aerugi-

nosa clinical isolate. Antimicrob Agents Chemother 1999;

43:1584–90.

4. Lee K, Lim JB, Yum JH, Yong D, Chong Y, Kim JM, et al. bla

VIM-2

cassette-containing novel integrons in metallo-β-lactamase–produc-

ing Pseudomonas aeruginosa and Pseudomonas putida isolates dis-

seminated in a Korean hospital. Antimicrob Agents Chemother

2002;46:1053–8.

5. Yan J-J, Hsueh P-R, Ko W-C, Luh E-T, Tsai S-H, Wu H-M, et al.

Metallo-β-lactamases in clinical Pseudomonas isolates in Taiwan and

identification of VIM-3, a novel variant of the VIM-2 enzyme.

Antimicrob Agents Chemother 2001;45:2224–8.

6. Tolman MA, Rolston K, Jones RN, Walsh TR. Molecular characteri-

zation of VIM-4, a novel metallo-beta-lactamase isolated from Texas:

report from the cancer surveillance program (2001). Proceedings of

the 42nd Interscience Conference on Antimicrobial Agents and

Chemotherapy. Abstract C1-1851. San Diego, California; 2002.

7. Poirel L, Naas T, Nicolas D, Collet L, Bellais S, Cavallo T-D, et al.

Characterization of VIM-2, a carbapenem-hydrolyzing metallo-β-lac-

tamase and its plasmid- and integron borne gene from a Pseudomonas

aeruginosa clinical isolate in France. Antimicrob Agents Chemother

2000;44:891–7.

8. Richet HM, Mohammed J, McDonald LC, Jarvis WR, INSPEAR.

Building communication networks: International Network for the

Study and Prevention of Emerging Antimicrobial Resistance. Emerg

Infect Dis 2001;7:319–22.

9. Molland ES, Black JA, Ourada J, Reisbid MD, Hanson ND, Thomson

KS. Occurrence of newer β-lactamases in Klebsiella pneumoniae iso-

lates from 24 U.S. hospitals. Antimicrob Agents Chemother

2002;46:3837–42.

10. Lee K, Lee HS, Jang SJ, Park AJ, Lee MH, Song WK, et al.

Antimicrobial resistance surveillance of bacteria in 1999 in Korea

with a special reference to resistance of enterococci to vancomycin

and gram-negative bacilli to third generation cephalosporin, imipen-

em, and fluoroquinolone. J Korean Med Sci 2001;16:262–70.

11. Kiska DL, Gilligan PH. Pseudomonas. In: Murray PR, Baron EJ,

Pfaller MA, Tenover FC, Yolken RH, editors. Manual of clinical

microbiology, 7th ed. Washington: American Society for

Microbiology; 1999. p. 517–25.

12. National Committee for Clinical Laboratory Standards. Performance

standards for antimicrobial susceptibility testing. Eleventh informa-

tional supplement. NCCLS document M100-S11. Wayne (PA): The

Committee, 2001.

13. Lee K, Chong Y, Shin HB, Kim YA, Yong D, Yum JH. Modified

Hodge test and EDTA-disk synergy tests to screen metallo-β-lacta-

mase-producing strains of Pseudomonas and Acinetobacter species.

Clin Microbiol Infect 2001;7:88–91.

14. Takahashi A, Yomoda S, Kobayashi I, Kubo T, Tsunoda M, Iyobe S.

Detection of carbapenemase-producing Acinetobacter baumannii in a

hospital. J Clin Microbiol 2000;38:526–9.

15. Rasmussen BA, Bush K. Carbapenem-hydrolyzing β-lactamases.

Antimicrob Agents Chemother 1997;41:223–32.

16. Cornaglia G, Mazzariol A, Lauretti L, Rossolini GM, Fontana R.

Hospital outbreak of carbapenem-resistant Pseudomonas aeruginosa

producing VIM-1, a novel transferable metallo-β-lactamase. Clin

Infect Dis 2000;31:1119–25.

Address for correspondence: Yunsop Chong, Department of Laboratory

Medicine and Research Institute of Bacterial Resistance, Yonsei

University College of Medicine, 134 Shinchon-dong Seodaemun-ku,

Seoul, Korea 120-752; fax: 82-2-313-0908; email: whonetkor@

yumc.yonsei.ac.kr

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 871

DISPATCHES

Leishmaniasis in

Germany

Gundel Harms,* Gabriele Schönian,*

and Hermann Feldmeier†

In 2000, a reference center was created to systemati-

cally record leishmaniases in Germany. We analyzed 58

cases of leishmaniases imported during a 2-year period.

These findings will serve as a baseline for the sandfly vec-

tor’s anticipated northward move because of global warm-

ing and as an advisory for immunocompromised persons

traveling to leishmaniasis-endemic areas.

eishmaniases compose a spectrum of protozoal dis-

eases currently endemic in 88 countries in Asia,

Africa, the Americas, and southern Europe. The geograph-

ic distribution of leishmaniases has widened, and the dis-

ease is reported in areas in which leishmaniasis was previ-

ously nonendemic (1). Apart from cutaneous, mucocuta-

neous/mucosal, and visceral leishmaniases, HIV-1–associ-

ated leishmaniasis acquired in southern Europe and in

other parts of the world have been observed in increasing

numbers (1,2).

Leishmaniasis is not notifiable in Germany. In

September 2000, a national advice and reference center

was created at the Institute of Tropical Medicine in Berlin;

the aim of the center was to monitor the frequency, origin,

and type of leishmaniases seen in Germany; to advise

physicians; and to improve information for travelers to dis-

ease-endemic areas. The healthcare professionals were

informed about the reference center by the Robert-Koch-

Institute, the center for surveillance of infectious diseases

in Germany, as well as through the Journal of the German

Medical Association, which is received by every registered

physician (3). Leishmaniasis was diagnosed if parasites

were detected in smears, culture, histologic sections, by

polymerase chain reaction (PCR) of lesion biopsy speci-

mens, bone marrow, or peripheral blood. For detection of

Leishmania-specific antigen the small subunit (ssu rRNA),

the internal transcribed spacer (ITS-1) region of the ribo-

somal RNA genes, or both, were amplified by PCR (4).

Leishmania complexes and species were determined by

digestion of the ribosomal ITS-1 PCR product with restric-

tion enzymes (4).

Within 2 years, 70 cases of leishmaniases (43 cutaneous

or mucocutaneous/mucosal; 27 visceral) were reported.

For 58 case-patients (35 cutaneous or mucocutaneous/

mucosal; 23 visceral), data were available on the age, sex,

residence, travel destination, possible exposure location,

reason for travel, duration of stay, duration and type of

symptoms, concomitant diseases or therapies, type of diag-

nosis, and treatment received.

Cutaneous and Mucosal Leishmaniasis

Of the 35 patients with cutaneous or mucocutaneous/

mucosal leishmaniasis, 30 were German tourists (Table 1).

The male-to-female ratio was 1.5:1. Ten had contracted

cutaneous or mucocutaneous/mucosal leishmaniasis in

Europe, 11 in Central and South America, 6 in Asia, and 3

in Africa. Two persons had been infected during work

stays of 1 to 4 months in French Guyana, one each in Peru

and Libya, and one patient had immigrated from

Afghanistan.

The median duration of lesions until the diagnosis of

leishmaniasis was made was 4 months (range 3 weeks to 2

years). Sixteen patients had more than one lesion (median

2, range 1–6 lesions). Seventeen lesions were located in

the face, including mouth and nose, 28 on the upper

extremities and 21 on the lower extremities. Lesions were

ulcerated in 39 cases, papular-nodular in 24, and plaque-

like in 3. Parasites were detected in 13 of 20 smears, in 9

of 10 cultures, and in 14 of 16 histologic sections; by using

PCR, Leishmania-specific DNA was detected in 16 of 16

biopsy specimens.

Patient 1 had lesions in the mouth caused by L. infan-

tum. She was under continuous immunosuppressive treat-

ment for severe bronchial asthma. Patient 2 had received

methotrexate and steroids for treatment of systemic col-

lagenosis for several weeks. Both patients were tested for

leishmanial infection in the blood; in both patients, the

Leishmania-specific PCR of the buffy coat of the blood

was positive. Patient 22 had mucocutaneous leishmaniasis

of the nasal septum. She had been treated for a skin lesion

caused by L. braziliensis 3 years earlier.

Visceral Leishmaniasis

A total of 18 of the 23 visceral leishmaniasis patients

were German tourists; 3 were immigrants from Angola,

Iran, and Togo; and 2 were visitors from Italy and Portugal

(Table 2). The male-to-female ratio was 6.7:1.

The median time between symptom onset and the cor-

rect diagnosis was 4 months (range 1–16 months). All

case-patients had fever, 17 (74%) had splenomegaly, 11

(48%) hepatomegaly, 20 (87%) anemia, 17 (74%) leukope-

nia, and 8 (35%) thrombocytopenia.

Bone marrow smears indicated Leishmania in 18 of 20,

bone marrow culture in 6 of 7, bone marrow histologic

sections in 7 of 8, PCR of the bone marrow in 8 of 9, and

PCR of the buffy coat of the blood in 7 of 7 cases.

Additionally, antibodies were detected in medium to high

concentration by an immunofluorescence test, enzyme-

872 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

*Charité, Humboldt University Berlin, Berlin, Germany; and †Free

University of Berlin, Berlin, Germany

linked immunosorbent assay (ELISA), or both, in 14 of 15

cases. Species was identified in 7 of 18 visceral cases con-

tracted in southern Europe and indicated Leishmania

belonging to the L. donovani complex, which implicated

infection with L. infantum.

Six cases of visceral leishmaniasis occurred in children

2 months of age to 11 years of age. Four German tourists

and two immigrants had long-known HIV infection (medi-

an duration 3 years, range 8 months–6 years). All HIV–co-

infected patients had CD4-cell counts below 200/µL

(median 108, range 23–185 CD4 cells/µL) when the diag-

nosis of visceral leishmaniasis was made. Of the remaining

11 patients, 1 had a thymoma with impaired T-helper-1 cell

function, 2 had received intermittent immunosuppressive

therapy (methotrexate and steroids) for rheumatologic dis-

ease, and 2 patients had their spleens removed. Three

patients were in an impaired general condition because of

combinations of diabetes, hypertonus, hypercholes-

terolemia, and emphysema. In the remaining three patients

(53–68 years of age), apart from hypertonus in one, no

impairing condition was detected.

Discussion

Information on single cases and a small case series of

imported leishmaniases in Germany is available, but sys-

tematic reporting on frequency, type, and origin of leish-

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 873

DISPATCHES

Table 1. Characteristics of patients with cutaneous and mucosal leishmaniasis, Germany

No.

Exposure

Status

Sex

Age (y)

Leishmania species

Treatment

Outcome

Europe

France

Tourist

L. donovani complex

No treatment

No cure

Italy

Tourist

L. donovani complex

Liposomal amphotericin B

Improved

Malta

Tourist

L. donovani complex

Perilesional pentavalent antimonials

Cured

Malta

Tourist

L. donovani complex

Perilesional pentavalent antimonials

Cured

Malta

Tourist

n.d.

Perilesional pentavalent antimonials

Cured

Malta

Tourist

n.d.

Perilesional pentavalent antimonials

Cured

Spain (Majorca)

Tourist

n.d.

Pentamidine isethionate

Cured

Spain (Majorca)

Tourist

L. donovani complex

Perilesional pentavalent antimonials

Cured

Spain

Tourist

L. donovani complex

No treatment

Unknown

Spain

Tourist

L. donovani complex

Antibiotic

Cured

Americas

Belize

Tourist

L. braziliensis complex

Liposomal amphotericin B

Unknown

Belize

Tourist

n.d.

IFN-gamma

Cured

Bolivia

Tourist

L. braziliensis complex

Liposomal amphotericin B

Unknown

Brazil

Tourist

n.d.

Systemic pentavalent antimonials

Cured

Brazil

Tourist

L. braziliensis complex

Liposomal amphotericin B

Unknown

Brazil

Tourist

L. braziliensis complex

Liposomal amphotericin B

Cured

Ecuador

Tourist

L. braziliensis complex

Liposomal amphotericin B

Cured

Ecuador

Tourist

n.d.

Liposomal amphotericin B

Cured

French Guyana

Work stay

L. braziliensis complex

Liposomal amphotericin B

Cured

French Guyana

Work stay

L. braziliensis complex

Liposomal amphotericin B

Cured

Guatemala

Tourist

L. mexicana

Ketoconazole

Cured

Peru

Work stay

L. braziliensis complex

Liposomal amphotericin B

Cured

Peru

Tourist

L. braziliensis complex

Liposomal amphotericin B

Cured

Peru

Tourist

L. braziliensis complex

Liposomal amphotericin B

Cured

Asia

Afghanistan

Immigrant

L. tropica

Aminosidine ointment

No cure

Afghanistan

Tourist

n.d.

Aminosidine ointment

Cured

United Arab Emirates

Tourist

n.d.

Perilesional pentavalent antimonials

Cured

Syria

Tourist

n.d.

Perilesional pentavalent antimonials

Unknown

Syria

Tourist

n.d.

Perilesional pentavalent antimonials

Unknown

Turkey

Tourist

n.d.

Antibiotic

Cured

Turkey

Tourist

L. donovani complex

Liposomal amphotericin B

Improved

Africa

Egypt

Tourist

L. tropica

Aminosidine ointment

No cure

Egypt

Tourist

L. tropica

Aminosidine ointment

Cured

Kenya

Tourist

n.d.

Perilesional pentavalent antimonials

Cured

Lybia

Work stay

n.d.

Aminosidine ointment

Cured

n.d., not done; IFN, interferon.

manial infections in Germany did not exist until 2000

(5–7). Our recent surveillance is dependent on passive

consultation and reporting and therefore may have selec-

tion bias because if visceral leishmaniasis, a potentially

fatal disease that requires hospitalization, is suspected,

advice on diagnosis and treatment is sought more often

than for the skin infection. We assume that our system cap-

tures approximately half of the visceral leishmaniasis

cases and approximately one third of the classical cuta-

neous cases imported to Germany.

A total of 47% of all cases, but 78% of the visceral

cases were contracted in the European Mediterranean area

and Portugal, and most of the infections indicated a species

of the L. donovani complex, most probably L. infantum, as

874 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table 2. Characteristics of patients with visceral leishmaniasis, Germany

No.

Exposure

Status

Sex

Age

Risk factor

Leishmania

species

Treatment

Outcome

Italy

Tourist

2 y

Child

n.d.

Liposomal amphotericin B

Cured

Italy

Tourist

5 y

Child

n.d.

Liposomal amphotericin B

Cured

Italy

Tourist

11 y

Child

n.d.

Liposomal amphotericin B

Cured

Spain

Tourist

8 mo

Child

n.d.

Liposomal amphotericin B

Cured

Spain

Tourist

9 mo

Child

L. donovani

complex

Liposomal amphotericin B

Cured

Iran

Immigrant

7 y

Child

L. donovani

complex

Liposomal amphotericin B

Cured

Spain

Tourist

43 y

HIV

L. donovani

complex

Liposomal amphotericin B;

maintenance therapy:

HAART plus liposomal

amphotericin B once monthly

No relapse for 6

months

Spain (Ibiza)

Tourist

48 y

HIV

L. donovani

complex

Liposomal amphotericin B;

maintenance therapy:

HAART plus liposomal

amphotericin B once monthly

No relapse for 8

months

Portugal

Visitor

29 y

HIV

n.d.

Liposomal amphotericin B;

maintenance therapy:

HAART plus liposomal

amphotericin B once monthly

unknown

France

Tourist

31y

HIV

L. donovani

complex

Liposomal amphotericin B;

maintenance therapy:

HAART plus liposomal

amphotericin B once monthly

Relapse after 4

months;

retreatment with

liposomal

amphotericin B; No

relapse for 3

months

Angola

Immigrant

40 y

HIV

L. donovani

complex

Systemic pentavalent

antimonials; maintenance

therapy: HAART

plus pentavalent antimonials

once monthly

unknown

Togo

Immigrant

37 y

HIV

L. donovani

complex

Liposomal amphotericin B

Cured

Italy (Sicily)

Visitor

31 y

Thymoma

n.d.

Liposomal amphotericin B

Cured

Italy (Ischia)

Tourist

67 y

Methotrexate/steroids

L. donovani

complex

Liposomal amphotericin B

Cured

Italy (Sicily)

Tourist

68 y

Methotrexate

n.d.

Liposomal amphotericin B

Cured

Italy (Ischia)

Tourist

70 y

Splenectomy

n.d.

Liposomal amphotericin B

Cured

Spain

Tourist

51 y

Splenectomy

L. donovani

complex

Liposomal amphotericin B

Cured

Greece

(Korfu)

Tourist

66 y

Diabetes mellitus

Hypertonus

n.d.

Liposomal amphotericin B

Cured

Spain

Tourist

52 y

Hypertonus

Hypercholesterolemia

n.d.

Liposomal amphotericin B

Cured

Greece

(Korfu)

Tourist

45 y

Diabetes mellitus

Emphysema

L. donovani

complex

Systemic pentavalent

antimonials

Cured

Tunisia

Tourist

53 y

Hypertonus

L. donovani

complex

Liposomal amphotericin B

Cured

Malta

Tourist

55 y

n.d.

Liposomal amphotericin B

Cured

China

Tourist

67 y

n.d.

Liposomal amphotericin B

Cured

n.d., not done; HAART, highly active anti-retroviral therapy.

the probable causative agent. Thirteen infections (22%)

were acquired on the Mediterranean islands of Ibiza,

Ischia, Majorca, Malta, Korfu, or Sicily.

This distribution reflects the fact, that the

Mediterranean countries, Spain, Italy, and the

Mediterranean islands, in particular, are the favorite vaca-

tion areas for Germans. Annually, Germans take 18 million

vacations to the European Mediterranean area (including 8

million to Spain and 6 million to Italy) with a median dura-

tion of 2 weeks. Sixty percent of travel to Italy and 90% of

travel to Spain are to Leishmania-endemic areas.

While leishmaniasis has always been endemic in the

Mediterranean countries, the maximum northern latitude

for sandfly survival is speculated to move further to the

North, beyond Germany (1) because of global warming. If

this scenario is correct, the imported cases may serve as a

potential substrate for the sandfly vector. Dogs that are

imported as pets from the disease-endemic areas of south-

western Europe or that contract the infection when accom-

panying their owners for vacation are another potential

substrate (8).

Infections with L. infantum in a child, as well as in a

horse who had never left Germany, have recently been

described and have led to speculations about an autochtho-

nous focus (9,10). Also recently, the first sandfly species,

Phlebotomus mascittii Grassi, 1908, was detected in south-

ern Germany, although its potential as a vector of

Leishmania remains to be demonstrated (11).

As expected, visceral leishmaniasis is often manifested

in persons with impaired immunocompetence because of

young age, HIV infection, immunosuppressive therapy and,

in our analysis, in older persons with concomitant diseases.

Notably, 12 (67%) of 18 of the visceral cases contracted

in the European Mediterranean area were in adults, thus

confirming a change in age groups affected. Formerly, vis-

ceral leishmaniasis was known mainly as a disease of chil-

dren (1,2). This change may partly be explained by the

increased proportion of Leishmania and HIV–co-infected

persons and partly by increased travel activities of otherwise

immunocompromised persons, including elderly persons.

Furthermore, even in patients with cutaneous leishma-

niasis, dissemination of parasites has to be excluded in

case of impaired immunocompetence (e.g., immunosup-

pressive treatment). In these cases, Leishmania-specific

PCR of the buffy coat of the peripheral blood is a sensitive

method for detecting parasite spread beyond the skin.

Parents of small children and persons with reduced

immunocompetence should be informed about their

increased susceptibility to infection with Leishmania when

traveling to disease-endemic areas. Measures to reduce the

exposure to sandflies, such as clothes, repellents, and mos-

quito nets as well as collars impregnated with repellents

for accompanying dogs, should be recommended.

Acknowledgments

We thank colleagues at German hospitals and health institu-

tions who discussed their patients with us or kindly contributed

to the data collected on leishmaniases.

Dr. Harms is a senior lecturer and researcher in tropical

medicine and international health at the Institute of Tropical

Medicine and at the Medical Faculty Charité, Humboldt

University Berlin, Germany. Her primary research interests are

the interaction of HIV/AIDS and parasitic infections, leishmani-

asis in particular.

References

1. Desjeux P. The increase in risk factors for leishmaniasis worldwide.

Trans R Soc Trop Med Hyg 2001;95:239–43.

2. World Health Organization. Leishmania/HIV co-infection: south-

western Europe, 1990–1998 (WHO/LEISH/2000.42). Geneva: The

Organization; 2000.

3. Harms G, Bienzle U. Leishmaniosen-importierte Krankheiten. Dtsch

Arztebl 2000;31/32:1589–92.

4. Schönian G, Schnur L, Fahri M. Genetic heterogeneity in the species

Leishmania tropica revealed by different PCR-based methods. Trans

R Soc Trop Med Hyg 2001;95:217–24.

5. Hohenschild S, Feldmeier H. Imported kala azar in children and

adults—comparison of medical history, clinical and diagnostic find-

ings. J Trop Pediatr 1995;41:378–9.

6. Harms G, Zenk J, Martin S, Kokozidou M, Püschel W, Bienzle U, et

al. Localized lymphadenopathy due to leishmanial infection.

Infection 2001;29:355–6.

7. Hölzer E, Kupferschmidt HG. Cutaneous leishmaniasis in East

German citizens. Z Arztl Fortbild (Jena) 1986;80:381–3.

8. Gothe R, Nolte I, Kraft W. Leishmaniasis in dogs in Germany: epi-

demiological case analysis and alternatives to conventional causal

therapy. Tierarztl Prax 1997;25:68–73.

9. Bogdan C, Schönian G, Banuls AL, Hide M, Pratlong F, Lorenz E, et

al. Visceral leishmaniasis in a German child who had never entered a

known endemic area: case report and review of the literature. Clin

Infect Dis 2001;32:302–6.

10. Koehler K, Stechele M, Hetzel U, Domingo M, Schönian G, Zahner

H, et al. Cutaneous leishmaniasis in a horse in southern Germany

caused by Leishmania infantum. Vet Parasitol 2002;109:9–17.

11. Naucke TJ, Pesson B. Presence of Phlebotomus (Transphlebotomus)

mascittii Grassi, 1908 (Diptera: Psychodidae) in Germany. Parasitol

Res 2000;86:335–6.

Address for correspondence: Gundel Harms, Institute of Tropical

Medicine Berlin, Spandauer Damm 130, 14050 Berlin, Germany; fax:

+49-30-30116-888; email: [email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 875

DISPATCHES

Probable Dengue

Virus Infection

among Italian

Troops, East

Timor, 1999–2000

Mario Stefano Peragallo,* Loredana Nicoletti,†

Florigio Lista,* and ‡Raffaele D’Amelio

for The East Timor Dengue Study Group

To investigate the attack rate and risk factors for prob-

able dengue fever, a cross-sectional study was conducted

of an Italian military unit after its deployment to East Timor.

Probable dengue was contracted by 16 (6.6%) of 241

army troops and caused half of all medical evacuations

(12/24); no cases were detected among navy and air force

personnel.

engue fever (DF), caused by dengue virus (DENV)

serotypes 1 to 4, is an emerging public health problem

in many tropical countries (1). Dengue hemorrhagic fever

(DHF) and dengue shock syndrome (DSS), the severe

manifestations of DENV infection, were first recognized

in the 1950s in Southeast Asia and are today a leading

cause of childhood illness and death in many tropical

countries. More recently, DHF and DSS have emerged in

Central and South America and in the Pacific region (2,3).

DF is also recognized as an emerging health problem for

international travelers (4,5) and for troops deployed to

tropical countries (6,7).

In 1999, following a United Nations Security Council

recommendation, the International Force for East Timor

(INTERFET) was formed to restore peace on the island. In

November 1999, INTERFET troops totaled 11,000 from

17 countries. The Italian Armed Forces contributed 640

soldiers.

DF is endemic in East Timor. The peak transmission

periods for DF are July–August and December–January,

corresponding to the rainy months (8). In 1998, at least

11% of hospital inpatient deaths in East Timor were attrib-

uted to DHF (9). In October 1999, a localized outbreak of

DF in a western district was attributed to serotype 3 (9) and

serotype 2 was isolated in December 1999 (10). Serotypes

2 and 3 were also responsible for DF cases among

Australian troops returning from East Timor in January–

February 2000 (11).

During deployment, a high attack rate of febrile illness

consistent with DF was reported among Italian troops. A

seroepidemiologic survey was therefore conducted in

February 2000 among soldiers returning home, in an

attempt to determine the cause of this outbreak and to

define infection rates and risk factors for infection.

The Study

All Italian troops eligible for deployment are routinely

vaccinated against diphtheria/tetanus, tetravalent meningo-

coccal meningitis, measles/mumps/rubella, hepatitis A and

B, polio (with inactivated virus), typhoid fever (orally), and

yellow fever (YF). In this situation, troops were also vacci-

nated against Japanese encephalitis (JE) (Nakajama strain, 3

doses on days 0, 7, and 14) just before landing in East Timor.

DF prevention consisted of the use of personal protec-

tion measures against mosquitoes (repellents applied to the

skin; permethrin-treated bed nets and uniforms) along with

environmental mosquito control. Adulticide spraying was

conducted weekly by pesticide-dispersal units but only

within the campsite and in its nearest surroundings, which

were also inspected daily to reduce or eliminate breeding

sites of vectors.

Italian troops were deployed in East Timor from late

September 1999 to mid-February 2000, and all 640 partic-

ipating military personnel were eligible for inclusion in the

study. Army soldiers were permanently based on the

ground and operated in Dili and surrounding areas, while

air force and navy personnel had only logistical tasks and

their presence in Dili was episodic, since they were main-

ly aboard ship or based in Darwin (Australia).

A seroepidemiologic survey was conducted February

15–28, 2000, among troops returning to Italy after their 3-

month period of duty in East Timor. After informed con-

sent was obtained, peripheral blood specimens were drawn

and a written questionnaire administered. The question-

naire asked for personal health data, including all symp-

toms experienced during deployment and information

about compliance with personal protection measures.

Immunization status and clinical data concerning febrile

illness cases consistent with DF were obtained from stan-

dardized records kept by medical personnel. Soldiers and

navy/air force personnel were studied according to their

serologic status and disease status during deployment.

876 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

*Centro Studi e Ricerche Sanità e Veterinaria Esercito, Rome,

Italy; †Istituto Superiore di Sanità, Rome, Italy; and ‡Stato

Maggiore della Difesa and Università “La Sapienza,” Seconda

Facoltà di Medicina, Rome, Italy

The East Timor Dengue Study Group: Roberto Biselli, Glauco

Calì, Maria Rosaria Capobianchi, Maria Grazia Ciufolini, Raffaele

D’Amelio, Paolo Di Zenzo, Giovanni Fascia, Cristiano Fiorentini,

Alberto Germani, Michele Giattino, Giuseppe Ippolito, Florigio

Lista, Fabio Magurano, Antonella Marchi, Loredana Nicoletti,

Mario Stefano Peragallo, Alessandro Polidori, Giuseppe Sarnicola,

and Antonio Stella.

All specimens were screened for antibodies to dengue

virus serotype 2 (DEN-2), yellow fever virus (YVF), and

West Nile virus (WNV) by hemagglutination-inhibition

test (HI). All serum specimens positive for DEN-2 were

tested by neutralization test (NT) for DEN-2. Additionally,

serum samples from participants who had experienced an

acute clinical syndrome suggestive of DF were directly

tested by NT for antibodies to DEN-2. Serum specimens

negative for DEN-2 were then tested for neutralizing anti-

bodies to dengue virus serotypes 1, 3, and 4 (DEN-1,

DEN-3 and DEN-4).

The HI test was performed by the method of Clarke and

Casals (12) and NT as 90% plaque reduction neutralization

test (PRNT) on Vero cells. Briefly, serum specimens

(twofold dilutions) and virus (10

PFU) were incubated

overnight at 4°C, injected onto monolayers of Vero cells,

and overlaid with 1% Tragacanth gum (Sigma-Aldrich

S.r.I., Milan, Italy). Seven days postinfection, cells were

washed with saline and stained with 1% crystal violet in

20% ethanol (DEN-2 and DEN-3) or by immunodetection

assay (DEN-1 and DEN-4) as described (13). Vero cells

were propagated in minimum essential medium with

Earle’s salts (EMEM), supplemented with nonessential

amino acids, 10% fetal calf serum, 100 IE/mL of penicillin

G, and 100 IE/mL of streptomycin.

The following viruses were used in the study: DEN-1

(Hawaii), DEN-2 (NGB), DEN-3 (H87), DEN-4 (H241),

YF (Asibi), and WN (Bratislava). Viruses were injected

into suckling mice by the intracerebral route. For NT, viral

stocks were prepared as 10% brain suspension in Hank’s

saline+7.5% bovine serum albumin (Sigma-Aldrich). For

HI, test antigens were prepared by sucrose-acetone extrac-

tion from mouse infected brains (12). Monoclonal antibod-

ies specific for DEN-1 or broadly reactive with flavivirus-

es were purchased from ATCC (ATCC HB112, ATCC

HB47) and used as mouse ascitic fluid after injection into

adult BALB/c mice.

Undetermined febrile illness was defined as an acute

clinical syndrome with temperature >38.5°C, unrelated to

diarrhea, malaria, or other identified infections. Suspected

dengue (14) was defined as an undetermined febrile illness

of 2–7 days’ duration, associated with two or more of the

following manifestations: headache, retroorbital pain,

myalgia, arthralgia, cutaneous rash. Antibody levels

1:1,280 dilutions by HI (1,15) for DEN-2 and >1:20 dilu-

tions by NT to at least one of the four DENV serotypes

were considered supportive serologic evidence of a recent

dengue infection. Probable dengue (1,14) was defined as a

case compatible with the clinical description of suspected

DF and serologic findings supportive of a recent dengue

infection.

The prevalence of undetermined febrile illness, sus-

pected dengue, and probable dengue was compared by chi-

square test among army and navy/air force personnel.

Since navy and air force personnel had a limited exposure

to the environment of East Timor, risk factors for probable

dengue were studied only in the army contingent. A uni-

variate analysis was first performed by Fisher exact test;

each risk variable was crossed with the prevalence of prob-

able dengue. Significance was tested at a level of α=0.05.

A multiple logistic regression model was used to deter-

mine the relationship between the outcome of probable

dengue and a set of explanatory variables, and test the sig-

nificance of each variable while simultaneously account-

ing for demographic and risk factors. The following vari-

ables were included in the model: age, rank, previous

deployments in dengue-endemic areas, YF/JE vaccination,

night guards, skin repellents/permethrin-treated uni-

forms/bed nets use, and operational versus logistic tasks.

To identify a subset of variables significantly related to

probable DF, the stepwise procedure was performed with

the likelihood ratio test, by using at each step the p value

of 0.05 as entry criterion and the p value of 0.10 as

removal criterion. Univariate statistical analysis was per-

formed with EpiInfo 6.04d software (Centers for Diseases

Control and Prevention, Atlanta, GA, January, 2001) and

multivariate analysis by SPSS 11.0 software (SPSS Inc.,

Chicago, IL).

Conclusions

Of 640 eligible participants (280 army, 93 air force, and

267 navy), 595 (93%) were included in the study: 241

army, 88 air force, and 266 navy personnel (Table 1).

Serum specimens and questionnaires were obtained within

2 weeks after the troops’ return, in late February 2000.

Some (14.5%) of the troops had previously been

deployed to DF-endemic areas, primarily Somalia and

Mozambique in 1992–1994. According to their immuniza-

tion status versus YF and JE viruses, 100 (41.5%) of the

241 army soldiers had received vaccinations against YFV

and JEV, 119 (49.4%) had been vaccinated against JEV

only, 2 (0.8%) against YF only, and 20 (8.3%) had not been

vaccinated.

Undetermined febrile illness was more frequently

reported (p<0.01) among army soldiers than among navy

and air force personnel: 85 (35.3%) of 241 versus 13

(3.7%) of 354 , respectively. All participants with suspect-

ed dengue (n=30), with serologic results supportive of a

recent dengue infection (n=27), and with a probable case

of dengue n=16), belonged to the army group (Table 2).

The 16 participants with probable dengue showed also

a significant increase (p<0.01) in HI antibody titers to YFV

(>1:1,280 in 15/16 infected soldiers vs. 14/225 uninfected

soldiers) and WNV (>1:1,280 in 10/16 vs. 6/225). The

average interval between the onset of clinical manifesta-

tions suggestive of DF and the date when blood samples

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 877

DISPATCHES

were taken was 36±25 days. All 16 case-patients with

probable DF had a fever >38.5°C; a saddle-back fever pat-

tern was recorded for 5 (31.3%). Other reported symptoms

included myalgia and rash in 13 (81.3%); headache in 11

(68.8%); retroorbital pain in 9 (56.3%), and adenopathy in

3 (18.8%). No patients had DHF/DSS.

The mean duration of probable DF cases was 7±3 days.

Moreover, 12 of the 16 patients with probable DF were

evacuated because of their clinical status. Univariate

analysis of risk factors for probable DF suggested a possi-

ble protective effect of JEV vaccination and personal pro-

tection measures (Table 3). However, logistic regression

analysis identified only a subset of variables significantly

related to probable dengue, whose risk was higher among

soldiers on duty in operational rather than logistic units,

and lower among participants with regular use of bed nets

(Table 4).

Since most of soldiers had been previously vaccinated

with a flavivirus vaccine (YFV, JEV, or both), their

immune response to an eventual dengue infection was

expected to be a secondary (anamnestic) response, with

high-titer antibodies cross-reacting with several DENV

serotypes, as well as other flaviviruses (15). Thus, in spite

of the lack of paired serum specimens, high antibody titers

to DEN-2 by HI (>1:1,280) (1,16) and to any of the four

dengue virus serotypes by NT (>1:20), after an average of

36 days from the onset of clinical manifestations compati-

ble with dengue infection, may be considered supportive

serology of a recent flavivirus infection, likely acquired

during deployment.

Overall, 6.6% of army soldiers contracted probable

dengue. No cases of probable DF were detected in the low-

exposure group of navy and air force personnel. The high

attack rate of probable dengue among the army contingent

may be due to several reasons. First, DF and DHF/DSS are

epidemic throughout Southeast Asia (3), including

Indonesia (17); in particular, the incidence of DF marked-

ly increased in East Timor in 1998–1999 (18). Secondly,

the multinational deployment to East Timor took place

during the rainy season (December–January), when the

risk of infection is high.

Approximately 60% of troops with supportive serolog-

ic evidence of a recent dengue infection showed the clini-

cal manifestations of classic DF, 20% had milder symp-

toms, and 20% were asymptomatic. This finding agrees

with the U.S. troops’ experience in Somalia in 1993, where

>85% of all DENV infections were symptomatic (6). In

contrast, the overall ratio of inapparent to clinical DENV

infections is quite high in persons living in disease-endem-

ic areas, as in Indonesia, where it has been reported to be

as high as 9.3 (17).

Performing duties outside the camp was associated

with a significantly higher risk of infection, probably

because vector control activities were regularly carried out

within the compound. Regular use of bed nets was the only

personal protection measure that significantly decreased

the risk of contracting probable dengue. This finding is not

new (6) and may have been because some of the troops

were frequently on duty at night and thus slept during the

day when the biting activity of dengue vectors is highest.

Otherwise, the regular use of repellents (applied to the

skin) and permethrin-treated uniforms seemed to decrease

the risk for dengue infection, but the differences between

those who did not follow these practices and those who did

were not significant statistically.

DF is therefore an emerging problem for troops

deployed to dengue-endemic areas, mainly because of the

lack of effective preventive measures, the high attack rate,

878 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Table 1. Characteristics of Italian military personnel, East Timor, 1999–2000

Feature

Army

Navy

Air Force

Total

No. participants

241

266

595

Mean age (years ± sd)

27 ± 7

28 ± 7

35 ± 7

–

Time of deployment

22/Sep/99–16/Feb/00

21/Oct/99–19/Feb/00

19/Sep/99–17/Feb/00

–

Mean duration of deployment (days ± sd)

100 ± 25

109 ± 14

41 ± 23

–

Person months

803

968

102

1,873

Presence in East Timor for >90 days (no. soldiers)

241

Episodical presence in East Timor (no. soldiers)

266

354

Table 2. Clinical and serologic findings of recent dengue infection among Italian troops

Serologic findings of recent dengue infection

Clinical assessment

No. supportive (%)

No. not supportive (%)

Total no. (%)

Undetermined febrile illness

6 (22.2)

49 (22.9)

55 (22.8)

Suspected dengue

16 (59.3)

14 (6.5)

30 (12.4)

Asymptomatic

5 (18.5)

151 (70.6)

156 (64.7)

Total

27 (100)

214 (100)

241 (100)

All military personnel with supportive serologic findings belonged to the army contingent (N=241). Probable dengue cases are represented by the 16 soldiers with

clinical manifestations compatible with DF (suspected dengue) and serologic findings supportive of a recent dengue infection.

Cases are defined in the section “Materials and Methods.”

the high symptomatic/inapparent infection ratio, and the

long period of being unfit for duty after the acute phase of

the disease. DF may thus seriously disrupt the readiness of

a military unit. Moreover, previously infected soldiers

redeployed to disease-endemic areas may be at increased

risk for DHF/DSS complications. Persons previously

infected by a DENV serotype may be at higher risk of

developing DHF/DSS, if they are subsequently infected by

a different serotype. Such risks should be taken into

account while planning international peace-keeping opera-

tions, and the risk of DHF among previously dengue-

infected military personnel should be evaluated.

Cross-reaction by antiflavivirus antibodies induced by

JEV vaccine may otherwise afford some cross-protection

against DF. JEV vaccine (Nakajama strain) seems to

decrease the attack rate of DHF and reduce the severity of

cases for a short time (19). More recently, researchers have

noted that prior vaccination of hamsters with a live, atten-

uated JEV vaccine strain (not licensed for human use) and

a St. Louis encephalitis virus wild strain seems to reduce

the severity of a subsequent WNV infection (20). Our data

suggest that prior vaccination with the commercially avail-

able JEV inactivated vaccine for human use (Nakajama

strain) may have some protective effect against subsequent

probable DF. The decrease was, however, not significant,

according to the multiple logistic regression model we

used.

Our data suggest that effectiveness of routine protective

measures against vector mosquitoes is far from satisfacto-

ry. A tetravalent dengue vaccine is needed to effectively

reduce the risk for DF and DHF/DSS among troops

deployed to tropical areas as well as to protect long-term

international travelers to dengue-endemic countries.

Acknowledgements

We thank David Vaughn, Ashley Croft, and Tom Jefferson

for critical review of the manuscript and Antonino Bella and

Fortunato “Paolo” D’Ancona for statistical analysis.

Dr. Peragallo is a researcher at the Centro Studi e Ricerche

di Sanità e Veterinaria of the Italian Army. His main research top-

ics are the epidemiology and control of infectious diseases, par-

ticularly in tropical settings.

References

1. Gibbons RV, Vaughn DW. Dengue: an escalating problem. BMJ

2002;324:1563–6.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 879

DISPATCHES

Table 3. Risk factors associated with probable dengue (univariate analysis)

Demographic and risk factors

No. cases of probable DF/no. soldiers exposed (%)

OR 95% CI

p value

Age, <26 y

Yes

10/124 (8.1)

6/117 (5.1)

1.62

(0.51 to 5.61)

0.26

Lower rank (enlisted men vs. NCOs/officers)

Yes

9/145 (6.2)

7/96 (7.3)

0.84

(0.27 to 2.76)

0.47

Previous deployments in dengue-endemic areas

Yes

3/44 (6.8)

13/197 (6.6)

1.04

(0.18 to 4.01)

0.59

YFV vaccination

Yes

5/102 (4.9)

11/139 (7.9)

0.60

(0.16 to 1.95)

0.26

JEV vaccination

Yes

9/219 (4.1)

7/22 (31.8)

0.09

(0.03 to 0.34)

<0.01

Night guard at least once a week

Yes

6/142 (4.2)

10/99 (10.1)

0.39

(0.11 to 1.25)

0.06

Skin repellents, regular use (at least once a day)

Yes

9/209 (4.3)

7/32 (21.9)

0.16

(0.05 to 0.56)

<0.01

Use of permethrin-treated uniform

Yes

9/186 (4.8)

7/55 (12.7)

0.35

(0.11 to 1.17)

0.05

Bed nets, regular use (every night)

Yes

9/223 (4.0)

7/18 (38.9)

0.07

(0.02 to 0.26)

<0.01

On duty in operational vs. logistic units

Yes

15/179 (8.4)

1/62 (1.6)

5.55

(0.82 to 238.67)

0.05

DF, dengue fever; OR, odds ratio; CI, confidence interval; NCOs, noncommissioned officers; YFV, yellow fever virus; JEV, Japane

se encephalitis virus.

Fisher exact test.

Table 4. Risk factors associated with probable dengue by

multivariate analysis

Risk factors

OR estimate

p value

On duty in operational vs. logistic units

11.29

<0.05

Bed nets, regular vs. nonregular use

0.04

<0.01

OR, odds ratio.

2. Gubler DJ and Clark GG. Dengue/dengue hemorragic fever: the

emergence of a global health problem. Emerg Infect Dis

1995;1:55–7.

3. World Health Organization. Dengue/dengue haemorragic fever: situ-

ation in 2000. Wkly Epidemiol Rec 2000;75:193–200.

4. Potasman I, Srugo I, Schwartz E. Dengue seroconversion among

Israeli travelers to tropical countries. Emerg Infect Dis 1999;5:824–7.

5. Jelinek T. Dengue fever in international travelers. Clin Infect Dis

2000;31:144–7.

6. Sharp TW, Wallace MR, Hayes CG, Sanchez JL, DeFraites RF, Arthur

RR, et al. Dengue fever in U.S. troops during operation Restore

Hope, Somalia, 1992–1993. Am J Trop Med Hyg 1995;53:89–94.

7. Trofa AF, DeFraites RF, Smoak BL, Kanesa-thasan N, King AD,

Burrous JM, et al. Dengue fever in US military personnel in Haiti.

JAMA 1997; 277:1546–8.

8. World Health Organization, Dili Office, East Timor. East Timor

health sector situation report: January-June 2000. Available from:

URL: http://www.who.int/eha/emergenc/etimor/14082000.htm

9. World Health Organization, Dili Office, East Timor. Weekly Report

45. 1999. Available from: URL: http://www.who.int/eha/emergenc/

etimor/191199.htm

10. World Health Organization, Dili Office, East Timor. Weekly Report

50–52 1999 & 01 2000. Available from: URL: http://www.who.int/

eha/emergenc/etimor/141200.htm

11. Hills S, Piispanen J, Foley P, Smith G, Humphreys J, Simpson J, et al.

Public health implications of dengue in personnel returning from East

Timor. Communicable Disease Intelligence [serial online]

2000;24:365–8. Available from: URL: http:// http://www.health.

gov.au/pubhlth/cdi/cdi2000.htm#december

12. Clarke DH, Casals J. Techniques for the hemagglutination and

hemagglutination-inhibition with arthropod-borne viruses. Am J Trop

Med Hyg 1958;7:561–77.

13. Desprès P, Frenkiel MP and Deubel V. Differences between cell

membrane fusion activities of two Dengue type-1 isolates reflect

modification of viral structure. Virology 1993;196:209–19.

14. World Health Organization. Recommended surveillance standards.

edition. Geneva: The Organization; 1999. p. 39.

15. World Health Organization. Laboratory diagnosis. In: Dengue

haemorragic fever. Diagnosis, treatment, prevention and control. 2nd

edition. Geneva: The Organization, 1997. p. 34–47.

16. Gubler DJ. Dengue and dengue hemorrhagic fever. Clin Microbiol

Rev 1998;11:480–96.

17. Corwin AL, Larasati RP, Bangs MJ, Wuryadi S, Arjoso S, Sukri N, et

al. Epidemic dengue transmission in southern Sumatra, Indonesia.

Trans R Soc Trop Med Hyg 2001;95:257–65.

18. Krause V. Increase in dengue fever notifications in visitors to East

Timor. Northern Territory Disease Control Bulletin [serial online]

2000;7:6-7. Available from: URL: http:// http://www.nt.gov.au/

health/cdc/bulletin/june_2000.pdf

19. Hoke CH, Nisalak A, Sangawhipa N, Jatanasen S, Laorakapongse T,

Innis BL, et al. Protection against Japanese encephalitis by inactivat-

ed vaccine. N Engl J Med 1988;319:608–14.

20. Tesh RB, Travassos da Rosa APA, Guzman H, Araujo TP, Xiao SY.

Immunization with heterologous Flaviviruses protective against fatal

West Nile encephalitis. Emerg Infect Dis 2002;8:245–51.

Address for correspondence: Mario Stefano Peragallo, Capo Primo

Reparto, Centro Studi e Ricerche di Sanità e Veterinaria dell’Esercito, Via

S. Stefano Rotondo 4, 00184 Rome, Italy; fax: ++39 06 7009005; email:

[email protected]

880 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

The print journal is available at no charge to public health professionals

YES, I would like to receive Emerging Infectious Diseases.

Please print your name and business

address in the box and return by fax to

404-371-5449 or mail to

EID Editor

CDC/NCID/MS D61

1600 Clifton Road, NE

Atlanta, GA 30333

Moving? Please give us your new address (in the box) and print the number of your old

mailing label here_________________________________________

Full text free online at

www.cdc.gov/eid

HIV Epidemic

among Young Thai

Men, 1991–2000

Kalyanee Torugsa,* Scott Anderson,†

Nucharee Thongsen,* Narongrid Sirisopana,*

Achara Jugsudee,‡ Pitak Junlananto,‡

Sorachai Nitayaphan,*

Suebpong Sangkharomya,*

and Arthur E. Brown*

Characterization of the HIV epidemic in Thailand has

benefited from the systematic testing of young men upon

entry into the military. These data, which have shown that

public health measures can reverse an HIV epidemic, have

been reanalyzed with current geographic information sys-

tems methods. The resulting maps are, thus far, the best

means of visualizing the geography of the dynamic HIV epi-

demic in Thailand.

n few nations is the HIV epidemic characterized as well

as it is in Thailand. As elsewhere, sentinel surveillance

of high-risk populations is performed and more general

populations are partially assessed by HIV testing of blood

donors and pregnant women. In addition, Thailand has

benefited from another measure to assess the HIV epidem-

ic in the general population. Since very early in the epi-

demic, young men have been tested for HIV at the time

they entered into military service. Military medicine and

public health leaders recognized that, in addition to using

this information to benefit individual men, maintaining

this information in a confidential database linked to area of

residence would allow monitoring of national demograph-

ic trends in the epidemic. Data generated in this way have

provided key evidence that public health policy has

reversed the trend of increasing HIV infections (1), pre-

venting an estimated 200,000 HIV infections by 2000 (2).

The HIV prevalence data on young Thai men is provid-

ed to public health policy makers on a regular basis. It has

also been published in the scientific literature (3,4) with

data presented that uses provinces as the scale of analysis.

More recently, this unique dataset has been analyzed by

using geographic information systems (GIS). GIS methods

allow enhanced visualization of the trends, both geograph-

ic and chronologic, within this well-documented HIV epi-

demic. Using a GIS approach, we present the Royal Thai

Army dataset, expanded to include the decade of

1991–2000 and data precision refined to the district level.

Methods

The methods of recruiting young men into the Thai mil-

itary have been described previously (3,4). Approximately

50,000–60,000 men, mostly 21 years of age, are selected

each year by lottery in their home province (the province

in which they are listed on the house registration). The sys-

tem produces a representative national sampling of Thai

men. Because of their young age, HIV prevalence in these

annual recruit classes may serve as a proxy for HIV inci-

dence. Induction into the military occurs in May (M) and

November (N) of each year. At the time when blood is col-

lected, recruits provide information about the location of

their main residence (including province and district) dur-

ing the previous 2 years (3). Although the actual locations

where infections occurred are unknown, residential data

enables analysis of the association between HIV preva-

lence and this key geographic marker.

To refine the HIV prevalence analyses, geographic

localization uses districts as the first administrative subunit

of provinces. When data were analyzed by using annual

classes grouped at the district level, calculations of the per-

centage testing positive for HIV were statistically unreli-

able because the number of men tested in some rural dis-

tricts was so small. Therefore, we merged data in two ways

to decrease variability attributable to the small sample size

of the prevalence figures: classes were combined across

time, and districts were combined across space. Sixteen

classes of men recruited from 1991 to 2000 were combined

temporally into four larger classes, each representing dis-

crete 2-year periods: 1) N91, M92, N92, M93; 2) N94,

M95, N95, M96; 3) N96, M97, N97, M98; and 4) N98,

M99, N99, M00. Data on classes recruited from the M91,

N93, M94, and N00 lotteries were not available for analy-

sis (M91, before full implementation; N93 and M94, pro-

tocol under revision; N00, completed after dataset closed).

However, even after combining classes into these 2-year

periods, a number of districts still had numbers too low for

statistical reliability.

We also merged some districts with neighboring dis-

tricts so that each had a minimum denominator of 20 in the

HIV prevalence calculation. Numbers of >

20 persons pro-

vide minimal, but acceptable, reliability in the percentage-

positive calculations. Districts with <20 were combined

with other districts according to the following protocol,

following a sequence of priorities: we combined districts if

they were in the same province, had historic connections

(formerly part of single larger district), had similarly small

numbers tested, had similar demographics, and had similar

topography or other geographic features.

For the GIS analysis, data tables provided in Excel files

(Microsoft Corp., Redmond, WA) by the Armed Forces

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 881

DISPATCHES

*Armed Forces Research Institute of Medical Sciences, Bangkok,

Thailand; †State University of New York at Cortland, Cortland,

New York, USA; and ‡Army Institute of Pathology, Bangkok,

Thailand

Research Institute of Medical Sciences were joined to dis-

trict-level GIS maps obtained from the Thai

Environmental Institute and National Statistical Office. We

used Arcview 3.2a software (ESRI, Redlands, CA) to cre-

ate dot density and choropleth maps. (A choropleth map

uses shades or colors to demonstrate the geographic distri-

bution of a range of values.)

Results and Discussion

Figure 1 provides visual evidence of the geographic

distribution of the epidemic. Each of the 10,043 dots rep-

resents a recruit who tested positive for HIV infection

(positive for HIV-1 by enzyme immunoassay and Western

blot) among the 442,923 recruits tested. Arcview software

uses an algorithm that allows the random placement of

dots within a geographic area (in this case a district). By

creating these dots at the district level but presenting them

at the province level, we show the vivid geographic pattern

of infection while illustrating the human impact of the epi-

demic. The dot distribution does not distinguish between

population density and HIV prevalence. Nevertheless,

these absolute numbers provide important information

regarding the potential HIV-related impact on the health-

care system and various other social structures.

The evolution of the HIV epidemic over the course of

the decade is seen in the four maps shown in Figure 2.

These maps use gradients of color shading to distinguish

the percentage of men who tested positive for HIV during

each 2-year period. The initial high prevalence of HIV in

men from the upper north region is illustrated, as is the

decline in prevalence over time (which likely occurred as

a result of public health interventions). The relative per-

sistence of HIV prevalence in some districts in the far

south of the country is also shown, suggesting that locale-

specific features of the epidemic need to be well under-

stood in that area, potentially leading to changes in public

health interventions.

Previous analyses conducted at the province level iden-

tified regional variation quite clearly, but at a less detailed

level (3,4). The maps from our study indicate considerable

variation even to the district level of analysis, which sug-

gests important local factors may be at work in determin-

ing rates of transmission. The HIV epidemic in Thailand

seems to have largely missed some districts and devastat-

ed others, even within the same province. Chiang Mai and

Chiang Rai Provinces are examples of this phenomenon.

These maps also show that, although the growth of the epi-

demic has slowed in most parts of the country, the epidem-

ic has not decreased in some districts in south Thailand.

Choropleth maps best demonstrate the changing preva-

lence of HIV among recruits from each district. The dot

map best demonstrates the intensity of the epidemic with-

in and across provinces. Both types of imagery are

required to describe and understand the epidemic and to

suggest locales in which public health initiatives are need-

ed. In areas where the intensity of the epidemic is highest,

health and social services may require additional resources

to serve persons with the disease. In areas where the preva-

lence rates are highest or are increasing, socially and cul-

turally appropriate interventions may be needed to

strengthen and focus HIV programs.

These GIS maps are the most accurate maps available

to date and are built upon the unique HIV prevalence

datasets collected over a decade by the Royal Thai Army.

GIS mapping at this descriptive level enhances visualiza-

tion of these data (5,6). More sophisticated methods of

multivariate GIS visualization could enhance this analysis

even more but would require the Royal Thai Army to

assess recruits at induction on other variables, especially

risk factors, or to use recruits’ home addresses rather than

districts or provinces of residence. Nonetheless, changes in

time and place of chronic viral infection can be added to

882 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

DISPATCHES

Figure 1. Dot density map of young men who tested positive for

HIV at time of entry into the Royal Thai Army, Thailand, November

1991–May 2000. Each dot represents one man. Location of dots

based on recruit’s residence during the previous 2 years. Data on

recruits entering in November 1993 and May 1994 are not avail-

able.

the recognized areas of GIS application in epidemiology,

transmission of vector-borne and water-borne diseases,

and environmental health (7). The maps also illustrate the

point that political borders may have no relationship to epi-

demiologic boundaries. Differences in HIV prevalence

often occur, not only between regions and provinces but

also within provinces and across provincial borders. When

an epidemic can be visualized at this level of detail and

accuracy, researchers can formulate and address questions

regarding the bases of these differences. At the same time,

policy makers can better direct intervention strategies and

more finely assess the outcomes of these interventions.

Coordination of data collection and joint GIS analysis by

neighboring countries would enhance regional understand-

ing of the emergence and spread of HIV epidemics and

monitoring of control programs.

Dr. Torugsa is a research scientist at the Armed Forces

Research Institute of Medical Sciences (AFRIMS) in Bangkok,

Thailand. Her research interests are in infectious disease epi-

demiology, and her current focus is the establishment of a disease

surveillance and outbreak response system in the Royal Thai

Army Medical Department.

References

1. Phoolcharoen W. HIV/AIDS prevention in Thailand: success and

challenges. Science 1996:280;1873–4.

2. World Bank. Thailand’s response to AIDS: building on success, con-

fronting the future. p. 1, Nov 3, 2000 [cited 2002 Jun 21]. Available

from: URL: http://www.worldbankor.th/social/index.html

3. Sirisopana N, Torugsa K, Mason CJ, Markowitz LE, Jugsudee A,

Supapongse T, et al. Correlates of HIV-1 seropositivity among young

men in Thailand. J Acquir Immune Defic Syndr Hum Retrovirol

1996:11;492–8.

4. Mason CJ, Kitsiripornchai S, Markowitz LE, Chanbancherd P,

Supapongse T, Jugsudee A, et al. Nationwide surveillance of HIV-1

prevalence and subtype in young Thai men. J Acquir Immune Defic

Syndr Hum Retrovirol 1998:19:165–73.

5. Albert DP, Gesler WM, Levergood B. Spacial analysis, GIS, and

remote sensing applications in the health sciences. Chelsea (MI): Ann

Arbor Press; 2000.

6. Gatrell AC, Loytonen M. GIS and health. GISDATA VI. London:

Taylor & Francis; 1998.

7. Clarke KC, McLafferty SL, Tempalski BJ. On epidemiology and geo-

graphic information systems: a review and discussion of future direc-

tions. Emerg Infect Dis 1996:2;85–92.

Address for correspondence: Arthur E. Brown, Department of

Retrovirology, AFRIMS, APO AP 96546; fax: 66-02-644-4824; email:

[email protected].mil

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 883

DISPATCHES

Figure 2. Choropleth maps of HIV prevalence in four classes of young men at time of entry into the Royal Thai Army, Thailand,

1991–2000. Location determined by residence during the previous 2 years. Prevalence is stratified by color and localized to district or

group of districts so that calculations are based on >20 men.

884 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

Taenia solium

Cysticercosis, Irian

Jaya, Indonesia

To the Editor: Cysticercosis, a tis-

sue infection caused by accidental

ingestion of eggs released from

humans harboring the pork tapeworm,

Taenia solium (TsCysti), is one of the

most serious reemerging parasitic dis-

eases worldwide (1). Taeniasis is an

intestinal infection caused by the

adult stage of the large tapeworm.

Carriers of T. solium acquire infection

through eating undercooked pork con-

taminated with cysticerci (larvae).

Although most Indonesian people are

Muslim and do not eat pork, infection

with T. solium has occurred in some

areas or islands where most local peo-

ple are Christian or Hindi.

The area most affected by this

infection is Irian Jaya, Indonesia, the

western half of New Guinea Island

(2–4). In field surveys conducted in

2000 and 2001, we found that 5

(8.6%) of 58 local people and 7 (11%)

of 64 local dogs living approximately

1 km from the local capital city,

Wamena, in Jayawijaya District, har-

bored adult tapeworms and cysticerci

of T. solium, respectively (5,6). We

have further seroepidemiologic data

from 1996 and molecular confirma-

tion of subcutaneous nodules (SCN)

as cysticerci of the T. solium Asian

genotype. We believe this organism is

an emergent problem in Irian Jaya.

We previously reported that

TsCysti was highly endemic in

Jayawijaya District, Irian Jaya (2–6).

A total of 96 local people >

18 years of

age from Assologaima, Jayawijaya

District, were chosen at random and

examined by serologic testing and by

administering questionnaires in

February 1996 after the local and

Indonesian governments gave their

ethical approval. The 96 persons were

divided into three groups on the basis

of a history of epileptic seizures (ES,

n=17), physical examination of SCN

by palpation (n=32), or good health

(including no ES or SCN; n=47). A

total of 14 subcutaneous nodules

removed from 14 men in both ES and

SCN groups were confirmed to be

cysticerci of T. solium by morpholog-

ic observation and to be T. solium

Asian genotype by mitochondrial

DNA analysis with cytochrome c oxi-

dase subunit 1 gene (3,7). For serolog-

ic analysis, we conducted an enzyme-

linked immunosorbent assay (ELISA)

that used glycoproteins purified from

cyst fluid of T. solium cysticerci by

preparative isoelectric focusing (frac-

tions of pH 9.1) (8) in 2001.

On the basis of serologic results,

12 (70.6%) of 17, 20 (62.5%) of 32,

and 12 (25.5%) of 47 of ES, SCN, and

healthy groups, respectively, were

infected with the larval stage of T.

solium. Serologically positive rates

increased to 83.3% (10/12) of people

with subcutaneous nodules in the ES

group. A follow-up study of seroposi-

tive persons in the healthy group in

1997 showed that five of eight per-

sons had ES (two persons), headache

(one person), or SCNs in upper arm

(two persons). Seropositive persons in

all three groups (ES, SCN, and health)

were considered to be infected with

TsCysti. Persons of the SCN and

healthy groups who showed optical

density values higher than the cut-off

value were considered to have asymp-

tomatic TsCysti cases.

The local persons we examined

ranged from 18 to 29 years of age

(n=30), 30–44 years of age (n=36),

and >45 years of age (n=30).

Seropositive persons (n=12) from the

ES group (n=17) were 18 to 29 years

of age (40.0%, 2/5), 30–44 years

(71.4%, 5/7), and >

45 years (100%,

5/5). The prevalence of TsCysti did

not vary statistically by sex (males

53.6% [37/69] versus females 33.3%

[9/27], Pearson’s chi-square test,

p=0.074).

That 14 persons confirmed to have

subcutaneous cysticerci of T. solium

were seropositive strongly suggests

that the serologic test (ELISA) is

highly reliable for detecting TsCysti

in patients, whether their infection is

symptomatic or asymptomatic. In

contrast, one of the following scenar-

ios was expected for cases in three of

five persons in the ES group who did

not have SCN and were seronegative:

1) the case was not due to TsCysti, 2)

the case was caused by TsCysti but

without antibody response, rather

common in cases of a solitary cyst, or

3) the case was caused by TsCysti

with calcified cysts and without anti-

body response. Twelve (approximate-

ly 40%) of seronegative persons from

the SCN group (n=32) were expected

to have cases of TsCysti without anti-

body response or to have calcified

cysts without antibody response.

Cases without antibody response

would be most expected because of

the heavily contaminated environ-

ment in Irian Jaya (3–6). However,

further evaluation with computed

tomography or magnetic resonance

imaging scans is necessary. Based on

serologic results and mitochondrial

DNA confirmation of T. solium Asian

genotype (3,7), we concluded that

47.9% (46/96) of local people exam-

ined at random, 53.6% of men (37/69)

and 33.3% of women (9/27) >

18 years

of age had TsCysti.

An additional 30 local people in non

–TsCysti-endemic Merauke District

underwent serologic testing. One

woman had an exceptionally high

antibody titer. She was a transmigrant

from another island (South Sulawesi

Province). Although Paniai, Jayawijaya,

and Manokwari Districts are contami-

nated with T. solium taeniasis and cys-

ticercosis (2–4), no additional critical

evidence exists to show that Merauke

District has already been contaminat-

ed with this parasite.

Taeniasis and cysticercosis may

have been accidentally introduced

into Irian Jaya in 1969 when the coun-

try was governed by Indonesia, since

the governing body came from Bali,

the only area in Indonesia where

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 885

LETTERS

TsCysti was exclusively endemic (2).

The contaminated areas in Irian Jaya

have increased from the central area

(Paniai), to the east (Jayawijaya) (3),

and then to the west (Manokwari),

where 54 TsCysti cases have been

reported (Papua Province Health

Office Services, 1997, unpub. data).

We wanted to know if taeniasis/cys-

ticercosis had been introduced into

the eastern half of New Guinea Island,

called Papua New Guinea (PNG) (9).

We had already serologically con-

firmed that 16 (3.0%) of 541 local res-

idents and Irianese refugees in Alice

River villages along the border in

PNG had asymptomatic TsCysti (Ito

et al., unpub. data). Follow-up sur-

veys will be crucial in several other

districts including Merauke District in

Irian Jaya, PNG, and other islands

such as Timor Island, where most of

the population is Christian and many

suspected cases have recently been

reported by the District Health Office

Services (10). Schoolchildren should

also be checked so that cases can be

detected and treated early. Sustainable

education of the local community in

Irian Jaya, Indonesia, and Papua New

Guinea is also necessary.

Financial support was provided by a

grant-in-aid from the Nissan Foundation

and the Ministry of Education, Japan

(09044279, 10557029, 11694259,

1255702414, and 14256001) (A.I.) and by

the Japan Health Science Foundation

(S.S.M.).

Toni Wandra,* Akira Ito,†

Hiroshi Yamasaki,†

Thomas Suroso,*

and Sri S. Margono‡

*Ministry of Health, Jakarta, Indonesia;

†Asahikawa Medical College, Asahikawa,

Japan; and ‡University of Indonesia,

Jakarta, Indonesia

References

1. Schantz PM, Wilkins PP, Tsang VWC.

Immigrants, imaging and immunoblot: the

emergence of neurocysticercosis as a sig-

nificant public health problem. In: Scheld

WM, Craig WA, Hughes GM, editors.

Emerging infections 2. Washington, DC:

ASM Press; 1998. p. 213–42.

2. Simanjuntak GM, Margono SS, Okamoto

M, Ito A. Taeniasis/cysticercosis in

Indonesia as an emerging disease. Parasitol

Today 1997;13:321–3.

3. Wandra T, Subahar R, Simanjuntak GM,

Margono SS, Suroso T, Okamoto M, et al.

Resurgence of cases of epileptic seizures

and burns associated with cysticercosis in

Assologaima, Jayawijaya, Irian Jaya,

Indonesia, 1991–95. Trans R Soc Trop Med

Hyg 2000; 94:46–50.

4. Subahar R, Hamid A, Purba W, Wandra T,

Karma C, Sako Y, et al. Taenia solium

infection in Irian Jaya (West Papua),

Indonesia: a pilot serological survey of

human and porcine cysticercosis in

Jayawijaya District. Trans R Soc Trop Med

Hyg 2001;95:388–90.

5. Ito A, Putra MI, Subahar R, Sato MO,

Okamoto M, Sako Y, et al. Dogs as alterna-

tive intermediate hosts of Taenia solium in

Papua (Irian Jaya), Indonesia confirmed by

highly specific ELISA and immunoblot

using native and recombinant antigens and

mitochondrial DNA analysis. J Helminthol

2002;76:311–4.

6. Margono SS, Ito A, Sato MO, Okamoto M,

Subahar R, Yamasaki H, et al. Taenia soli-

um taeniasis/cysticercosis in Papua,

Indonesia in 2001: detection of human

worm carriers. J Helminthol 2003;

77:39–42.

7. Nakao M, Okamoto M, Sako Y, Yamasaki

H, Nakaya K, Ito A. A phylogenetic hypoth-

esis for the distribution of two genotypes of

the pig tapeworm Taenia solium world-

wide. Parasitology 2002;124:657–62.

8. Ito A, Plancarte A, Ma L, Kong Y, Flisser

A, Cho SY, et al. Novel antigens for neuro-

cysticercosis: simple method for prepara-

tion and evaluation for serodiagnosis. Am J

Trop Med Hyg 1998;59:291–4.

9. McManus DP. Improved diagnosis as an aid

to better surveillance of Taenia solium cys-

ticercosis, a potential health threat to Papua

New Guinea. Papua New Guinea Medical

Journal 1995;38:287–94.

10. Suroso T. Petunjuk pemberantasan taenia-

sis/sistiserkosis di Indonesia. Depkes RI,

Ditjen. Jakarta: PPM & PL; 2000. p. 1–30

(in Indonesian).

Address for correspondence: Akira Ito,

Department of Parasitology, Asahikawa

Medical College, Midorigaoka-Higashi 2-1-1-

1, Asahikawa 078-8510, Japan; fax: +81-

(0)166-68-2429; email: akiraito@asahikawa-

med.ac.jp

Recombinant

Vaccine–Derived

Poliovirus in

Madagascar

To the Editor: Between October

2001 and April 2002, five cases of

acute flaccid paralysis associated with

vaccine-derived poliovirus (VDPV)

type 2 isolates were reported in the

southern province of the Republic of

Madagascar. The first patient, an 11-

year-old child from the urban district

of Toliara, first experienced paralysis

on October 29, 2001. Three other chil-

dren, 6, 9, and 14 months of age from

Ebakika village, in a rural district of

Taolagnaro (250 miles east of Toliara),

showed signs of poliomyelitis

between March 21 and March 26,

2002. The last case-patient, a 20-

month-old child from Ambanihazo

village (6 miles north of Ebakika),

came into contact with one of the three

case-patients in Ebakika in March

2002, and symptoms developed on

April 12, 2002 (1). None of the

patients had been fully vaccinated

against poliomyelitis.

Nine type 2 poliovirus (PV) strains

were isolated. A restriction fragment

length polymorphism (RFLP) assay,

with three different genomic regions

amplified by reverse transcription–

polymerase chain reaction (RT-PCR)

and four different restriction enzymes

(HinfI, DpnII, RsaI, and DdeI) were

used to characterize the PV isolates at

the molecular level (2). The RFLP

profiles of all of the isolates in the two

capsid protein regions were identical

to that of the type 2 strain of the oral

polio vaccine (OPV) in the VP1-2A

region (nucleotides 2,872 to 3,647)

but slightly different in the VP3-VP1

region (nucleotides 1,915 to 2,883).

The observed differences allowed us

886 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

to distinguish two groups (isolates

from Toliara and isolates from

Taolagnaro) and two subgroups (iso-

lates from March and isolates from

April). The RFLP profiles of isolates

in the noncapsid region, at the 3′-ter-

minal end of the genome (polymerase

3D and 3′ noncoding regions:

nucleotides 6,535 to 7,439) also con-

firmed the presence of two separate

groups. These last profiles were com-

pletely different from those of the

three reference vaccine strains, sug-

gesting recombination with other

enteroviruses.

Partial genomic sequencing con-

firmed these observations. The entire

VP1 region (903 nucleotides) of the

type 2 PV strains from Toliara and

Taolagnaro differed from the type 2

OPV strain by 1% and 2.5%

nucleotides, respectively. This differ-

ence may indicate that the two strains

had been multiplying or circulating

for approximately 1 and 2.5 years,

respectively. Taolagnaro strains are

closely related to each other (<1%

nucleotide difference) but appear to be

very different from Toliara strains

(2.9% nucleotide difference), indicat-

ing the existence of two genetic line-

ages. The sequencing of the noncapsid

region (440 nucleotides corresponding

to nucleotide positions 6,705 to 7,144

of the Sabin 2 genome) confirms the

existence of two lineages derived

from different recombination events

with two nonidentified enteroviruses

of the phylogenetic cluster C. This

cluster, based on sequence similarity,

includes some coxsackieviruses and

all PV strains (3).

We tried to identify the donor

strains for sequences in the 3’ terminal

end of these recombinant strains by

aligning the nonidentified sequences

with homologous enterovirus

sequences available in a nucleotide

sequence database (FASTA, version

3.3 applied to GenBank) (4). The

highest percentages of nucleotide

sequence identity were those with

PVs and with most other cluster C

enteroviruses available in the data-

base (87% to 91% nucleotide identi-

ties). No wild PV strains have been

isolated in Madagascar since 1997

despite surveillance and investigation

of viral causes of acute flaccid paraly-

sis cases (5). Thus, that the detected

VDPVs were the product of recombi-

nation between OPV strains and two

nonpolio enteroviruses is more likely

than that they were the product of

OPV strains and two different unde-

tected wild PV strains. However, we

cannot exclude the possibility that

wild PVs were imported or circulating

silently for a while.

In response to the outbreak, the

local health authorities conducted

house-to-house vaccination with

OPV. Further field investigations

were carried out to determine the

extent to which VDPV had spread and

to search actively for other cases.

Data analysis is in progress.

As with the other epidemics in

Egypt and Hispaniola, VDPV circu-

lated in a province of Madagascar

with low OPV coverage (6,7).

Because a high OPV coverage rate

helps prevent the circulation of both

VDPVs and wild PVs, obtaining and

maintaining high rates of immuniza-

tion coverage are essential (8).

Moreover, two recombinant VDPV

lineages in Madagascar indicate that

recombination is frequent between

OPV and cluster C enteroviruses.

Similar recombinant VDPVs have

been implicated in the epidemics in

Hispaniola and in the Philippines

(6,9). Determining whether the neu-

rovirulence and transmissibility of

these VDPVs could be the result of

The print journal is available at no charge to public health professionals

YES, I would like to receive Emerging Infectious Diseases.

Please print your name and business

address in the box and return by fax to

404-371-5449 or mail to

EID Editor

CDC/NCID/MS D61

1600 Clifton Road, NE

Atlanta, GA 30333

Moving? Please give us your new address (in the box) and print the number of your old

mailing label here_________________________________________

Full text free online at

www.cdc.gov/eid

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 887

LETTERS

the recombination with nonpolio

enteroviruses is important. These

VDPVs have major implications for

the cessation of immunization with

OPV after certification that wild PV

has been eradicated.

Acknowledgments

We thank R. Crainic for support

and encouragement, O. Kew for pro-

viding unpublished results; and E. de

Gourville, O. Tomori, D. Wood, and

F. Colbere-Garapin for their interest

and advice.

Dominique Rousset,*

Mala Rakoto-Andrianarivelo,*

Richter Razafindratsimandresy,*

Bakolalao Randriamanalina,†

Sophie Guillot,‡ Jean Balanant,‡

Philippe Mauclère,*

and Francis Delpeyroux‡

*Institut Pasteur de Madagascar,

Antananarivo, Republic of Madagascar;

†Ministry of Health, Antananarivo, Republic

of Madagascar; and ‡Institut Pasteur,

Paris, France

References

1. Centers for Disease Control and

Prevention. Poliomyelitis—Madagascar

2002. MMWR Morb Mortal Wkly Rep

2002;51:622.

2. Guillot S, Caro V, Cuervo N, Korotkova E,

Combiescu M, Persu A, et al.] Natural

genetic exchanges between vaccine and

wild poliovirus strains in humans. J Virol

2000;74:8434–43.

3. Pringle CR. Virus taxonomy at the XIth

International Congress of Virology, Sydney,

Australia, 1999. Arch Virol

1999;144:2065–-70.

4. Pearson WR, Lipman DJ. Improved tools

for biological sequence comparison. Proc

Natl Acad Sci U S A 1988;85:2444–8.

5. World Health Organization. Vaccines,

immunization, and biologicals. Accessed

May 13, 2003. Available from: URL:

http://www.who.int/vaccines/casecount/afp

extractnew.cfm

6. Centers for Disease Control and

Prevention. Circulation of a type 2 vaccine-

derived poliovirus—Egypt, 1982–1993.

MMWR Morb Mortal Wkly Rep 2001;

50:41–2.

7. Kew O, Morris-Glasgow V, Landaverde M,

Burns C, Shaw J, Garib Z, et al. Outbreak

of poliomyelitis in Hispaniola associated

with circulating type 1 vaccine-derived

poliovirus. Science 2002;296:356–9.

8. Wood DJ, Sutter RW, Dowdle WR.

Stopping poliovirus vaccination after eradi-

cation: issues and challenges. Bull World

Health Organ 2000;78:347–57.

9. Centers for Disease Control and

Prevention. Acute flaccid paralysis associ-

ated with circulating vaccine-derived

poliovirus—Philippines, 2001. MMWR

Morb Mortal Wkly Rep 2001;50:874–5.

Address for correspondence: Mala Rakoto-

Andrianarivelo, Unité de Virologie, Institut

Pasteur de Madagascar, BP 1274, Antananarivo

101, Madagascar; fax: (261.20) 22.415.34;

email: mala@pasteur.mg

West Nile Virus

Infection in

Crocodiles

To the Editor: Recently West Nile

virus (WNV) infection has been

reported in three alligators (Alligator

sp.) from central Florida (1) and one

captive crocodile monitor (Varanus

salvadori) with neurologic signs from

the District of Columbia and

Maryland area (2). These first reports

of the virus in American reptiles high-

light the possible role of this group of

vertebrates in the WNV life cycle. To

our knowledge, WNV in a reptile was

reported only once before in a sero-

survey conducted in Israel from 1965

to 1966, in which 22 reptiles and 96

amphibians were tested for hemagglu-

tination-inhibiting antibodies against

several viruses, including WNV; one

turtle (Clemmys caspica) was

seropositive (3). Experimental infec-

tion of the lake frog (Rana ridibunda)

with a Russian strain of WNV result-

ed in high levels of viremia (4). At

present, the role of reptiles and

amphibians in the life cycle and epi-

demiology of WNV is not known.

We report, for the first time, WNV

infection in crocodiles (Crocodylus

niloticus). To assess the potential role

of crocodiles in the life cycle of WNV

in Israel, serum specimens were col-

lected from 20 healthy crocodiles on a

commercial farm in the Negev Desert,

in southern Israel (31°14′N, 34°19′E).

The crocodiles came from two sepa-

rate breeding farms (32°03′N,

35°26′E and 30°18′N, 35°07′E) in the

Syrian-African Rift Valley, which is

on the main route of bird migration

from Africa to Europe. Five males and

15 females, 1–2.5 years of age, were

examined. Blood was withdrawn from

the crocodiles’ ventral caudal vein,

separated by centrifugation, and kept

at –20°C until analyzed. Neutralizing

antibody titers were determined

against WN-goose-98 (5) and

attempts to isolate the virus were per-

formed by using Vero cell culture (6)

and by using direct reverse transcrip-

tion–polymerase chain reaction (RT-

PCR) on the serum specimens. To

eliminate the possibilities of nonspe-

cific reaction, all serum samples were

concurrently tested for the only other

flavivirus known to be present in

Israel; Israeli turkey meningo-

encephalitis virus (ITV) (7). Because

ITV does not produce cytopathic

effects (CPE) in Vero cells, virus neu-

tralization was conducted on BHK

cells for both WNV and ITV by using

WN-goose-98 and ITV (vaccine

strain). In this case, the virus stocks

(10

-4.2

50% tissue culture infective

dose) were diluted 1:400, and virus

neutralization titers were checked 3

days later.

Viral RNA was extracted from

serum samples with the QIAamp

RNA blood kit (QIAGEN, Valencia,

CA) , according to the manufacturer’s

protocol and resuspended in 30 µl of

RNase-free water. The primer pair

WN240-Kun848 (respective genome

positions 5′: 848 and 1,645) was used

to synthesize an 800-bp product in the

E gene region (8,9). The resulting

DNA fragment was visualized on

888 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

1.5% agarose gel stained with ethidi-

um bromide.

The seroprevalence rate in the first

set of virus neutralization assays in

Vero cells was 14/20 (70%, with titers

ranging from 1:20 to 1:320 [3x1:20,

3x1:40, 3x1:80, 2x1:160, 3x1:320]).

No differences were discernible in

either the seroprevalence rate or in the

average titers of crocodiles from two

different breeding farms. In BHK

cells, a similar seroprevalence rate

was observed, with titers ranging from

1:40 to 1:1,280 (3x1:40, 2x1:80,

1x1:160, 4x1:320, 3x1:640,

1x1:1280). All serum samples, except

one, were <1:10 against ITV virus,

which had a titer of 1:640 against

WNV and 1:10 against ITV. Viremia

was not detected in any of the 20 sam-

ples in Vero cell culture or by RT-PCR.

These results demonstrate a high

rate of infection with WNV in croco-

diles in Israel. The crocodiles may

have been exposed to the virus during

the summer at their present location,

since no difference in prevalence was

seen between the two groups (which

differed only in the farm of origin)

and since the younger crocodiles had

been hatched in the spring of 2002.

Furthermore, a cross-reaction with the

other prevalent flavivirus in Israel,

ITV, was ruled out. Preliminary

results from an equine seroprevalence

study (involving 800 horses over a 3-

year period) of virus neutralization

antibodies to WNV collected during

fall 2002, indicate that most horses

sampled in Israel’s Arava Valley (a

desert in the Syrian-African Rift near

the Jordanian and Egyptian borders)

and the Gulf of Aqaba/Eilat (30°59′N,

35°18′E to 29°34′N, 34°57′E) (85%,

79/90) were positive (A. Steinman

and S. Tal, unpub. data,). WNV was

also isolated from mosquitoes in the

same region (10). The seroprevalence

of WNV antibodies among horses and

local birds from the Negev Desert is

not known nor is the time when the

horses acquired WNV infection.

However, the isolation of WNV from

mosquitoes (10) and the presence of

antibodies to WNV in young croco-

diles demonstrate arboviral activity in

this region in the summer of 2002,

although clinical cases were few. That

virus was not isolated from crocodiles

in late November (past outbreaks of

WNV in Israel mainly occurred

between August and October) (6,11).

WNV has been endemic in Israel

since the early 1950s (12). More

recently, in the summer of 2000, an

extensive outbreak occurred, affecting

hundreds of people (11), dozens of

horses (6), and several flocks of geese

(5). However, no deaths of crocodiles

were reported. This contrasts with the

report from Florida (1), where WNV

was isolated from dead alligators, and

where hundreds of cases of sudden

death had been reported in previous

years; these deaths are now suspected

to result, at least in part, from WNV

disease.

The role of various reptile species

in the epidemiology of other

arboviruses such as western equine

encephalitis, eastern equine

encephalitis, and Venezuelan equine

encephalitis is well documented

(13–15). At present, the role of rep-

tiles and amphibians in the life cycle

and epidemiology of WNV is not

known, and further research is neces-

sary.

Acknowledgments

We thank Kubbi Ofer for assistance

in the collection of serum samples from

the crocodiles.

Amir Steinman,* Caroline Banet-

Noach,† Shlomit Tal,‡ Ohad Levi,*

Lubov Simanov,† Shimon Perk,†

Mertyn Malkinson,†

and Nahum Shpigel*

*Hebrew University of Jerusalem Koret

School of Veterinary Medicine, Rehovot,

Israel; †Kimron Veterinary Institute, Bet

Dagan, Israel; and ‡Tel Aviv University

Sackler Faculty of Medicine, Tel Aviv, Israel

References

1. ProMED-mail. Florida: West Nile virus

identified in alligators for the first time.

ProMED-mail 2002; 14 Nov:

20021114.5797. Available from: URL:

http://www.promedmail.org

2. Travis D, McNamara T, Glaser A,

Campbell R. A national surveillance system

for WNV in zoological institutions.

Available from: URL: http://www.cdc.gov/

ncidod/dvbid/westnile/conf/ppt/1a-travis.

ppt

3. Nir Y, Lasowski Y, Avivi A, Goldwasser R.

Survey for antibodies to arboviruses in the

serum of various animals in Israel during

1965–1966. Am J Trop Med Hyg

1969;18:416–22.

4. Kostiukov MA, Gordeeva ZE, Bulychev

VP, Nemova NV, Daniiarov OA. The lake

frog (Rana ridibunda)—one of the food

hosts of blood-sucking mosquitoes in

Tadzhikistan—a reservoir of the West Nile

Dejà Vu

"I must have made this molecule before;

It is familiar to the core!"

Said a yeast cell emerging from mitosis,

With no experience yet in synthesis.

Then, guided by a transmigrant human gene,

It assembled that "alien" protein.

Boghos L. Artinian

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 889

LETTERS

fever virus. Med Parazitol (Mosk)

1985;3:49–50.

5. Malkinson M, Banet C, Weisman Y,

Pokamunski S, King R, Drouet MT, et al.

Introduction of West Nile virus in the

Middle East by migrating white storks.

Emerg Infect Dis 2002;8:392–7.

6. Steinman A, Banet C, Sutton GA, Yadin H,

Hadar S, Brill A. Clinical description of

equine West Nile encephalomyelitis during

the outbreak of 2000 in Israel. Vet Rec

2002;151:47–9.

7. Ianconescu M. Turkey meningoen-

cephalitis: a general review. Avian Dis

1976;20:135–8.

8. Berthet FX, Zeller HG, Drouet MT, Rauzier

J, Digoutte JP, Deubel V. Extensive

nucleotide changes and deletions within the

envelope glycoprotein gene of Euro-

African West Nile viruses. J Gen Virol

1997;78:2293–7.

9. Savage HM, Ceianu C, Nicolescu G,

Karabatsos N, Lanciotti RS, Vladimirescu

A, et al. Entomologic and avian investiga-

tions of an epidemic of West Nile fever in

Romania in 1996, with serologic and

molecular characterization of a virus isolate

from mosquitoes. Am J Trop Med Hyg

1999;61:600–11.

10. Ministry of the Environment, State of

Israel. Available from: URL: http://

www.sviva.gov.il

11. Weinberger M, Pitlik SD, Gandacu D, Lang

R, Nassar F, Ben David D, et al. West Nile

fever outbreak, Israel, 2000: epidemiologic

aspects. Emerg Infect Dis 2001;7:686–91.

12. Bernkopf H, Levine S, Nerson R. Isolation

of West Nile virus in Israel. J Infect Dis

1953;93:207–18.

13. Bowen G.S. Prolonged western equine

encephalitis viremia in the Texas tortoise

(Gopherus berlandieri). Am J Trop Med

Hyg 1977;26:171–5.

14. Thomas LA, Eklund CM, Rush WA.

Susceptibility of garter snakes

(Thamnophis spp.) to western equine

encephalomyelitis virus. Proc Soc Exper

Biol Med 1958;99:698–700.

15. Walder R, Suarez OM, Calisher CH.

Arbovirus studies in the Guajira region of

Venezuela: activities of Eastern equine

encephalitis and Venezuelan equine

encephalitis viruses during an interepizoot-

ic period. Am J Trop Med Hyg

1984;33:699–707.

Address for correspondence: Amir Steinman,

Koret School of Veterinary Medicine, Hebrew

University of Jerusalem, P.O.B. 12, Rehovot

76100, Israel; fax: 972-3-9688539; email:

[email protected]

Rickettsia

aeschlimannii in

Spain: Molecular

Evidence in

Hyalomma

marginatum and

Five Other Tick

Species that Feed

on Humans

To the Editor: Rickettsia aeschli-

mannii is a pathogenic spotted fever

group rickettsia first isolated from

Hyalomma marginatum ticks collect-

ed in Morocco in 1997 (1). Later

found in H. marginatum ticks from

Zimbabwe, Niger, Mali, and Portugal

(2), R. aeschlimannii has also been

found in a Rhipicephalus appendicu-

latus tick attached to the right thigh of

a patient in South Africa (3). These

data suggest a broad geographic dis-

tribution for R. aeschlimannii and the

possibility that tick species other than

H. marginatum may also be suitable

vectors for this rickettsia.

The pathogenicity of Rickettsia

aeschlimannii in humans has been

demonstrated by Raoult et al. (4) in a

French patient who became ill after

returning from a trip to Morocco. The

patient exhibited symptoms similar to

those of Mediterranean spotted fever

(MSF) produced by R. conorii, with a

tache noire–like eschar on his ankle,

fever (39.5°C), and a generalized

maculopapular skin rash. The second

documented and most recent case of

human infection by R. aeschlimannii

occurred in a South African man who

was bitten by R. appendiculatus; an

eschar also developed around the tick

attachment site on this patient (3). He

was aware of the risk for tick-trans-

mitted disease, so he removed the tick

and administered doxycycline; he did

not develop additional symptoms.

Over the past 6 years, throughout

the region of Castilla y León, north-

western Spain, we have collected and

identified 3,059 ticks belonging to 15

species (unpub. data) that were

attached to persons living in this terri-

tory. We have systematically analyzed

the ticks by polymerase chain reaction

(PCR) to detect those infected with

Borrelia burgdorferi, Anaplasma

phagocytophila, and Rickettsia spp.

This procedure enabled us to identify,

for the first time in Spain, R. aeschli-

mannii in 35 tick specimens belong-

ing to H. marginatum and to another

five species.

During the 6-year study, ticks

found on patients who sought medical

advice in the hospitals and healthcare

centers of Castilla y León were

removed and referred to our laborato-

ry for identification and analysis. Each

tick was first disinfected by immersion

in 70% alcohol, rinsed in sterile water,

and dried on sterile filter paper. We

then extracted DNA in 5% Chelex-

100, according to the method of

Guttman et al. (5). In searching for

Rickettsia spp., we proceeded as

described by La Scola and Raoult (6):

All DNA samples were first tested for

a fragment of the rickettsial gltA gene

(7), and then, in the gltA-positive sam-

ples, a fragment of the rickettsial

ompA gene (8) was amplified,

sequenced, and compared with gene

databases for identification. The gltA

amplicon was sequenced only when

the ompA was not successfully ampli-

fied. To prevent DNA contamination

and the carryover of amplified prod-

ucts, we used sterile tools at all times

and carried out each step of the analy-

sis (extracting DNA, preparing the

reaction mixture, and amplifying and

analyzing the PCR product) in separat-

ed work areas. Two negative controls

(Milli-Q water and DNA from labora-

tory-reared noninfected ticks) were

included in each amplification trial.

These controls were never amplified.

We obtained 21 ompA amplicons

(629–632 bp) from 21 ticks. One

890 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

amplicon, from a Haemaphysalis

punctata tick, had 100% sequence

identity with the ompA of R. aeschli-

mannii (GenBank accession no.

U43800). The nucleotide sequences

of the remaining 20 ompA amplicons

shared >99% similarity with the

ompA of R. aeschlimannii. These 20

amplicons were obtained from nine

Hyalomma marginatum, five

Rhipicephalus bursa, three R. turani-

cus, one R. sanguineus, and two

Ixodes ricinus ticks.

In an additional 14 ticks (10 H.

marginatum, 2 R. bursa, 1 R. san-

guineus, and 1 I. ricinus) we

sequenced 14 gltA amplicons (382

bp), which were 100% identical to the

gltA of R. aeschlimannii (GenBank

accession no. U59722). No other tick-

borne pathogens were detected in the

35 R. aeschlimannii–containing ticks.

R. aeschlimannii has never been

detected in Spain; therefore, our study

constitutes its first citation in this

country. Because R. aeschlimannii

had been already detected in ticks in

Portugal, we believe that its presence

in Spain was expected and our finding

is not surprising. However, the high

number of tick species in which we

found this rickettsia was unexpected:

six species belonging to four genera.

For these six species, the ratio

between the specimens infected and

the specimens analyzed (as well as the

infection rates) were as follows: I.

ricinus (3/1,320; 0.23%), H. mar-

ginatum (19/324; 5.86%), H. punctata

(1/106; 0.94%), R. bursa (7/425,

1.64%), R. sanguineus (2/102;

1.96%), and R. turanicus (3/330;

0.91%). Although H. marginatum was

the fourth most anthropophilic species

in our study, this species simultane-

ously showed the highest number of

infected specimens and the highest

infection rate, making H. marginatum

the main vector of R. aeschlimannii in

our region. The next most important

vectors are Rhipicephalus spp., and in

particular R. bursa.

The 35 R. aeschlimannii–positive

ticks were removed in the first 6 to 12

hours after attachment, before they

could have ingested any blood, thus

indicating that they were previously

infected with the bacterium. Persons

bitten by these specimens had never

had symptoms of spotted fever, and

they remained asymptomatic after the

bite, suggesting that, as expected

because of the rapid removal of the

ticks, they did not acquire the infection.

Although MSF is endemic in

Castilla y León (9), we only found

one tick infected with R. conorii

(0.03%) among the 3,059 analyzed

(unpub. data), whereas R. aeschliman-

nii was much more prevalent in these

same ticks (1.14%). Hence, in accor-

dance with what was proposed by

Raoult et al. (4) for MSF cases in

Morocco, we suspect that many cases

of MSF in Castilla y León may really

have been due to R. aeschlimannii.

Our findings show that R. aeschli-

mannii is present in Castilla y León,

the largest region in Spain, in six tick

species that frequently feed on

humans. Our observations expand the

geographic distribution of this bac-

terium and the range of its potential

tick vectors.

Acknowledgments

We thank Rufino Álamo-Sanz for his

invaluable contributions and N. Skinner

for revising the English version of the

manuscript.

This work was supported by the

Consejería de Sanidad y Bienestar Social

of the Junta de Castilla y León (Spain).

Pedro Fernández-Soto,*

Antonio Encinas-Grandes,*

and Ricardo Pérez-Sánchez†

*Universidad de Salamanca, Salamanca,

Spain; and †Instituto de Recursos

Naturales y Agrobiología, Consejo Superior

de Investigaciones Científicas, Salamanca,

Spain

References

1. Beati L, Meskini M, Thiers B, Raoult D.

Rickettsia aeschlimannii sp. nov., a new

spotted fever group rickettsia associated

with Hyalomma marginatum ticks. Int J

Syst Bacteriol 1997;47:548–54.

2. Parola P, Raoult D. Ticks and tickborne

bacterial diseases in humans: an emerging

infectious threat. Clin Infect Dis

2001;32:897–928.

3. Pretorius A, Birtles RJ. Rickettsia aeschli-

mannii: a new pathogenic spotted fever

group rickettsia, South Africa. Emerg Infect

Dis 2002;8:874.

4. Raoult D, Fournier PE, Abboud P, Caron F.

First documented human Rickettsia aeschli-

mannii infection. Emerg Infect Dis

2002;8:748–9.

5. Guttman D S, Wang PW, Wang IN, Bosler

EM, Luft BJ, Dykhuizen DE. Multiple

infections of Ixodes scapularis ticks by

Borrelia burgdorferi as revealed by single-

strand conformation polymorphism analy-

sis. J Clin Microbiol 1996;34:652–6.

6. La Scola B, Raoult D. Laboratory diagnosis

of rickettsioses: current approaches to diag-

nosis of old and new rickettsial diseases. J

Clin Microbiol 1997;35: 2715–27.

7. Regnery RL, Spruil CL, Plikaytis BD.

Genotypic identification of Rickettsiae and

estimation of intraspecies sequence diver-

gence for portions of two rickettsial genes.

J Bacteriol 1991;173:1576–89.

8. Roux V, Furnier PE, Raoult D.

Differentiation of spotted fever group rick-

ettsiae by sequencing and analysis of

restriction fragment length polymorphism

of PCR-amplified DNA of the gene encod-

ing the protein rOmpA. J Clin Microbiol

1996;34:2058–65.

9. Boletines Epidemiológicos de Castilla y

León 1997–2002. (Spain): 2000;16:9–12.

Address for correspondence: Ricardo Pérez-

Sánchez, Departamento de Patología Animal,

Instituto de Recursos Naturales y Agrobiología,

Consejo Superior de Investigaciones

Científicas, C/ Cordel de Merinas, 40-52,

37007 Salamanca, Spain; fax: +34 923 219609;

email: [email protected]

The opinions expressed by authors con-

tributing to this journal do not necessari-

ly reflect the opinions of the Centers for

Disease Control and Prevention or the

institutions with which the authors are

affiliated.

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 891

LETTERS

Hantaviruses in

São Paulo State,

Brazil

To the Editor: Hantavirus pul-

monary syndrome (HPS) is an emerg-

ing health problem in Brazil. This

syndrome was first reported in 1993

in three persons living in a rural area

of Juquitiba County; two of them died

of acute respiratory failure (1).

Although Juquitiba County is part of

the metropolitan area of greater São

Paulo City, patients lived in a recently

deforested region. From 1993 through

2002, approximately 200 HPS cases

were reported in Brazil, with a 40%

case-fatality ratio (Ministry of Health

of Brazil, Report on Hantavirus cases

1993–2002, unpub. data).

The wild rodent Bolomys laziurus

is believed to be the most important

hantavirus reservoir in the State of

São Paulo, based on high levels of

specific antibodies observed in serum

from captured specimens (L.E.

Pereira, Adolpho Lutz Institute, pers.

comm., 2001). The economy of the

inland region of Ribeirão Preto in the

State of São Paulo, with its 3.5 million

inhabitants, is based on the sugar cane

agroindustry. The region has been

almost completely deforested, with

important consequences to the envi-

ronment and wild rodent ecology.

Twenty HPS cases were reported in

Ribeirão Preto in the last 5 years, with

a 60% case-fatality ratio. Review of

medical records showed that a pro-

dromic fever occurred in all 14 case-

patients studied; dyspnea, cough,

hypotension, and tachycardia

occurred in about two thirds of

patients; and hemorrhagic phenomena

(hematuria, melena, and hypermenor-

rhea) in about one third.

Thrombocytopenia was observed in

all the patients, elevated hematocrit in

about three fourths, and leukocytosis

with neutrophilia and a left shift in the

differential count in about two thirds.

Serum creatinine levels were also

increased (average level 2 mg/dL).

Chest radiographs showed diffuse

alveolar flocculant infiltrates in most

cases (2,3). Laboratory diagnosis of

HPS was made by serologic testing

(enzyme-linked immunosorbent assay

[ELISA]) in 18 cases and by reverse

transcription–polymerase chain reac-

tion (RT-PCR) in 11 cases; for 7 cases,

both techniques were used. We per-

formed a nucleotide sequence analy-

sis of the N gene of hantavirus

(residues 236–477) obtained from the

blood of 11 of the 20 patients. This

analysis showed that the infections

were caused by Araraquara virus, a

previously known hantavirus that had

been detected by RT-PCR in the

serum of an HPS patient living in a

nearby county (4). Thus, Araraquara

virus is the causative agent of a severe

form of HPS, with a high death rate.

This high death rate could also be

related to the lack of adequate initial

therapy provided by clinicians who

probably did not immediately suspect

HPS and may have not recommended

hospitalization in intensive-care units.

In addition, some hospitalized

patients were in shock when first seen

and were rehydrated with massive

quantities of fluids, which may have

aggravated pulmonary edema and

contributed to death.

The occurrence of 10% of the

Brazilian HPS reported cases in

Ribeirão Preto indicates that this

region is suitable for studying the epi-

demiology of hantavirus infections. A

serologic survey conducted in the

region in 1999, which included 567

primary-care patients from Ribeirão

Preto, Guariba, and Jardinópolis

Counties, found that 7 (1.23%) of

them had immunoglobulin (Ig) G

antibodies to Sin Nombre virus by

ELISA and that 5 of those lived in

Jardinópolis (population 30,000), a

county where a fatal case of HPS

occurred in 1999 (5). Thus,

Jardinópolis County was chosen for a

population-based survey. In May

2001, we obtained personal informa-

tion and collected fingerprick blood

samples from 818 participants, 15–70

years of age, living in urban and rural

areas of the county. IgG antibodies to

the N recombinant protein of Andes

virus were detected by ELISA in the

blood samples of 14.3% of the partic-

ipants (5). Even though all HPS cases

in Ribeirão Preto were associated

with rural activity and rodent expo-

sure, these serologic data suggest that

hantavirus infections are common in

Jardinópolis County, independent of

sex, profession, or history of contact

with rodents. None of the 14.3% par-

ticipants with IgG antibodies to han-

tavirus had a history of HPS-like dis-

ease, and the ELISA test showed

cross-reactions with most of the South

American hantaviruses, including

Araraquara. Persons living in the

urban area had higher levels of anti-

bodies to hantavirus than those from

rural areas. In Ribeirão Preto, the

physical boundaries of cities have

expanded to incorporate other areas,

encroaching upon rural areas with

many popular subsidized housing

complexes. Work-related and recre-

ational rural activities in that region

are also frequent, which makes it dif-

ficult to interpret these data. These

results suggest that in this region of

southeast Brazil, hantaviruses may be

causing undiagnosed asymptomatic or

clinically minor infections in addition

to typical HPS. This finding envokes

important questions. Is more than one

hantavirus circulating in this region,

causing mostly benign infections? Is

Araraquara virus widespread, causing

mostly inapparent infections and only

rarely causing HPS? Would HPS be

associated with some predisposing

condition in the infected person? If

more than one hantavirus is circulat-

ing in the region, could urban rodents

be reservoirs?

Further studies are necessary to

better understand the epidemiology

and clinical signs and symptoms of

hantavirus infection in the region of

Ribeirão Preto. Such studies should

892 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

emphasize determining the reservoirs,

the modes of virus transmission to

people, and the possible distinct clini-

cal forms of hantavirus infections.

Acknowledgments

We thank the employees of the

Health Bureau of the County of

Jardinópolis for the dedicated collabora-

tion on the serologic survey and Paula

Padula for supplying antigen for enzyme-

linked immunoabsorbent assays.

This work was supported by the State

of São Paulo Research Foundation

(FAPESP).

Luiz T.M. Figueiredo,*

Marcos L. Moreli,*

Gelse M. Campos,*

and Ricardo L.M. Sousa*

*University of São Paulo, São Paulo, Brazil

References

1. Iversson LB. Hantavirus pulmonary

syndrome in the rural area of Juquitiba, São

Paulo Metropolitan Area, Brazil. Rev Inst

Med Trop Sao Paulo 1997;39:237–8.

2. Figueiredo LTM, Moreli ML, Kashima S,

Almeida VSO, Félix PR, Bruno JC, et al.

Hantavirus pulmonary syndrome (HPS) in

Guariba, SP, Brazil. Report of 2 cases. Rev

Inst Med Trop Sao Paulo 1999;41:131–7.

3. Figueiredo LTM, Campos GM, Rodrigues

FB. Síndrome pulmonar e cardiovascular

por hantavirus: aspectos epidemiológicos,

clínicos, do diagnóstico laboratorial e do

tratamento. [in Portuguese]. Rev Soc Bras

Med Trop 2001;34:15–27.

4. Johnson AM, de Souza LT, Ferreira IB,

Pereira LE, Ksiazek TG, Rollin PE, et al.

Genetic investigation of novel hantaviruses

causing fatal HPS in Brazil. J Med Virol

1999;59:527–35.

5. Holmes RR, Boccanera R, Figueiredo

LTM, Mançano SR, Pane C.

Seroprevalence of human hantavirus

infection in Ribeirão Preto Region of São

Paulo State, Brazil. Emerg Infect Dis

2000;6:560–1.

Address for correspondence: Luiz Tadeu

Moraes Figueiredo, Faculdade de Medicina de

Ribeirão Preto–USP, Av. Bandeirantes 3900,

Ribeirão Preto, SP, 14049-900, Brazil; fax: 55

16 6336695; email: [email protected]

Israeli Spotted

Fever Rickettsia

in Sicilian

Rhipicephalus

sanguineus Ticks

To the Editor: Mediterranean

spotted fever (MSF) is endemic in

Italy, where it is a reportable disease.

From 1992 to 1998, the Italian

Ministry of Health was notified of

approximately 8,500 cases of human

rickettsioses presumed to be MSF.

MSF occurs more commonly in some

central (Lazio) and southern

(Sardinia, Sicily, and Calabria)

regions (1,2); in 1998, an average of

8.8 cases occurred for every 100,000

persons in Sicily, compared with the

national average of 1.6 cases per

100,000 persons. Rickettsia conorii

has been thought to be the only path-

ogenic Rickettsia of the spotted fever

group in Sicily (3,4) or the western

Mediterranean area. Recently, three

different spotted fever group rickettsi-

ae, including R. helvetica, were

detected in Ixodes ricinus ticks from

central and northern Italy. This find-

ing suggests that bacteria other than

R. conorii are involved in rickettsial

diseases in Italy (5).

To investigate whether unusual

tick-transmitted rickettsiae are also

present in Sicily, we used molecular-

sequence–based identification tech-

niques to study two strains isolated

from the hemolymph of

Rhipicephalus sanguineus ticks col-

lected in 1990 in western Sicily. These

isolates had been previously identi-

fied by serologic tests as belonging to

the spotted fever group rickettsiae. We

obtained bacterial DNA and per-

formed polymerase chain reaction

(PCR) for ompA gene and restriction

analysis under conditions previously

described by Roux et al. (6). Our

observation of a peculiar PstI profile

allowed a presumptive identification

of one of the two tick isolates as

belonging to the Israeli spotted fever

rickettsiae, while the other showed a

restriction profile corresponding to

that of R. conorii strain Seven. To

confirm the identification of the

Israeli spotted fever Rickettsia isolate,

we sequenced the PCR-amplified

fragment of ompA gene (MWG-

Biotech AG, Ebersberg, Germany)

and aligned sequence data with

homologous sequences of reference

strains of the spotted fever group rick-

ettsiae retrieved from the GenBank

database. Sequence analysis showed

100% similarity with the homologous

sequence of Israeli spotted fever

Rickettsia reference strain ISTT

CDC1 (GenBank accession no.

U43797). The Israeli spotted fever

Rickettsia belongs to the R. conorii

complex (7,8) and was first isolated in

1974 from ticks and humans. Initially,

Israeli spotted fever rickettsiae distri-

bution appeared to be restricted to

Israel (9), but more recently the

organism has also been isolated from

patients with MSF in Portugal (10).

Our finding of Israeli spotted fever

Rickettsia infection in a R. sanguineus

tick, the main vector for MSF in

Sicily, also suggests that the geo-

graphic distribution of Israeli spotted

fever might be wider than previously

thought, including not only Israel and

the Iberian Peninsula but also Italy.

Molecular analysis of spotted

fever group Rickettsia isolates from

Sicilian MSF patients is under way to

verify this hypothesis. Because initial

signs and symptoms of Israeli spotted

fever are particularly uncharacteristic,

awareness of the presence of Israeli

spotted fever Rickettsia in our geo-

graphic area may hasten provision of

the appropriate treatment. The

Sicilian ompA gene sequence

described in this study has been

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 893

LETTERS

deposited in the GenBank database

(accession no. AY197565).

Giovanni M. Giammanco,*

Serafino Mansueto,*

Pietro Ammatuna,*

and Giustina Vitale*

*Università di Palermo, Palermo, Italy

References

1. Scaffidi V. Current endemic expansion of

boutonneuse fever in Italy. Minerva Med

1981;72:2063–70.

2. Tringali G, Intonazzo V, Perna AM,

Mansueto S, Vitale G, Walker DH.

Epidemiology of Boutonneuse fever in

western Sicily. Distribution and prevalence

of spotted fever group rickettsial infection

in dog ticks (Rhipicephalus sanguineus).

Am J Epidemiol 1986;123:721–7.

3. Mansueto S, Vitale G, Lavagnino A, Di

Rosa S, Merulla R. Rickettsiae of the spot-

ted fever group in dog fleas

(Ctenocephalides spp.) in western Sicily.

Ann Trop Med Parasitol 1989;83:325.

4. Vitale G, Di Stefano R, Damiani G,

Mansueto S. Characterization of Sicilian

strains of spotted fever group rickettsiae by

using monoclonal antibodies. J Clin

Microbiol 1989;27:1081–5.

5. Beninati T, Lo N, Noda H, Esposito F,

Rizzoli A, Favia G, et al. First detection of

spotted fever group Rickettsiae in Ixodes

ricinus from Italy. Emerg Infect Dis

2002;8:983–6.

6. Roux V, Fournier PE, Raoult D.

Differentiation of spotted fever group

Rickettsiae by sequencing and analysis of

restriction fragment length polymorphism

of PCR-amplified DNA of the gene

encoding the protein rOmpA. J Clin

Microbiol 1996;34:2058–65.

7. Regnery RL, Spruill CL, Plikaytis BD.

Genotypic identification of Rickettsiae and

estimation of interspecies sequence diver-

gence for portions of two rickettsial genes.

J Bacteriol 1991;173:1576–89.

8. Fournier PE, Roux V, Raoult D.

Phylogenetic analysis of spotted fever

group Rickettsiae by study of the outer sur-

face protein rOmpA. Int J Syst Bacteriol

1998;48:839–49.

9. Roux V, Raoult D. Phylogenetic analysis

and taxonomic relationships among the

genus Rickettsia. In: Raoult D, Brouqui P,

editors. Rickettsiae and rickettsial diseases

at the turn of the third millennium. Paris:

Elsevier; 1999. p. 52–66.

10. Bacellar F, Beati L, Franca A, Pocas J,

Regnery R, Filipe A. Israeli spotted fever

rickettsia (Rickettsia conorii complex)

associated with human disease. Emerg

Infect Dis 1999;5:835–6.

Address for correspondence: Giovanni M.

Giammanco, Department of Hygiene and

Microbiology, Università di Palermo, via del

Vespro 133, I-90127, Palermo, Italy; fax: +39

0916553676; email: [email protected]

Co-feeding

Transmission and

Its Contribution to

the Perpetuation of

the Lyme Disease

Spirochete

Borrelia afzelii

In Reply: Richter et al. (1) have

asked an important question: To what

extent does the transmission of non-

systemic infections of the Lyme borre-

liosis spirochete (Borrelia afzelii)

between co-feeding nymphal and lar-

val Ixodes ricinus ticks apply to natu-

ral tick infestations on wild rodents?

The authors conclude that the trans-

mission of infections 3 days after inoc-

ulation by tick bite is >100 times less

efficient than the transmission of infec-

tions that have lasted at least 14 days.

That answer depends on a critical cal-

culation based on experimental results

combined with field observations.

Unfortunately, this calculation is incor-

rect by a factor of approximately 20.

When hairless laboratory mice

were restrained within wire mesh

tubes and larvae were allowed to

attach at random over their bodies,

13.6% of these larvae became infect-

ed with B. afzelii if they fed 3 days

after the attachment of a single infect-

ed nymph (i.e., transmission probabil-

ity of 0.136, as used below). By con-

trast, 85.4% of larvae that fed 14 days

after the nymph became infected (1).

At three sites in Germany and France,

over the period April–October in each

of the years from 1993 through 1995,

17.6% of mice (Apodemus flavicollis

and A. sylvaticus) and voles

(Clethrionomys glareolus) fed larval

and nymphal ticks together, while

1.5% fed nymphs alone. Of these

nymphs, 26.4% were infected with B.

burgdorferi s.l. before attachment.

The probability of a larva’s acquiring

an infection equals the product of 1)

the probability of transmission from

host to larva and 2) the probability of

the host’s being infected, while the

larva feeds, via an infected nymphal

tick bite. For a short-lived (3-day)

infection, the probability is 0.136 ×

0.176 × 0.264 = 0.0063; for longer-

lived (14-day) infections, the proba-

bility is 0.854 × (0.176 + 0.015) ×

0.264 = 0.0431. The ratio is therefore

1:6.8. Richter et al. erroneously con-

cluded that the ratio was 1:116

because they did not take into account

the probability of wild rodents’

acquiring a long-lived, ”systemic”

infection; the authors assumed the

probability was 1. A greater propor-

tion of garden dormice (Eliomys

quercinus) carried ticks and so would

yield much higher transmission prob-

abilities but in almost the same ratio,

1:6.4.

In fact, how much of the increase

from 13.6% transmission at day 3 to

85.4% at day 14 was due to the devel-

opment of systemic infections (i.e.,

disseminated to parts of the hosts’

bodies >2 cm from the infected tick

bite) is not clear because the feeding

sites of the larvae attached ad libitum

on the hairless mice were not report-

ed. In the original discovery of co-

feeding transmission of B. burgdor-

feri s.l. (2), the infection prevalence in

larvae feeding close to infected

nymphs increased from 33% on day 2

to 96% on day 11 and 100% on day 14

(3; see Figure 2 therein) in the demon-

strated absence of a systemic infec-

tion. Mice skin and ticks feeding at

distant sites remained uninfected.

Only after day 14 had a systemic

894 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

infection developed (2). Because

spirochetes are not transmitted to the

host until at least 17.6 h after an

infected nymph starts feeding (4–6)

and then disseminate only slowly

from the feeding site (7), co-feeding

in space rather than in time is the cru-

cial feature in Lyme borreliosis (2,8)

(so-called “extended co-feeding”[3]).

Larvae that attach to hosts simultane-

ously with infected nymphs rarely

acquire spirochetes (1,9), wherever

they attach. This pattern is distinct

from the more immediate and short-

lived co-feeding transmission of tick-

borne viruses (10–12). In both cases,

however, the key feature is a nonsys-

temic infection.

Despite the uncertainties in Richter

et al.’s study, their corrected ratio is

very similar to that (1:5.7) calculated

(3) with a “synthetic model . . . based

on major assumed parameters” (sic)

(1). In that model we assumed that

50% of larvae were likely to be feed-

ing within 1 to 2 cm of any infected

nymph, the distance over which co-

feeding ticks can pick up nonsystemic

infections (1,2), because in the wild

very few rodents carry nymphs in the

absence of larvae (1,13), and >95% of

all immature stage ticks feed in aggre-

gations, mostly on the ears and also

around the eyes or on the snouts of

mice and voles. Considerable risks

exist in using laboratory experimental

results to quantify the epidemiologic

importance of nonsystemic infections

in the wild because of differences

between host species, unnatural spa-

tial distributions of introduced ticks

on hosts, and the subtleties of natural

tick-host relationships. Coincident

aggregated distributions of larvae and

nymphs among their rodent hosts,

whereby the same individual hosts

carry the largest numbers of both

stages, increase the number of larvae

co-feeding with any infected nymph,

and so augment the potential amplifi-

cation of infection prevalence in ticks

(13). Nevertheless, in the case of

rodents, nonsystemic infections are

soon rendered redundant by the much

longer lived systemic infections. In

contrast, in the case of host species in

which systemic infections do not

develop, the transmission of nonsys-

temic infections between co-feeding

ticks is the only way in which infec-

tion prevalence can be amplified in

feeding ticks. Field data suggest that

this route of transmission occurs on

wild Sika deer (Cervus nippon) (14).

Natural experimental systems have

confirmed that on sheep this transmis-

sion pathway exists and is sufficient

alone to maintain enzootic cycles of

Lyme borreliosis (8).

Sarah Randolph* and Lise Gern†

*University of Oxford, Oxford, United

Kingdom, and †Université de Neuchâtel,

Neuchâtel, Switzerland

References

1. Richter D, Allgöwer R, Matuschka F-R.

Co-feeding transmission and its contribu-

tion to the perpetuation of the Lyme disease

spirochete Borrelia afzelii. Emerg Infect

Dis 2002;8:1421–5.

2. Gern L, Rais O. Efficient transmission of

Borrelia burgdorferi between cofeeding

Ixodes ricinus ticks (Acari: Ixodidae). J

Med Entomol 1996;33:189–92.

3. Randolph SE, Gern L, Nuttall PA. Co-feed-

ing ticks: epidemiological significance for

tick-borne pathogen transmission. Parasitol

Today 1996;12:472–9.

4. Kahl O, Janetzki-Mittmann C, Gray JS,

Jonas R, Stein J, de Boer R, et al. Risk of

infection with Borrelia burgdorferi sensu

lato for a host in relation to the duration of

nymphal Ixodes ricinus feeding and the

method of tick removal. Zentbl Bakt

1998;287:41–52.

5. Crippa M, Rais O, Gern L. Investigations

on the mode and dynamics of transmission

and infectivity of Borrelia burgdorferi ss

and Borrelia afzelii in Ixodes ricinus ticks.

Vector-borne and Zoonotic Diseases

2002;2:3–9.

6. Piesman J. Dispersal of the Lyme disease

spirochete Borrelia burgdorferi to salivary

glands of feeding nymphal Ixodes scapu-

laris (Acari: Ixodidae). J Med Entomol

1995;32:519–21.

7. Shih C-M, Pollack RJ, Telford SR,

Spielman A. Delayed dissemination of

Lyme disease spirochetes from the site of

deposition in the skin of mice. J Infect Dis

1992;166:827–31.

8. Ogden NH, Nuttall PA, Randolph SE.

Natural Lyme disease cycle maintained via

sheep by co-feeding ticks. Parasitology

1997;115:591–9.

9. Piesman J, Happ CM. The efficacy of co-

feeding as a means of maintaining Borrelia

burgdorferi: a North American model sys-

tem. J Vector Ecol 2001;26:216–20.

10. Jones LD, Davies CR, Steele GM, Nuttall

PA. A novel mode of arbovirus transmis-

sion involving a nonviraemic host. Science

1987;237:775–7.

11. Alekseev AN, Chunikhin SP. Exchange of

tick-borne encephalitis virus between

Ixodidae simultaneously feeding on the ani-

mals with sub-threshold levels of viraemia.

Med Parazitol Parazit Bolezni

1990;2:48–50.

12. Labuda M, Jones LD, Williams T,

Danielova V, Nuttall PA. Efficient transmis-

sion of tick-borne encephalitis virus

between cofeeding ticks. J Med Entomol

1993;30:295–9.

13. Randolph SE, Miklisová D, Lysy J, Rogers

DJ, Labuda M. Incidence from coinci-

dence: patterns of tick infestations on

rodents facilitate transmission of tick-borne

encephalitis virus. Parasitology 1999;

118:177–86.

14. Kimura K, Isogal E, Isogal H, Kamewaka

Y, Nishikawa T, Ishii N, et al Detection of

Lyme disease spirochetes in the skin of nat-

urally infected wild Sika deer (Cervus nip-

pon yesoensis) by PCR. Appl Environ

Microbiol 1995;61:1641–2.

Address for correspondence: Sarah Randolph,

University of Oxford, South Parks Road,

Oxford OX1 3PS, UK; fax: +44 1865 310447;

email: sarah.randolph@zoology.ox.ac.uk

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 895

LETTERS

Co-feeding

Transmission and

Its Contribution to

the Perpetuation of

the Lyme Disease

Spirochete

Borrelia afzelii

In Reply to Randolph and Gern:

Although transmission between co-

feeding vector ticks may perpetuate

particular tick-borne viruses, this

mode of transmission plays no role in

the epizootiology of Lyme disease

spirochetes (1,2). In their letter,

Randolph and Gern defend their sug-

gestion that tick-borne pathogens per-

petuate effectively by direct passage

from one feeding tick to another by

criticizing our analysis (3). These

researchers mainly address our com-

parison of the transmission efficiency

between simultaneously feeding ticks

with that between ticks feeding

sequentially on a persistently infected

rodent. Our experiments demonstrate

that approximately six times as many

larvae (85.4%) acquire Borrelia

afzelii spirochetes from a systemical-

ly infected mouse than from a mouse

on which an infected nymph is feed-

ing simultaneously (13.6%) (1). In

nature, however, larval ticks rarely

co-feed with nymphs on mice or

voles; only approximately one fifth

(18.8%) of these hosts harbor both

subadult stages simultaneously. And

of the nymphs, only approximately

one quarter (26.4%) are infected by

Lyme disease spirochetes. As a result,

the natural transmission efficiency

between simultaneously feeding ticks

would be only one twentieth (5%) of

that observed in the laboratory.

Multiplying the experimentally

observed efficiency of co-feeding

transmission (13.6%) by the likeli-

hood of larval and nymphal ticks co-

infesting small rodents, as well as by

the prevalence of infected nymphal

ticks, reduces the efficiency of co-

feeding transmission in nature to

<1%. Although Randolph and Gern

commit several minor mathematical

errors, their calculations support our

argument that few larval vector ticks

would acquire spirochetal infection

directly from an infected nymph (3).

Randolph and Gern err, however,

by applying the same mathematical

modifications to the transmission effi-

ciency by which larvae acquire spiro-

chetes from a persistently infected

host (3). Whereas the efficiency of co-

feeding transmission observed in the

laboratory must be modified to pay

tribute to the rare event of larvae co-

feeding with an infected nymph in

nature, the efficiency by which larvae

acquire infection from a persistently

infected host is independent of such

limiting parameters. Because a com-

petent rodent host remains infectious

to larval ticks throughout its life, the

proportion of hosts infested by partic-

ular subadult stages of the vector is

irrelevant. Thus, the transmission effi-

ciency on a persistently infected host

is unchanged in the laboratory and the

field. Almost 85.4% of larvae feeding

on mice or voles in nature would,

therefore, acquire spirochetal infec-

tion—far more than by co-feeding.

We are correct in stating that natural

896 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

LETTERS

transmission by sequentially feeding

ticks is more efficient than transmis-

sion between co-feeding ticks.

Randolph and Gern suggest that

we could have recorded the distance

between the feeding ticks to clarify

whether the increase from a 13.6%

transmission efficiency between co-

feeding ticks to a transmission effi-

ciency of 85.4% from a persistently

infected host is associated with the

development of a systemic infection.

Our experimental observation (Table

1 in our article [1]), as well as a study

on the movement of spirochetes

through their host’s skin (4), conclu-

sively demonstrates that the increase

in transmission efficiency is due to the

progressive dissemination of spiro-

chetes from the site of inoculation.

The likelihood of a larva’s acquiring

spirochetes from any site of its host’s

skin increases with the passage of

time since the infected nymph

attached. To compare the two modes

of transmission in terms of efficiency

(Table 2 in our article [1]), we permit-

ted the larvae to attach randomly to

their rodent hosts, just as they would

do in nature.

In the epizootiology of Lyme dis-

ease spirochetes, “simultaneous”

transmission between co-feeding ticks

(<1%) is some two orders of magni-

tude less efficient than sequential

transmission between ticks feeding on

persistently infected reservoir rodents

(85.4%). The two studies that relied

on natural infestation densities and

refrained from using artificial feeding

capsules conclusively demonstrated

that transmission of Lyme disease

spirochetes between ticks feeding

simultaneously and in close proximity

contributes little to the perpetuation of

this pathogen, either in North America

or in Europe (1,2). We are correct in

concluding that Lyme disease spiro-

chetes are maintained in nature main-

ly by sequential attachment of ticks to

persistently infected reservoir hosts.

Dania Richter,* Rainer Allgöwer,*

and Franz-Rainer Matuschka*

*Humboldt-Universität zu Berlin, Berlin,

Germany

References

1. Richter D, Allgöwer R, Matuschka F-R.

Co-feeding transmission and its contribu-

tion to the perpetuation of the Lyme disease

spirochete Borrelia afzelii. Emerg Infect

Dis 2002;8:1421–5.

2. Piesman J, Happ CM. The efficacy of co-

feeding as a means of maintaining Borrelia

burgdorferi: a North American model sys-

tem. J Vector Ecol 2001;26:216–20.

3. Randolph S, Gern L. Reply to Richter et al:

Co-feeding transmission and its contribu-

tion to the perpetuation of the Lyme disease

spirochete Borrelia afzelii. Emerg Infect

Dis 2003;9:893–4.

4. Shih CM, Telford SR, Pollack RJ, Spielman

A. Rapid dissemination by the agent of

Lyme disease in hosts that permit fuminat-

ing infection. Infect Immun

1993;61:2396–9.

Address for correspondence: Dania Richter,

Abteilung Parasitologie, Institut für Pathologie,

Charité, Humboldt-Universität zu Berlin,

Malteserstraße 74-100, 12249 Berlin,

Germany; fax: 49 30 776 2085; email:

[email protected]

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 897

NEWS & NOTES

Conference Summary

West Nile Virus

Southeast

Conference

On January 14–15, 2003, more

than 60 scientists, public health offi-

cials, and clinicians from throughout

the southeastern United States gath-

ered in Atlanta to present data from

the 2001 and 2002 West Nile virus

(WNV) epidemics. The aim of the

daylong conference, hosted by the

Southeastern Center for Emerging

Biological Threats at Emory

University, was to assemble a diverse

perspective on WNV by sharing

knowledge, identifying key questions

for research about the disease and pre-

vention, and fostering collaborations

between epidemiologists, veterinari-

ans, laboratorians, and clinicians.

In the overview presentation, the

experiences of state health officials,

clinical spectrum and pathogenesis,

laboratory diagnostics, veterinary

issues, surveillance, and vector control

were discussed. The 2002 epidemic

produced the largest annual number of

cases to date, and four novel modes of

transmission were discovered: 1)

transplanted organs or tissue, includ-

ing blood; 2) breast milk from an

infected mother; 3) percutaneous

exposure to infected tissue or serum

among laboratory and hospital work-

ers; and 4) transplacental exposure to

fetuses in utero, resulting in birth

defects. In addition, WNV-associated

acute flaccid paralysis was discussed;

the paralysis is caused by localized

infection of the anterior horn cells of

the spinal cord, resulting in signs and

symptoms similar to poliomyelitis.

State epidemiologists or their rep-

resentatives presented information

about the state epidemics in the south-

eastern United States. The experi-

ences in four states (Florida, Georgia,

Louisiana, and Mississippi) were

common in their themes of expansion

of WNV epidemics, concentrated

nature of outbreaks, importance of

protection from mosquito bites, limi-

tations of diagnostic methods, and

dynamics of WNV spread in the

United States from 2001 to 2002.

Significant differences also emerged

regarding the observed benefits of

mosquito control activities and the

value of animal surveillance as an

early detection system. Also, while

the states unanimously agreed that

collaboration among local, state, and

federal public health agencies, aca-

demic research institutions, and other

nongovernmental organizations is

critical to responding effectively to

WNV, the degree to which such col-

laborations actually occurred and the

existence of previously established

relationships varied.

Several clinicians discussed the

pathogenesis and clinical aspects of

the disease. The pathogenesis of

WNV was reviewed; clinicians sug-

gested that infections of the central

nervous system can demonstrate any

or all of three distinct characteristics:

neuroinvasiveness (the ability to enter

the nervous system), neurotropism

(the ability to infect neural cells), and

neurovirulence (the ability to cause

neurologic disease). WNV possesses

all three. The virus can enter the nerv-

ous system, as shown by the

encephalitis that occurs in approxi-

mately 1 in 150 infected persons; it

has been shown to infect neurons;

approximately 10% of patients in

which the virus has invaded the nerv-

ous system eventually die. One pre-

senter suggested that the virus had

changed in the last 60 years, evolving

into a more virulent strain. The neuro-

logic manifestations of the disease

were described, including weakness

and flaccid paralysis (which can occur

even without fever or meningoen-

cephalitis).

WNV infection in patients with

HIV was also described. Two HIV-

infected patients, with CD4 cell

counts below 200 cells/µL were iden-

tified with WNV by the presence of

immunoglobulin M antibodies to

WNV. The first was a 50-year-old

homeless man co-infected with tuber-

culosis. Despite treatment including

intravenous acyclovir, the patient con-

tinued to deteriorate, and died 18 days

after admission. An autopsy revealed

meningitis. The second HIV-positive

patient diagnosed with WNV was a

48-year-old man who arrived at the

emergency room with fever,

headache, and confusion; he reported

feeling “slow,” and indeed was slow

to respond to questions. This patient

improved rapidly and was discharged

after 3 days.

Information about intracellular

host-virus interactions was summa-

rized. Studies in mice have identified

a genetic allele that apparently con-

fers resistance to flaviviruses.

Although humans do not have a direct

genetic homologue, studies to identify

genetic differences to explain differ-

ent clinical outcomes should be pur-

sued. Results of a case-control study

suggested that the greatest increases

in risk were related to environmental

factors favoring mosquito popula-

Presentations: James J. Sejvar, Centers for Disease Control and Prevention (CDC); Mary Currier, Mississippi Department of Health;

Raoult Ratard, Louisiana Department of Health & Human Services; Lisa Conti, Florida Department of Health; Susan Lance-Parker,

Georgia Department of Human Resources; Richard T. Johnson, Johns Hopkins University School of Medicine; A. Arturo Leis, University

of Mississippi Medical Center; Margo Brinton, Georgia State University; Sally Slavinski, Mississippi State Health Department; Carlos del

Rio, Emory University; Sharif R. Zaki, CDC; Lillian Stark, Florida Department of Health; Elizabeth Howerth, University of Georgia; Brigid

Elchos, Mississippi Department of Health; Jonathan Day, University of Florida; Tom Bevan, Georgia Institute of Technology; Mike

Bunning, CDC; and David Stallknecht, University of Georgia.

898 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

NEWS & NOTES

Conference Summary

West Nile Virus and

Wildlife Health

The West Nile Virus and Wildlife

Health Workshop, hosted by the

Smithsonian Institution, National

Audubon Society, U.S. Geological

Survey, and U.S. Department of

Agriculture, was held February 5–7,

2003, at the Smithsonian Environ-

mental Research Center in Edgewater,

Maryland. The event was attended by

more than 100 scientists, who heard

29 speakers and participated in strate-

gy discussions during the 2-day meet-

ing. The main focus of the conference

was the present and future impact of

West Nile virus on wildlife popula-

tions. Talks and discussions empha-

sized how basic research, public

health, and land management can con-

tribute to our understanding of the dis-

ease’s impact and spread. A primary

objective of this meeting was to devel-

op future research priorities from both

basic and applied perspectives.

The conference centered around

four main themes: 1) host, vector, and

pathogen interactions (disease ecolo-

gy); 2) vertebrate behavior and ecolo-

gy; 3) vector behavior and ecology;

and 4) modeling and spatial statistics.

We describe some of the findings

from the meeting. For an in-depth

summary of this meeting, please visit

the conference website for meeting

abstracts and a downloadable confer-

ence white paper (available from:

URL: www.serc.si.edu/migratory-

birds/migratorybirds_index.htm).

West Nile virus (WNV) has spread

rapidly across North America since its

probable introduction to the New

York City area in 1999 (D.J. Gubler,

Centers for Disease Control and

Prevention, Fort Collins, CO). By

December 2002, the Canadian

provinces of Saskatchewan, Quebec,

Ontario, Nova Scotia, and Manitoba

reported dead birds that tested posi-

tive for WNV. By winter 2002, only

four states in the continental United

States remained free of confirmed

WNV infection; the virus was expect-

ed to reach the West Coast later in the

year. WNV has also found its way into

tropical regions. One case in a person

was reported in 2001 from the

Cayman Islands. Additionally, resi-

dent birds from Jamaica (January

2002) and the Dominican Republic

tions. These findings in turn raised

questions about factors involved in

human susceptibility, risks of pesti-

cide exposure, efficacy of mosquito

control, the value of sentinel animals

in surveillance, and the roles played

by various species in virus transmis-

sion and amplification.

The ability to diagnose WNV in

the laboratory emphasized the role of

pathology, including histopathology,

electron microscopy, immunohisto-

chemistry, polymerase chain reaction,

and virus isolation. Few of these tests

are entirely sensitive or specific. The

challenges involved in the pathologic

diagnosis in animals because of the

large diversity of infected species

were discussed. The single most

urgent concern, repeatedly empha-

sized by state public health officials

and clinicians, is the need for a faster,

simpler diagnostic test for WNV that

would ease the amount of work by

public health laboratories, assist

physicians in correctly diagnosing

infected patients, and improve surveil-

lance by identifying subclinical cases.

The problems of vector control

and the best application of these meth-

ods were also emphasized. The cost of

these programs and proof of effective-

ness will require careful research,

including the identification of specific

mosquito vectors and assessment of

the long-term safety of pesticides and

personal repellant applications. The

most important principle in attempt-

ing emergency vector control is to

consider it early, before the epidemic

has evolved.

Data from in vitro studies evaluat-

ed the interaction of viral vectors and

amplification hosts. These studies

might elucidate the importance of

birds, horses, and household pets in

maintenance of epidemics. Because

wild birds play a key role in the

spread of WNV, the exact nature of

that role must be clarified to predict

the development and expansion of

future epidemics.

The variation in changing epi-

demiology of the states’ experiences

to date with WNV, even within the

southeastern United States, clearly

demonstrates that research needs to be

replicated in numerous localities;

what succeeds in one state may not

prove successful in another. Whether

epidemics will continue to expand in

size and geographic distribution or

whether a more sporadic pattern of

occurrence will emerge is still

unclear. Controlling WNV in the

southeastern United States will take a

concerted, cohesive effort. The con-

tinued collaboration of the diverse

scientists in this meeting will aid in

this effort. Presentations from this

conference are available on the Web

site for Southeastern Center for

Emerging Biological Threats (avail-

able from: URL: www.secenterbio-

threats.org).

David Rimland,* Jeffrey Koplan,†

David S. Stephens*†‡

*Emory University School of Medicine,

Atlanta, Georgia, USA; †Southeastern

Center for Emerging Biologic Threats,

Atlanta, Georgia, USA; and ‡Centers for

Disease Control and Prevention, Atlanta,

Georgia, USA

Address for correspondence: David Stephens,

Southeastern Center for Emerging Biologic

Threats, 1440 Clifton Road NE, Suite 410,

Atlanta, GA 30322, USA; fax: 404-639-3059;

email: dstep01@emory.edu

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 899

NEWS & NOTES

(November 2002) have shown WNV

antibodies. Recent reports note that

the virus has also reached Mexico’s

Yucatan peninsula. Since 1999, WNV

has killed thousands of birds and

other wildlife, and the impact on

regional wildlife populations is

unclear.

A primary theme of the meeting

was that we still have much to learn

about how WNV is dispersed, trans-

mitted, and amplified by competent

vectors and still relatively unknown

reservoir hosts. At the time of the con-

ference, WNV had been detected in

37 mosquito species, 157 bird species,

horses, 16 other mammals, and alliga-

tors (D.J. Gubler, Centers for Disease

Control and Prevention, Fort Collins,

CO). The Culex genus, particularly

Culex pipiens in the northern United

States and C. quinquefasciatus in the

southern United States, appears to be

the most important mosquito group

for the avian vector amplification

cycle. However, opportunistic mos-

quito species are probably important

bridge vectors to humans, horses, and

other deadend hosts. While the avian

amplification cycle appears to the

most dominant, other cycles may also

be occurring at the same time (i.e., in

mammal and ticks). Reptiles, amphib-

ians, and associated mosquito vectors

may also play important roles (M.J.

Turell, The United States Army

Medical Research Institute for

Infectious Diseases, Fort Detrick,

MD, and E. Jacobson, University of

Florida, Gainesville, FL). Although

mosquitoes appear to be the main vec-

tor, other ectoparasites such as ticks,

louse flies, and fleas should also be

examined as potential vectors.

Several scientists reported that

transmission of WNV is more compli-

cated than previously thought. The

presence of WNV in avian reproduc-

tive organs suggests that vertical

transmission may be a possibility

(T.S. McNamara, Wildlife Conser-

vation Society, Bronx, NY). WNV in

the kidneys leads to cloacal excretion,

which may lead to cloacal-oral mouth

infection. Bird-to-bird transmission

has been demonstrated in the labora-

tory and may be an important infec-

tion route among social birds like the

American crow (R.G. McLean, U.S.

Department of Agriculture, Fort

Collins, CO). Evidence suggests that

ingesting infected vertebrates and

mosquitoes can infect birds.

The impact of WNV on animal

populations is another unknown area.

Data from individually marked popu-

lations of crows in New York State

and Oklahoma (K.J. McGowan,

Cornell University, Ithaca, NY; A.

Clark, State University of New York-

Binghamton, Binghamton, NY; and

C.L. Caffrey, National Audubon

Society, Ivyland, PA) show that these

populations are experiencing impor-

tant declines after the initial WNV

outbreak. Analysis of breeding bird

surveys and annual winter bird cen-

suses (Christmas bird count) from a

wide array of passerine bird species

showed local declines in WNV

“hotspots” but no declines at the

range-wide scale that can be attrib-

uted to WNV (J. Sauer, United States

Geological Survey, Laurel, MD; P.P.

Marra, Smithsonian Environmental

Research Center, Edgewater, MD; and

W. Hochachka, Cornell University,

Ithaca, NY).

Another important issue discussed

at this conference was the secondary

impact that pest management might

have on organisms not pinpointed for

WNV, especially in aquatic environ-

ments. This issue is especially impor-

tant in nature reserves (W.K. Reisen,

University of California, Davis, CA).

Modelers attending the meeting

stressed the importance of standardiz-

ing sampling methods, such as the

dead bird surveillance programs oper-

ated across the nation by many state

health departments. These programs

must consistently and conscientiously

monitor sampling efforts and report

the total sample sizes of dead birds

collected, including the number of

birds that test negative (D.J. Rogers,

Oxford, UK). In addition, a better

understanding of the real-world per-

sistence of WNV antibodies in live

bird surveillance programs would be

useful for virus dispersal models.

Scientists at the meeting felt

strongly that we need to closely mon-

itor how WNV impacts organisms in

tropical regions, including humans

and the many endemic avian species

already threatened or endangered.

Species in Hawaii, many of which are

still endangered after malaria’s centu-

ry-old invasion, should be of special

concern. WNV is not the first and will

not be the last virus to enter our bor-

ders. By developing techniques to sur-

vey, monitor, and control WNV in

wildlife, we prepare ourselves for the

next pathogen species. Our experi-

ences with WNV emphasize the need

to strengthen and integrate animal

monitoring programs with basic

research on population and disease

ecology. A conference white paper,

several review articles, and a list of

research priorities are planned as

products of this meeting.

Peter P. Marra,* Sean M. Griffing,*

and Robert G. McLean†

*Smithsonian Environmental Research

Center, Edgewater, Maryland, USA; and

†National Wildlife Research Center, Fort

Collins, Colorado, USA

Address for correspondence: Peter P. Marra,

Smithsonian Environmental Research Center,

P.O. Box 28, Edgewater, MD 21037, USA; fax:

443-482-2380; email: [email protected]

900 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

NEWS & NOTES

Upcoming Infectious

Disease Conferences

July 7–10, 2003

ASM Conference on Bio-,

Micro- and Nanosystems

American Society for

Microbiology

New York, New York

Contact: ASM Conferences

Phone: 202-942-9261

Fax: 202-942-9340

Email: [email protected]

July 13–16, 2003

The 2nd IAS Conference on

HIV Pathogenesis and Treatment

Paris, France

Contact: JCD Conseil/IAS 2003

Phone: 33 1 40 64 2000

Fax: 33 1 40 64 2050

Website: http://www.ias2003.org

July 16–18, 2003

ISAAR 2003—4th International

Symposium on Antimicrobial

Agents and Resistance

Antimicrobial Treatment in the

21st Century: Current Challenges

and Future Strategies

Seoul, Korea

Contact: Ms. Susan Chung

Phone: 82-2-3410-0327

Fax: 82-2-3410-0023

Email: [email protected]

Website: http://www.ansorp.org/

isaar2003/intro.htm

July 16–19, 2003

11th Annual Conference:

Boards of Health: Stewards

of Public Health

National Association of Local

Boards of Health

Baltimore, MD

Website: http://www.nalboh.org/

confer/confmain.htm

Emerging Infectious Diseases • Vol. 9, No. 7, July 2003 901

As the Gods began one world, and man another,

So the snakecharmer begins a snaky sphere

With moon-eye, mouth pipe.

He pipes. Pipes green. Pipes water

Sylvia Plath, “Snakecharmer”

ritics called him “naïve,” the term for painters with no

formal training in art. Henri Rousseau, self-made late-

bloomer from Laval, France, fit the definition. But his

work proved that in art as in all ventures, training, though

valuable, is not the key ingredient—not as key perhaps as

talent, inspiration, or originality. Untrained in art but not

uneducated by the standards of his day, Rousseau was a

teacher and a military man. He was interested in politics

and the realm of ideas. He knew music and poetry and

even tried his hand as playwright. Dubbed “Le Douanier”

(customs officer) after his main occupation outside art, he

struggled in anonymity until near the end of his life, when

he was discovered by Pablo Picasso and others and was

recognized for his powerful individual style (1).

Like other naïve or primitive artists, Rousseau found art

late in life. He took up painting as a hobby and soon retired

from his job in the customs office to devote time to this

new vocation. He copied the masters, struggled to learn

their craft (particularly the academic style of Ingres), and

aspired to paint like them. He exhibited often at the Salon

des Indépendents in Paris, where artists could show their

paintings without selection restrictions, and what he might

have lost to technical clumsiness he seemed to make up in

ingenuous charm. Although his work is difficult to catego-

rize, he seems to have been influenced by his contempo-

rary Paul Gauguin and his followers, the Nabis, who pro-

moted directness of feeling and color harmony (2).

Rousseau eventually found his own formal language and

style, but what elevated his mature work to greatness were

perhaps the very oblivion of convention, the freshness of

approach, and the depth of discovery that comes from a

truly unique perspective.

Rousseau’s exotic compositions owe nothing to tradi-

tional art methods yet defy modern labels. The fantastic

vegetation in his jungle paintings (for which he is best

known) has no equivalent in nature. These exotic land-

scapes, oversized and filled with exuberant color, were

entirely imaginary. Although often inhabited by half-con-

cealed wild beasts and laced with conflict, they exuded an

eerie stillness. The images, smooth, vivid, and clearly

defined, were flat and fluid against dense but dimension-

less greenery, and although unreal and extraordinary, were

rendered in meticulous botanical detail.

The Snake Charmer, on this month’s cover of Emerging

Infectious Diseases, is one of Rousseau’s finest and most

celebrated works. Like his other jungle paintings, it is

filled with lush greenery. Punctuated by an uncoiling rep-

tile at arm’s length, the thick vegetal screen that makes up

most of the landscape is live with tension. The dark, undu-

lating figure of the snake charmer dances ambiguously

amidst a tangle of wildlife. Nature, framed by “a wave of

flickering-grass tongues,” (3) looms in the foreground

immediate and tangible, yet dreamlike and distant as the

moon. In a trance, the animals are guided (it seems as

much by the glossy stream as by the snake charmer’s reed)

into a tight ecological web, where unbeknownst to them,

they share more than the music.

Rousseau’s imagination, like that of many of his con-

temporaries in Paris, succumbed to the allure of exotic

lands, where plants grew larger than life, wild animals held

unknown powers and magnetism, and humans lounged in

“Eden’s navel” (3) amidst all that was lost in the fall from

grace. To explain the products of his inflamed imagination,

Rousseau falsely claimed that he had visited Mexico. But

unlike Gauguin, who went to Tahiti in search of inspira-

tion, Rousseau traveled only vicariously and found his

models in local gardens and the Paris zoo.

Exotic lands have become prosaic to us. What remains

naïve and primitive is our knowledge of the forest’s archi-

tecture and the perils of its convergence with human habi-

tat. But like the uncoiling snake in Rousseau’s painting,

out of the impenetrable jungle comes knowledge about

pestilences, piece by piece: a favorable environment, a sta-

ble population, a reservoir host, the agent. The emergence

of West Nile virus in North America is a case in point. We,

modern snake charmers, must pipe the pieces (bird, horse,

reptile) into a knowable, harmonious fabric.

Polyxeni Potter

1. Shattuck R, Behar H, Hoog M, Lanchner C, Bubin W. Henri

Rousseau. New York: New York Graphic Society; 1985.

2. Janson HW, Janson AF. History of art. New York: Harry N. Abrams

Inc.; 2001.

3. Plath S. Snakecharmer [cited 2003 May 14]. Available from: URL:

http://www.eng.fju.edu.tw/English_Literature/us_poetry/Plath/Plath_

poem.html

ABOUT THE COVER

Henri Rousseau—known as Le Douanier Rousseau (1844–1910) The Snake Charmer (1907). Oil on canvas, 169 cm x 189.5 cm

Musée d’Orsay, Paris, France. Credit: Réunion des Musées Nationaux/Art Resource, NY

902 Emerging Infectious Diseases • Vol. 9, No. 7, July 2003

NEWS & NOTES

A Peer-Reviewed Journal Tracking and Analyzing Disease Trends Vol.9, No.8, August, 2003

Upcoming Issue

For a complete list of articles included in the August issue,

and for articles published online ahead of print publication,

see http://www.cdc.gov/ncidod/eid/upcoming.htm

Look in the August issue for the following topics:

Porcine Reproductive and Respiratory Syndrome Virus: Origin Hypothesis

Detecting Bioterror Attack by Screening Blood Donors: Best-Case Analysis

Survival of Batrachochytrium dendrobatidis in

Water: Quarantine and Disease Control Implications

Legionnaires' Disease Outbreak in Murcia, Spain

Enzootic Transmission of Yellow Fever Virus in Peru

NmcA Carbapenem-hydrolyzing Enzyme in

Enterobacter cloacae in North America

Echovirus 13 Associated with Aseptic Meningitis, Spain

Molecular Characteristics of a Non–Babesia divergens

Organism Causing Zoonotic Babesiosis in Europe

Multidrug-Resistant Tuberculosis in HIV-Negative Patients,

Buenos Aires, Argentina Invasive Group A

Streptococcal Disease: Risk Factors for Adults

Severe Tungiasis in Underprivileged Communities: Case Series from Brazil

Instructions to Authors

Manuscript Preparation. For word processing, use MS Word. Begin

each of the following sections on a new page and in this order: title page,

keywords, abstract, text, acknowledgments, biographical sketch, refer-

ences, tables, figure legends, appendixes, and figures. Each figure should

be in a separate file.

Title Page. Give complete information about each author (i.e., full name,

graduate degree(s), affiliation, and the name of the institution in which

the work was done). Clearly identify the corresponding author and pro-

vide that author's mailing address (include phone number, fax number,

and e-mail address). Include separate word counts for abstract and text.

Keywords. Include up to 10 keywords; use terms listed in Medical

Subject Headings Index Medicus.

Text. Double-space everything, including the title page, abstract, refer-

ences, tables, and figure legends. Printed manuscript should be single-

sided, beginning with the title page. Indent paragraphs; leave no extra

space between paragraphs. After a period, leave only one space before

beginning the next sentence. Use 12-point Times New Roman font and

format with ragged right margins (left align). Italicize (rather than under-

line) scientific names when needed.

Biographical Sketch. Include a short biographical sketch of the first

author—both authors if only two. Include affiliations and the author's pri-

mary research interests.

References. Follow Uniform Requirements (www.icmje.org/index.html).

Do not use endnotes for references. Place reference numbers in parenthe-

ses, not superscripts. Number citations in order of appearance (including

in text, figures, and tables). Cite personal communications, unpublished

data, and manuscripts in preparation or submitted for publication in

parentheses in text. Consult List of Journals Indexed in Index Medicus for

accepted journal abbreviations; if a journal is not listed, spell out the jour-

nal title. List the first six authors followed by “et al.” Do not cite refer-

ences in the abstract.

Tables and Figures. Create tables within MS Word’s table tool. Do not

format tables as columns or tabs. Send graphics in native, high-resolution

(200 dpi minimum) .TIF (Tagged Image File), or .EPS (Encapsulated

Postscript) format. Graphics should be in a separate electronic file from

the text file. For graphic files, use Arial font. Convert Macintosh files into

the suggested PC format. Figures, symbols, letters, and numbers should

be large enough to remain legible when reduced. Place figure keys with-

in the figure. For more information see EID Style Guide (http://www.

cdc.gov/ncidod/EID/style_guide.htm).

Manuscript Submission. Include a cover letter indicating the proposed

category of the article (e.g., Research, Dispatch) and verifying that the

final manuscript has been seen and approved by all authors. To submit a

manuscript, access Manuscript Central from the Emerging Infectious

Diseases website (www.cdc.gov/eid).

Manuscript Types

Perspectives. Articles should be under 3,500 words and should include

references, not to exceed 40. Use of subheadings in the main body of the

text is recommended. Photographs and illustrations are encouraged.

Provide a short abstract (150 words) and a brief biographical sketch of

first author. Articles in this section should provide insightful analysis and

commentary about new and reemerging infectious diseases and related

issues. Perspectives may also address factors known to influence the

emergence of diseases, including microbial adaptation and change,

human demographics and behavior, technology and industry, economic

development and land use, international travel and commerce, and the

breakdown of public health measures. If detailed methods are included, a

separate section on experimental procedures should immediately follow

the body of the text.

Synopses. Articles should be under 3,500 words and should include ref-

erences, not to exceed 40. Use of subheadings in the main body of the text

is recommended. Photographs and illustrations are encouraged. Provide a

short abstract (150 words) and a brief biographical sketch of first

author—both authors if only two. This section comprises concise reviews

of infectious diseases or closely related topics. Preference is given to

reviews of new and emerging diseases; however, timely updates of other

diseases or topics are also welcome. If detailed methods are included, a

separate section on experimental procedures should immediately follow

the body of the text.

Research Studies. Articles should be under 3,500 words and should

include references, not to exceed 40. Use of subheadings in the main body

of the text is recommended. Photographs and illustrations are encour-

aged. Provide a short abstract (150 words) and a brief biographical sketch

of first author—both authors if only two. Report laboratory and epidemi-

ologic results within a public health perspective. Although these reports

may be written in the style of traditional research articles, they should

explain the value of the research in public health terms and place the find-

ings in a larger perspective (i.e., "Here is what we found, and here is what

the findings mean").

Policy and Historical Reviews. Articles should be under 3,500 words

and should include references, not to exceed 40. Use of subheadings in

the main body of the text is recommended. Photographs and illustrations

are encouraged. Provide a short abstract (150 words) and a brief biogra-

phical sketch. Articles in this section include public health policy or his-

torical reports that are based on research and analysis of emerging disease

issues.

Dispatches. Articles should be 1,000–1,500 words and need not be divid-

ed into sections. If subheadings are used, they should be general, e.g.,

“The Study” and “Conclusions.” Provide a brief abstract (50 words); ref-

erences (not to exceed 15); figures or illustrations (not to exceed two);

and a brief biographical sketch of first author—both authors if only two.

Dispatches are updates on infectious disease trends and research. The

articles include descriptions of new methods for detecting, characterizing,

or subtyping new or reemerging pathogens. Developments in antimicro-

bial drugs, vaccines, or infectious disease prevention or elimination pro-

grams are appropriate. Case reports are also welcome.

Commentaries. Thoughtful discussions (500–1,000 words) of current top-

ics. Commentaries may contain references but should not include figures

or tables.

Another Dimension. Thoughtful essays, short stories, or poems on philo-

sophical issues related to science, medical practice, and human health.

Topics may include science and the human condition, the unanticipated

side of epidemic investigations, or how people perceive and cope with

infection and illness. This section is intended to evoke compassion for

human suffering and to expand the science reader's literary scope.

Manuscripts are selected for publication as much for their content (the

experiences they describe) as for their literary merit.

Letters. This section includes letters that present preliminary data or

comment on published articles. Letters (500–1,000 words) should not be

divided into sections, nor should they contain figures or tables.

References (not more than 10) may be included.

Book Reviews. Short reviews (250–500 words) of recently published

books on emerging disease issues are welcome. The name of the book,

publisher, and number of pages should be included.

Announcements. We welcome brief announcements (50–150 words) of

timely events of interest to our readers. (Announcements may be posted

on the journal Web page only, depending on the event date.)

Conference Summaries. (500–1,000 words) of emerging infectious dis-

ease conferences may provide references to a full report of conference

activities and should focus on the meeting's content rather than on indi-

vidual conference participants.

Editorial Policy and Call for Articles

Emerging Infectious Diseases is a peer-reviewed journal established expressly to promote the recognition of new and reemerging infectious diseases

around the world and improve the understanding of factors involved in disease emergence, prevention, and elimination.

The journal has an international scope and is intended for professionals in infectious diseases and related sciences. We welcome contributions from

infectious disease specialists in academia, industry, clinical practice, and public health, as well as from specialists in economics, demography, sociol-

ogy, and other disciplines. Inquiries about the suitability of proposed articles may be directed to the Editor at 404-371-5329 (tel), 404-371-5449 (fax),

or [email protected] (e-mail).

Emerging Infectious Diseases is published in English and features the following types of articles: Perspectives, Synopses, Research Studies, Policy

and Historical Reviews, Dispatches, Commentaries, Another Dimension, Letters, Book Reviews, and News and Notes. The purpose and requirements

of each type of article are described in detail below. To expedite publication of information, we post journal articles on the Internet as soon as they are

cleared and edited.

Chinese, French, and Spanish translations of some articles can be accessed through the journal’s home page at http://www.cdc.gov/eid.