Draft Human Health Risk Assessment for Formaldehyde CASRN 50

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

EPA Document #EPA-740-D-24-003

March 2024

United States Office of Chemical Safety and

Environmental Protection Agency Pollution Prevention

Draft Human Health Risk Assessment for Formaldehyde

CASRN 50-00-0

March 2024

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 2 of 151

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ..................................................................................................................... 6

EXECUTIVE SUMMARY ...................................................................................................................... 8

1 INTRODUCTION....................................................................................................................... 16

1.1 Background ................................................................................................................................. 16

1.2 Risk Evaluation Scope ................................................................................................................ 16

Life Cycle and Production Volume ....................................................................................... 17

Conditions of Use .................................................................................................................. 19

Other Sources of Formaldehyde in Air .................................................................................. 28

1.3 Chemistry, Fate, and Transport Assessment Summary .............................................................. 29

1.4 Environmental Release Assessment ........................................................................................... 32

1.5 Human Health Assessment Scope .............................................................................................. 34

Conceptual Exposure Models ................................................................................................ 34

1.5.1.1 Industrial and Commercial Activities and Uses ............................................................. 34

1.5.1.2 Consumer Activities and Uses ........................................................................................ 36

1.5.1.3 Indoor Air Exposures ..................................................................................................... 38

1.5.1.4 General Population Exposures from Environmental Releases ....................................... 40

Potentially Exposed or Susceptible Subpopulations.............................................................. 42

2 HUMAN EXPOSURE ASSESSMENT SUMMARY .............................................................. 44

2.1 Occupational Exposure Assessment ........................................................................................... 44

Inhalation Exposure Assessment ........................................................................................... 45

Dermal Exposure Summary ................................................................................................... 46

2.2 Consumer Exposure Assessment ................................................................................................ 47

2.3 Indoor Air Exposure Assessment ............................................................................................... 51

Indoor Air Exposure Monitoring Results .............................................................................. 52

Indoor Air Exposure Modeling Results ................................................................................. 56

2.3.2.1 Aggregate Indoor Air Exposure ..................................................................................... 59

2.4 Ambient Air Exposure Assessment ............................................................................................ 59

Monitoring for Ambient Air Concentrations ......................................................................... 59

Modeling Ambient Air Concentrations ................................................................................. 60

2.4.2.1 Integrated Indoor/Outdoor Air Calculator Model (IIOAC) ............................................ 60

2.4.2.2 AirToxScreen ................................................................................................................. 63

2.4.2.3 Human Exposure Model (HEM) .................................................................................... 64

Integrating Various Sources of Formaldehyde Data ............................................................. 67

2.5 Weight of Scientific Evidence and Overall Confidence in Exposure Assessment .................... 68

Overall Confidence in Occupational Exposure Assessment ................................................. 69

Overall Confidence in the Consumer Exposure Assessment ................................................ 70

Overall Confidence in the Indoor Air Exposure Assessment ................................................ 71

Overall Confidence in the Ambient Air Exposure Assessment ............................................. 71

3 HUMAN HEALTH HAZARD SUMMARY ............................................................................ 73

3.1 Summary of Hazard Values ........................................................................................................ 73

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 3 of 151

3.2 Weight of Scientific Evidence and Overall Confidence in Hazard Assessment ........................ 76

Overall Confidence in the Acute Inhalation POD ................................................................. 76

Overall Confidence in the Chronic, Non-cancer Inhalation POD ......................................... 76

Overall Confidence in the Chronic IUR ................................................................................ 77

Overall Confidence in the Dermal POD ................................................................................ 77

Overall Confidence in the Subchronic and Chronic Oral PODs ........................................... 77

4 HUMAN HEALTH RISK CHARACTERIZATION .............................................................. 79

4.1 Risk Characterization Approach ................................................................................................. 79

Estimation of Non-cancer Risks ............................................................................................ 80

Estimation of Cancer Risks ................................................................................................... 81

4.2 Risk Estimates ............................................................................................................................ 81

Risk Estimates for Workers ................................................................................................... 81

4.2.1.1 Risk Estimates for Inhalation Exposures ........................................................................ 81

4.2.1.2 Overall Confidence in Worker Inhalation Risks ............................................................ 86

4.2.1.3 Risk Estimates for Dermal Exposures ............................................................................ 88

4.2.1.4 Overall Confidence in Worker Dermal Risks ................................................................ 88

Risk Estimates for Consumers ............................................................................................... 88

4.2.2.1 Risk Estimates for Inhalation Exposure to Formaldehyde in Consumer Products ........ 89

4.2.2.2 Risk Estimates for Dermal Exposure to Formaldehyde in Consumer Products ............. 93

Risk Estimates for Indoor Air ................................................................................................ 94

4.2.3.1 Risk Estimates Based on Indoor Air Monitoring Data ................................................... 94

4.2.3.2 Risk Estimates Based on Indoor Air Modeling for Specific TSCA COUs .................... 95

4.2.3.3 Integration of Modeling and Monitoring Information and Consideration of Aggregate

Risk ................................................................................................................................. 97

Risk Estimates for Ambient Air ............................................................................................ 97

4.2.4.1 Risk Estimates Based on Ambient Air Monitoring ........................................................ 98

4.2.4.2 Risk Estimates Based on Modeled Concentrations near Releasing Facilities ................ 99

4.2.4.3 Integration of Modeling and Monitoring Information .................................................. 104

4.2.4.4 Overall Confidence in Exposures, Risk Estimates, and Risk Characterizations for

Ambient Air .................................................................................................................. 105

Comparison of Non-cancer Effect Levels and Air Concentrations ..................................... 106

Potentially Exposed or Susceptible Subpopulations............................................................ 107

4.3 Aggregate and Sentinel Exposures ........................................................................................... 111

5 NEXT STEPS ............................................................................................................................ 113

REFERENCES ...................................................................................................................................... 114

100

APPENDICES ....................................................................................................................................... 123

101

Appendix A ABBREVIATIONS AND ACRONYMS ..................................................................... 123

102

Appendix B LIST OF DOCUMENTS AND SUPPLEMENTAL FILES ...................................... 125

103

Appendix C DETAILED EVALUATION OF POTENTIALLY EXPOSED AND

104

SUSCEPTIBLE SUBPOPULATIONS ....................................................................... 127

105

PESS Based on Greater Exposure ............................................................................................ 127

106

PESS Based on Greater Susceptibility ..................................................................................... 131

107

Appendix D AMBIENT AIR RISK ESTIMATES – COMMERCIAL USES .............................. 140

108

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 4 of 151

Appendix E DRAFT OCCUPATIONAL EXPOSURE VALUE DERIVATION ........................ 142

109

Draft Occupational Exposure Value Calculations .................................................................... 143

110

Summary of Air Sampling Analytical Methods Identified ...................................................... 145

111

Appendix F ACUTE AND CHRONIC (NON-CANCER AND CANCER) OCCUPATIONAL

112

INHALATION EQUATIONS ..................................................................................... 147

113

Appendix G DERMAL EXPOSURE APPROACH ........................................................................ 151

114

115

LIST OF TABLES

116

Table 1-1. Categories and Subcategories of Use and Corresponding Exposure Scenario in the Risk

117

Evaluation for Formaldehyde ........................................................................................... 20

118

Table 1-2. Physical and Chemical Properties of Formaldehyde and Select Transformation Products .... 30

119

Table 2-1. Indoor Air Monitoring Concentrations for Formaldehyde ...................................................... 53

120

Table 2-2. Formaldehyde Monitored in U.S. Commercial Buildings from 2000 to Present .................... 54

121

Table 2-3. Representative Residential Indoor Air Exposure Scenarios According to COUs ................... 58

122

Table 2-4. Overall Monitored Concentrations of Formaldehyde from AMTIC Dataset .......................... 60

123

Table 3-1. Hazard Values Identified for Formaldehyde ........................................................................... 74

124

Table 4-1. Use Scenarios, Populations of Interest, and Toxicological Endpoints Used for Acute and

125

Chronic Exposures ............................................................................................................ 79

126

Table 4-2: Population Summary for Cancer Risk Estimates Derived from HEM Modeling of TRI

127

Releases Formaldehyde to Air ........................................................................................ 102

128

Table 4-3. Demographic Details of Population with Estimated Cancer Risk Higher than or Equal to 1 in

129

1 Million, Compared with National Proportions ............................................................ 103

130

Table 4-4. Summary of PESS Considerations Incorporated throughout the Analysis and Remaining

131

Sources of Uncertainty .................................................................................................... 108

132

133

LIST OF FIGURES

134

Figure 1-1. Risk Evaluation Document Summary Map ............................................................................ 17

135

Figure 1-2. Lifecycle Diagram of Formaldehyde ..................................................................................... 18

136

Figure 1-3. Chemical Equilibria for Formaldehyde in Aqueous Solutions .............................................. 31

137

Figure 1-4. Conceptual Model for Industrial and Commercial Activities and Uses: Potential Exposure

138

and Hazards ....................................................................................................................... 35

139

Figure 1-5. Formaldehyde Conceptual Model for Consumer Activities and Uses: Potential Exposures

140

and Hazards ....................................................................................................................... 37

141

Figure 1-6. Formaldehyde Conceptual Model for Indoor Air: Residential Exposures and Hazards from

142

Article Off-Gassing ........................................................................................................... 39

143

Figure 1-7. Formaldehyde Conceptual Model for Environmental Releases and Wastes: General

144

Population Exposures and Hazards ................................................................................... 41

145

Figure 1-8. Industrial Releases to the Environment and Pathways by Which Exposures to the General

146

Population May Occur ...................................................................................................... 42

147

Figure 2-1. Summary of 15-Minute Peak Consumer Inhalation Concentrations (Based on CEM) ......... 49

148

Figure 2-2. Summary of Average Daily Consumer Inhalation Concentrations, per Year (Based on CEM)

149

........................................................................................................................................... 50

150

Figure 2-3. Summary of Acute Consumer Dermal Concentrations (Based on Thin Film Model) ........... 51

151

Figure 2-4. Long-Term Average Daily Concentrations of Formaldehyde According to Air Monitoring

152

Data Source ....................................................................................................................... 55

153

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 5 of 151

Figure 2-5. Modeled Formaldehyde Average Daily Inhalation Concentrations in Indoor Air (According

154

to CEM) ............................................................................................................................ 57

155

Figure 2-6. Exposure Concentrations by TSCA COU for the 95th Percentile Release Scenario and 95th

156

Percentile Modeled Concentration between 100 and 1,000 m from Industrial Facilities

157

Releasing Formaldehyde to the Ambient Air ................................................................... 62

158

Figure 2-7. Distributions of 2019 AirToxScreen Modeled Data for All Sources, Secondary Production

159

Sources, Point Sources, and Biogenic Sources for the Contiguous United States ........... 64

160

Figure 2-8. Map of Contiguous United States with HEM Model Results for TRI Releases Aggregated

161

and Summarized by Census Block. .................................................................................. 66

162

Figure 2-9. Median and Maximum Concentrations (Fugitive, Stack, and Total Emissions) across the 11

163

Discrete Distance Rings Modeled in HEM ....................................................................... 67

164

Figure 2-10. Distributions of AMTIC Monitoring Data, IIOAC Modeled Data, and AirToxScreen

165

Modeled Data .................................................................................................................... 68

166

Figure 4-1. Acute, Non-cancer Occupational Inhalation and Dermal Risk by TSCA COU .................... 83

167

Figure 4-2. Chronic, Non-cancer Occupational Inhalation Risk by TSCA COU ..................................... 84

168

Figure 4-3. Chronic Cancer Occupational Inhalation Risk by TSCA COU ............................................. 85

169

Figure 4-4. Peak 15-Minute Inhalation Risk by COUs in Consumer Products ........................................ 90

170

Figure 4-5. Chronic Non-cancer Inhalation Risks for Consumer Products by COU ................................ 91

171

Figure 4-6. ADAF-Adjusted Chronic Inhalation Cancer Risk by COUs in Consumer Products ............. 92

172

Figure 4-7. Acute Dermal Loading Risk by High-End Exposure Scenarios in Consumer Products........ 93

173

Figure 4-8. ADAF-Adjusted Lifetime Cancer Inhalation Risk by Indoor Air Monitoring Data Source .. 95

174

Figure 4-9. Chronic Non-cancer Inhalation Risk Based on Modeled Air Concentrations for Specific

175

TSCA COUs ..................................................................................................................... 96

176

Figure 4-10. ADAF-Adjusted Cancer Risk for Monitoring and Modeling Ambient Air Data ................ 98

177

Figure 4-11. Risk Estimates by TSCA COU for the 95th Percentile Release Scenario and 95th Percentile

178

Modeled Concentration between 100 and 1,000 m from Industrial Facilities Releasing

179

Formaldehyde to the Ambient Air .................................................................................. 100

180

Figure 4-12. Comparison of Non-cancer Health Effect Levels Reported in People and Indoor and

181

Outdoor Air Concentrations ............................................................................................ 106

182

183

LIST OF APPENDIX TABLES

184

Table_Apx C-1. PESS Based on Greater Exposure................................................................................ 128

185

Table_Apx C-2. Susceptibility Category, factors, and evidence for PESS susceptibility ...................... 132

186

Table_Apx E-1. Limit of Detection (LOD) and Limit of Quantification (LOQ) Summary for Air

187

Sampling Analytical Methods Identified ........................................................................ 145

188

Table_Apx F-1. Appendix F Formulae – Symbols, Values, and Units .................................................. 148

189

Table_Apx F-2. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+) ...... 149

190

Table_Apx F-3. Median Years of Tenure with Current Employer by Age Group ................................. 150

191

192

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 6 of 151

ACKNOWLEDGEMENTS

193

This report was jointly developed by the United States Environmental Protection Agency (U.S. EPA or

194

the Agency), Office of Chemical Safety and Pollution Prevention (OCSPP), Office of Pollution

195

Prevention and Toxics (OPPT), and the Office of Pesticide Programs (OPP).

196

197

Acknowledgements

198

The OPPT and OPP Assessment Teams gratefully acknowledge the participation, input, and review

199

comments from OPPT, OPP, and OCSPP senior managers and science advisors and assistance from

200

EPA contractor SRC (Contract No. 68HERH19D0022) and ERG (Contract No. 68HERD20A0002).

201

OPPT and OPP also gratefully acknowledge systematic review work conducted by staff in the Data

202

Gathering and Analysis Division. Special acknowledgement is given for the contributions of technical

203

experts from EPA’s Office of Research and Development (ORD).

204

205

As part of an intra-agency review, the draft formaldehyde risk evaluation was provided to multiple EPA

206

Program Offices for review. Comments were submitted by ORD, Office of Air and Radiation, and the

207

Office of Children’s Health Protection.

208

209

Docket

210

Supporting information can be found in public docket, Docket ID (EPA-HQ-OPPT-2018-0438).

211

212

Disclaimer

213

Reference herein to any specific commercial products, process or service by trade name, trademark,

214

manufacturer or otherwise does not constitute or imply its endorsement, recommendation, or favoring by

215

the United States Government.

216

217

Authors: Shawn Shifflett (Assessment Lead), Rochelle Bohaty (Management Lead and Branch Chief),

218

Whitney Hollinshead, Giorvanni Merilis, Kevin Vuilleumier, and Susanna Wegner

219

220

Contributors: John Allran, Edwin Arauz, Marcy Card, Bryan Groza, Grant Goedjen, and Myles Hodge

221

222

Technical Support: Mark Gibson, Hillary Hollinger.

223

224

This draft risk evaluation was reviewed by OPPT, OPP, and OCSPP leadership.

225

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 7 of 151

Formaldehyde – Human Health Risk Characterization - Key Points

Formaldehyde is a highly reactive gas that is ubiquitous in indoor and outdoor environments. It is

widely used in a range of industrial applications, consumer products, and building materials (e.g.,

composite wood products, plastics, rubber, various adhesives and sealants). It also occurs as a product

of combustion, a product of normal metabolism in the human body, and is formed naturally through the

decomposition of organic matter (i.e., biogenic sources).

Health effects of concern for formaldehyde include cancer and respiratory effects such as increased

asthma prevalence, reduced asthma control, and reduced lung function. People may be exposed to

formaldehyde at work, through indoor air, through use of consumer products, and through outdoor air

near sources of formaldehyde. People are often exposed to many sources of formaldehyde

concurrently, some of which are regulated under the Toxic Substances Control Act (TSCA) some of

which are regulated under other laws, and some of which are not regulated at all (for example, the

decomposition of leaves).

This draft human health risk assessment for formaldehyde evaluates the risks of formaldehyde

exposures for workers, consumers, and the general population resulting from TSCA conditions of use

(COUs).

Risk estimates include inherent uncertainties and the overall confidence in specific risk estimates

varies. The analysis provides support for the Agency to make a determination about whether

formaldehyde poses an unreasonable risk to human health and to identify drivers of unreasonable risk

among exposures for people (1) with occupational exposure to formaldehyde, (2) with consumer

exposure to formaldehyde, (3) with exposure to formaldehyde in indoor air, and (4) who live or work

in proximity to locations where formaldehyde is released to air. Concurrent with this draft TSCA Risk

Evaluation, EPA is releasing a preliminary risk determination for formaldehyde.

While EPA is making this risk determination, EPA will consider the standard risk benchmarks

associated with interpreting margins of exposure and cancer risks. However, EPA cannot solely rely on

those risk values. The Agency also will consider naturally occurring sources of formaldehyde (i.e.,

biogenic, combustion, and secondary formation) and associated risk levels therefrom, and consider

contributions from all sources as part of a pragmatic and holistic evaluation of formaldehyde hazard

and exposure in making its unreasonable risk determination. If an estimate of risk for a specific

exposure scenario exceeds the benchmarks, then the decision of whether those risks are formally

unreasonable under TSCA must be both case-by-case and context driven in the case of formaldehyde.

EPA is taking the risk estimates of the human health risk assessment (HHRA), in combination with a

thoughtful consideration of other sources of formaldehyde, to interpret the risk estimates in the context

of making an unreasonable risk determination.

226

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 8 of 151

EXECUTIVE SUMMARY

227

Sixty-two conditions of use of formaldehyde were determined to be within the scope of TSCA and were

228

assessed by OPPT. These conditions of use were identified as part of the Final Scope for the Risk

229

Evaluation for Formaldehyde 50-00-0 (U.S. EPA, 2020c) and recently updated to better reflect the

230

Agency’s understanding of the sources of formaldehyde. Examples of the conditions of use considered

231

in the TSCA risk evaluation are listed below with a comprehensive list provided in the Draft Conditions

232

of Use for the Formaldehyde Risk Evaluation (U.S. EPA, 2024c); these include

233

• manufacturing of formaldehyde,

234

• processing and manufacturing of articles and products,

235

• composite wood products,

236

• plastics used in toys,

237

• rubber materials, and

238

• various adhesives and sealants.

239

Readily available information indicates that formaldhyde is released to air, land, and water from various

240

TSCA conditions of use. Although the draft formaldehyde risk evaluation considered each of these

241

pathways of exposure, some of these releases result in negligible exposure based on the chemistry, fate,

242

and transport properties of formaldehyde. Formaldehyde exposures by those pathways were not assessed

243

further. These include

244

• sediment and water including drinking water, and

245

• soils, biosolids, and landfills.

246

Similarly, some conditions of use were considered for consumer scenarios and result in negligible

247

exposure based on the chemistry, fate, and transport properties of formaldehyde. Other conditions of use

248

could not be assessed due to the limitation of available models and data. These conditions of use are

249

• water treatment,

250

• laundry detergent, and

251

• lawn and garden products.

252

This Draft Human Health Risk Assessment for Formaldehyde focuses on human exposure to

253

formaldehyde from industrial, occupational, and consumer activities via inhalation of indoor and

254

outdoor air and dermal (skin) routes. Exposure to workers, consumers and people within the general

255

population have been assessed under specific conditions of use. Not all conditions of use result in

256

formaldehyde exposure for all populations. Among the populations assessed are potentially exposed or

257

susceptible subpopulations (PESS), which are people who have higher exposures or are more susceptible

258

so may be at greater risk of adverse health effects from formaldehyde. Example populations (including

259

PESS), routes of exposure, and conditions of use include the following:

260

• worker inhalation and dermal exposure during manufacturing, processing, distribution, use and

261

disposal of formaldehyde;

262

• consumer (based on highest expected exposure among all ages) inhalation and dermal exposure

263

from use of paint, laundry detergents, hand and dishwashing soaps, drain and toilet cleaners,

264

textile and leather finishing products, varnishes and floor finishes, rubber mats, adhesives, caulks

265

and sealants, liquid photographic processing solutions, and non-spray lubricants that contain

266

formaldehyde;

267

• general population (all ages) inhalation exposure to indoor air from products used in new

268

construction of homes and mobile homes (e.g., wood materials, furniture seat covers,); and

269

automobiles with products that contain formaldehyde; and

270

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 9 of 151

• general population (all ages) inhalation exposure to outdoor air near industrial facilities that

271

release formaldehyde.

272

As mentioned, there are many formaldehyde sources. Not all sources are considered in the Draft TSCA

273

Risk Evaluation, either because they occur naturally or because they are regulated under other statutes.

274

These include

275

• forest fires;

• combustion

;

• tail-pipe emissions from cars, trucks, and

other vehicles;

• plastic products used for food storage

and distribution;

• animal feed;

• biogenic sources (like trees and wood

chips);

• secondary formation

;

• drugs for fisheries and hatcheries;

• pesticides and other formaldehyde uses

regulated by the Food and Drug

Administration;

• pacifiers and baby bottles; and,

• embalming or as a preservative from

funeral homes and taxidermy.

These other sources can produce substantial amounts of formaldehyde resulting in exposures in the

276

occupational, indoor, and outdoor environments. For example, biogenic concentrations can contribute

277

upwards of 25 percent of the total formaldehyde concentration and secondary formation can account for

278

as much as 80 percent in ambient air, depending on the circumstances.

279

280

Hazard Values

281

Human health hazard data for this draft assessment were obtained through many sources including

282

collaboration with ORD and OPP as well as through the TSCA systematic review process. In addition,

283

OSCPP is relying on the peer reviews provided by the National Academies of Sciences, Engineering,

284

and Medicine and the Human Studies Review Board on certain aspects of the human hazard assessment.

285

286

OPPT is using the inhalation unit risk for nasopharyngeal cancer as derived in the draft EPA IRIS

287

Toxicological Review of Formaldehyde – Inhalation (U.S. EPA, 2022b). Although inhaled

288

formaldehyde has been associated with multiple types of cancer in humans, including nasopharyngeal

289

and myeloid leukemia, the myeloid leukemia findings are not sufficient to develop quantitative estimates

290

of cancer risk. While there may be uncertainty on the extent to which other mechanisms contribute to the

291

carcinogenicity of formaldehyde, the IRIS assessment concluded that a mutagenic action contributes to

292

risk of nasopharyngeal cancer from inhaled formaldehyde. To account for the potential increased

293

susceptibility that may be associated with early life exposure to formaldehyde, OPPT modified this

294

cancer value using age-dependent adjustment factors for exposure scenarios that include early life.

295

296

Formaldehyde exposure is also associated with a range of respiratory and non-respiratory health effects

297

in humans—including reduced pulmonary function, increased asthma prevalence, decreased asthma

298

Formaldehyde can be emitted from many types of combustion, from naturally occurring wildfires and burning candles to

household appliance and industrial combustion turbines. These sources can also include tailpipe emissions (including cars,

trucks, and boats); and emissions from fires (including wildfires, accidental fires, and agricultural burning). Some

combustion activities could be included in the evaluation of other conditions of use under TSCA such as processing or other

similar industrial use. However, given the number of potential sources of formaldehyde production in the home, occupational

settings and in the environment, EPA did not consider formaldehyde from the combustion sources independent of other

TSCA COUs due to their abundant nature.

Formaldehyde is also largely found in the environment due to secondary formation of the chemical after degradation of

other compounds, for example when a chemical undergoes chemical reactions in the air and forms formaldehyde. Some

secondary formation may be a result of TSCA conditions of use but these cannot be distinguished from all other secondary

formations because they are so abundant.

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 10 of 151

control, allergy-related conditions, sensory irritation, male and female reproductive toxicity, and

299

developmental effects. OPPT is using a chronic point of departure for pulmonary function in children

300

derived from the draft EPA IRIS Toxicological Review of Formaldehyde – Inhalation. Sensory irritation

301

(e.g., eye irritation) observed in adults is the critical effect for non-cancer respiratory effects from

302

breathing formaldehyde for more than 15 minutes. Skin sensitization observed in adults is the critical

303

effect for assessing formaldehyde exposure via the dermal routes.

304

305

Oral hazard data are also available for formaldehyde but were not used in the risk assessment because

306

exposure was not expected.

307

308

Exposure for Workers and the General Population

309

Many data sources were used to evaluate exposures to humans (workers; consumers and general

310

population, both including children) from indoor and outdoor air as well as dermal exposures. These

311

include measured and model estimated concentrations data. There are many conditions of use and many

312

different exposure scenarios for each population assessed.

313

314

Workers

315

Worker exposure to formaldehyde via inhalation and dermal are expected to result in the highest

316

formaldehyde exposures among the assessed populations. Workplace concentrations of formaldehyde

317

vary based on activities performed (i.e., manufacturing, processing, industrial, and commercial settings).

318

Individuals in workplaces whose duties are not directly associated with manufacturing, processing, or

319

use of formaldehyde (i.e., occupational non-users [ONUs] such as supervisors] who may be near or

320

within the same workspace (i.e., breathing the same air) are also expected to be exposed to

321

formaldehyde at similar concentrations.

322

323

Inhalation exposures were estimated based largely on measured formaldehyde concentrations in

324

occupational settings. Monitoring data were available for many scenarios. However, monitoring data are

325

not available for three conditions of use in commercial settings and were thus modeled. These model

326

estimates generally fell within the range of monitored workplace concentrations. Across all conditions of

327

use, full work shift (8 to 12 hours) inhalation exposure estimates were between 7.5 to 17,353.3 µg/m

328

Peak inhalation estimates for workers were between 86 to 237,902 µg/m

across all conditions of use.

329

The highest inhalation exposure was based on modeled estimates for use of formulations containing

330

formaldehyde in automotive care products. Occupational exposure concentrations, as expected, are

331

generally higher than modeled and measured outdoor and indoor formaldehyde air concentrations. EPA

332

has an overall medium confidence in the reported exposure estimates because most of the values are

333

based on recent (1992 to 2020) real workplace monitoring data from multiple sources and therefore are

334

expected to be reflective of current industrial practices. The Agency does not have higher overall

335

confidence in the reported exposure estimates because the sources did not always provide supplemental

336

information such as worker activities and associated process conditions. Therefore, EPA made

337

assumptions in integrating monitoring data.

338

339

Short-term dermal exposures were estimated based on liquid contact with formulations containing

340

formaldehyde. Dermal exposure estimates ranged from 0.56 to 3,090 µg/cm

. The highest dermal

341

exposure was estimated during spray application of products such as paints and automotive care

342

products. EPA has medium confidence in the dermal exposure estimates because the estimates were

343

derived using a standard peer-review model based on measured data on the retention of liquids on the

344

skin surface. EPA does not have higher confidence in the reported values because the Agency did not

345

have monitored formaldehyde dermal exposure data to ground truth these exposure estimates.

346

347

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 11 of 151

General Population – Consumer Exposures in Residential Settings

348

Frequent users of products containing formaldehyde are anticipated to be the next highest population

349

effected due to its use in products and articles that are available to most people for purchase. Some

350

examples of these consumer products that contain formaldehyde include automotive care products;

351

fabrics, textiles, and leather products; and adhesives or sealants. Exposure estimates for these products

352

varies due to the different durations (or activity) of use along with formaldehyde amount acquired from

353

safety data sheets. This assessment considered concentrations of formaldehyde during and following use

354

of consumer products in residential settings. Specifically, peak (15-minute) and long-term (annual

355

average) inhalation exposures as well as short-term dermal exposures were estimated. For a subset of

356

conditions of use, longer-term or lifetime exposure scenarios were assessed based on known consumer

357

use activities.

358

359

Seven conditions of use were evaluated for peak inhalation exposures. Fifteen-minute concentration

360

estimates ranged from 1.72 to 2,500 µg/m

. The highest concentrations were for products like floor

361

covering, foam seating, and bedding. Four conditions of use were evaluated for chronic consumer

362

inhalation exposure to formaldehyde. These conditions of use were selected because the uses are

363

expected to be the most substantial contributors to long-term inhalation exposures based on the expected

364

consumer activity profile and formaldehyde concentrations in the product. Annual estimated

365

formaldehyde concentrations ranged from 0.04 to 23.83 µg/m

. The highest concentrations were for arts,

366

crafts, and hobby materials. EPA has medium confidence in the inhalation exposure estimates based on

367

the number of monitoring data sources, use of the EPA’s Exposure Factors Handbook (U.S. EPA, 2011)

368

and survey data on consumer behavior and activities, and chemical amounts report on product-specific

369

safety data sheets. Monitoring data that can be tied to specific consumer conditions are not available.

370

Formaldehyde concentrations from consumer products are expected to be represented in the available

371

indoor air monitoring data as an aggregate concentration with other consumer and indoor air sources.

372

373

Dermal short-term exposures for consumers were estimated based on contact with products containing

374

formaldehyde. Nineteen conditions of use were evaluated with estimated short-term formaldehyde

375

dermal loading rates ranging from 1.03 to 3,090 µg/cm

. The highest concentrations were estimated to

376

be for exterior car waxes and polishes followed by photographic processing solutions. EPA has medium

377

confidence in these estimates because there are no monitoring data available to ground truth these

378

concentration estimates.

379

380

General Population – Indoor Air Exposures in Residential and Vehicular Settings

381

There are many sources of formaldehyde within residences (homes and mobile homes) and vehicles. As

382

mentioned, these include both TSCA sources such as building materials, finishes such as wood flooring

383

and paint, and foam cushions on furniture, and other sources such as combustion sources like candles,

384

fireplaces, and stoves. Additionally, consumer products containing formaldehyde may also contribute to

385

indoor concentrations of formaldehyde.

386

387

The highest formaldehyde concentrations from TSCA sources are expected in newly constructed homes

388

and mobile homes. In these settings, multiple sources of formaldehyde contribute to total indoor air

389

concentrations especially during the peak product emission period when new formaldehyde containing

390

articles and products are introduced. These concentrations substantially diminish within the first two

391

years of the product life based on open literature data. The peak exposure to formaldehyde from these

392

products is expected to occur within one year of use or manufacture. Indoor air concentrations can also

393

be high when new materials like hardwood floors or wallpaper are installed in homes. Similarly, fabric

394

in new furniture may also release formaldehyde in indoor environments after being introduced.

395

Therefore, formaldehyde concentrations in indoor environments are expected to vary over longer time

396

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 12 of 151

periods (e.g., an individual’s lifetime) and are highly dependent on an individual’s propensity to move to

397

new homes as well as their purchasing behaviors.

398

399

Four conditions of use in both automobiles and homes were evaluated. The estimated average daily

400

concentrations of formaldehyde for these conditions of use ranged from 5.19 to 423 µg/m

. The highest

401

concentration comes from construction and building materials that cover large surface areas like

402

hardwood floors. These modeled concentrations represent high-end estimates for each condition of use.

403

Furthermore, many of the products that fall within this condition of use are subject to the new emission

404

standards under TSCA Title VI (15 U.S.C. §2697) which have not been fully implemented.

405

406

Monitoring data from the American Healthy Homes Survey II suggests that concentrations of

407

formaldehyde range from 0.27 to 124.2 µg/m

for all homes, with 95 percent of homes having

408

concentrations below 46 µg/m

. Thus, indoor exposures to formaldehyde are in general agreement

409

across available data and sources of formaldehyde; however, monitoring values represent all sources of

410

formaldehyde in indoor air (including sources that are not subject to TSCA) and cannot be attributed to a

411

single TSCA condition of use. Similarly, measured concentrations are not expected to reflect full

412

implementation of the TSCA Title VI (15 U.S.C. §2697), which have not been fully implemented as of

413

the time of publication of this draft risk evaluation. Therefore, it is reasonable to expect that less

414

formaldehyde will be released from many wood products in the future than occurred in the past.

415

416

EPA has medium confidence in the indoor air concentration estimates because the values are based on

417

product-specific emission rates and product-specific formulations of formaldehyde. However, EPA does

418

not have high confidence in the indoor air concentration estimates because available monitoring data

419

could not corroborate the full range of estimates. In addition, the Agency does not have high confidence

420

because (1) dissipation rates of formaldehyde cannot be determined for indoor air for all types of

421

furniture, wood, or other products; and (2) the available monitoring data cannot be directly tied to

422

specific products (e.g., wood and fabric products) and associated conditions of use.

423

424

General Population – Outdoor Air Exposures

425

As mentioned at the beginning of this summary, formaldehyde exposures in outdoor air (ambient air)

426

come from many sources including biogenic sources, secondary formation, and conditions of use.

427

Outdoor air exposures are lower than those in any other setting. However, TSCA condition of use

428

contributions are highly variable across the United States and only exceed other sources in specific

429

locations. The outdoor air exposure assessment only considered exposures from inhalation for

430

populations living within a half mile of release facilities. This assessment considered short-term (daily

431

average) and long-term (annual average) inhalation exposures. After evaluating all durations, only long-

432

term durations appeared to be substantial and relevant for this Draft TSCA Risk Evaluation. Estimated

433

annual ambient air concentration ranged from 0.0001 to 5.75 µg/m

. The highest potential exposures

434

come from operations with nonmetallic mineral product manufacturing as well as textile, apparel, and

435

leather manufacturing.

436

437

Monitoring data from Ambient Monitoring Technology Information Center, based on data collected

438

between 2015 to 2020, range from 0 to 60.1 µg/m

with a median of 1.6 µg/m

across more than

439

300,000 monitored values from 214 sites. Monitoring data could not be linked to specific conditions of

440

use.

441

442

Since monitored concentrations represent total aggregated concentrations from all contributing sources,

443

while these values are not directly comparable to IIOAC modeled concentrations alone, by considering

444

multiple data sources (modeled concentrations, biogenic and secondary sources), EPA found

445

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 13 of 151

considering these three primary contributors together represent a large portion of the total monitored

446

concentrations and does not result in concentrations outside of or well above any monitored

447

concentration.

448

449

EPA has high confidence in the outdoor air concentration estimates because the values are based on

450

reported formaldehyde releases from EPA databases, uses standard risk assessment approaches and

451

utilizes more refined models to better understand population and demographics near releasing facilities.

452

453

Risk Characterization

454

People are regularly exposed to formaldehyde in their workplace, in their vehicles, and in their homes.

455

People may also be exposed to formaldehyde due to its natural formation in the environment and as a

456

natural part of human metabolism.

457

458

Worker Risk Characterization

459

Based on available occupational monitoring data and exposure modeling estimates, worker exposure to

460

formaldehyde is expected to be higher than exposures from naturally occuring sources. This assessment

461

does not assume personal protective equipment use to account for a range of possible workplaces. Both

462

high-end and central tendency exposure estimates were used with the available hazard data to calculate

463

worker risk for acute, chronic non-cancer, and cancer inhalation effects along with the potential to cause

464

dermal sensitization.

465

466

Results indicate that effects to workers are more likely to be for acute and chronic non-cancer inhalation

467

effects. Workers may experience sensory irritation from short-term exposures and decreased pulmonary

468

function or other respiratory effects from longer-term exposures. The hazard values are largely based on

469

studies in children, but adults may also experience adverse effects at similar concentrations. At high-end

470

exposure scenarios, results indicate workers may also be at increased risk for nasopharyngeal cancer.

471

Cancer effects are based on human studies in occupational settings.

472

473

The risk estimates for occupational exposures reflect use of standard risk assessment approaches

474

considering an abundance of high-quality workplace monitoring data that clearly exceed concentrations

475

of formaldehyde from other sources including natural sources and human hazard data. Likewise, risk

476

estimates are generally consistent across central tendency and high-end exposure scenarios for workers.

477

While there are some uncertainties in the assessment, these uncertainties are not expected to change risk

478

estimates enough to shift the overall risk assessment conclusions but may be great enough to change risk

479

estimates for specific conditions of use.

480

481

Results indicate that effects to workers from dermal exposure that could lead to sensitization with

482

repeated exposure for all conditions of use except one. All exposure estimates were based on standard

483

modeling approaches including the assumption of the amount of liquid left on the skin after contact

484

which is not specific to formaldehyde. The hazard data for skin sensitization is based on controlled

485

human exposures in adult volunteers and is corroborated by animal and in vitro evidence. The dermal

486

sensitization data are based on controlled human exposures studies in adults.

487

488

Consumer Exposure Risk Characterization

489

Consumer risk estimates were calculated for acute, chronic non-cancer, and cancer inhalation effects, as

490

well as dermal sensitization.

491

492

Consumers may experience acute sensory irritation (eye irritation) when inhaling peak concentrations of

493

formaldehyde in their residences when using products that contain high amounts of formaldehyde for

494

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 14 of 151

short durations. These acute effects are based on a robust dataset of evidence for sensory irritation in

495

humans, including several high-quality controlled exposure studies with relevance for acute exposure

496

scenarios. The risk estimates reflect use of standard risk assessment approaches and best available data.

497

498

Consumers inhaling formaldehyde may also experience decreased pulmonary function and other chronic

499

effects when those products are used frequently. These effects are based on data from humans at

500

sensitive lifestages, but it is unclear whether exposure scenarios represent how all people use these

501

products and articles containing formaldehyde. EPA has substantial data on use patterns of these

502

products based on surveys conducted on consumer activities and behaviors. Similarly, EPA’s Exposures

503

Factors Handbook was used to support consumer exposure analyses. Lastly, safety data sheets were

504

used to identify concentrations of formaldehyde in consumer products. It is worth noting that

505

conservative estimates from these data sources may not represent exposures to all consumers using

506

products and articles containing formaldehyde. The risk estimates reflect use of standard risk assessment

507

approaches considering best available data for consumers who frequently use products containing

508

formaldehyde; but understanding the commonness of these practices has some uncertainty because it is

509

unclear how older data from surveys represents current behaviors and uses.

510

511

At high-end exposure scenarios, results indicate consumers may have increased risk for developing

512

nasopharyngeal cancer, but this is expected to be rare in the general population. The data for cancer

513

effects are based on human studies that are corroborated in animal studies. EPA believes these risk

514

estimates are for consumers who frequently use products containing formaldehyde over the course of

515

many years. However, the Agency does not have information on how common it is that consumers

516

would use these products for this length of time, and it is unclear how older data from surveys represents

517

current behaviors and uses.

518

519

Consumers using products containing formaldehyde may experience dermal sensitization after acute

520

exposures to their skin. The hazard data for skin sensitization is based on controlled human exposures in

521

adult volunteers and is corroborated by animal and in vitro evidence. Risk estimates for these dermal

522

exposures is based on estimated dermal loading from models. Monitoring data are not available to

523

determine how common these exposures may be for consumers. Thus, EPA has less certainty in how

524

common these exposures result in skin sensitization for consumers in the general population.

525

526

Indoor Air Exposure Characterization

527

Indoor air risk estimates were calculated for chronic non-cancer inhalation effects. People who are living

528

in homes where high concentrations are present may experience decreased pulmonary function and other

529

chronic effects. These effects are based on data from humans at sensitive lifestages. However, the

530

exposure scenarios where these effects are seen are mostly limited to homes where high surface area

531

products like hardwood floors and wallpaper may be introduced. Similarly, these effects may occur in

532

new homes and mobile homes where all new products may be contributing to high concentrations of

533

formaldehyde in air. As previously mentioned, the dissipation rate of formaldehyde from these TSCA

534

conditions of use could not be fully characterized. However, concentrations are anticipated to decrease

535

with time and ventilation. Generally, new products are expected to have substantially reduced

536

formaldehyde emissions within 2 years.

537

538

In addition to TSCA sources, other sources of formaldehyde may contribute substantially to indoor air

539

concentrations of formaldehyde. Formaldehyde concentrations from candles, incense, cooking, wood

540

combustion, and air cleaning devices fall within the range of formaldehyde concentrations from TSCA

541

conditions of use. Furthermore, the range of concentrations estimated fall within the range of available

542

monitoring data.

543

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 15 of 151

Many of these other sources of formaldehyde represent temporary emission sources, which may affect

544

the overall impact on indoor air quality. Further, qualities such as the frequency and duration of use of

545

these temporary formaldehyde sources (e.g., burning candles or the use of a fireplace), age of the home

546

and formaldehyde-containing home finishes and furnishings, and ventilation rate will impact the total

547

concentration of formaldehyde in indoor air and the relative contribution of TSCA and other sources to

548

the indoor air. Combined, the many factors that may contribute to overall indoor air concentrations and

549

relative concentrations from TSCA and other uses introduce a significant source of uncertainty in the

550

indoor air exposure assessment.

551

552

EPA has medium confidence in the conclusion of the inhalation risk assessment for indoor air. This is

553

because the assessment is based on product-specific emission rates, data, and standard methods. While

554

the monitoring data cannot be tied to individual conditions of use, it is expected to represent aggregate

555

exposure to formaldehyde resulting from multiple sources. As such, EPA has confidence it is not

556

underestimating formaldehyde exposure resulting from TSCA conditions of use or across all sources of

557

formaldehyde.

558

559

Ambient Air Risk Characterization

560

Based on modeling estimates, individuals of the general population living within half mile of a releasing

561

facility may be exposed to formaldehyde concentrations greater than naturally occuring sources in the

562

outdoor environment but are generally within the range of concentrations from natural sources like

563

biogenic sources. Acute, chronic non-cancer, and cancer inhalation risk estimates were calculated. Non-

564

cancer risk estimates are based on chronic respiratory effects observed in people at sensitive lifestages

565

and acute sensory irritation observed in controlled human exposures in adults. Cancer risk estimates are

566

based on effects observed in human studies and corroborated in animal studies.

567

568

Results indicate that the general population is not likely to experience sensory irritation from short-term

569

exposures or decreased pulmonary function or increased asthma prevalence from longer-term exposures

570

when compared to other formaldehyde exposures; however, in some locations some individuals may be

571

at increased risk for developing nasopharyngeal and other cancer types. However, this is contingent on

572

the assumption that an individual lives within a half mile of a releasing facility their entire life. EPA

573

conducted a higher tier analysis to identify locations where TSCA releases contributed to formaldehyde

574

concentrations exceeding background concentrations of formaldehyde.

575

576

EPA has high confidence in the conclusion of the inhalation risk assessment for the general population.

577

EPA has this confidence because the assessment is based on a large amount for formaldehyde reported

578

release data and standard methods. Furthermore, the range of concentrations estimated fall within the

579

range of available monitoring data. Although the monitoring data cannot be tied to individual conditions

580

of use, it is expected to represent aggregate exposure to formaldehyde resulting from multiple sources.

581

As such, EPA has confidence it is not underestimating formaldehyde exposure resulting from TSCA

582

conditions of use or across all sources of formaldehyde.

583

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 16 of 151

1 INTRODUCTION

584

1.1 Background

585

Formaldehyde is a high priority chemical undergoing the Toxic Substances Control Act (TSCA) risk

586

evaluation process after passage of the Frank R. Lautenberg Chemical Safety for the 21st Century Act in

587

2016. It is concurrently undergoing a hazard assessment in EPA’s Integrated Risk Information System

588

(IRIS) program and a risk assessment under the Federal Insecticide, Fungicide, and Rodenticide Act

589

(FIFRA). This Draft Human Health Risk Assessment is a TSCA-specific assessment that will serve to

590

support risk management needs by the Office of Pollution Prevention and Toxics (OPPT) and is one of

591

many documents comprising the draft formaldehyde risk evaluation.

592

593

In April 2022, EPA’s IRIS program released a draft Toxicological Review of Formaldehyde – Inhalation

594

(U.S. EPA, 2022b) (also called “draft IRIS assessment”) for public comment and peer review. OPPT

595

and OPP have relied upon the hazard conclusions and dose-response analysis presented in the draft IRIS

596

assessment for inhalation and have coordinated to evaluate additional information on environmental fate

597

and transport, human health hazard, and environmental hazard consistently across programs.

598

599

A list of the regulatory history of formaldehyde can be found in Appendix D of the Final Scope for the

600

Risk Evaluation for Formaldehyde 50-00-0 (U.S. EPA, 2020c), which includes regulation under the

601

Clean Air Act, Clean Water Act, Resource Conservation and Recovery Act, and other EPA regulatory

602

programs and non-EPA programs.

603

604

Following publication of the final scope document, EPA considered and reviewed reasonably available

605

information in a systematic and fit-for-purpose approach to develop this draft formaldehyde risk

606

evaluation, leverage existing EPA assessment work, collaborate across offices, rely on best available

607

science, and base it on the weight of the scientific evidence as required by EPA’s Risk Evaluation Rule

608

under TSCA. Reasonably available information was reviewed, and the quality evaluated in accordance

609

with EPA’s Draft Systematic Review Protocol Supporting TSCA Risk Evaluations for Chemical

610

Substances (U.S. EPA, 2021b), which underwent external peer review by the Science Advisory

611

Committee on Chemicals (SACC) in July 2021.

612

1.2 Risk Evaluation Scope

613

The draft formaldehyde risk evaluation comprises a series of modular assessments. Each module

614

contains sub-assessments that inform adjacent, “downstream” modules. A basic diagram showing the

615

layout and relationships of these assessments is provided below in Figure 1-1. In some cases, modular

616

assessments were completed jointly under TSCA and FIFRA. These modules are shown in dark gray.

617

This human health risk assessment is shaded blue. High level summaries of each relevant module are

618

presented in this risk assessment. Detailed information for each module can be found in the

619

corresponding documents/modules.

620

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 17 of 151

621

Figure 1-1. Risk Evaluation Document Summary Map

622

623

These modules leveraged the data and information sources already identified in the Final Scope of the

624

Risk Evaluation for Formaldehyde 50-00-0 (U.S. EPA, 2020c). OPPT conducted a comprehensive

625

search for “reasonably available information” to identify relevant formaldehyde data for use in the risk

626

evaluation. In some modules, data utilized were also located in collaboration with other EPA offices. As

627

previously noted, OPPT is relying on the EPA’s IRIS draft Toxicological Review of Formaldehyde –

628

Inhalation (U.S. EPA, 2022b) in the formaldehyde risk evaluation (shaded light gray in Figure 1-1). The

629

draft IRIS assessment is not part of the TSCA risk evaluation bundle. The approach used to identify

630

specific relevant risk assessment information was discipline-specific and is detailed in Systematic

631

Review Protocol for the Draft Formaldehyde Risk Evaluation (U.S. EPA, 2023a), or as otherwise noted

632

in the relevant modules.

633

Life Cycle and Production Volume

634

The Life Cycle Diagram (LCD)—which depicts the conditions of use that are within the scope of the

635

risk evaluation during various life cycle stages, including manufacturing, processing, use (industrial,

636

commercial, consumer), distribution and disposal—is shown below in Figure 1-2. The LCD has been

637

updated since it was included in the Final Scope of the Risk Evaluation for Formaldehyde CASRN 50-

638

00-0 (U.S. EPA, 2020c). The commercial and consumer uses for agricultural use products (non-

639

pesticidal) have been included; it was inadvertently omitted under the industrial, commercial, and

640

consumer uses lifecycle stage in the diagram in the final scope document (U.S. EPA, 2020c).

641

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 18 of 151

642

Figure 1-2. Lifecycle Diagram of Formaldehyde

643

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 19 of 151

The current domestic formaldehyde production volume is 453 million to 2.3 billion kg/year. This is

644

based on the Chemical Data Reporting (CDR) Rule under TSCA, which requires U.S. manufacturers

645

(including importers) to provide EPA with information on the chemicals they manufacture or import into

646

the United States every 4 years. For the 2020 CDR cycle, data collected for formaldehyde is further

647

detailed in the Use Report for Formaldehyde (CASRN 50-00-0) (EPA-HQ-OPPT-2018-0438).

648

Conditions of Use

649

The formaldehyde COUs included in the scope of the draft formaldehyde risk evaluation are reflected in

650

Table 1-1 and the LCD (Figure 1-2) and include industrial, commercial, and consumer applications such

651

as textiles, foam bedding/seating, semiconductors, resins, glues, composite wood products, paints,

652

coatings, plastics, rubber, resins, construction materials (including roofing), furniture, toys, and various

653

adhesives and sealants.

654

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 20 of 151

Table 1-1. Categories and Subcategories of Use and Corresponding Exposure Scenario in the Risk Evaluation for Formaldehyde

655

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

Manufacturing

Domestic manufacturing

Manufacturing of formaldehyde

Importing

Import and/or repackaging of

formaldehyde

Processing

Reactant

Adhesives and sealant chemicals in: Plastic and resin

manufacturing; Wood product manufacturing; Paint and coating

manufacturing; basic organic chemical manufacturing

Processing as a reactant

Processing

Reactant

Intermediate in: Pesticide, fertilizer, and other agricultural

chemical manufacturing; Petrochemical manufacturing; Soap,

cleaning compound, and toilet preparation manufacturing; All

other basic organic chemical manufacturing; Plastic materials

and resin manufacturing; Adhesive manufacturing; chemical

product and preparation manufacturing; Paper manufacturing;

Paint and coating manufacturing; Plastic products

manufacturing; Synthetic rubber manufacturing; Wood product

manufacturing; Construction; Agriculture, forestry, fishing, and

hunting

Processing

Reactant

Functional fluid in: oil and gas drilling, extraction, and support

activities

Processing

Reactant

Processing aids, specific to petroleum production in all other

basic chemical manufacturing

Processing

Reactant

Bleaching agent in wood product manufacturing

Processing

Reactant

Agricultural chemicals in agriculture, forestry, fishing, and

hunting

Processing

Incorporation into an article

Finishing agents in textiles, apparel, and leather manufacturing

Textile finishing

Leather tanning

Processing

Incorporation into an article

Paint additives and coating additives not described by other

categories in transportation equipment manufacturing (including

aerospace)

Use of coatings, paints, adhesives, or

sealants (non-spray applications)

Use of coatings, paints, adhesives, or

sealants (spray or unknown

applications)

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 21 of 151

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

Processing

Incorporation into an article

Additive in rubber product manufacturing

Rubber product manufacturing

Processing

Incorporation into an article

Adhesives and sealant chemicals in wood product

manufacturing; Plastic material and resin manufacturing

(including structural and fireworthy aerospace interiors);

Construction (including roofing materials); paper manufacturing

Composite wood product

manufacturing

Paper manufacturing

Plastic product manufacturing

Other composite material

manufacturing

Processing

Incorporation into a

formulation, mixture, or

reaction product

Petrochemical manufacturing, petroleum, lubricating oil and

grease manufacturing; Fuel and fuel additives; Lubricant and

lubricant additives; Basic organic chemical manufacturing; All

other petroleum and coal products manufacturing

Processing of formaldehyde into

formulations, mixtures, or reaction

products

Incorporation into a

formulation, mixture, or

reaction product

Asphalt, paving, roofing, and coating materials manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Solvents (which become part of a product formulation or

mixture) in paint and coating manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Processing aids, specific to petroleum production in: oil and gas

drilling, extraction, and support activities; chemical product and

preparation manufacturing; and basic inorganic chemical

manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Paint additives and coating additives not described by other

categories in: paint and coating manufacturing; Plastic material

and resin manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Intermediate in: all other basic chemical manufacturing; all

other chemical product and preparation manufacturing; plastic

material and resin manufacturing; oil and gas drilling,

extraction, and support activities; wholesale and retail trade

Incorporation into a

formulation, mixture, or

reaction product

Solid separation agents in miscellaneous manufacturing

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 22 of 151

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

Processing

Incorporation into a

formulation, mixture, or

reaction product

Agricultural chemicals (non-pesticidal) in: agriculture, forestry,

fishing, and hunting; pesticide, fertilizer, and other agricultural

chemical manufacturing

Processing of formaldehyde into

formulations, mixtures, or reaction

products

Incorporation into a

formulation, mixture, or

reaction product

Surface active agents in plastic material and resin manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Ion exchange agents in adhesive manufacturing and paint and

coating manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Lubricant and lubricant additive in adhesive manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Plating agents and surface treating agents in all other chemical

product and preparation manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Soap, cleaning compound, and toilet preparation manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Laboratory chemicals

Incorporation into a

formulation, mixture, or

reaction product

Adhesive and sealant chemical in adhesive manufacturing

Incorporation into a

formulation, mixture, or

reaction product

Bleaching agents in textile, apparel, and leather manufacturing

Repackaging

Sales to distributors for laboratory chemicals

Import and/or repackaging of

formaldehyde

Recycling

Distribution

Distribution in Commerce

Storage and retail stores

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 23 of 151

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

Industrial Use

Non-incorporative activities

Process aid in: oil and gas drilling, extraction, and support

activities; process aid specific to petroleum production,

hydraulic fracturing

Use of formaldehyde for oilfield well

production

Industrial Use

Non-incorporative activities

Used in: construction

Furniture manufacturing

Industrial Use

Non-incorporative activities

Oxidizing/reducing agent; processing aids, not otherwise listed

Processing aid

Industrial Use

Chemical substances in

industrial products

Paints and coatings; adhesives and sealants; lubricants

Use of coatings, paints, adhesives, or

sealants (non-spray applications)

Use of coatings, paints, adhesives, or

sealants (spray or unknown

applications)

Industrial use of lubricants

Foundries

Commercial

Use

Chemical substances in

furnishing treatment/care

products

Floor coverings; Foam seating and bedding products; Furniture

& furnishings including stone, plaster, cement, glass and

ceramic articles; metal articles; or rubber articles; Cleaning and

furniture care products; Leather conditioner; Leather tanning,

dye, finishing impregnation and care products; Textile (fabric)

dyes; Textile finishing and impregnating/ surface treatment

products.

Installation and demolition of

formaldehyde-based furnishings and

building/construction materials in

residential, public, and commercial

buildings, and other structures

Textile finishing

Leather tanning

Chemical substances in

treatment products

Water treatment products

Use of formulations containing

formaldehyde for water treatment

Chemical substances in

treatment/care products

Laundry and dishwashing products

Use of formulations containing

formaldehyde in laundry and

dishwashing products

Chemical substances in

construction, paint,

electrical, and metal

products

Adhesives and Sealants; Paint and coatings

Use of coatings, paints, adhesives, or

sealants (non-spray applications)

Use of coatings, paints, adhesives, or

sealants (spray or unknown

applications)

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 24 of 151

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

Commercial

Use

Chemical substances in

furnishing treatment/care

products

Construction and building materials covering large surface

areas, including wood articles; Construction and building

materials covering large surface areas, including paper articles;

metal articles; stone, plaster, cement, glass and ceramic articles

Installation and demolition of

formaldehyde-based furnishings and

building/construction materials in

residential, public and commercial

buildings, and other structures

Chemical substances in

electrical products

Machinery, mechanical appliances, electrical/electronic articles;

Other machinery, mechanical appliances, electronic/electronic

articles

Use of electronic and metal products

Chemical substances in

metal products

Construction and building materials covering large surface

areas, including metal articles

Chemical substances in

automotive and fuel

products

Automotive care products; Lubricants and greases; Fuels and

Category

Subcategories

Chemical substances in

products not described by

other codes

Laboratory chemicals

General laboratory use

Consumer Uses

Chemical substances in

furnishing treatment/care

products

Floor coverings; Foam seating and bedding products; Cleaning

and furniture care products; Furniture & furnishings including

stone, plaster, cement, glass and ceramic articles; metal articles;

or rubber articles

Varnishes and floor finishes

Plastic articles: foam insulation (living

room)

Plastic articles: foam insulation

(automobile)

Drain and toilet cleaners

Textile and leather finishing products

Furniture & furnishings – wood

articles: furniture

Consumer Uses

Chemical substances in

furnishing treatment/care

products

Fabric, textile, and leather products not covered elsewhere

(clothing)

Fabrics: furniture covers, car seat

covers, tablecloth (automobiles)

Fabrics: furniture covers, car seat

covers, tablecloth (living room)

Fabrics: clothing

Consumer Uses

Chemical substances in

treatment products

Water treatment products

Drinking water treatment

Consumer Uses

Chemical substances in

treatment/care products

Laundry and dishwashing products

Laundry detergent (liquid)

Hand Dishwashing Soap/ Liquid

detergent

Consumer Uses

Chemical substances in

construction, paint,

electrical, and metal

products

Adhesives and Sealants; Paint and coatings

Water-based wall paint

Solvent-based wall paint

Glues and adhesives, small scale

Caulk (Sealants)

Consumer Uses

Chemical substances in

construction, paint,

Construction and building materials covering large surface

areas, including wood articles; Construction and building

Building/construction materials – wood

articles: hardwood floors

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 26 of 151

Conditions of Use

Occupational/Consumer Exposure

Scenario Mapped to COU

Life Cycle

Stage

Category

Subcategories

electrical, and metal

products

materials covering large surface areas, including paper articles;

metal articles; stone, plaster, cement, glass and ceramic articles

Liquid concrete

Consumer Uses

Chemical substances in

electrical products

Machinery, mechanical appliances, electrical/electronic articles;

Other machinery, mechanical appliances, electronic/electronic

articles

Electronic appliances

Consumer Uses

Chemical substances in

automotive and fuel

products

Automotive care products; Lubricants and greases; Fuels and

Category

Subcategories

The repackaging scenario covers only those sites that purchase formaldehyde or formaldehyde containing products from domestic and/or foreign suppliers

and repackage the formaldehyde from bulk containers into smaller containers for resale. Sites that import and directly process/use formaldehyde are assessed in

the relevant occupational exposure scenario (OES). Sites that that import and either directly ship to a customer site for processing or use or warehouse the

imported formaldehyde and then ship to customers without repackaging are assumed to have no exposures or releases and only the processing/use of

formaldehyde at the customer sites are assessed in the relevant OES.

Each of the TSCA COU of formaldehyde may generate waste streams of the chemical that are collected and transported to third-party sites for disposal,

treatment, or recycling. Industrial sites that treat, dispose, or directly discharge onsite wastes that they themselves generate are assessed in each COU

assessment. This section only assesses wastes of formaldehyde that are generated during a COU and sent to a third-party site for treatment, disposal, or

recycling.

656

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 28 of 151

Other Sources of Formaldehyde in Air

657

Formaldehyde is ubiquitous in both indoor and outdoor (ambient) air because it is formed naturally in

658

the environment and from numerous anthropogenic sources, which include both TSCA (Section 1.2.2)

659

and other activities. As a result, people are routinely exposed to formaldehyde in indoor and outdoor air,

660

with indoor air generally having higher concentrations than outdoor air. Robust monitoring data are

661

available to estimate the concentrations of formaldehyde across common outdoor and indoor

662

environments. However, attributing measured concentrations to TSCA versus other sources is complex.

663

This section will provide an overview of these data sources and seeks to differentiate between sources

664

when possible. This section is not intended to be a comprehensive review of the scientific literature on

665

this topic but instead provides context for understanding and interpreting the exposures of formaldehyde

666

from a variety of sources as part of risk characterization and risk determination of COUs under TSCA.

667

668

Formaldehyde has been measured in outdoor air across the country. EPA’s Ambient Monitoring

669

Technology Information Center (AMTIC) maintains a database of spatially and temporally diverse air

670

quality monitoring data that meet specified collection and quality assurance criteria. The Agency used

671

monitoring data extracted from EPA’s AMTIC (U.S. EPA, 2022a) from 2015 through 2021 to

672

contextualize modeled values as well as characterize total aggregate exposures to formaldehyde from all

673

possible contributing sources—including sources associated with TSCA COUs and other sources out of

674

scope for this assessment and not associated with TSCA COUs (e.g., biogenic sources (decay of organic

675

matter), secondary formation, combustion byproduct formation, other byproduct formation, mobile

676

sources, and others).These data are described in detail in Sections 2.4.1 and 3.3.2 of the Draft Ambient

677

Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024a). In addition, satellite data have measured

678

formaldehyde concentrations across the United States, providing insights on temporal and geographic

679

trends that help to characterize ambient formaldehyde concentrations (Wang et al., 2022; Harkey et al.,

680

2021; Zhu et al., 2017).

681

682

Comprehensive modeling efforts have been undertaken to characterize formaldehyde concentrations that

683

vary across the county. EPA’s AirToxScreen is one example that uses release data with chemical

684

transport and dispersion models to estimate average annual outdoor ambient air concentrations of air

685

toxics across the U.S. and is validated against available monitoring data. For formaldehyde, this model

686

estimates concentrations from different sources contributing to ambient air concentrations including

687

biogenic sources, secondary formation, and point sources. Other sources of formaldehyde are included

688

but may not be relevant to the scope of this draft risk evaluation for formaldehyde. Accordingly, the

689

2019 AirToxScreen estimates that secondary formation of formaldehyde accounts for 80 percent of

690

formaldehyde in ambient air and direct biogenic sources contribute 15 percent. Based on the 2019

691

AirToxScreen estimates, the calculated ninety-fifth percentile biogenic concentration of formaldehyde in

692

ambient air was 0.28 µg/m

(e.g., Ninety-five percent of estimated concentrations of formaldehyde in

693

ambient air attributable to biogenic sources based on the 2019 AirToxScreen data all biogenic sources of

694

formaldehyde are below 0.28 µg/m

.).

695

696

Much like outdoor air, many efforts have been made to characterize formaldehyde in the indoor

697

environment. Draft data from a recent national survey provides a representative sample of formaldehyde

698

concentrations in indoor air, showing average residential levels an order of magnitude higher than

699

outdoor concentrations. The American Healthy Homes Survey II (AHHS II) survey, sponsored by the

700

U.S. Department of Housing and Urban Development (HUD) along with EPA, was conducted from

701

March 2018 through June 2019 and measured indoor air concentrations of formaldehyde in U.S. homes

702

of various ages, types, conditions, and climates (QuanTech, 2021). Across all housing, the weighted-

703

mean concentration is 23.2 µg/m

(95% confidence interval 21.6–25.2 µg/m

) with 10 percent of homes

704

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 29 of 151

higher than 41.8 µg/m

. Formaldehyde is introduced into residential indoor air from numerous TSCA

705

sources (e.g., building materials, finishes such as flooring and paint, and furniture) and other sources

706

(e.g., fireplaces, gas stoves, candles, photocatalytic air purifiers, and tobacco use). The TSCA sources

707

are expected to consistently release formaldehyde over long periods of time, with release rates

708

decreasing over time as the materials age. In contrast, many of the other sources are temporary emission

709

sources and contribute formaldehyde to the indoor air intermittently. Overall, due to differences in the

710

ages of building materials, home finishes, and furnishings and differences in presence and use patterns

711

of other formaldehyde sources in the residence, the relative contributions of formaldehyde from TSCA

712

and other sources to residential indoor air varies both among homes and over time within a single home.

713

Thus, despite the availability of quality monitoring data, it remains difficult to discern source

714

apportionment for the residential environment and there are uncertainties related to assessing exposures

715

tied to specific TSCA COUs based on this monitoring data. OPPT will solicit comment from the SACC

716

and the public on additional sources of information that could inform the attribution of other sources of

717

formaldehyde to support risk characterization.

718

1.3 Chemistry, Fate, and Transport Assessment Summary

719

EPA considered reasonably available information identified by the Agency through its systematic

720

review process under TSCA and submissions under FIFRA to characterize the physical and chemical

721

properties as well as the environmental fate and transport of formaldehyde. This was done as a joint

722

effort with the OPP. Physical and chemical properties of formaldehyde, as well as some known

723

environmental transformation products (methylene glycol, paraformaldehyde), are provided in Table

724

1-2. Formaldehyde is expected to be a gas under most environmental conditions. Due to the reactivity of

725

formaldehyde, it is not expected to be present in most environmental media but may be abundant in air

726

due to continual release from multiple sources including from TSCA releases, biogenic sources, and

727

formation from secondary sources.

728

729

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 30 of 151

Table 1-2. Physical and Chemical Properties of Formaldehyde and Select Transformation

730

Products

731

Chemical Properties

Formaldehyde

Methylene Glycol

Paraformaldehyde

Molecular formula

(OH)

HO(CH

(n = 8–100)

CASRN

50-00-0

463-57-0

30525-89-4

Molecular weight

30.026 g/mol

48.02 g/mol

(30.03)

g/mol (Varies)

Physical form

Colorless gas

Colorless liquid

White crystalline solid

Melting point

−92.0 to −118.3 °C

−43.8 °C

120 to 170 °C

Boiling point

−19.5 °C

131.6 °C

None identified

Density

0.815 g/cm

at 20 °C

1.20 g/cm

1.46 g/cm

at 15 °C

Vapor pressure

3,890 mmHg at 25 °C

3.11 mmHg at 25°C

1.45 mmHg @ 25 °C

Vapor density

1.067 (air = 1)

None identified

1.03 (air = 1)

Water solubility

<55%; 400 to 550 g/L

Miscible

Insoluble

Octanol/water partition

coefficient (log K

)

0.35

−0.79

N/A

Henry’s Law constant

3.37E−07 atm/m

·mol at

25 °C

1.65E−07

atm/m

·mol at 25 °C

N/A

Physical and chemical properties for formaldehyde, methylene glycol, and paraformaldehyde are considered

best estimates. Because the chemical substance often exists in a mixture at varying concentrations, these

properties can vary based on the equilibration with other chemical substances present.

732

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 31 of 151

In water, formaldehyde quickly hydrates to form methylene glycol, which can polymerize to form

733

oligomers of various chain lengths and paraformaldehyde (U.S. EPA, 2024b)—all structurally different

734

compounds when compared to formaldehyde (Figure 1-3). Formaldehyde is not expected to be found in

735

aquatic systems (U.S. EPA, 2024e).

736

737

738

Figure 1-3. Chemical Equilibria for Formaldehyde in Aqueous Solutions

739

Adapted from (Boyer et al., 2013).

740

741

In soil, formaldehyde is also expected to quickly transform to products that are structurally dissimilar to

742

parent formaldehyde; thus, formaldehyde is not expected to be found in soil (U.S. EPA, 2024b).

743

Formaldehyde can be formed in the early stages of plant residue decomposition in soil and is reportedly

744

degraded by bacteria in the soil (U.S. EPA, 2024b). Formaldehyde is expected to undergo abiotic

745

(hydration and nucleophilic addition) chemical reactions in soils to form other compounds.

746

747

In air, formaldehyde is susceptible to direct and indirect photolysis; however, it may be present in air

748

environments with low or no sunlight (e.g., nighttime, indoor). As such, the primary exposure route for

749

formaldehyde is expected to be the air pathway (U.S. EPA, 2024e). More specifically, the half-life of

750

formaldehyde in air depends on the intensity and duration of sunlight and ambient conditions such as

751

temperature and humidity. Under direct sunlight, formaldehyde will undergo photolysis with a half-life

752

up to 4 hours yielding mainly hydroperoxyl radical (HO

), carbon monoxide (CO), and hydrogen (H

753

In the absence of sunlight, formaldehyde can persist with a half-life up to 114 days.

754

755

Due to the physical and chemical properties of formaldehyde including a log K

(0.35), which is

756

associated with low bioconcentration and bioaccumulation are not expected (U.S. EPA, 2024b).

757

Therefore, human exposure to formaldehyde via consumption of fish was not expected and therefore not

758

assessed.

759

760

EPA has high confidence in the overall fate and transport profile of formaldehyde and

761

paraformaldehyde; however, EPA is less confident in the overall fate and transport of the transformation

762

products methylene glycol and poly(oxy)methylene glycol. Key sources of uncertainty for this

763

assessment are related to formaldehyde equilibrium in various media and subsequent transformation. In

764

cases where there are little fate and transport data, EPA relied on physical and chemical properties to

765

describe the expected fate and transport of the respective chemical. As such, although EPA has some

766

uncertainty in the precision of a specific parameter value, it has confidence in the overall fate and

767

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 32 of 151

transport profile of formaldehyde. Additional details can be found in the Chemistry, Fate, and Transport

768

Assessment for Formaldehyde (U.S. EPA, 2024b).

769

1.4 Environmental Release Assessment

770

Formaldehyde is directly released to all three environmental media (air, land, and water) from TSCA

771

COUs (U.S. EPA, 2024g). It is also released to the environment during regulated other uses (e.g., use as

772

a pesticide and U.S. Food and Drug Administration uses), as a transformation product of different parent

773

chemicals, and from combustion sources.

774

775

EPA reviewed release data from the Toxics Release Inventory or TRI (data from 2016 to 2021),

776

Discharge Monitoring Report (DMR; data from 2016 to 2021), and the 2017 National Emissions

777

Inventory (NEI) to identify releases to the environment that are relevant to the formaldehyde TSCA

778

COUs, as stated in Table 1-1. Based on a review of these databases, waste streams containing

779

formaldehyde are directly discharged to surface water, indirectly discharged to publicly owned treatment

780

works (POTW) or other wastewater treatment (WWT) plants, disposed of via different land disposal

781

methods (e.g., landfills, underground injection), sent to incineration, and emitted via fugitive and stack

782

releases.

783

784

Based on TRI and DMR reporting from 2016 to 2021, less than 150,000 kg each year of formaldehyde

785

are directly discharged to surface water for TSCA-related activities based on reporting from 168

786

facilities. Approximately 2 million kg each year are transferred to POTW/WWT plants for treatment

787

based on reporting from 168 facilities (U.S. EPA, 2024g). For these wastewater streams transferred to

788

POTW or WWT plants, biological wastewater treatment systems have shown a mean removal efficiency

789

of 99.9 percent for formaldehyde based on literature and 92 percent removal of methylene glycol

790

through biodegredation based on EPISuite™ estimates (U.S. EPA, 2024b). These disposal methods

791

provide additional time for formaldehyde and methylene glycol to further transform to chemically

792

dissimilar products in the presence of water and chemical, biological, and physical treatment processes

793

prior to being discharged to surface water.

794

795

Based on TRI reporting from 2016 to 2021, most waste of formaldehyde is disposed of via land disposal

796

methods. The most significant method of land disposal of formaldehyde is via underground injection

797

with 22 sites disposing of more than 5 million kg of formaldehyde annually. The amount of waste

798

reported to be disposed of in RCRA Subtitle C landfills and other landfills varies across the reporting

799

years from 200 facilities reporting a total of 423,517 kg/year in 2016 to the most recent year (RY2021)

800

of 127,348 kg/year. Other land disposal methods (e.g., surface impoundments, solidification/

801

stabilization) are also reported at lower levels. Formaldehyde is not expected to persist in water or soils;

802

thus, EPA determined that additional analyses of releases to water or land were not needed and targeted

803

its review of release information to fugitive and stack emissions of formaldehyde from TSCA COUs.

804

805

EPA identified more than 150,000-point source emission data (includes unit-level estimates) for

806

formaldehyde across the two EPA databases (TRI data from 2016 to 2021and 2017 NEI). To

807

characterize this amount of data, EPA utilized the self-reported NAICS codes to assign sites into CDR

808

industrial sectors (IS). These industrial sectors can be directly correlated with the TSCA COUs, as

809

further discussed in the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA, 2024g).

810

Most TSCA COUs indicate one or more industrial sectors, and in some cases an industrial sector can

811

appear in more than one TSCA COU. Therefore, an industrial sector may be associated with multiple

812

formaldehyde TSCA COUs.

813

814

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 33 of 151

For this fit-for-purpose TSCA risk assessment, EPA targeted its review of environmental releases to

815

point sources, and did not review the road, nonroad, and other automotive exhaust information

816

identified, as formaldehyde produced from combustion sources is not assessed as an independent COU

817

subcategory in this draft risk evaluation. EPA focused its environmental release assessment on total

818

facility emissions which can include emission from both uses of formaldehyde and combustion sources

819

at the same facility or, potentially, only combustion sources from that facility.

820

821

EPA categorizes the facilities and corresponding release information by industrial sectors that can be

822

directly correlated to the TSCA industrial COUs. For commercial TSCA COUs, EPA used professional

823

judgement to assign the industrial sector to commercial TSCA COUs, where applicable. For a few

824

TSCA COUs (Commercial use – chemical substances in treatment/care products – laundry and

825

dishwashing products; Commercial use – chemical substances in treatment products – water treatment

826

products; Commercial use – chemical substances in outdoor use products – explosive materials; and

827

Commercial use – chemical substances in products not described by other codes – other: laboratory

828

chemicals), releases were only qualitatively assessed due to limited use information. Additional details

829

are provided in the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA, 2024g).

830

831

In the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA, 2024g), EPA identified

832

approximately 800 TRI facilities between 2016 and 2021 and approximately 50,000 NEI facilities in

833

2017 with reported air releases of formaldehyde (U.S. EPA, 2024g). From these facilities, EPA

834

identified the maximum release reported through TRI was 10,161 kg/year-site (IS: Paper

835

Manufacturing) for a fugitive release reported in 2019 and 158,757 kg/year-site (IS: Wood Product

836

Manufacturing) for a stack release reported in 2017. The NEI program identified sites reporting as high

837

as 138,205 kg/year-site (IS: Wholesale and Retail Trade) for fugitive releases and 1,412,023 kg/year-site

838

(IS: Oil and gas drilling, extraction and support activities) for stack releases reporting in 2017, in which

839

the higher releases are associated with sectors not required to report to TRI. The high release sites in

840

NEI program were associated with natural gas compressor stations and airport operations, which EPA

841

expects is due to formaldehyde produced from combustion sources. EPA analyzed the release

842

information by the industrial sector, providing the minimum, median, 95th percentile, and maximum

843

releases across the entire distribution of reported releases within each industrial sector, as further

844

discussed in the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA, 2024g)

845

846

In general, EPA has medium to high confidence in environmental releases for industrial TSCA COUs

847

and low to medium confidence in commercial TSCA COUs.

EPA has high data quality ratings for TRI

848

and NEI, which are supported by numerous facility-reported estimates. Some sites that emit

849

formaldehyde may not be included in these databases if the release does not meet the reporting criteria

850

for the respective program. EPA used total emissions per site, which may combine formaldehyde

851

emissions from multiple TSCA COUs if the site’s formaldehyde-generating processes are applicable to

852

more than one TSCA COU. For example, a facility may manufacture formaldehyde as well as process

853

formaldehyde as a reactant. In some cases, the formaldehyde-generating process may also fall outside of

854

scope of the draft risk evaluation.

855

856

EPA categorizes the facilities and corresponding release information by industrial sectors that can be

857

directly correlated to the TSCA industrial COUs. For commercial COUs, EPA used professional

858

judgement to assign the industrial sector to commercial COUs, where applicable. For a few COUs

859

(Commercial use – chemical substances in treatment/care products – laundry and dishwashing products;

860

TSCA COUs that are included under the life cycle stage of manufacturing, processing, and industrial use.

TSCA COUs that are included under the life cycle stage of commercial uses.

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 34 of 151

Commercial use – chemical substances in treatment products – water treatment products; Commercial

861

use – chemical substances in outdoor use products – explosive materials; and Commercial use –

862

chemical substances in products not described by other codes – other: laboratory chemicals), releases

863

were only qualitatively assessed due to limited use information. For distribution in commerce,

864

formaldehyde released accidently during transit has occurred based on available information, but it was

865

not quantified due to uncertainties in the frequency or volume that may occur in the future. Additional

866

details are provided in the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA,

867

2024g).

868

1.5 Human Health Assessment Scope

869

Generally, EPA expects inhalation to be a major route of exposure for occupational, consumer, indoor

870

air, and ambient air based on the volatility and presence of formaldehyde in air. Dermal sensitization

871

from formaldehyde exposure is a rapid effect. Thus, for occupational and consumer COUs where

872

dermal contact to formaldehyde may occur, EPA expects the dermal route to be another significant

873

route of exposure to formaldehyde.

874

875

A quantitative assessment of the water pathway was not conducted in this risk assessment given the

876

relatively limited release of formaldehyde directly to surface water, and due to the rapid transformation

877

of formaldehyde in water based on the physical and chemical properties governing the environmental

878

fate of formaldehyde in water. Water monitoring data, while limited, demonstrate formaldehyde is not

879

detected in water as described in more detail in the environmental exposure assessment (U.S. EPA,

880

2024e). Based on these lines of evidence, EPA does not expect human exposure to formaldehyde will

881

occur via surface water. In addition, formaldehyde is not expected to persist in land or leach to

882

groundwater that may be sourced for drinking water based on the physical and chemical properties

883

governing the environmental fate of formaldehyde in land. Therefore, EPA does not expect human

884

exposure to formaldehyde will occur via soil, land, or groundwater.

885

Conceptual Exposure Models

886

1.5.1.1 Industrial and Commercial Activities and Uses

887

The conceptual model in Figure 1-4 presents the exposure pathways, exposure routes and hazards to

888

people from industrial and commercial activities and uses of formaldehyde. EPA evaluated exposures to

889

workers and occupational non-users (ONU) via inhalation routes and exposures to workers via dermal

890

routes, as shown in Figure 1-4. Oral exposure may occur through wood or textile dust that deposit in the

891

upper respiratory tract that is then ingested; however, formaldehyde will continue to evaporate and there

892

is uncertainty on the amount inhaled that is ingested. For this draft risk evaluation, these exposures were

893

evaluated as an inhalation exposure.

894

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 35 of 151

895

896

Figure 1-4. Conceptual Model for Industrial and Commercial Activities and Uses: Potential Exposure and Hazards

897

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 36 of 151

Note that fugitive air emissions, as described in Figure 1-4, are those that are not stack emissions and

898

include fugitive equipment leaks from valves, pump seals, flanges, compressors, sampling connections

899

and open-ended lines; evaporative losses from surface impoundment and spills; and releases from

900

building ventilation systems.

901

1.5.1.2 Consumer Activities and Uses

902

Formaldehyde is found in consumer products and articles that are readily available for public purchase

903

at common retailers and through online shopping venues. Formaldehyde may be either a chemical

904

ingredient in a consumer product or a component in material(s) utilized in the manufacturing of

905

consumer products or articles (adhesives, resins, glues, etc.) or both. Use of such product is expected to

906

result in exposures to both consumers who use a product (consumer user) and bystanders (individuals

907

who are not directly using a product but are exposed while the product is being used by someone else).

908

909

Figure 1-5 presents the conceptual model for consumer activities and uses that are in scope for the

910

TSCA formaldehyde risk evaluation. Formaldehyde-containing consumer products include textiles,

911

foam bedding/seating, semiconductors, resins, glues, composite wood products, paints, coatings,

912

plastics, rubber, resins, construction materials (including roofing), furniture, toys, and various adhesives

913

and sealants. EPA identified these formaldehyde COUs from information reported to EPA through CDR

914

and TRI reporting, published literature, and consultation with stakeholders for products currently in

915

production or not discontinued.

916

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 37 of 151

917

918

Figure 1-5. Formaldehyde Conceptual Model for Consumer Activities and Uses: Potential Exposures and Hazards

919

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 38 of 151

Some consumer products assessed may also have commercial applications. See Table 1-1 for categories

920

and subcategories of COUs. Inhalation is the primary expected route of exposure for formaldehyde

921

resulting from consumer activities, however, dermal exposures are also expected. EPA considered

922

potential oral exposure pathways associated with TSCA COUs, including lawn and garden products and

923

oral mouthing behaviors in infants and young children. However, because EPA lacks sufficient data to

924

quantify exposures and risks for any of these pathways, oral exposures were qualitatively assessed for

925

relevant COUs (e.g., lawn and garden products). Section 2.2 for the Consumer Exposure Module (U.S.

926

EPA, 2024d) provides more detail about the COUs within the scope of this draft risk evaluation.

927

1.5.1.3 Indoor Air Exposures

928

EPA expects formaldehyde exposure to occur in the indoor air environments from several sources via air

929

including from off-gassing of formaldehyde from various consumer articles. The separation of the

930

consumer exposure assessment and the indoor air exposure assessment is intentional; each assessment

931

represents a different context of exposures. The conceptual model in Figure 1-6 presents the exposure

932

pathways, exposure routes and hazards to people from emitters of formaldehyde in indoor air. For

933

example, a passenger may be exposed to formaldehyde through inhalation for the duration of a taxi ride

934

due to formaldehyde off-gassing to air from seat covers within the vehicle.

935

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 39 of 151

936

937

Figure 1-6. Formaldehyde Conceptual Model for Indoor Air: Residential Exposures and Hazards from Article Off-Gassing

938

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 40 of 151

1.5.1.4 General Population Exposures from Environmental Releases

939

Environmental releases of formaldehyde are reported to occur into the ambient air, ambient water, and

940

land environmental media. (U.S. EPA, 2024g). General population exposures to formaldehyde occur

941

when individuals encounter these releases through interaction with one or more of these media (e.g.,

942

breathing ambient air into the body (inhalation), incidental skin contact through swimming (dermal), or

943

ingestion of soil (oral)).

944

945

Figure 1-7 provides a detailed conceptual model of all pathways and all routes of exposure by which

946

exposures to the general population may occur. While releases are reported to all three environmental

947

media, formaldehyde is not expected to be present in water or land based on the chemical, fate, and

948

transport properties of formaldehyde as described in the Draft Chemistry, Fate, and Transport

949

Assessment for Formaldehyde (U.S. EPA, 2024b) and discussed in Section 1.2.3. As such, EPA does not

950

expect general population exposure to formaldehyde to occur via either the water or land media and

951

therefore did not quantitatively assess exposures via these media in this draft risk assessment. This is

952

depicted in Figure 1-7 by the dashed lines.

953

954

While formaldehyde is susceptible to direct and indirect photolysis, it is expected to be present in the

955

ambient air for at least several hours in direct sunlight (and many more hours in no sunlight) based on

956

the chemical, fate, and transport properties of formaldehyde as described in the Draft Chemistry, Fate,

957

and Transport Assessment for Formaldehyde (U.S. EPA, 2024b) and Section 1.2.3. Formaldehyde is

958

consistently present in ambient air based on monitoring and testing programs implemented under the

959

Clean Air Act and other EPA programs and statutes. Additional modeling and data from the 2019

960

AirToxScreen supports the ubiquity and consistent presence of formaldehyde in ambient air from

961

multiple sources (including TSCA and other sources). Considering these multiple lines of evidence,

962

EPA expects general population exposure to formaldehyde from industrial releases to be predominantly

963

via the ambient air pathway. Therefore, EPA quantitatively assessed the ambient air pathway in this risk

964

assessment. This is depicted in Figure 1-7 by a solid line.

965

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 41 of 151

966

967

Figure 1-7. Formaldehyde Conceptual Model for Environmental Releases and Wastes: General Population Exposures and Hazards

968

969

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 42 of 151

Figure 1-8 provides a simplified visual representation of industrial releases to ambient air by which

970

exposure to the general population occurs. In general, formaldehyde is released from industrial facilities

971

as uncontrolled fugitive releases (e.g., process equipment leaks, process vents, building windows,

972

building doors, roof vents) and stack releases that may be either uncontrolled (e.g., direct releases out a

973

stack) or controlled with some pollution control device prior to release to the ambient air (e.g.,

974

baghouse, scrubber, thermal oxidizer). Once released to the ambient air, the releases move off-site into

975

the surrounding ambient air where exposure to the general population occurs through inhalation. For

976

purposes of this risk assessment, EPA focuses on formaldehyde exposures to individuals living nearby

977

industrial facilities associated with TSCA COUs that are releasing formaldehyde to the ambient air.

978

979

980

Figure 1-8. Industrial Releases to the Environment and Pathways by Which Exposures to

981

the General Population May Occur

982

Potentially Exposed or Susceptible Subpopulations

983

This assessment considers potentially exposed or susceptible subpopulation (PESS), a group of

984

individuals within the general population identified by the Administrator who, due to either greater

985

susceptibility or greater exposure, may be at greater risk than the general population of adverse health

986

effects from exposure to a chemical substance or mixture. There are many factors that may contribute to

987

increased exposure or biological susceptibility to a chemical, including life stage (e.g., infants, children,

988

pregnant women, elderly), pre-existing disease, lifestyle activities (e.g., smoking, physical activity),

989

occupational and consumer exposures (including workers and ONUs, consumers and bystanders),

990

geographic factors (living in proximity to a large industrial source of formaldehyde), socio-demographic

991

factors, unique activities (e.g., subsistence fishing), aggregate exposures, and other chemical and non-

992

chemical stressors.

993

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 43 of 151

Considerations related to PESS may influence the selection of relevant exposure pathways, the

994

sensitivity of derived hazard values, the inclusion of populations, and/or the discussion of uncertainties

995

throughout the assessment.

996

997

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 44 of 151

2 HUMAN EXPOSURE ASSESSMENT SUMMARY

998

This section summarizes the formaldehyde exposures to occupational workers, ONUs, consumers,

999

bystanders, and general population from both indoor air and ambient air. Detailed information

1000

supporting each subsection are available in the associated technical support modules included as

1001

supplemental files to this draft human health risk assessment for formaldehyde.

1002

1003

Each exposure assessment considers peak and long-term inhalation exposures. When available, the

1004

highest 15-minute average concentrations are used to represent peak exposures while annual average

1005

concentrations or 8-hour time-weighted averages (TWA) are used to represent longer-term exposure

1006

durations. The long-term exposure duration depends on the exposure scenario being assessed.

1007

Specifically, exposure durations for cancer assessment are based on 31 (central tendency) and 40 (high-

1008

end) working years for occupational exposure. Exposure durations for cancer assessment are based on

1009

12- or 57-year residency time and 78-year lifetime exposure for consumer and general population. Acute

1010

dermal exposures were estimated for workers and consumers and are based on short-term durations, see

1011

Appendix G for additional information on the dermal approaches.

1012

1013

Each exposure assessment integrates modeling methodologies previously peer reviewed as well as

1014

monitoring data to assess exposures to the respective populations. The exposure assessment also

1015

integrates information from the Draft Chemistry, Fate, and Transport Assessment for Formaldehyde

1016

(U.S. EPA, 2024b) and the Draft Environmental Release Assessment for Formaldehyde (U.S. EPA,

1017

2024g).

1018

1019

Due to the magnitude of available scientific information on formaldehyde coupled with its complex

1020

toxicology and exposure profiles, EPA acknowledges that the evaluation of formaldehyde exposure is

1021

challenging. The Agency is at a critical point in the development of the draft risk evaluation where

1022

SACC and public input will be essential. For example, OPPT will seek input on its use of inputs and

1023

assumptions in the exposure assessments for occupational, consumer, outdoor air, and indoor air

1024

scenarios, in part to understand whether its approach may compound one conservative assumption upon

1025

another in a manner that leads to unrealistic or un-addressable outcomes. Following SACC and public

1026

comments, EPA will revise the draft risk evaluation and issue a final evaluation that will include a

1027

determination of whether, under its conditions of use, formaldehyde presents unreasonable risk to health

1028

or the environment.

1029

2.1 Occupational Exposure Assessment

1030

EPA identified 49 TSCA COUs under manufacturing, processing, industrial/commercial uses, and

1031

disposal. In the Draft Occupational Exposure Assessment for Formaldehyde (U.S. EPA, 2024k), EPA

1032

evaluated occupational exposure scenarios (OESs) based on the COUs with expected worker activities,

1033

inhalation exposure estimates, and dermal exposure estimates for each OES (U.S. EPA, 2024k). Several

1034

of the TSCA COU categories and subcategories were grouped and assessed together into a single OES

1035

due to similarities in the processes or lack of data to differentiate between them. This grouping

1036

minimized repetitive assessments. In other cases, TSCA COU subcategories were further delineated into

1037

multiple OESs based on expected differences in processes and associated releases/exposure potentials

1038

between facilities. This resulted in assessing 36 OESs for inhalation and dermal exposure. For additional

1039

details on the approaches and results, please refer to Draft Occupational Exposure Assessment for

1040

Formaldehyde (U.S. EPA, 2024k).

1041

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 45 of 151

Inhalation Exposure Assessment

1042

To assess inhalation exposures from formaldehyde, EPA reviewed workplace inhalation monitoring data

1043

from government agencies such as Occupational Safety and Health Administration (OSHA), inhalation

1044

monitoring data found in peer-reviewed literature, and other inhalation monitoring data submitted to

1045

EPA. Where monitoring data were reasonably available, EPA used these data to characterize central

1046

tendency and high-end peak (15-minute) and 8-hour TWA (i.e., full-shift) inhalation exposures for each

1047

scenario (OES) to workers and ONUs. In some cases, EPA did not identify 15-minute peak exposure

1048

data but identified task-based monitoring data that was used in lieu of 15-minute peak data. The quality

1049

of the monitoring data was evaluated using the data quality review evaluation metrics and the

1050

categorical ranking criteria described in the Draft Systematic Review Protocol Supporting TSCA Risk

1051

Evaluation for Chemical Substances (U.S. EPA, 2021b). Relevant data were assigned an overall quality

1052

determination of high, medium, low, or uninformative. For evidence integration, preference was given to

1053

monitoring data sampled after the latest update of the OSHA permissible exposure limit (PEL) of

1054

formaldehyde in 1992 to 937 µg/m

(0.75 ppm) and short-term exposure limit (STEL) to 2,498 µg/m

1055

(2.0 ppm). This reduces uncertainties with relying on data that may not reflect current regulatory

1056

requirements for TSCA COUs.

1057

1058

For many cases, EPA did not have monitoring data to estimate inhalation exposure for ONUs. In such

1059

cases for full-shift exposures, EPA used the central tendency of worker exposure estimates. However,

1060

EPA did not quantify peak exposures for ONUs. In general, EPA expects ONU exposures to be less than

1061

worker exposures.

1062

1063

For some of the OESs, inhalation monitoring data were not identified. For these cases, EPA utilized

1064

models including using a Monte Carlo simulation and Latin Hypercube sampling method to estimate

1065

inhalation exposures. Where available, the EPA used generic scenarios or emission scenario documents

1066

for relevant exposure points and model input parameters. The Agency then used either monitoring data

1067

or modeling results to develop a high-end and central tendency estimates for short-term exposures and

1068

8-hour TWAs for each OES.

1069

1070

Monitoring data were available to support exposure estimates for all COUs except for three COUs that

1071

relied on modeled estimates:

1072

• Commercial use – chemical substances in automotive and fuel products – automotive care

1073

products; lubricants and greases; fuels and related products;

1074

• Commercial use – chemical substances in agriculture use products – lawn and garden products;

1075

and

1076

• Commercial use – chemical substances in treatment products – water treatment products.

1077

Across TSCA COUs for peak exposure estimates, the central tendency of air concentration estimates

1078

ranged from 86 to 2,002 µg/m

(0.07 to 1.63 ppm) and high-end of air concentration estimates ranged

1079

from 86 to 237,902 µg/m

(0.07 to 193.7 ppm). The TSCA COU of Manufacturing showed

1080

formaldehyde concentrations above other scenarios, with high-end and central tendency of air

1081

concentration results of 237,902 µg/m

and 590 µg/m

, respectively. The underlying scenario was based

1082

on monitoring data from manufacturing sites within the United States, which included tasks where the

1083

workers wore respiratory protection.

1084

1085

Across TSCA COUs for full-shift estimates, the central tendency of air concentration estimates ranged

1086

from 7.5 to 499.3 µg/m

(0.01 to 0.40 ppm) and high-end of air concentration estimates ranged from 7.5

1087

to 17,353.3 µg/m

(0.01 to 13.9 ppm). The TSCA COU of Commercial use – chemical substances in

1088

automotive and fuel products – automotive care products; lubricants and greases; fuels and related

1089

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 46 of 151

products showed formaldehyde concentrations above other scenarios. The underlying scenario was

1090

modeled using a Monte Carlo simulation and assumed that no engineering controls were present. The

1091

first modeling approach resulted in a high-end and central tendency of air concentrations results of

1092

17,353.3 µg/m

and 499.3 µg/m

, respectively and assumes that formaldehyde within the automotive

1093

care product is completely evaporated during duration of application. This results in a very conservative

1094

high-end estimate, well above the current OSHA PEL. EPA also used a second modeling approach using

1095

industry monitoring data on total volatile organic compounds to estimate 1,874 µg/m

and 371 µg/m

1096

1097

EPA uses peak exposure concentration estimates to calculate acute exposure concentrations (AECs),

1098

which is used to estimate acute, non-cancer risks. The full-shift (8- or 12-hour TWA concentrations) are

1099

used to calculate average daily concentrations (ADCs) and lifetime average daily concentrations

1100

(LADCs). The ADC is used to estimate chronic, non-cancer risks and the LADC is used to estimate

1101

chronic, cancer risks. These calculations required additional parameter inputs, such as years of exposure

1102

(31 or 40 year worker tenure), exposure duration and frequency (167 or 250 days), and lifetime years

1103

(78 years). See Appendix F for more information about parameters and equations used to calculate acute

1104

and chronic exposures.

1105

Dermal Exposure Summary

1106

Dermal exposure data were not reasonably available for any of the formaldehyde OESs. Therefore, the

1107

EPA modeled dermal exposure to workers using a modified version of the EPA Dermal Exposure to

1108

Volatile Liquids Model. As the health effect of concern for formaldehyde is the result of exposure at the

1109

point of contact, as opposed to the chemical absorbing into the skin, the absorption factor, body weight,

1110

and surface area were not necessary for the calculation of dermal exposure. The calculation reduces to

1111

an assumed amount of liquid on the skin during one contact event per day adjusted by the weight

1112

fraction of formaldehyde in the liquid to which the worker is exposed.

1113

1114

EPA only evaluated dermal exposures for workers since ONUs are not assumed to directly handle

1115

formaldehyde. EPA did not quantify dermal exposure for two COUs: Distribution in commerce and

1116

Commercial use – chemical substances in packaging, paper, plastic, hobby products – paper products;

1117

plastic and rubber products; toys, playground, and sporting equipment as dermal contact was expected

1118

with solid articles that may contain low residual formaldehyde concentrations.

1119

1120

EPA used the maximum formaldehyde concentrations, which is the highest concentration level of

1121

formaldehyde that a worker handles throughout the process. EPA used concentration data from

1122

published literature and CDR to develop high-end and central tendency dermal exposure estimates.

1123

1124

The dermal exposure estimates ranged from 0.56 to 840 µg/cm

for central tendency exposures, and 0.84

1125

to 3,090 µg/cm

for high-end exposures. The high-end dermal retained dose for four COUs had a value

1126

of 3,090 µg/cm

, which is well above the other dermal exposure estimates:

1127

• Commercial use – chemical substances in automotive and fuel products – automotive care

1128

products; lubricants and greases; fuels and related products and

1129

• Processing – incorporation into an article – paint additives and coating additives not described by

1130

other categories in transportation equipment manufacturing [including aerospace];

1131

• Industrial use – paints and coatings; adhesives and sealants; lubricants; and

1132

• Commercial use – chemical substances in construction, paint, electrical, and metal products –

1133

adhesives and sealants; paint and coatings.

1134

For manual spray applications, EPA expects dermal exposures to be higher. Spray applications are

1135

expected for the use of automotive care products and coatings, paints, adhesives, or sealants. In addition,

1136

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 47 of 151

during the use of automotive care products, workers may use immerse rags in the detailing products,

1137

which could lead to higher dermal loading. For both OESs, EPA assumed an immersive dermal loading

1138

(HE: Q

of 10.3 mg/cm

) on the skin during the exposure scenario. For other OESs, EPA calculated

1139

dermal exposures assuming lower dermal loadings based on expected worker activities (HE: Q

of 2.1

1140

mg/cm

1141

2.2 Consumer Exposure Assessment

1142

To assess consumer exposures, EPA identified 30 exposure scenarios (from 12 formaldehyde TSCA

1143

COUs) that may lead to consumer or bystander exposures. EPA’s Consumer Exposure Model (CEM)

1144

Version 3.0 was used to estimate the 15-minute peak and lifetime average daily concentration for

1145

inhalation exposures to consumer users and bystanders, and the dermal loading during relevant product

1146

and article use. The key conclusions of the consumer exposure assessment are summarized in the CEM

1147

(U.S. EPA, 2024d) and below.

1148

1149

EPA only quantified exposures for plausible exposure pathways, routes, and timespans of exposure and

1150

exposure scenarios for which EPA had at least a medium level of confidence. This means that for some

1151

COUs (i.e., solid products) a dermal loading estimate was not generated since it was not deemed

1152

appropriate (e.g., dermal loading from machinery, mechanical appliances, electrical/electronic articles)

1153

given the best available tools and data. This also means that the total number of COUs assessed for acute

1154

and chronic inhalation scenarios (e.g., 15-minute peak compared to lifetime average daily concentration

1155

estimations) varied according to the relevance of the exposure assessment. However, as presented in

1156

Table 1-1 of the Draft Consumer Exposure Assessment for Formaldehyde (U.S. EPA, 2024d), EPA

1157

quantified exposures for all relevant COUs for at least one route of exposure.

1158

1159

Of note, when potential exposures to the machinery, mechanical appliances, electrical/electronic articles

1160

were assessed, CEM did not yield any expected inhalation exposures via estimates of 15-minute peak

1161

and average daily concentration. Modeled estimates for adhesives and sealants were used as surrogates

1162

for the exposures to electronic products because adhesives and sealants are used in the binding of

1163

internal components and especially at the seams of electronic products. Similarly, EPA does not expect

1164

dermal (skin loading) or oral exposures from reasonably foreseen use of such products, as these

1165

exposures are expected to be negligible.

1166

1167

In addition, EPA did not quantify exposures for COUs in which EPA had a low exposure assessment

1168

confidence. EPA did, however, qualitatively assess the following COUs:

1169

• Water treatment products: No supporting products could be identified other than a fish tank

1170

cleaning solution and because formaldehyde is highly reactive in water; therefore, these

1171

exposures are expected to be negligible.

1172

• Laundry and dish washing products: Formaldehyde is highly reactive in water. EPA believes

1173

these preliminary CEM modeling results are implausible.

1174

• Lawn and garden products: The non-pesticidal exposure scenario for this TSCA COU is unclear

1175

because when mixed in water, formaldehyde is highly reactive. In addition, EPA’s CEM

1176

assumes no inhalation exposure from such products. This is likely due to the default assumption

1177

that such activities typically occur outdoors where the chemical would be diluted in the ambient

1178

air during and after use.

1179

• Foam insulation: Formaldehyde exposures from foam insulation products were not quantified as

1180

consumer exposures to these products are expected to be minimal. During the public comment

1181

period for the draft high priority designation of formaldehyde, the North American Insulation

1182

Manufacturers Association stated “for those insulation products in which formaldehyde is a

1183

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 48 of 151

component of the binder, the products are cured at high temperatures during the manufacturing

1184

process after the binder has been applied, virtually eliminating the free formaldehyde content.

1185

Any free formaldehyde released from the binder during cure is destroyed either during the cure

1186

process or by emissions control equipment required by the Maximum Achievable Control

1187

Technology (MACT) standard. Therefore, formaldehyde off-gassing from the majority of

1188

finished products is highly unlikely” (Docket ID EPA-HQ-OPPT-2019-0131). Given this

1189

information, EPA expects formaldehyde exposures to foam insulation to be negligible.

1190

Given that each TSCA COU may comprise multiple exposure scenarios and multiple scenarios may be

1191

applicable to multiple COUs, representative scenarios were identified for each TSCA COU per relevant

1192

exposure assessment. Representative scenarios were identified according to the highest estimated

1193

exposure estimate per assessment. Refer to Appendix B of the Draft Consumer Exposure Assessment for

1194

Formaldehyde (U.S. EPA, 2024d) for a list of representative consumer exposure scenarios according to

1195

TSCA COUs.

1196

1197

CEM uses a two-zone representation of the building of use when predicting indoor air concentrations.

1198

Zone 1 represents the room where the consumer product is used; Zone 2 represents the remainder of the

1199

building. Each zone is considered well-mixed. CEM allows further division of Zone 1 into a near-field

1200

and far-field to accommodate situations where a higher concentration of product is expected very near

1201

the product user when the product is used. Zone 1-near-field represents the breathing zone of the user at

1202

the location of the product use while Zone 1-far-field represents the remainder of the Zone 1 room.

1203

1204

Inhalation exposure is estimated in CEM based on zones and pre-defined activity patterns. The

1205

simulation run by CEM places the product user within Zone 1 for the duration of product use while the

1206

bystander is placed in Zone 2 for the duration of product use. Following the duration of product use, the

1207

user and bystander follow one of three predefined activity patterns established within CEM, based on

1208

modeler selection. The selected activity pattern takes the user and bystander in and out of Zone 1 and

1209

Zone 2 for the period of the simulation. The user and bystander inhale airborne concentrations within

1210

those zones, which will vary over time, resulting in the overall estimated exposure to the user and

1211

bystander.

1212

1213

Modeled formaldehyde concentrations depend upon the room of use, amount of the chemical in the

1214

product and consumer use patterns (e.g., amounts used). Consumer users of products and articles

1215

generally had higher peak and long-term inhalation exposures, in comparison with bystanders. Across

1216

all relevant age groups and exposure scenarios, the highest estimated 15-minute peak TWA

1217

formaldehyde air exposure was for consumer users of floor coverings; foam seating and bedding

1218

products; cleaning and furniture care products; furniture & furnishings including stone, plaster, cement,

1219

glass and ceramic articles; metal articles; or rubber articles, while the lowest 15-minute peak exposure

1220

was for individuals using textiles or clothing that emit formaldehyde (Figure 2-1). Consumer users of

1221

adhesives and sealants; paint and coatings were estimated to have the highest estimated average daily air

1222

exposure to formaldehyde (Figure 2-2), while consumer users of automotive care products had the

1223

lowest average daily exposure.

1224

1225

The highest acute dermal loading for consumer users resulted from use of automotive care products. The

1226

lowest acute dermal loading resulted from use of arts, crafts, and hobby materials (Figure 2-3). For the

1227

dermal assessment, the estimated dermal loading was based on weight fraction identified in the literature

1228

and safety data sheets (SDSs).

1229

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 49 of 151

1230

Figure 2-1. Summary of 15-Minute Peak Consumer Inhalation Concentrations (Based on CEM)

1231

For some products, air concentrations were modeled for near-field and far-field (generally describing differences in exposure within the same room), while

1232

for other products, concentrations were modeled for zones 1 and 2 (generally describing different rooms). Risks from near-field and zone 1 exposures

1233

generally represent risks from direct exposures to consumer users while far-field and zone 2 tend to represent risks to consumer bystanders. The x-axis

1234

presents the 15-minute peak inhalation non-cancer concentration and the y-axis presents the modeled TSCA COU.

1235

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 50 of 151

1236

Figure 2-2. Summary of Average Daily Consumer Inhalation Concentrations, per Year (Based on CEM)

1237

The x-axis presents the chronic inhalation average daily concentration, and the y-axis presents the modeled exposure TSCA COU.

1238

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 51 of 151

1239

Figure 2-3. Summary of Acute Consumer Dermal Concentrations (Based on Thin Film Model)

1240

The x-axis presents dermal loading concentration, and the y-axis presents the modeled TSCA COUs. The term

1241

“High” in the figure refers to high-end scenarios as described above.

1242

2.3 Indoor Air Exposure Assessment

1243

A detailed analysis for indoor air can be found in the Draft Indoor Air Exposure Assessment for

1244

Formaldehyde (U.S. EPA, 2024j). The separation of the consumer exposure assessment and the indoor

1245

air exposure assessment is intentional; each assessment represents a different context of exposures.

1246

Generally, exposures to most consumer products occur over a relatively short period of time (minutes to

1247

hours per day) and the duration of exposure from those uses within a residence are expected to be short

1248

relative to continuous sources of exposure such as flooring or furniture. Thus, the indoor air exposure

1249

assessment represents exposures mainly resulting from the presence of articles or materials within a

1250

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 52 of 151

residential household which typically off-gas formaldehyde over an extended period (particularly the

1251

first several years after an article or material is manufactured). The indoor air exposure assessment also

1252

incorporates aspects of ongoing exposures to populations in office or commercial settings and therefore

1253

is more expansive and inclusive than the consumer exposure assessment.

1254

1255

Formaldehyde is a chemical ingredient in many products, which release formaldehyde into the indoor

1256

air. Indeed, indoor air studies of formaldehyde (IPCS, 2002; ATSDR, 1999) demonstrate that the indoor

1257

environment, including homes and automobiles, can be a major source of formaldehyde exposure. This

1258

is because formaldehyde is used ubiquitously for the manufacturing of various consumer products (e.g.,

1259

wallpaper, hardwood floors, seat covers used in numerous articles) and because formaldehyde is formed

1260

as a combustion byproduct from sources such as fireplaces, ovens, stoves, and tobacco smoke.

1261

1262

Given the number of TSCA and other sources contributing to formaldehyde in indoor air, indoor air

1263

concentrations reported in monitoring studies are generally considered a reflection of aggregate

1264

exposures. Any reported average indoor air monitoring for formaldehyde in American homes is

1265

expected to be a result of off-gassing from articles or materials, or long-term emissions (e.g., from

1266

fireplaces or stoves), from multiple TSCA COUs and other sources. While intermittent product or article

1267

use may briefly contribute to indoor air formaldehyde concentrations, generally EPA assumes that most

1268

formaldehyde indoor air exposures occur over an extended period spanning several months to multiple

1269

years (U.S. EPA, 2016b).

1270

1271

In the Draft Indoor Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024j), EPA considered

1272

available monitoring data from commercial, residential, and automobile environments (Section 2.3.10).

1273

EPA also used CEM to model chronic indoor air exposure resulting from TSCA COUs that are expected

1274

to be the largest contributors of formaldehyde to indoor air primarily due to off-gassing (Section 0).

1275

EPA incorporated TSCA COU-specific emission rates extracted from the literature, when available, into

1276

its modeling to better approximate real-world conditions. Residential indoor air modeled and measured

1277

concentrations of formaldehyde were generally within the same order of magnitude.

1278

Indoor Air Exposure Monitoring Results

1279

EPA identified over 800 monitoring studies, 290 of which are specific to the indoor air environment and

1280

associated with the 12 TSCA COUs subject to this risk evaluation (see Appendix A of the Draft Indoor

1281

Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024j)). As was presented in Section 3.2.2 of

1282

the 2016 Formaldehyde Exposure Assessment Report TSCA Title VI Final Rule (U.S. EPA, 2016b), EPA

1283

presents a supplemental summary of formaldehyde concentrations identified from several well-

1284

established residential (Table 2-1, Figure 2-4) and commercial (Table 2-2) indoor air monitoring studies

1285

to provide additional context to the TSCA formaldehyde indoor air exposure assessment. From a

1286

comparison of residential (Table 2-1) and commercial (Table 2-2) indoor air monitoring, residential

1287

indoor air exposures to formaldehyde are generally expected to be higher compared to commercial

1288

buildings due to expected lower room volumes and air exchange rates in residences relative to

1289

commercial buildings.

1290

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 53 of 151

Table 2-1. Indoor Air Monitoring Concentrations for Formaldehyde

1291

Reference

Monitoring Study Description

Formaldehyde Concentrations (µg/m

)

Central Value

Range/Percentiles

American Healthy

Home Survey

(QuanTech, 2021)

Nationally representative sample of

688 U.S. homes of various ages,

types, conditions, and climates

Mean: 23.2

Range (lower/upper 95% tiles

of mean): 21.4–25.0

(CARB, 2004)

Portable and traditional classrooms in

67 California schools (Phase II study)

Arithmetic mean:

18.42 (portable)

14.74 (traditional)

95th Percentile:

31.93 (portable)

27.02 (traditional)

(Gilbert et al.,

2005)

59 homes in Prince Edward Island,

Canada

Geometric mean:

33.16

Range:

5.53–87.33

(Gilbert et al.,

2006)

96 homes in Quebec City, Canada

Geometric mean:

29.48

Range: 9.58–89.91

(Hodgson et al.,

2004)

4 new relocatable classrooms

Unspecified mean:

9.83 (indoor-

outdoor)

Range:

4.91–14.74 (indoor-outdoor)

(Hodgson et al.,

2000)

New homes in eastern/SE U.S.:

4 new manufactured homes

7 new site-built homes

Geometric mean:

41.76

44.22

Range:

25.79–57.73

17.2–71.24

(Liu et al., 2006)

234 homes in Los Angeles County,

CA; Elizabeth, NJ; and Houston, TX

Median:

20.02

Range:

12.53–32.43

(5th–95th percentiles)

(LBNL, 2008)

4 FEMA camper trailers

Unspecified mean:

568.67

Range:

330.39–924.85

(Murphy et al.,

2013)

Sample:

All structures (519)

Travel trailers (360)

Park models (90)

Mobile homes (69)

Geometric mean:

94.57

99.49

54.04

70.01

Range:

3.68–724.65

3.68–196.52

13.51–393.03

(Offermann et al.,

2008)

108 new SF homes in CA

Median:

38.2

Range:

4.67–143.33

(Sax et al., 2004)

Inner-city homes:

NY City (46) – winter (W), summer

(S)

Los Angeles (41) – winter (W), fall

(F)

Median:

12.28 (W), 18.42 (S)

18.42 (W), 14.74 (F)

Range:

4.91–22.11 (W), 6.14–50.36

(S)

7.37–55.27 (W), 7.37–31.93

(F)

1292

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 54 of 151

Table 2-2. Formaldehyde Monitored in U.S. Commercial Buildings from 2000 to Present

1293

Reference

Monitoring Study Description

Formaldehyde

Concentrations

(µg/m

)

Descriptor

(Ceballos and

Burr, 2012)

Office space indoor air monitoring for

formaldehyde in a commercial building

24.56

Average

(U.S. EPA,

2023k)

Indoor air monitoring across 100

randomly selected U.S. commercial

buildings

3.68

5th percentile

14.74

50th percentile

30.71

95th percentile

(Page and Couch,

2014)

Indoor air U.S. government offices

<61.41

Maximum

(Lukcso et al.,

2014)

12.28

Geometric mean

56.50

Maximum

(Dodson et al.,

2007)

Classrooms in school buildings in the

United States

17.69

Median

1294

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 55 of 151

1295

1296

Figure 2-4. Long-Term Average Daily Concentrations of Formaldehyde According to Air Monitoring Data Source

1297

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 56 of 151

Monitoring data from the American Healthy Homes Survey II suggests that concentrations of

1298

formaldehyde may range from 0.27 to 124.2 µg/m

for all homes (including new homes at the time of

1299

survey), with 95 percent of homes having concentrations below 47 µg/m

(QuanTech, 2021). Those data

1300

include formaldehyde produced from both TSCA sources (Section 3.1.1 of the Draft Indoor Air

1301

Exposure Assessment for Formaldehyde (U.S. EPA, 2024j) and other sources of formaldehyde such as

1302

tobacco smoke or the use of fireplaces, gas-burning appliances, candles, and air purifiers (QuanTech,

1303

2021). These other sources do not contain formaldehyde but rather lead to the formation of

1304

formaldehyde during use.

1305

1306

For other sources of formaldehyde in indoor air, simulated 50th percentile room concentrations ranged

1307

from 12.3 to 44.2 μg/m

individually for candles, incense, cooking, wood combustion, and air cleaning

1308

devices, and up to 152.2 μg/m

for ethanol fireplaces (ECHA, 2019). Air cleaning devices such as

1309

photocatalytic air purifiers can produce formaldehyde from irradiation of air contaminants, leading to

1310

increased indoor air concentrations of formaldehyde (Salthammer, 2019). Formaldehyde production

1311

associated with cooking depends on many factors, including cooking temperature and type of oil and

1312

variety of food being cooked. Select gas-oven cooking tests involving a variety of cooking parameters

1313

resulted in formaldehyde concentrations ranging from 36.5 to 417.3 μg/m

(Salthammer, 2019). Tobacco

1314

smoke is also known to be a contributor to formaldehyde concentrations within all indoor air

1315

environments (U.S. EPA, 2016b; Girman et al., 1982), although according to the World Health

1316

Organization, tobacco smoke primarily increases formaldehyde concentrations in indoor air

1317

environments where the rates of smoking are high with minimal ventilation (IPCS, 2002).

1318

Indoor Air Exposure Modeling Results

1319

EPA used CEM to model indoor air concentrations in American homes and vehicles based on TSCA

1320

COU-specific emission rates, providing an estimate of TSCA COU-specific contributions to

1321

formaldehyde in indoor air. Central tendency estimates were generated as discussed in Section 2.1.1.1.3

1322

of the Indoor Air Exposure Module (U.S. EPA, 2024j) for comparability with AHHS II monitoring data

1323

and to estimate common indoor air concentrations for most American households. For the TSCA COUs

1324

identified in Section 1.1 of the Indoor Air Exposure Module (U.S. EPA, 2024j), EPA estimated chronic

1325

average daily indoor air exposures. Through a review of key products known to be significant and

1326

persistent emitters of formaldehyde, EPA identified four TSCA COUs as potentially significant

1327

contributors to residential indoor air environment.

1328

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 57 of 151

1329

1330

Figure 2-5. Modeled Formaldehyde Average Daily Inhalation Concentrations in Indoor Air (According to CEM)

1331

The x-axis presents the average daily concentration, and the y-axis presents the modeled TSCA COUs.

1332

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 58 of 151

EPA generated estimated indoor air exposures using the CEM for four TSCA COUs (see Section 2.1.1

1333

of the Indoor Air Exposure Module (U.S. EPA, 2024j)). The Agency used emission rates and fluxes

1334

identified from the literature and compared the estimated indoor air concentrations in homes and

1335

vehicles with air monitoring concentrations from the literature (Table 2-3 of the Indoor Air Exposure

1336

Module (U.S. EPA, 2024j)). Modeled concentrations of formaldehyde are within the same order of

1337

magnitude as reported in monitoring studies, including the American Healthy Homes Survey II (see

1338

Section 3.2 of the Indoor Air Exposure Module (U.S. EPA, 2024j)).

1339

1340

The estimated formaldehyde indoor air exposures likely represent exposures from new articles added to

1341

a resident (e.g., wood products). Given each COU may comprise multiple exposure scenarios and

1342

multiple scenarios may be applicable to multiple COUs, representative exposure scenarios were

1343

identified according to the highest estimated exposure estimate per scenario in a room of use, for each

1344

COU (Table 2-3).

1345

1346

Table 2-3. Representative Residential Indoor Air Exposure Scenarios According to COUs

1347

Conditions of Use

CEM Exposure Scenarios

Construction and building materials covering large surface areas,

including wood articles; Construction and building materials covering

large surface areas, including paper articles; metal articles; stone, plaster,

cement, glass and ceramic articles

Building/Construction

Materials – Wood Articles:

Hardwood Floors (residential)

Fabric, textile, and leather products not covered elsewhere

Seat Covers (automobile)

Furniture Seat Covers

(residential)

Fabrics: Clothing (residential)

Floor coverings; Foam seating and bedding products; Cleaning and

furniture care products; Furniture & furnishings including stone, plaster,

cement, glass and ceramic articles; metal articles; or rubber articles

Furniture & Furnishings –

Wood Articles: Furniture

(residential)

Paper products; Plastic and rubber products; Toys, playground, and

sporting equipment

Paper-Based Wallpaper

(residential)

Representative exposure scenarios, as noted in Section 2.1.1, are bolded as these scenarios had the highest

estimated concentrations per COU.

Within this COU, the Clothing (residential) scenario is identified as the representative scenario despite a lower

estimated concentration compared to Seat covers (automobile), since residential indoor air environments are of

primary interest in this indoor air assessment.

1348

Over the span of a year, the highest TSCA COU contributor to the residential indoor air environment

1349

was building wood products. Additionally, while several of the modeled COUs may occur

1350

simultaneously, aggregating exposures for all four TSCA COUs may not be reflective of actual exposure

1351

scenarios encountered over a lifetime as the combination of these TSCA COU likely differ from home to

1352

home and overtime. Additionally, while several of the modeled COUs may occur simultaneously,

1353

aggregating exposures for all four TSCA COUs may not be reflective of actual exposure scenarios

1354

encountered over a lifetime because the combination of these TSCA COUs likely differ both from home

1355

to home and over time.

1356

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 59 of 151

2.3.2.1 Aggregate Indoor Air Exposure

1357

EPA defines aggregate exposure as “the combined exposures to an individual from a single chemical

1358

substance across multiple routes and across multiple pathways (40 CFR § 702.33).” Theoretically, the

1359

reported formaldehyde concentrations from the monitoring data may represent aggregate formaldehyde

1360

indoor air concentrations in vehicles per the Lawryk et al. study (Lawryk and Weisel, 1996; Lawryk et

1361

al., 1995) and across U.S. households per the AHHS II study (QuanTech, 2021), assuming at least a 3-

1362

hour TWA; or the typical indoor air concentration of formaldehyde in these environments.

1363

1364

EPA considered aggregating modeled air concentrations for plausible combinations of COUs expected

1365

to co-occur in specific indoor air environments (e.g., combinations of products likely to be present in

1366

mobile homes, new homes or automobiles), but concluded that, due to variability among homes and over

1367

time within a given home, uncertainties were too great to support a quantitative aggregate analysis

1368

across multiple COUs.

1369

2.4 Ambient Air Exposure Assessment

1370

The ambient air exposure assessment for formaldehyde quantitatively evaluates exposures resulting

1371

from industrial releases of formaldehyde to ambient air that are associated with TSCA COUs. This

1372

assessment focuses on a subset of the general population who reside near releasing facilities by utilizing

1373

both modeling approaches and ambient monitoring data to assess and characterize ambient air

1374

concentrations and exposures to formaldehyde. A detailed summary of all the analyses conducted,

1375

methodologies used, and all exposure concentration results for formaldehyde are provided in the Draft

1376

Ambient Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024a) and associated supplemental

1377

files.

1378

Monitoring for Ambient Air Concentrations

1379

EPA identified and summarized monitoring data for formaldehyde from EPA’s Ambient Monitoring

1380

Technology Information Center (AMTIC) (U.S. EPA, 2022a). The Agency also identified and

1381

summarized outside monitoring data during EPA’s systematic review process (U.S. EPA, 2023a). These

1382

results are presented in the Draft Ambient Air Exposure Assessment for Formaldehyde (U.S. EPA,

1383

2024a).

1384

1385

This assessment summarizes monitoring data from EPA’s AMTIC (U.S. EPA, 2022a)to understand

1386

aggregate or total formaldehyde concentrations in ambient air. The AMTIC data are also used to

1387

characterize modeled concentrations of formaldehyde with recognition of the differences between these

1388

information sources. That is, modeled environmental concentrations only include releases that can be

1389

associated with TSCA COUs while monitoring data does not differentiate between concentrations

1390

associated with TSCA COUs and concentrations from all other sources. These differences can limit

1391

direct comparison, although EPA conducted some analyses to inform specific local impacts where both

1392

modeled and monitored ambient air concentrations are available based on locations of monitoring sites

1393

and industrial facilities releasing formaldehyde to the ambient air.

1394

1395

The AMTIC dataset for formaldehyde includes 195 monitoring sites from 36 different states. Data were

1396

extracted across 6 years (2015 through 2020) and include a total of 306,529 observations. EPA

1397

calculated summary statistics for all samples, samples by state, samples by census tract, samples by

1398

monitoring site, samples by monitoring site and year, and samples by monitoring site and year and

1399

quarter. For purposes of this ambient air exposure assessment, EPA used the overall statistics across all

1400

samples to characterize exposures and characterize exposures to the general population (Table 2-4).

1401

Monitoring locations and annual summary statistics are provided in the ambient air exposure module

1402

(U.S. EPA, 2024a).

1403

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 60 of 151

1404

The last 5 years of available AMTIC data were selected for use in the formaldehyde assessment. (2015

1405

to 2020). This dataset includes a total of 233,961 entries for formaldehyde within the five-year duration

1406

from 20 air monitoring programs covering 32 states within the contiguous United States. Any entries

1407

with missing key data were omitted from the analysis (e.g., concentrations, concentration units, method

1408

detection limits, methodology used). All concentration and method detection limit (MDL) values were

1409

converted to micrograms per cubic meter (µg/m

) for unit uniformity between submitting programs.

1410

Method detection limits were provided along with sample concentrations on a submission-by-

1411

submission basis by submitting agencies, from 0.000011 to 1.2 µg/m

, and varied by sample based on

1412

the sampling and analysis methodology. Entries with reported concentrations below the method

1413

detection limit were substituted with a value of 0 µg/m

. Concentrations of formaldehyde ranged from

1414

below the method detection limit to 60.1 µg/m

and a median value of 1.6 µg/m

. A summary of the

1415

statistics extracted from the overall dataset are provided in Table 2-4.

1416

1417

Table 2-4. Overall Monitored Concentrations of Formaldehyde from AMTIC Dataset

1418

Monitored Concentrations (µg/m

)

Aggregation

Count

Minimum

(non-zero)

Median

Mean

Maximum

All Entries

233,961

0.002

1.6

2.1 ± 2.2

Daily Mean

3,843

0.011

2.5

3.0 ± 2.0

18.4

Annual Mean

1.4

2.9

3.0 ± 1.1

6.5

1419

The individual site data collected by AMTIC represents various sampling techniques sample collection

1420

duration ranging from 5 minutes to 24 hours. When using these data for comparison to the presented

1421

formaldehyde models, the concentrations were converted to daily and annual averages. AMTIC

1422

concentration values were used to calculate daily or annual average only when there was greater than 75

1423

percent sample coverage over the averaged timeframe when converting from sub-hour samples to hourly

1424

averages and again for hourly samples to daily averages. Each annual quarter required a minimum of

1425

seven valid daily averages and each annual mean required a minimum of three valid quarterly averages

1426

per year per site. The high standards for coverage resulted in a drastic reduction in the data available for

1427

conversion to daily and annual averages. Of the original 233,961 complete entries, there were 64 site-

1428

years and 3,843 site-days with sufficient coverage to calculate daily and annual average statistics (Table

1429

2-4). EPA is investigating additional methods under OAR guidance to better estimate daily and annual

1430

average statistics to increase the number of available sites and data available for use in model

1431

comparison.

1432

Modeling Ambient Air Concentrations

1433

2.4.2.1 Integrated Indoor/Outdoor Air Calculator Model (IIOAC)

1434

EPA used the Integrated Indoor-Outdoor Air Calculator (IIOAC) Model to estimate daily- and annual-

1435

averaged formaldehyde concentrations for a suite of exposure scenarios at three predefined distances

1436

from a facility releasing formaldehyde to the ambient air. EPA’s modeling evaluated industrial releases

1437

of formaldehyde that are associated with COUs from two separate databases (TRI and NEI). EPA

1438

compared releases and modeled concentrations from the two databases and found results were within the

1439

same estimated distribution range. Therefore, to provide a clearer picture of findings, the Agency only

1440

presents results from the TRI dataset in this draft human health risk assessment. Nonetheless, results

1441

from all exposure scenarios and datasets evaluated are provided in the “Draft Ambient Air Exposure

1442

Assessment Results and Risk Calcs Supplement A.”

1443

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 61 of 151

EPA utilized the 95th percentile release value reported to TRI by Industry Sector (mapped to respective

1444

TSCA COUs) and the 95th percentile modeled daily-averaged and annual-averaged air concentrations

1445

from the IIOAC output file at a distance of 100 to 1,000 m from the release facility to characterize

1446

exposures and derive risk estimates (see Section 4.2.4.2). Additionally, the exposure scenario used for

1447

this Draft Human Health Risk Assessment assumes an industrial facility releasing formaldehyde to the

1448

ambient air operates 24 hours/day, 7 days/week, 365 days/year, which is likely a conservative

1449

assumption.

1450

1451

The 95th percentile release scenario and modeled concentrations were used to represent a more national

1452

level exposure estimate based on actual reported releases. The operating scenario was selected because it

1453

is representative of typical operating conditions under which industrial facilities involved with

1454

formaldehyde manufacturing, processing, etc. operate. Although this scenario is representative of a high-

1455

end exposure scenario that is inclusive of more sensitive and locally impacted populations, it is not a

1456

maximum worst-case exposure scenario and thus considered more representative of an overall

1457

community or nationally representative exposure scenario.

1458

1459

Because of the exposure scenario used (365 days per year, 24 hrs/day, 7 days per week), the daily-

1460

averaged modeled concentration and annual-averaged modeled concentration output values from the

1461

IIOAC Model are the same. Results from this exposure scenario are summarily presented independently

1462

in the “Draft Ambient Air Exposure Assessment Results and Risk Calcs Supplement B.” The reason for

1463

the same modeled concentrations is a math exercise based on the way annual-averaged concentrations

1464

are calculated as an arithmetic average of all daily-averaged concentrations. If the daily-averaged

1465

concentrations are based on 365 days of exposure, then the annual average will be the average of the

1466

same values and result in the same modeled concentration. However, EPA also ran 250 days of exposure

1467

(although not presented here, modeled concentrations are included in the supplemental files), and for

1468

this 250-day exposure scenario, the daily-averaged and annual-averaged concentrations are different.

1469

The reason for that is the annual-averaged concentrations will also include zero concentration days, and

1470

therefore result in a different arithmetic average of the daily modeled concentrations.

1471

1472

Results for acute and chronic exposures across all industry sectors and associated COUs ranged from

1473

0.0001 to 5.7 µg/m

for the exposure scenario described above. Results are presented for each TSCA

1474

COU in Figure 2-6. These results represent the highest exposure concentration across all industry sectors

1475

associated with the respective formaldehyde TSCA COU. The presented results also represent both the

1476

acute and chronic exposure concentrations, which are the same, as described above. Additional details

1477

on these results, including the industry sectors with the highest estimated exposure concentrations and

1478

associated TSCA COUs are provided in the Draft Ambient Air Exposure Assessment for Formaldehyde

1479

(U.S. EPA, 2024a).

1480

1481

1482

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 62 of 151

1483

1484

Figure 2-6. Exposure Concentrations by TSCA COU for the 95th Percentile Release Scenario and 95th Percentile Modeled

1485

Concentration between 100 and 1,000 m from Industrial Facilities Releasing Formaldehyde to the Ambient Air

1486

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 63 of 151

2.4.2.2 AirToxScreen

1487

EPA used 2019 AirToxScreen to understand the relative contributions of other sources to overall

1488

formaldehyde concentrations in the ambient air. AirToxScreen is an EPA screening tool used to evaluate

1489

air toxics from all known sources across the United States and estimates air concentration and associated

1490

health risk at the census tract level nationwide using a combination of models and data sources (Scheffe

1491

et al., 2016). For formaldehyde specifically, AirToxScreen integrates atmospheric chemistry for

1492

predicting the production and decay over larger extents using the Community Multiscale Air Quality

1493

(CMAQ) model (Luecken et al., 2019). The 2019 AirToxScreen data are shown in Figure 2-7. The

1494

figure shows the range of concentrations across all sources of formaldehyde, as well as contributions

1495

from biogenic sources, secondary sources, and point sources.

1496

1497

Secondary production of formaldehyde is the largest contributor of formaldehyde to ambient air with

1498

modeled concentrations ranging from 0.085 to 1.8 µg/m

(mean ± 1SD: 0.86 ± 0.25 µg/m

) according to

1499

the AirToxScreen data. Secondary production is the atmospheric formation of formaldehyde from

1500

natural and manmade compounds. This can include the degradation of isoprene (a compound naturally

1501

produced by animals and plants) to formaldehyde and other complex air chemistry. AirToxScreen is not

1502

able to apportion the relative contributions from different secondary sources (source apportion).

1503

1504

Biogenic sources also have a higher contribution to total concentration with a range of 0.0014 to 0.67

1505

µg/m

(mean ± 1SD: 0.13 ± 0.072 µg/m

) based on the AirToxScreen data. Biogenic sources include

1506

those emissions from trees, plants, and soil microbes.

1507

1508

It is noteworthy that the AirToxScreen data cannot be attributed to COUs but do show relative

1509

distributions of various sources. The point source estimates; however, are expected to include

1510

contributions from COUs. Point sources contributions to total formaldehyde concentrations range from

1511

0.0 to 0.88 µg/m

(mean ± 1SD: 0.0070 ± 0.014 µg/m

). However, as described above, the

1512

AirToxScreen data are averaged across census tracts, which can result in a considerable underestimation

1513

of exposures relative to a source-specific contribution to which populations living nearby releasing

1514

facilities are exposed and thus not comparable to the modeled concentrations from IIOAC.

1515

1516

Figure 2-7 does not include AirToxScreen data for on-road sources, near-road sources, off-road sources,

1517

wildfire sources, etc. However, these sources would be captured in the results shown for all sources.

1518

1519

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 64 of 151

1520

Figure 2-7. Distributions of 2019 AirToxScreen Modeled Data for All Sources, Secondary

1521

Production Sources, Point Sources, and Biogenic Sources for the Contiguous United States

1522

2.4.2.3 Human Exposure Model (HEM)

1523

EPA used the Human Exposure Model (HEM 4.2)

to estimate formaldehyde concentrations on a site-

1524

specific basis at multiple distances from releasing facilities. HEM 4.2 has two components: (1) an

1525

atmospheric dispersion model, AERMOD, with included regional meteorological data; and (2) U.S.

1526

Census Bureau population data at the Census block level. The current HEM version utilizes 2020

1527

Census data—including all 50 states, the District of Columbia, Puerto Rico, and the U.S. Virgin Islands.

1528

AERMOD estimates the magnitude and distribution of chemicals concentrations in ambient air in the

1529

vicinity of each releasing facility within a user-defined radial distances out to 50 km (about 30 miles).

1530

HEM also provides chemical concentrations in ambient air at the centroid of over 8 million census

1531

blocks across the United States. This higher tier model was selected to expand on the IIOAC results by

1532

providing more granularity in modeling individual facilities and more discrete distances, geospatial data

1533

associated with modeling results for mapping and further analysis, and population data associated with

1534

modeled results.

1535

1536

Ambient air concentrations at the census block level were modeled by HEM and are shown in Figure

1537

2-8. These aggregated concentrations are the summed stack and fugitive modeled concentrations, which

1538

can include the summation of multiple adjacent facilities, at specific locations. The site-specific

1539

concentration results represent the expected annual average ambient air concentration attributable from

1540

all modeled TRI releases of TSCA COUs, in some census blocks accounting for concentrations from

1541

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 65 of 151

multiple releasing facilities. Concentrations ranged from 0 to 8.9 µg/m³. Census blocks with modeled

1542

total concentrations below the 95th percentile biogenic formaldehyde threshold of 0.28 µg/m³ are

1543

presented in grey. Turquoise dots show census blocks with concentrations ranging from 1 to 5 times the

1544

biogenic threshold, purple dots show concentrations from 5 to 10 times the biogenic threshold, and pink

1545

dots show values greater than 10 times the biogenic threshold. Across the country, a total population of

1546

105,463 people (based on 2020 Census data) live in census blocks shown with ambient air.

1547

1548

Elevated ambient air concentrations of formaldehyde from industrial releases appear most densely

1549

concentrated in the southeastern United States. Census blocks with elevated concentrations are found

1550

throughout the country, with some regions showing fewer overall TRI facilities, and fewer releases

1551

resulting in elevated air concentrations.

1552

1553

Patterns in the relative contribution of stack and fugitive releases, and the distribution of results at

1554

varying radial distances from the releasing facility were examined (Figure 2-9). Each vertical bar and

1555

median line indicate the shape of the distribution of concentrations by release type for individual

1556

facilities. These results indicate that concentrations resulting from fugitive emissions are greater than

1557

those from stack emissions closer to the releasing facility, but concentrations from stack emissions tend

1558

to become greater at further distances. As many facilities report only a single release type (either

1559

fugitive or stack), the total concentration distributions represent a greater number of facilities than the

1560

corresponding fugitive and stack distributions and tend to fall somewhere between the fugitive and stack

1561

values. Total modeled concentrations tend to reach their maximum within 1,000 m of a facility. Values

1562

represented in this analysis are directly modeled at the 16 radial points around each distance ring, rather

1563

than census block centroids, and can therefore be located much closer to the releasing facility and

1564

represent much higher concentrations. These points are not associated with population estimates, and in

1565

some cases the modeled distances may still be within a facility property boundary.

1566

1567

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 66 of 151

1568

Figure 2-8. Map of Contiguous United States with HEM Model Results for TRI Releases Aggregated and Summarized by Census

1569

Block

1570

Census blocks with modeled total concentrations below the 95th percentile biogenic formaldehyde threshold of 0.28 µg/m³ are presented in

1571

grey. Turquoise dots show census blocks with concentrations ranging from 1 to 5 times the biogenic threshold, purple dots show

1572

concentrations from 5 to 10 times the biogenic threshold, and pink dots show values greater than 10 times the biogenic threshold.

1573

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 67 of 151

1574

Figure 2-9. Median and Maximum Concentrations (Fugitive, Stack, and Total Emissions) across

1575

the 11 Discrete Distance Rings Modeled in HEM

1576

Integrating Various Sources of Formaldehyde Data

1577

Monitoring data from AMTIC, modeled exposures calculated from IIOAC, and modeled data from

1578

AirToxScreen were compiled to understand how exposures from COUs fit into the broader context of

1579

available information on formaldehyde. Figure 2-10 shows the distributions of data from these datasets.

1580

As shown these distributions overlap. At the national scale, populations are exposed to many different

1581

sources of formaldehyde (COUs, secondary, biogenic, etc.). Modeled exposure estimates downwind

1582

from TSCA COU releases are variable across COUs and locations. In some locations the concentrations

1583

from COUs dominate total concentrations of formaldehyde in ambient air. In most of the country

1584

however, ambient air concentrations are dominated by other sources (secondary, biogenic, etc.)

1585

according to AirToxScreen. All populations are exposed to concentrations between the various sources

1586

of formaldehyde.

1587

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 68 of 151

1588

Figure 2-10. Distributions of AMTIC Monitoring Data, IIOAC Modeled Data, and AirToxScreen

1589

Modeled Data

1590

1591

EPA recognizes that the different model estimates are not directly comparable. For example, the IIOAC

1592

results represent a 95th percentile annual average concentration between 100 to 1,000 m from the release

1593

point. In contrast, AirToxScreen concentrations represent annual average concentrations at the census

1594

tract scale. Given the spatial scale difference it is expected that AirToxScreen results could

1595

underestimate concentrations on a smaller scale (i.e., near facilities) or have lower concentration

1596

estimates than IIOAC and this difference can be seen in Figure 2-10. Additionally, only point source

1597

data within AirToxScreen may represent a broader set of formaldehyde releases that include releases

1598

associated with TSCA COUs.

1599

1600

Furthermore, the AMTIC data represent a range of samples collected at various locations (independent

1601

of TSCA releases of formaldehyde) and collection durations are much shorter than a year (5 minutes to

1602

24 hours). Despite these uncertainties, these data suggest that formaldehyde concentrations from TSCA

1603

sources are higher than formaldehyde concentrations that are expected to occur due to natural formation.

1604

These higher concentrations will be driven by the location of release. These COUs are listed in Section

1605

2.4.2.1 and this conclusion is further supported by the HEM analysis.

1606

2.5 Weight of Scientific Evidence and Overall Confidence in Exposure

1607

Assessment

1608

As described in the 2021 Draft Systematic Review Protocol (U.S. EPA, 2021b), the weight of scientific

1609

evidence supporting exposure assessments is evaluated based on the availability and strength of

1610

exposure scenarios and exposure factors, measured and monitored data, estimation methodology and

1611

model input data, and, if appropriate, comparisons of estimated and measured exposures. The strength of

1612

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 69 of 151

each of these evidence streams can be ranked as either robust, moderate, slight, or indeterminate. For

1613

each component of this exposure assessment, EPA evaluated the weight of scientific evidence for

1614

individual evidence streams and then used that information to evaluate the overall weight of evidence

1615

supporting each set of exposure estimates. General considerations for evaluating the strength of evidence

1616

for each evidence stream are summarized in Table 7-6 of the Draft Systematic Review Protocol

1617

Supporting TSCA Risk Evaluations for Chemical Substances (U.S. EPA, 2021b). Specific examples of

1618

how these considerations can be applied to overall weight of scientific evidence conclusions are

1619

provided in Table 7-7 of the Draft Systematic Review Protocol (U.S. EPA, 2021b). The weight of

1620

scientific evidence supporting each element of the human health exposure assessment are discussed in

1621

the occupational exposure assessment (U.S. EPA, 2024k) consumer exposure assessment (U.S. EPA,

1622

2024d), indoor air assessment (U.S. EPA, 2024j) and ambient air assessment (U.S. EPA, 2024a)

1623

modules.

1624

1625

Overall confidence descriptions of high, medium, or low are assigned to the exposure assessment based

1626

on the strength of the underlying scientific evidence. When the assessment is supported by robust

1627

evidence, overall confidence in the exposure assessment is high; when supported by moderate evidence,

1628

overall confidence is medium; when supported by slight evidence, overall confidence is low.

1629

1630

OPPT will seek input on its use of inputs and assumptions in the exposure assessments for occupational,

1631

consumer, outdoor air, and indoor air scenarios, in part to understand whether its approach may

1632

compound one conservative assumption upon another in a manner that leads to unrealistic or un-

1633

addressable outcomes.

1634

Overall Confidence in Occupational Exposure Assessment

1635

The confidence in the occupational exposure assessment varies from low to high, the confidence is

1636

based on the strengths, limitations, and uncertainties associated with the exposure estimates for each

1637

individual occupational exposure scenario. Most COUs have medium confidence based on moderate to

1638

robust and moderate weight of scientific evidence conclusions. The primary strength of most of the

1639

inhalation assessments is that it uses monitoring data that is chemical-specific and is directly applicable

1640

to the exposure scenario. The use of applicable monitoring data is preferable to other assessment

1641

approaches such as modeling or the use of occupational exposure limits. The principal limitation of the

1642

monitoring data is the uncertainty in the representativeness of the data due to some scenarios having

1643

limited exposure monitoring data in the literature or the available monitoring data lacking additional

1644

contextual information. Additionally, different sampling objectives may introduce uncertainty since

1645

OSHA and other studies may target workers with the highest expected exposures. For many of the

1646

COUs, the EPA received aggregated data from industry; therefore, EPA was unable to distinguish each

1647

site’s contribution to the exposure estimates. EPA also assumed 8 exposure hours per day and 250

1648

exposure days per year based on continuous formaldehyde exposure for each working day for a typical

1649

worker schedule. It is uncertain whether this captures actual worker schedules and exposures.

1650

1651

Some of the COUs lacked monitoring data; therefore, EPA used models to estimate inhalation

1652

exposures. EPA addressed variability in inhalation models by identifying key model parameters to apply

1653

a statistical distribution that mathematically defines the parameter’s variability. EPA defined statistical

1654

distributions for parameters using documented statistical variations where available. Where the

1655

statistical variation was unknown, assumptions were made to estimate the parameter distribution using

1656

available literature data, such as General Scenario (GS) and Emission Scenario Document (ESDs).

1657

However, there is uncertainty as to the representativeness of the parameter distributions with respect to

1658

the modeled scenario because the data are often not specific to sites that use formaldehyde. In general,

1659

the effects of these uncertainties on the exposure estimates are unknown, as the uncertainties may result

1660

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 70 of 151

in either overestimation or underestimation of exposures depending on the actual distributions of each of

1661

the model input parameters.

1662

1663

As described in the Draft Occupational Exposure Assessment for Formaldehyde (U.S. EPA, 2024k),

1664

EPA has low confidence in the inhalation estimates for the four COUs below based on a slight weight of

1665

scientific evidence:

1666

• Industrial use – non-incorporative activities – process aid in: oil and gas drilling, extraction, and

1667

support activities; process aid specific to petroleum production, hydraulic fracturing

1668

• Commercial use – chemical substances in treatment/care products – laundry products and

1669

dishwashing products

1670

• Commercial use – chemical substances in outdoor use products – explosives materials

1671

• Commercial use– chemical substances in packaging, paper, plastic, hobby products – paper

1672

products; plastic and rubber products; toys, playground, and sporting equipment

1673

This was mainly due to the low number of monitoring samples available, lack of information specific to

1674

formaldehyde usage for the given COUS and uncertainties with the representativeness of the monitoring

1675

data. However, EPA concluded that the underlying data still provides a plausible estimate of exposures

1676

for these OESs.

1677

1678

EPA had moderate weight of scientific evidence conclusions for all dermal scenarios assessed. The

1679

primary strength of the dermal assessment is that most of the data that EPA used to inform the modeling

1680

parameter distributions have overall data quality determinations of either high or medium from EPA’s

1681

systematic review process, such as the 2020 CDR (U.S. EPA, 2020b). A limitation of the assessment is

1682

that some COUs lacked formaldehyde weight concentration data.

1683

Overall Confidence in the Consumer Exposure Assessment

1684

EPA has medium confidence in the inhalation exposure assessment for consumers. As detailed in

1685

Section 3.2 of the Draft Consumer Exposure Assessment for Formaldehyde (U.S. EPA, 2024d), the

1686

inhalation exposure assessment is supported by a robust monitoring dataset and robust modeling

1687

approaches.

1688

1689

Aside from the potential exposures to water treatment, laundry and dish washing, and lawn and garden

1690

products, EPA has medium confidence in the consumer inhalation modeling approaches and model input

1691

data—including TSCA COU-specific product weight fractions identified from SDS of consumer

1692

products currently on the market, the quality and applicability of the CEM for the assessment of realistic

1693

consumer exposure scenarios that are representative of COUs, common consumer use patterns (e.g.,

1694

TSCA COU-specific amount used, duration and frequency of use (U.S. EPA, 2019)) according to the

1695

EPA Exposure Factors Handbook (U.S. EPA, 2011) and the 1987 Westat survey (Westat, 1987) and

1696

applicable to most population groups. EPA also has medium confidence in the quality and

1697

representativeness of air monitoring data. This use of TSCA COU-specific monitoring information

1698

increases confidence in estimated inhalation exposures.

1699

1700

EPA has medium confidence in the dermal exposure assessment for consumers. As detailed in Section

1701

3.2 of the Draft Consumer Exposure Assessment for Formaldehyde (U.S. EPA, 2024d), EPA has

1702

medium confidence in the Thin Film Model, which EPA used to estimate dermal loading from spray and

1703

liquid consumer products, and in default model input values used in the dermal exposure assessment of

1704

realistic consumer exposure scenarios, which are representative of COUs, common consumer use

1705

patterns, and applicable to most population groups. EPA has high confidence in the TSCA COU-specific

1706

product weight fractions identified from SDSs of consumer products currently on the market and

1707

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 71 of 151

medium confidence in the applied quantity remaining on skin (Q

) constant. Although a Q

of 10.3

1708

mg/cm

(used to approximate hand immersion and wiping experiments using oil-based products (U.S.

1709

EPA, 1992)) is assumed to be realistic and protective of most liquid product consumer dermal exposures

1710

to formaldehyde, it is conceivable that a lower Q

may be applicable for some consumer exposure

1711

scenarios (e.g., consumer uses liquid product with personal protective equipment [PPE] that prevents

1712

immersion or development of thin film of formaldehyde on the skin). No monitoring data are available

1713

on dermal exposures for consumers.

1714

Overall Confidence in the Indoor Air Exposure Assessment

1715

EPA has medium confidence in the overall findings for the indoor air exposure assessment. As detailed

1716

in Section 3.2.1 of the Draft Indoor Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024j), the

1717

exposure assessment is supported by a robust monitoring dataset and robust modeling approaches. EPA

1718

has medium confidence that the exposure scenarios evaluated in this assessment are reasonable and

1719

representative of people who spend most time indoors. The indoor air exposure scenario assumes

1720

continuous exposure to indoor air over a lifetime.

1721

1722

EPA has medium confidence in the quality and representativeness of indoor air monitoring data. The set

1723

of 16 studies used as an indication of indoor air concentrations and as a basis for comparison to modeled

1724

concentrations were rated high quality. This dataset includes the American Healthy Homes Survey II, a

1725

quality nationally representative formaldehyde residential indoor air monitoring study administered by

1726

EPA and the U.S. Department of Housing and Urban Development (HUD). EPA also has medium

1727

confidence in the indoor air modeling approaches and model input data, including the quality and

1728

applicability of the Consumer Exposure Model and the emission rates and fluxes from quality product

1729

emission studies used to refine the model. The set of nine studies incorporated into indoor air modeling

1730

were, altogether, rated medium quality.

1731

1732

EPA considered concordance between monitored and modeled concentrations. Monitored concentrations

1733

are expected to reflect aggregate concentrations resulting from multiple sources of formaldehyde and are

1734

therefore not directly comparable to modeled concentrations estimated for specific sources. In addition,

1735

CEM does not incorporate chemical half-life. Therefore, it is unclear whether the modeling results are

1736

reflective of most indoor air home environments in American residences. However, the fact that

1737

modeled concentrations are within the same order of magnitude of monitored concentrations increases

1738

confidence in modeled concentrations. The availability of both modeled concentrations and monitoring

1739

data provides information about both the aggregate exposures from all sources contributing to indoor air

1740

concentrations as well as information about the relative contributions of individual TSCA COUs.

1741

1742

Based on consideration of the weight of scientific evidence, EPA has medium confidence in the overall

1743

findings for the indoor air exposure assessment (U.S. EPA, 2024j) due to a high confidence in the CEM

1744

used and emission fluxes and rates from quality product emission studies used to refine the model, in

1745

comparison with American Healthy Homes Survey II.

1746

Overall Confidence in the Ambient Air Exposure Assessment

1747

EPA has high confidence in the overall characterization of exposures for the ambient air exposure

1748

assessment. As described in the Draft Ambient Air Exposure Assessment for Formaldehyde (U.S. EPA,

1749

2024a), exposure estimates rely upon direct reported releases and peer-reviewed models to derive

1750

exposure concentrations at distances from releasing facilities where individuals within the general

1751

population reside for many years. Furthermore, ambient monitoring data supports the presence of

1752

formaldehyde in the ambient air and shows comparable monitored values to EPA’s modeled

1753

concentrations.

1754

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 72 of 151

1755

For industrial TSCA COUs, EPA has a moderate to robust weight of scientific evidence as the databases

1756

have high data quality scores and are supported by numerous data points. A primary strength of TRI and

1757

NEI data is that these programs compile the best readily available release data for large facilities.

1758

Limitations are that these programs may not cover some sites that emit formaldehyde as both programs

1759

have conditions that must be met prior to being required to report releases. For formaldehyde, the

1760

potential contribution of combustion sources is an uncertainty and use of the full facility data complicate

1761

singular TSCA COU estimates, such that emissions at one site may include multiple sources under

1762

multiple COUs that include combustion sources and non-combustion sources.

1763

1764

In general, for commercial COUs, EPA has a moderate weight of scientific evidence as TRI and NEI

1765

have high data quality and generic scenarios that have a medium to high data quality rating. EPA relied

1766

upon professional judgement in mapping TRI and NEI industrial sectors to commercial COUs. There is

1767

some uncertainty that a commercial TSCA COU may occur across several industrial sectors beyond the

1768

industrial sector used for analysis. In addition, some industrial sectors cover both industrial and

1769

commercial operations, so they may overestimate air releases occurring in a commercial setting. Four

1770

commercial COUs either lacked sufficient data or was supported by a slight weight of evidence:

1771

• Commercial use – chemical substances in treatment/care products – laundry and dishwashing

1772

products;

1773

• Commercial use – chemical substances in treatment products – water treatment products;

1774

• Commercial use – chemical substances in outdoor use products – explosive materials; and

1775

• Commercial use – chemical substances in products not described by other codes – other:

1776

laboratory chemicals.

1777

EPA estimated the exposed population to modeled releases to ambient air; however, these estimates are

1778

considered an underestimate of total exposed population. EPA limited this modeling to the 810 TRI

1779

facilities directly reporting with Form R. As indicated in the TRI reporting, the ambient air releases

1780

reported to EPA are from different estimation approaches (e.g., emission factors) and may not be from

1781

active stack monitoring. These TRI emissions are a subset of the approximately 49,000 distinct facilities

1782

with estimated emissions in NEI but are of greater confidence due to the direct reporting rather than the

1783

indirect, state-specific reporting currently used to develop the NEI. Finally, the exposed population

1784

estimates from HEM are derived by averaging the modeled annual concentration at the proximate census

1785

block centroids across the census block, using site-specific meteorological conditions. EPA did not

1786

make facility-specific adjustments to modeling receptor files based on land use analysis to capture the

1787

highest proximate populations in this analysis, therefore population estimates are biased against

1788

capturing the populations of the most highly exposed residents within rural (and therefore larger) census

1789

blocks. Therefore, while EPA has a high confidence in the methods used, based on the expected

1790

underestimation of the exposed population estimates, the confidence is medium.

1791

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 73 of 151

3 HUMAN HEALTH HAZARD SUMMARY

1792

EPA’s OPP and OPPT collaborated to develop a joint hazard assessment for formaldehyde (U.S. EPA,

1793

2024i). This joint assessment evaluated available human health hazard and dose-response information

1794

for formaldehyde and identified hazard values to support risk assessments in both offices.

1795

1796

For cancer and non-cancer hazards associated with chronic inhalation exposures, the joint hazard

1797

assessment relies upon the analysis already completed in the draft IRIS assessment on formaldehyde

1798

inhalation (U.S. EPA, 2022b) and peer reviewed by the National Academies of Sciences, Engineering,

1799

and Medicine (NASEM) (NASEM, 2023). The systematic review literature searches, data quality

1800

review, evidence integration, dose-response analyses, and peer review performed in support of the IRIS

1801

assessment reflect the best available science on formaldehyde hazards from chronic inhalation exposures

1802

and are consistent with the needs of both OPP and OPPT.

1803

1804

To identify additional available hazard and dose-response information for acute inhalation, dermal, and

1805

oral formaldehyde exposures, EPA used a fit-for-purpose systematic review protocol, integrating the

1806

needs and approaches of both OPP and OPPT. Details of the fit-for-purpose systematic review protocol

1807

used in OPPT's work on this assessment are described in the Systematic Review Protocol for the Draft

1808

Risk Evaluation for Formaldehyde (U.S. EPA, 2023a). This approach is based in part on the OPPT

1809

systematic review approach described in the Draft Systematic Review Protocol Supporting TSCA Risk

1810

Evaluations for Chemical Substances (U.S. EPA, 2021b).

1811

1812

EPA identified a range of factors that may increase susceptibility to formaldehyde and considered

1813

susceptibility throughout the hazard assessment. Descriptions of how EPA incorporated PESS due to

1814

greater biological susceptibility into the risk evaluation are provided in Appendix C. Factors that may

1815

increase susceptibility to formaldehyde exposures include chronic respiratory disease, lifestage, sex, and

1816

co-exposure to chemical and non-chemical stressors that influence the same health outcomes.

1817

3.1 Summary of Hazard Values

1818

The non-cancer and cancer hazard values identified for inhalation, dermal, and oral exposures to

1819

formaldehyde in the joint hazard assessment (U.S. EPA, 2024i) are summarized in Table 3-1.

1820

Consistent with the recommendations of the Human Studies Review Board (HSRB), OPPT will seek

1821

input on its hazard assessment, particularly with regards to the PODs and uncertainty/extrapolation

1822

factors for acute and chronic non-cancer assessment and the extent to which the draft hazard assessment

1823

for formaldehyde appropriately considered recommendations from other federal advisory committees

1824

(e.g., NASEM, HSRB).

1825

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 74 of 151

Table 3-1. Hazard Values Identified for Formaldehyde

1826

Exposure

Scenario

Hazard Value

Uncertainty

Factors

Total

Uncertainty

Factor

Study and Toxicological Effects

Inhalation

Acute

(15-minute

duration)

NOAEC and BMCL =

0.5 ppm

(0.62 mg/m

) as a 15-

minute peak exposure

= 10

Total UF = 10

Kulle et al, (1987); supported by:

LOAEC = 1 ppm (mg/m

) based on eye irritation in adult volunteers

Mueller et al. (2013)

LOAEC = 0.3 ppm over 4 hours, with 15-minute peaks of 0.6 ppm, based on eye

irritation in hypersensitive adult volunteers

Lang et al. (2008)

LOAEC= 0.5 ppm over 4 hours, with peaks of 1 ppm (0.62/1.23 mg/m

), based

on eye irritation in adult volunteers

Inhalation

Chronic non-

cancer

(Long-term, >6

months)

BMCL

= 0.017 ppm

(0.021 mg/m

)

= 3

Total UF = 3

POD is derived from the IRIS RfC (U.S. EPA, 2022b). The specific BMCL

value used here is based on reduced pulmonary function in children in

Krzyzanowski et al. (1990), but is consistent with the RfC, derived based on,

pulmonary function, allergy-related conditions, asthma (prevalence and degree of

asthma control) in people, as reported in Annesi-Maesano et al. (2012),

Matsunaga et al. (2008), Venn et al. (2003), and Krzyzanowski et al. (1990).

Inhalation

Chronic

Cancer

Adult-based IUR:

0.0079 ppm

-1

(6.4 × 10

-6

(µg/m

)

-1

)

ADAF-adjusted IUR:

0.013 ppm

-1

(1.1 × 10

–5

(µg/m

)

-1

)

N/A

IUR established by IRIS (U.S. EPA, 2022b) based on data on nasopharyngeal

cancer in people reported in Beane-Freeman et al. (2013).

Dermal

Acute

Induction:

= 0.4% (100

μg/cm

) in 4:1

acetone:olive oil

= 10

Total UF= 100

Basketter et al., (2003)

based on induction of dermal sensitization in mice

Elicitation:

BMDL

= 10.5

µg/cm

(0.035%)

= 10

Total UF = 10

Flyvholm et al., (1997)

based on threshold for elicitation of dermal sensitization in people

Oral

Short-Term/

subchronic

(1-30 days),

HED= 6 mg/kg-day

= 3

= 10

Total UF = 30

Til (1988)

NOAEL= 25 mg/g-day; LOAEL = 125 mg/kg-day based on gastrointestinal

histopathology in rats

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 75 of 151

Exposure

Scenario

Hazard Value

Uncertainty

Factors

Total

Uncertainty

Factor

Study and Toxicological Effects

Oral

Chronic

HED = 3.6 mg/kg-day

= 3

= 10

Total UF = 30

Civo Inst.(1987); Til (1989)

NOAEL= 15 mg/g-day; LOAEL = 82 mg/kg-day based on gastrointestinal

histopathology in rats

This value is used to estimate risks from both sub-chronic and chronic occupational exposures.

Point of departure (POD) = A data point or an estimated point that is derived from observed dose-response data and used to mark the beginning of extrapolation to

determine risk associated with lower environmentally relevant human exposures; NOAEL = no-observed adverse-effect level; LOAEL = lowest-observed adverse-

effect level; UF = uncertainty factor; UF

= extrapolation from animal to human (interspecies); UF

= potential variation in sensitivity among members of the

human population (intraspecies). UF

= use of a LOAEL to extrapolate a NOAEL. UF

= use of a short-term study for long-term risk assessment. UF

= to account

for the absence of key data (i.e., lack of a critical study). IUR= inhalation unit risk; ADAF-adjusted IUR = IUR for calculating cancer risks associated with a full

lifetime of exposure

1827

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 76 of 151

3.2 Weight of Scientific Evidence and Overall Confidence in Hazard

1828

Assessment

1829

As described in the Draft Systematic Review Protocol Supporting TSCA Risk Evaluations for Chemical

1830

Substances (U.S. EPA, 2021b), the weight of scientific evidence supporting hazard assessment and dose

1831

response is evaluated based on the quality of the key studies, consistency of effects across studies, the

1832

relevance of effects for human health, confidence in the dose-response models, and the coherence and

1833

biological plausibility of the effects observed. The weight of evidence and overall confidence in chronic

1834

inhalation hazard values derived by IRIS are described in the draft IRIS assessment (U.S. EPA, 2022b).

1835

The weight of evidence and sources of confidence and uncertainty in dermal, oral, and acute inhalation

1836

hazard values derived by OCSPP are described in the hazard assessment (U.S. EPA, 2024i). This section

1837

summarizes overall confidence and sources of uncertainty in the hazard values used to develop risk

1838

estimates in this risk characterization.

1839

Overall Confidence in the Acute Inhalation POD

1840

Overall confidence in the acute inhalation POD is medium. As described in the joint hazard assessment

1841

(U.S. EPA, 2024i), the acute POD is based on a robust dataset of evidence for sensory irritation in

1842

humans, including several high-quality controlled exposure studies with relevance for acute exposure

1843

scenarios. Concordance of reported sensory irritation effects and the effect levels reported across acute

1844

exposure studies increases confidence in the final POD. Variability across individuals’ response

1845

contributes to uncertainty around effect levels that are protective across the population. A 10x

1846

uncertainty factor is applied to account for uncertainty related to intraindividual variability.

1847

1848

This acute POD focuses on defining peak threshold exposure concentrations rather than average 8- or

1849

24-hour exposure concentrations. There is some uncertainty around the degree to which duration

1850

influences effect levels because there are no studies available that provide direct evidence that effect

1851

levels following 8- or 24-hour exposures are the same as effects following 2 to 5 hours of exposure.

1852

1853

Immune-mediated respiratory effects like asthma may also have relevance for acute hazard, but

1854

available studies do not provide sufficient information to characterize dose-response relationships for

1855

acute inhalation exposures. Although this may be a source of uncertainty for the acute POD, dose-

1856

response data for these additional respiratory endpoints are used as the basis for the chronic inhalation

1857

POD.

1858

Overall Confidence in the Chronic, Non-cancer Inhalation POD

1859

As described in the draft IRIS assessment (U.S. EPA, 2022b), overall confidence in the chronic non-

1860

cancer inhalation POD is high. The chronic POD derived by IRIS is supported by a robust database of

1861

evidence for a range of endpoints in humans and animals. The overall POD is informed by dose-

1862

response information in humans across multiple respiratory endpoints and reflects concordance in effect

1863

levels identified across those endpoints. EPA also considered dose-response information for

1864

reproductive and developmental effects in selection of the overall POD. While there is more uncertainty

1865

around the PODs derived for these endpoints, the overall POD is expected to be protective of these

1866

reproductive and developmental effects in humans. Many of the observational epidemiology studies

1867

providing the quantitative basis for the chronic POD reflect relevant human exposure scenarios in homes

1868

and schools. In addition, several of the studies include children with asthma or other sensitive groups.

1869

1870

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 77 of 151

Overall Confidence in the Chronic IUR

1871

As described in the draft IRIS assessment (U.S. EPA, 2022b), overall confidence in the preferred unit

1872

risk estimate is medium. The IUR derived for nasopharyngeal cancer is informed by a robust dataset of

1873

both human and animal data. The availability of human data eliminates the need to extrapolate from

1874

animal studies, increasing the confidence in the IUR. In addition, the IUR derived from animal data is

1875

similar to the IUR derived from human evidence, further increasing confidence in the IUR. Sources of

1876

uncertainty in the IUR include reliance on extrapolation from high doses that occur in occupational

1877

settings to lower doses that may occur in the general population, reliance on data from a single high

1878

quality occupational cohort study that may not capture the sensitivity of susceptible populations or

1879

lifestages, and reliance on mortality data as a surrogate for cancer incidence.

1880

1881

EPA was not able to derive IURs for all tumor sites associated with formaldehyde exposure. This is a

1882

source of uncertainty and may lead to an underestimate of risk. Although EPA was able to derive an

1883

IUR for myeloid leukemia, the lack of confidence in the dose-response data and IUR for myeloid

1884

leukemia is a source of uncertainty. The cancer risk estimates presented in this risk characterization do

1885

not include risks for myeloid leukemia and other tumor sites. Based on the IUR estimated for myeloid

1886

leukemia in the draft IRIS document, IRIS estimated that consideration of myeloid leukemia may

1887

increase the age-dependent adjustment factor (ADAF)-adjusted IUR by as much as four-fold.

1888

Overall Confidence in the Dermal POD

1889

Overall confidence in the dermal POD is medium. As described in the OCSPP joint hazard assessment

1890

(U.S. EPA, 2024i), the dermal POD is derived from an extensive dataset on dermal sensitization in

1891

human, animal, and in vitro studies. Multiple streams of evidence from studies evaluating elicitation

1892

thresholds in sensitive people and induction thresholds in animal and in in vitro assays arrive at similar

1893

effect levels. While there are some uncertainties associated with the human studies related to lack of

1894

clarity in methods and data reporting, concordance in effect levels across multiple streams of evidence

1895

increases confidence in the POD. The potential impact of methanol present in available dermal

1896

formaldehyde studies is a source of uncertainty in the POD. While there is substantial variation in

1897

sensitization responses across individuals, application of a 10× uncertainty factor is used to account for

1898

uncertainty related to intraindividual variability.

1899

Overall Confidence in the Subchronic and Chronic Oral PODs

1900

Overall confidence in the subchronic and chronic oral PODs is medium. As described in the OSCPP

1901

joint hazard assessment (U.S. EPA, 2024i), the subchronic and chronic oral PODs rely on a limited

1902

database of animal studies but are supported by three studies that report consistent patterns of

1903

gastrointestinal damage at similar dose levels.

1904

1905

Due to technical challenges around generating pure and stable formaldehyde treatments for oral

1906

exposure, most of the available animal studies have major limitations and uncertainties. Among the

1907

available studies that are not confounded by the presence of methanol, gastrointestinal effects are the

1908

most sensitive endpoint evaluated. Reduced drinking water intake in the high dose groups reduced

1909

confidence in each of the chronic studies when considered in isolation. However, when considered in

1910

conjunction with the results of the 28-day study that included water-restricted controls, EPA has

1911

confidence that the reported effects are attributable to formaldehyde exposure.

1912

1913

There is very limited information on reproductive, developmental, and immune endpoints following oral

1914

exposure to formaldehyde. Although there are some studies that suggest effect levels for these endpoints

1915

may be more sensitive than those used as the basis for the POD, the only studies that evaluate

1916

reproductive, developmental, and immune endpoints are confounded by the presence of methanol.

1917

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 78 of 151

Evidence of reproductive and developmental effects reported in humans and animals following

1918

inhalation exposure to formaldehyde indicates that such effects are possible following formaldehyde

1919

exposure. Similarly, the available data do not evaluate factors that may increase susceptibility to oral

1920

formaldehyde exposure in sensitive groups or lifestages. The lack of data on these endpoints and

1921

sensitive groups and lifestages following oral exposure could be perceived as uncertainty; however, the

1922

likelihood of a lower POD being identified based on these outcomes is low given the effect used as the

1923

basis of the current PODs (gastrointestinal effects) are close to the portal of entry, first pass metabolism

1924

via the oral route, and the reactivity of formaldehyde.

1925

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 79 of 151

4 HUMAN HEALTH RISK CHARACTERIZATION

1926

4.1 Risk Characterization Approach

1927

The exposure scenarios, populations of interest, and toxicological endpoints used for evaluating risks

1928

from acute and chronic exposures are summarized below in Table 4-1. EPA estimated cancer and non-

1929

cancer risks from occupational, consumer, and general population exposures as described below.

1930

1931

While EPA will consider the standard risk benchmarks shown in Table 4-1 associated with interpreting

1932

margins of exposure and cancer risks, EPA cannot solely rely on those risk values. Risk estimates

1933

include inherent uncertainties and the overall confidence in specific risk estimates varies. The analysis

1934

provides support for the Agency to make a determination about whether formaldehyde poses an

1935

unreasonable risk to human health and to identify drivers of unreasonable risk among exposures for

1936

people (1) with occupational exposure to formaldehyde, (2) with consumer exposure to formaldehyde,

1937

(3) with exposure to formaldehyde in indoor air, and (4) who live or work in proximity to locations

1938

where formaldehyde is released to air. Concurrent with this draft TSCA Risk Evaluation, EPA is

1939

releasing a preliminary risk determination for formaldehyde.

1940

1941

The Agency also will consider naturally occurring sources of formaldehyde (i.e., biogenic, combustion,

1942

and secondary formation) and associated risk levels from, and consider contributions from all sources as

1943

part of a pragmatic and holistic evaluation of formaldehyde hazard and exposure in making its

1944

unreasonable risk determination. If an estimate of risk for a specific scenario exceeds the benchmarks,

1945

then the decision of whether those risks are unreasonable is both case-by-case and context driven. In the

1946

case of formaldehyde, EPA is taking the risk estimates of this draft human health risk assessment

1947

(HHRA) in combination with a thoughtful consideration of other sources of formaldehyde, to interpret

1948

the risk estimates in the context of an unreasonable risk determination.

1949

1950

Table 4-1. Use Scenarios, Populations of Interest, and Toxicological Endpoints Used for Acute and

1951

Chronic Exposures

1952

Populations

of Interest

and

Exposure

Scenarios

Workers

Acute – Adolescent (≥16 years old) and adult workers exposed to formaldehyde in a single

workday for 15 min or longer

Chronic – Adolescent (≥16 years old) and adult workers exposed to formaldehyde over a full-shift

workday for 250 days per year for 40 working years

Consumers and Bystanders

Acute – Consumers across all age groups (depending on the product or article) exposed to

formaldehyde result from product or article use. Exposures are estimated to be 15-minute peak

concentrations. It should be noted that the 15-minute peak concentration for a given TSCA COU

and exposure scenario may occur several hours after product use.

Chronic – Consumers across all age groups (depending on the product or article) exposed to

formaldehyde result from product or article use up to 78 years.

General Population Indoor Ambient Air Exposure

Chronic – People across all age groups exposed to formaldehyde through ambient air continuously

up to 78 years.

General Population Outdoor Ambient Air Exposure

Chronic – People across all age groups exposed to formaldehyde through ambient air near

industrial release site continuously up to 78 years.

Health

Effects,

Hazard

Non‐cancer Acute Hazard Values

Acute inhalation health effect: sensory irritation

• Acute inhalation POD (15-minute duration) = 0.5 ppm (0.62 mg/m

)

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 80 of 151

Values and

Benchmarks

• Uncertainty Factors (Benchmark MOE) = 10 (UF

= 1; UF

= 10; UF

= 1; UF

=1;

=1)

Acute dermal health effect: sensitization (elicitation)

• Acute POD = 10.5 µg/cm

• Uncertainty factors (Benchmark MOE) = 10 (UF

= 1; UF

= 10; UF

= 1; UF

=1;

=1)

Acute oral health effect: no acute oral PODs identified

Non-cancer Subchronic Hazard Values

Subchronic oral health effects: Gastrointestinal effects

• Oral HED = 6 mg/kg-day

• Uncertainty Factors (Benchmark MOE) = 30 (UF

= 3; UF

= 10; UF

= 1; UF

=1;

=1)

Non-cancer Chronic Hazard Values

Chronic inhalation health effects: Respiratory effects, including reduced pulmonary function,

allergy-related conditions, asthma (prevalence and degree of asthma control), and sensory

irritation

• Inhalation HEC = 0.017 ppm (0.021 mg/m

)

• Uncertainty Factors (Benchmark MOE) = 3 (UF

= 1; UF

= 3; UF

= 1; UF

Chronic oral health effects: Gastrointestinal effects

• Oral HED = 3.6 mg/kg-day

• Uncertainty Factors (Benchmark MOE) = 30 (UF

= 3; UF

= 10; UF

= 1; UF

= 1)

Cancer Hazard Values

Inhalation cancer hazard for formaldehyde is based on nasopharyngeal cancers

• IUR = 0.0079 ppm

−1

(6.4×10

-6

(µg/m

)

−1

)

• ADAF applied for early life exposures

Oral and dermal cancer hazards are not quantified because there is insufficient data to support

derivation of cancer slope factors for these routes of exposure.

Adult workers (≥16 years old) include both female and male workers.

Inhalation exposures are described in terms of air concentrations and do not include lifestage-specific adjustments; risk

estimates based on air concentrations are intended to address risks to all lifestages.

MOE = margin of exposure; UF

= Interspecies uncertainty factor for animal-to-human extrapolation; UF

= Intraspecies

uncertainty factor for human variability; UF

= LOAEC-to-NOAEC uncertainty factor for reliance on a LOAEC as the POD

Estimation of Non-cancer Risks

1953

EPA used a margin of exposure (MOE) approach to identify potential non-cancer risks. The MOE is the

1954

ratio of the non-cancer POD divided by a human exposure dose. Acute and chronic MOEs for non-

1955

cancer inhalation and dermal risks were calculated using Equation 4-1:

1956

1957

Equation 4-1.

1958







  



1959

1960

Where:

1961

MOE = Margin of exposure (unitless)

1962

Hazard value (POD) = HEC (ppm) or HED (mg/kg-d)

1963

Human Exposure = Exposure estimate (in ppm or mg/kg-d)

1964

1965

MOE risk estimates may be interpreted in relation to benchmark MOEs. Benchmark MOEs are typically

1966

the total UF for each non‐cancer POD. If the numerical value of the MOE is less than the benchmark

1967

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 81 of 151

MOE, this relationship is a starting point to determine if there are unreasonable non-cancer risks. On the

1968

other hand, if the MOE estimate is equal to or exceeds the benchmark MOE, risk is not indicated.

1969

Typically, the larger the MOE, the more unlikely it is that a non‐cancer adverse effect occurs relative to

1970

the benchmark. When determining whether a chemical substance presents unreasonable risk to human

1971

health or the environment, calculated risk estimates are not “bright-line” indicators of unreasonable risk,

1972

and EPA has discretion to consider other risk-related factors apart from risks identified in risk

1973

characterization.

1974

Estimation of Cancer Risks

1975

Extra cancer risks for repeated inhalations exposures to formaldehyde were estimated using Equation

1976

4-2:

1977

1978

Equation 4-2.

1979

    

1980

1981

Where:

1982

Risk = Extra cancer risk (unitless)

1983

Human exposure = Exposure estimate (LADC in ppm)

1984

IUR = Inhalation unit risk

1985

1986

EPA has concluded that “the evidence is sufficient to conclude that a mutagenic mode of action of

1987

formaldehyde is operative in formaldehyde-induced nasopharyngeal carcinogenicity” (U.S. EPA,

1988

2022b). To account for increased nasopharyngeal cancer risks from early life exposures to

1989

formaldehyde, EPA applies an ADAF.

1990

1991

Estimates of extra cancer risks are interpreted as the incremental probability of an individual developing

1992

cancer over a lifetime following exposure (i.e., incremental, or extra individual lifetime cancer risk).

1993

4.2 Risk Estimates

1994

Risk Estimates for Workers

1995

EPA estimated cancer and non-cancer risks for workers exposed to formaldehyde based on the

1996

occupational exposure estimates that were described in Section 2.1. For many TSCA COUs, EPA did

1997

not identify inhalation exposure data for ONUs, and therefore evaluated chronic risks using the central

1998

tendency estimates for workers. EPA did not identify information for potential peak exposures by ONUs

1999

and therefore did not quantify acute inhalation risks for ONUs. Risks to ONUs are assumed to be equal

2000

to or less than risks to workers who handle materials containing formaldehyde as part of their job.

2001

2002

These risk estimates are based on exposures to workers in the absence of PPE such as gloves or

2003

respirators. Section 2.5.1 contains an overall discussion on strengths, limitations, assumptions, and key

2004

sources of uncertainty for the occupational exposure assessment. Additionally, the Draft Occupational

2005

Exposure Assessment for Formaldehyde (U.S. EPA, 2024k) contains a comprehensive weight of

2006

scientific evidence summaries, which presents an OES-by-OES discussion of the key factors that

2007

contributed to each weight of scientific evidence conclusion.

2008

4.2.1.1 Risk Estimates for Inhalation Exposures

2009

EPA estimated acute, sub-chronic and chronic non-cancer and chronic cancer risks to workers and

2010

ONUs from inhalation. Generally, EPA expects workers to be exposed at higher formaldehyde

2011

concentrations comparative to other populations. Across occupational exposure scenarios for full-shift

2012

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 82 of 151

estimates, the central tendency of air concentrations estimates ranged from 7.5 to 499.3 µg/m

(0.006 to

2013

0.40 ppm) and high-end of air concentrations estimates ranged from 7.5 to 17,353.3 µg/m

(0.006 to

2014

13.9 ppm), which is generally higher than the modeled estimates of ambient air (up to 5.7 µg/m

) and

2015

measured indoor air concentrations (~40 µg/m

at the 95th percentile of concentrations measured in

2016

AHHS II).

2017

2018

Risk estimates vary across OESs/COUs. As shown in Figure 4-1, acute non-cancer risk estimates for

2019

worker inhalation exposure range from 2.58×10

−3

to 11.6 for both high-end and central tendency

2020

exposures. For COUs with multiple OESs or estimation approaches, the estimate with the highest high-

2021

end value was illustrated. For the formaldehyde risk assessment, acute occupational risks were estimated

2022

using 15-minute monitoring data, which in most cases is expected to represent activities with the highest

2023

exposure potential for the scenario. Acute risk estimates below indicate that exposure is greater than the

2024

hazard POD identified for 15-minute peak exposures based on sensory irritation reported in controlled

2025

human exposure studies in healthy adult volunteers. All TSCA COUs except one COU have acute risk

2026

estimates below an MOE of 10, and 39 TSCA COUs have acute risk estimates below an MOE of 1.

2027

2028

EPA did not identify inhalation exposure data for peak exposures for the industrial use as process aid in:

2029

Oil and gas drilling, extraction, and support activities; process aid specific to petroleum production,

2030

hydraulic fracturing. Of note, the Commercial use – laundry and dishwashing products COU only had

2031

one identified data point for peak exposures, and therefore one risk value is provided.

2032

2033

As shown in Figure 4-2, chronic non-cancer risk estimates for worker inhalation exposure range from

2034

2.42×10

−3

to 6.4 for both high-end and central tendency exposures. For COUs with multiple OESs or

2035

estimation approaches, the scenario with the highest central tendency value was illustrated. Chronic non-

2036

cancer risk estimates below 1 indicate that exposure is greater than the hazard point of departure based

2037

on respiratory effects in children. While some healthy adult workers may be less susceptible to

2038

formaldehyde at those concentrations, MOEs below 1 may be a concern for susceptible workers such as

2039

those with chronic respiratory disease or those with co-exposures that contribute to similar respiratory

2040

effects. Of the 49 TSCA COUs evaluated, 48 TSCA COUs have chronic risk estimates below an MOE

2041

of 3, and 47 TSCA COUs have chronic risk estimates below an MOE of 1. Sub-chronic, non-cancer risk

2042

estimates follow a similar risk profile and are not separately illustrated.

2043

2044

Worker cancer risk estimates for inhalation exposure range from 4.05×10

−6

to 1.3×10

−2

for both high-

2045

end and central tendency exposures, as shown in Figure 4-3. For COUs with multiple OESs or

2046

estimation approaches, the scenario with the highest central tendency value was illustrated. The cancer

2047

risk estimates calculated for workers do not include risks for myeloid leukemia and other tumor sites

2048

because EPA was not able to quantify those risks with confidence. Cancer risk estimates may therefore

2049

underestimate risks. Of the 49 TSCA COUs evaluated, 46 TSCA COUs have chronic risk estimates

2050

greater than 1 in 10,000. All risk estimates including for all exposure scenarios evaluated are provided in

2051

the “Supplemental file: Occupational Risk Calculator.”

2052

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 83 of 151

2053

Figure 4-1. Acute, Non-cancer Occupational Inhalation and Dermal Risk by TSCA COU

2054

Acute non-cancer MOE risk estimates based on peak occupational exposure estimates (15-minute) with lower

2055

MOE values indicating greater risks. For COUs with multiple OESs or estimation approaches, the estimate with

2056

the highest high-end value was illustrated.

2057

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 84 of 151

2058

2059

Figure 4-2. Chronic, Non-cancer Occupational Inhalation Risk by TSCA COU

2060

Non-cancer MOE risk estimates based on occupational exposure with lower MOE values indicating greater risks.

2061

For COUs with multiple OESs or estimation approaches, the scenario with the highest central tendency value was

2062

illustrated.

2063

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 85 of 151

2064

Figure 4-3. Chronic Cancer Occupational Inhalation Risk by TSCA COU

2065

Cancer risk estimates based on occupational exposure with higher values indicating greater risks. For COUs with

2066

multiple OESs or estimation approaches, the scenario with the highest central tendency value was illustrated.

2067

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 86 of 151

4.2.1.2 Overall Confidence in Worker Inhalation Risks

2068

Overall confidence in risk estimates for workers via inhalation exposure varies per COU, depending on

2069

the confidence in the hazard and the exposure assessment for each OES as provided in the Draft

2070

Occupational Exposure Assessment for Formaldehyde (U.S. EPA, 2024k).

2071

2072

EPA’s occupational exposure assessment is supported by a large body of workplace monitoring data

2073

specific to the exposure scenarios assessed. A limitation of the monitoring data is the uncertainty in the

2074

representativeness of the data. Some monitoring data was limited in additional contextual information

2075

such as site identification, worker activities and process conditions, such that EPA used other

2076

information to assign to the respective exposure scenario. For scenarios based on limited monitoring

2077

data, the assessed exposure levels are less likely to be representative of worker exposure across the

2078

entire job category or industry. For many exposure scenarios, EPA incorporates OSHA CEHD data.

2079

This data source does not provide job titles or worker activities associated with the sample. As the

2080

OSHA CEHD data were apportioned to OESs based on their NAICS code, there is an uncertainty in the

2081

representativeness of the mapped OSHA CEHD data for the corresponding exposure scenario.

2082

2083

The effects of these uncertainties on the occupational exposure assessment are unknown, as the

2084

uncertainties may result in either overestimation or underestimation of exposures depending on the

2085

actual distribution of formaldehyde air concentrations and the variability of work practices among

2086

different sites. In some scenarios where monitoring data were available, EPA did not find sufficient data

2087

to determine complete statistical distributions. Ideally, EPA will present 50th and 95th percentiles for

2088

each exposed population. In the absence of percentile data for monitoring, the mean or midpoint of the

2089

range may serve as a substitute for the 50th percentile of the actual distributions. Similarly, the highest

2090

value of a range may serve as a substitute for the 95th percentile of the actual distribution. However,

2091

these substitutes are uncertain. The effects of these substitutes on the occupational exposure assessment

2092

are unknown, as the substitutes may result in either overestimation or underestimation of exposures

2093

depending on the actual distribution. Although the weight of scientific evidence varies, EPA has

2094

concluded that the underlying data still provide plausible estimates of exposures for all OESs.

2095

2096

EPA has medium confidence in the acute inhalation POD. It is based on evidence in healthy adults in

2097

controlled exposures. Generally, EPA has medium confidence in the exposure estimates for peak

2098

exposures, but it varies from low to high across the OESs assessed. For most exposure scenarios, EPA

2099

estimated peak exposures using 15-minute workplace monitoring data from the OSHA CEHD database.

2100

However, in some cases, EPA may not have information on the worker activities sampled and whether

2101

these activities would be expected to result in peak levels of formaldehyde. For many scenarios, there is

2102

a high level of non-detects integrated within exposure estimates, which can bias the exposure estimate.

2103

Generally, the limit of detection for the 15-minute samples were higher than the calculated occupational

2104

exposure value for acute effects (see Appendix E.1). For example, acute risks are greatest for the below

2105

COUs, in which EPA has an overall medium confidence in the individual risk estimates:

2106

• Commercial use – chemical substances in automotive and fuel products – automotive care

2107

products; lubricants and greases; fuels and related products: EPA has medium confidence in

2108

the risk estimates for this COU. Three occupational exposure scenarios are estimated for this

2109

COU, the exposure scenario with the highest central tendency exposure estimate was selected for

2110

risk characterization of this condition of use. The automotive care products OES was modeled

2111

for the worker activity of applying a detailing product containing formaldehyde. The scenario

2112

was modeled using two approaches: an approach that model complete evaporation of the

2113

expected formaldehyde contained in the detailing product during application, and an approach

2114

using measured VOC data. To account for variability, EPA performed 100,000 Monte Carlo

2115

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 87 of 151

iterations where parameters were varied based on industry defaults such as number of cars

2116

detailed per site, amount of product used, and formaldehyde specific information, concentration

2117

of formaldehyde in the product. EPA calculated vapor generation using the chemical properties

2118

of formalin as well as reported VOC emissions in the automotive detailing industry. A limitation

2119

of this modeled estimate is that it does not account for if any engineering controls are used

2120

during application.

2121

• Manufacturing-manufacturing: EPA has medium confidence in the risk estimates for this

2122

COU. Acute inhalation risk estimates were derived using 15 personal breathing zone sample data

2123

collected at two U.S. formaldehyde manufacturing facilities in 1992 and one U.S. formaldehyde

2124

manufacturing facility in 2020. Due to a limited amount of recent monitoring data, there is some

2125

uncertainty in the representativeness of the estimates at current manufacturing facilities.

2126

For chronic inhalation risks, EPA has medium confidence in the cancer inhalation unit risk underlying

2127

these risk estimates and high confidence in the chronic, non-cancer hazard POD. The chronic, non-

2128

cancer hazard POD is supported by a robust database of evidence in humans and animals that

2129

demonstrates concordance in effect levels across multiple endpoints and it includes evidence in children

2130

with asthma and other sensitive groups.

2131

2132

Generally, EPA has medium confidence in the exposure estimates for full-shift exposures but confidence

2133

for individual scenarios varies from low to high across the OESs assessed. For most exposure scenarios,

2134

EPA estimated full-shift exposures by integrating discrete data identified from peer-reviewed literature

2135

and other sources. As discussed earlier, OSHA CEHD does not provide all of the meta-data associated

2136

with the sampled data. For estimation of full-shift exposures, EPA establish a cut-off total sampling

2137

duration of 5.5 hours to reduce uncertainties by using data most expected to represent full-shift

2138

exposures. EPA then calculated an 8-hour TWA assuming that unsampled time was zero. This approach

2139

may lead to underestimation of full-shift exposures if workers were still exposed to formaldehyde for the

2140

unsampled time. A sensitivity analysis on these assumptions were included in the Draft Occupational

2141

Exposure Assessment for Formaldehyde (U.S. EPA, 2024k).

2142

2143

For calculation of the ADC and LADC, EPA assumes that workers are exposed for 250 days per year for

2144

chronic and 22 days per month for sub-chronic risk estimates across all scenarios. For LADC, the

2145

assumption of worker tenure is important, in which EPA uses 31 years for central tendency risk

2146

estimates and 40 years for high-end risk estimates. These parameters may vary by individual workers. A

2147

principal limitation of the ADC and LADC used is that these exposure estimates assume no exposure to

2148

formaldehyde outside of the workplaces. In Section 4.3, EPA considers how aggregate exposures to

2149

formaldehyde from multiple sources, across multiple routes, or across pathways may increase the overall

2150

risk for some people.

2151

2152

Although the weight of scientific evidence varies, EPA has concluded that the underlying data still

2153

provide plausible estimates of exposures for all OESs. As examples, chronic risks are greatest for the

2154

below COUs, in which EPA has an overall medium confidence in the risk estimates:

2155

• Commercial use – chemical substances in automotive and fuel products – automotive care

2156

products; lubricants and greases; fuels and related products: EPA has medium confidence in

2157

the risk estimates for this COU. Three occupational exposure scenarios are estimated for this

2158

COU, the exposure scenario with the highest central tendency exposure estimate was selected for

2159

risk characterization of this condition of use. The automotive care products OES was modeled

2160

for the worker activity of applying a detailing product containing formaldehyde. The model

2161

assumes that as the detailing product containing formaldehyde is applied, that the formaldehyde

2162

evaporates during application. To account for variability, EPA performed 100,000 Monte Carlo

2163

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 88 of 151

iterations where parameters were varied based on industry defaults such as number of cars

2164

detailed per site, amount of product used, and formaldehyde specific information, concentration

2165

of formaldehyde in the product. A limitation of this modeled estimate is that it does not account

2166

for if any engineering controls are used during application. EPA calculated vapor generation both

2167

using the chemical properties of formalin as well as reported VOC emissions in a similar

2168

industry.

2169

• Processing – processing as a reactant (COU Group): EPA has medium to high confidence in

2170

the risk estimates for this COU. The underlying occupational exposure scenario covers, in

2171

general, processes that use formaldehyde as a reactant for a variety of downstream products. This

2172

scenario integrates data from a variety of sources (e.g., industry submissions, OSHA CEHD

2173

data) for a total of 192 8-hr TWA samples. Limitations within the monitoring data is a lack of

2174

additional details on worker activities for the individual samples. There is some uncertainty on

2175

the representativeness of the 50th and 95th percentiles towards the true distribution for the

2176

exposed population for this scenario.

2177

4.2.1.3 Risk Estimates for Dermal Exposures

2178

Acute non-cancer risk estimates for dermal exposure range from 3.24×10

−3

to 18 (benchmark MOE of

2179

10) for central tendency exposures and high-end exposures. Risk estimates are greatest for TSCA COUs:

2180

Commercial use – chemical substances in automotive and fuel products – automotive care products;

2181

lubricants and greases; fuels and related products; and TSCA COUs: Processing – incorporation into an

2182

article – paint additives and coating additives not described by other categories in transportation

2183

equipment manufacturing (including aerospace); Industrial use – paints and coatings; adhesives and

2184

sealants; lubricants; commercial use – chemical substances in construction, paint, electrical, and metal

2185

products – adhesives and sealants; paint and coatings. Both OESs assumed an immersive dermal loading

2186

on the skin during the exposure scenario.

2187

2188

Dermal risk estimates were not provided for Distribution in commerce and commercial use – packaging,

2189

paper, and hobby products COUs. These COUs involve the handling of solid articles with low

2190

concentrations of formaldehyde in which the dermal modeling approaches were not suitable. EPA

2191

expects the primary concern for these products is inhalation exposures from formaldehyde off-gassing.

2192

4.2.1.4 Overall Confidence in Worker Dermal Risks

2193

Overall confidence in risk estimates via dermal exposure is medium. As described in Section 3.2, overall

2194

confidence in the dermal hazard value is medium. As described in Section 2.5.1, overall confidence in

2195

dermal occupational exposures is medium based on a moderate weight of scientific evidence for all

2196

scenarios assessed. All scenarios used a modified version of the EPA Dermal Exposure to Volatile

2197

Liquids Model, which reduced to two parameters: an activity-based dermal loading and a maximum

2198

weight concentration of formaldehyde in the formulations handled. For many scenarios, maximum

2199

concentration information from sources such as the 2020 CDR (U.S. EPA, 2020b) have overall data

2200

quality determinations of either high or medium from EPA’s systematic review process. Some scenarios

2201

lacked sufficient information on the maximum concentrations expected and industry-specific or

2202

surrogate scenarios were used to inform calculations. There is some uncertainty on the range of

2203

concentrations of formaldehyde within certain processes and products whose impact is unknown and

2204

may either result in an overestimation or underestimation of exposures.

2205

Risk Estimates for Consumers

2206

EPA estimated cancer and non-cancer risks for exposure to formaldehyde resulting from exposure to

2207

formaldehyde in consumer products. For this analysis, EPA relied on the consumer exposure estimates

2208

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 89 of 151

modeled in the Draft Consumer Exposure Assessment for Formaldehyde (U.S. EPA, 2024d) and

2209

summarized in Section 2.2.

2210

4.2.2.1 Risk Estimates for Inhalation Exposure to Formaldehyde in Consumer

2211

Products

2212

EPA estimated cancer and non-cancer risks to consumers and bystanders from inhalation of

2213

formaldehyde in consumer products.

2214

2215

Acute inhalation risk estimates range from 4.65×10

−4

to 1.31 (Figure 4-4). These acute risk estimates are

2216

calculated using high-end air concentrations modeled for a 15-minute period based a set of high-end

2217

model input assumptions and TSCA COU-specific assumptions about exposure frequency and duration.

2218

Acute risk estimates below 1 indicate that exposure is greater than the hazard point of departure

2219

identified for 15-minute peak exposures based on sensory irritation reported in controlled human

2220

exposure studies in healthy adult volunteers.

2221

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 90 of 151

2222

Figure 4-4. Peak 15-Minute Inhalation Risk by COUs in Consumer Products

2223

Acute non-cancer risk estimates are based on high-end consumer and bystander exposure estimates. Acute non-cancer MOEs are based on modeled air

2224

exposure estimates and are interpreted relative to a benchmark MOE of 10. Lower MOE values indicate greater risks. For some products, air

2225

concentrations were modeled for near-field and far-field (generally describing differences in exposure within the same room) while for other products

2226

concentrations were modeled for zones 1 and 2 (generally describing different rooms). Risks from near-field and zone 1 exposures generally represent

2227

risks from direct exposures to consumer users while far-field and zone 2 tend to represent risks to consumer bystanders. For instance, an individual

2228

applying floor coverings: Varnishes and floor finishes in a living room can be described as a consumer of that product in zone 1 or near-field of the

2229

application area. On the other hand, while the product is being applied there may be someone else either also in the room of use and assumed to be away

2230

from the immediate application area (or in the far-field), or in a completely different room from where the product is being applied (also known as zone 2).

2231

The x-axis presents the 15-minute peak inhalation non-cancer concentration, and the y-axis presents the modeled TSCA COUs.

2232

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 91 of 151

Chronic non-cancer risk estimates for consumers based on modeled chronic inhalation exposures range

2233

from 5.70×10

−1

to 7.64, with lower values indicating greater risks (Figure 4-5). Non-cancer risk

2234

estimates below 1 indicate that exposure is greater than the hazard point of departure based on

2235

respiratory effects in sensitive groups, including children. Chronic ADAF-adjusted lifetime cancer risk

2236

estimates based on modeled chronic inhalation range from 2.36×10

−11

to 4.82×10

−4

(Figure 4-6), with

2237

larger numbers indicating increasing risk. The risk estimates for chronic exposures presented here are

2238

based on central tendency air concentrations modeled for a set of mid-range model input assumptions

2239

and TSCA COU-specific assumptions about exposure frequency and duration. Risk estimates presented

2240

here represent risks to consumers who frequently use products containing formaldehyde and are based

2241

on the consumer activity and use patterns described in the Draft Consumer Exposure Assessment for

2242

Formaldehyde (U.S. EPA, 2024d). For example, cancer risk estimates for the arts, crafts, and hobby

2243

material COU presented here are not representative of all arts and crafts products. They are based on an

2244

assumption of exposure to a specific set of products that contain 0.1 percent formaldehyde used an

2245

average of 15 minutes/day, 300 days each year, over a period of 57 years which are standard CEM

2246

temporal inputs primarily based upon the 1987 Westat survey of consumer activities and use patterns

2247

(U.S. EPA, 2021a, 2019; Westat, 1987).

2248

2249

2250

Figure 4-5. Chronic Non-cancer Inhalation Risks for Consumer Products by COU

2251

Chronic risk estimates are based on consumer and bystander exposure estimates that rely on central tendency

2252

assumptions about product use duration and frequency. Non-cancer MOEs are based on modeled air exposure

2253

estimates and are interpreted relative to a benchmark MOE of 3. Lower MOE values indicate greater risks. The x-

2254

axis presents risk estimates for chronic inhalation exposure estimates, and the y-axis presents the modeled TSCA

2255

COUs.

2256

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 92 of 151

2257

Figure 4-6. ADAF-Adjusted Chronic Inhalation Cancer Risk by COUs in Consumer Products

2258

ADAF-adjusted lifetime cancer risk estimates are based on consumer and bystander central tendency exposure

2259

estimates. Higher cancer risk estimates indicate greater risk. The x-axis presents the ADAF-adjusted lifetime

2260

cancer risk and the y-axis presents the modeled TSCA COUs.

2261

2262

Overall confidence in inhalation risk estimates for consumer products is medium for chronic non-cancer

2263

risks and medium for cancer risk and acute non-cancer risk. As described in Section 3.2.1.1 of the

2264

Consumer Exposure Module, the overall confidence in monitoring data used in the indoor air assessment

2265

is high due to reliance on 41 high quality formaldehyde air exposure studies relevant to TSCA COUs,

2266

and CEM modeling assumptions and inputs, which have been peer reviewed and used in previous

2267

existing chemical risk evaluations. While EPA relied on available survey data on product use patterns,

2268

there is uncertainty around the applicability of the generic survey data for current use patterns for

2269

specific product types. For example, for some inputs relied on the use and activity patterns reported in

2270

the Westat survey from 1987 (Westat, 1987). Although this is a robust dataset it may not be reflective of

2271

current use patterns for the specific product types assessed. As described in Section 3.2, overall

2272

confidence in the chronic, non-cancer hazard POD is high because it is supported by a robust database of

2273

evidence in humans and animals that demonstrates concordance in effect levels across multiple

2274

endpoints and it includes evidence in children with asthma and other sensitive groups. Overall

2275

confidence in the inhalation unit risk for formaldehyde is medium. The cancer risk estimates presented

2276

here do not include risks for some of the tumor sites. While the draft IRIS assessment concluded that the

2277

evidence demonstrates that formaldehyde inhalation causes myeloid leukemia and sinonasal cancer in

2278

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 93 of 151

humans, EPA was not able to quantify those risks with confidence. The draft IRIS assessment estimated

2279

that the IUR used to estimate lifetime cancer risks may underestimate total cancer risk by as much as 4-

2280

fold. EPA has medium confidence in the acute inhalation POD based on evidence in healthy adult

2281

volunteers in controlled exposure conditions.

2282

4.2.2.2 Risk Estimates for Dermal Exposure to Formaldehyde in Consumer Products

2283

EPA estimated non-cancer risks for acute dermal exposure to formaldehyde in consumer products.

2284

2285

Dermal risk estimates were calculated based on low, central tendency and high-end exposure estimates.

2286

The estimated dermal risks based on high-end exposures range from 3.24×10

−3

to 9.71 and are presented

2287

in Figure 4-7. Risk estimates below 1 indicate that exposures are above the POD based on skin

2288

sensitization responses observed in adults. There is uncertainty surrounding the assumption of occlusion

2289

or immersion of hands using liquid or spray consumer products, which may overestimate exposures and

2290

risks for some consumer exposure scenarios.

2291

2292

2293

Figure 4-7. Acute Dermal Loading Risk by High-End Exposure Scenarios in Consumer Products

2294

Dermal non-cancer MOE risk estimates are based on consumer exposure estimates and are interpreted relative to a

2295

benchmark MOE of 10. Lower MOE values indicate greater risks. The x-axis presents the acute dermal loading

2296

MOE, and the y-axis presents the modeled scenarios written as TSCA COU followed by relevant exposure

2297

scenario in parentheses.

2298

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 94 of 151

Overall confidence in risk estimates for dermal exposure is medium. As described in Section 3.2.1.1 of

2299

the Consumer Exposure Module, the overall confidence in monitoring data used in the indoor air

2300

assessment is medium due to no formaldehyde dermal exposure studies identified through systematic

2301

review; though other highly rated supplemental studies were used to identify loading of formaldehyde to

2302

skin (U.S. EPA, 2019; Delmaar et al., 2013; IPCS, 2002; ATSDR, 1999) and product specific modeling

2303

assumptions and weight fractions identified via safety data sheets reviewed and used in previous existing

2304

chemical risk evaluations. As described in Section 3.2, overall confidence in the dermal hazard value is

2305

medium.

2306

Risk Estimates for Indoor Air

2307

EPA estimated cancer and non-cancer risks for exposure to formaldehyde in indoor air. For this analysis,

2308

EPA considered available indoor air monitoring data as well as air concentrations modeled based on

2309

specific TSCA COUs, as described in the Draft Indoor Air Assessment for Formaldehyde (U.S. EPA,

2310

2024j). Monitoring data provide an indication of aggregate exposure and risks in a range of indoor

2311

environments while modeled air concentrations can provide information about the contributions of

2312

specific TSCA COUs to indoor air concentrations.

2313

4.2.3.1 Risk Estimates Based on Indoor Air Monitoring Data

2314

Monitoring data provide information about actual concentrations of total formaldehyde in indoor air, but

2315

the data reflect aggregate concentrations from all TSCA and other sources present. Monitoring data are

2316

therefore a good indication of aggregate formaldehyde exposures and risks in a range of indoor

2317

environments, but do not provide information about the relative contributions of each source.

2318

2319

EPA estimated cancer and non-cancer risks based on levels of formaldehyde detected in indoor air in

2320

monitoring studies representing a range of indoor air environments. The American Healthy Home

2321

Survey II is a survey published in 2021 that is representative of residential indoor air conditions across a

2322

wide range of American households (QuanTech, 2021). It is the most current nationally representative

2323

survey of formaldehyde in indoor air in American homes and is likely the best representation of the

2324

current range of aggregate exposures and risks from all sources of formaldehyde in indoor air. Other

2325

monitoring datasets considered in this analysis generally target indoor environments that typically have

2326

higher formaldehyde concentrations, such as trailers and mobile homes. Available indoor air monitoring

2327

datasets likely do not represent current conditions in indoor air following Title VI regulation of wood

2328

products. Figure 4-8 summarizes ADAF-adjusted lifetime cancer risk estimates based on indoor air

2329

monitoring data, relying on the assumption that these monitored concentrations could represent average

2330

exposures in indoor air and that exposure to these concentrations may be experienced continuously over

2331

a 78-year lifetime. This may be a conservative assumption for high end indoor air exposures, as

2332

concentrations in a particular home change over time and people typically live in multiple homes over

2333

the course of their lives.

2334

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 95 of 151

2335

Figure 4-8. ADAF-Adjusted Lifetime Cancer Inhalation Risk by Indoor Air Monitoring Data

2336

Source

2337

2338

Cancer risk estimates are based on air concentrations reported in monitoring data and rely on the

2339

assumption that individuals may be consistently exposed to these concentrations over a 78-year lifetime.

2340

Higher cancer risk estimates indicate greater risk. Air monitoring data sources listed on the y-axis are

2341

described in more detail in the Draft Indoor Air Assessment for Formaldehyde (U.S. EPA, 2024j).

2342

2343

Among all residence types and commercial environments, lifetime cancer risk estimates based on indoor

2344

air monitoring data ranged from 2.74×10

−6

to 9.46×10

−3

. These ranges of risk estimates correspond to

2345

measured minimum concentrations of 2.18×10

−4

ppm by the American Healthy Home Survey II

2346

(QuanTech, 2021), and a measured maximum concentration of 7.53×10

−1

ppm from a study of four

2347

FEMA camper trailers (LBNL, 2008), respectively. Chronic non-cancer risk estimates based on the

2348

same indoor air monitoring data range from 77.8 to 0.02, with lower values indicating greater risk.

2349

4.2.3.2 Risk Estimates Based on Indoor Air Modeling for Specific TSCA COUs

2350

Indoor air concentrations modeled for specific COUs provide an indication of the contributions of

2351

individual COUs to formaldehyde exposure and risk. EPA estimated chronic non-cancer risks based on

2352

formaldehyde concentrations modeled based on long-term emissions associated with specific COUs, as

2353

described in Section 2.3. The modeled air concentrations used as the basis for chronic risk estimates for

2354

indoor air were designed to estimate concentrations at the central tendency. As described in the Draft

2355

Indoor Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024j), there is substantial uncertainty

2356

related to the degree of dissipation of formaldehyde over time and how exposures from specific products

2357

change over the course several years. For this reason, EPA has low confidence in exposure estimates

2358

modeled over longer than a year for specific TSCA COUs contributing to formaldehyde in indoor air.

2359

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 96 of 151

EPA therefore did not calculate cancer risk based on chronic indoor air exposures resulting from specific

2360

TSCA COUs.

2361

2362

Non-cancer risk estimates based on indoor air concentrations modeled for specific COUs range from

2363

0.05 to 4. Risk estimates below 1 indicate that exposure is greater than the hazard point of departure

2364

based on respiratory effects in sensitive groups, including children. Figure 4-9 summarizes chronic non-

2365

cancer risk estimates based on modeled average indoor air concentrations estimated to result from

2366

specific TSCA COUs over the course of the first year of product use. These risk estimates account for

2367

dissipation that occurs over time due to the depletion of formaldehyde from the article and air exchange

2368

but do not account for the half-life of formaldehyde.

2369

2370

2371

Figure 4-9. Chronic Non-cancer Inhalation Risk Based on Modeled Air Concentrations for

2372

Specific TSCA COUs

2373

Chronic non-cancer risk estimates are based on indoor air exposure estimates. Lower MOEs indicate greater risk.

2374

The y-axis presents the modeled scenarios written as TSCA COU followed by relevant exposure scenario.

2375

2376

Overall confidence in risk estimates by individual TSCA COU modeling is medium. In general, EPA

2377

has medium confidence in CEM’s ability to assess formaldehyde exposures in indoor air and the

2378

supporting monitoring data. The inability to account for half-life in the model decreases confidence in

2379

the exposure estimates. It is unclear whether the modeling results are reflective of most indoor air home

2380

environments in American residences. EPA has medium confidence in the applicability of the modeling

2381

results used to assess indoor air exposures to formaldehyde. As described in Section 3.2.1.1 of the Draft

2382

Indoor Air Exposure Assessment Module, the overall confidence in modeling used in the indoor air

2383

assessment is high due to medium quality studies used to incorporate TSCA COU-specific emission

2384

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 97 of 151

rates and due to the use of a high quality CEM modeling inputs and formulas used to generate TSCA

2385

COU-specific indoor air concentrations.

2386

2387

Monitoring data reflect total concentrations from a wider range of sources and are therefore not directly

2388

comparable to modeled estimates. However, in general, modeled and monitored indoor air formaldehyde

2389

concentrations are within the same order of magnitude that increases the confidence in the modeled

2390

formaldehyde indoor air exposures underlying these risk estimates.

2391

2392

As described in Section 3.2, overall confidence in the chronic non-cancer hazard POD is high. It is

2393

supported by a robust database of evidence in humans and animals that demonstrates concordance in

2394

effect levels across multiple endpoints and it includes evidence in children with asthma and other

2395

sensitive groups.

2396

4.2.3.3 Integration of Modeling and Monitoring Information and Consideration of

2397

Aggregate Risk

2398

Risk estimates based on modeled air concentrations provide information about the contribution of

2399

specific COUs to exposures and risks from formaldehyde in indoor air. However, given the ubiquity of

2400

formaldehyde in indoor environments, risks from individual sources rarely occur in isolation. EPA has

2401

therefore also considered monitoring data as an indication of aggregate exposure and risks from all

2402

sources contributing to formaldehyde in indoor air.

2403

2404

While monitoring data does not distinguish between risk contributions from TSCA and other sources, it

2405

offers a way to interpret risks from individual COUs in the context of aggregate risks from all co-

2406

occurring sources.

2407

2408

As previously noted, the AHHS II is the most current nationally representative survey of formaldehyde

2409

in indoor air in American homes. Therefore, among all monitoring sources, it is likely the most

2410

appropriate source for the estimation of aggregate risks in American residential indoor air across all

2411

households, including old and new homes. Using the maximum estimated monitoring indoor air estimate

2412

for formaldehyde in AHHS II (including contributions from both TSCA and other sources), it may be

2413

assumed that indoor air aggregate non-cancer MOEs are as low as 1.681×10

−1

and cancer MOEs are as

2414

high as 1.271×10

−3

in typical U.S. The same can be inferred from mobile home, classroom, and other

2415

monitoring indoor air risk estimates.

2416

Risk Estimates for Ambient Air

2417

EPA evaluated cancer risks resulting from human exposure to formaldehyde via the ambient air pathway

2418

using previously peer-reviewed methodologies along with multiple lines of evidence including multiple

2419

release estimates from two separate databases (TRI and NEI), several peer-reviewed models (IIOAC,

2420

HEM, AirToxScreen), and monitoring data (AMTIC) from EPA’s ambient monitoring network. When

2421

looking at direct analysis of formaldehyde release data from TRI using IIOAC to represent a more

2422

localized exposure, 26 of 29 TSCA COUs evaluated have risk estimates greater than 11×10

−6

, and 19

2423

COUs have risk estimates greater than 11×10

−5

. Additionally, 21 of the 29 TSCA COUs have risk

2424

estimates greater than relative risk estimates for biogenic sources. As expected, modeled concentrations

2425

using IIOAC fall within the lower range of monitoring data from AMTIC (although not amortized as

2426

annual averages) since AMTIC represents a total formaldehyde concentration from all sources rather

2427

than localized impacts near industrial facilities releasing formaldehyde to the ambient air and associated

2428

with COUs evaluated with IIOAC. Nonetheless, cancer risk estimates based on monitoring data from

2429

AMTIC range from 7.11×10

−8

to 6.1×10

−4

. Figure 4-10 shows the ADAF-adjusted cancer risk estimates

2430

for all AMTIC monitoring data, IIOAC modeled data, and AirToxScreen modeled data, based on the

2431

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 98 of 151

assumption that these concentrations reflect average exposures that occur continuously over a 78-year

2432

lifetime.

2433

2434

Figure 4-10. ADAF-Adjusted Cancer Risk for Monitoring and Modeling Ambient Air Data

2435

2436

EPA recognizes that the different model estimates are not directly comparable. For example, the IIOAC

2437

results represent a risk estimate between 100 to 1,000 m from the release point. In contrast,

2438

AirToxScreen concentrations represent risk estimates at the census tract scale; only point source data

2439

may represent some releases of formaldehyde from TSCA COUs. Given the spatial scale difference, it is

2440

expected that AirToxScreen results could underestimate concentrations on a smaller scale (i.e., near

2441

facilities) or have lower concentration estimates than IIOAC and this difference can be seen in Figure

2442

2-10. Additionally, only point source data within AirToxScreen may represent a broader set of

2443

formaldehyde releases that include releases associated with TSCA COUs.

2444

4.2.4.1 Risk Estimates Based on Ambient Air Monitoring

2445

There is abundant monitoring data on formaldehyde in ambient air. As described in Section 2.4.1,

2446

monitoring data from EPA’s AMTIC (U.S. EPA, 2022a) include a range of air monitoring data collected

2447

across the country under a range of experimental designs across heterogenous environments. EPA

2448

considers the available monitoring data for formaldehyde to reflect the range of aggregate formaldehyde

2449

concentrations under a range of outdoor environments from both TSCA and other sources of

2450

formaldehyde.

2451

2452

EPA calculated chronic cancer risks based on air concentrations reported in AMTIC, relying on the

2453

assumption that monitored concentrations could represent chronic exposure (as shown at the top of

2454

Figure 4-10). However, because some monitoring efforts included in the dataset capture a snapshot of

2455

air concentrations at a single timepoint, there is uncertainty around the extent to which the available

2456

monitoring data are an accurate representation of long-term chronic exposures.

2457

2458

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 99 of 151

Given the ubiquity of formaldehyde and the diversity of sources, monitoring data does not provide clear

2459

information on the contributions of specific TSCA or other sources of formaldehyde. Risk estimates

2460

based on the available monitoring data provide an indication of the aggregate risk from all sources

2461

contributing to ambient air concentrations of formaldehyde, which may be present in the real world and

2462

provide context for risks from individual TSCA COUs.

2463

4.2.4.2 Risk Estimates Based on Modeled Concentrations near Releasing Facilities

2464

EPA estimated risks associated with acute and chronic non-cancer exposure to formaldehyde in the

2465

ambient air. EPA utilized the 95th percentile release value reported to TRI by Industry Sector (mapped

2466

to respective COUs) and the 95th percentile modeled annual-averaged air concentrations from the

2467

IIOAC output file at 100 to 1,000 m from the release point as described in the Draft Ambient Air

2468

Exposure Assessment for Formaldehyde (U.S. EPA, 2024a) to derive risk estimates. All derived risk

2469

estimates for acute and chronic non-cancer effects were above relative MOE benchmarks. Therefore,

2470

while all risk estimates are included in the “Draft IIOAC Assessment Results and Risk Calcs

2471

Supplement A for Ambient Air,” EPA focuses on cancer risk estimates as described below for purposes

2472

of risk characterization in this draft human health risk assessment.

2473

2474

EPA estimated cancer risks associated with continuous chronic exposure to formaldehyde in the ambient

2475

air over a 78-year lifetime. EPA utilized the 95th percentile release value reported to TRI by Industry

2476

Sector (mapped to respective TSCA COUs) and the 95th percentile modeled annual-averaged air

2477

concentrations from the IIOAC output file at a distance of 100 to 1,000 m from the release facility

2478

described in the Draft Ambient Air Exposure Assessment for Formaldehyde (U.S. EPA, 2024a) and in

2479

Section 2.4.2.1, to derive cancer risk estimates. Risk estimates are presented by TSCA COU in Figure

2480

4-11. As described in Section 4.1.2, higher cancer risk estimates indicate higher risks.

2481

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 100 of 151

2482

Figure 4-11. Risk Estimates by TSCA COU for the 95th Percentile Release Scenario and 95th Percentile Modeled Concentration

2483

between 100 and 1,000 m from Industrial Facilities Releasing Formaldehyde to the Ambient Air

2484

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 101 of 151

Across all TSCA COUs, cancer risk estimates ranged from 1.1×10

−9

to 5.9×10

−5

. The three highest

2485

cancer risk estimates are 5.9×10

−5

, 4.5×10

−5

, and 3.4×10

−5

. These three cancer risk estimates represent

2486

three industry sectors and seven TSCA COUs.

2487

2488

The three industry sectors with the highest cancer risk estimates associated with TSCA COUs are:

2489

• Non-metallic mineral product manufacturing (5.9×10

−5

);

2490

• Textiles, apparel, and leather product manufacturing (4.5 ×10

−5

); and

2491

• Transportation equipment manufacturing (3.4 ×10

−5

2492

Together, these three industry sectors are associated with seven formaldehyde TSCA COUs (i.e.,

2493

individual industry sector results are used to represent multiple formaldehyde TSCA COUs as shown

2494

below). Those COUs are:

2495

• Processing – incorporation into an article-adhesives and sealant chemicals (5.9×10−5);

2496

• Processing as a reactant-intermediate (5.9×10

−5

);

2497

• Processing – incorporation into a formulation, mixture, or reaction product-intermediate

2498

(5.9×10

−5

);

2499

• Processing – incorporation into article-finishing agent (4.5×10

−5

µg/m

);

2500

• Processing – incorporation into a formulation, mixture, or reaction product-bleaching agents

2501

(4.5×10

−5

);

2502

• Processing-incorporation into an article-paint additives and coating additives (3.4×10

−5

); and

2503

• Industrial use-chemical substances in industrial products-paints and coatings; adhesives and

2504

sealants, lubricants (3.4×10

−5

2505

In total, 19 of the 29 TSCA COUs (65.5%) have cancer risk estimates within the same order of

2506

magnitude greater than 1×10

−5

. An additional seven TSCA COUs have cancer risk estimates within the

2507

same order of magnitude greater than 1×10

−6

and less than 1×10

−5

. Two COUs have cancer risk

2508

estimates within the same order of magnitude greater than 1×10

−7

and less than 1×10

−6

, and one TSCA

2509

COU has a cancer risk estimate in the 1×10

−9

range.

2510

2511

Recognizing the ubiquity of formaldehyde in ambient air occurs from multiple sources including other

2512

sources like biogenic/natural sources and secondary formation, EPA compared the calculated risk

2513

estimates for modeled concentrations from IIOAC to the calculated risk estimate for the 95th percentile

2514

concentration of attributable to biogenic sources. Across all 29 TSCA COUs evaluated, 21 TSCA COUs

2515

have risk estimates greater than the risk estimate for biogenic sources (2.85×10

−6

). Eighteen TSCA

2516

COUs have calculated risk estimates greater than 5 times the calculated risk estimate for biogenic

2517

sources (1.42×10

−5

). Seven TSCA COUs have calculated risk estimates greater than 10 times the

2518

calculated risk estimate for biogenic sources (2.85×10

−5

). Eight TSCA COUs have calculated risk

2519

estimates less than the risk estimate for biogenic sources.

2520

2521

For the industry sector of Oil and Gas Drilling, Extraction, and Support Activities, results were not

2522

available from the TRI program. Although many of the NAICS codes for this industry sector are not

2523

covered by the TRI program, the sites are well represented in the NEI database. This industry sector is

2524

associated with the following formaldehyde TSCA COUs:

2525

• Processing as a reactant-functional fluid;

2526

• Processing – incorporation into a formulation, mixture, or reaction product – processing aids,

2527

specific to petroleum production;

2528

• Processing – incorporation into a formulation, mixture, or reaction product – intermediate; and

2529

• Industrial use – non-incorporative activities – process aid.

2530

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 102 of 151

Upon further review, the emission source information provided in the NEI database indicated that the

2531

majority of emissions within this industry are combustion sources (e.g., reciprocating engines), with a

2532

limited number of emission sources related to storage tanks, amine processes, and unclassified units with

2533

emission sources typically less than 100 kg/year. These releases are lower than the median for the

2534

industry sector, which have cancer risks below the 1×10

−5

. Therefore, EPA did not include the oil and

2535

gas drilling, extraction, and support activities industry sector as the primary emissions are outside of the

2536

scope of this draft risk evaluation.

2537

2538

Overall, these results indicate that while releases, exposures, and associated risk estimates may vary

2539

across industry sectors and TSCA COUs, the results presented in Figure 4-11 are generally

2540

representative of risks to individuals residing near industrial facilities releasing formaldehyde into the

2541

ambient air that are associated with TSCA COUs.

2542

2543

Risks estimates calculated by the HEM model at census blocks were also considered to inform EPA’s

2544

understanding of how modeled results intersected with populated areas and demographic characteristics.

2545

Overall, HEM modeling estimated a total population of 1,023,773 people experiencing a lifetime cancer

2546

risk of at least one in one million. These cancer risk estimates are based solely on formaldehyde

2547

emissions from facilities reporting to TRI, and do represent the aggregation of exposures from multiple

2548

nearby facilities. A full breakdown of estimated population by level of risk estimate with stratification

2549

by demographics is presented in Table 4-2. At higher levels of estimated risk, 6,935 people were

2550

estimated to experience risk greater than 10 in 1 million, and 19 were estimated to experience risk

2551

greater than 100 in 1 million. No estimated risks exceeded 200 in 1 million. Across the entire modeling

2552

domain, which included census blocks within 50 km of any TRI facility reporting formaldehyde

2553

releases, the average risk to the entire population of 232,907,302 people was estimated to be 0.04 in 1

2554

million. This average risk was slightly higher for the African American and Native American

2555

demographics included in the modeling, at an estimate of 0.06 in 1 million. While population counts are

2556

summarized at the census block level, the demographic information is summarized by census block

2557

group, and applied to each block within the block group. In order to avoid double counting, the

2558

“Hispanic or Latino” category is treated as a distinct demographic category for these analyses. A person

2559

is identified as one of five racial/ethnic categories presented below: White, African American, Native

2560

American, Other and Multiracial, or Hispanic/Latino.

2561

2562

Table 4-2: Population Summary for Cancer Risk Estimates Derived from HEM Modeling of TRI

2563

Releases Formaldehyde to Air

2564

Range of Lifetime

Individual Cancer

Risk

Number of People within 50 km of any Facility in Different Ranges for Lifetime

Cancer Risk

Total

Population

White

African

American

Native

American

Other and

Multiracial

Hispanic or

Latino

< 1 in 1 million

232,907,302

140,083,682

30,322,675

881,180

21,243,988

40,375,778

1 to <5 in 1 million

1,023,773

665,609

171,444

7,929

54,384

124,408

5 to <10 in 1

million

40,652

26,742

5,429

542

2,884

5,055

10 to <20 in 1

million

6,935

4,430

1,057

246

1,181

20 to <30 in 1

million

2,692

1,901

388

331

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 103 of 151

Range of Lifetime

Individual Cancer

Risk

Number of People within 50 km of any Facility in Different Ranges for Lifetime

Cancer Risk

Total

Population

White

African

American

Native

American

Other and

Multiracial

Hispanic or

Latino

30 to <40 in 1

million

509

359

40 to <50 in 1

million

555

379

117

50 to <100 in 1

million

338

202

101

100 to <200 in 1

million

≥200 in 1 million

Total population

within model

domain

233,982,775

140,783,315

30,501,287

889,684

21,301,603

40,506,886

Average risk

(chance in 1

million)

0.04

0.06

0.03

2565

Further breakdown of relative population demographics compared to national averages is presented in

2566

Table 4-3. This summary of results shows that among the population with estimated cancer risk modeled

2567

by HEM to be higher than 1 in 1 million, some population groups are disproportionately represented,

2568

which would be indicated by a higher percentage of a population group experiencing elevated risk than

2569

the overall nationwide percentage of the population representing that group. These groups include white,

2570

African American, and Native American demographics, as well as those with income below the poverty

2571

level and those aged over 25 years without a high school diploma.

2572

2573

Table 4-3. Demographic Details of Population with Estimated Cancer Risk Higher than or Equal

2574

to 1 in 1 Million, Compared with National Proportions

2575

Demographic

Nationwide

Population with Cancer Risk Higher than or Equal to 1 in

1 Million (Estimated by HEM Modeling of TRI Releases)

Total Population

329,824,950

1,075,473

Race and ethnicity by percent

White

59.5%

65.1%

African American

12.1%

16.6%

Native American

0.6%

0.8%

Other and Multiracial

8.8%

5.4%

Hispanic or Latino

19.0%

12.2%

Income by percent

Below Poverty Level

12.8%

15.7%

Above Poverty Level

87.2%

84.3%

Below Twice Poverty

Level

30.2%

34.9%

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 104 of 151

Above Twice Poverty

Level

69.8%

65.1%

Education by percent

Over 25 and without a

High School Diploma

11.6%

12.3%

Over 25 and with a

High School Diploma

88.4%

87.7%

Linguistically isolated by percent

Linguistically Isolated

5.2%

2.2%

2576

Overall confidence in risk estimates based on modeled air concentrations is high for non-cancer risk

2577

estimates and medium for cancer risk estimates. As described in Section 2.4.2, overall confidence in

2578

modeling for exposures used to derive risk estimates for ambient air is high because modeling relies

2579

upon direct reported releases from multiple years and databases that received a high-quality rating from

2580

EPA’s systematic review process. Peer-reviewed modeling approaches and methods with IIOAC were

2581

used to estimate concentrations to derive risk estimates at distances from releasing facilities where

2582

individuals typically reside for many years. Use of additional peer-reviewed models (AirToxScreen and

2583

HEM) along with monitoring data (AMTIC) to further contextualize ambient air concentrations of

2584

formaldehyde, which also present a consistent picture of exposures when compared to IIOAC results,

2585

provide added strength and confidence to the risk estimates.

2586

2587

As described in Section 3.2, overall confidence in the acute and chronic, non-cancer hazard POD is high

2588

while overall confidence in the inhalation unit risk for formaldehyde is medium. The cancer risk

2589

estimates presented here do not include risks for some of the tumor sites. While the draft IRIS

2590

assessment concluded that the evidence demonstrates that formaldehyde inhalation causes myeloid

2591

leukemia and sinonasal cancer in humans, EPA was not able to quantify those risks with confidence.

2592

The draft IRIS assessment estimated that the IUR used to estimate lifetime cancer risks may

2593

underestimate total cancer risk by as much as 4-fold.

2594

4.2.4.3 Integration of Modeling and Monitoring Information

2595

EPA evaluated and characterized exposures and risks to the general population from industrial releases

2596

of formaldehyde to the ambient air using actual reported releases and peer reviewed models to estimate

2597

exposures at select distances from releasing facilities. EPA also evaluated and characterized exposures

2598

and risks to the general population based on ambient monitoring data obtained from AMTIC.

2599

2600

Modeling and monitoring results show comparable exposures and risks to the general population from

2601

formaldehyde in the ambient air. However, direct comparisons between modeled and monitored

2602

concentrations and associated risks should be made with caution because each approach represents

2603

different contributions to the overall exposures and associated risks.

2604

2605

EPA’s modeling approaches use actual reported releases of formaldehyde, required to be reported by

2606

statute to peer-reviewed databases, as direct inputs to peer-reviewed models. The models are then used

2607

to estimate exposures used to derive risk estimates and characterize risks. Because the modeling

2608

approaches use actual reported releases from real facilities, each release can be mapped to a

2609

representative TSCA COU. This allows EPA to estimate exposures, derive risk estimates, and

2610

characterize risks to its TSCA COU as required by statute and is a strength of the modeling approaches

2611

used. However, since some modeling inputs require assumptions that may be conservative in nature and

2612

retain some uncertainty results from modeling may overestimate exposures to the chemical modeled and

2613

thus overestimate risk. While this may be seen as a limitation to the relevance of modeling to estimate

2614

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 105 of 151

exposures and associated risks, the modeling approaches are not overly conservative (based on a series

2615

of sensitivity analyses) and provide a more health protective estimate for use in risk characterization,

2616

risk determination, and regulatory decisions.

2617

2618

In addition to modeled concentrations of formaldehyde in ambient air, EPA relied upon monitoring data

2619

from EPA’s ambient air monitoring network. The monitoring network samples on a regular, and

2620

sometimes continuous, basis concentrations of a variety of chemicals in the ambient air. The monitoring,

2621

sampling, and analysis methods follow EPA reference methods, which have been rigorously peer

2622

reviewed and often promulgated in the Code of Federal Regulations (CFR). Monitored concentrations,

2623

therefore, represent actual measured concentrations of chemicals in the ambient air that contrasts with

2624

modeled concentrations that are estimated based on a series of assumptions and input parameters.

2625

However, ambient monitoring also measures the total concentration of the chemical in the ambient air,

2626

which can be due to multiple sources (TSCA COUs, secondary formation, biogenic formation, and other

2627

sources that cannot readily be mapped to a single TSCA COU). Since monitored concentrations

2628

represent a total concentration of a chemical in ambient air, in a given location, at a given period in time,

2629

monitoring data may be more representative of a total aggregate exposure of the general population to

2630

formaldehyde in the ambient air rather than an independent exposure from a single source over a

2631

continuous exposure period.

2632

4.2.4.4 Overall Confidence in Exposures, Risk Estimates, and Risk Characterizations

2633

for Ambient Air

2634

Confidence in the characterization of exposures for the general population utilized to derive these risk

2635

estimates is high as exposures are based on actual reported releases required by statute to be reported by

2636

industry to peer-reviewed databases. Additionally, peer-reviewed models are used to model ambient air

2637

concentrations at distances from releasing facilities where individuals within the general population

2638

typically reside for many years. Finally, the TRI database undergoes repeatable quality assurance and

2639

quality control reviews and is a high-quality database under EPA’s systematic review process.

2640

2641

For formaldehyde, the potential contribution of combustion sources is an uncertainty and use of the full

2642

facility data complicate singular TSCA COU estimates, such that emissions at one site may include

2643

multiple sources under multiple COUs that include combustion sources and non-combustion sources.

2644

For industrial COUs, EPA has a moderate to robust weight of scientific evidence as the databases have

2645

high data quality scores and are supported by numerous data points. EPA targeted its assessment to

2646

industrial COUs as it expects industrial releases to be the largest proportion of TSCA-related releases.

2647

For commercial COUs, EPA used TRI and NEI results to inform the potential ranges of ambient air risk

2648

estimates in Appendix D. EPA has a moderate weight of scientific evidence for the commercial COUs.

2649

2650

Overall confidence in risk estimates based on air concentrations modeled near release sites is high for

2651

non-cancer estimates and moderate for cancer estimates based on the hazard values. As described in

2652

Section 3.2, overall confidence in the chronic, non-cancer hazard POD is high, while overall confidence

2653

in the inhalation unit risk for formaldehyde is medium. The cancer risk estimates presented here do not

2654

include risks for some of the tumor sites. Although the draft IRIS assessment concluded that the

2655

evidence demonstrates that formaldehyde inhalation causes myeloid leukemia and sinonasal cancer in

2656

humans, EPA was not able to quantify those risks with confidence. The draft IRIS assessment estimated

2657

that the IUR used to estimate lifetime cancer risks may underestimate total cancer risk by as much as 4-

2658

fold.

2659

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 106 of 151

Comparison of Non-cancer Effect Levels and Air Concentrations

2660

Hazard and risk assessments often lack human data on the specific concentrations at which an effect

2661

occurs in people and risk estimates often incorporate a substantial amount of uncertainty. In the case of

2662

formaldehyde, a robust database of epidemiology studies provides information about the air

2663

concentrations of formaldehyde that have been associated with respiratory effects in people and supports

2664

hazard values with minimal uncertainty.

2665

2666

Figure 4-12 indicates that the respiratory effects of formaldehyde in people can occur within the range of

2667

air concentrations reported in monitoring studies. This comparison suggests that chronic exposure to

2668

some of the indoor and outdoor air concentrations captured in available monitoring data are at levels that

2669

may be expected to result in adverse health effects based on available human evidence.

2670

2671

2672

Figure 4-12. Comparison of Non-cancer Health Effect Levels Reported in People and Indoor and

2673

Outdoor Air Concentrations

2674

Indoor air monitoring data summarized here are the American Healthy Homes Survey II data described in Section

2675

2.3.1 and reflect the range of typical indoor air concentrations. Outdoor air monitoring data summarized here are

2676

the AMTIC dataset and include a diverse range of outdoor air monitoring sources. Black shapes indicate air

2677

concentrations at which adverse health effects were reported in epidemiology studies or controlled human

2678

exposure studies (LOAEL or BMC), grey circles and squares indicate concentrations at which no significant

2679

health effects were reported (NOAEL or BMCL), and grey bars indicate the total uncertainty factors identified for

2680

each study. Effect levels (LOAEL, BMC, NOAEL and BMCL) and composite uncertainty factors for each study

2681

are presented as reported in the draft IRIS assessment.

2682

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 107 of 151

Potentially Exposed or Susceptible Subpopulations

2683

EPA considered PESS throughout the exposure and hazard assessments supporting this analysis. Table

2684

4-4 summarizes how PESS were incorporated into the risk evaluation through consideration of increased

2685

exposures and/or increased biological susceptibility. The table also summarizes the remaining sources of

2686

uncertainty related to consideration of PESS. Appendix C provides additional details on PESS

2687

considerations for the formaldehyde risk evaluation.

2688

2689

The available data suggest that some groups or lifestages have greater exposure to formaldehyde. For

2690

example, people exposed to formaldehyde at work, those who frequently use consumer products

2691

containing high concentrations of formaldehyde, people living or working near facilities that emit

2692

formaldehyde, and people living in mobile homes and other indoor environments with high

2693

formaldehyde concentrations are expected to have greater exposures. In this assessment, EPA evaluated

2694

risks anticipated for a range of scenarios under TSCA COUs where exposures are expected to be

2695

greatest. In addition to high exposures associated with COUs, some people will have greater exposure to

2696

formaldehyde through sources that are not being assessed under TSCA. For example, those living near

2697

major roadways, people living in areas with frequent exposure to wildfire smoke, smokers, and people

2698

exposed to second-hand smoke, are expected to have greater exposures to formaldehyde. For these

2699

groups, higher exposures from other sources of formaldehyde may increase susceptibility to additional

2700

exposures from TSCA sources. As described in Section 4.3, EPA assessed risks from several aggregate

2701

exposure scenarios; however, the wide range of possible combinations of aggregate sources are expected

2702

to be highly variable across individuals and are a remaining source of uncertainty.

2703

2704

Some groups or lifestages may be more susceptible to the health effects of formaldehyde exposures. For

2705

example, children have developing respiratory systems and narrower airways that may make them more

2706

susceptible to the respiratory effects of formaldehyde. The chronic inhalation hazard value is derived in

2707

part based on dose-response information in children with asthma and is supported by dose-response

2708

information on lifestage-specific reproductive and developmental effects in humans and animals. The

2709

chronic inhalation hazard value incorporates information on several sensitive groups; therefore, EPA

2710

used a value of 3 for the UF

to account for human variability.

2711

2712

Other factors that may increase susceptibility to formaldehyde include chronic disease, co-exposures,

2713

sex, lifestyle, sociodemographic status, and genetic factors. People with chronic respiratory diseases

2714

(e.g., asthma) may be more susceptible to the respiratory effects of formaldehyde. Co-exposure to other

2715

chemical or non-chemical stressors that increase risk of asthma, reduced pulmonary function,

2716

reproductive and/or developmental toxicity, nasopharyngeal cancer or myeloid leukemia, may increase

2717

susceptibility to the effects of formaldehyde on the same health outcomes. While these factors are not

2718

quantitatively accounted for in the hazard characterization, EPA used values of 3 or 10 for the human

2719

variability UF

to account for increased susceptibility when quantifying risks from exposure to

2720

formaldehyde. The Risk Assessment Forum, in A Review of the Reference Dose and Reference

2721

Concentration Processes (U.S. EPA, 2002), discusses some of the evidence for choosing the default

2722

factor of 10 when data are lacking—including toxicokinetic and toxicodynamic factors as well as greater

2723

susceptibility of children and elderly populations. U.S. EPA (2002), however, did not discuss many of

2724

the factors presented in Appendix CError! Reference source not found.

2725

2726

As described in Section 4.1.2 and in the draft IRIS assessment (U.S. EPA, 2022b), EPA concluded that a

2727

mutagenic mode of action is operative in formaldehyde-induced nasopharyngeal carcinogenicity. EPA

2728

therefore applied ADAFs to lifetime cancer risk estimates to account for increased susceptibility to

2729

nasopharyngeal cancer following inhalation exposure during early life.

2730

2731

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 108 of 151

Table 4-4. Summary of PESS Considerations Incorporated throughout the Analysis and Remaining Sources of Uncertainty

2732

PESS

Category

Subcategory

Increased Exposure from

OtherSources

Increased Exposure from TSCA COUs

Quantitative Data Sources

Lifestage

Embryo/fetus

• EPA did not identify other sources of

increased exposure anticipated for this

lifestage.

• EPA did not identify sources of increased

TSCACOU exposure anticipated for this

lifestage.

• EPA did not quantify exposures

specific to this lifestage.

Pregnant people

• EPA did not identify other sources of

increased exposure anticipated for this

lifestage.

• EPA did not identify sources of increased

TSCA COU exposure anticipated for this

lifestage.

• EPA did not quantify exposures

specific to this lifestage

Children

(infants, toddlers)

• EPA did not identify other sources of

increased exposure anticipated for this

lifestage.

• For air exposures, the impacts of lifestage

differences were not able to be adequately

quantified and so the air concentrations are

used for all lifestages.

• Consumer exposure scenarios include

lifestage-specific exposure factors for

adults, children, and infants (U.S. EPA,

2024d)

• Based on pchem properties and a lack of

studies evaluating potential for

accumulation in human milk following

inhalation, dermal or oral exposures, EPA

did not quantitatively evaluate the human

milk pathway. This is a remaining source

of uncertainty.

• In the consumer exposure assessment, EPA

also considered potential oral exposure

associated with mouthing behaviors in

infants and young children (U.S. EPA,

2024d), however EPA did not have

sufficient information on this exposure

route to quantify risks.

• Lifestage specific consumer exposure

scenarios for infants, children, and

adults are based on information from

U.S. EPA (2005a)and U.S. EPA

(2011).

Older Adults

• EPA did not identify other sources of

increased exposure anticipated for this

lifestage.

• EPA did not identify sources of increased

COU or pathway specific exposure for

this lifestage.

• EPA did not quantify exposures

specific to this lifestage.

Sociodemo-

graphic

factors

Race/Ethnicity

• EPA did not identify specific data on

other sources of increased exposure

associated with race/ethnicity.

• EPA did not identify specific data on

increased COU or pathway specific

exposure associated with race/ethnicity.

• EPA did not quantify exposures

associated with race/ethnicity.

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 129 of 151

Category

Subcategory

Increased Exposure from

OtherSources

Increased Exposure from TSCA COUs

Quantitative Data Sources

Socioeconomic

status

• EPA did not identify specific data on

other sources of increased exposures

associated with socioeconomic status.

• EPA did not identify specific data on

increased COU or pathway specific

exposure associated with socioeconomic

status.

• EPA did not directly quantify

exposures associated with

socioeconomic status.

Unique

Activities

Subsistence Fishing

• EPA did not identify other sources of

increased exposure associated with

subsistence fishing or other exposure

scenarios unique to tribes or other groups.

• EPA did not identify sources of increased

COU or pathway specific exposure for

subsistence fishing or other exposure

pathways unique to tribes or other groups.

• EPA did not quantify exposures

associated with subsistence fishing.

Lifestyle

Smoking

• EPA identified smoking as an additional

other source of exposure to formaldehyde

that may increase aggregate exposure for

smokers and people exposed to second-

hand smoke. To some degree,

formaldehyde exposure from smoking is

indirectly accounted for in some indoor

air monitoring data described in Section

5.2.3.1, but it is not directly quantified.

• EPA did not identify sources of increased

COU or pathway specific exposure for

smoking or other lifestyle factors.

• EPA did not directly quantify

exposures associated with smoking.

Geography

Living in proximity

to sources of

formaldehyde

releases to outdoor

air

• EPA identified living near major

roadways or in areas with frequent

exposure to wildfire smoke as potential

sources of increased exposure to

formaldehyde for some populations. To

some degree, ambient air monitoring data

may indirectly account for some of these

sources but they are not directly

quantified. These other sources of

formaldehyde are a source of uncertainty

that is not directly incorporated into risk

estimates for outdoor air exposures.

• EPA evaluated risks to communities in

proximity to sites where formaldehyde is

released to ambient air (Section 5.2.4). In

the environmental release assessment,

EPA mapped tribal lands in relation to air,

surface water and ground water releases

of formaldehyde to identify potential for

increased exposures for tribes due to

geographic proximity (U.S. EPA, 2024g).

• EPA quantified exposures for

communities in proximity to release

sites using air concentrations modeled

based on releases reported to TRI, as

described in U.S. EPA (2024g) and

Section 5.2.4

• EPA did not directly quantify

exposures associated with living near

roadways or other sources of

formaldehyde in outdoor air.

Other

chemical and

non-chemical

stressors

Built Environment

• EPA identified the built environment

(including building materials and other

products) as source of increased exposure

to formaldehyde associated with other

sources. Indoor air concentrations

assessed in Section 4.2.3 incorporate both

TSCA and other sources of formaldehyde

in indoor air.

• EPA identified the built environment

(including building materials and other

products) as a source of increased

exposure to formaldehyde associated with

COUs. Indoor air concentrations assessed

in Section 4.2.3 incorporate both TSCA

and other sources of formaldehyde in

indoor air.

• EPA quantified exposures associated

with specific TSCA COUs based on

2016 and 2020 Chemical Data

Reporting (U.S. EPA, 2020a, 2016a),

the Formaldehyde and

Paraformaldehyde Use Report (U.S.

EPA, 2020d) and product weight

fractions and densities reported in

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 130 of 151

Category

Subcategory

Increased Exposure from

OtherSources

Increased Exposure from TSCA COUs

Quantitative Data Sources

chemical safety data sheets (SDSs)

identified through product-specific

internet searches; EPA quantified

exposures and risks associated with

aggregate indoor air based on a range

of monitoring data described in the

Indoor Air Assessment (U.S. EPA,

2024j).

• EPA did not directly quantify indoor

air exposures associated with other

sources.

Occupational

Workers and

occupational non-

users

• EPA identified firefighters as an

occupational group with increased

exposure to formaldehyde associated with

combustion containing building materials

with high concentrations to formaldehyde.

While combustion exposures are beyond

the scope of this assessment, this is a

remaining source of uncertainty in

characterizing aggregate exposures for

some groups.

• EPA identified all occupational exposure

scenarios as a potential source of exposure

to formaldehyde. Those with higher

frequency or higher duration exposures

are expected to have the greatest

exposures and risks. EPA evaluated risks

for a range of occupational exposure

scenarios that increase exposure to

formaldehyde, including manufacturing,

processing, and use of formulations

containing formaldehyde. EPA evaluated

risks for central tendency and high-end

exposure estimates for each of these

scenarios (Section 5.2.1).

• EPA quantified occupational

exposures associated with TSCA

COUs based on a range of COU-

specific data, including monitoring

data from OSHA and NIOSH and

modeled air concentrations. Specific

data sources are described in detail in

the Draft Occupational Exposure

Assessment (U.S. EPA, 2024k).

Consumer

High frequency

consumers

• EPA identified dietary exposures through

food, food packaging, drugs, and personal

care products that contain formaldehyde

as other sources that may contribute to

total formaldehyde exposure. These

exposures are beyond the scope of this

assessment and are a source of uncertainty

in characterizing aggregate exposures.

• Consumer products designed for children

(e.g., children’s toys) may lead to elevated

exposures for children and infants.

• EPA identified all consumer exposure

scenarios involving TSCA COUs as

potential sources of exposure to

formaldehyde. Those with higher

frequency and/or higher duration

exposures are expected to have the

greatest exposures and risks.

• EPA quantified consumer exposure

(U.S. EPA, 2024d) based on the

Formaldehyde and Paraformaldehyde

Use Report (U.S. EPA, 2020d) and

the Exposure Factors Handbook (U.S.

EPA, 2011) (Ch. 17).

High duration

consumers

3372

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 131 of 151

PESS Based on Greater Susceptibility

3373

In this section, EPA addresses subpopulations and lifestages expected to be more susceptible to

3374

formaldehyde exposure than others. This discussion draws heavily from the recent summary of

3375

susceptible populations and lifestages included in the draft IRIS assessment. Table_Apx C-2. presents

3376

the data sources that were used in the PESS analysis evaluating susceptible subpopulations and identifies

3377

whether and how the subpopulation was addressed quantitatively in the risk evaluation of formaldehyde.

3378

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 132 of 151

Table_Apx C-2. Susceptibility Category, factors, and evidence for PESS susceptibility

3379

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Lifestage

Embryos/

fetuses/infants

Direct quantitative human and

animal evidence for developmental

toxicity following inhalation

exposure (e.g., decreased fertility,

increased spontaneous abortions and

changes in brain structures).

Taskinen et al. (1999)

John et al. (1994)

Sarsilmaz et al.

(2007)

Aslan et al. (2006)

–

Hazard value for chronic inhalation is

supported in part by dose-response

information on female reproductive

effects and developmental toxicity and

is expected to be protective of these

endpoints

Infants and

children

In some studies, children appear to

be more susceptible than adults to

respiratory effects of formaldehyde.

Early life exposures to chemicals

with a mutagenic mode of action

may increase cancer risk. EPA has

concluded that the evidence is

sufficient to conclude that a

mutagenic mode of action of

formaldehyde is operative in

formaldehyde-induced

nasopharyngeal carcinogenicity.

Bateson and

Schwartz (2008)

Venn et al. (2003)

Annesi-Maesano et

al. (2012)

Krzyzanowski et al.

(1990).

U.S. EPA (2005b)

Developing lungs

until age 6-8,

narrower airways

Different expression

of enzymes

responsible for

metabolizing

formaldehyde

Bateson and

Schwartz (2008)

Thompson et al.

(2009)

Hazard value for chronic inhalation is

based in part on dose-response

information on asthma

prevalence/asthma control in children.

ADAFs are applied to nasopharyngeal

cancer risk estimates to account for

increased susceptibility to cancer

following exposure during early life.

Pregnant

women

No direct evidence identified

–

Pregnant women may

have increased

sensitivity to the

development and

exacerbation of atopic

eczema following

exposure to

formaldehyde

Matsunaga et al.

(2008)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Males of

reproductive

age

Direct quantitative evidence in

humans and animals evidence for

reduced fertility following inhalation

exposure

–

Possible contributors

to male reproductive

effects/infertility (see

also factors in other

rows):

• Enlarged veins of

testes

• Trauma to testes

CDC (2023b)

Hazard value for chronic inhalation is

supported in part by dose-response

information on male reproductive

toxicity and is expected to be

protective of these endpoints

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 133 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Lifestage

• Anabolic steroid

or illicit drug use

Cancer treatment

Older adults

No direct evidence identified

–

Older adults may have

reduced metabolism

and higher rates of

chronic diseases that

may increase

susceptibility

–

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Pre-existing

disease or

disorder

Health

outcome/

target organs

A few epidemiological studies found

that individuals with asthma and

allergies were more susceptible to

the deterioration of respiratory

function after being exposed to

formaldehyde than those without

these conditions.

Evidence from human and animal

studies indicated that individuals

with pre-existing nasal damage or a

history of respiratory issues were

more susceptible to developing

formaldehyde induced nasal cancer.

Krzyzanowski et al.

(1990)

Kriebel et al. (1993)

Woutersen et al.

(1989)

Appelman et al.

(1988)

Falk et al. (1994)

Individual variations

in nasal anatomy and

soluble factors in the

upper respiratory tract

can potentially

influence the uptake

of highly reactive

gases like

formaldehyde. This

variability could

possibly lead to

differences in the

distribution of inhaled

formaldehyde and

susceptibility to its

health effects.

ICRP (1994)

Santiago et al.

(2001)

Singh et al.

(1998)

Acute inhalation hazard values are

based in part on dose-response

information in humans already

identified as sensitive to

formaldehyde in dermal patch test

studies.

No direct quantitative adjustment to

chronic inhalation, oral or dermal

hazard values or risk estimates; Use of

Lifestyle

activities

Smoking

No direct evidence identified

–

Heavy smoking may

increase susceptibility

to formaldehyde

toxicity. However, it

is unclear if this

increased sensitivity

is due to additional

formaldehyde

exposure or other

chemicals in cigarette

smoke.

Fishbein (1992)

CDC (2023a)

CDC (2023b)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 134 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Lifestyle

activities

Alcohol

consumption

No direct evidence identified

–

Chronic alcohol

consumption may

affect the

susceptibility to

reproductive and

cancer related health

outcomes.

CDC (2023a)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Physical

activity

Studies observed that prolonged

physical activity increased an

individual’s susceptibility to

formaldehyde induced respiratory

impairments. These studies

demonstrated that those who were

exposed to formaldehyde after 15

minutes of exercise experienced

more significant declines in lung

function compared to those who had

shorter exercise sessions or no

exercise at all.

Green et al. (1987)

Green et al. (1989)

Insufficient activity

may increase

susceptibility to

multiple health

outcomes

Overly strenuous

activity may also

increase

susceptibility.

CDC (2022)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Sociodemo-

graphic status

Race/ethnicity

An epidemiological study suggests a

racial difference in susceptibility to

formaldhyde toxicity, as nonwhite

individuals were found to have

higher mortality rates for

nasopharyngeal cancer and multiple

myeloma compared to their white

counterparts.

Hayes et al. (1990)

–

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Socio-

economic

status

No direct evidence identified

–

Individuals with

lower socioeconomic

status may experience

adverse health

ODPHP (2023b)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 135 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Sociodemo-

graphic status

outcomes due to

unmet social needs,

environmental factors,

and limited access to

healthcare services.

Sex/gender

A higher prevalence of burning or

tearing eyes was observed among

women compared to men, suggesting

that women may be more sensitive to

the irritant properties of

formaldehyde on the eyes and upper

respiratory tract.

Several animal studies showed that

males exhibit a higher incidence of

lesions in the upper respiratory tract

than females.

Evidence from epidemiological

studies and animal models indicates

that formaldehyde exposure can lead

to male reproductive impairments,

reduced fertility, and increased risk

of miscarriage in women

Liu et al. (1991)

Woutersen et al.

(1987)

Zwart et al. (1988)

Maronpot et al.

(1986)

Kerns et al. (1983)

Taskinen et al. (1999)

John et al. (1994)

Wang et al. (2015)

–

Both acute and chronic inhalation

hazard values are based in part on

epidemiological studies include that

include both male and female

subjects,

Nutrition

Diet

No direct evidence identified

–

An antioxidant

deficient diet may

exacerbate

inflammatory

responses, primarily

due to formaldehyde’s

well-known

inflammatory

properties.

Obesity can increase

susceptibility to

cancer.

CDC (2023a)

CDC (2020)

CDC (2023c)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Malnutrition

No direct evidence identified

–

Micronutrient

malnutrition can

result in various

CDC (2021)

CDC (2023c)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 136 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Nutrition

conditions, such as

birth defects, maternal

and infant mortality,

preterm birth, low

birth weight, poor

fetal growth,

childhood blindness,

and undeveloped

cognitive ability.

Deficiencies in

micronutrients may

increase an

individual’s

susceptibility to the

adverse health effects

of formaldehyde,

particularly

respiratory

impairments. This is

due to the critical role

of micronutrients in

maintaining robust

immune function,

potent antioxidant

defenses, and the

structural integrity of

the respiratory

system.

Genetics/

epigenetics

Target organs

No direct evidence identified

–-

Genetic disorders,

such as Klinefelter’s

syndrome, Y-

chromosome

microdeletion,

myotonic dystrophy

can affect male

reproduction/fertility

CDC (2023b)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Toxicokinetics

Studies suggested that certain genetic

variants could impair the activity of

ADH and ALDH enzyme. This

Wu et al. (2007)

Hedberg et al. (2001)

–

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 137 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Genetics/

epigenetics

potential impairment could reduce

the clearance of formaldehyde,

thereby increasing susceptibility to

adverse health effects associated

with formaldehyde exposure.

Studies have demonstrated that

genetic variations in ADH3 and

ALDH2 genes have been associated

to higher susceptibility to asthma and

CNS toxicity, while polymorphism

in genes related to DNA repair, such

as XRCC3, have been shown to

impact susceptibility to formaldhyde

induced genotoxicity.

Studies in experimental animals with

genetically modified ALDH2 and

ALDH5 genes, responsible for

eliminating endogenous

formaldhyde, suggested that

variations in these genes could

potentially increase susceptibility to

genotoxicity.

Although some studies have

suggested that specific genetic

variants may influence susceptibility

to formaldehyde toxicity, their

findings have not been conclusive.

Deltour et al. (1999)

Tan et al. (2018)

Nakamura et al.

(2020)

Dingler et al. (2020)

Other

chemical and

nonchemical

stressors

Built

environment

No direct evidence identified

–

Poor quality housing

often contains

environmental

triggers of asthma

such as pests, mold,

dust, building

materials that may

exacerbate reduced

asthma control

associated with

ODPHP (2023a)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 138 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

Other

chemical and

nonchemical

stressors

formaldehyde

exposure

Social

environment

No direct evidence identified

–

Poverty, violence, as

well as other social

factors may make

some populations

more susceptible to

the health effects

associated with

formaldehyde

exposure.

CDC (2023d)

ODPHP (2023c)

No direct quantitative adjustment to

hazard values or risk estimates; Use of

Chemical co-

exposures

Several studies have demonstrated

that co-exposure to formaldehyde

and other substances, including

environmental pollutants and dietary

components, could potentially affect

respiratory health, hypersensitivity

reactions, or lung function.

While studies have indicated that

certain dietary components, such as

methanol and caffeine can contribute

to the endogenous production of

formaldehyde in non-respiratory

tissues, the extent to which this

influences susceptibility to inhaled

formaldehyde remains unclear.

Environmental tobacco smoke

exposure has been associated with an

increased likelihood of

hypersensitivity responses in

individuals concurrently exposed to

formaldehyde. Studies suggest that

exposure to tobacco smoke may

potentiate the effects of

formaldehyde or even trigger such

responses at lower formaldehyde

Besaratinia et al.

2014

Fang et al. 2004

Gavriliu et al. 2013

Hohnloser et al.

(1980)

Riess et al. (2010)

Summers et al.

(2012)

Krzyzanowski et al.

(1990)

–

No direct quantitative adjustment to

hazard values or risk estimates; Use of

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 139 of 151

Susceptibility

Category

Specific

Factors

Direct Evidence this Factor

Modifies Susceptibility to Formaldehyde

Indirect Evidence of Potential Impact

through Target Organs or Biological

Pathways Relevant to Formaldehyde

Incorporation of Each Factor into

the Risk Evaluation

Description of Interaction

Key Citations

Description of

Interaction

Key Citations

concentrations, particularly in

children and nonsmoking adults

3380

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 140 of 151

Appendix D AMBIENT AIR RISK ESTIMATES – COMMERCIAL

3381

USES

3382

The ambient air exposure assessment for formaldehyde quantitatively evaluates exposures resulting

3383

from industrial releases of formaldehyde to ambient air. EPA expects that releases resulting from TSCA

3384

industrial COUs have larger point source emissions than the air emissions resulting from commercial

3385

uses.

3386

3387

As discussed in the Environmental Release Assessment (U.S. EPA, 2024g), where available, EPA used

3388

TRI and NEI to inform air releases from commercial COUs. However, facilities are only required to

3389

report to TRI if the facility has 10 or more full-time employees; is included in an applicable North

3390

American Industry Classification System (NAICS) code; and manufactures, processes, or uses the

3391

chemical in quantities greater than a certain threshold. Reporting to NEI depends on submissions

3392

voluntarily provided by state, local, and tribal agencies and is supplemented by data from other EPA

3393

programs. For NEI, the general threshold for major source is the potential to emit more than 10 tons per

3394

year for a single Hazardous Air Pollutant (HAP), or 25 tons/year for any combination of HAPs.

3395

3396

Due to these limitations, commercial sites that use formaldehyde and/or formaldehyde-containing

3397

products may not report to TRI or NEI and are therefore not included in these datasets.

3398

3399

EPA did not quantify releases and therefore ambient air risk estimates for the following COUs:

3400

• Distribution in commerce

3401

• Commercial use – chemical substances in treatment/care products – laundry and dishwashing

3402

products

3403

• Commercial use – chemical substances in treatment products – water treatment products

3404

• Commercial use – chemical substances in outdoor use products – explosive materials

3405

• Commercial use – chemical substances in products not described by other codes – other:

3406

laboratory chemicals; and

3407

• Commercial use – chemical substances in automotive and fuel products- automotive care

3408

products; lubricants and greases; fuels and related products.

3409

EPA discusses the release potential for each COU in in the Draft Environmental Release Assessment for

3410

Formaldehyde (U.S. EPA, 2024g) based on the available information. In general, EPA expects industrial

3411

COUs to be the drivers of risk for ambient air from the TSCA COUs within the scope of this draft risk

3412

evaluation.

3413

3414

For the following commercial COUs

3415

• Commercial use – chemical substances in furnishing treatment/care products- floor coverings;

3416

foam seating and bedding products; furniture and furnishings not covered elsewhere; cleaning

3417

and furniture care products; fabric, textile, and leather products not covered elsewhere-

3418

construction

3419

• Commercial Use – chemical substances in construction, paint, electrical, and metal products-

3420

adhesives and sealants; paint and coatings

3421

Use of fuels may be associated with petroleum refinery and utilities, however, note formaldehyde from combustion sources

is not assessed as an independent COU subcategory in this risk evaluation.

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 141 of 151

• Commercial Use – chemical substances in furnishing treatment/care products –

3422

building/construction materials – wood and engineered wood products; building/ construction

3423

materials not covered elsewhere

3424

EPA expects emissions may be similar to the construction sector, which has cancer risk estimate lower

3425

than 1×10

−6

based on 100 to 1,000 m from the release site for the 95th percentile annual reported release

3426

amount.

3427

3428

For the following commercial COUs

3429

• Commercial use – chemical substances in electrical products – electrical and electronic products

3430

• Commercial use – chemical substances in metal products – metal products not covered elsewhere

3431

EPA expects emissions may be similar to the electrical equipment, appliance, and component

3432

manufacturing and fabricated metal product manufacturing sector, which has cancer risk estimate lower

3433

than 1×10

−6

based on 100 to 1,000 m from the release site for the 95th percentile annual reported release

3434

amount.

3435

For the following commercial COU, Commercial use – chemical substances in agriculture use products

3436

– lawn and garden products, EPA expects emissions may be similar to the agriculture, forestry, fishing,

3437

and hunting sector, which has risk estimate lower than 1×10

−6

based on 100 to 1,000 m from the release

3438

site for the 95th percentile annual reported release amount.

3439

3440

For the following commercial COUs

3441

• Commercial use – chemical substances in packaging, paper, plastic, hobby products – paper

3442

products; plastic and rubber products; toys, playground, and sporting equipment

3443

• Commercial use – chemical substances in packaging, paper, plastic, hobby products- arts, crafts,

3444

and hobby materials

3445

• Commercial use – chemical substances in packaging, paper, plastic, hobby products- ink, toner,

3446

and colorant products; photographic supplies

3447

EPA expects emissions may be similar to the Printing and Related Support Activities & Photographic

3448

Film Paper, Plate, and Chemical Manufacturing sector, which have risk estimates lower than 1×10

−6

3449

based on 100 to 1,000 m from the release site for the 95th percentile annual reported release amount.

3450

EPA does, however, note that printing operations that use printing ink, toner, or colorant products

3451

containing formaldehyde may occur at industrial sites such as those included in Paper Manufacturing,

3452

which has a cancer risk estimate of 1.24×10

−5

3453

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 142 of 151

Appendix E DRAFT OCCUPATIONAL EXPOSURE VALUE

3454

DERIVATION

3455

EPA has calculated a draft 8-hour existing chemical occupational exposure value to summarize the

3456

occupational exposure scenario and sensitive health endpoints into a single value. EPA calculated the

3457

draft value rounded to 0.011 ppm (14 µg/m

) for inhalation exposures to formaldehyde as an 8-hour

3458

TWA and for consideration in workplace settings (see Appendix E.1) based on the chronic and

3459

intermediate non-cancer hazards value for respiratory effects.

3460

3461

TSCA requires risk evaluations to be conducted without consideration of costs and other non-risk

3462

factors, and thus this draft occupational exposure value represents a risk-only number. If risk

3463

management for formaldehyde follows the final risk evaluation, EPA may consider costs and other non-

3464

risk factors, such as technological feasibility, the availability of alternatives, and the potential for critical

3465

or essential uses. In general, any existing chemical exposure limit (ECEL) used for occupational safety

3466

risk management purposes could differ from the draft occupational exposure value presented in this

3467

appendix based on additional consideration of exposures and non-risk factors consistent with TSCA

3468

section 6(c), and this is certain to be the case for formaldehyde. The unique challenge associated with

3469

this evaluation is that the formaldehyde released from activities and products that are subject to TSCA is

3470

mixed in with the formaldehyde released from all sources as described in the executive summary, which

3471

could raise a challenge if/when an implementable regulatory occupational exposure limit is designed.

3472

More specifically, the draft occupational exposure value of 14 µg/m

for formaldehyde is below ~20 -

3473

40 µg/m

(50th to 95th percentile of concentrations measured in AHHS II for indoor air in residential

3474

settings)for indoor air. EPA must therefore consider this unique challenge if it ultimately designs and

3475

proposes a regulatory limit for occupational inhalation exposures to formaldehyde.

3476

3477

This calculated draft value for formaldehyde represents the exposure concentration below which

3478

workers and occupational non-users are not expected to exhibit any appreciable risk of adverse

3479

toxicological outcomes, accounting for potentially exposed and susceptible populations (PESS). It is

3480

derived based on the most sensitive human health effect relative to benchmarks and standard

3481

occupational scenario assumptions of 8 hours/day, 5 days/week exposures for a total of 250 days

3482

exposure per year, and a 40-year working life.

3483

3484

EPA expects that at the draft occupational exposure value of 0.011 ppm (14 µg/m

), a worker or ONU

3485

also would be protected against respiratory effects resulting from chronic exposures. In addition, this

3486

calculated draft value would protect against excess risk of nasopharyngeal cancer above the 1×10

−4

3487

benchmark value resulting from lifetime exposure if ambient exposures are kept below this draft

3488

occupational exposure value. The acute exposure limit is unchanged for all durations of a single

3489

exposure and also serves as the short-term exposure limit (STEL) to protect against 15-minute

3490

exposures.

3491

3492

Of the identified occupational monitoring data for formaldehyde, there have been measured workplace

3493

air concentrations below the calculated draft exposure value. A summary table of available monitoring

3494

methods from the Occupational Safety and Health Administration (OSHA) and the National Institute for

3495

Occupational Safety and Health (NIOSH) is included in Appendix E.2. The table covers validated

3496

methods from governmental agencies and is not intended to be a comprehensive list of available air

3497

monitoring methods for formaldehyde. The calculated draft exposure value is above the limit of

3498

detection (LOD) and limit of quantification (LOQ) using at least one of the monitoring methods

3499

identified.

3500

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 143 of 151

The Occupational Safety and Health Administration (OSHA) set a permissible exposure limit (PEL) as

3501

an 8-hour TWA for formaldehyde of 0.75 ppm in 1992 (https://www.osha.gov/annotated-pels), with an

3502

action level of 0.5 ppm. In addition, OSHA has set a STEL of 2 ppm. OSHA’s PEL must undergo both

3503

risk assessment and feasibility assessment analyses before selecting a level that will substantially reduce

3504

risk under the Occupational Safety and Health Act. EPA’s calculated draft exposure value is a lower

3505

value and is based on newer information and analysis from this risk evaluation.

3506

3507

There are also recommended exposure limits established for formaldehyde by other governmental

3508

agencies and independent groups. The American Conference of Governmental Industrial Hygienists

3509

(ACGIH) set a Threshold Limit Value (TLV) at 0.1 ppm TWA and 0.3 ppm STEL in 2017. This

3510

chemical also has a NIOSH Recommended Exposure Limit (REL) of 0.016 ppm TWA and 15-minute

3511

Ceiling limit of 0.1 ppm (https://www.cdc.gov/niosh/npg/).

3512

Draft Occupational Exposure Value Calculations

3513

This appendix presents the calculations used to estimate draft occupational exposure values using inputs

3514

derived in this draft risk evaluation. Multiple values are presented below for hazard endpoints based on

3515

different exposure durations. For formaldehyde, the most sensitive occupational exposure value is based

3516

on respiratory effects and the resulting 8-hour TWA is rounded to 14 µg/m

. The human health hazard

3517

values used in these equations are based on the inhalation non-cancer hazard values and the IUR

3518

summarized in Table 3-1.

3519

3520

Draft Intermediate Non-cancer Occupational Exposure Value

3521

The draft exposure value was calculated for the occupational non-cancer repeat-dose human equivalent

3522

concentration for respiratory effects as the concentration at which the chronic margin of exposure

3523

(MOE) would equal the benchmark MOE for 8-hour intermediate occupational exposures with

3524

Equation_Apx E-1:

3525

3526

Equation_Apx E-1.

3527































3528

3529









  

  















 

3530

3531









 



















 







3532

3533

Where:

3534

Molar Volume = 24.45 L/mol, the volume of a mole of gas at 1 atm and 25 °C

3535

MW = Molecular weight of formaldehyde (30.026 g/mole)

3536

3537

Draft Acute/Short-Term, Non-cancer Occupational Exposure Value

3538

The acute occupational exposure value (EV

acute

), equivalent to the 15-minute STEL, was calculated as

3539

the concentration at which the acute MOE would equal the benchmark MOE for acute occupational

3540

exposures using Equation_Apx E-2:

3541

3542

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 144 of 151

Equation_Apx E-2.

3543





















   







3544

3545

Draft Chronic, Non-cancer Occupational Exposure Value

3546

The chronic occupational exposure value (EV

chronic

) can be calculated as the concentration at which the

3547

chronic MOE would equal the benchmark MOE for chronic occupational exposures. However, for

3548

purposes of risk management, EPA has determined that because the same critical health effect applies to

3549

both in both intermediate and chronic exposure contexts, the relevant averaging time should be

3550

considered equivalent across both exposure scenarios. Therefore, the resulting EV

chronic

would be the

3551

same as the draft exposure value based on intermediate exposures.

3552

3553

Draft Lifetime Cancer Occupational Exposure Value

3554

The EV

cancer

is the concentration at which the extra cancer risk is equivalent to the benchmark cancer

3555

risk of 1×10

−4

3556



















    











3557







  



















 













 

















3558

= 0.108 ppm = 1.33







3559

3560

Where:

3561

HECrepeat

= Averaging time for the POD/HEC used for evaluating non-cancer,

3562

intermediate and chronic occupational risk, based on study

3563

conditions and/or any HEC adjustments (24 hr/day for 30 days)

3564

(see Section 4.2.2.1)

3565

HECacute

= Averaging time for the POD/HEC used for evaluating non-cancer,

3566

acute occupational risk, based on study conditions and/or any HEC

3567

adjustments (24 hr/day) (see Section 4.2.2.1)

3568

IUR

Averaging time for the cancer IUR, based on study conditions and

3569

any adjustments (24 hr/day for 365 days/year) and averaged over a

3570

lifetime (78 years) (Supplemental File: Releases and Occupational

3571

Exposure Assessment; Appendix B).

3572

Benchmark MOE

acute

= Acute non-cancer benchmark margin of exposure, based on the

3573

total uncertainty factor of 10 (see Table 3-7)

3574

Benchmark MOE

repeat

= Short term non-cancer benchmark margin of exposure, based on

3575

the total uncertainty factor of 100 (see Table 3-8)

3576

Benchmark

Cancer

= Benchmark for excess lifetime cancer risk

3577

acute

= Exposure limit based on acute effects

3578

intermediate

= Existing chemical exposure limit (mg/m

), based on non-cancer

3579

effects following repeat exposures

3580

chronic

= Existing chemical exposure limit (mg/m

), based on non-cancer

3581

effects following repeat exposures

3582

cancer

= Exposure limit based on excess cancer risk

3583

ED = Exposure duration (8 hr/day) (see Table 3-8)

3584

EF = Exposure frequency (250 days/yr), (see Section 4.2.2.1)

3585

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 145 of 151

HEC

acute or repeat

= Human equivalent concentration for acute or intermediate/chronic

3586

occupational exposure scenarios, respectively

(see Tables 3-7 and

3587

3-8)

3588

IUR = Inhalation unit risk (per ppm) (see Table 3-6)

3589

IR = Inhalation rate (default is 1.25 m

/hr for workers and 0.6125 m

/hr

3590

for general population at rest)

3591

WY = Working years per lifetime at the 95th percentile (40 years

3592

(Supplemental File: Releases and Occupational Exposure

3593

Assessment; Appendix B)

3594

3595

Unit conversion:

3596

1 ppm = 1.23 mg/m

(based on molecular weight of 30.026 g/mol for formaldehyde)

3597

Summary of Air Sampling Analytical Methods Identified

3598

EPA conducted a search to identify relevant NIOSH and OSHA analytical methods used to monitor for

3599

the presence of formaldehyde in air (see Table_Apx E-1). This table covers validated methods from

3600

governmental agencies and is not intended to be a comprehensive list of available air monitoring

3601

methods for formaldehyde. The sources used for the search included the following:

3602

1. NIOSH Manual of Analytical Methods (NMAM), 5th Edition;

3603

2. NIOSH NMAM 4th Edition; and

3604

3. OSHA Index of Sampling and Analytical Methods.

3605

3606

Table_Apx E-1. Limit of Detection (LOD) and Limit of Quantification (LOQ) Summary for Air

3607

Sampling Analytical Methods Identified

3608

Air Sampling

Analytical Methods

Year

Published

LOD

LOQ

Notes

Source

NIOSH Method 2016

2016

0.012

ppm

N/A

Estimated LOD is 0.07

µg/sample. The working

range is 0.012 to 2.0 ppm

for a 15-L sample.

NIOSH Manual of

Analytical Methods

(NMAM 2016)

NIOSH Method 2541

1994

0.24 ppm

N/A

Estimated LOD is 1

µg/sample. The working

range is 0.24 to 16 ppm

for a 15-L sample.

NIOSH Manual of

Analytical Methods,

4th Edition

(NMAM 2541)

NIOSH Method 3500

1994

0.02 ppm

N/A

Estimated LOD is 0.5

µg/sample. The working

range is 0.02 to 4 ppm for

an 80-L sample.

NIOSH Manual of

Analytical Methods,

4th Edition

(NMAM 3500)

NIOSH Method 5700

1994

0.0004

mg/m

(0.0003

ppm)

N/A

Estimated LOD is 0.08

µg/sample. The working

range is 0.0004 to 3.8

mg/m

for a 1,050-L

sample. Used for

determination of

formaldehyde in both

textile and wood dusts.

NIOSH Manual of

Analytical Methods,

4th Edition

(NMAM 5700)

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 146 of 151

Air Sampling

Analytical Methods

Year

Published

LOD

LOQ

Notes

Source

OSHA Method 52

1989

16 ppb

Detection limit and

reliable quantification

limit is 482 ng per sample

(16 ppb for 24 L)

OSHA Index of

Sampling and

Analytical Methods

(OSHA 52)

OSHA Method 1007

https://www.osha.gov

/sites/default/files/met

hods/osha-1007.pdf

2005

0.56, 1.70, or

0.17 ppb

(Sampler –

ChemDisk-

AL, UMEx

100, DSD-

DNPH,

respectively)

1.88, 5.68, or

0.58 ppb

(Sampler –

ChemDisk-

AL, UMEx

100, DSD-

DNPH,

respectively)

Method reports

LOD/LOQ of overall

procedure as 0.56/1.88

ppb for ChemDisk-AL

samplers, 1.70/5.68 ppb

for UMEx 100 samplers,

and 0.17/0.58 for DSD-

DNPH samplers

OSHA Index of

Sampling and

Analytical Methods

(OSHA 1007)

ppm = parts per million; ppb = parts per billion; ppt = parts per trillion

EPA has additional air sampling methods targeted for measurement of ambient and indoor air, the methods

listed in this table are air sampling for occupational exposures.

These sources cover a range of LOD including both below and above the preliminary occupational exposure

value.

The method is suitable for the simultaneous determinations of acrolein and formaldehyde.

This is the most sensitive formaldehyde method in the NIOSH Manual of Analytical Methods and is able to

measure ceiling levels as low as 0.1 ppm (1 5-L sample). It is best suited for the determination of formaldehyde

in area samples.

Results should be considered separately from vapor-phase formaldehyde exposure; Method measures both

“released” and formaldehyde equivalents.

Recommends use of OSHA Method 52 when monitoring exposures resulting from the use of formalin

solutions.

3609

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 147 of 151

Appendix F ACUTE AND CHRONIC (NON-CANCER AND

3610

CANCER) OCCUPATIONAL INHALATION

3611

EQUATIONS

3612

This assessment provides estimates of 15-minute peak air concentrations, short-term air concentrations,

3613

and full-shift (8- or 12-hour) concentrations. For calculation of risk, these exposure estimates are

3614

incorporated with additional parameter inputs, such as working years, exposure duration and frequency,

3615

and lifetime years.

3616

3617

AC is used to estimate workplace inhalation exposures for acute risks (i.e., risks occurring after less than

3618

one day of exposure), per Equation_Apx F-1, Equation_Apx F-2, and Equation_Apx F-3 below.

3619

3620

Equation_Apx F-1.

3621

 

    





3622

Where:

3623

AC = Acute exposure concentration

3624

C = Contaminant concentration in air (TWA)

3625

ED = Exposure duration (hr/day), 0.25 hr/day

3626

BR = Breathing rate ratio (unitless), 1

3627

acute

= Acute averaging time (hr), 0.25 hr

3628

3629

ADC and LADC are used to estimate workplace exposures for non-cancer and cancer risks, respectively.

3630

These exposures are estimated per Equation_Apx F-2, as follows:

3631

3632

Equation_Apx F-2.

3633

 

        



3634

    

day

month

 

day

3635

    

day

 

day

3636

3637

Where:

3638

ADC = Average daily concentration used for chronic non-cancer risk calculations

3639

ED = Exposure duration (hr/day)

3640

EF = Exposure frequency (day/yr)

3641

BR = Breathing rate ratio (unitless),

3642

WY = Working years per lifetime (yr)

3643

= Averaging time (hr) for sub-chronic, non-cancer risk

3644

AT = Averaging time (hr) for chronic, non-cancer risk

3645

3646

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 148 of 151

Equation_Apx F-3.

3647

 

        



3648



   

day

 

day

3649

Where:

3650

LADC = Lifetime average daily concentration used for chronic cancer risk

3651

calculations

3652

ED = Exposure duration (hr/day)

3653

EF = Exposure frequency (day/yr)

3654

WY = Working years per lifetime (yr),

3655

= Averaging time (hr) for cancer risk

3656

LT = Lifetime years (yr) for cancer risk, 78 yr

3657

For exposure duration, frequency, and working years used in this appendix, see Table_Apx F-1.

3658

Table_Apx F-1. Appendix F Formulae – Symbols, Values, and Units

3659

Symbol

Value

Unit

8 or 12

hour/day

250 or 167

day/year

(CT)

years

(HE)

years

(CT)

271,560

hours

(HE)

350,400

hours



683,280

hours

3660

Worker Years

3661

EPA has developed a triangular distribution for working years. EPA has defined the parameters of the

3662

triangular distribution as follows:

3663

• Minimum value: BLS CPS tenure data with current employer as a low-end estimate of the

3664

number of lifetime working years: 10.4 years;

3665

• Mode value: The 50th percentile tenure data with all employers from SIPP as a mode value for

3666

the number of lifetime working years: 36 years; and

3667

• Maximum value: The maximum average tenure data with all employers from SIPP as a high-end

3668

estimate on the number of lifetime working years: 44 years.

3669

This triangular distribution has a 50th percentile value of 31 years and a 95th percentile value of 40

3670

years. EPA uses these values for central tendency and high-end ADC and LADC calculations,

3671

respectively.

3672

3673

The BLS (U.S. BLS, 2014) provides information on employee tenure with current employer obtained

3674

from the CPS, which is a monthly sample survey of about 60,000 households that provides information

3675

on the labor force status of the civilian non-institutional population age 16 and over. CPS data are

3676

released every 2 years. The data are available by demographics and by generic industry sectors but are

3677

not available by NAICS codes.

3678

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 149 of 151

The U.S. Census’ (U.S. Census Bureau, 2019a) SIPP provides information on lifetime tenure with all

3679

employers. SIPP is a household survey that collects data on income, labor force participation, social

3680

program participation and eligibility, and general demographic characteristics through a continuous

3681

series of national panel surveys of between 14,000 and 52,000 households (U.S. Census Bureau, 2019b).

3682

EPA analyzed the 2008 SIPP Panel Wave 1, a panel that began in 2008 and covers the interview months

3683

of September 2008 through December 2008 (U.S. Census Bureau, 2019a, b). For this panel, lifetime

3684

tenure data are available by Census Industry Codes, which can be cross walked with NAICS codes.

3685

3686

SIPP data include fields for the industry in which each surveyed, employed individual works

3687

(TJBIND1), worker age (TAGE), and years of work experience with all employers over the surveyed

3688

individual’s lifetime. Census household surveys use different industry codes than the NAICS codes used

3689

in its firm surveys, so these were converted to NAICS using a published crosswalk (Census Bureau,

3690

2012b). EPA calculated the average tenure for the following age groups: (1) workers aged 50 and older,

3691

(2) workers aged 60 and older, and (3) workers of all ages employed at time of survey. EPA used tenure

3692

data for age group “50 and older” to determine the high-end lifetime working years, because the sample

3693

size in this age group is often substantially higher than the sample size for age group “60 and older.” For

3694

some industries, the number of workers surveyed, or the sample size, was too small to provide a reliable

3695

representation of the worker tenure in that industry. Therefore, EPA excluded data where the sample

3696

size is less than five from our analysis.

3697

3698

Table_Apx F-2 summarizes the average tenure for workers aged 50 years and older from SIPP data.

3699

Although the tenure may differ for any given industry sector, there is no significant variability between

3700

the 50th and 95th percentile values of average tenure across manufacturing and non-manufacturing

3701

sectors.

3702

3703

Table_Apx F-2. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+)

3704

Industry Sectors

Working Years

Average

50th

Percentile

95th

Percentile

Maximum

Manufacturing sectors (NAICS 31–33)

35.7

Non-manufacturing sectors (NAICS 42–81)

36.1

Source: (U.S. Census Bureau, 2019a).

Note: Industries where sample size is less than five are excluded from this analysis.

3705

BLS CPS data provides the median years of tenure that wage and salary workers had been with their

3706

current employer. Table_Apx F-3 presents CPS data for all demographics (men and women) by age

3707

group from 2008 to 2012. To estimate the low-end value on number of working years, EPA uses the

3708

most recent (2014) CPS data for workers aged 55 to 64 years, which indicates a median tenure of 10.4

3709

years with their current employer. The use of this low-end value represents a scenario where workers are

3710

only exposed to the chemical of interest for a portion of their lifetime working years, as they may

3711

change jobs or move from one industry to another throughout their career.

3712

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 150 of 151

Table_Apx F-3. Median Years of Tenure with Current Employer by Age Group

3713

Age

January 2008

January 2010

January 2012

January 2014

16 years and over

4.1

4.4

4.6

16 to 17 years

0.7

18 to 19 years

0.8

1.0

0.8

20 to 24 years

1.3

1.5

1.3

25 years and over

5.1

5.2

5.4

5.5

25 to 34 years

2.7

3.1

3.2

3.0

35 to 44 years

4.9

5.1

5.3

5.2

45 to 54 years

7.6

7.8

7.9

55 to 64 years

9.9

10.0

10.3

10.4

65 years and over

10.2

9.9

10.3

3714

PUBLIC RELEASE – DO NOT CITE OR QUOTE

March 2024

Page 151 of 151

Appendix G DERMAL EXPOSURE APPROACH

3715

The dermal load (Q

) is the quantity of chemical on the skin after the dermal contact event. This value

3716

represents the quantity remaining after the bulk chemical formulation has fallen from the hand that

3717

cannot be removed by wiping the skin (e.g., the film that remains on the skin). To estimate the dermal

3718

load for formaldehyde for occupational and consumer uses, EPA used dermal loading based on A

3719

Laboratory Method to Determine the Retention of Liquids on the Surface of the Hands (U.S. EPA, 1992)

3720

and formaldehyde weight concentrations relevant to the occupational use or consumer product. In

3721

addition, only acute exposures were quantitatively assessed given the identified dermal skin sensitization

3722

POD is likely only relevant to acute exposures (U.S. EPA, 2024i). The supporting study measured liquid

3723

retention on the surface of hands based on indirect (i.e., contact with saturated object) contact and direct

3724

(i.e., immersive) contact.

3725

3726

For consumer exposures, EPA assumes the product used may involve immersion into a liquid and that a

3727

pool of a liquid product was formed on the skin, or that a rag was used that reduced the evaporation of

3728

formaldehyde during use. A Qu of 10.3 mg/cm

was used to approximate hand immersion and wiping

3729

experiments, using oil-based products expected to have longer residence times on the skin relative to

3730

water-based products, as reported in (U.S. EPA, 1992). While this is the most protective value for

3731

consumer usage of oil-based products, it may overestimate exposures in some cases including when

3732

using water-based liquid products. Dermal exposures are only reasonably foreseen for consumers but not

3733

bystanders.

3734

3735

Owing to volatility and expected use patterns, dermal loading of formaldehyde from solid products is

3736

unlikely, except for certain textiles including clothing that are treated with formaldehyde in dyeing and

3737

wrinkle prevention step in the textile manufacturing process (Herrero et al., 2022). EPA could not

3738

identify supporting evidence for dermal loading exposures from the handling or wear of fabrics. The

3739

Agency also could not identify a diffusion coefficient of formaldehyde for clothing. Therefore, EPA had

3740

a low level of confidence in the estimation of dermal loading from textiles including clothing. Thus, a

3741

qualitative assessment is reported for this product type in the Draft Consumer Exposure Assessment for

3742

Formaldehyde (U.S. EPA, 2024d).

3743

3744

For occupational exposures, EPA uses the guidance in Updating CEB’s Method for Screening-Level

3745

Assessments of Dermal Exposure (U.S. EPA, 2013) on selection of Q

values. EPA assumes routine and

3746

incidental contact with liquids occur for workers during routine maintenance activities, manual cleaning

3747

of equipment, filling drums, connecting transfer lines, sampling, and bench-scale liquid transfers. For

3748

this event, the memorandum uses values of 0.7 to 2.1 mg/cm

-event for routine liquid contact. EPA uses

3749

the maximum value of the range from the memorandum to estimate high-end dermal loads. EPA also

3750

included a central tendency liquid dermal loading values, EPA used the 50th percentile of the dermal

3751

loading results from the underlying study (U.S. EPA, 1992). The 50th percentile value was 1.4 mg/cm

3752

event for routine/incidental contact with liquids.

3753

3754

EPA assumes routine and immersive contact with liquids occur for workers during manual spray

3755

applications or contact with very wet surfaces. For this event, the memorandum uses values of 1.3 to

3756

10.3 mg/cm

-event for liquid contact. EPA uses the maximum value of the range from the memorandum

3757

to estimate high-end dermal loads. EPA also included a central tendency liquid dermal loading values,

3758

EPA used the 50th percentile of the dermal loading results from the underlying study (U.S. EPA, 1992).

3759

The 50th percentile value was 3.8 mg/cm

-event for routine/incidental immersive contact with liquids.

3760

The dermal exposure estimates do not consider the use of gloves or other protective equipment.

3761