FLUORIDE IN DRINKING-WATER
6
A few investigations on skeletal fluorosis or the risk of fractures include quantitative estimates
of the dose–response relationship. Studies in China and India report an increased prevalence
of skeletal fluorosis above the level of 1.4 mg fluoride/litre in drinking-water (Jolly et al.,
1968; Choubisa et al., 1997; Xu et al., 1997). However, such studies suffer from limitations:
the diagnostic criteria are not always specified or consist of self-reported symptoms, and only
drinking-water is considered as a source of exposure. The latter problem is likely to be
important, since other studies (Liang et al., 1997; Ando et al., 1998) estimate that, at least in
some regions of China and India, the contribution from food can greatly exceed that from
water. Therefore, one cannot rule out that high rates of skeletal fluorosis associated with a
level greater than 1.4 mg/litre in drinking-water are due to other exposures. While there is a
clear excess of skeletal fluorosis in these studies for a total intake of 14 mg/day, the
quantitative relationship between total intake of fluoride from different sources and the risk of
skeletal fluorosis cannot be estimated because of substantial uncertainties in the prevalence of
effects in the range of intakes between 3 and 14 mg/day.
The studies on fractures are also difficult to interpret, but for different reasons:
Few studies report an interpretable range of exposures to fluoride.
Results tend to be contradictory, with no clear-cut trend in both men and women.
Total intake of fluoride is not estimated.
One exception is represented by a study in China (Li et al., 2001) in which different sources of
exposure have been considered and an estimate of total intake is presented. In this study, there
is an upward trend for the risk of total fractures above an exposure of 1.45 mg fluoride/litre in
drinking-water, but only for the highest level of exposure (i.e., >4.32 mg fluoride/litre in
drinking-water) was the relative risk statistically significant (RR = 1.47; P = 0.01). In the
concentration range of 1.45–2.19 mg fluoride/litre in drinking-water, corresponding to a total
intake of 6.54 mg/day, there was a relative risk for all fractures of 1.17 and for hip fractures of
2.13 (both not statistically significant).
In summary, estimates based on studies from China and India indicate that:
for a total intake of 14 mg/day, there is a clear excess risk of skeletal adverse effects; and
there is suggestive evidence of an increased risk of effects on the skeleton at total fluoride
intakes above about 6 mg/day.
Several epidemiological studies are available on the possible association between
fluoride in drinking-water and cancer rates among the population. IARC evaluated
these studies in 1982 and 1987 and considered that they provided inadequate evidence
of carcinogenicity in humans. Subsequently, IPCS (2002) considered all new data and
concluded that overall the evidence of carcinogenicity in laboratory animals is
inconclusive and that the weight of evidence does not support the hypothesis that
fluoride causes cancer in humans; however, the data on bone cancer are relatively
limited.
The results of several epidemiological studies on the possible adverse effects of
fluoride in drinking-water on pregnancy outcome indicate that there is no apparent
relationship between the rates of Down syndrome or congenital malformation and the
consumption of fluoridated drinking-water (IPCS, 1984, 2002; US EPA, 1985a;
Janssen et al., 1988).
It is known that persons suffering from certain forms of renal impairment have a
lower margin of safety for the effects of fluoride than the average person. The data