You are on page 1of 7

Drinking Water and Cancer

Robert D. Morris
Department of Family and Community Medicine, Medical College of Wisconsin, Milwaukee, Wisconsin
Any and all chemicals generated by human activity can and will find their way into water supplies. The types and quantities of carcinogens present in
drinking water at the point of consumption will differ depending on whether they result from contamination of the source water, arise as a
consequence of treatment processes, or enter as the water is conveyed to the user. Source-water contaminants of concern include arsenic,
asbestos, radon, agricultural chemicals, and hazardous waste. Of these, the strongest evidence for a cancer risk involves arsenic, which is linked to
cancers of the liver, lung, bladder, and kidney. The use of chlorine for water treatment to reduce the risk of infectious disease may account for a
substantial portion of the cancer risk associated with drinking water. The by-products of chlorination are associated with increased risk of bladder
and rectal cancer, possibly accounting for 5000 cases of bladder cancer and 8000 cases of rectal cancer per year in the United States. Fluoridation of
water has received great scrutiny but appears to pose little or no cancer risk. Further research is needed to identify and quantify risks posed by
contaminants from drinking-water distribution pipes, linings, joints, and fixtures and by biologically active micropollutants, such as microbial agents.
We need more cost-effective methods for monitoring drinking-water quality and further research on interventions to minimize cancer risks from
drinking water. - Environ Health Perspect 1 03(Suppl 8):225-232 (1995)
Key words: arsenic, asbestos, bladder cancer, cancer, chlorination, environment, fluoridation, rectal cancer, water

Introduction source water, arise as a consequence of asbestos, radon, agricultural chemicals,


Few things tie humans so directly to the treatment processes, or enter as the water is and hazardous waste.
natural environment as drinking water. conveyed to the user. Many different car- Some of the strongest evidence for a
The contamination of water is a direct cinogens may contaminate source waters, cancer risk associated with source-water
reflection of the degree of contamination but they usually exist in drinking water at contamination involves arsenic. Epidemi-
of the environment. After flushing airborne low concentrations. On the other hand, ologic studies from Taiwan have suggested
pollutants from the skies, rainwater literally chemicals that enter drinking water during that arsenic in drinking water poses sub-
washes over the entire human landscape the course of water treatment are limited in stantial risks of liver, lung, bladder, and
before running into the aquifers, streams, number, but these chemicals appear in kidney cancer as listed in Table 1 (1,2).
rivers, and lakes that supply our drinking- drinking-water supplies with greater fre- Although toxicologic studies do not pro-
water. Any and all of the chemicals gener- quency than most source water contami- vide unequivocal evidence of carcinogenic-
ated by human activity can and will find nants. Finally, the compounds contained in ity (3), occupational studies, as well as
their way into water supplies. Evaluating the pipes, joints, and fixtures of the water other epidemiologic studies, support the
possible links between drinking water distribution system may contaminate findings of the Taiwanese studies(4).
and cancer means identifying those chemi- treated water on its way to the consumer. Estimates of attributable risk based on the
cals that appear in enough water supplies Similarities in the construction of drinking- data in Table 1 suggest that an average
at sufficient concentrations to pose a water distribution systems mean that any level of arsenic 2.5 pg/l in drinking water
substantial attributable cancer risk. carcinogen entering through this pathway in the United States of causes approxi-
Contaminants may enter water supplies may be widespread and can pose substantial mately 3000 cases of cancer per year (4).
at many points before reaching the tap. The attributable risks of cancer. The following Although asbestos is a proven carcino-
types and quantities of carcinogens present discussion reviews the attributable risks gen, the attributable risks associated with
in drinking water at the point of consump- for contaminants entering at each of these asbestos in drinking water do not appear to
tion may result from contamination of the points. Data gaps are identified and be substantial. An early study in California
emerging areas of concern are discussed. (5) suggested that there may be an
Source-Water Contaminants Table 1. Estimated mortality risk ratios by arsenic
This paper was presented at the President's Except for naturally occurring minerals levels in drinking water in Taiwan.
Cancer Panel Conference on Avoidable Causes of such as calcium carbonate, contaminants
Cancer held 7-8 April 1994 in Bethesda, Maryland. Cancer Back- Water levels, pg/l p-Value
Manuscript received 9 March 1995; manuscript that enter the water supply through the site Sex ground 170 470 800 for trend
accepted 24 March 1995. source water generally occur at low con-
The author gratefully acknowledges R. Munasinghe centration levels. Source-water contami- Liver M 1.0 1.2 1.5 2.5 <0.001
for his assistance in preparation of the maps included F 1.0 1.6 2.1 3.6 <0.001
in this paper. nants of concern either are sufficiently
Address correspondence to Dr. Robert D. Morris, potent carcinogens to pose risks at Lung M 1.0 1.8 3.3 4.5 <0.001
Department of Family and Community Medicine, extremely low concentrations or cause F 1.0 2.8 4.3 8.8 <0.001
Medical College of Wisconsin, 8701 Watertown Bladder M 1.0 5.1 12.1 28.7 <0.001
Plank Road, Milwaukee, WI 53226. Telephone: local contamination at high concentra-
(414) 456-8382. Fax: (414) 266-8502. E-mail: tions. The source-water contaminants F 1.0 11.9 25.1 65.4 <0.001
rmorris@post.its.mcw.edu that have been the focus of concern Kidney M 1.0 4.9 11.9 19.6 <0.001
Abbreviations: IARC, International Agency for F 1.0 4.0 13.9 37.0 <0.001
Research on Cancer; DDT, dichlorodiphenyl- among those individuals investigating
trichloroethane. environmental cancer risks include arsenic, Data from Chen et al. ( 1).

Environmental Health Perspectives 225


R.D. MORRIS

elevation in colorectal cancer risk associ- cancer risks are difficult to detect because be made as pure as mountain spring water.
ated with asbestos in drinking water. It of the relatively low incidence of site- Chlorination destroys all animal and
appears that these findings are limited to specific neoplasms and the typically small microbial life, leaving no trace of itself
situations in which naturally occurring lev- size of exposed populations (16). An eco- afterwards" (21).
els are high. A subsequent, more detailed logic study in New Jersey found weak evi- This statement represented the prevail-
study of asbestos in source water, together dence for a positive association between ing wisdom until about 20 years ago when
with studies of asbestos leached from water volatile organic compounds in drinking halogenated organic compounds, particu-
distribution systems, suggests that, when water and leukemia (17). In a national larly chloroform, were identified in chlori-
asbestos is present at levels commonly ecologic study, Griffith et al. (18) found nated drinking water (22). A subsequent
found in drinking water, it does not pose a evidence of elevated cancer rates in the survey of water supplies showed that these
major cancer risk (6,7). vicinity of hazardous waste sites. Limita- compounds were common in water sup-
Radon is also a known carcinogen; tions on ecologic data urge caution in the plies throughout the United States and that
however, the evidence linking consumption interpretation of such findings. Contami- concentrations were far higher in treated
of radon-contaminated water to human nation of wells associated with hazardous surface water than in treated groundwater
cancer is weak (8). The relationship waste disposal in Woburn, Massachusetts, (23). With these revelations came a shift in
between ionizing radiation and cancer is was ultimately linked to elevations in their the basis of our definition of cleanliness in
well understood. This information, coupled incidence of leukemia (19). Although this drinking water. New concerns about cancer
with measured levels of radon in drinking investigation was arguably the most thor- risks associated with chemical contamina-
water, suggests that fewer than 100 cases of ough study of this kind, questions were tion from chlorination by-products have
cancer occur each year in the United States raised about the magnitude of the risk given rise to 25 epidemiologic studies.
as a consequence of consuming radon in (20). There are numerous factors that Table 2 summarizes the results of a
drinking water (9). make it difficult, if not impossible, to esti- metaanalysis of the cohort and case-control
Farm runoff containing agricultural mate the attributable risks associated with studies that have been conducted to evalu-
chemicals and manure may lead to local or hazardous wastes on a national level, ate the association between consumption
regional contamination of source waters including the wide variety of chemicals of chlorinated drinking water and cancer at
with insecticides, fungicide, rodenticides, present in hazardous waste sites, the various sites (24). For each cancer site, the
herbicides, and fertilizers, which contain difficulties in assessing exposure, the obsta- pooled results from available studies show
phosphorous and nitrogen. Although some cles to establishing links between exposure elevations in risk, and the risk estimates
pesticides are carcinogens, drinking-water and cancer even when links are present, the achieved statistical significance for bladder
contamination resulting from their agricul- small size of exposed populations, and the and rectal cancer. Further analyses in this
tural application has not been directly asso- uncertainties concerning future risks. study suggested that risks increased with
ciated with cancer in epidemiologic studies. increasing exposure and that improvements
Emerging evidence, however, indicates that Cancer Risks Associated in exposure assessment yielded higher
fertilizers may pose cancer risks. with Water Treatment estimates of risk. Confounding could con-
Studies in China among populations Until this century, concerns about the ceivably explain the observed pattern of
exposed to high levels of nitrates in drink- cleanliness of drinking water focused association, but stratification into studies
ing water have suggested links between almost exclusively on the presence or that adjusted or did not adjust for con-
nitrate contamination and stomach and absence of pathogens. Ironically, the chlo- founders does not support such an asser-
liver cancer (10). In these studies, the rine used to reduce the risk of infectious tion. Studies that adjusted for population
histology of the gastric lesions has been disease may account for a substantial por- density, smoking, or occupation, did not
linked to the level of nitrates in the water tion of the cancer risk associated with demonstrate a difference in relative risk
(11) and cancer rates increased with the in drinking water. estimates. Although it is still possible that
vitro mutagenicity of the drinking water Chlorination of drinking water played a the pattern of associations could represent
(12). Nitrates may act as carcinogens central role in the reduction in the mortal-
through the formation of N-nitroso com- ity rates associated with waterborne
pounds (13). When human volunteers pathogens. Water chlorination was first Table 2. Results from metaanalysis of chlorination
were given proline, which is a secondary introduced at the Jersey City Water Works by-products and cancer.
amine, those participants in areas with in Boonton, New Jersey. The relative ease Relative risk 95% Confidence
higher levels of nitrate in their drinking of use of water chlorination, together with Site n8 estimate interval p
water had higher levels of N-nitrosoproline its potent bactericidal action, lead to the Bladder 7 1.21 1.09 1.34 <0.0001
in their urine than volunteers residing in rapid dissemination of this treatment tech- Brain 2 1.29 0.53 3.14 0.56
places with low nitrate levels in their drink- nology throughout the United States. Breast 4 1.18 0.90 1.54 0.24
ing water (14). Although an epidemiologi- Overshadowed by the clear benefits to Colon 7 1.11 0.91 1.35 0.32
cal study in France failed to demonstrate public health, the potential health risks Colorectal 8 1.15 0.97 1.37 0.10
an association between nitrates in drinking associated with water chlorination received Esophagus 5 1.11 0.85 1.45 0.43
Kidney 4 1.16 0.89 1.51 0.23
water and cancer (15), current evidence is little attention. This view is evident in an Liver 4 1.15 0.94 1.40 0.16
sufficient to warrant further study of this article heralding the opening of the Lung 5 1.01 0.86 1.18 0.94
potential carcinogen. Boonton waterworks, which appeared on Pancreas 6 1.05 0.91 1.22 0.48
Few examples of significant links the back page of the New York Times (21). Rectum 6 1.38 1.01 1.87 0.04
between hazardous waste in drinking water The brief article claimed that, with this Stomach 6 1.14 0.94 1.38 0.19
and cancer have been reported. Elevated process, "any municipal water supply can 'Number of studies evaluating specific cancer site.

226 Environmental Health Perspectives


DRINKING WATER AND CANCER

some systematic bias in the available studies, change its results. Nonetheless, these materials are used in these systems. Pipes
no specific bias has emerged to explain the apparent inconsistencies may reflect impor- can be made from metals, primarily iron,
observed results. tant differences in the carcinogenicity of copper and lead; plastics, such as polyvinyl
In summary, the available studies gen- the exposures experienced among the chloride and polyethylene; and concrete or
erally support the notion that by-products various study populations. The complex asbestos/concrete aggregates. These pipes
of chlorination are associated with mixture of compounds that comprise chlo- may be plated or lined with a variety of
increased cancer risks. The precise charac- rination by-products, the multiple path- compounds including zinc, coal tar,
terization of these risks is somewhat less ways of exposure to those compounds, and asphalt, or vinyl. In addition, bacteria and
clear. The broad category of chlorination the potential for synergy with diet and organic matter frequently coat the inside of
by-products includes many different com- other exposures may well explain the appar- pipes within the distribution systems (33).
pounds, and the carcinogens among these ent inconsistencies that exist among the All of these can be sources of new contami-
compounds have not been clearly identi- studies included in the metaanalysis. nation, or they can combine with chemicals
fied. Trihalomethanes are the most preva- To stop chlorination of drinking-water already in the water to alter the health risks
lent compounds and, given the evidence to eliminate the elevated cancer risks from posed by drinking water. In 1979, a study
suggesting that they are animal carcino- chlorination by-products would be fool- of several medium-size water systems
gens, have been the focus of research and hardy. Nonetheless, the data provide strong demonstrated increases in mutagenicity of
regulation. The chlorination by-products evidence to support expanded efforts in drinking water after passage through the
that have been specifically identified, how- research and development of alternatives to distribution system (34). This study did
ever, account for only about half of the chlorination for the disinfection of drink- not isolate specific contaminants that might
bound chlorine in finished drinking water. ing-water. Chlorination is particularly effec- be responsible. Perhaps the most extensively
Other compounds present in far smaller tive in preventing recontamination during studied contaminant associated with drink-
quantities may pose substantial cancer risks distribution. Alternatives must provide a ing water distribution is asbestos, which can
by virtue of high potency (25). similar level of protection. The capacity of leach from asbestos-concrete pipes. The
The goal of precise characterization of chemical disinfectants to kill pathogens gen- available research suggests that asbestos
the cancer risk posed by each of the chlori- erally reflects their strong tendency to react from this source does not pose significant
nation by-products will probably prove to with organic chemicals. The production of human cancer risks (35-37). A study by
be unrealistic. A quantitative dose-response by-products may, therefore, be inherent to Ashengrau et al. (38) showed an increase in
relationship has not been well described for the chemical disinfection of drinking water. leukemia in association with trichlorethyl-
any individual compound, much less the For example, ozone produces aldehydes ene, which had leached from a plastic liner
entire complex mixture. The relative con- including formaldehyde and bromate if used in concrete pipes. Other than the
tributions of different exposure pathways the source waters contain bromine. These negative results of the asbestos studies, the
vary among the by-products and have not compounds pose a cancer risk that is not available research does not allow for strong
been well characterized. Nonetheless, given yet fully quantified (29). Before the wide- conclusions concerning the magnitude of
the large number of people who consume spread introduction of any new method of cancer risks relating to contamination from
chlorinated surface water, the number of water treatment, the carcinogenicity of the distribution system. Further research is
cases of cancer potentially attributable to by-products should be carefully evaluated. needed to identify and quantify risks posed
this exposure is substantial. The numbers Of the other compounds routinely by contamination that occurs during
derived from the metaanalysis suggest that added during the course of drinking-water drinking-water distribution.
5000 (95% CI = 2000-7000) cases of blad- treatment, fluoride has received the greatest
der cancer per year and 8000 cases of rectal scrutiny as a potential carcinogen. The Emerging Concerns and
cancer per year (95% CI = 200-14,000) International Agency for Research on Potential Cancer Risks
may be associated with consumption of Cancer ([ARC) Working Group on Cancer Water is among the most basic require-
chlorinated drinking water. Although these Risks from Fluoridated Drinking Water has ments for human survival, therefore,
figures do not provide a precise estimate of concluded that available ecologic studies emerging health threats related to drinking-
risk, the true risk is probably within an have been consistent in finding no risk but water contamination demand careful con-
order of magnitude of these values. stopped short of suggesting that flouride sideration. Although the identification of
Since the publication of the meta- was not carcinogenic because the studies potential threats to human health requires a
analysis, a number of other studies have were all ecologic in design (30). One ani- certain degree of speculation, protection of
been completed. McGeehin et al. (26) mal study (31) and one case-control study public health requires a willingness to occa-
found an elevated risk for bladder cancer (32) suggested that fluoridated water could sionally err in the name of caution. Cancer
comparable in magnitude to the summary be linked with osteosarcoma, but these risks may emerge from the micropollutants
estimate of the metaanalysis. Kuovaslo et findings will require further confirmation and microbial contaminants that can enter
al. (27) found a similar estimate of risk for to be considered suggestive of causality. It our drinking-water supply. Less direct
bladder cancer but did not find an elevated appears that if flouride poses any cancer effects may also pose risks.
risk for rectal cancer. Kantor (28), on the risk, the attributable risk is relatively small. One focus of current concerns about
other hand, found a risk for rectal cancer the potential for micropollutants to cause
similar to that in the metaanalysis, but an Cancer Risks Associated with cancer involves those compounds that
increase in bladder cancer risk associated Drinking-Water Distribution mimic naturally occurring, biologically
with chlorination by-products was only The chemical components of pipes, joints, active compounds. Biologically active
observed among smokers. Including these and fixtures can contaminate drinking micropollutants or endocrine disrupters
findings within the metaanalysis does not water after treatment. A broad range of appear to have the ability to disturb normal

Volume 103, Supplement 8, November 1995 227


R.D. MORRIS

intercellular communications. For example,


evidence from wildlife biologists, toxi- - -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.-
cologists, endocrinologists, and epidemiol-
ogists demonstrate the potential for
estrogenic effects of environmental conta-
minants among humans (39,40). Meta-
bolites of DDT are estrogenic in vivo and
have been associated with the development
of breast cancer in epidemiologic studies
(41,42). Nonyl-phenol, a common chemi- ~~~~~~~~o-
cal surfactant, increases proliferation in
breast tumor cell cultures (43). The poten-
tial risks from drinking-water contami-
nants acting through these mechanisms
have not been evaluated.
Because of the complex mixture of con- 'Ic
taminants, examining cancer risks for each
individual compound may not give a com-
plete picture of cancer risks associated with
drinking water. An alternative approach is | Rate per 100,000
0.00-2091 2091-2783 _ 2783-3384
to look at the geographic distribution of 3384-4041 >4041
neoplasms that might be associated with
drinking water. These include cancer of the Figure 1. Rank sum map of incidence rates among persons 65 and older for cancers of the esophagus, stomach,
gastrointestinal tract and bladder cancer colon, rectum, and bladder from 1988 to 1989 (based on Appendix I).
(i.e., neoplasms of the mucosal epithe-
lium). Figure 1 provides maps showing
clustering of the incidence of site-specific
neoplasms among the elderly. By ranking
the incidence of the neoplasms of the
mucosal epithelium and combining those
ranks, we can see where this group of neo-
plasms might be elevated. A map of the
clustering of elevated cancer rates is shown
in Figure 2. This map indicates a
significant elevation of these cancers in the
northeastern United States. To draw con-
clusions about the link between the geo-
graphic distribution and drinking water
would, of course, be premature, but any
effort to explain this pattern should con-
sider drinking-water contamination to be a
possible contributing factor.
Microbial contaminants also have
carcinogenic potential. For example,
Schistosoma haematobium is waterborne,
although it is not transmitted by drinking-
water, and has been linked to cancer of the ,-- ----Not significant miSignificance and rate c med _ Significance and rate > med
urinary bladder (44). Tumor promotion by
algal toxins has already been suggested in Figure 2. Areas of significant clustering of elevated rates of mucosal neoplasms. Significance of Moran's l:
literature (45). Bacteria, parasites, and p<O.OOOl
viruses appear sporadically in most water
supplies. The possibility that currently
unidentified pathogens in drinking water Contamination of fishing grounds may deplete natural fisheries. A dietary shift
can cause cancer should not be overlooked. pose both direct and indirect cancer risks. from fish to red meat, either because of
Water pollution may pose cancer risks Persistent, potentially carcinogenic com- diminished fish stocks or fear of contami-
other than the direct, toxic effects of expo- pounds, such as polychlorinated biphenyls, nants, could also increase diet-associated
sure to contaminated water. Causal links accumulate in the fatty tissues of fish (46). cancer risks.
for the effects described below have not Fish consumption is a major exposure Under conditions of average tempera-
been clearly established, but they are plau- pathway for these compounds. In addition, ture, humidity, and activity, the human
sible and should be considered in evaluating contamination or destruction of spawning body loses and, therefore, must replace about
cancer risks from drinking water. grounds may combine with over-fishing to 2.3 liters of water each day. Two-thirds of

228 Environmental Health Perspectives


DRINKING WATER AND CANCER

this consumption is in the form of water or risks are not clear. Hazardous waste and Table 3. Carcinogenic risk associated with major
some other beverage. Concerns about the pesticides may contaminate waters locally contaminants of drinking water.
health risks or taste of drinking water may and regionally, but the attributable cancer Attributable risk Contaminant Point of Entry
induce those who consume tap water to risk is difficult to quantify. Nitrates are
shift to bottled water, or other beverages. more widespread contaminants and more Substantial Chlorination Treatment
by-products
These beverages may include sweetened soft closely linked to human cancer, but evi- Arsenic Source water
drinks and alcoholic beverages, which can dence is incomplete. Contamination dur- Small to Asbestos Source water,
pose health risks greater than those associ- ing drinking-water distribution may pose moderate distribution
ated with drinking water. In addition, the cancer risks, but the epidemiologic Radionuclides Source water
production and disposal of containers for evidence is extremely limited. Less conven- Little or none Fluoride Treatment
alternative beverages, including bottled tional cancer risk factors, such as biologi- Unknown Hazardous waste Source water
Pesticides Source water
cally active micro pollutants and pathogens,
water, may lead to the release of carcinogens. Tetrachloroethylene Distribution
only present the possibility of risk at Nitrates Source water
Summary and Prevention present but may emerge as important Pathogens Source water,
Strategies carcinogens in the future. distribution
The cancer risks associated with the major Cancer-prevention strategies must focus Biologically active Source water,
contaminants of drinking water are listed on source-water purity. In particular, micropollutants distribution (?)
in Table 3. The weight of the evidence strong source-water protection efforts pro-
suggests that chlorination by-products pose vide a barrier to emerging cancer risks that
substantial cancer risks that should be have not been identified or fully character- for monitoring drinking-water quality need
reduced. A growing body of evidence sup- ized. Furthermore, failure to protect source to be aggressively developed. Finally,
ports the possibility that arsenic in drink- water purity will necessitate more extensive drinking-water research should be a prior-
ing water may also carry unacceptable water treatment and, in most cases, heavier ity. The consequences of a lack of vigilance
cancer risks. The cancer risks from radon chlorination. Drinking-water treatment with respect to emerging threats in drink-
and asbestos in drinking water are less sub- technologies should be evaluated with ing water were felt with devastating impact
stantial but may require remediation where extreme care and should be reevaluated on in Milwaukee, Wisconsin, in 1993, when
local conditions dictate. The available a regular basis. The concept of continuous 400,000 people fell ill during a waterborne
evidence does not support assertions of quality improvement should be fully inte- outbreak of cryptosporidiosis (47). We
cancer risks associated with fluoridation of grated into drinking-water treatment and should view this as a warning and an
drinking water. should include ongoing efforts to develop, opportunity for timely intervention to
For most other compounds present in evaluate, and implement new treatment minimize health risks from drinking water.
drinking water, the attributable cancer technologies. More cost-effective methods

Appendix I surrounding areas. This can be expressed The mean and standard deviation of the
The map of cancer incidence rates for quantitatively for each analytic area i, as the
distribution of RSAC can be reasonably
mucosal cancer was based on the applica- weighted covariance of its disease rate (xi) approximated by a normal distribution
tion of the method described below. with the rates for the rest of the analytic with an expectation of zero and a standard
areas in the study region (xj) as given by deviation of a/(n-2)1/2 where a is the stan-
Assessing Cancer Incidence Rates dard deviation of xi and n is the number of
The incidence of cancer of the esophagus, wij (xi-X-)(xj-X-) [1] analytic areas.
stomach, colon, and urinary bladder for n-1 The RSAC was calculated for each ana-
persons over 64 years of age for the period lytic area, and the theoretical mean and
1988 through 1989 was estimated using where the weights (w,1) are the inverse of standard deviation were used to test for
Medicare hospital admissions data. The the distance between population centroids significance. Analytic areas that have
method used to estimate cancer incidence of the analytic areas (4). significantly high RSACs were further
with this database is reported elsewhere If the sizes of the study areas are not classified into two groups based on
(1-3). Briefly stated, all patients with a hos- homogeneous across the study regions, whether their disease rates were greater or
pital admission for cancer were identified. the weights corresponding to two adjacent less than the median rate. Analytic areas
Patients with no admissions for the site areas will vary according to the size of with significant RSACs and disease rates
specific cancer diagnosis in the previous 4 those areas. After modification to accom- greater than the median were defined as
years were considered to represent incident modate variations in region size, the analytic areas with clustering of elevated
cases. From these, age and sex adjusted, regional spatial autocorrelation coefficient disease rates or high clusters. These ana-
race-specific cancer rates were determined. (RSAC) for analytic area i, Ri, becomes lytic areas were shaded black in the map.
Analytic areas with significant RSACs and
Localizing Diseawe Clusters E(xi )2x(wii ) disease rates less than the median were
A disease cluster can be defined as a group Ri (xi-
= defined as analytic areas with clustering of
of geographic areas that are close to one -E _2(Wi_) low disease rates or low clusters. When
another with disease rates that are simi- the value of the RSAC was not significant,
larly increased or decreased relative to analytic areas were not shaded and

Volume 103, Supplement 8, November 1995 229


R.D. MORRIS

A B~~~~~~~~~~~~~~~~~~~~~~~
C D

Not significant _ Significance and rate < med Significance and rate > med

Figure Al. Regional spatial autocorrelation coefficient (RSAC) maps showing localized areas containing statistically significant disease clusters. (A) Malignant neoplasm of
the esophagus. (B) Malignant neoplasm of the bladder. (C) Malignant neoplasm of the colon. (D) Malignant neoplasm of the stomach. Significance of Moran's l: p<O.OOOl.

represented random spatial structures. Appendix References 137(2):226-234 (1993).


Maps depicting the results of these analy- 1. Morris RD, Munasinghe RL. 3. Whittle J, Steinberg EP. Accuracy of
ses (RSAC maps) were created to evaluate Aggregation of existing geographic Medicare claims data for estimation of
the use of this method as a visual aid to regions to diminish spurious variabil- cancer incidence and resection rates
ity of disease rates. Stat in Med among elderly Americans. Med Care
localize areas that contain disease clusters. 12:1915-1929 (1993). 29(12):1226-1236 (1991).
These methods are described in detail else- 2. McBean AM, Babish JD, Warren JL. 4. Munasinghe RL, Morris RD.
where (1,4). The resulting maps are Determination of lung cancer Localization of disease clusters using
shown in Figure Al. incidence in the elderly using regional measures of spatial autocorre-
Medicare claims data. Am J Epidemiol lation. Stat Med (in press).

REFERENCES
1. Chen CJ, Kuo TL, Wu MM. Arsenic and cancers (letter). Hertz-Picciotto I, Duggan HM, Woed R, Kosnett M, Smith
Lancet:1(8582)414-415 (1988). MT. Cancer risks from arsenic in drinking water. Environ
2. Wu MM, Kuo TL, Hwang YH, Chen CJ. Dose-response rela- Health Perspect 97:259-267 (1992).
tion between arsenic well water and mortality from cancer. Am 5. Kanarek MS, Conforti PM, Jackson LA, Cooper RC, Murchio
J Epidemiol 130:1123-1132 (1989). JC. Asbestos in drinking water and cancer incidence in the San
3. IARC. IARC Monographs on Evaluation of Carcinogenic Risks Francisco Bay area. Am J Epidemiol 112:54-72 (1980).
of Chemicals to Humans; Overall Evaluation of Carcinogenicity: 6. Polissar L, Severson RK, Boatman ES. A case-control study of
an Updating of IARC Monographs, Vols 1-42, Suppl 7. asbestos drinking water and cancer risk. Am J Epidemiol
Lyon:International Agency for Research on Cancer, 1987. 119:456-471 (1984).
4. Smith AH, Hopenhayn-Rich C, Bates MN, Goeden HM, 7. U.S. Department of Health and Human Services Working

230 Environmental Health Perspectives


DRINKING WATER AND CANCER

Group. Report on cancer risks associated with the ingestion of 28. Kantor KP, Lynch CF, Hildesheim M. Chlorinated drinking
asbestos. Environ Health Perspect 72:253-265 (1987). water and risk of bladder, colon, and rectal cancers: a case-
8. Neuberger JS. Residential radon exposure and lung cancer: an control study in Iowa, USA. Epidemiology 6:S30 (1995).
overview of published studies. Cancer Detect Prev 15:435-443 29. Bull RJ, Kopfler KC. Health Effects of Disinfectants and
(1991). Disinfectant By-products. Denver:American Water Works
9. Crawford-Brown DJ, Cothern CR. A Bayesian analysis or sci- Association, 1992.
entific judgment of uncertainties in estimating risk due to 30. IARC. Inorganic fluorides used in drinking water and dental
222Rn in U.S. public drinking-water supplies. Health Phys preparations. In: IARC Monographs on the Evaluation of
53:11-21 (1987). Carcinogenic Risks to Humans. Vol 27: Some Aromatic
10. Tao XG, Zhu HG. Yu SZ, Zhao ZY, Wang JR, Wu GD, You Amines, Anthraquinones and Nitroso Compounds, and
XF, Li C, Zhi WL, Bao JP, Sun ZQ. Effects of drinking water Inorganic Fluorides Used in Drinking Water and Dental
from the lower reaches of the Huangpu River on the risk of Preparations. Lyon:International Agency for Research on
male stomach and liver cancer death. Public Health Rev Cancer, 1982.
19:229-236 (1991-1992). 31. Bucher JR, Hejtmancik MR, Toft JD, Persing RL, Eustis SL,
11. Xu G, Song P, Reed PI. The relationship between gastric Haseman JK. Results and conclusions of the National
mucosal changes and nitrate intake via drinking water in a Toxicology Program's rodent carcinogenicity studies with
high-risk population for gastric cancer in Moping County, sodium fluoride. Int J Cancer 48:733-737 (1991).
China. Eur J Cancer Prev 1:437-443 (1992). 32. Cohn, P. Osteosarcoma and Chlorinated Drinking Water.
12. Tao XG, Zhu HG, Yu SZ, Zhao QY, Wang JR, Wu GD, You Trenton, NJ:New Jersey Department of Health, 1993.
XF, Li C, Zhi WL, Bao JP, Sun ZQ. Pilot study on the rela- 33. National Academy of Science. Biological quality of water in the
tionship between male stomach and liver cancer death and the distribution system. In: Drinking Water and Health. Vol 4.
mutagenicity of drinking water in the Huangpu River area. Washington:National Academy Press, 1982.
Public Health Rev 19:219-227 (1991-1992). 34. Schwartz DJ, Saxena J, Kopfler FC. Water distribution system,
13. Mirvish SS, Huang Q, Chen SC, Birt DF, Clark GW, Hinder a new source of mutagens in drinking waters. Environ Sci
RA, Amyrk TC, DeMeester TR. Metabolism of carcinogenic Technol 13:1138-1141 (1979).
nitrosamines in the rat and human esophagus and induction of 35. Harrington JM, Craun GF, Meigs JW, Landrigan PJ, Flannery
esophageal adenocarcinoma in rats. Endoscopy 25:627-631 JT, Woodhull RS. An investigation of the use of asbestos
(1993). cement pipe for public water supply and the incidence of gas-
14. Mirvish SS, Grandjean AC, Moller H, Fike S, Maynard T, trointestinal cancer in Connecticut, 1935-1973. Am J
Jones L, Rosinsky S, Nie G. N-Nitrosoproline excretion by Epidemiol 107:96-103 (1978).
rural Nebraskans drinking water of varied nitrate content. 36. Howe HL, Wolfgang PE, Burnett WS, Nasca PC, Youngblood
Cancer Epidemiol Biomarkers Prev 1:455-461 (1992). L. Cancer incidence following exposure to drinking water with
15. Leclerc H, Vincent P, Vandevenne P. Nitrates de l'eau de bois- asbestos leachate. Public Health Rep 104:251-256 (1989).
son et cancer. Ann Gastroenterol Hepatol 27:326-32 (1991). 37. Sadler TD, Rom WN, Lyon JL, Mason JO. The use of
16. National Research Council. Environmental Epidemiology. Vol asbestos-cement pipe for public water supply and the incidence
1: Public Health and Hazardous Wastes. Washington:National of cancer in selected communities in Utah. J Commun Health
Academy Press, 1991. 9:285-293 (1984).
17. Fagliano JM, Berry M, Bove F, Burke T. Drinking-water cont- 38. Aschengrau A, Ozonoff D, Paulu C, Coogan P, Vezina R,
amination and the incidence of leukemia: an ecologic study. Heeren T, Zhang Y. Cancer risk and tetrachforoethylene-cont-
AmJ Public Health 80:1209-1212 (1990). aminated drinking water in Massachusetts. Arch Environ
18. Griffith J, Duncan RC, Riggan WB, Pellom AC. Cancer mor- Health 48:284-297 (1993).
tality in US countries with hazardous waste sites and ground 39. Colborn T, Vom Saal FS, Soto AM. Developmental effects of
water pollution. Arch Environ Health 44:69-74 (1989). endocrine-disrupting chemicals in wildlife and humans.
19. Lagakos SW, Wessen BJ, Zelen M. An analysis of contami- Environ Health Perspect 101:378-384 (1993).
nated well water and health effects in Woburn, Massachusetts. 40. Davis DL, Bradlow HL, Wolff M, Woodruff T, Hoel G,
J Am Stat Assoc 81:583-596 (1986). Anton-Culver H. Medical hypothesis: xenoestrogens as pre-
20. MacMahon B. Comment. J Am Stat Assoc 81:597-599 (1986). ventable causes of breast cancer. Environ Health Perspect
21. Anonymous. New water purifier: chlorine used in reservoir that 101:372-377 (1993).
supplies Jersey City. New York Times, 7 December 1908; 16. 41. Falck F Jr, Ricci A Jr, Wolff MS, Godbold J, Deckers P.
22. Rook JJ. Formation of halogforms during chlorination of nat- Pesticides and polychlorinated biphenyl residues in human
ural waters. Soc Water Treat Exam 23:234-243 (1974). breast lipids and their relation to breast cancer. Arch Environ
23. Symons JM, Bellar TA, Carswell JK, DeMarco J, Kropp KL, Health 47:143-146 (1992).
Robeck GG, Seeger DR, Slocum CJ, Smith BL, Stevens AA. 42. Wolff MS, Toniolo PG, Lee EW, Rivera M, Dubin N. Blood
National organics reconnaissance survey for halogenated organ- levels of organochlorine residues and risk of breast cancer. J
ics. Am Water Works Assoc 67:634-646 (1975). Natl Cancer Inst 85:648-652 (1993).
24. Morris RD, Audet AM, Angelillo IF, Chalmers TC, Mosteller 43. Soto AM, Justicia H, Wray JW, Sonnenschein C. p-Nonyl-
F. Chlorination, chlorination by-products, and cancer: a meta- phenol: an estrogenic xenobiotic released from "modified"
analysis. Am J Public Health 82:955-963 (1992). polystyrene. Environ Health Perspect 92:167-173 (1991).
25. Kronberg L, Holmboom B, Tikkanen L. Identification of the 44. Gelfand M, Weinberg RW, Castle WM. Relation between car-
strong mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy- cinoma of the bladder and infestation with Schistosoma haema-
2(5H)-furanone and its geometric isomer E-2-chloro-3- tobium. Lancet 502:1249-1251 (1967).
(dichloromethyl)-4-oxy-butenoic acid in mutagenic fractions of 45. Falconer IR, Buckley TH. Tumor promotion by Microcystis sp.,
chlorine-treated humic water and in drinking water. In: Water a blue-green alga occurring in water supplies [letter]. Med J
Chlorination Chemistry, Environmental Impact and Health Aust 150:351 (1989).
Effects (Jolley RL, ed). Chelsea, MI:Lewis Publishers, 46. U.S. EPA. National Study of Chemical Residues in Fish.
1987;153-159. EPA/823-R-92-008b. Washington:US Environmental
26. McGeehin MA, ReifJS, Becher JC, Mangione EJ. Case-control Protection Agency, 1992.
study of bladder cancer and water disinfection methods in 47. Mackenzie WR, Hoxie NJ, Proctor ME, Gradus MS, Blair KA,
Colorado. Am J Epidemiol 138:492-501 (1993). Peterson DE, Kazmierczak JJ, Addiss DG, Fox KR, Rose JB,
27. Koivausalo M, Jaakkola JJ, Varitiainen T, Jakulinen T, Davis JP. A massive outbreak in Milwaukee of Crypotosporidium
Karjalainen S, Pukkala E, Tuomisto J. Drinking-water muta- infection transmitted through the public water supply. N Engl J
genicity and gastrointestinal and urinary tract cancers: an ecologi- Med 331:161-167 (1994).
cal study in Finland. Am J Public Health 84:1223-1228 (1994).

Volume 103, Supplement 8, November 1995 231