Epidemiologic studies have suggested a causal association between exposure to chlorinated drinking water and risk of cancer, particularly cancer of the urinary bladder.1-19 However, as a result of limitations in measuring exposure and estimating outcome, the evidence is not conclusive.
Our objective was to address some of these identified limitations in a study of whether exposure to disinfection byproducts in chlorinated drinking water (measured as trihalomethane concentration) altered the frequency of micronucleated urinary bladder epithelial cells in humans. Trihalomethanes are used as the exposure measure in this study because they are routinely monitored by most water utilities and often used as an indicator of health effects. Exposure was determined at an individual level, adjusting for variations in ingestion, inhalation, and dermal exposure. The need for long-term follow up was overcome by using micronuclei, a preclinical biomarker of genotoxicity.
A cohort study was conducted in 1997 in 3 communities in Australia. Bungendore, a town in New South Wales, had an unchlorinated water supply. Canberra in the Australian Capital Territory and the northwestern suburbs of Adelaide in South Australia had varying levels of trihalomethane in the water supply.
Study Subjects and Data Collection
We sent recruitment letters to all 412 households connected to the community water supply in Bungendore. In Canberra and Adelaide, the local water utilities identified 27 routinely monitored sampling sites for which average total trihalomethane for the preceding year exceeded 110 μg/L. We randomly selected 25 households within a 1-km radius of each of these sampling points and sent them letters. Informed consent was obtained from eligible persons willing to participate in the study.
Eligibility was restricted to men aged 30 to 65 years who had resided at their current address for at least 6 months and had never been diagnosed with cancer other than skin cancer. Regular swimmers in chlorinated pools were not eligible to participate.
We obtained measures of each participant’s available dose and intake dose. Available dose was the average concentration of 4 trihalomethane readings taken over the 2-week study period from the water supply of each participant’s home without adjustment for individual variation in intake. SA Water (Adelaide) and Ecowise Environmental (Canberra), Australia, undertook the trihalomethane assays using headspace gas chromatography with electron capture detection method (Teckmar Head Space Autosampler 7000, Ohio, connected to a Hewlett Packard 5890 Series II Gas Chromatograph, Wilmington, DE, with Electron Capture Detector). The detection limit of this method was 1 μg/L.
Intake dose was estimated by adjusting available dose for individual variations in ingestion, inhalation, and dermal exposure. We collected data on participants’ water ingestion over the 2-week study period using a fluid intake diary. Oral dose was estimated as recommended by Jo et al.20 Inhalation exposure was estimated based on the same experimental study models. Because shower air concentration was not measured in this study, we estimated chloroform using data provided by Jo et al. using modeling techniques. We obtained individual showering or bathing time during the 2-week study period. This information was incorporated when calculating dermal dose according to the method recommended by Jo et al.20 Intake dose was the sum of ingestion, inhalation, and dermal doses.
The study period was defined as 2 weeks, based on evidence that micronuclei levels respond to changes in exposure within 7 to 10 days.21
The Outcome Measure
The outcome of DNA damage to bladder cells was estimated using the prevalence of micronuclei in exfoliated bladder epithelial cells, reported as the number of micronuclei per 1000 normal (not degenerated) cells. Micronuclei have been shown to increase in frequency with exposure to carcinogens, and the assay has been used to assess the risk of cancer.21-24 At the end of the 2-week study period, we collected the entire second and third voids of urine for the day. Scoring of the prepared cells was done by the School of Public Health, University of California, Berkeley, in accordance with their published protocol.25
We administered a questionnaire by telephone to all participants to identify relevant family and medical history and exposure to other known bladder cancer risk factors. Venous blood samples were obtained for the determination of plasma vitamin B12 and folate levels, because these affect cell integrity.
The prevalence of micronuclei in exfoliated bladder epithelial cells was assessed separately for available and intake doses of individual trihalomethane compounds and total trihalomethane. Random-effect Poisson regression models were fitted adjusting for significant interaction effects and confounding variables. In the models examining the 4 trihalomethane compounds, for compounds that were correlated with coefficients of 0.80 or above, 1 of the correlated compounds was chosen for inclusion in the model. Potentially confounding factors found to alter the relative risk estimate by more than 10% were included in the multivariate models. Relative risks were estimated for DNA damage to bladder cells for each increase in exposure of 10 μg/L.
Approval for this study was obtained from the Australian National University Human Research Ethics Committee.
Altogether, 1087 recruitment letters were mailed. Telephone contact was established with 884 households, and 529 persons were identified as eligible to participate. Of these, 353 persons (66.7%) were recruited to the study, 348 (65.8%) completed the study, and 228 (85 unexposed and 143 exposed) had slides suitable for scoring. Slides for the remaining 120 persons were found to be unsuitable for scoring because of insufficient cells, or because of debris or bacteria covering the cells.
The average age of participants with micronuclei scores was 47 years. Adelaide participants were older (mean age, 52 y) than participants from Bungendore and Canberra (mean age, 46 y for both).
Bungendore had a higher proportion of smokers (32%) than Canberra (15%) and Adelaide (22%). Smoking was not associated with the frequency of micronuclei; for univariate analysis the relative risk was 1.04 (95% confidence interval [CI] = 0.74-1.45) and when controlled for disinfection byproduct exposure the relative risk was 0.97 (95% CI = 0.63-1.50). Based on existing support for the effect of age and smoking on frequency of micronuclei, these variables were included as potential confounders in the multiple regression models. Other factors were not found to be confounders in this study.
Trihalomethanes were not detected in the water supply of the unexposed (Bungendore) group. The levels of trihalomethane in the water supply for the exposed group were all within Australian Drinking Water Guidelines (Table 1). 26 Adelaide had higher levels of trihalomethane than Canberra, except for chloroform, which was found to be higher in Canberra.
When adjusted for individual variations in intake, the intake dose for chloroform was higher in Canberra (Table 1). The other 3 compounds and total trihalomethane were higher in Adelaide.
The prevalence of micronuclei in bladder epithelial cells is shown in Table 2. The proportions of abnormal cells did not differ among the 3 communities. The median frequency of micronuclei (not adjusted for potential confounders) was higher for the unexposed community, with lowest levels observed in the highest exposed community.
Association Between Exposure and Outcome
When adjusted for potential confounders and stratified by smoking status, the relative risk estimates for the associations between exposure indices and outcome were approximately 1.0 (Table 3).
Variables used in the multivariate models were categorized and examined for dose-response relations (Table 4), recognizing that this does not take into account exposure to complex mixtures. The risk was reduced marginally for the higher concentrations of chloroform, whereas the risk increased for bromoform when compared with those not exposed. However, all confidence intervals included 1.0 and overall no obvious dose-response patterns were observed.
Unlike most previous studies, exposure in this study was determined at an individual level, adjusting for variations in ingestion, inhalation, and dermal exposure, with information collected on known potentially confounding factors. The need for long-term follow up was overcome by using micronuclei, a preclinical biomarker of genotoxicity. Micronuclei are markers of acute exposure, responding to changes in exposure within 7 days. Therefore, exposure could be measured prospectively at an individual level. Micronuclei prevalence has been used to assess genotoxicity of potential carcinogens such as arsenic, benzene, radiation, and smoking.22,23,27-32 The appearance of micronuclei is a preclinical stage in the pathway to cancer development and is therefore a biologically plausible outcome measure to assess the carcinogenic potential of a substance. However, micronuclei might not be sufficiently sensitive to serve as an indicator of carcinogenicity in a relatively small study. The small size of this study could also explain the lack of significant association between smoking and the frequency of micronuclei. The higher prevalence of micronuclei in our unexposed community might have been the result of the higher prevalence of cigarette smoking in this community compared with the other groups.
Despite methodologic advances in previous studies, this study failed to demonstrate an association between trihalomethane concentration in drinking water and DNA damage to bladder epithelial cells. The null finding from this study supports previous evidence that trihalomethanes (chloroform in particular) are not genotoxic.33-36
Water and urine analysis was carried out by Ecowise Environmental (Canberra, Australia) and SA Water (Adelaide, Australia). Scoring of exfoliated bladder cells for micronuclei was undertaken at the School of Public Health, University of California, Berkeley, California, under the supervision of Nina Titenko-Holland. We thank Malcolm Mearns for assistance in data collection, Robyn Attewell for assistance with statistical analysis, and Erin O’Neil for assistance in cell preparation.
1.Cantor KP, Hoover R, Mason TJ, et al. Associations of cancer mortality with halomethanes in drinking water. J Natl Cancer Inst
2.Wilkins JR, Comstock G. Source of drinking water at home and site-specific cancer incidence in Washington County, Maryland. Am J Epidemiol
3.Isacson P, Bean JA, Splinter R, et al. Drinking water and cancer incidence in Iowa. Am J Epidemiol
4.Koivusalo M, Jaakkola JJ, Vartiainen T, et al. Drinking water mutagenicity and gastrointestinal and urinary tract cancers: an ecological study in Finland. Am J Public Health
5.Yang C-Y, Chiu H-F, Cheng M-F, et al. Chlorination of drinking water and cancer mortality in Taiwan. Environ Res
. 1998;78(section A):1–6.
6.Gottlieb MS, Carr JK, Morris DT. Cancer and drinking water in Louisiana: colon and rectum. Int J Epidemiol
7.Gottlieb MS, Carr JK, Clarkson JR. Drinking water and cancer in Louisiana: a retrospective mortality study. Am J Epidemiol
8.Young TB, Kanarek MS, Tslatis AA. Epidemiological study of drinking water chlorination and Wisconsin female cancer mortality. J Natl Cancer Inst
9.Kannio A, Ridanpaa M, Koskinen H, et al. A molecular and epidemiological study on bladder cancer: p53 mutations, tobacco smoking, and occupational exposure to asbestos. Cancer Epidemiol Biomarkers Prev
10.Zierler S, Danley RA, Feingold L. Types of disinfectant in drinking water and patterns of mortality in Massachusetts. Environ Health Perspect
11.Cantor KP, Hoover R, Hartage P, et al. Bladder cancer, drinking water source, and tap water consumption: a case-control study. J Natl Cancer Inst
12.McGeehin MA, Reif JS, Becher JC, et al. Case-control study of bladder cancer and water disinfection methods in Colorado. Am J Epidemiol
13.King WD, Marrett LD. Case-control study of bladder cancer and chlorination by-products in treated water (Ontario, Canada). Cancer Causes Control
14.Freedman DM, Cantor KP, Lee NL, et al. Bladder cancer and drinking water: a population-based case-control study in Washington County, Maryland (United States). Cancer Causes Control
15.Cantor KP, Lynch CF, Hildesheim ME, et al. Drinking water source and chlorination by-products—I. Risk of bladder cancer. Epidemiology
16.Koivusalo M, Hakulinen T, Vartiainen T, et al. Drinking water mutagenicity and urinary tract cancers: a population-base case-control study in Finland. Am J Epidemiol
17.Koivusalo M, Pukkala E, Vartiainen T, et al. Drinking water chlorination and cancer—a historical cohort study in Finland. Cancer Causes Control
18.Neutra RR, Ostro B. An evaluation of the role of epidemiology in assessing current and future disinfection technologies for drinking water. Sci Total Environ
19.Morris RD, Audet A-M, Angelillo IF, et al. Chlorination, chlorination by-products, and cancer: a meta-analysis. Am J Public Health
20.Jo WK, Weisal CP, Lioy PJ. Chloroform exposure and the health risk associated with multiple uses of chlorinated tap water. Risk Anal
21.Rosin MP. The use of the micronucleus test on exfoliated cells to identify anti-clastogenic action in humans: a biological marker for the efficacy of chemopreventative agents. Mutat Res
22.Smith AH, Hopenhayn-Rich C, Warner M, et al. Rationale for selecting exfoliated bladder cell micronuclei
as potential biomarkers for arsenic genotoxicity. J Toxicol Environ Health
23.Cossman J. Molecular Genetics in Cancer Diagnosis
. New York: Elsevier Science Publishing Co Inc; 1990.
24.Warner ML, Moore LE, Smith MT, et al. Increased micronuclei
in exfoliated bladder cells of individuals who chronically ingest arseniccontaminated water in Nevada. Cancer Epidemiol Biomarkers Prev
25.Titenko-Holland N, Moore LE, Smith MT. Measurement and characterisation of micronuclei
in exfoliated human cells by fluorescence in-situ hybridization with a centrometric probe. Mutat Res
26.NHMRC (National Health and Medical Research Council). Australian Drinking Water Guidelines.
Canberra, Australia: Commonwealth of Australia; 1996.
27.Biggs ML, Kalman DA, Moore LE, et al. Relationship of urinary arsenic to intake estimates and a biomarker of effect, bladder cell micronuclei
. Mutat Res
28.Da Cruz AD, McArthur AG, Silva CC, et al. Human micronucleus counts are correlated with age, smoking, and cesium-137 dose in the Goiania (Brazil) radiological accident. Mutat Res
29.Gonsebatt ME, Vega L, Salazar AM, et al. Cytogenetic effects in human exposure to arsenic. Mutat Res
30.Holmen A, Karlsson A, Bratt I, et al. Increased frequencies of micronuclei
in T8 lymphocytes of smokers. Mutat Res
31.Moore LE, Warner ML, Smith AH, et al. Use of the fluorescent micronucleus assay to detect the genotoxic effects of radiation and arsenic exposure in exfoliated human epithelial cells. Environ Mol Mutagen
32.Moore LE, Smith AH, Hopenhayn-Rich C, et al. Micronuclei
in exfoliated bladder cells among individuals chronically exposed to arsenic in drinking water. Cancer Epidemiol Biomarkers Prev
33.Bull RJ, Birnbaum LS, Cantor KP, et al. Water chlorination: essential process or cancer hazard? (Symposium overview). Fundam Appl Toxicol
34.Golden RJ, Holm SE, Robinson DE, et al. Chloroform mode of action: implications for cancer risk assessment. Regul Toxicol Pharmacol
35.Melnick RL, Kohn MC, Portier CJ. Implications for risk assessment of suggested non-genotoxic mechanisms of chemical carcinogenesis. Environ Health Perspect
. 1996;104(suppl 1):123–134.
36.World Health Organization. International Programme on Chemical Safety (IPCS), Environmental Health Criteria 216: Disinfectants and Disinfectants By-products.
Geneva: WHO; 1997.