Tobacco Documents Online

216 HERMAN F. KRAYBILL Five detailed procedures have been described elsewhere (3, 8) which, in essence, emphasize that conventional means of assessing the risk of individual constituents do not appear feasible and that none of the experimental data cited to date allow a reasonable quantitation of the health hazard associated with drinking water. The procedures now recommended for testing should provide some evidence of the efficacy of GAC (granular activated carbon) treatment of'drinking water to reduce organic carcinogenic contaminants and possibly reduce any contribution to the total cancer burden. There are about 12 reported epidemiologic studies.. Four hypotheses have emerged from them to date: namely, that the risk of cancer may be greater in the following situations: (a) surface water compared with ground water, (b) chlori- nated water compared with nonchlorinated water, (c) high levels of trihalomethanes (.THM) compared with low levels of THM, and (d) water from highly polluted surface water such as the Mississippi compared with surface water from less polluted waters. The THM studies were given greater weight by the National Academy of Sciences' Epidemiology Subcommittee of the Safe Drinking Water Committee (13). All except 1 of the 12 studies are based on an ecologic study. The THM studies suggested some association between this parameter and increased frequency of bladder cancer. The results do not establish causality and estimates of increased or decreased risk are extremely crude since additionally confounding factors, such as cigarette smoking, have not been ascertained. Moreover, a positive association with bladder cancer was small with a large mar- gin of error not just on a statistical basis but because of the nature of the study. All of the studies are limited by the absence of data on past exposures which would be quite relevant to cancer that has been diagnosed. Another deficiency was the possibility of identifying a stable population in areas where water quality was studied. None of the studies provide a critical appraisal of the correlation or regression coefficients obtained, in essence, their results demonstrate the prob- lem of establishing relationships between the aforementioned environmental vari- ables and health statistics. The subcommittee cautions, therefore, on the interpretation of the studies. Large populations would have to be studied to demon- strate any statistical associations because, for example, the concentrations of THMs are not large nor is there any current ideas as to magnitude of differences. SUMMARY, OR, IS THERE A HAZARD? Carcinogens and mutagens have been identified in drinking water. Data col- lected over the past few years have demonstrated that material obtained by con- centrating drinking water was mutagenic in Sahnonella and BALB 3T3 cells (I). Minimally concentrated drinking water samples (100- to 200-fold) treated with disinfectants such as O:~, HOCI, and CINH~ have been shown to produce mouse skin papillomas in two-stage "initiation promotion studies (3). While these findings were achieved with less than 1 mg of material applied, it is conceivable that the predominant risk might reside in the nonvolatile products and/or the products of disinfectant reactions with chemicals naturally occurring. Additive effects of car- cinogens cannot be overlooked and the fact that single chemical analysis in ro- dents indicates carcinogenicity raises suspicions that the multiple stress from carcinogens in water could contribute to the cancer load in man. However, the

CONFERENCE: PRIMARY PREVENTION OF CANCER 217 fact that some carcinogens from drinking water may persist in body tissues makes quantification difficult (10). The experimental investigations thus far provide only presumptive evidence and this situation may remain until better procedures are developed to measure the total carcinogenic insult of multiple components in water by oral administration in rodents or fish as animal models. A clearer under- standing of the actions these low-level insults have on DNA and target cells of liver, kidney, or bladder should provide more pertinent evidence of the car- cinogenic insult. Physiologic and biochemical pharmacokinetic studies should be useful in defining the potential.carcinogenic hazard. Quantification of risk assessment, while having some appeal, may be limited by certain factors such as (a) diverse mathematical models which yield extreme vari- ation in estimates of risk, (b) studies with different species sensitivities which can also yield a wide range in risk estimates, and (c) the failure to consider either additi~'e, synergistic, or even antagonistic effects that may underestimate the true risk. While the concentration of a single chemical carcinogen in drinking water will occur at microgram levels, the concentration of a composite of all carcinogens has been calculated to be at the milligram level, a dose capable of achieving a positive response in either animal models or in man. The NAS Subcommittee on Epidemiology of the Safe Drinking .Water Commit- tee considered the two types of epidemiologic studies. None of those reviewed could adequately .take into account many of the well-established risk factors. That is, even for bladder cancer it is necessary to control for occupation, cigarette smoking, use of alcohol and drugs, nonaqueous sources of trihalomethanes, coffee consumption, socioeconomic status, and ethnic origin. Another problem is that the trihalomethane levels in water were only inferred rather than actually mea- sured. Also, the studies had to assume that the present exposure to trihalomethanes reflected lifetime esposures of the study populations. These pre- liminary studies have provided adequate exploration and have generated hypoth- eses, but future studies will have to correct for previous deficiencies. While these experimental and epidemiologic studies provide some suggestive evidence on the potential hazard of these carcinogenic biorefractories, one cannot describe, with assurance, any causality for cancer but the probability remains that these ingested water contaminants may contribute to the total cancer burden. With this in mind it seems prudent to control and/or reduce the exposure to carcinogenic/mutagenic contaminants in drinking water. REFERENCES I. Berg, J. W., and Burbank, F. Correlations between carcinogenic trace metals in water supplies and cancer mortality. Attlt. N. )'. A¢'ad. St'i. 191,249 (1972). 2. Brown, E. R., Sinclair, T., Keith, L., Beamer, P., Hazdra, J. J., Hair, V., and Callaghan, O. Chemical pollutants in relation to diseases in fish. Proceedings of Conference on Aquatic Pollutants and Biological Effects with Emphasis on Neoplasia. Ann. N.Y. Acad. S~'i. 298, 535-547 ( 19771. 3. B.ull, R.J. "'Alternative Approaches to Assessing Hazard Associated with l~rinking Water and Efficacy of Granular Activated Carbon in Reducing This Risk." Unpublished report, Environ- mental Protection Agency, Cincinnati, Ohio ( 19791. 4. Correa, P., Haenszel,.W., Cuello, C., Tannenbaum, S., and Archer, M. A model for gastric cancer epidemiology, l.,n~'et 2, 58-60 (19751. 5. Dunham, L. J., O'Gara, R. W., and Taylor, F. B. Studies on pollutants from processed waters:

218 IIERMAN F, KR&YBII i collection from three stations and biologic testing for toxicity and carcinogenesis. Amer. J. I'uhl. Ih'alth 57, 2178- 2185 ( 19671. 6. Hueper, W.C. "'OcctJpational and Environmental Can~:ers of the Respiratory System." Springer-Verlag. New York, 1966. 7. Hueper. W. C.. and Ruchhoft. C. C. Carcinogenesis stadies on adsorbates of industrially polluted rilw and finished w;.iter supplies. Art'h. Iml, II3",~,~. Ot't'ltl,. ,'tie,/. 9. 488-495 (1954L 8, Kraybill, H. F. "Animal Models and Systems for Risk Evaluation of Low Level Carcinogenic Contaminants in Wa~ter." For presentation al Ihe Water Chlorination Environmental Impact i|lltl }-Ieldlh Effccts Confclcncc. ill Colorado Springs. Colo., Oclobcr 28-- November 2, 1979. 9. Kraybill, H. F., tlelmes. T. C., and Sigman. C. C. Biomedical aspects of biorefractories in water. io "'Aquatic Pollutants Transformation and Biological Effects". (O. Hutzinger. L. H. VanLeyl- veld, and B. C. J. Zoetem:m. Eds.). pp. 419-a,59. Pergamon Press. New York. 1978. 10. Laseter. J. L., and Dowty. B.J. Associ:tlitm of biorefraclories in drinking waler and body burden of people. Proceedings of Conference on Aquatic Pollulants and Biological Effects with Emphasis on Neopiasia. Am~. :%'. )'. Arad. SH, 298, 547-556 (1977). II. Loper. J. C.. t,ang. I). R.. Smith. C. C., Schoeny. R. S., Kopller. F. C.. and Tardiff. R. G. In vitro mutagenesis ~md carcinogenesis testing of residual organics in drinking water, in "'Aquatic Pollutants Transformution and Biological Effecls" (O. Hutzinger, L. H. VanLeylveld. and R. C. J. Zoeteman, Eds.). pp. 405-419. Pergamon Press. New York. 1978. 12. National Academy of Sciences. Advisory Cenler on Toxicology, Assembly of Life Sciences, National Research Council Safe Drinking Water Committee. "Drinking Waler and Heallh.'" NAS pt~blicalion. Washington, D.C. (19771. 13. National Academy of Sciences, National Research Cotmcil Safe Drinking Water Subcommittee on Epidemiology. "'Epidemiological Studies of Cuncer Frequency and Certain Organic Con- stiluents of Drinking Water: A Review of Recent Literature Published :rod Unpublished." Report for the Environmental Protection Agency. Washington, D.C. (1978). 14. New York Academy of Sciences. "Annals of Conference on Aquatic Pollutants and Biological Effects wilh Emphasis on Neoplasia'" (H. F. Kraybill. R. G. Tardiff, C. J. Dawe, and J. C. Harshbarger. Eds.)," Vol 298. NYAS publication, New York, 1977. 15. Pojasek, R. B. How to protect drinking water sources. Em'iroo. Sci. "l'e~'hnol. I I. 342-347 (1977). 16. Tseng, W. F. Prevalence of skin cancer in an endemic area of chronic arsenicism in Taiwan. ,/. N~tt. Cao~'er Itt.~t. 40, 453 (1968).