A case-control study of asbestos in drinking water and cancer risk

The authors conducted a case-control, interview-based study of the risk of developing cancer from asbestos in drinking water. The Everett, Washington area was selected for the study because of the unusually high concentration of chrysotile asbestos in the drinking water it draws from the Sultan River (200 X 10(6) fibers/liter). Through a population-based tumor registry, the authors identified 382 individuals with cancer of the buccal cavity, pharynx, respiratory system, digestive system, bladder, or kidney, diagnosed between 1977 and 1980, and then interviewed them or their next-of-kin. The authors conducted similar interviews of a control group of 462 individuals. Finally, interviews were validated in several ways, including comparing the collected data with that from secondary sources. Estimates of exposure to asbestos in drinking water were based on residence and workplace history, and on individual water consumption. Four different measures of exposure were used. Cancer risk was estimated by logistic regression and other methods. The authors found no convincing evidence for cancer risk from imbibed asbestos. Exposure was similar between cases and controls. Confidence intervals for relative odds for almost all sites included unity. Out of 84 dependent estimates of risk by sex, site, and exposure measure, 63 indicated a protective effect and 21 indicated an increased risk. In instances where relative odds differed appreciably from unity for both males and females, the effect was protective. The relative odds of cancer for 20 years of exposure to Sultan River drinking water varied from 0.92 to 0.99 for females and 0.82 to 1.01 for males for all study sites grouped. For cancer of the digestive system, the corresponding range was 1.03 and 1.08 for females and 0.85 to 1.00 for males. There were six statistically significant associations (p less than 0.05). All involved male stomach (eight cases) and male pharynx (four cases), and indicate elevated risk. This number of significant associations is close to that expected (for the number of comparisons made). The female risks for these two sites indicate a protective effect, and it is concluded that the male results are probably due to chance.