Several hypothesis-generating ecologic studies have been based on cancer incidence data provided by the Iowa cancer registry for 1969 to 1978. The data were taken from municipalities that had populations of at least 1000 and public water supplied by a single source that had remained stable for a minimum of 14 years. Rates for cancers of the lung and rectum among males and females were higher for municipalities using chlorinated surface water compared to those with groundwater sources (Bean et al., 1982a). Subsequently Isacson et al. (1985) studied cancer incidence data for the years 1969-1981 from the Iowa cancer registry for towns with a public water supply from a single stable ground source and the levels of volatile organic compounds and metals found in the finished drinking water of these towns in 1979. This study lessened one of the problems common to ecologic investigations, namely, misclassification of exposure. Early studies of cancer and drinking water associated consumption of water and rates of cancer by comparing the proportion of county or parish residents supplied by surface water sources with overall cancer mortality rates for the total area. In these early studies, the lack of data on individual consumption patterns hampered interpretation of the results. Isacson et al. (1985) used cancer incidence by municipality, along with a survey on drinking water habits, in order to reduce misclassification of exposure. Associations between 1,2dichloroethane and the incidence of cancers of the colon and rectum and between nickel and cancers of the bladder and lung were most clearly seen in males. Although nickel, 1,2-dichloroethane and trichloroethane are known to be carcinogens, the levels of these materials found in this study were well below the nondetectable-response level estimated from the experimental literature on these compounds. The researchers concluded that one plausible explanation of their result is that the mere presence of these industrial effluents in groundwater indicates that exposures have occurred from anthropogenic sources. Thus, the measured substances do not necessarily account for the increased rate of cancer, but rather indicate the presence of other materials.
In an earlier report from the Iowa cancer registry, Bean et al. (1982b) considered the contribution of waterborne radioactivity to differences in cancer incidence. They found that incidence rates of cancers of the lung and bladder among males and of cancers of the breast and lung among females were higher in towns where the water supply contained more than 5.0 picocuries per liter of radium-226. A gradient of increasing incidence associated with rising radioactivity levels for