The interference of geographical changes of cancer risk in tumor etiology in Japan.

An attempt was made to elucidate the relation between various human neoplasias in their geneses by comparatively investigating the geographical changes of cancer risk either between 2 sexes with one neoplasia or between two neoplasias for each sex with the data of multiple local populations of Japan as well as of the world. Specific cancer death rates for 11 sites in 46 prefectures of Japan, 1969-1971, were used in the domestic study, and age-adjusted cancer incidence rates for the corresponding 11 sites in 40 selected cancer registration areas of the world were used in the international study. The relation between 2 cancer risks was analyzed by calculating a 1st order regression equation using log-transformed data sets. Results obtained are as follows: 1) both the domestic and international data sets confirmed positive correlations of a high degree between the male and the female cancer risks in 6 out of 7 cancer sites tested (esophagus, stomach, colon, rectum, liver and bladder). A positive intersex correlation was also observed in lung cancer with the international data, but not with the domestic one. It was indicated that the male and female populations of Japan as well as of the world were under the influence of a common risk modifier in the course of geographical changes of cancer risk for each of 6 neoplasias except lung cancer. 2) Intercancer correlation was lacking in 11 out of 23 cancer pairs in the domestic study, whereas all 22 intercancer correlations but one in the international study were found to be statistically significant. Further insight into the correlation data of the international study led to the conclusion that the observed combination of positive and negative correlation in the 22 cancer pairs were explainable in terms of the differential cancer risk changes of Western type and Non-Western type cancers due to the transition of life style between the Western extreme and the Non-Western extreme in the world. In contrast, a high proportion of non-correlations (11 out of 23 cancer pairs) and emergence of opposite or new correlations (3 out of 23 pairs) were observed in the domestic study, a finding which indicates that the domestic cancer risk modifier, as evidenced in the intersex correlation test, is site-specific, and is to be differentiated from the Westernization-sensitive cancer risk modifiers of the international study.(ABSTRACT TRUNCATED AT 400 WORDS)