Geographical variation of cancer mortality in Italy.

Abstract

The distribution of death certification rates from various cancers or groups of cancers in broad Italian geographical areas (north/centre/south) was analysed. In both sexes, total cancer mortality was considerably elevated in the north of the country compared to southern regions (around 70% for males and 30% for females in the truncated 35-64 rate), and generally intermediate in central areas. Northern mortality rates were higher for respiratory cancers and other tobacco related neoplasms (excluding bladder), with a north/south ratio ranging from 1.5 for lung and most respiratory sites to about 4.0 for oesophageal cancer in males. There was little tendency towards a leveling of these differences in younger (40-49 year old) males. Northern areas showed higher death certification rates for cancers of the stomach, large bowel, liver and most other digestive sites. The lower gastric cancer mortality registered in southern Italy is curious, since this is the poorest part of the country. Death certification rates from all other common neoplasms (uterus apart) were also elevated in the north. The geographical variation, however, appeared more limited for non-epithelial neoplasms. The substantial differences in cancer mortality between various Italian geographical areas can hardly be dismissed as due to lower death certification accuracy in the south. Some of the differences can be explained in terms of available knowledge of the causes of cancer (eg reproductive factors for breast and ovarian neoplasms, alcohol plus tobacco for oesophageal cancer). However, the lower mortality from respiratory cancers in southern areas can only with some difficulty be totally explained in terms of tobacco consumption. Likewise, the north/south variation cannot be related to non-specific consequences of industrialization, since cancer mortality was similarly elevated in highly industrialized and chiefly rural northern areas. It is conceivable that dietary factors may also explain some of the differences. However, at present, there is no obvious general explanation for this quite peculiar geographical distribution of cancer mortality within a single country.