The relationship of brain infarction to diabetes and alcoholism was studied in a 10-year follow-up of a defined general population (n = 159,200) of native, urban Swedes. Confounding bias due to high age and lower socioeconomic conditions was reduced with a new epidemiological technique which was used instead of conventional mathematical multivariate procedures. We observed 6-13-fold and 4-6-fold excess rates of subjects with brain infarction among diabetics (P less than 0.001) and alcoholics (P less than 0.001), respectively. In addition, diabetes and alcoholism were often found to be associated. The distribution of these diseases varied with the topography of the city and demography of the population. Hence, in future studies into the pathogenesis of brain infarction, study samples should be made homogeneous not only from a clinical, but also from an epidemiological point of view. We conclude that diabetes, alcoholism and both in combination, associate with brain infarction.