We studied regional patterns of thyroid cancer incidence in relation to iodine intake and iodinization in Sweden using 5,838 incident cases diagnosed in the period 1958-1981. Region was defined either by iodine status, urban-rural status or health-care region (internal controls). Age, period and cohort models were fitted to evaluate trends and regional variation in incidence by histopathologic type. In iodine-deficient areas, the relative risk (RR) of developing thyroid cancer was 0.92 for all histologic types combined, 0.80 for papillary cancer and 0.87 for anaplastic carcinoma. Residence in iodine-deficient regions was associated with a 2-fold increased risk of follicular cancer in men (RR 1.98) and a 17% increase in risk in women (RR 1.17). Regional differences in iodine intake fell after iodinization of the food supply, which was started in 1936 and enhanced in 1966. Nevertheless, incidence of both papillary and follicular carcinoma increased during the study period, with largely similar trends in iodine-deficient and iodine-sufficient areas. Overall, residence in urban or rural areas was not an important determinant of incidence, though trends in the incidence of papillary, follicular and anaplastic cancer did vary between urban and rural areas. The occurrence of thyroid cancer differed only marginally between the 6 health-care regions in Sweden, suggesting that the observations in regions defined by iodine intake were unlikely to be artifactual. Our data suggest that iodinization of the food supply is not associated with adverse trends in the occurrence of thyroid cancer.