The relationship of thyroid autoantibodies and elevated TSH level to indices of cardiovascular diseases was studied in two population series monitored for 5 years and in a cross-sectional hospital series. In a cohort of 1105 males, initially 55-74 years of age, deaths due to cardiovascular causes occurred in 19% of subjects with thyroid autoantibodies and in 11% of controls matched for age (P less than 0.05). In another cohort of 1045 males and 1223 females, initially 40-64 years of age, no difference emerged in males, while 6 out of 20 females who died of cardiovascular causes had thyroid autoantibodies, compared with 18% in the whole series. In a series of 97 hospital patients with myocardial infarction, 7 patients had thyroid autoantibodies as opposed to 12 antibody-positive subjects among controls matched for age and sex. Elevated TSH level appeared to be no better an indicator of cardiovascular morbidity or mortality than thyroid autoantibodies. It is concluded that thyroid autoimmunity may act as a cardiovascular risk factor under certain circumstances, but it does not have any general significance and the mechanism of action remains unclear.