To evaluate temporal patterns of cause-specific mortality after long-term exposure to the alpha-particle-emitting contrast medium Thorotrast, we investigated a cohort consisting of 693 Swedish patients with neurological disorders who received Thorotrast during cerebral angiography, with follow-up ending in 1993. Standardized mortality ratios (SMRs) were calculated as the ratio of observed cases in the cohort to expected cases in the general population. Persons exposed to Thorotrast had significant excesses of all causes of death (SMR = 2.8; 95% CI 2.5-3.0), with similar increases noted for men and women. The largest risks were observed for deaths from hematological causes (SMR = 16.4; n = 8), cerebrovascular diseases (SMR = 10.1; n = 18), gastrointestinal disorders including liver cirrhosis (SMR = 5.2; n = 36), and tumors (SMR = 4.7; n = 187). There was a significant decrease in SMR with time since injection for cerebrovascular and circulatory diseases, indicative of the impact of underlying neurological disorders. In contrast, the SMR increased significantly with time for tumors and gastrointestinal diseases, suggestive of a detrimental effect of cumulative radiation dose. A significant dose-response relationship was found for all causes of death and malignant tumors among all age groups, and since SMR increased with time for the latter category, this is consistent with an effect of cumulative radiation exposure on cancer development. However, the findings should be treated with caution, since selection bias may have influenced some of the results.