Mortality and cancer incidence following cerebral angiography with thorotrast

Abstract

Epidemiological studies are important tools in assessing radiation-related health effects in humans. The most common form of radiation exposure is low-dose alpha-radiation such as that from radon. However, the long-term health effects of low-dose radiation are incompletely known. Studies of patients exposed to Thorotrast provide valuable information, since the administration of this radioactive contrast medium resulted in lifelong exposure to alpha particles at a low dose rate. Thorotrast was used worldwide for visualization of vascular structures from about 1930 to the beginning of 1950. It proved to be one of the most carcinogenic substances ever utilized in medical practice, causing very high rates of liver cancer. In Sweden, 1,115 patients were exposed to Thorotrast through cerebral angiography performed between 1932-48. The principal objective of these studies was to evaluate associated health effects.

Incident cancers (paper IV and V) and site-specific causes of death (paper I-II) compared to the general population were assessed through the Swedish Cancer and Cause of Death Registers (SCR and SCDR) for the exposed patients alive when these registers were started. A comparison group of 1,058 sex- and age-matched non-exposed patients admitted to the same medical departments during the same period of time was selected for Cox regression analyses of the influence of time and volume on mortality and cancer incidence (paper II and V). The influence of underlying diseases, gender, time since injection, volume of Thorotrast, occupation, and lifestyle factors were investigated (paper V). Cohorts of Thorotrast-exposed in Sweden, Denmark, and the United States were combined to perform a pooled analysis of mortality among 1,736 exposed patients (paper III).

Patients exposed to Thorotrast had a significant three-fold increase in overall mortality compared to the general population (standardized mortality ratio=SMR 2.8), with similar excesses for men and women. The largest risks were observed for deaths from benign haematological diseases, cerebrovascular disorders, liver cirrhosis, and malignant and benign tumors. Trends with time and injected volume of Thorotrast suggested an effect of cumulative radiation exposure on tumor development, gastrointestinal diseases including liver cirrhosis, and circulatory diseases.

A total of 170 cancers occurring in 152 individuals were reported, whereas only 57 cases were expected in the general population. The standardized incidence ratio (SIR) was increased three-fold for cancer at all sites, with the largest excesses noted for primary liver and gallbladder cancer (SIR 39.2). Significantly increased risks were also seen for liver cancer not specified as primary, and cancers of the small intestine, stomach, kidney, pancreas, as well as leukemia. The association with time and volume showed that cumulative radiation was directly related to carcinogenesis in the liver and gallbladder, and possibly to other solid tumors and leukemia.

Reducing selection and information bias by comparing the Thorotrast-exposed patients with a selected control group resulted in a reduction of the excesses in non-malignant mortality. An association between continuous radiation exposure and leukemia, cancers of the liver and gallbladder, and liver cirrhosis remained. Gender, occupation, smoking, and alcohol consumption did not influence the results.

Improving statistical power in the pooled analysis gave similar results with significantly elevated all-cause mortality (relative risk=RR 1.7). Cancer, benign blood, and digestive diseases showed the highest radiation-related increases. Our study stresses the importance of reducing the influence of bias by using a comparable control group when severely ill individuals like the Thorotrast-exposed are studied.