Anticonvulsant drugs and cancer. A cohort study in patients with severe epilepsy.

Over 2000 epileptic patients admitted to the Chalfont Centre for Epilepsy between 1931 and 1971 and taking anticonvulsants were followed up to the end of 1977. Mortality between 1951 and 1977 was greatly in excess of that in the general population of England and Wales in that period allowing for age and sex. Some of the excess was directly attributable to epilepsy, but there were also more deaths from suicide and circulatory, respiratory, and malignant disease than would be expected. Apart from the brain and central nervous system, no particular site had a significant excess of tumours. In particular, there were no liver tumours (and only one gallbladder carcinoma). This makes it unlikely that the liver tumours produced on feeding phenobarbitone to mice are indicators of major human risk.