Background: People with epilepsy may experience an altered pattern of mortality due to cancer. We determined cancer mortality in two distinct populations with epilepsy: (i) an institutionalised cohort with severe epilepsy (SEC) and (ii) another cohort with presumed milder epilepsy (MEC).
Methods: Mortality data from 1977 until 2003 were obtained for the two cohorts (SEC, n=1358 and MEC, n=4494). Standardised mortality ratios (SMRs) for all causes, all cancers and specific cancer sites were calculated.
Findings: The SMR for all cancers was significantly elevated in the SEC (SMR: 1.42; 95%CI: 1.18-1.69) but not in the MEC (SMR: 0.93; 95%CI: 0.84-1.03). The SMR for brain and CNS neoplasms was significantly elevated in the MEC; most cancer deaths in younger ages (<45 years) in this cohort were due to brain and CNS neoplasms. There was also a deficit in deaths due to lung cancer (SMR: 0.68; 95%CI: 0.54-0.84) in this cohort. In the SEC, SMRs for oesophageal (SMR: 2.81, 95%CI: 1.35-5.17), liver (SMR: 4.12, 95%CI: 1.12-10.56) and gall bladder (SMR: 7.32, 95%CI: 1.51-21.38) cancers and non-Hodgkin's lymphoma (SMR: 2.67, 95%CI: 0.98-5.80) were elevated.
Conclusions: The modest excess of all cancer deaths and large excesses of deaths due to cancers at certain sites in the SEC and the lack thereof in the MEC suggests dissimilarity in the risk factor profile of the two populations. These risk factors, whether life style attributes or antiepileptic drugs, need to be identified.