Page 365 - ILAE_Lectures_2015

than to the epilepsy itself. The lowest SMRs are reported in the 75+ age group; this relates in part
to the fact that this age group in the population has a high mortality rate.

Few studies have looked at the risk of mortality after a single seizure. Two retrospective studies
have investigated mortality after a single unprovoked seizure; one provided an SMR of 2.3 (95%
confidence interval [CI] 1.5, 3.3)5 and the other an SMR of 1.1 (95% CI 0.1, 4.0) for single
idiopathic seizures6. In the NGPSE, people with an acute symptomatic seizure (provoked seizure)
had an SMR of 3.2 (CI 2.4, 4.3) after more than 20 years of follow up7. Overall the SMR in
people with a newly diagnosed unprovoked seizure ranges from 2.5 to 4.1 with the highest rates
in children and those with symptomatic aetiology8.

Reported SMRs in mortality studies from developed countries range from 1.6 to 4.19. In the
Rochester study5, the SMR for the total group after 29 years’ follow-up was 2.3, with the most
significant increase in the first 10 years. In the NGPSE the SMR was 2.5 after median 6.9 years
with the highest SMR in the first year (5.1)10. The SMR further decreased to 2.1 after 11 to 14
years of follow up11, and remained stable but persistently elevated after a median follow-up of
22.8 years7. The highest SMRs were estimated in people with remote symptomatic epilepsy
(SMR 3.7; 95% CI 3.1, 4.6) and epilepsy due to a congenital neurological deficit (SMR 19; 95%
CI 7.0, 49.7)7, which remained elevated throughout follow-up. In contrast people with
idiopathic/cryptogenic epilepsy (defined as aetiology not determined) did not have a significantly
increased long-term mortality rate (SMR 1.3; 95% CI 0.9, 1.9)11 in the initial stages of follow-
up, a finding that has been replicated in community-based studies in Iceland6 and France12. The
French study, which examined the short-term mortality in people with epilepsy, is the only study
to have used the categories idiopathic and cryptogenic epilepsy as defined by the ILAE, with no
significant differences between the two groups. Interestingly, mortality was significantly elevated
in people with idiopathic/cryptogenic epilepsy in the NGPSE during the last 10 years of follow-
up7.

Data on premature mortality in people with epilepsy is more limited from resource-poor
countries. Findings from a prevalent cohort study with a follow-up of 6.1 years gave an overall
SMR of 2.9 (95% CI 2.6, 3.4)13, although the SMR reported earlier, after the first 25 months of
follow-up, was higher (SMR 3.9, 95% CI 3.8, 3.9)14. A much higher SMR was found in young
people (aged 10–29 years) (SMRs 28 to 37). Death from drowning was a significant risk (overall
SMR 39; 95% CI 26.4, 55.5), but was more critical for people living in a waterside area than for
those living in the mountains (HR 3.9; 95% CI 1.7, 9.2, P = 0.002)14. A similar prevalent cohort
with a follow-up of median 28 months found an overall SMR of 4.9 (95% CI 4.0, 6.1), with
higher SMRs in young people15.

Overall, people with epilepsy have been found to have a reduction of life expectancy which is
greatest at the time of diagnosis. This reduction can be up to two years in people with
idiopathic/cryptogenic epilepsy and up to 10 years in people with symptomatic epilepsy16.

Mortality in population-based studies is summarised in Table 1.

Causes of death

Causes of death in people with epilepsy can be divided into epilepsy-related and non-epilepsy-
related deaths. For people with symptomatic epilepsy (both remote and progressive) the excess
mortality risk relates primarily to the underlying cause of the epilepsy rather than to the epilepsy
itself. In a study of 692 children with epilepsy followed up over an average of 13 years, the SMR