Page 364 - ILAE_Lectures_2015

Chapter 37

The mortality of epilepsy

AIDAN NELIGAN1 and GAIL S. BELL2

1Homerton University Hospital Foundation Trust, London, and 2UCL Institute of Neurology,
Queen Square, London, and Epilepsy Society, Chalfont St Peter, Buckinghamshire
____________________________________________________________________________

It has been consistently shown in population studies that the risk of premature death is two to
three times higher in people with epilepsy than in the general population. This mortality risk is
highest in the early years following diagnosis. The risk is stratified by aetiology with people with
remote symptomatic epilepsy and neurological deficits having persistently higher risks. Other
factors of relevance have been gender, age, a previous episode of status epilepticus, frequency
and severity of seizures and antiepileptic drug (AED) compliance.

Long-term population-based prospective incident cohort studies provide the most reliable means
of examining the risk of premature mortality and the way it changes over the course of the
condition1, although there are very few studies with follow-up of more than 20 years.

The estimates of the risk of premature death have varied between studies, and case ascertainment
can be an issue depending on the methodology used. Mortality studies in epilepsy should be
community-based studies of incident cohorts. Studies of people with prevalent epilepsy may
underestimate the short-term mortality (as the mortality in people with epilepsy has consistently
been shown to be highest in the early years following diagnosis) while simultaneously
overestimating the long-term mortality (as those who have gone into remission may not be
included in the cohort)2. The risk of premature death in people with epilepsy has been studied
using death certificates, hospital or institutional records and through follow-up of community
cohorts. Death certificates have been shown to be an unreliable source, with epilepsy being
recorded on the death certificate in only 7% of patients known to have had seizures3. This figure
increased to 17% in those with frequent seizures. In a community-based study of mortality in
children with epilepsy, epilepsy was recorded on the death certificate in 55% of deaths directly
attributable to epilepsy4.

The most commonly reported measures of mortality in epilepsy studies comparing deaths
between the study and a control population are the proportional mortality ratio (PMR) and the
standardised mortality ratio (SMR). The PMR gives the proportion of deaths caused by a specific
cause in the cohort and compares it with a control group. This is not a direct measure of mortality
but rather gives the proportion of deaths due to one specific cause and can be influenced by the
rates of other causes of death. The SMR is the ratio of the observed deaths in the study population
to the expected deaths if the group had experienced the same age and sex-specific death rates as
the population from which they came.

What is the risk and who is at risk?

Studies have consistently shown that males with epilepsy have higher mortality rates, with no
clear explanation for this difference. The SMR tends to be high in children but this relates
principally to the underlying cause of the epilepsy (remote symptomatic, perinatal insults) rather