Life expectancy in epilepsy

Does a diagnosis of epilepsy reduce a person’s life expectancy? According to Athanasios Gaitatzis and colleagues the answer is yes, but under certain circumstances and to a variable extent. These authors recently estimated life expectancy in people with epilepsy, with data from the prospective community-based UK National General Practice Study of Epilepsy (NGPSE), and made comparisons with the general population. Life expectancy can be reduced by up to 10 years when there is a known cause of the epilepsy, the estimated reduction being highest at the time of diagnosis. These observations are hardly surprising considering the wealth of literature that shows increased mortality rates in people with epilepsy.

Traditionally, mortality has been expressed as the ratio of the observed and expected numbers of death: the standardised mortality ratio. Expected deaths are calculated by applying the death rates of an external reference population to the age distribution of the study population. The epilepsy population has an standardised mortality ratio of 2-3 (ie, a mortality that is 2-3 times higher than that of the general population). People with epilepsy of unknown cause have at most only a slight increase in mortality, while those with epilepsy as a symptom of a known underlying cause largely account for the overall increased mortality. The increase is evident during the first years after the onset of epilepsy, and mortality then declines to levels close to those in the general population. However, some studies show an increase in mortality later, a decade or more after disease onset. Relative survivorship, defined as the proportion of observed to expected number of survivors, provides a different perspective of mortality. The relative survivorship at 5, 10, and 15 years after diagnosis was 91%, 85% and 83%, respectively, in a pioneering study. In epilepsy of unknown cause relative survivorship is as high as 96% 25 years after diagnosis.

Gaitatzis and colleagues introduce yet another way of analysing data on mortality in patients with epilepsy. The estimated life expectancy of people with new-onset non-febrile seizures from the UK followed up for a median of 15 years was compared with that in people of the same age and sex in the general population. The difference in life expectancy between these populations gives the estimated years of life lost, which is presented at different intervals after the diagnostic seizure–for men and women, young and old, and epilepsy of known and unknown cause. With this analysis, the authors have expressed earlier knowledge in a new form, which serves to refine our understanding of an important prognostic aspect of epilepsy.

The critical question is what the reasons might be for the reduced life expectancy in people with new-onset epilepsy. Gaitatzis’ data, as indeed other population-based studies, strongly suggest that the increased mortality is related to the underlying cause of epilepsy rather than to the seizures. One may therefore ask, for example, how different life expectancy would be for a person with poststroke epilepsy compared with someone who has had a stroke but not developed epilepsy. Indeed, such patients might constitute the most relevant control group to clarify the contribution of the epilepsy itself. In fact, the reported causes of death in the NGPSE cohort, mainly cancer, ischaemic heart disease, cerebrovascular disease, and pneumonia, also indicate that mortality as a consequence of seizures is rare in newly diagnosed patients.

Although the estimates of life expectancy are of considerable interest for researchers and physicians, they are probably less useful in counselling patients and relatives. First, it is difficult to generalise from these data because Gaitatzis and colleagues included patients with single seizures and acute symptomatic seizures, while excluding those with brain tumours. Second, epilepsy is a heterogeneous disorder with multiple causes, each affecting the prognosis differently. Because of the limited size of the cohort, patients with different causes were lumped into one group: symptomatic epilepsy. However, it does not make much sense to estimate life expectancy for an individual with epilepsy after a traumatic brain injury on the basis of data from patients with underlying cerebrovascular disease. Third, the concept of years of lost life might be difficult to communicate to patients. A person with idiopathic/cryptogenic epilepsy would probably consider an estimated reduction in life expectancy of up to 2 years highly significant rather than minimal.

Although this extended analysis of the NGPSE data provides a new and interesting perspective on mortality risks, we now need to find ways to estimate the contribution of the epilepsy and seizures themselves and the causes of such deaths, and to analyse to what extent treatment can minimise risks and reduce years of lost life for people with epilepsy. Finally, we should acknowledge that the available data on mortality in epilepsy derive almost exclusively from western industrialised countries. We lack information from the rest of the world where the vast majority of the global epilepsy population resides and where epilepsy-related mortality and life expectancy is likely to be different.

Source: Torbjörn Tomson, Lars Forsgren

The Lancet
March 15, 2005

Original web page at The Lancet