SUDEP: Overview of definitions and review of incidence data

Annegers JF and Coan SP (1999) SUDEP: Overview of definitions and review of incidence data. Seizure 8:6 347–52

Abstract: The classification, occurrence, and predictors of sudden unexpected and unexplained death in individuals with epilepsy (SUDEP) have received considerable attention over the last few years. Specific criteria for the classification of definite, probable, possible, and not SUDEP implemented in United States epidemiologic studies are presented. The incidence of SUDEP in different epilepsy populations is presented. SUDEP is a real phenomenon, because the occurrence of such deaths, especially at relatively young ages, among individuals with epilepsy is far greater (perhaps 40-fold) than among those without epilepsy. SUDEP incidence rates are lower in population-based studies, higher in referral populations and clinical trials of adjunct drugs for complex partial epilepsy, and highest for surgical series. Seizure severity appears to be the strongest risk factor for SUDEP because higher rates are reported from studies of individuals with intractable epilepsy. Other potential risk factors, including sex, seizure etiology, younger age at onset, and partial-onset seizures, are unresolved.

Article

Introduction

For over a century, clinical series of the proportionate mortality in individuals with epilepsy suggest an unusual frequency of sudden unexplained death in individuals with epilepsy. Interest in sudden unexpected unexplained death in epilepsy (SUDEP) in the United States was rekindled by Jay and Leestma during the early 1980s. It has become apparent that many such deaths among individuals with epilepsy were being ascribed to asphyxia, aspiration, status epilepticus, seizure, and cardiac arrest. All-cause mortality in epilepsy is about two to three times higher than in the general population, but the ratio is higher at younger ages (Table 1). The major causes for increased mortality rates in epilepsy are the causes of epilepsy, e.g. traumatic brain injury, cerebrovas- cular diseases, brain tumors, and accidents (mostly drowning) or, more rarely, of status epilepticus due to seizures and SUDEP (Fig. 1).

Materials and Methods

During the 1990s, the question of whether individuals in clinical trials of new antiepileptic drugs (AEDs) had an increased incidence of sudden death refocused attention on SUDEP. In 1993, the United States Food and Drug Administration (FDA) and Burroughs Wellcome established an advisory committee to develop criteria for SUDEP in clinical trials and compassionate use of lamotrigine:

1. The victim suffered from epilepsy, defined as recurrent unprovoked seizures.
2. The victim died unexpectedly while in a reasonable state of health.
3. The death occurred ‘suddenly’ (in minutes), when known.
4. The death occurred during normal activities (e.g. in or around bed, at home, at work) and benign circumstances.
5. An obvious medical cause of death was not found.
6. The death was not the direct result of the seizure or status epilepticus.

The criteria excluded accidental deaths secondary to seizures but included deaths with or without evidence of a seizure near the time of death. The panel did not preclude any specific circumstances of death from consideration for SUDEP, such as possible drowning (as suggested by Nashef), but reviewed and classified each death.

‘Definite SUDEP’ cases meet all criteria and have sufficient descriptions of the circumstances of the death and a postmortem report. ‘Probable SUDEP’ cases meet all criteria but lack postmortem data. ‘Possible SUDEP’ includes cases in which SUDEP cannot be ruled out but there is insufficient evidence regarding the circumstances of death and no postmortem report available. ‘Not SUDEP’ includes deaths for which other causes are clearly established or the circumstances make SUDEP highly improbable. Because only about 14% of all deaths in North America undergo a postmortem examination, working definitions for SUDEP must include ‘definite plus probable SUDEP’, and perhaps an upper range of rates may be calculated including even the possible cases. Possible SUDEP has two distinct meanings. (a) The information regarding the circumstances of the death is suggestive of SUDEP but is too limited to document with sufficient certainty as probable or definite SUDEP. (b) Adequate or complete documentation is available but there is a plausible competing explanation of the death, e.g. in cases found dead in the bath. Perhaps future criteria should consider these as separate classifications. Any classification system, however, must respect the need for a working definition that considers the variable nature of the documentation.

In studies applying these criteria, two or more panel members must classify each death as definite, probable, possible or not SUDEP, and discrepancies are adjudicated in a consensus conference.

Results from Studies

Rates of SUDEP

SUDEP rates from North American and British studies are presented in Table 2. Direct comparison of these studies is ill-advised because the study populations, levels of documentation, and SUDEP definitions vary considerably. Nevertheless, these studies show that SUDEP is a real phenomenon, because the occurrence of such deaths, especially at relatively young ages, among individuals with epilepsy is far greater than among those without epilepsy. As shown in Fig. 2, the lowest rate is from the only study of population-based incidence cases of all epilepsy, higher in referral populations and clinical trials of adjunct drugs for complex partial epilepsy, and highest for surgical series.

Population Rates of Sudden Death

There have been many United States studies of sudden death in the general population. Sudden unexpected death in the general population is extremely rare in young adults (<45 years) with an incidence of 5–10/100 000 person-years (Fig. 3) while the rates climb steeply with age and reach around 300/100 000 person-years in the elderly.

There were an estimated 251 000 sudden deaths in the United States in 1986. The incidence of sudden death has fallen from the 1950s and 1960s to the 1990s in parallel with incidence and mortality rates due to cardiovascular disease and cerebrovascular disease. The definition of sudden death in these studies differs from that of SUDEP because sudden deaths with post-mortem findings are not excluded. The major difference was the inclusion in the first recognized manifestation of ischemic heart disease of sudden death, even when the postmortem revealed advanced arteriosclerotic heart disease.

Relative Risks for SUDEP

SUDEP studies present the rates of occurrence, but do not provide the relative risk among individuals with epilepsy. Two studies estimate the relative risk for sudden death in individuals with and without epilepsy. Ficker et al. (Table 3), found the rate of SUDEP in the Rochester, Minnesota, epilepsy incidence cases to be 23.7 times that of the general population. In Allegheny County, Pennsylvania, there were 108 sudden deaths in the 14–21 age group of which 25 in people with epilepsy (overall rate 5.6/100 000 person-years), from 1972 to 1980 in individuals aged 14–21 years. Applying the prevalence of epilepsy of 7/100030, to this study population, the incidence of sudden death for individuals with epilepsy was estimated at 188.6/100000 compared with only 4.6/100 000 for the general population (Table 4). This produces a relative risk 40.2-fold higher for individuals with epilepsy compared with the general population.

Discussion

SUDEP mortality rates in the general population and those with intractable and non-intractable epilepsy are shown in Fig. 4. General population rates of sudden deaths climb logarithmically from around 5/100000 in young adults to about 300/100 000 in the elderly. The incidence of sudden death in young adults with intractable epilepsy is many times that of the general population, whereas in older age groups the relative increased incidence of SUDEP is small or perhaps nonexistent. The extent to which the incidence of SUDEP is elevated in people with remitting epilepsy is unknown.

Seizure severity appears to be the strongest risk factor because higher rates are reported from studies of individuals with intractable epilepsy. This may be partly due to the frequent proximate occurrence of seizure and SUDEP. The risk appears higher for acquired epilepsy, primarily from traumatic brain injury or encephalitis/meningitis than for idiopathic/cryptogenic epilepsy. Other potential risk factors, including male sex, younger age at onset, and partial-onset seizures, are unresolved.

SUDEP does not appear to be related to the use of any or specific AEDs. The higher rate of SUDEP in clinical trials of new drugs is presumably due to the high-risk patients eligible for such trials rather than to adverse AED effects. The age-specific rates and relative risks of SUDEP are not established. There may be an upper age limit at which the incidence of SUDEP does not differ from sudden death in the general population. Currently, several large case-control studies are being conducted to determine predictors of SUDEP among individuals with epilepsy.