SUDEP is the sudden, unexpected death of someone with epilepsy, but who was otherwise healthy, and for whom no other cause of death can be found. Because of its nature, the cause of death in SUDEP is unknown, but researchers are beginning to develop a better understanding.
In 1868 Bacon, an eminent physician, noted the occurrence of ‘sudden death in a fit’ and almost 40 years later Spratling, one of the earliest American neurologists, recognised epilepsy as ‘a disease which destroys life suddenly and without warning through a single brief attack.’ Despite this, in the 1960’s it was suggested that ‘there is no reason why …someone with epilepsy… should not live as long as he would if he did not have epilepsy’ (Livingstone 1963). SUDEP is sudden unexpected death in someone with epilepsy, who was otherwise well, and in whom no other cause for death can be found, despite thorough post mortem examination and blood tests. The definition excludes people dying in status epilepticus and those who drown.
Awareness of SUDEP has increased over recent years, yet in many countries the medical profession has been reluctant to consider SUDEP. Indeed, there is little information on the number of cases in different countries. It has been estimated that the risk of sudden death is almost 24 times higher than for someone without epilepsy. Most people with newly diagnosed epilepsy will stop having seizures, and SUDEP is very rare amongst them. Searching for risk factors in this group would require meticulous follow up of large numbers of people. Studies of SUDEP have therefore usually been conducted in groups of people with more severe forms of epilepsy, such as specialist clinic populations, hospital inpatients or residential groups. The risk of SUDEP is elevated in these populations. It is estimated as between 1:500 and 1:1000 patient-years in community based populations with epilepsy, and even higher in people considered for surgery.
The cause of death in SUDEP is currently, by definition, unknown, but various risk factors have been suggested. These include young adulthood, presence of convulsive attacks, poor seizure control and poor adherence with antiepileptic drugs (AEDs). Other suggested risk factors are male gender, use of more than one AED, frequent changes of dose or type of AED, alcohol abuse and certain epilepsy syndromes.
Studies of age at death in SUDEP can be problematic because its definition requires negative findings at post mortem examination. Many elderly people have evidence of vascular disease and it is often difficult to exclude this as a cause of death. Most studies have found young adults as the group at higher risk. The age of onset of seizures is lower in people with epilepsy who died of SUDEP than in those dying of other causes.
SUDEP is usually unwitnessed but, when witnessed, often follows a convulsion. Evidence for a seizure prior to death is frequently, but not always, found at post mortem examination. Studies have found that most cases of SUDEP in whom the seizure type was known, had a history of convulsions. Many people dying of SUDEP are found in or near the bed. A recent study found that supervision at night appears to be protective.
Higher rates of SUDEP are found in studies of people with severe epilepsy, suggesting that people with frequent, severe seizures are most at risk. Some authors have suggested that seizure frequency is not a risk factor but a recent study found it to be the most significant risk factor.
There is conflicting evidence concerning the use of more than one AED as a risk factor. One study which showed increased risk associated with multiple AEDs suggested this may be because these people have more severe seizures, while another found the increased risk associated with more than two AEDs was still significant after adjusting for seizure frequency. Other studies have not found any association.
People have debated whether SUDEP is more likely in men than women, but a recent study found the rate of SUDEP to be identical in men and women.
A number of studies suggest that SUDEP is largely a peri-ictal event. These include case-control studies and accounts of witnessed deaths and detailed circumstances of death. They also include incidence studies showing an association with less well controlled epilepsy, particularly where there is a history of generalised tonic clonic seizures.
Risk factors identified so far, however, only tell part of the story. Sometimes individuals with infrequent seizures die, while others with more frequent and apparently more severe seizures do not. Some may be more at risk because of social factors, lifestyle or sub-optimal management; others may have additional biological susceptibility.
Of particular interest are risk factors we can influence or prevent, such as tonic clonic seizures. One study also identified polytherapy and frequent treatment changes as independent risk factors. While these risk factors may still be surrogate markers for epilepsy severity, theoretically, both could make SUDEP more likely. Medication changes, prescribed or otherwise, particularly if abrupt, could result in instability, while (poly)therapy could, theoretically, affect post ictal respiratory depression and/or cardiac autonomic function. Of note, however, is that one study found untreated epilepsy to be a risk factor, with those who had never had drug therapy at increased risk.
Another consistent observation is that most deaths occur in bed, presumably from sleep. Again, are nocturnal seizures different pathophysiologically – for example, associated with increased vagal tone or respiratory depression – or is it simply that assistance is not to hand, as Delasiauve wrote a century and a half ago? There is some evidence to suggest that he was at least partly right. A study looking at position in death found 71% prone and only 4% supine. Another case control study showed decreased risk associated with using listening devices or sharing the room with someone capable of giving assistance. Respiratory compromise frequently occurs in more severe seizures. Central and obstructive apnoea, excess bronchial and oral secretions, pulmonary oedema and hypoxia are all well documented. That assistance in the community provides some protection favours a role for respiratory factors which, to some extent, can be influenced by relatively unskilled intervention such as airway protection, re-positioning or stimulation.
Primary or secondary cardiac mechanisms are also likely to be important. One can postulate a number of hypotheses, which are not mutually exclusive:
That malignant cardiac tachy/brady arrhythmias or cardiac failure occur secondary to the ictal discharge and/or apnoea/hypoxia. A study of long-term cardiac rhythm monitoring using an implanted device resulted in 4/19 patients with intractable epilepsy being paced because of recorded periods of asystole, thought to have occurred in a very small proportion of seizures reported.
That long-term cardiac changes secondary to uncontrolled epilepsy may occur. Pathological studies provide limited support for this.
That, theoretically, the same underlying process which causes epilepsy could predispose to sudden cardiac death in a proportion of people with epilepsy. For example, ion channel disorders are implicated in genetic epilepsies and in conditions predisposing to sudden cardiac death; many are co-expressed in heart and brain.
That some individuals with epilepsy may have a co-existing unrelated ‘mild’ genetic susceptibility to sudden death, through cardiac conduction/ion channel disorders, which then manifests because of uncontrolled seizures. Advances in the studies of complex minor genetic susceptibility may further research into SUDEP.
Strategies to reduce SUDEP include focusing on seizure prevention, optimising medical and surgical management and assistance during seizures, aided by future advances in seizure prediction/detection. Further research into mechanisms exploring the above hypotheses may also lead to specific interventions. An analogy between SIDS and SUDEP is often drawn. Success in preventing some potential SIDS cases followed advice on optimal positioning of infants. More recently individual susceptibility in a small proportion of cases (‘SIDS’ genes which include channelopathies) is being addressed. How important these turn out to be remains to be seen. Many lessons can be learnt from the SIDS story.
A systematic review was performed to provide evidence-based estimates about risk factors and incidence of SUDEP (Tellez-Zenteno, Ronquillo & Wiebe 2005). We searched Index Medicus, Medline, EMBASE, and the Cochrane database, for retrospective or prospective cohort and case-control English language studies exploring the risk factors and incidence of SUDEP in adults and children from 1966 to 2003. Of 83 potentially eligible articles 36 fulfilled eligibility criteria (29 cohort and 8 case-control studies). Salient findings follow:
Researchers do not use standard definitions of SUDEP. Standard definitions are essential for meaningful communication about clinical conditions. Some authors did not state a definition, some crafted their own, and others adhered to published definitions. A definition of SUDEP was explicit in 65% of studies, not clear in 29%, and not given in 6%. In many studies, the diagnosis of SUDEP was probable, not definite, because the most used definitions require post mortem data, and autopsies are performed infrequently in many settings. This suggests that the requirement of post mortem examination for a definitive SUDEP diagnosis may require revision. Specifically, how much certainty and precision does the post mortem examination add to the diagnosis of SUDEP?
The risk factors for SUDEP depend on the type of comparison. Studies exploring risk factors use one of two main comparison groups, i.e. non-SUDEP epilepsy deaths, and live people with epilepsy (PWE). Comparisons with non-SUDEP deaths explore best the circumstances surrounding death (e.g. seizures preceding death, place of death, AED levels at the time of death). Comparisons with live PWE explore best the lifestyle and clinical variables related to SUDEP (e.g. frequency of seizures, number of AEDs, use of other drugs). Therefore, the seemingly disparate risk factors found in these studies are really complementary.
High-risk and low-risk groups for SUDEP are identified. The risk of SUDEP is expressed as number of deaths per 1000 person-years. The risk is highest in studies of candidates for epilepsy surgery and epilepsy referral centres (2.2 to 10 per 1000), intermediate in studies including patients with mental retardation (3.4 to 3.6 per 1000), and lowest in children (0 to 0.2 per 1000). The incidence was similar in autopsy series (0.35 to 2.5 per 1000) and in studies of epilepsy patients in the general population (0 to 1.35 per 1000). PWE in the high-risk group typically suffer from more severe epilepsy, have frequent seizures, and require many AEDs. These factors are consistently associated with SUDEP (Langan & Nashef 2003; Lhatoo & Sander 2002; Stollberger & Finsterer 2004). The reasons for a lower risk in children require further investigation.
The contribution of SUDEP to overall mortality varies by risk group. As expected, in higher-risk groups SUDEP is a more frequent cause of death than in lower-risk groups. In studies of children, general population, epilepsy registers and autopsy series, SUDEP explained from 0 to 14% of deaths. In studies of epilepsy clinics, drug trials, epilepsy surgery or surgical candidates, and registers of patients with refractory epilepsy SUDEP explained from 29 to 75% of deaths.
Several aspects of SUDEP require further research. Prospective studies from seizure monitoring units could be fruitful. Standardisation of case ascertainment, definitions, and description of population sources is necessary to improve the analysis and interpretation of data. International panels could review the SUDEP definition and scientific journals could encourage researchers to adhere to standard definitions. Finally, researchers need to assess the impact that learning about SUDEP has on patients and their families, while exploring optimum risk communication and coping strategies for this infrequent, but devastating event.