Many people with epilepsy (PWE) receive their care from non-specialist physicians or neurologists and neuro-paediatricians without specific epilepsy expertise. Every physician caring for PWE needs to know the risks associated with epilepsy in order to provide appropriate counselling about how to reduce those risks. Reduced life expectancy is a major concern. Epilepsy is associated with a two to three-fold mortality increase when compared to the general population (Wicks and Fountain, 2012). This increased risk is partly due to co-morbidities and the aetiology underlying the epilepsy, and partly due to seizures (Duncan et al., 2006; Schuele et al., 2007).
Among the major causes of death related to epilepsy, which include accidents and status epilepticus, the most common is sudden unexpected death in epilepsy (SUDEP) (Sander and Bell, 2004). There is a clear need to develop strategies to identify those at increased risk and improve the management of PWE, with the aim of reducing premature mortality, in particular SUDEP.
Further, as epilepsy is a chronic condition, PWE who have been in remission may experience relapses making them vulnerable to seizure-related death. This is also true for those who have had successful epilepsy surgery. It has been shown that the presence of pre-operative generalised seizures is a risk factor for late recurrence, thus a risk factor for SUDEP (Schwartz et al., 2006).
Although there are currently no established evidence-based prevention strategies (Maguire et al., 2016; McLean et al., 2016), this seminar will summarize available evidence and suggest how PWE may be counselled about SUDEP.
When is a death classified as SUDEP?
SUDEP is the sudden and unexpected death of a PWE when complete autopsy and toxicology does not identify a cause of death. Based on this definition, SUDEP is a diagnosis of exclusion (box 1). Definite SUDEP requires an autopsy to confirm no anatomical or toxicological cause. Probable SUDEP is applied when autopsy is not performed but the circumstances of death are otherwise very suggestive, and possible SUDEP is when autopsy is not performed and there is a potential competing cause of death. SUDEP Plus is used for cases which would otherwise fulfil the definition of SUDEP in the presence of another condition which could have contributed to the death (e.g. coronary insufficiency with no evidence of myocardial infarction or long-QT syndrome with no documented terminal primary ventricular arrhythmia). Near-SUDEP includes cases in which cardiorespiratory arrest is reversed by resuscitation efforts with subsequent survival for more than one hour (Nashef et al., 2012).
Generalised tonic-clonic seizures (GTCS) convey the greatest risk of SUDEP, though there are exceptions and SUDEP can rarely occur in the absence of a history of tonic-clonic seizures in patients presenting focal seizures with impaired awareness (Sperling et al., 1999; Langan et al., 2000) and in the absence of terminal seizures (Lhatoo et al., 2016).
Epidemiology of SUDEP
In the UK, SUDEP accounts for approximately 500 of the 1,200 epilepsy-related deaths a year (Shankar et al., 2013). It has been more difficult to estimate the number of SUDEPs in the USA. A large population-based study of PWE (Ficker et al., 1998) reported an estimated incidence of SUDEP of 0.35 cases per 1,000 person-years of follow-up while another reported 2.7 per 1,000 person-years (Leestma et al., 1989). Other studies (Neuspiel and Kuller,1985; Donner et al., 2001; Hitiris et al., 2007) in North America have shown differences possibly due to variation in patient selection criteria and methods of study and analysis (Tomson et al., 2016).
The risk of SUDEP is about 1 in 10,000 person-years in population-based studies of newly diagnosed epilepsy and 1 in 1,000 person-years in people with chronic epilepsy. SUDEP risk, however, increases with less well controlled epilepsy to 1 in 200-300 person-years in cohorts seen in specialist centres and up to almost 1 in 100 person-years in those with severe treatment-resistant epilepsy, being particularly high among those with uncontrolled tonic-clonic seizures (Tomson et al., 2005).SUDEP risk is in general lower in children, at around 0.2 per 1,000 children with epilepsy (Donner et al., 2001; Ackers, 2011; Berg et al., 2013), but no studies have evaluated SUDEP rates in subpopulations of children with epilepsy. In the US, the public health burden of SUDEP in terms of life-years lost is estimated to be second only to stroke among neurological conditions (Devinsky et al., 2016). The cumulative risk of SUDEP in a population-based follow-up study of 40 years was estimated to be 7-12% (Sillanpää and Shinnar, 2010).
The majority of reported witnessed SUDEP cases occurred in conjunction with a GTCS. Current evidence, drawn from deaths occurring in epilepsy monitoring units, suggests that the final terminal pathway involves severe compromise of centrally mediated cardiopulmonary function as a result of the seizure, with terminal apnoea usually preceding asystole (Ryvlin et al., 2013). That approximately 80% of all SUDEP cases appear to be unwitnessed suggests that witnessed seizures are less likely to be fatal (Nashef et al., 1998). It is, however, not known why one specific seizure is fatal when the PWE may have previously had numerous apparently similar non-fatal seizures, nor why only some individuals with treatment-resistant epilepsy die of SUDEP. While some differences may relate to modifiable environmental or treatment-related factors, there may also be genetic susceptibility (Goldman et al., 2016; Nashef and Sander, 2016).
Most SUDEPs are unwitnessed, with the deceased often found in bed, more often in the prone position than would be expected by chance (Kloster and Engelskjon, 1999), and with evidence suggestive of a seizure in the majority of cases (Nashef et al., 1998).
Modifiable risk factors for SUDEP
Case-control studies have identified clinical factors associated with increased SUDEP risk among PWE (Tomson et al., 2016).By far the most important clinical risk factor is frequency of GTCS, but nocturnal seizures, early age at epilepsy onset (before the age of 16 years), male gender, and long duration of epilepsy (over 15 years) have been identified as additional risk factors. Lack of antiepileptic drug treatment has also been associated with increased SUDEP risk, while the presence of someone capable of providing assistance at night was reported as protective in one case control study (Langan et al., 2005).
Preventive measures are directed at potentially modifiable risk factors although there have been no studies to show that modifications of risk factors result in a reduction in SUDEP incidence. Such studies would face major methodological and ethical challenges (Tomson et al., 2016). Despite the lack of direct evidence, it is likely that there are measures that can be taken to reduce the risk of SUDEP.
As stated, the most established risk factor for SUDEP is uncontrolled tonic-clonic seizures (Tomson et al., 2008 ; Surges et al., 2009; Tomson et al., 2016; Shankar et al., 2016).If the seizure frequency is higher, the risk of SUDEP is greater. Compared to being free from GTCS, having 1-2 GTCS per year was associated with a five-fold increased risk of SUDEP, and three or more GTCS per year a 15-fold increase (Hesdorffer et al., 2011). It is therefore imperative that ongoing seizures or any increase in seizure frequency or shift from non-generalised to GTCS be closely reviewed and managed with proactive measures, including treatment optimisation and referral to specialised services. There is a paucity of studies correlating changes in seizure severity to SUDEP. The limited evidence (Shankar et al., 2014) suggests a possible association. This area would benefit from further systematic study.
A study which pooled data from randomised placebo-controlled trials in patients with refractory epilepsy showed that treatment with adjunctive AEDs at efficacious doses may have reduced the incidence of definite or probable SUDEP by more than seven times compared with placebo in patients with previously uncontrolled seizures.This provides evidence that, at least in a clinical trial setting, active treatment review and management in patients with refractory epilepsy may reduce SUDEP rates (Ryvlin et al., 2011).
While nocturnal seizures may carry a lower risk of accidental injury, they are a risk factor for SUDEP (Lamberts et al., 2012). The reported protective effect of sharing a room with someone capable of providing assistance has already been referred to (Langan et al., 2005). Evidence from the MORTEMUS study (Ryvlin et al., 2013) demonstrated that delayed resuscitation, when events occur outside of daytime hours, was noted in all SUDEP cases, indicating that at least some of the risk associated with nocturnal seizures relates to lack of supervision (Sander, 2013). Thus, noting and managing nocturnal seizures needs to be part of the standard assessment at epilepsy care visits, including discussing ways to improve night time supervision for persons with frequent nocturnal GTCS. Practical measures, such as co-habiting with a friend or a relative and the use of audio monitors if seizures are uncontrolled, can be considered. There is, however, very limited and poor-quality evidence for most personal-use devices which are claimed to monitor and protect PWE in the community (Jory et al., 2016).
It may be hypothesized that successful epilepsy surgery can reduce SUDEP risk in association with a reduction in seizure frequency (Sperling et al., 2005). Refractory PWE who did not undergo surgery had a six-fold increased rate of death compared to those who underwent epilepsy surgery. Post-surgery, the operated PWE had a standardised mortality rate equal to the general population (Sperling et al., 2005).Pregnant women with epilepsy are also at risk of SUDEP (Saving Mothers Lives, 2011).
Comorbidities and SUDEP risk
Whether comorbidities such as cardiac and respiratory conditions, including obstructive sleep apnoea, have an impact on SUDEP risk is unknown. However, inadequately managed sleep apnoea is associated with increased seizure frequency, as is sleep deprivation from any other cause.
Intellectual disability (ID) is common among PWE. It is estimated that around 25% of PWE have ID (De Boer et al., 2008). There have been suggestions of over-representation of SUDEP in ID populations (Kiani et al., 2013), but data are conflicting (Shankar et al., 2016; Young et al., 2015). This group is particularly challenging with a high proportion being treatment resistant. Difficulties with communication may make informed choice difficult. The clinician thus needs to have an appreciation of ID-specific issues, as applied to epilepsy. There is an urgent need for studies assessing the role of ID and other comorbidities of epilepsy in SUDEP risk.
Given the importance of seizure control for SUDEP risk reduction, patient counselling should include a discussion of seizure provoking factors, such as alcohol, sleep deprivation and non-adherence to treatment, and the need to avoid these. Excessive or harmful use of alcohol is a documented trigger for seizures in PWE (Shankar et al., 2016) and alcohol or substance abuse should be explored in relevant settings. Poor adherence to AED medication has been associated with significantly increased overall mortality, though SUDEP was not specifically analysed (Faught et al., 2008; Ostler et al., 2015; Ridsdale, 2015).
As triggers may provoke seizures in some but not others, an individualised assessment of seizure triggers is needed (box 2). It is important that PWE recognise the risks to their safety and make informed choices around treatment adherence.
SUDEP in children
Like adults, children with epilepsy are at a significantly increased risk of death compared to the general population. The majority of premature mortality in children with epilepsy is not seizure related, most often due to respiratory illness in association with severe neurological disability, chronic epilepsy, and comorbid conditions.
SUDEP is the most common cause of seizure-related death also in children, however, compared with adults, rates are considerably lower in children, affecting approximately 0.2 to 0.3 per 1,000 children with epilepsy per year (Donner et al., 2001; Ackers et al., 2011; Berg et al., 2013).The majority of reported SUDEP deaths occur after the age of 20 years (Thurman et al., 2014). When SUDEP does occur in childhood, many of the risk factors previously identified appear to be present, including treatment-resistant GTCS and nocturnal seizures. The limited studies examining risk factors specific for SUDEP in children identify developmental delay and drug-resistant epilepsy as important factors (Callenbach et al., 2001; Donner et al., 2001; Weber et al., 2005).Mortality is an important consideration in the childhood-onset epileptic encephalopathies because of the combination of severe drug-resistant seizures and developmental delay. Dravet syndrome, most often associated with mutations in the sodium channel-encoding gene SCN1A, has been specifically associated with an increased risk of SUDEP, although it is not clear if a specific genetic mutation confers an increased risk beyond the already mentioned factors (Skluzacek et al., 2011).The situation is similar in other conditions such as isodicentric chromosome 15 syndrome.
While the literature regarding specific paediatric risk factors for SUDEP is sparse, there is a relatively strong literature on the wishes of parents regarding SUDEP risk disclosure.Qualitative and quantitative research supports the finding that parents of children with epilepsy want to be informed of SUDEP risk (Gayatri et al., 2010; Ramachandrannair et al., 2013). Health care providers are therefore encouraged to discuss this risk with families and young persons, where appropriate.SUDEP risk discussion can be incorporated into a general discussion about safety in the context of seizures and is also appropriate at times when young persons are having difficulty with treatment concordance, are contemplating surgical referral for epilepsy surgery, or facing lifestyle changes, such as moving out of the family home.When SUDEP risk is lower, such as the case with fully controlled seizures or absence seizures, a discussion of risk can offer reassurance to the family.When risk is higher, families can be counselled on how to mitigate that risk by reducing seizure burden and improving seizure safety, with the use of audio devices, improved treatment adherence, and referral to specialised epilepsy clinics for comprehensive assessment and treatment.
SUDEP communication and the role of the neurologist
Routine clinical epilepsy care should include a comprehensive person-centred risk assessment and communication of all identified risks, with a focus on modifiable or avoidable risk. It may be difficult to identify the appropriate time to address SUDEP with PWE and their families.While some advocate discussion at the first clinic visit, others recognize this may be difficult.Most agree, however, that discussing SUDEP, as well as other seizure-related risk factors and risk management, sometime early in the course of the patient's epilepsy is important. A proposed pathway (table 1 table 1) is shown, and the key points to discuss are listed in table 2 table 2. Table 3 summarises the benefits and risks of having this discussion. Given the sensitivity of the discussion, there is the potential for some PWE to become distressed by it (Tonberg et al., 2015). Box 1 lists key facts related to SUDEP to use along with box 2 for a discussion on this sensitive issue.
A fatal accident inquiry into the SUDEP deaths of two young women in Scotland, UK (Scotland Judiciary, 2016) showed neither had been advised of the risk of SUDEP by their doctor. There was no recorded SUDEP discussion. The judge concluded that their deaths might have been avoided had they been informed of the risks and taken precautions to minimise these risks. In his report, the judge made a series of recommendations. These include the recommendation that the majority of people with epilepsy should be told about SUDEP when first diagnosed. Any decision not to do this should be noted in medical records (http://www.scotland-judiciary.org.uk/).
Although surveys have not shown this to be commonly carried out (Waddell et al., 2013), it is now considered best practice to conduct a person-centred discussion of SUDEP preferably at an early appointment. If such a discussion is deemed inappropriate, it is important for the reasons to be recorded and a plan made to identify when and by whom the topic would be re-visited. The onus is on the PWE's responsible direct clinical team (nurse, doctor, etc.) who have competency in epilepsy care to discuss the risks of SUDEP along with other epilepsy-related risks. The clinical practice guidance in the UK from The National Institute of Clinical Excellence (NICE) (Clinical Guideline 137, NICE) has clearly addressed this.
However, despite patient expectations and practice guidelines, there remain barriers to meaningful discussions of SUDEP with PWE (Waddell et al., 2013).Principal among them is a lack of clarity about how to have a structured and patient-centred conversation. The risk of SUDEP can initially be included in the discussion when other risks related to epilepsy are addressed and when the importance of preventing seizures by avoiding triggers and through successful treatment is discussed. This can be backed up by general patient information leaflets, many of which address SUDEP. A possible model (Brown et al., 2013; Shankar et al., 2013, 2014, 2015a, 2016) is to incorporate a brief semi-structured risk assessment in routine clinical practice, such as the SUDEP and seizure safety checklist(https://www.sudep.org/checklist) (SUDEP, 2016). This can be a catalyst for open discussion and could help guide treatment by identifying areas for modification of risk status, supporting clinician interventions, and providing a better understanding of those risk factors which lie within the control of the patient.
SUDEP and premature mortality in epilepsy have become a major focus in the epilepsy community (Devinsky et al., 2016; Tomson et al., 2016). Leading experts and advocacy organizations call for systemic strategies to reduce mortality.While there has been much progress in understanding and managing epilepsy as a condition, it is increasingly recognised that systematic person-centred communication of risk (Shankar et al., 2015b) remains a neglected area. The physician has a critical role in patient education, to encourage epilepsy self-management and ensure that PWE are fully informed of the risks of epilepsy and how to modify those risks. It is only with clearly communicated and accurate information about epilepsy that PWE can make informed decisions to reduce their risk of mortality in epilepsy.
Summary didactic slides are available on the www.epilepticdisorders.com website.
RS has received institutional and research support and personal fees from UCB, Eisai, Janssen, Lilly, GSK, Servier, Astra Zeneca and Desitin outside the submitted work. BM has received institutional and research support and personal fees from UCB, Eisai, GSK and Desitin outside the submitted work. TT has received speaker's honoraria to his institution from Eisai, UCB, BMJ India, and Actavis, honoraria to his institution for advisory boards from UCB and Eisai, and received research support from Stockholm County Council, CURE, and GSK. LN has attended scientific meetings as a guest of Bial and Eisai, and received speaker's fees/honoraria from Eisai; UCB have sponsored departmental meetings, and her department has received trigeminal nerve stimulators and electrodes from NeuroSigma. ED has no disclosures.