Sudden Unexpected Death in Epilepsy Incidence Rates and Risk Factors Practice Guideline Sudden Unexpected Death in Epilepsy Incidence Rates and Risk Factors Report by: Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology and the American Epilepsy Society

Practice Guideline Endorsement and Funding This practice guideline was co-developed with the American Epilepsy Society and endorsed by the International Child Neurology Association. This practice guideline was developed with financial support from the American Academy of Neurology (AAN). Authors who serve as AAN subcommittee members or methodologists (C.H., D.G., J.A.F.) were reimbursed by the AAN for expenses related to travel to subcommittee meetings where drafts of manuscripts were reviewed.

Sharing This Information The AAN develops these presentation slides as educational tools for neurologists and other health care practitioners. You may download and retain a single copy for your personal use. Please contact guidelines@aan.com to learn about options for sharing this content beyond your personal use.

Presentation Objectives To present the evidence systematically reviewed in the AAN guideline on sudden unexpected death in epilepsy (SUDEP) Determine incidence rates in different epilepsy populations Address whether risk factors for SUDEP have been identified To present practice recommendations

Overview Introduction Clinical questions AAN practice guideline process Methods Conclusions Practice recommendations

Introduction SUDEP is a poorly understood and catastrophic risk of epilepsy The sensitive nature of discussions of this infrequent but important risk with patients with epilepsy and their families has prompted the need for evidence-based information about SUDEP The goal of this practice guideline is to examine evidence regarding the SUDEP incidence rate in different epilepsy populations and any prognostic factors for SUDEP occurrence

Clinical Questions What is the incidence rate of SUDEP in different epilepsy populations? Clinical Question 1 Are there specific risk factors for SUDEP? Clinical Question 2

AAN Practice Guideline Process* Clinical Question Evidence Conclusions Modified Delphi Consensus Recommendations *Practice guideline broadly developed using the 2004 AAN Clinical Practice Guideline Process Manual. However, 2011 process used to formulate the conclusions and recommendations.

Literature Search/Review Rigorous, Comprehensive, Transparent Searched: EMBASE, MEDLINE 1,068 abstracts 35 rated articles Inclusion criteria: Randomized studies pertinent to the questions n > 10 SUDEP definition provide by Nashef 1997, e3, Annegers 1997, e4, and Leetstma et al 1997e5 Exclusion criteria: Studies in animals Languages other than English

AAN Classification of Evidence Screening Scheme (2011) Study of a cohort of patients at risk for the outcome from a defined geographic area (i.e., population based) The outcome is objective Also required: a. Inclusion criteria defined b. At least 80% of patients undergo the screening of interest

AAN Classification of Evidence Screening Scheme (2011) Class II A non–population-based nonclinical cohort (e.g., mailing list, volunteer panel) or a general medical, neurology clinic/center without a specialized interest in the outcome. Study meets criteria a and b (see Class I) The outcome is objective Class III A referral cohort from a center with a potential specialized interest in the outcome Class IV Did not include person at risk for the outcome Did not statistically sample patients or patients specifically selected for inclusion by outcome Undefined or unaccepted screening procedure or outcome measure No measure of frequency or statistical precision calculable

AAN Classification of Evidence Prognostic Scheme (2011 amended in 2015) Cohort survey with prospective data collection Includes a broad spectrum of persons at risk for developing the outcome Outcome measurement is objective or determined without knowledge of risk factor status Also required: a. Inclusion criteria defined b. At least 80 percent of enrolled subjects have both the risk factor and outcome measured

AAN Classification of Evidence Prognostic Scheme (2011 amended in 2015) Class II Cohort study with retrospective data collection or case-control study. Study meets criteria a and b (see Class I) Includes a broad spectrum of persons with and without the risk factor and the outcome The presence of the risk factor and outcome are determined objectively or without knowledge of one another Class III Cohort or case control study Narrow spectrum of persons with or without the disease The presence of the risk factor and outcome are determined objectively, without knowledge of the other or by different investigators Class IV Did not include persons at risk for the outcome Did not include patients with and without the risk factor Undefined or unaccepted measures of risk factors or outcomes No measures of association or statistical precision presented or calculable *Objective outcome measurement: an outcome measure that is unlikely to be affected by an observer’s (patient, treating physician, investigator) expectation or bias (e.g., blood tests, administrative outcome data).

AAN Conclusions and Recommendation Confidence in evidence anchored to the studies’ risk of bias Highly likely or highly probable* = high confidence level Likely or probable = moderate confidence level Possibly = low confidence level Insufficient evidence = very low confidence level Recommendations informed by premises Evidence systematically reviewed Strong evidence derived from related conditions Axiomatic principles of care Inferences made from one or more statements in the recommendation rationale Clinician level of obligation assigned (modified Delphi) Must (or must not) = Level A (strong) Should (or should not) = Level B (moderate) May (or may not) = Level C (weak) Should not = Level R (restricted to research setting only) No recommendation made = Level U (insufficient evidence) *Italics denotes language that would appear in the conclusions or recommendations

Clinical Question 1 What is the incidence of SUDEP in different epilepsy populations?

Conclusions Because of imprecision in the incidence results for general populations of persons with epilepsy, authors performed a random-effects meta-analysis for several subpopulations To explore reasons for heterogeneity in the absolute risk, the panel conducted a meta-analysis of subgroups of studies including different groups of persons with epilepsy

Conclusions What is the incidence of SUDEP in childhood? Based on moderate confidence in the evidence from 3 Class I studies, SUDEP risk in children with epilepsy is 0.22/1,000 patient-years (95% CI 0.16–0.31).

Conclusions What is the incidence of SUDEP in people with childhood-onset epilepsy, including SUDEP occurrences after adulthood? Based on low confidence in the evidence due to considerable imprecision across 5 Class I studies,e6,e7­­– e9,e16 SUDEP risk increases in adults to 1.2/1,000 patient-years (95% CI 0.64–2.32).

Conclusions What is the incidence of SUDEP in the general epilepsy population? Based on low confidence in the evidence owing to considerable imprecision across 12 Class I studies, the overall SUDEP risk is 0.58/1,000 patient-years (95% CI 0.31–1.08).

Conclusions What is the incidence of SUDEP in the general population of persons with epilepsy who at some point in the course of their disease were likely difficult to treat? Based on moderate confidence in the evidence from 1 Class I study, the general population of people with epilepsy who at some point in the course of their disease were likely difficult to treat, the overall SUDEP incidence rate is 1.5/1,000 patient-years (95% CI, 0.8–2.7).

Conclusions What is the incidence of SUDEP in pregnant women with epilepsy? Based on moderate confidence in the evidence from 1 Class I study, the SUDEP incidence rate for pregnant women with epilepsy is 0.79/1,000 pregnancies (95% CI 0.6–1.0).

Table e-2. Conclusions for SUDEP incidence Population SUDEP/1,000 patient-years (CI) Confidence level Overall 0.58 (0.31–1.08) Low Childhood 0.22 (0.16–0.31) Moderate Adulthood 1.2 (0.64–2.32)

Recommendations: Clinical Question 1 SUDEP incidence in children Level B Clinicians caring for children with epilepsy should inform the children’s parents or guardians that There is a rare risk of SUDEP In 1 year, SUDEP typically affects 1 in 4,500 children with epilepsy; in other words, annually, 4,499 of 4,500 children will not be affected by SUDEP

Recommendations: Clinical Question 1 SUDEP incidence in adults Level B Clinicians should inform their adult persons with epilepsy that There is a small risk of SUDEP In 1 year, SUDEP typically affects 1 in 1,000 adults with epilepsy; in other words, annually, 999 of 1,000 adults will not be affected by SUDEP

Clinical Question 2 Are there any risk factors for SUDEP?

Most concerning risk factors Odds ratio (CI) Confidence level Presence of GTCS vs lack of GTCS 10 (7–14) Moderate Frequency of GTCS OR 5.07 (2.94–8.76) for 1–2 GTCS per y, and OR 15.46 (9.92–24.10) for >3 GTCS per y High Not being seizure free for 1–5 y 4.7 (1.4–16) Not adding an AED when patients are medically refractory 6 (2–20) Nocturnal supervision (risk reduction) 0.4 (0.2–0.8) Use of nocturnal listening device (risk reduction) 0.1 (0.0–0.3)

Conclusions Presence or absence of (GTCS) It is likely that GTCS occurrence (vs no GTCS occurrence) increases SUDEP risk, based on moderate confidence in the evidence from 2 Class II studies. Although 1 study shows significance and the other does not, the CI of the smaller studye21 does not exclude an important effect encompassed by the CI of the larger study.e20 The larger study also shows a dose-response effect of GTCS occurrence vs not having GTCS on SUDEP risk,e20 which results in an upgrade of the conclusion to high from moderate.

Conclusions Seizure-associated factors: GTCS frequency It is highly likely that GTCS frequency is associated with an increased SUDEP risk (based on 2 Class II studies upgraded to high from moderate because of magnitude of the effect). SUDEP risk increases 3-fold at a GTCS frequency of >3/y, compared with a GTCS frequency of 1–2/y.

Conclusions Seizure-associated factors: Seizure freedom/seizure remission It is likely that having a seizure within the last year increases SUDEP risk (moderate confidence in the evidence based on 2 Class II studies), as does having a seizure in the previous 5 years (moderate confidence in the evidence based on 1 Class I study) compared with being seizure free.

Conclusions Highly refractory epilepsy There is insufficient evidence to support or refute the prognostic value of refractory epilepsy for SUDEP (vs newly diagnosed epilepsy) (very low confidence in the evidence, 2 Class I studies, both with insufficient precision to drive recommendations because 95% CIs included both important and unimportant differences).

Conclusions Nocturnal supervision It is likely that nocturnal supervision defined as the presence in the bedroom of another individual of normal intelligence and at least 10 years old is associated with a decreased SUDEP risk (confidence in the evidence from 1 Class II study upgraded to moderate from low because of magnitude of the effect of 20%–80% reduction).

Conclusions Nocturnal listening device It is likely that use of regular checks throughout the night or a nocturnal listening device is associated with a decreased SUDEP risk (confidence in the evidence from 1 Class II study upgraded to moderate from low because of magnitude of the effect of at least 70% reduction).

Additional Conclusions Please see manuscript for full discussion of conclusions. The evidence is low that the following factors are associated with altering SUDEP risk: Nocturnal seizures (associated with increased risk) Any specific AED (none associated specifically with increased risk) LTG use in women (associated with increased risk) Never having been treated with an AED (associated with increased risk) Number of AEDs used overall (associated with increased risk) Heart rate variability (not associated with increased risk) Extratemporal epilepsy (associated with increased risk) Intellectual disability (associated with increased risk) Male gender (associated with increased risk) Anxiolytic drug use (associated with increased risk)

Additional Conclusions Please see manuscript for full discussion of conclusions. The evidence is very low or conflicting that the following factors are associated with altering SUDEP risk: Overall seizure frequency when evaluated by using all seizure types Medically refractory epilepsy vs not having well-controlled seizures defined as no seizures for the past year Monotherapy vs polytherapy CBZ, PHT, or VPA levels that are above, below, or within the reference range Psychotropic drug use Mental health disorders, lung disorders, or alcohol use LTG use in people with highly refractory epilepsy Frequent changes in AEDs Therapeutic drug monitoring

Additional Conclusions (continued) Please see manuscript for full discussion of conclusions. The evidence is very low or conflicting that the following factors are associated with altering SUDEP risk: Undergoing a resective epilepsy surgical procedure (although current research does not rule out the possibility of a beneficial effect or, further, the potential effect of epilepsy surgery on reducing GTCS frequency and epilepsy severity on reducing SUDEP risk) Engel outcome of epilepsy surgery (although current research does not rule out the possibility of a beneficial effect and, further, the potential effect of epilepsy surgery on reducing GTCS frequency and epilepsy severity on reducing SUDEP risk) VNS use for more than 2 years (however, current research does not rule out the possibility of a beneficial effect and, further, the potential effect of epilepsy surgery on reducing GTCS frequency and epilepsy severity on reducing the risk of SUDEP)

Additional Conclusions (continued) Please see manuscript for full discussion of conclusions. The evidence is very low or conflicting that the following factors are associated with altering SUDEP risk: Epilepsy etiology, whether idiopathic or localization related Structural lesion on MRI Duration of epilepsy Age at epilepsy onset Postictal EEG suppression

Recommendations: Clinical Question 2 Risk factor of GTCS Level B For persons with epilepsy who continue to experience GTCS, clinicians should continue to actively manage epilepsy therapies to reduce seizure occurrences and the risk of SUDEP while incorporating patient preferences and weighing the risks and benefits of any new approach

Recommendations: Clinical Question 2 Risk factor of lack of nocturnal supervision Level C For persons with frequent GTCS and nocturnal seizures, clinicians may advise selected patients and families, if permitted by their individualized epilepsy and psychosocial circumstances, to use nocturnal supervision or other nocturnal precautions, such as the use of a remote listening device, to reduce SUDEP risk

Recommendations: Clinical Question 2 Risk factor of uncontrolled epilepsy Level B Clinicians should inform their persons with epilepsy that seizure freedom, particularly freedom from GTCS (which is more likely to occur with medication adherence), is strongly associated with a decreased risk of SUDEP

Recommendations for Future Research Systematic methods should be developed to identify and report the incidence of SUDEP in different epilepsy populations in order to obtain a better understanding of the incidence and causes of this devastating condition. Educational efforts are needed to improve the forensic knowledge of SUDEP among professionals such as medical examiners, coroners, and pathologists in order to help determine, and document on death certificates, the etiology in individuals, and in order to improve overall knowledge of this condition. Research to identify preventable risk factors should be supported and encouraged so that future clinical trials will be conducted to reduce SUDEP occurrence. Of particular importance is to better understand (a) the relationship between the nature, severity, and duration of epilepsy and the occurrence of SUDEP, and (b) whether current treatments affect the risk of developing SUDEP. Because of (a) risks identified with frequent GTCS, (b) the fact that one study shows more SUDEP events occur in people in placebo arms of trials, and (c) increased SUDEP risk, serious consideration should be given to avoid assigning people with frequent GTCS to placebo for long periods.

References References cited here can be found in the published practice guideline, available at AAN.com/guidelines.

Access Practice Guideline and Summary Tools To access the practice guideline and related summary tools, visit AAN.com/guidelines. Practice guideline article Summary for clinicians and summary for patients/families

Questions?