Abstract
Objective To estimate the incidence of sudden unexpected death in epilepsy (SUDEP) in people with intellectual disabilities in residential care settings and to ascertain the effects of nocturnal seizures and nocturnal supervision on SUDEP risk.
Methods We conducted a nested case-control study reviewing records of all people who died at 2 residential care settings over 25 years. Four controls per case were selected from the same population, matched on age (±5 years) and residential unit. Nocturnal supervision was graded in 3 categories: (1) no supervision; (2) a listening device or a roommate or physical checks at least every 15 minutes; and (3) 2 of the following: a listening device, roommate, additional device (bed motion sensor/video monitoring), or physical checks every 15 minutes. Outcome measures were compared using Mann-Whitney U tests and Fisher exact tests.
Results We identified 60 SUDEP cases and 198 matched controls. People who died of SUDEP were more likely to have nocturnal convulsive seizures in general (77% of cases vs 33% of controls, p < 0.001) and a higher frequency of nocturnal convulsive seizures. Total SUDEP incidence was 3.53/1,000 patient-years (95% confidence interval [CI] 2.73–4.53). The incidence differed among centers: 2.21/1,000 patient-years (95% CI 1.49–3.27) vs 6.12/1,000 patient-years (95% CI 4.40–8.52). There was no significant difference in nocturnal supervision among cases and controls, but there was a difference among centers: the center with a lowest grade of supervision had the highest incidence of SUDEP.
Conclusions Having nocturnal seizures, in particular convulsions, may increase SUDEP risk. Different levels of nocturnal supervision may account for some of the difference in incidence.
Glossary
- CCE=
- Chalfont Centre for Epilepsy;
- SEIN=
- Stichting Epilepsie Instellingen Nederland;
- SUDEP=
- sudden unexpected death in epilepsy
Sudden unexpected death in epilepsy (SUDEP) is one of the most frequent causes of death in people with epilepsy, particularly chronic epilepsy. It may account for up to a fifth of all cases of premature mortality among people with epilepsy.1 The pathologic mechanisms are unclear, but multiple SUDEP risk factors have been identified. Having frequent convulsive seizures is the greatest risk factor.2,3 Circadian factors seem important, with higher risk for those with nocturnal sleep-related convulsive seizures,4 but this still needs confirmation.
SUDEP is mostly a sleep-related and unwitnessed event.4,–,6 Conversely, nocturnal supervision seems to lower SUDEP risk.7 This poses a problem for residential care facilities housing people with refractory epilepsy and intellectual disabilities.8 The incidence of SUDEP is substantial (3.6–3.8 per 1,000 person-years),9,10 but recommendations for nocturnal supervision are lacking.11 It is also well recognized that nocturnal seizures may go unnoticed. A recent survey indicated that one-third of nocturnal seizures in a residential population were missed despite the use of an acoustic detection system.12
This raises the question whether enhancing nocturnal supervision may lower SUDEP risk. We aimed at estimating the incidence of SUDEP in residential care settings and at determining the effects of nocturnal seizures and nocturnal supervision on the risk of SUDEP.
Methods
Selection of cases and controls
We selected all cases of SUDEP in 2 epilepsy residential care facilities: Stichting Epilepsie Instellingen Nederland (SEIN) and Chalfont Centre for Epilepsy (CCE), which provide residential care facilities for people with complex needs. People are assigned to a specific residential unit, depending on epilepsy severity, behavioral problems, and comorbidities.
For SEIN, we reviewed all people who died while in residential care from 1987 to 2012. For the CCE, we reviewed all deaths from 1989 to 2014. Deaths at the CCE between 1989 and 2009 were previously reported.13
All cases of definite and probable SUDEP were selected. Deaths of those older than 60 years were excluded because of a higher risk of competing causes of death. Four controls per case were matched on age (±5 years) and residential unit. The date of death of the case was used as an index date for matching and extraction of individual attributes of the controls.
Identification of probable SUDEP cases
All sudden deaths were reviewed by M.v.d.L. and classified as non-SUDEP deaths or as possible, probable, and definite SUDEP according to the Unified SUDEP definitions.14 As an additional criterion to portray the diagnosis of “probable SUDEP,” we developed a point score based on established SUDEP circumstances. Factors increasing the probability of SUDEP included age <40 years,15 signs of a recent seizure,16,–,18 and an unwitnessed and sleep-related death.4,–,6 Circumstances making SUDEP less likely included the following: having a fever or illness at the time of death, a history of heart disease, or being seizure-free for over a year19 (table 1). These criteria were weighted on effect, and points were assigned to all cases. The cutoff score for the differentiation of probable SUDEP vs probable non-SUDEP sudden death was determined by applying the point scores to the SUDEP and non-SUDEP cases with a postmortem examination.
Classification system for sudden unexpected deaths
Sudden deaths with a score ≥1 were presumed to be probable SUDEP, and those with a score <1 were categorized as non-SUDEP sudden deaths. An expert panel (R.D.T. and J.W.S.) discussed cases if the initial clinical assessment (M.v.d.L.) was not consistent with the test score, or if >3 variables were unknown.
Degree of nocturnal supervision
Until the late nineties, both care facilities had twin rooms and small dormitories (up to 4 people per room), but from 2000 onward, all residents moved to single-room accommodations. In 1999, SEIN deployed a central nocturnal acoustic detection system covering all residents. Bed motion sensors were added in 2008 and video monitoring in 2010 for those in whom there was a suspicion of unwitnessed nocturnal events. The number of staff allocated to ensure proper monitoring of seizure detection systems increased to one carer per 14 residents. Individuals were physically checked at least once or twice a night, and additional checks are performed on a need basis.
No central nocturnal seizure detection system was used at the CCE, although individual devices such as listening devices and the so-called exit alarms (sensors which alert when someone leaves their bed) were used in some. Over the years, the number of carers increased from one per 12 to one per 6 residents. The institutional protocol recommends that all residents were physically checked once every 15 or 30 minutes.
We graded the intensity of nocturnal supervision for all cases and controls (table 2). The grading entails various factors including the presence and type of seizure detection devices, the presence of a roommate (who may alert the nursing staff in case of a seizure), and the frequency of physical checks. A survey of video-EEG-recorded SUDEP cases indicated that terminal apnea occurred within 11 minutes postictally.17 We therefore only assigned a higher grade of nocturnal supervision to those with physical checks, occurring once every 15 minutes.
Grades of nocturnal supervision
Individual characteristics
For all SUDEP cases and controls, medical records were abstracted for the following variables: age, sex, age at onset of epilepsy, seizure types, seizure frequency, epilepsy classification, number of antiepileptic drugs, use of benzodiazepines, IQ, type of nocturnal supervision, and number of residents per carer at night.
The nocturnal and diurnal seizure frequencies were estimated for the year before the death of the case, as an average per month, specified by seizure type (convulsions or other seizure types).
Analysis
The incidence of (probable) SUDEP was estimated by dividing the number of SUDEP cases by the total number of person-years per center. A nested case-control study compared SUDEP cases with controls. Mann-Whitney U tests and Fisher exact tests were used to analyze differences in SUDEP risk factors between cases and controls. Grade of supervision was corrected for epilepsy severity (number of nocturnal convulsive seizures) by fitting a multivariable logistic regression model.
Standard protocol approvals, registrations, and consents
The study was approved by the Leiden University Medical Center Ethics Committee with a waiver of informed consent. It was also registered as a service evaluation in the United Kingdom.
Data availability
Anonymized data will be shared by request from any qualified investigator.
Results
Altogether, we identified 60 SUDEP cases: 25 at center 1 (8 SUDEP and 17 probable SUDEP) and 35 in center 2 (22 SUDEP and 13 probable SUDEP cases). All definite SUDEP and probable SUDEP cases scored >1 on the SUDEP probability score. Two cases could not be assessed with the SUDEP score because of missing values but were classified as probable SUDEP by the expert panel.
The total duration of follow-up was 17,016.5 patient-years, resulting in a total SUDEP incidence of 3.53 per 1,000 patient-years (95% confidence interval [CI] 2.73–4.53). At center 1, over 26 years, we accumulated 11,302 patient-years, resulting in a SUDEP incidence of 2.21/1,000 patient-years (95% CI 1.49–3.27). In center 2, a total 5,714.5 patient-years was accumulated, leading to an incidence of 6.12/1,000 patient-years (95% CI 4.40–8.52).
Cases more often had convulsive seizures (p = 0.019) and had a higher IQ compared with controls (p = 0.005) (table 3). Compared with controls, SUDEP cases more often had nocturnal convulsive seizures (p < 0.001). Cases also had a higher frequency of convulsive seizures (p = 0.001) and nocturnal convulsive seizures in particular (p < 0.001) (table 4 and figure 1). There was no significant difference in the grade of supervision; also after correction for epilepsy severity (number of nocturnal convulsive seizures logistic regression nocturnal supervision grade 1: odds ratio [OR] 0.73; 95% CI 0.15–3.46; p = 0.693; grade 2: OR 0.37; 95% CI 0.08–1.83; p = 0.225).
Characteristics of SUDEP cases vs controls
Sudden unexpected death in epilepsy risk factors and grade of supervision of cases and controls
When comparing cases of center 1 and center 2, no significant differences were seen for the presence of convulsive seizures, presence of nocturnal convulsive seizures, frequency of convulsive seizures, and frequency of nocturnal convulsive seizures. Cases from center 1 had a significantly higher grade of nocturnal supervision compared with cases from center 2 (p < 0.001) (table 5). The same applied to the controls from center 1 vs center 2 (p < 0.001) (table 5 and figure 2). Patient characteristics for cases and controls per center are available in table 6.
SUDEP risk factors and grade of nocturnal supervision: center 1 cases vs center 2 cases and center 1 controls vs center 2 controls
Patient characteristics for cases and controls per center
Discussion
We found that the presence and frequency of nocturnal seizures increased SUDEP risk in 2 residential populations of people with refractory epilepsy and intellectual disabilities. The incidence of SUDEP differed significantly among sites. This difference could not be explained by markers of epilepsy severity. Instead, differences in institutional policies on nocturnal supervision seemed the most plausible explanation with more SUDEP victims in the center with less supervision.
There was a difference in supervision among the sites, but nocturnal supervision was similar for cases and controls within each site. This is likely explained by the fact that cases were matched to controls from the same site. Most supervision systems were implemented per residential unit. Cases and controls were from the same unit, resulting in no significant differences between cases and controls within a center. There were clear differences in institutional policies and reimbursement systems among the sites, resulting in different resources. Accordingly, the level of supervision was also different when comparing controls from the sites, thus explaining why no differences were found between cases and controls at the same site.
To explain difference in the incidence of SUDEP among sites, we compared known SUDEP risk factors such as the presence and frequency of diurnal and nocturnal convulsive seizures among cases and found no difference. The only difference we identified was the grade of nocturnal supervision, and this seems to be the only likely explanation for the differing SUDEP incidence among sites.
Our work has several limitations. We lacked autopsies in some cases. To avoid misclassification of probable SUDEP and avoid diagnostic bias, we restricted our analysis to those aged <60 years and developed a SUDEP point score based on clinical criteria of the postmortem cases. Because of the retrospective design, there were missing values precluding a matched analysis. For most people, extensive seizure charts were available for the extraction of seizure frequency. Still, we had to rely on carer's reports for nocturnal seizure frequency. Lack of supervision may have led to an underestimation of nocturnal seizure frequency,12,20 affecting our results. Convulsive seizures, the most relevant seizure type for SUDEP risk, are less likely to be missed than other seizure types.12,20 Even if these seizures missed, they can still be reported, as there are often signs such as tongue bite. We compared all major SUDEP risk factors among centers; however, additional factors such as different policies and different medication regimes should also be considered. Different antiepileptic regimes might have led to different seizure frequencies. The seizure frequencies, however, were similar among centers, suggesting that this was not a major factor.
Our data on the incidence of SUDEP in this population with epilepsy and complex needs are comparable to previous reports on similar but smaller populations.9,10 Our number of person-years followed is over 5 times more than previous reports, and this enabled us to compare confidently incidence numbers in the 2 sites. Our data confirmed well-established SUDEP risk factors such as convulsive seizures and a high frequency of these seizures in general.2,3,21 We also confirmed that the presence and high frequency of nocturnal convulsive seizures to be independent SUDEP risk factors.4
The strong association between SUDEP and sleep is explained by the interaction with 2 environmental factors: prone position and the absence of a witness. Most SUDEP cases are found in the prone position,6,17,22 which is remarkable as people seldom are prone after nonfatal convulsive seizures.23,24 When an individual is in the prone position after a seizure, respiratory dysfunction may lead to apnea and asystole, which should normally evoke an arousal response.25 Postictal coma might, however, prevent arousal and thus the resumption of ventilation, consequently leading to SUDEP.26 Nursing interventions such as repositioning and oxygen administration have been reported to significantly shorten the duration of respiratory dysfunction after a convulsive seizure.27,28 Further study of the mechanisms involving nursing interventions (other than cardiopulmonary resuscitation) may help prevent postictal coma or even SUDEP.29
Our study and 2 previous reports suggest that nocturnal supervision is protective for SUDEP. A case-control study showed that SUDEP cases less often had a roommate or a listening device compared with the controls.7 In a cohort study of children with severe epilepsy and intellectual disabilities, all 14 SUDEP deaths occurred while the young students were not under the supervision of the boarding school once they had left the school or were on leave.30
Although we provide some support for the protective effect of nocturnal supervision, specific recommendations to reduce SUDEP risk require further research.31,32 Many seizure detection systems are available,33,34 but what system works best for individuals in different populations is yet unknown. The variation in nocturnal supervision among the sites in our study was predominantly explained by the implementation of an acoustic detection system in one center. Acoustic detection systems are often useful, as in 85% of tonic-clonic seizures, an ictal cry is heard35 and at least half of major seizures are captured with a listening device.36 In case these measures failed, additional individually tailored devices such as a bed motion sensor and video monitoring were deployed in center 1. This only concerned a small subgroup with nocturnal supervision grade 3 and is thus unlikely to be a major determinant for the differences in the incidence of SUDEP. We have previously shown that residential care center video monitoring facilitated the detection of 10% of all seizures requiring an intervention.12 Video monitoring, however, has large implications for privacy and is costly. This study underlines the importance of nocturnal supervision and reliable seizure detection systems for different populations.
Author contributions
M. van der Lende, D.C. Hesdorffer, J.W. Sander, and R.D. Thijs conceptualized the study and designed it. M. van der Lende did the data acquisition and prepared the first draft. All authors were involved in the interpretation of data, critically reviewed the manuscript, and approved the final version.
Study funding
This work was funded by Christelijke Vereniging voor de Verpleging van Lijders aan Epilepsie, The Netherlands, the Dutch Epilepsy Foundation, project number 15-10, and the AC Thomson Foundation. This work was partly conducted at NHHR University College London Hospitals/University College London Comprehensive Research Center, which receives a proportion of funding from the UK Department of Health's National Institute for Health Research Biomedical Research Centers funding scheme. J.W. Sander receives research support from the Dr. Marvin Weil Epilepsy Research Fund and the UK Epilepsy Society.
Disclosure
M. van der Lende reports no disclosures relevant to the manuscript. D. Hesdorffer was a consultant for Cyberonics, is an Associate Editor of Epilepsia, and is the advisor to the Mount Sinai Injury Prevention Center. J.W. Sander received honoraria from UCB, Eisai, Bial, and Janssen Cilag, which are involved in the manufacturing of antiepileptic drugs. R. Thijs receives research support from the Dutch National Epilepsy Fund, The Netherlands Organisation for Health Research and Development (ZonMW), NUTS Ohra Fund, Medtronic, and AC Thomson Foundation and has received fees for lectures from Medtronic, UCB, and GSK. Go to Neurology.org/N for full disclosures.
Publication history
Received by Neurology January 23, 2018. Accepted in final form July 10, 2018.
Acknowledgments
The authors thank Mrs. Annette Russell, Dr. Gail Bell, and Prof. Mathias Koepp for their help with data acquisition at the CCE; Dr. Jan Novy for sharing data from a previous study of the mortality at the CCE; and Prof. Job van der Palen for his advice on statistical methods and for her help in data acquisition.
Footnotes
Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Editorial page 731
- Received January 23, 2018.
- Accepted in final form July 10, 2018.
- © 2018 American Academy of Neurology
References
Letters: Rapid online correspondence
- Author response to Dr. Sethi
- Nocturnal supervision and SUDEP risk
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
