Abstract
Association of Prone Position with Sudden Unexpected Death in Epilepsy
Liebenthal JA, Wu S, Rose S, Ebersole JS, Tao JX. Neurology 2015;84:703–709.
OBJECTIVE: To examine the association between prone position and sudden unexpected death in epilepsy (SUDEP). METHODS: We conducted a systematic review and meta-analysis based on a literature search from databases PubMed, Web of Science, and Scopus, using keywords “SUDEP” or “sudden unexpected death in epilepsy” or “sudden unexplained death syndromes in epilepsy.” Twenty-five publications met the inclusion and exclusion criteria and were enrolled in this study. RESULTS: Body positions were documented in 253 cases of SUDEP. Of these patients, 73.3% (95% confidence interval [CI] 5 65.7%, 80.9%) died in the prone position, whereas 26.7% (95% CI 5 16.3%, 37.1%) died in nonprone positions. Binary random-effects analysis showed that prone position is significantly associated with SUDEP, as compared with nonprone position (p, 0.001). In addition, the prone position was reported in all 11 cases of video-EEG–monitored SUDEP. Moreover, in a subgroup of 88 cases of SUDEP in which demographics and circumstances of death were documented, the prone position was observed in 85.7% (95% CI 5 74.6%, 93.3%) of patients aged 40 years or younger, but in only 60% (95% CI 5 38.7%, 78.9%) of patients older than 40 years. Statistical analysis confirmed that the prone position was significantly more prevalent in the younger patient group, as compared with the older patient group (odds ratio 3.9; 95% CI 5 1.4%, 11.4%; p 5 0.009). CONCLUSION: There is a significant association between prone position and SUDEP, which suggests that prone position is a major risk factor for SUDEP, particularly in patients aged 40 years and younger. As such, SUDEP may share mechanisms similar to sudden infant death syndrome.
Commentary
Liebenthal et al. put forth their findings that the prone position is associated with the occurrence of sudden unexplained death in epilepsy (SUDEP). In their extensive literature search of SUDEP patients and body position at death, many more SUDEP patients for whom these details were available were found in a prone position than in non-prone positions. This finding fits well into the hypothesis that SUDEP has a similarity to sudden infant death syndrome (SIDS), since avoiding the prone position during sleep reduces SIDS risk (1). The association between these two entities is strengthened by a molecular mechanism involving serotonin. As the authors discuss in some depth, SIDS victims may have a serotonin-deficiency that impairs arousal. On the epilepsy side, epileptic DBA/2 mice are susceptible to death during a seizure in the setting of serotonin antagonism. Death during seizures in this model is mitigated, however, by a serotonin reuptake inhibitor, fluoxetine. Another hypothesis invoking the role of the prone position in SUDEP occurrence is its participation as a cardinal feature of the cascade leading to SUDEP. The authors identify other cascade features as 1) generalized tonic clonic seizures (GTCS) and 2) postictal generalized electroencephalographic suppression (PGES).
There is much connecting of dots in these hypotheses, and some of these connections are threadbare. For example, as the authors point out, PGES associates with GTCS but does not associate clearly with SUDEP occurrence. This raises the question as to whether PGES belongs in the cascade described previously. A less tenuous connection starts with the authors generally assuming that all instances of SUDEP are preceded by GTCS, which is implied but not clearly stated in the article. That GTCSs likely precede SUDEP is now widely accepted and well supported by the recordings of SUDEP and near-SUDEP in epilepsy monitoring units (2). GTCS occurrence increases SUDEP risk, as does nocturnal GTCS occurrence (3).
From details provided in this article, only 4 of 11 patients who had video-EEG-recorded SUDEP were sleeping in a prone position before terminal seizure onset. However, all 11 were reported as having a terminal seizure and were found dead in a prone position. During a GTCS, the patient could become prone due to versive or other motor phenomenon, as described in three patients from the same video-EEG-recorded SUDEP series. Therefore, it cannot be determined from these data whether the prone position observed after death had any additional contribution to the SUDEP risk imparted by a GTCS. Since the mechanism of SUDEP is not known, ending up in a prone position after a GTCS may contribute. However, sleeping in a supine position would likely not decrease the risk of SUDEP if the patient's GTCS results in their lying in a prone position, but preventing the prone position at the end of a GTCS may be a strategy to reduce SUDEP.
As with PGES, the prone position risk factor for SUDEP needs to be compared in the setting of its susceptible condition, a GTCS. The question is this: What is the relative risk of SUDEP when the GTCS ends with the patient in a prone versus non-prone position? The need for this comparison actually mechanistically separates SUDEP from SIDS. The susceptible condition for SIDS is not an intermittent, unpredictable, or even pathological event; it is something that occurs normally every day for many hours—sleep.
Our work as scientists requires that we attempt to connect the dots. The validity of this approach is reinforced in the most highly satisfying manner with every episode of Law and Order, which some of us watch way too often. Of course, we rarely eliminate a disease as simply as the television lawyers send the bad guy (usually played by a striving New York City actor) to jail, but with SUDEP, the attempt to identify risk factors is enhanced and heightened by an infusion of terror and tragedy. When looked at after stepping outside this feverish milieu, the conclusion that a prone position as a risk factor for SUDEP is far from proven. To achieve scientific proof, which will support a clinical truth, sources of bias that could erroneously sway the results must be minimized. These authors were specifically concerned about publication bias, which Wikipedia defines as a “is a bias with regard to what is likely to be published, among what is available to be published.” The authors state that they managed publication bias by only including reports in which the total number of SUDEP patients found in various body positions must be documented. While it remains unclear if publication bias was mitigated, reporting bias, defined as “selective revealing or suppression of information” (Wikipedia), could surely have been present. The authors identified 413 SUDEP patients within the 25 articles meeting their inclusion criteria, but position at death was noted only for 253, and these patients were included in the study. This means that data on 160 SUDEP patients, or 39% of the total SUDEP patients, were missing. Why would these data be missing? It is likely not because of subversive reasons but more likely because of practical reasons, such as the patient being moved before the position was noted. In any case, this proportion of missing data contributes to a marked risk of reporting bias. If many patients for whom data were missing died of SUDEP while in a supine position, this report would not be supporting the prone position as SUDEP risk factor and possibly would not exist at all, therefore contributing to publication bias.
The nature of SUDEP is such that patients and families will go to great lengths and expense to prevent it. While the magnitude of risk reduction may be less important in discussing this most unforgiving outcome, feasibility, cost, and availability are important to consider when planning SUDEP prevention strategies with patients and families. The use of a nocturnal monitoring device or having a family member sleep in the room with the patient are both likely helpful in reducing SUDEP risk (4), but their implementation must be an individualized decision. Should the worst case scenario occur, the added burden of guilt caused by an unfounded perception that SUDEP could have been prevented is also a concern. The mechanism by which SUDEP occurs is still elusive and is likely, as the authors state herein, multifactorial. However, it is clear that SUDEP can be prevented if GTCS are stopped. This remains our most effective line of attack against SUDEP.