Novel SUDEP-3 Inventory Shows Potential to Predict SUDEP Risk

SAP Partner | <b>Vickie and Jack Farber Institute for Neuroscience</b>

The SUDEP-3 outperformed the prior SUDEP-7 variation on all measures, and simplifies the risk assessment with a 3-item weighted scoring while enhancing the clinical practicality.

A newly proposed inventory tool for predicting sudden unexpected death in epilepsy (SUDEP) has been shown in a small study to demonstrate the limitations of the currently used SUDEP-7 inventory and better prediction of SUDEP.

The novel tool includes weighting scoring for 3 items—generalized tonic-clonic seizures (GTCS) within the last year, any seizures within the last year, and intellectual disability—which can result in a score from 0 to 4 (TABLE). The data from this study showed resulting scores from SUDEP-3 that were significantly able to predict SUDEP (P <.001), accounting for 28% of its variance, which outperformed the 7-item SUDEP inventory.

Notably, for each additional point gained on the SUDEP-3, the odds of SUDEP increased by 180%, compared to the per point odds increase of 40% associated with the SUDEP-7. In receiver-operating characteristic (ROC) analyses, the area under the curve (AUC) was 0.75 for the SUDEP-3 compared with 0.66 for the SUDEP-7. Model quality was also deemed better for the SUDEP-3.

“We believe that this inventory provides a simple way to identify patients at highest risk of SUDEP. Additionally, it can also help track whether there is a reduction in the patient’s SUDEP risk over time, since the scoring is dynamic and does not depend upon the duration of epilepsy (as does the SUDEP-7),” study author Maromi Nei, MD, vice chair of faculty affairs, and director, Epilepsy Fellowship Program, Jefferson University, told NeurologyLive. “We hope that using this inventory may help clinicians counsel patients regarding their individual SUDEP risk. Such information may also help to highlight the importance of medication adherence and support consideration of using other means to achieve seizure control, such as epilepsy surgery, when appropriate, to lower SUDEP risk.”

She and colleagues wrote, as well, that “although the SUDEP-3 improves upon the SUDEP-7, there is ongoing opportunity for improvement, particularly with regard to optimizing sensitivity and specificity and validating the inventory in a larger, population-based cohort.” Nei et al added that further study is required to validate this new inventory, as well as to better clarify the cutoff values for components such as seizure frequency and to include additional variables reflective of SUDEP mechanisms like cardiac and pulmonary function.

The SUDEP-7 inventory was developed in 2010 by Christopher M. DeGiorgio, MD, professor and vice chairman, Department of Neurology, UCLA, and colleagues, and was then revised in 2015 to include new scoring methodology.2,3 Despite this, there remains a need for a validated clinical tool to accurately identify which patients are at highest risk of SUDEP, with Nei et al. pointing out that the only recent attempt to validate the SUDEP-7 identified no associations between SUDEP-7 score and SUDEP occurrence. “Because the SUDEP-7 inventory is used as a tool for estimating the risk of SUDEP both in practice and clinical research, validation of the instrument is essential,” Nei and colleagues wrote.

This retrospective study of the SUDEP-3 included 48 individuals with epilepsy who underwent video-electroencephalography (EEG) monitoring at Thomas Jefferson University Hospital and subsequently died of definite or probably SUDEP between 1989 and 2017. For each individual who died of SUDEP, 2 matched controls with epilepsy were identified, all of whom underwent video-EEG between 2008 and 2015 at the same location. The final analysis included 28 individuals in the SUDEP group (women, n = 13; median age at admission, 32 years [range, 14-54]) and 56 in the control group (women, n = 26; median age at admission, 31 years [range, 15-56]). Of note, at both the time of admission and at last follow-up, the SUDEP-7 scores were significantly different between the SUDEP (admission: 3.65 [standard deviation (SD), 1.62]; last follow-up: 3.65 [SD, 2.18]) and control groups (admission: 2.65 [standard deviation (SD), 1.47], P = .024; last follow-up: 2.09 [SD, 1.82], P = .016).

WATCH NOW: Dravet Syndrome: SUDEP and Increase Risk of Premature Mortality

Several components of the SUDEP-7 inventory were independently associated with a higher risk of SUDEP, including the occurrence of more than 3 GTCS (P =.002), 1 or more GTCS (P = 0.001), or 1 or more seizures of any type within the last year (P = .013), and intellectual disability (P = .031). When conducting stepwise regression models, however, Nei et al found that SUDEP-7 scores did not improve SUDEP prediction over either GTCS frequency or seizure frequency alone. The SUDEP-3 inventory, though—comprising again of GTCS frequency, seizure frequency, and intellectual disability (P <.001)—outperformed the SUDEP-7 inventory in predicting SUDEP (P = .010).

Additionally, Nei and colleagues wrote that “the SUDEP-3 inventory demonstrated consistently better sensitivity and specificity than the SUDEP-7 inventory. For the SUDEP-3, a cutoff score of ≥3 yielded the optimal balance of sensitivity (0.57) and specificity (0.75; Youden's J = 0.32). A SUDEP-7 cutoff score of ≥3 yielded comparable sensitivity (0.61) but lower specificity (0.54). For both measures, sensitivity fell below 0.50 at higher cutoff values.”

The investigators noted that a minor disadvantage of the novel inventory exists in the presence of the fewer cutoffs from which to choose with the SUDEP-3. Although, they justified that the simple weighting and the small number of risk factors also simplifies the risk prediction while enhancing the clinical practicality of the SUDEP-3. “

“While this inventory was valid in our cohort, additional validation in other, particularly larger groups, is also needed,” Nei told NeurologyLive. “We hope that this tool will be useful for both researchers and clinicians. In SUDEP research, a simple and reliable indicator of SUDEP risk is needed, and we hope this inventory may serve that purpose. Reliable differentiation of low- versus high-risk SUDEP patients may help in the identification of specific SUDEP biomarkers and aid in the design of prospective studies, which may ultimately lead to specific interventions that could prevent SUDEP.”

Improving SUDEP education and discussion has been an ongoing effort in the field of epilepsy care in recent years. Just recently, NeurologyLive spoke with Fabio Nascimento, MD, clinical fellow, Massachusetts General Hospital, to discuss the motivation behind a study he conducted that identified an alarming gap in knowledge of SUDEP among medical trainees, as well as the importance behind identifying these gaps in epilepsy education. Watch below as he shares his insight into the findings and his concerns about these education and communication gaps.

1. Rasekhi RT, Devlin KN, Mass JA, et al. Improving prediction of sudden unexpected death in epilepsy: From SUDEP-7 to SUDEP-3. Epilepsia. Published online June 4, 2021. doi: 10.1111/epi.16928.
2. DeGiorgio CM, Miller P, Meymandi S, et al. RMSSD, a measure of vagus-mediated heart rate variability, is associated with risk factors for SUDEP: the SUDEP-7 inventory. Epilepsy Behav. 2010;19(1):78–81. doi: 10.1016/j.yebeh.2010.06.011.
3. Novak JL, Miller PR, Markovic D, Meymandi SK, DeGiorgio CM. Risk assessment for sudden death in epilepsy: the SUDEP-7 inventory. Front Neurol. 2015;9(6):252. doi: 10.3389/fneur.2015.00252.