Sudden Unexpected Death in Epilepsy: A Personalized Prediction Tool

Objective: To develop and validate a tool for individualised prediction of
Sudden Unexpected Death in Epilepsy (SUDEP) risk, we re-analysed data from
one cohort and three case-control studies undertaken 1980-2005. Methods:
We entered 1273 epilepsy cases (287 SUDEP, 986 controls) and 22 clinical
predictor variables into a Bayesian logistic regression model. Results:
Cross-validated individualized model predictions were superior to baseline
models developed from only average population risk or from generalised
tonic-clonic seizure frequency (pairwise difference in
leave-one-subject-out expected log posterior density = 35.9, SEM +/-12.5,
and 22.9, SEM +/-11.0 respectively). The mean cross-validated (95%
Bootstrap Confidence Interval) Area Under the Receiver Operating Curve was
0.71 (0.68 to 0.74) for our model versus 0.38 (0.33 to 0.42) and 0.63
(0.59 to 0.67) for the baseline average and generalised tonic-clonic
seizure frequency models respectively. Model performance was weaker when
applied to non-represented populations. Prognostic factors included
generalized tonic-clonic and focal-onset seizure frequency, alcohol
excess, younger age of epilepsy onset and family history of epilepsy.
Anti-seizure medication adherence was associated with lower risk.
Conclusions: Even when generalised to unseen data, model predictions are
more accurate than population-based estimates of SUDEP. Our tool can
enable risk-based stratification for biomarker discovery and
interventional trials. With further validation in unrepresented
populations it may be suitable for routine individualized clinical
decision-making. Clinicians should consider assessment of multiple risk
factors, and not only focus on the frequency of convulsions.