Authors :
Presenting Author: Jad El Ahdab, MD – Cleveland Clinic
Tyler Bare, MD – Cleveland Clinic
Matheus Lima de Araujo, PhD – Cleveland Clinic
James Bena, MS – Cleveland Clinic
Ryan Delaney, MS – Cleveland Clinic
Madeleine Grigg-Damberger, MD – University of New Mexico
Nancy Foldvary-Schaefer, DO, MS – Cleveland Clinic
Rationale:
Obstructive Sleep Apnea (OSA) is highly prevalent and an important, often overlooked comorbidity in adults with epilepsy (AWE). Screening tools including STOP (Snoring, Tiredness, Observed apneas, high blood Pressure) and STOP-BANG (+BMI, Age, Neck circumference (NC), and Gender) are widely used in general populations. Novel models like STOP-BAG2 may improve OSA risk stratification. Performance characteristics of OSA instruments have not been studied in large AWE populations.
Methods:
We analyzed 1048 AWE with polysomnography (PSG) and complete STOP-BANG responses at Cleveland Clinic. Primary outcome was apnea-hypopnea index (AHI) ≥ 5. We compared STOP, STOP-BANG, STOP-BAG (excluding NC), and STOP-BAG2, a logistic regression tool created at Cleveland Clinic and validated in stroke populations. Sensitivity, specificity, NPV, PPV, and AUC were estimated with 95% confidence intervals. AUC values > 0.5 indicated better-than-chance prediction, with higher values indicating better discrimination for AHI > 5 versus AHI < 5. Seizure severity and frequency were assessed by Liverpool Seizure Severity Scale (LSSS). Instrument performance was compared by estimating differences in AUC using a significance level of 0.05.