Identifying obstructive sleep apnea in patients with epilepsy: A cross-sectional multicenter study

Several screening questionnaires for obstructive sleep apnea (OSA) have been introduced. No study has compared the performance of different questionnaire in the same patients with epilepsy. Herein, we compare the performance characteristics of four common questionnaires for assessing the probability of OSA in patients with epilepsy.

This cross-sectional multicenter study was conducted among adult epilepsy patients attending neurology and general medical clinics in Thailand. Before performing full polysomnography (PSG), all participants completed the STOP-BANG, STOP-BAG, SA-SDQ, and NoSAS questionnaires. OSA was defined by apnea/ hypopnea index (AHI) criteria of AHI: ≥ 5, ≥ 15, and ≥ 30. Discriminatory ability was assessed by area under the receiver operating characteristics (ROC) curve (AUC) and likelihood ratio. To improve discriminative ability, we created 3 ranges of the score to predict lower, middle, and higher probability of OSA as defined by each diagnostic criterion.

A total 166 patients with epilepsy were included. OSA prevalence was 38%. Overall, NoSAS had higher AUC at all AHI criteria but not significantly higher than that of other scales. Using prediction cut-points for NoSAS of ≥ 4 and ≥ 8 to predict OSA defined by AHI ≥ 5, the likelihood ratios for the 3 ranges were 0.37, 2.22 and 9.81 respectively.

Among the 4 scales, the 2-cut-point NoSAS score had the highest discriminatory ability at each AHI cutoff.