O068 Slow frequency wake EEG and hypoxemia during sleep predicts vigilance performance following extended wakefulness in patients with OSA

STUDY OBJECTIVES: Obstructive sleep apnea is highly prevalent, but heterogeneous making identifying patients at risk of driving impairment challenging. Resting wake electroencephalography (qEEG) is associated with vigilance performance in healthy subjects. We examined if qEEG and clinical measures predict vigilance impairment in OSA. METHODS: Fifty-four patients underwent baseline PSG and 28hr extended wakefulness (EW) with 2 hourly psychomotor vigilance test (PVT) and Karolinska drowsiness test and driving simulator performance at baseline and following EW. Resting wake EEG was subjected to fast Fourier transform analysis to derive EEG power in delta, theta, alpha and beta ranges. Cluster analysis assigned participants into groups of either resistant (n=39) or vulnerable (n=15) to alertness failure based on PVT and driving performance. Backward stepwise regression models were used to determine the strongest clincal and wake EEG predictors vigilance impairment. RESULTS: Compared with resistant patients, the vulnerable group exhibited increased wake EEG delta (p<0.001) and theta (p=0.003) power during eyes open across all testing periods. The only significant predictors of vulnerability to alertness failure (PVT and driving) included O2Nadir from baseline PSG and baseline wake EEG theta and delta (eyes open) respectively explaining 42 and 32% of the variance in vigilance outcome in models that included Age, BMI, ESS, SOL, TST, AHI, O2 Nadir. CONCLUSIONS: Slow frequency wake EEG (delta and theta) and hypoxemia are predictive of future alertness failure and driving impairments in patients with OSA. These findings are important for fitness to drive assessment in OSA but require validation in independent sample.