1107 Objective-Subjective Discrepancy in Sleep Assessment in OSA Patients with a History of Depression

Abstract Introduction Depression is associated with sleep disturbance on both PSG and self-reported measures, such as PSQI. However, PSQI and PSG variables may relate differently to depression symptoms. Additionally, a discrepancy between objective and subjective sleep in depression was found to correlate with OSA severity. This study analyzes objective-subjective mismatch in patients with depression undergoing OSA evaluation. Methods From 2001 patients referred for diagnostic PSGs, 49 patients with depression (30 women, 33 Caucasian, MAge=45.8±16.9) were selected. Major medical and other psychiatric conditions were excluded. PSQI estimates of habitual total sleep time (TST), sleep latency (SL) and sleep efficiency (SE), and post-PSG estimates of in-lab TST, SL and WASO were used to calculate mismatch indices (MI) with respective PSG values (MI=objective-subjective/objective). Multiple regressions were used to predict each MI based on objective TST, SL, SE, N1%, N3%, REM%, REM latency, WASO, #awakenings, total arousal index (TAI), AHI, SpO2nadir, and self-reported PSQI-Global, ISI, ESS, Center for Epidemiologic Studies Depression Scale-Revised (CESDR). Age, sex, BMI were used as covariates. Results Relative to PSG, PSQI estimates showed significant underestimation of TST (p=0.028) and SE (p=0.008), with greater underestimation of SE predicted by lower AHI (p 0.001) and, marginally, by higher PSQI-Global (p=0.059). MIs for PSQI-estimated TST and SL did not relate to any explanatory variables. Post-PSG estimates showed significant overestimation of WASO (p 0.001) and marginal underestimation of SL (p=0.077). In regressions of post-PSG MIs, overestimation of WASO was predicted by lower ESS (p=0.039) and, marginally, by lower CESDR (p=0.054); underestimation of TST was predicted by lower AHI (p=0.038); overestimation of SL was predicated by higher PSQI-Global (p=0.003) but lower ISI (p=0.024). Conclusion This sample of patients with depression evidenced disparities between objective and subjective sleep measures. Underestimation of SE on PSQI and underestimation of TST on post-PSG self-report were both predicted by lower AHI. Roles of depression symptoms and standardized sleep questionnaires in objective-subjective mismatch were inconsistent. Mechanisms linking lower AHI with perception of shorter/less efficient sleep in depression should be further explored. Support (if any) none