1031 Objective-Subjective Discrepancy in Sleep Assessment in OSA Patients with a History of Tinnitus

Abstract Introduction Approximately 50-70% of tinnitus patients report disturbed sleep and insomnia on PSQI and ISI. However, a PSG evaluation showed higher total sleep time (TST) and N3% and lower total arousal index (TAI) and AHI in a tinnitus group compared to no-tinnitus. This study explores reasons for mismatch between objective and subjective sleep in tinnitus patients evaluated for OSA. Methods From 2001 consecutive diagnostic PSGs, we selected 118 patients without major medical or psychiatric history who reported tinnitus on a screener questionnaire (61 women, 43 Caucasian, M[Age]=48.2±15.6). Tinnitus Handicap (TH) Inventory yielded 4 scores: Functional-TH, Emotional-TH, Catastrophic-TH, Total-TH. Objective-subjective mismatch indices (MI=[objective-subjective]/objective) were calculated for PSQI reports of habitual TST, sleep latency (SL) and sleep efficiency (SE), and for post-PSG self-reports of in-lab TST, SL and WASO. MIs were outcomes in multiple regressions. Predictor variables were: objective TST, SL, SE, N1%, N3%, REM%, REM latency, WASO, #awakenings, total arousal index (TAI), AHI, SpO2nadir, self-reported PSQI-Global, ISI, ESS, Center for Epidemiologic Studies Depression Scale-Revised (CESDR), TH scores. Age, sex, BMI were covariates. Results Greater underestimation of TST on PSQI was predicted by lower Functional-TH (p=0.006), higher Emotional-TH (p=0.015), higher N3% (p=0.039), higher objective-SE (p=0.043), more objective awakenings (p=0.012), higher PSQI-Global (p< 0.001). Greater post-PSG underestimation of TST was predicted by higher Emotional-TH (p=0.052), lower Catastrophic-TH (p=0.006), longer REM latency (p=0.006), lower objective-WASO (p=0.001), lower objective-SE (p=0.028), shorter objective-SL (p=0.005). Greater PSQI overestimation of SL was predicted by higher objective-WASO (p=0.026), higher objective-SE (p=0.021), higher PSQI-Global (p=0.001), lower ISI (p=0.019). Greater post-PSG overestimation of SL was predicted by higher objective-WASO (p=0.004), higher objective-SE (p=0.011), lower N1% (p=0.037). Greater PSQI underestimation of SE was predicted by shorter objective-SL (p=0.026), longer REM latency (p=0.043), higher PSQI-Global (p< 0.001). Post-PSG estimation of WASO did not relate to any predictor variables. Conclusion Three aspects of tinnitus handicap contributed to objective-subjective mismatch in TST. Overall, MIs showed inconsistent relationships with objective and subjective sleep variables, which should be clarified in future studies. There were no relationships of MIs with depression symptoms. Associations of PSQI-Global with several MIs suggest a cautious interpretation of self-reported sleep disturbance in tinnitus patients. Support (if any) none