0419 Phenotyping of Insomnia Based on a Single vs. Two Nights of Polysomnography

Abstract Introduction In 2009, Vgontzas and his colleagues proposed phenotyping of insomnia with short sleep duration (ISSD) vs. normal sleep duration (INSD) using a cutoff point of total sleep time < 6 hours from a single night of polysomnography (PSG). Multiple studies since then have shown that the ISSD phenotype is associated with adverse health outcomes, including hypertension, diabetes, heart disease, stroke and cognitive impairment. However, a limitation of this approach is the potential misclassification based on a single night due to the “first night effect”. In this study we examined how phenotype assignment differs based on a single night vs. the mean of two nights of PSG. Methods SONO is an ongoing 4-site clinical trial UG3/UH3 (Hershey, Pittsburg, Denver and Quebec City) designed to examine the modifying role of insomnia phenotyping on the efficacy of cognitive-behavioral therapy for insomnia (CBT-I) and the efficacy of trazodone vs. placebo among non-remitters to CBT-I. All participants meet diagnostic criteria for chronic insomnia disorder and report an insomnia severity index score of ≥ 10. In this study, we included a subset of the first participants (n=191) enrolled in the SONO study (67% females, median age 54 (19,91) years) who had two nights of PSG at baseline. We used cross-tabulations to examine the degree of misclassification based on the first night vs. two nights (gold standard). Results We observed a strong agreement between night 1 vs the mean of 2 nights of PSG (85.3%). Further, in a sensitivity analysis after excluding those that the first night TST was within ± 15 minutes of the 6 hours PSG threshold (n=31, 16.2%), agreement between night 1 and the mean of 2 nights was 91.2%. Conclusion A single night of PSG is highly reliable in classifying the two insomnia phenotypes. Misclassification based on a single night PSG was between 8.8% and 14.7%, similar to other tests used routinely in sleep medicine to determine sleep apnea or daytime sleepiness (e.g., multiple sleep latency test). Finally, in-lab TST is a stable, trait-like characteristic that probably reflects a person’s habitual sleep. Support (if any) UH3HL161342 and 5U24HL161359