0371 The Odds Ratio Product: A Novel Tool to Identify Features of Comorbid Insomnia and Sleep Apnea
Umaer Hanif,Alyssa Cairns,Veronica Guadagni,Kari Lambing,Guy Duke,Vincent Mysliwiec
DOI: https://doi.org/10.1093/sleep/zsae067.0371
IF: 6.313
2024-04-20
SLEEP
Abstract:Abstract Introduction Many patients presenting with either OSA or insomnia symptoms exhibit signs of the other condition. This overlap, known as COMISA, is often undetected due to differing diagnostic methods PSG for OSA and clinical symptoms for insomnia. The lack of objective diagnostic criteria for COMISA contributes to its low diagnosis rate. We sought to address this gap by investigating whether the Odds Ratio Product (ORP) algorithm can accurately distinguish between individuals with COMISA, OSA patients, and controls. Methods 2000 patients (50% male/female) were selected from a large clinical cohort, equally grouped into 5 categories of 400 patients: clinic controls (AHI < 5, no insomnia), mild OSA (AHI 5-15), mild OSA and insomnia (defined by presence of self-reported complaints of difficulty initiating sleep, maintaining sleep, experiencing restless sleep, and daytime fatigue), moderate-severe OSA (AHI ≥ 15, ODI ≥ 15), and moderate-severe OSA and insomnia. The OSA and COMISA subgroups were matched by AHI. The ORP was computed using a 3-second sliding window, averaged across each 30-second epoch of PSG total recording time (TRT). Extracted ORP features included mean, minimum, and percent TRT spent in ORP deciles 1-10 (1=deepest sleep; 10=fully awake). Univariate ordinal regression was utilized to classify disease subgroups using each ORP feature as input. Results Each disease subgroup was compared to clinic controls (reference group) and statistically presented in the following rank-ordered format (mild OSA, mild COMISA, moderate-severe OSA, moderate-severe COMISA). Significant discriminatory ORP features for disease subtype were mean ORP in NREM (ORs: 3.51, 9.35, 17.3, 43.8, p < 0.001), minimum ORP (ORs: 5.67, 6.33, 5.62, 12.3, p < 0.001), and percent TRT spent in decile 2 “deep sleep” (ORs: 0.982, 0.969, 0.963, 0.947, p < 0.01) and 7 “transitional sleep” (ORs: 1.05, 1.08, 1.09, 1.11, p < 0.01). Conclusion This study demonstrates the potential utility of the ORP algorithm to identify and discriminate between OSA and COMISA subgroups, the latter being underdiagnosed and inadequately treated in the routine sleep clinic paradigm. These findings suggest that ORP may be a valuable tool in the polysomnographic assessment of comorbid sleep disorders, offering a pathway towards an objective assessment and targeted treatment strategies. Support (if any)
neurosciences,clinical neurology