0656 Psychometric Validation of the PROMIS Fatigue 8a, PROMIS Sleep Impairment 8a and ESS in Adults with Mild to Severe Obstructive Sleep Apnea

Abstract Introduction AD109, an investigational oral combination therapy of the antimuscarinic, aroxybutynin (2.5mg) with the selective norepinephrine reuptake inhibitor, atomoxetine (75mg), significantly improved objective measures of airway obstruction and oxygenation vs placebo (PBO) in two phase 3 clinical trials in obstructive sleep apnea (OSA). Patient-reported outcomes (PROs) were also assessed. This study aimed to expand the existing literature on their fit for purpose use in OSA by assessing the psychometric properties of the Patient Reported Outcome Measurement Information System (PROMIS) - Fatigue (F) 8a, PROMIS-Sleep Impairment (SI) 8a, and Epworth Sleepiness Scale (ESS) in the AD109 Phase 3 trials. Methods Analyses were conducted using SynAIRgy (NCT05813275; N=646) and LunAIRo (NCT0811247; N=660) randomized, double-blind, PBO-controlled, parallel-arm phase 3 trials of AD109 vs PBO in adults with mild-to-severe OSA (AHI>5) intolerant to or refusing positive airway pressure (PAP) therapy. PROs included PROMIS-F 8a, PROMIS-SI 8a, ESS, Patient Global Impression of Severity-Fatigue (PGI-S) and Patient Global Impression of Change-Fatigue (PGI-C). Psychometric analyses included internal consistency, test-retest reliability (TRTR), convergent and known-groups validity, responsiveness to change and meaningful within-patient change (MWPC) thresholds. Results Reliability of all three PROs was evidenced by excellent internal consistency, with McDonald’s ω and Cronbach’s α estimates consistently exceeding acceptable thresholds (> 0.70). TRTR was consistently moderate to strong, particularly when stability was defined with PGI- anchors of fatigue (intraclass correlation coefficients >0.70). Convergent validity was supported by strong correlations with conceptually related measures with all correlations meeting or exceeding predefined thresholds. Known-groups validity analyses supported the ability to differentiate levels of symptoms severity across PGI-S responses, and correlations of change over time demonstrated that each PRO is sensitive to detect change at week 26. Anchor-based analyses of clinically MWPC improvement thresholds were - 8.40 (supplementary range, 6.0 to -9.0) for PROMIS-F 8a T-score; -9.0 (-7.5 to -10.0) for PROMIS-SI 8a T-score; and -2.0 (-2.0 to -3.0) for ESS in the OSA population. Conclusion PROMIS-Fatigue 8a, PROMIS-SI 8a and ESS are reliable and valid measures of OSA-related symptoms, confirming their fitness for the purpose of supporting clinical trial endpoints in the OSA population, in accordance with the FDA PRO guidance. Support (if any) Apnimed, Inc.