STUDY OBJECTIVES: To assess the impact of the non-respiratory arousal burden at baseline polysomnography (PSG) on residual daytime sleepiness in positive airway pressure (PAP)-treated obstructive sleep apnea (OSA). METHODS: We included OSA patients from the European Sleep Apnea Database registry with available arousal data who had at least 2 treatment follow-up visits. The primary outcome was the Epworth Sleepiness Scale (ESS) score under PAP. The non-respiratory arousal ratio (NRAR) was defined as the ratio of non-respiratory to total arousals at baseline PSG. A linear mixed model tested the effect of NRAR tertiles on residual sleepiness. Baseline variables that differed significantly between groups were included as covariates. RESULTS: 800 patients with OSA (69.6 % male, mean age 57.1 ± 12.0 years, mean NRAR 0.22 ± 0.20) were evaluated during three follow up visits at a mean of 197.4, 499.3, and 731.6 days after PAP initiation. The interaction between time and NRAR tertile was statistically significant (F = 4.55, p = 0.001). The lowest NRAR tertile was associated with lower residual sleepiness over time compared to the highest NRAR tertile. The associations were independent of sex, comorbidities, body mass index, blood pressure, baseline apnea-hypopnea index, and baseline ESS score. CONCLUSIONS: NRAR at baseline PSG predicts residual sleepiness in PAP-treated OSA patients. The findings offer new insights into OSA phenotyping and have important implications for patient care.