Clinical and polysomnographic predictors of suboptimal auto-adjusting CPAP titration in adult OSA patients: a single-center study.

Abstract

To examine potential clinical, demographic, anthropometric, and polysomnographic predictors of successful auto-adjusting continuous positive airway pressure (CPAP) titration for treatment of obstructive sleep apnea (OSA). This cross-sectional study was conducted in adults diagnosed with moderate-to-severe OSA (baseline apnea-hypopnea index [AHI] ≥ 15.0/h), who underwent auto-adjusting CPAP titration (S9 or S10 AutoSet ResMed®) in a sleep laboratory setting while wearing a nasal or pillow mask. Participants were then grouped into two groups: optimal CPAP titration (residual AHI < 5.0/h) or suboptimal CPAP titration (residual AHI ≥ 5.0/h). Multivariate logistic regression analysis was used to assess possible independent predictive factors for suboptimal CPAP titration. A total of 1222 adults consisting of 874 subjects with optimal CPAP titration (71.5%) and 348 subjects with suboptimal CPAP titration (28.5%) were evaluated. Multivariate analysis resulted in a model with an adequate calibration (Hosmer-Lemeshow chi-square-test: 7.088; p = 0.527), with male sex, higher values of baseline AHI, therapeutic pressure (95th percentile), and mask leak (95th percentile) emerging as significant and independent predictors for suboptimal CPAP titration: adjusted odds ratio (OR): 1.456 (95% confidence interval [CI] 1.076-1.971; p = 0.015), OR: 1.009 (95% CI 1.002-1.016; p = 0.013), OR: 1.281 (95% CI 1.206-1.361; p < 0.001), and 1.035 (1.026-1.043; p < 0.001), respectively. In a large cohort of adults undergoing auto-adjusting CPAP titration due to moderate-to-severe OSA, male sex, increased values of baseline AHI, pressure requirements, and mask leak were significant predictors for less than optimal CPAP titration.