Loop gain predicts residual sleep apnoea among people using positive airway pressure

Abstract

Rationale Residual sleep apnoea-defined by an Apnoea-Hypopnoea Index >= 10 events/hour-affects similar to 20% of obstructive sleep apnoea patients treated with positive airway pressure therapy and poses a major clinical challenge. Ventilatory control instability is a plausible cause of residual sleep apnoea. Elevated loop gain (LG), a measure of ventilatory instability, may be a risk factor, but this has not been rigorously tested.

Objective To assess whether high LG at baseline is associated with residual sleep apnoea on positive airway pressure therapy in two large, randomised control trials: Apnoea Positive Pressure Long-Term Efficacy Study (APPLES) and Randomised Intervention with CPAP in Coronary Artery Disease and OSA (RICCADSA).

Methods LG was estimated from baseline polysomnography using a validated method. Residual sleep apnoea was defined using polysomnography on positive airway pressure at 2 months (APPLES) or device downloads at 3 months (RICCADSA). Logistic regression estimated the odds of residual sleep apnoea with high LG (highest quartile), adjusting for confounders. A sensitivity analysis was performed using linear regression, where both the exposure and outcome were defined continuously.

Measurements and main results In the unadjusted analysis, high LG was associated with threefold odds of residual sleep apnoea in both samples. After adjustment, elevated odds persisted in both APPLES (2.17 (1.24-3.78)) and RICCADSA (3.31 (1.33-8.24)). Associations remained after accounting for measures of central sleep apnoea. Linear regression confirmed the association of LG and residual Apnoea-Hypopnoea Index.

Conclusions High LG is a significant risk factor for residual sleep apnoea on positive airway pressure therapy. Ventilatory control instability identified at baseline may warrant closer monitoring or the initiation of adjunctive therapies aimed at reducing LG and improving the therapeutic response.