Physiological Characteristics of Central versus Obstructive Sleep Apnea in Patients with Heart Failure

Central (CSA) and obstructive (OSA) sleep apnea are common in heart failure (HF) patients, and treatment options depend on differential classification. To better characterize CSA vs OSA, we evaluated the discriminative value of established and novel polysomnographic traits in patients with HF.

We analyzed LOFT-HF trial participants with reduced left ventricular ejection fraction (LVEF) and apnea-hypopnea index (AHI) >15 events/hr. Central versus obstructive events were classified with expert investigator input. Multivariable logistic regression tested whether CSA (>50% of central/mixed events; otherwise OSA) was associated with increased loop gain, and other mechanistic traits (arousal threshold, collapsibility, muscle compensation). Mechanistic traits together with novel clinically-informed traits (e.g. event periodicity, flow limitation) were modeled to provide a probability score "Pr(CSA)" for objective CSA-versus-OSA discrimination.

Among 122 HF patients (CSA:OSA=84:38), CSA was associated with higher loop gain (OR=4.4[2.2-9.1]), lower arousal threshold (6.4[2.4-16.9]), and greater muscle compensation (2.5[1.0-6.2]). Novel traits-greater event periodicity, larger event-related drive reduction, and reduced flow limitation-further improved discrimination (likelihood ratio test P=0.013 vs. mechanistic model). The final model (AUC=0.88) identified a subgroup enriched for true CSA (OR=8.4[3.5-20.1]). Sensitivity analyses using stricter central apnea-based definitions yielded similar or stronger associations with loop gain and modestly improved discrimination, although fewer individuals met CSA criteria.

Patients with HF and CSA exhibit a distinct set of polysomnographic characteristics compared to OSA, which may facilitate their objective physiology-based characterization along the CSA-OSA continuum.