Can residual obstructive sleep apnea persist during positive airway pressure therapy?

Residual sleep-disordered breathing can persist despite positive airway pressure (PAP), and device-reported AHI alone may not establish physiologic control. The claim is substantially supported for pressure mismatch and the downstream potential for intermittent hypoxia and sleep fragmentation; leak and sleep position are credible contributors. Oxygen instability itself should not be viewed as a cause of residual obstruction.

Clinical and practical evidence

• Pressure mismatch is a well-supported cause of residual obstruction. PAP titration guidance directs pressure increases for persistent obstructive apneas, hypopneas, respiratory-effort–related arousals (RERAs), and snoring, aiming for residual respiratory disturbance below 5 events/hour. With bilevel PAP, residual obstructive apneas generally call for EPAP adjustment, whereas hypopneas/RERAs may require IPAP adjustment.
• Leak is a plausible contributor and a major measurement problem. Greater unintentional leak has been associated with lower delivered auto-PAP pressure and more residual device-detected events. Intermittent leak underestimated device AHI by 33.8% in bench testing; therefore, reassuring download summaries can miss clinically important residual disease.
• Supine sleep can expose inadequate pressure. Mean optimal CPAP was 10.0 cm H₂O supine versus 7.6 cm H₂O lateral. A pressure effective laterally may therefore fail after a patient turns supine.

Mechanisms and consequences

• Residual respiratory events may sustain hypoxic burden and arousal-related sleep fragmentation. In CPAP users, device AHI exceeded 5/hour in 32.3%, while independent testing can identify REM-related events, RERAs, and hypoxemia missed by device algorithms.
• Leak may also promote treatment-emergent central apnea, potentially through proximal CO₂ washout, reinforcing the need to distinguish obstructive from central residual events.

Oxygenation interpretation

• Persistent desaturation is clinically important but is not an established driver of residual obstructive apnea. Even with device AHI <5/hour, 28.3% of nights had ≥5 minutes below 90% saturation and 10.6% had time below 88%, potentially reflecting hypoventilation, cardiopulmonary disease, REM physiology, or under-detection.

Bottom line

• Review pressure adequacy, leak patterns, event phenotype, REM/supine vulnerability, and oxygenation; persistent symptoms or discordant data warrant oximetry and often attended PAP polysomnography rather than relying on device AHI alone.