Persistent Sleepiness on PAP Despite a 'Normal' Report
A 52-year-old man with obesity (BMI 38) and hypertension was diagnosed with severe OSA (AHI 62, nadir SpO2 71%). He was started on auto-PAP at 5-15 cmH2O with a nasal mask. At 3 months, the device download shows 7 nights of use per week, mean usage 6.8 h/night, residual AHI 3.1, mask leak 24 L/min (95th percentile), and pressure 95th percentile 13 cmH2O. He returns saying he is still 'exhausted.' He reports no morning headaches, no nocturia, and no witnessed apneas. On detailed questioning, he admits to falling asleep at red lights about once a week and scores 16 on the Epworth Sleepiness Scale. He takes amlodipine and hydrochlorothiazide. He has gained 4 kg since titration.
1. Is the residual AHI of 3.1 sufficient to conclude his PAP therapy is adequately controlling his disease? What specific features of the download should be scrutinized before accepting this as treatment-emergent central sleep apnea or as simply undertreated OSA?
Read the worked answer
A residual AHI of 3.1 is conventionally 'acceptable' (<5), but it is not proof of effective therapy. Key checks:
- Leak: 24 L/min at the 95th percentile is at the upper limit of acceptable (large-leak threshold often ~24 L/min). Significant leak can cause under-scoring of events and patient-ventilator asynchrony. Review the leak trace for sustained periods above threshold rather than the summary statistic.
- Residual event composition: A residual index of 3 can be dominated by central apneas. Treatment-emergent central sleep apnea (TECSA) may appear after PAP initiation, especially with high pressure and in patients with cardiac or opioid risk factors. Distinguish obstructive vs central events on the flow/pressure waveform; TECSA often improves with time but may require ASV if persistent and symptomatic (and ASV is contraindicated in symptomatic HFrEF with LVEF ≤45%).
- Pressure adequacy: 95th percentile of 13 cmH2O with a wide range suggests the auto algorithm may be under-titrating during REM/supine. Consider a formal in-lab titration or review of pressure vs. event timing.
- Mask/interface and sleep architecture: Poor sleep quality, insufficient total sleep time, or a competing disorder (e.g., insufficient sleep, circadian delay, depression) can produce sleepiness despite good AHI control.
So the AHI alone does not settle the question; the download must be interpreted in context, and the persistent sleepiness warrants evaluation for alternative causes.
2. He remains sleepy (ESS 16) with good adherence. Outline a stepwise management plan, including what you would NOT do and why.
Read the worked answer
Step 1 - Optimize PAP: Address leak (mask refit, consider nasal pillows or full-face), verify pressure adequacy with a titration study, and inspect for residual central events. Reassess after 4–6 weeks.
Step 2 - Exclude other causes of sleepiness:
- Insufficient sleep syndrome (sleep log/actigraphy).
- Circadian phase delay.
- Depression/anxiety and sedating medications (none obvious here, but review).
- Restless legs/PLMD.
- Consider MSLT only after optimizing PAP and documenting adequate sleep (≥7 h) and no significant circadian disruption; a diagnostic MSLT on PAP is appropriate if narcolepsy/idiopathic hypersomnia is suspected.
Step 3 - Weight and comorbidity: 4 kg gain is relevant; refer to weight management. Screen for hypothyroidism and diabetes.
What NOT to do:
- Do not simply add a stimulant (modafinil/armodafinil) before excluding residual OSA, TECSA, and insufficient sleep. Wake-promoting agents are not a substitute for effective PAP.
- Do not empirically increase pressure without data; higher pressure can worsen central events and leak.
- Do not order an MSLT while the patient is on suboptimal PAP or sleep-deprived - it will be uninterpretable.
- Do not diagnose narcolepsy without cataplexy/MSLT criteria.