Double-click to edit...
There is a clinical scenario that comes up more often than most people expect: a patient with obstructive sleep apnea who is using their CPAP consistently, whose respiratory events are controlled, and who is still not sleeping.
The device is working. The apnea is being treated. And yet the patient lies awake at 2am, exhausted, frustrated, and no better off than before the diagnosis.
This is not a CPAP failure. It is likely a second sleep disorder that was present all along and never addressed.
Sleep Apnea and Insomnia Often Coexist
Obstructive sleep apnea and insomnia are the two most common sleep disorders in adults. They also frequently occur together — a combination researchers call COMISA, for comorbid insomnia and sleep apnea.
Studies suggest that approximately 30 to 50 percent of patients presenting for OSA treatment also report clinically significant insomnia symptoms. That is not a small overlap. It means that for a substantial portion of the patients being treated for sleep apnea, the diagnosis is correct but incomplete.
The clinical implication is straightforward: treating OSA does not necessarily resolve coexisting insomnia. A patient whose AHI is well-controlled may still have difficulty initiating sleep, maintaining sleep, or both — not because the CPAP is failing, but because the insomnia was never the apnea’s fault in the first place.
The Relationship Goes Both Ways
What makes COMISA clinically interesting is that the relationship between the two conditions is bidirectional.
Most clinicians are aware that untreated sleep apnea disrupts sleep architecture and produces insomnia-like symptoms. Less widely known is the evidence that insomnia may actively worsen sleep apnea.
In a randomized controlled trial of patients with both insomnia and moderate-to-severe OSA, CBT-I reduced the apnea-hypopnea index by approximately 15 percent compared with controls (Sweetman et al., 2020), with a 2025 study finding a similar result (Brooker et al., 2025). The effect is modest, and CBT-I is not a substitute for PAP therapy — but it suggests that the two conditions are not simply parallel. They interact.
The proposed mechanism involves the respiratory arousal threshold — a physiological trait that determines how easily a person wakes in response to minor breathing disturbances during sleep. Patients with a low arousal threshold, common in insomnia, wake readily to the same disruptions that OSA produces. Treating the insomnia appears to increase slow-wave sleep, a stage during which breathing is more stable, which may reduce the frequency of apneic events.
There is also a practical benefit: patients who completed CBT-I before starting PAP therapy used their CPAP approximately one hour more per night than those who did not (Sweetman et al., 2019). Insomnia and CPAP non-adherence are closely linked. Treating one may meaningfully improve the other.
What This Means Clinically
A few things worth keeping in mind when a patient’s sleep remains disrupted despite adequate OSA treatment:
A controlled AHI does not mean insomnia has resolved. The two conditions require separate assessment and, when both are present, separate treatment.
Asking about sleep initiation and maintenance is worth doing explicitly. Patients are often not reporting insomnia symptoms because no one has asked, or because they attribute all their sleep difficulties to the apnea. A brief validated measure like the Insomnia Severity Index takes minutes and can clarify the clinical picture considerably.
CBT-I has demonstrated efficacy in patients with comorbid insomnia and OSA, including those on PAP therapy. It is not contraindicated by the apnea diagnosis and does not need to wait until OSA is fully optimized.
For newly diagnosed OSA patients who also report insomnia — particularly those who are hesitant about starting PAP — CBT-I before or alongside CPAP initiation may improve both sleep outcomes and device adherence.
A Note on Referral
Sleep medicine and behavioral sleep medicine address different parts of the same problem. A patient whose respiratory events are controlled but whose sleep is still impaired is not a PAP failure — they may simply need the other half of the treatment picture.
References
Sweetman, A., Lack, L., McEvoy, R. D., Smith, S., Catcheside, P. G., Antic, N. A., Chai-Coetzer, C. L., Douglas, J., O’Grady, A., Dunn, N., Robinson, J., Paul, D., & Clements, J. (2019). Bi-directional relationships between co-morbid insomnia and sleep apnea (COMISA). Sleep Medicine Reviews, 45, 37–46. https://doi.org/10.1016/j.smrv.2019.03.005
Sweetman, A., Lack, L., Bastien, C., Knolle, A., Catcheside, P., Donn, N., Paul, D., & McEvoy, R. D. (2020). Cognitive behavioural therapy for insomnia reduces sleep apnea severity: A randomized controlled trial. ERJ Open Research, 7(1). https://doi.org/10.1183/23120541.00160-2020
Brooker, E. J., Sweetman, A., Reynolds, A. C., & Adams, R. J. (2025). Effect of CBT-I on OSA endotypes and slow-wave sleep. Journal of Clinical Sleep Medicine. [Add DOI when available]
Morin, C. M., & Espie, C. A. (2003). Insomnia: A clinical guide to assessment and treatment. Springer.