I read with interest the article by Davis et al.¹ on early diagnosis and treatment of obstructive sleep apnea (OSA) after stroke. In the article, the authors exhaustively argue the negative prognostic role of OSA on stroke clinical outcome and on its recurrence, as well as evidence of the positive effect of CPAP on patients' outcome. The reasons for patients' adherence to CPAP treatment were oversimplified. The authors neglected that some pathophysiologic mechanisms responsible for sleep-disordered breathing, such as a high loop gain, can be responsible for patients' unresponsiveness to CPAP.² In fact, a central sleep apnea–hypopnea can be detected in cases with OSA-hypopnea when treated with CPAP.³ This phenomenon has been reported to occur in stroke patients, and can verify spontaneously in patients with congestive heart failure.⁴ Since CPAP treatment of stroke cases has given inconclusive results, the question arises whether clinical or diagnostic parameters, or both, are able to predict unresponsive cases; a high loop gain has been proposed as one of those parameters. A schematic representation of proposed responsible mechanisms is reported in the figure.

Authors Respond:

We appreciate the responses and agree that sleep-disordered breathing is common after stroke and often not just simple OSA. In one series of 147 patients admitted to rehabilitation after a first-time stroke, 44% had a respiratory disturbance index ≥10, with problems mainly due to obstructive or predominately obstructive events vs predominately central in only 6.1%.⁵ Another study of patients with a first-ever stroke (73%) or transient ischemic attack (24%) found that in the first 48 to 72 hours after admission, 72% of patients had an apnea-hypopnea index >10, with 52.2% having predominantly obstructive events and 38.5% with predominantly central.⁶ As Dr. Sethi notes, the best time after stroke to diagnose and treat sleep apnea remains uncertain. Patient adherence to CPAP is highest when patients are treated early and while still on inpatient rehabilitation.⁷ While sleep-disordered breathing may improve spontaneously poststroke, OSA usually persists. In the study noted above, sleep studies repeated after 3 months showed substantial reductions in the central apnea index (6.2 to 3.3) when compared to the acute phase, but the obstructive apnea index was unchanged.⁶ Thus, repeating sleep studies and retitrating CPAP in the chronic phases poststroke may be justified. Dr. Sacchetti's letter accurately highlights that CPAP adherence in some cases can be related to pathophysiologic mechanisms such as high loop gain, which effectively leads to excessive ventilatory response to changes in Pco₂.⁸ High loop gain may result in development of significant central apneas with the application of CPAP in 3% to 20% of patients with OSA, termed complex sleep apnea or CPAP-emergent central sleep apnea.^9,10 While the majority of these central apneas resolve after months of continued therapy, in a minority of patients, 1.5% in one of the largest studies,⁹ they persist and contribute to intolerance and ineffectiveness of CPAP therapy. It remains difficult to predict which patients have a high loop gain phenotype and are at the highest risk for developing complex sleep apnea refractory to CPAP. While stroke is a known risk factor for central sleep apnea, whether stroke is also a risk factor for high loop gain is unknown. Thus, although complex sleep apnea may limit adherence, whether stroke is associated with an increased risk of this phenomena requires further study.