Comparison of empirical continuous positive airway pressure (CPAP) treatment versus initial portable sleep monitoring followed by CPAP treatment for patients with suspected obstructive sleep apnoea

Abstract

Background: Polysomnography is labour-intensive for diagnosing obstructive sleep apnoea (OSA). We compared two algorithms for initiating continuous positive airway pressure (CPAP) treatment for patients with suspected OSA.

Methods: Symptomatic OSA patients were randomised into either algorithm I or II. Algorithm I consisted of an empirical CPAP trial whereas algorithm II utilised an Apnea Risk Evaluation System, a wireless device applied on the forehead, for establishing a diagnosis before a CPAP trial for 3 weeks. Primary outcome was success of CPAP trial, defined as CPAP usage > 4 h/night and willingness to continue CPAP. Subjective usefulness of CPAP, accuracy of Apnea Risk Evaluation System versus polysomnography and CPAP adherence at 6 months were secondary end-points.

Results: Altogether 138 patients in algorithm I and 110 patients in algorithm II completed the CPAP trial. There were no significant differences between these algorithms with respect to the primary end-point. The sensitivity and specificity of algorithm I versus II as a diagnostic test for OSA were 0.3, 0.8 versus 0.31, 1.00 respectively. In predicting CPAP adherence at 6 months, the likelihood ratio positive for algorithms I and II was 2.7 and 5.27 respectively. The mean (SE) time taken from the first consultation to the end of CPAP trial in algorithm I and algorithm II was 60 (2) and 98 (5) days, respectively, P < 0.01.

Conclusion: An ambulatory approach with portable sleep monitoring for diagnosing OSA before a CPAP trial can identify more patients who would adhere to CPAP at 6 months than empirical CPAP treatment alone.