Rationale: Obstructive sleep apnea (OSA) is an independent risk factor for the development of insulin resistance (IR). Treatment with continuous positive airway pressure (CPAP) for OSA has shown conflicting results on IR.
Objectives: To conduct a meta-analysis of randomized controlled trials (RCTs) that have evaluated the effect of CPAP on a validated index of IR, the homeostasis model assessment of insulin resistance (HOMA-IR).
Methods: PubMed and Embase were searched through August 10, 2012. Two independent reviewers screened citations to identify trials of the effect of CPAP on HOMA-IR. Data were extracted for postintervention HOMA-IR values.
Measurements and main results: A total of five studies that enrolled 244 subjects (83% male) met the inclusion criteria. None of the subjects in the included studies had diabetes. The pooled estimate of the difference in means in HOMA-IR between the CPAP and sham CPAP/control groups was -0.44 (95% confidence interval, -0.82 to -0.06; P = 0.02). The funnel plot does not suggest the presence of any publication bias. The I-squared index for the data on difference in means in HOMA-IR between the CPAP and sham CPAP/control groups was 0.00 (P = 0.61).
Conclusions: The pooled estimate of RCTs shows a favorable effect of CPAP on insulin resistance as measured by HOMA-IR in patients with OSA without diabetes. The effect size on HOMA-IR is modest, but not insignificant, when compared with the effects of thiazolidinedione in nondiabetic patients with metabolic syndrome. Further research and RCTs are warranted involving a larger number of patients and longer treatment periods to determine the beneficial effects of CPAP on IR.