Patients with untreated obstructive sleep apnea hypopnea (OSAH) are predisposed to developing hypertension, and therapy with continuous positive airway pressure (CPAP) may reduce blood pressure (BP). The purpose of this study was to assess the impact of CPAP therapy on BP in patients with OSAH. We performed a comprehensive literature search up to July 2006 [Medline, PubMed, EMBASE, Cochrane Database of Systematic Reviews (CDSR), Cochrane controlled trials register (CCTR), and Database of Abstract and Reviews of Effect (DARE)] to identify clinical studies and systemic reviews that examined the impact of CPAP on BP. Studies were included if they (1) were randomized controlled trials with an appropriate control group, (2) included systolic and diastolic BP measurements before and after CPAP/control in patients with OSAH, and (3) contained adequate data to perform a meta-analysis. To calculate pooled results, studies were weighted by inverse variances, with either a fixed or a random effects model used depending on the presence of heterogeneity (assessed with Q test). Ten studies met our inclusion criteria (587 patients): three studies were crossover (149 patients) and seven were parallel in design. Seven studies (421 patients) used 24-h ambulatory BP and three used one-time measurements. Two studies were of patients with heart failure (41 patients). Overall, the effects of CPAP were modest and not statistically significant; CPAP (compared to control) reduced systolic BP (SBP) by 1.38 mmHg (95% CI: 3.6 to -0.88, p = 0.23) and diastolic BP (DBP) by 1.52 mmHg (CI: 3.1 to -0.07; p = 0.06). Six of the trials studied more severe OSAH (mean AHI > 30/h, 313 patients); in these six trials, CPAP reduced SBP by 3.03 mmHg (CI 6.7 to -0.61; p = 0.10) and DBP by 2.03 mmHg (CI: 4.1 to -0.002; p = 0.05). There was a trend for SBP reduction to be associated with CPAP compliance. In unselected patients with sleep apnea, CPAP has very modest effects on BP. However, we cannot exclude the possibility that certain subgroups of patients may have more robust responses-this may include patients with more severe OSAH or difficult-to-control hypertension. Future randomized controlled trials in this area should potentially concentrate on these subgroups of patients.