Based on meta-analyses, the BP-lowering effect of CPAP therapy in patients with OSA is reported to be approximately 2 to 3mmHg. This figure is derived from heterogeneous trials, which are often limited by poor CPAP adherence, and thus the treatment effect may possibly be underestimated. We analyzed morning BP data from three randomized controlled CPAP withdrawal trials, which included only patients with optimal CPAP compliance. Within the three trials, 149 patients with OSA who were receiving CPAP were randomized to continue therapeutic CPAP (n= 65) or to withdraw CPAP (n= 84) for 2weeks. Morning BP was measured at home before and after sleep studies in the hospital. CPAP withdrawal was associated with a return of OSA (apnea-hypopnea index [AHI] at a baseline of 2.8/h and at follow-up of 33.2/h). Office systolic BP (SBP) increased in the CPAP withdrawal group compared with the CPAP continuation group by+5.4mmHg (95%CI, 1.8-8.9mmHg; P= .003) and in the home SBP group by+9.0mmHg (95%CI, 5.7-12.3mmHg; P< .001). Office diastolic BP (DBP) increased by+5.0mmHg (95%CI, 2.7-7.3mmHg; P< .001), and home DBP increased by+7.8mmHg (95%CI,5.6-10.4mmHg; P< .001). AHI, baseline home SBP, use of statin drugs, sex, and the number of antihypertensive drugs prescribed were all independently associated with SBP change in multivariate analysis, controlling for age, BMI, smoking status, diabetes, and sleepiness. CPAP withdrawal results in a clinically relevant increase in BP, which is considerably higher than in conventional CPAP trials; it is also underestimated when office BP is used. Greater OSA severity is associated with a higher BP rise in response to CPAP withdrawal. ClinicalTrials.gov; No.: NCT01332175 and NCT01797653) URL: www.clinicaltrials.gov and ISRCTN registry (ISRCTN 93153804) URL: http://www.isrctn.com/.