Obesity hypoventilation syndrome (OHS) with concomitant severe obstructive sleep apnea (OSA) is treated with CPAP or noninvasive ventilation (NIV) during sleep. NIV is costlier, but may be advantageous because it provides ventilatory support. However, there are no long-term trials comparing these treatment modalities based on OHS severity.To determine if CPAP have similar effectiveness when compared to NIV according to OHS severity subgroups.Post hoc analysis of the Pickwick randomized clinical trial in which 215 ambulatory patients with untreated OHS and concomitant severe OSA, defined as apnoea-hypopnea index (AHI)≥30events/h, were allocated to NIV or CPAP. In the present analysis, the Pickwick cohort was divided in severity subgroups based on the degree of baseline daytime hypercapnia (PaCO2 of 45-49.9 or ≥50mmHg). Repeated measures of PaCO2 and PaO2 during the subsequent 3 years were compared between CPAP and NIV in the two severity subgroups. Statistical analysis was performed using linear mixed-effects model.204 patients, 97 in the NIV group and 107 in the CPAP group were analyzed. The longitudinal improvements of PaCO2 and PaO2 were similar between CPAP and NIV based on the PaCO2 severity subgroups.In ambulatory patients with OHS and concomitant severe OSA who were treated with NIV or CPAP, long-term NIV therapy was similar to CPAP in improving awake hypercapnia, regardless of the severity of baseline hypercapnia. Therefore, in this patient population, the decision to prescribe CPAP or NIV cannot be solely based on the presenting level of PaCO2.