We have investigated pulmonary hemodynamics in a large series of consecutive, unselected patients with obstructive sleep apnea syndrome (OSAS). The aims of this study were to evaluate the frequency of pulmonary artery hypertension (PH) in OSAS and to analyze, as far as possible, its mechanisms. Two hundred twenty patients were included on the basis of a polysomnographic diagnosis of OSAS (apnea+hypopnea index > 20). PH, defined by a resting mean pulmonary artery mean pressure (PAP) of at least 20 mm Hg, was observed in 37 of 220 patients (17%). Patients with PH differed from the others with regard to pulmonary volumes (vital capacity [VC], FEV1) and the FEV1/VC ratio that were significantly lower (p < 0.001); PaO2 (64.4 +/- 9.3 vs 74.7 +/- 10.1 mm Hg; p < 0.001); PaCO2 (43.8 +/- 5.4 vs 37.6 +/- 3.9 mm Hg; p < 0.001), apnea+hypopnea index (100 +/- 33 vs 74 +/- 32; p < 0.001), and mean nocturnal arterial oxygen saturation (SaO2) (88 +/- 6% vs 94 +/- 2%; p < 0.001). Patients with PH were also more overweight (p < 0.001). Multiple regression analysis showed that 50% of the variance of PAP could be predicted by an equation including PaCO2 (accounting for 32% of the variance), FEV1 (12%), airway resistance (4%), and mean nocturnal SaO2 (2%). In conclusion, PH is observed, in agreement with previous studies, in less than 20% of OSAS patients. PH is strongly linked to the presence of an obstructive (rather than restrictive) ventilatory pattern, hypoxemia, and hypercapnia, and is generally accounted for by an associated obstructive airways disease. In this regard, the severity of OSAS plays only a minor role.