To define the parameters of respiratory insufficiency in OSA, 114 consecutive patients (108 men, six women) were prospectively studied. In addition to standard polysomnography, they underwent pulmonary function tests, right heart catheterization, and ventilatory response tests to hypercapnia. Nineteen patients (19 percent) had a resting PAP greater than or equal to 20 mm Hg. Multiple regression analysis showed that FEV1 and PaO2 (both with a negative coefficient) and PaCO2 (with a positive coefficient) significantly contributed to PAP. Thirteen patients (12 percent) had a PaCO2 greater than or equal to 45 mm Hg. A multiple regression analysis showed that FEV1 and the minute ventilation at PETCO2 = 60 mm Hg (both with a negative coefficient) and the cumulative apnea duration (with a positive coefficient) significantly contributed to PaCO2. Thirty-seven patients (33 percent) had a PaO2 less than or equal to 65 mm Hg. A multiple regression analysis showed that FEV1 (with a positive coefficient) and the hypopnea + apnea index (with a negative coefficient) significantly contributed to PaO2. These data confirm that impaired daytime pulmonary function (diffuse airway obstruction) contributes to the development of daytime pulmonary hypertension, hypoxemia, and hypercapnia in OSA patients. They show that the amount of sleep-related breathing disorders also plays a significant role.