To compare the efficacy of standard medical therapy (ST) and noninvasive mechanical ventilation additional to standard medical therapy in hypercapnic acute respiratory failure (HARF). Single center, prospective, randomized, controlled study. Pulmonary medicine directed critical care unit in a university hospital. Between March 1993 and November 1996, 30 HARF patients were randomized to receive ST or noninvasive positive pressure ventilation (NPPV) in addition to ST. NPPV was given with an air-cushioned face via a mechanical ventilator (Puritan Bennett 7200) with initial setting of 5 cm H2O continuous positive airway pressure and 15 cm H2O pressure support. At the time of randomization, patients in the ST group had (mean+/-SD) PaO2 of 54+/-13 mm Hg, PaCO2 of 67+/-11 mm Hg, pH of 7.28+/-0.02, and respiratory rate of 35.0+/-5.8 breaths/min. Patients in the NPPV group had PaO2 of 55+/-14, PaCO2 of 69+/-15, pH of 7.27+/-0.07, and respiratory rate of 34.0+/-8.1 breaths/min. With ST, there was significant improvement of only respiratory rate (p < 0.05). However, with NPPV, PaO2 (p < 0.001), PaCO2 (p < 0.001), pH (p < 0.001), and respiratory rate (p < 0.001) improved significantly compared with baseline. Six hours after randomization, pH (p < 0.01) and respiratory rate (p < 0.01) in NPPV patients were significantly better than with ST. Hospital stay for NPPV vs ST patients was, respectively, 11.7+/-3.5 and 14.6+/-4.7 days (p < 0.05). One patient in the NPPV group required invasive mechanical ventilation. The conditions of six patients in the ST group deteriorated and they were switched to NPPV; this was successful in four patients, two failures were invasively ventilated. This study suggests that early application of NPPV in HARF patients facilitates improvement, decreases need for invasive mechanical ventilation, and decreases the duration of hospitalization.