BACKGROUND--There are no controlled trials of the use of different modes of nasal intermittent positive pressure ventilation (NIPPV) in patients with exacerbations of chronic obstructive pulmonary disease (COPD). This study describes the effect on blood gas tensions of four different modes of nasal ventilation. METHODS--Twelve patients with acute exacerbations of COPD were studied (mean (SD) FEV1 0.59 (0.13) l, PaO2 (air) 5.10 (1.12) kPa, PaCO2 9.28 (1.97) kPa, pH 7.32 (0.03)). Each patient underwent four one-hour periods of nasal ventilation in randomised order: (a) inspiratory pressure support 18 cm H2O; (b) pressure support 18 cm H2O+positive end expiratory pressure (PEEP) 6 cm H2O (IPAP+EPAP); (c) continuous positive airway pressure (CPAP) 8 cm H2O; and (d) volume cycled NIPPV. Arterial blood samples were obtained before each period of ventilation and at one hour. RESULTS--Pressure support, CPAP, and volume cycled NIPPV all produced significant improvements in PaO2; there was no difference between these three modes. The change in PaO2 with IPAP+EPAP did not reach statistical significance. None of the modes produced significant changes in mean PaCO2; patients with higher baseline levels tended to show a rise in PaCO2 whereas those with lower baseline levels tended to show a fall. CONCLUSIONS--Although PaO2 improved in all patients there are differences in efficacy between the modes, while the changes in PaCO2 were variable. The addition of EPAP conferred no advantage in terms of blood gas tensions.