Acute exacerbations of chronic obstructive airways disease (COAD) are a common cause of admission to hospital, and have a high mortality. Nasal intermittent positive pressure ventilation (NIPPV) has been used successfully in patients with respiratory failure due to neuromuscular and skeletal disorders, but the outcome of treatment in patients with COAD is less well known. We carried out a prospective randomised controlled trial of conventional treatment versus conventional treatment plus NIPPV, in 60 patients with acute ventilatory failure due to exacerbations of COAD. For the NIPPV group there was a rise in pH, compared with a fall in the controls (mean difference of change between the groups 0.046 [95% CI 0.06-0.02, p < 0.001]), and a larger fall in PaCO2 (mean difference in change between the groups 1.2 kPa [95% CI 0.45 to 2.03, p < 0.01]). Median visual analogue scores over the first 3 days of admission showed less breathlessness in the NIPPV group (2.3 cm [range 0.1-5.5]) than in the control group (4.5 cm [range 0.9-8.8]) (p < 0.025). Survival rates at 30 days were compared for intention-to-treat and efficacy populations. In the efficacy mortality comparison, mortality in the NIPPV group was reduced: 1/26 vs 9/30 (relative risk = 0.13, CI = 0.02-0.95, p = 0.014). This effect was less in the intention-to-treat analysis: 3/30 vs 9/30 (relative risk = 0.33, CI = 0.10-1.11, p = 0.106). In patients with acute ventilatory failure due to COAD who received NIPPV there was a significant rise in pH, a reduction in PaCO2 and breathlessness, and reduced mortality.