BACKGROUND: Noninvasive positive pressure ventilation (NPPV) provides ventilatory support without the need for an invasive airway. Interest has emerged in using NPPV to facilitate earlier removal of the endotracheal tube and decrease complications associated with prolonged intubation. OBJECTIVES: To summarize the evidence comparing NPPV and invasive positive pressure ventilation (IPPV) weaning on clinical outcomes in intubated adults with respiratory failure. SEARCH STRATEGY: We searched the Cochrane Central Register of Controlled Trials (The Cochrane Library, issue 2, 2003), MEDLINE (January 1966 to July 2003) and EMBASE (January 1980 to July 2003) for randomized controlled trials comparing NPPV and IPPV weaning. Additional data sources included personal files, conference proceedings and author contact. SELECTION CRITERIA: Randomized and quasi-randomized studies comparing early extubation with immediate application of NPPV to IPPV weaning in intubated adults with respiratory failure. DATA COLLECTION AND ANALYSIS: Two reviewers independently assessed trial quality and abstracted data according to prespecified criteria. Sensitivity and subgroup analyses were planned to assess the impact of (i) excluding quasi-randomized trials and (ii) the etiology of respiratory failure on outcomes. MAIN RESULTS: We identified eleven trials, of which five were included, involving 171 participants with predominantly chronic obstructive pulmonary disease. Overall, the included studies were of moderate to good quality. Compared to the IPPV strategy, the NPPV strategy decreased mortality (RR 0.41, 95% CI 0.22 to 0.76), the incidence of ventilator associated pneumonia (RR 0.28, 95% CI 0.09 to 0.85), intensive care unit length of stay (WMD -6.88 days, 95% CI -12.60 to -1.15), hospital length of stay (WMD -7.33 days, 95%CI -14.05 to -0.61), total duration of mechanical support (WMD -7.33 days, 95% CI -11.45 to -3.22) and the duration of endotracheal mechanical ventilation (WMD -6.79 days, 95% CI -11.70 to -1.87). There was no effect of NPPV on weaning failures or the duration of mechanical support related to weaning and insufficient data to pool adverse events or quality of life. Excluding a single quasi-randomized trial maintained the significant reduction in mortality and ventilator associated pneumonia. Subgroup analyses suggested that the mortality benefit of the NPPV approach is greater in patients with chronic obstructive pulmonary disease. REVIEWER'S CONCLUSIONS: Summary estimates from five studies of moderate to good quality demonstrated a consistent positive effect on overall mortality. At present, use of NPPV to facilitate weaning in mechanically ventilated patients, with predominantly chronic obstructive lung disease, is associated with promising, although insufficient, evidence of net clinical benefit.