Eleven randomised controlled trials (RCTs) and one quasi-randomised controlled trial (n=530) were included in the review. Sample sizes ranged from 21 to 90 patients. Included studies were of moderate to high quality.
Non-invasive weaning was more effective than invasive weaning in reducing mortality (RR 0.55, 95% CI: 0.38, 0.79, p=0.001, 12 trials) and ventilator associated pneumonia (RR 0.29, 95% CI: 0.19, 0.45, p<0.001, 11 trials). There was no evidence of statistical heterogeneity.
Non-invasive weaning reduced the length of stay in intensive care (WMD -6.27 days, 95% CI: -8.77, -3.78) and hospital stay (WMD: -7.19 days, 95% CI: -10.8, -3.58). Non-invasive weaning also reduced total duration of mechanical ventilation by 5.6 days and duration of invasive ventilation by 7.8 days but there was evidence of statistical heterogeneity. There were no significant differences in benefit between invasive and non-invasive weaning on weaning failure or weaning time. Adverse events were reported in the review.
Sensitivity analysis and post-hoc analysis did not significantly alter the results. Subgroup analysis showed a significant benefit of non-invasive weaning on mortality in studies enrolling 50% or more patients with chronic obstructive pulmonary disease (p=0.02). No statistically significant benefits were reported for other subgroup analyses.
There was evidence of publication bias on inspection of the funnel plot, which suggests that studies with non-significant results may have been missed.