First author [Ref.] | Bilevel NIPPV | Control | Weight % | WMD (random) 95% CI | ||
n | Mean±sd | n | Mean±sd | |||
Casanova 11 | 20 | 51.10±8.80 | 24 | 52.30±6.10 | 18.77 | −1.20 (−5.76–3.36) |
Clini 23 | 23 | 54.00±5.62 | 24 | 59.00±4.87 | 22.05 | −5.00 (−8.01– −1.99) |
Diaz 19 | 18 | 48.37±3.97 | 18 | 54.37±6.00 | 21.42 | −6.00 (−9.32– −2.68) |
Garrod 24 | 17 | 43.30±6.68 | 20 | 44.20±9.07 | 17.64 | −0.90 (−5.99–4.19) |
Gay 21 | 4 | 57.50±14.40 | 6 | 50.20±4.30 | 5.50 | 7.30 (−7.23–21.83) |
Renston 22 | 9 | 52.00±9.00 | 8 | 44.00±4.24 | 14.62 | 8.00 (1.43–14.57) |
Total | 91 | 100 | 100 | −1.20 (−5.05–2.65) |
NIPPV: noninvasive positive pressure ventilation; WMD: weighted mean difference; CI confidence interval. Comparison: RCT trials of bilevel NIPPV versus all modalities (long-term oxygen therapy, sham ventilation, exercise). Outcome: arterial carbon dioxide tension cmH2O. Test for heterogeneity: Chi squared = 18.76, df = 5 (p = 0.002), I2 = 73.3%. Test for overall effect: Z = 0.61 (p = 0.54).