Combined analysis for arterial carbon dioxide tension in chronic obstructive pulmonary disease nonrandomised controlled trials
| First author [Ref.] | Favours bilevel n | Favours control n | MD±se | Weight % | MD (random) 95% CI |
| Ambrosino 25 | 7 | 7 | −7.0000±3.2071 | 10.46 | −7.00 (−13.29– −0.71) |
| Krachman 20 | 6 | 6 | −3.2700±3.9438 | 7.67 | −3.27 (−11.00–4.46) |
| Lien 26 | 11 | 11 | −0.8000±4.9179 | 5.33 | −0.80 (−10.44–8.84) |
| Lin 27 | 10 | 12 | −0.9000±1.4299 | 24.29 | −0.90 (−3.70–1.90) |
| Maragoni 28 | 14 | 14 | −6.0000±2.8885 | 12.08 | −6.00 (−11.66– −0.34) |
| Meecham Jones 29 | 14 | 14 | −4.5000±1.9744 | 18.72 | −4.50 (−8.37– −0.63) |
| Nava 17 | 6 | 7 | −7.7000±2.7889 | 12.65 | −7.70 (−13.17– −2.23) |
| Strumpf 30 | 7 | 7 | 3.0000±3.6061 | 8.80 | 3.00 (−4.07–10.07) |
| Total | 75 | 78 | 100 | −3.52 (−5.93– −1.11) |
MD: mean difference; CI: confidence interval. Comparison: crossover trials of bilevel noninvasive positive pressure ventilation versus all modalities. Outcome: arterial carbon dioxide tension cmH2O/end tidal carbon dioxide tension cmH2O. Test for heterogeneity: Chi squared = 11.16, df = 7 (p = 0.13), I2 = 37.3%. Test for overall effect: Z = 2.86 (p = 0.004).