Table 4

Exercise tolerance results

StudyNo of patientsOutcome measure1Mean on opioidMean on placeboPooled SDMean differenceSE2 crude/correctedSMD (95% CI)
1Change from baseline or post-treatment measurement.
2 “Crude” standard errors ignore crossover designs (or assume a correlation of 1); “corrected” standard errors are from paired analyses.
*Standard errors imputed using correlations observed in other studies.
†Standard errors based on assumptions of zero correlation coefficient.
Oral and parenteral opioid studies
Chua1012Post51245593.55738.17/12.85–0.61 (–0.88 to –0.34)
Eiser11 (a)10Post223.5216.0118.17.552.83/22.45*–0.06 (–0.44 to 0.31)
Eiser11 (b) 8Post272263141.4970.71/14.48–0.06 (–0.26 to 0.14)
Johnson1218Post249213133.13644.38/15.39–0.27 (–0.49 to –0.05)
Light 13 7Change0.200.070.650.130.35/0.35†–0.20 (–1.25 to 0.85)
Poole1414Change–35.121.666.6–56.725.15/24.850.85 (0.12 to 1.58)
Woodcock1512Change53439.24916.02/16.02†–1.25 (–2.05 to –0.45)
Woodcock1611Post387368103.21944.01/18.70*–0.18 (–0.54 to 0.17)
Nebulised opioid studies
Beauford17 8Change0–0.10.580.10.29/0.29†–0.17 (–1.15 to 0.81)
Harris-Eze20 5Post42543736.9–1221.33/9.06*0.32 (–0.16 to 0.81)
Leung2210Change2711–54.92/5.220.45 (–0.48 to 1.39)
Young2511Change64690.15838.44/20.70–0.64 (–0.89 to –0.39)
Meta-analysisNo of studiesTest for heterogeneityTest for overall effectSMD (95% CI)
All studies12Q=42.5(p<0.0001)p=0.09–0.20 (–0.42 to 0.03)
Non-nebulised studies 8Q=25.96 (p=0.001)p=0.08–0.22 (–0.47 to 0.03)
Nebulised studies 4Q=15.0 (p=0.002)p=0.86–0.06 (–0.70 to 0.58)
COPD studies10Q=30.6 (p<0.0001)p=0.12–0.19 (–0.43 to 0.05)