Mortality and drug therapy in patients with chronic obstructive pulmonary disease: a network meta-analysis

David A Scott, Bethan Woods, Juliette C Thompson, James F Clark, Neil Hawkins, Mike Chambers, Bartolome R. Celli, Peter Calverley

Increasing evidence suggests pharmacological treatments may impact on overall survival in Chronic Obstructive Pulmonary Disease (COPD) patients. Individual clinical trials are rarely powered to detect mortality differences between treatments and may not include all treatment options relevant to healthcare decision makers. A systematic review was conducted to identify RCTs of COPD treatments reporting mortality; evidence was synthesised using network meta-analysis (NMA). The analysis included 40 RCTs; a quantitative indirect comparison between 14 treatments using data from 55,220 patients was conducted. The analysis reported two treatments reducing all-cause mortality; salmeterol/fluticasone propionate combination (SFC) was associated with a reduction in mortality versus placebo in the fixed effects (HR 0.79; 95 % Crl 0.67, 0.94) but not the random effects model (0.79; 0.56, 1.09). Indacaterol was associated with a reduction in mortality versus placebo in fixed (0.28; 0.08 to 0.85) and random effects (0.29; 0.08, 0.89) models. Mean estimates and credible intervals for hazard ratios for indacaterol versus placebo are based on a small number of events; estimates may change when the results of future studies are included. These results were maintained across a variety of assumptions and provide evidence that SFC and indacaterol may lead to improved survival in COPD patients. Results of an NMA of COPD treatments suggest that SFC and indacaterol may reduce mortality. Further research is warranted to strengthen this conclusion.