Trial Duration and Risk Reduction in Combination Therapy Trials for Pulmonary Arterial Hypertension: A Systematic Review

Background Relative risk (RR) and number needed-to-treat (NNT) are frequently time-dependant measures. We performed a systematic review and meta-analysis to assess whether trial duration influenced the relative and absolute risk of worsening in randomized controlled trials (RCTs) comparing combination therapy (CT) of pulmonary arterial hypertension (PAH)-specific therapies vs monotherapy (MT). Methods We searched MEDLINE, Embase, and the Cochrane Library (January 1990 to September 2016) for RCTs assessing CT compared with MT in PAH. The primary outcome was the risk of clinical worsening. We assessed whether trial duration correlated with RR and NNT using weighted meta-regression with mixed effects. Changes in NNT overtime were also assessed using data from long-term event-driven trials. Results There were 3,801 patients throughout 15 studies included. The RR for clinical worsening positively correlated with trial duration ( R 2  = 0.67, P  = .0002), whereas the NNT did not (mean NNT, 7; R 2  = 0.02; P  = .65). Among long-term event-driven trials, the mean NNT progressively decreased until 52 weeks of follow-up, being stable thereafter. Conversely, the mean RR progressively increased from approximately 0.40 at week 16 to approximately 0.68 at week 104. Conclusions Absolute risk reduction of clinical worsening was relatively constant beyond 6 to 12 months of treatment in clinical trials comparing CT with MT in PAH. These results question the need for CT trials of very long duration in PAH.