Comparative efficacy and safety of treatments for localised prostate cancer: an application of network meta-analysis.

Context: There is ongoing uncertainty about the optimal management of patients with localised prostate cancer. Objective: To evaluate the comparative efficacy and safety of different treatments for patients with localised prostate cancer. Design: Systematic review with Bayesian network meta-analysis to estimate comparative ORs, and a score (0-100%) that, for a given outcome, reflects average rank order of superiority of each treatment compared against all others, using the Surface Under the Cumulative RAnking curve (SUCRA) statistic. Data sources: Electronic searches of MEDLINE without language restriction. Study selection: Randomised trials comparing the efficacy and safety of different primary treatments (48 papers from 21 randomised trials included 7350 men). Data extraction: 2 reviewers independently extracted data and assessed risk of bias. Results: Comparative efficacy and safety evidence was available for prostatectomy, external beam radiotherapy (different types and regimens), observational management and cryotherapy, but not high-intensity focused ultrasound. There was no evidence of superiority for any of the compared treatments in respect of all-cause mortality after 5 years. Cryotherapy was associated with less gastrointestinal and genitourinary toxicity than radiotherapy (SUCRA: 99% and 77% for gastrointestinal and genitourinary toxicity, respectively). Conclusions: The limited available evidence suggests that different treatments may be optimal for different efficacy and safety outcomes. These findings highlight the importance of informed patient choice and shared decision-making about treatment modality and acceptable trade-offs between different outcomes. More trial evidence is required to reduce uncertainty. Network meta-analysis may be useful to optimise the power of evidence synthesis studies once data from new randomised controlled studies in this field are published in the future.