A Bayesian network meta-analysis of three different surgical procedures for the treatment of humeral shaft fractures

BACKGROUND The optimal surgical procedure for humeral shaft fractures remains a matter of debate. We aimed to establish the optimum procedure by performing a Bayesian network meta-analysis. METHODS PubMed, EMBASE, the Cochrane Library, and Medline were searched for both randomized controlled trials and prospective studies of surgical treatment for humeral shaft fractures. The quality of the included studies was assessed according to the Cochrane Collaboration's "Risk of bias". RESULTS Seventeen RCTs or prospective studies were included in the meta-analysis. The pooled results showed that the occurrence rate of radial nerve injury was lowest for minimally invasive plate osteosynthesis (MIPO; SUCRA probability, 95.1%), followed by open reduction and plate osteosynthesis (ORPO; SUCRA probability, 29.5%), and was highest for intramedullary nailing (IMN; SUCRA probability, 25.4%). The aggregated results of pairwise meta-analysis showed no significant difference in radial nerve injury rate when comparing ORPO versus IMN (OR, 1.92; 95% CI, 0.96 to 3.86), ORPO versus MIPO (OR, 3.38; 95% CI, 0.80 to 14.31), or IMN versus MIPO (OR, 3.19; 95% CI, 0.48 to 21.28). Regarding the nonunion, SUCRA probabilities were 90.5%, 40.2%, and 19.3% for MIPO, ORPO, and IMN, respectively. The aggregated results of a pairwise meta-analysis also showed no significant difference for ORPO versus IMN (OR, 0.83; 95% CI, 0.41 to 1.69), ORPO versus MIPO (OR, 2.42; 95% CI, 0.45 to 12.95), or IMN versus MIPO (OR, 2.49; 95% CI, 0.35 to 17.64). CONCLUSION The current evidence indicates that MIPO is the optimum choice in the treatment of humeral shaft fractures and that ORPO is superior to IMN.