Non-surgical treatments for post-burn scars: A network meta-analysis

Background and aim: Post-burn scarring is a prevalent condition, and the existing non-surgical treatments exhibit varying degrees of efficacy. There is limited evidence available to determine the effective non-surgical treatment for post-burn scars. This study employs a multi-index network meta-analysis to conduct a comprehensive evaluation and comparative ranking of non-surgical treatments for post-burn scars. The aim is to identify the most effective treatment methods, thereby providing a robust, evidence-based foundation to guide clinical decision-making. Methods: PubMed, Web of Science, Cochrane Library, PEDro, and Embase were systematically searched for eligible randomized controlled trial studies, and the network meta-analysis was performed via a frequentist approach. The primary outcomes assessed were Vancouver Scar Scale score, scar thickness and Visual Analogue Scale score. Results: A total of 17 studies and 1,013 participants were included in this analysis. The treatment ranking revealed that massage therapy demonstrated the most significant efficacy in reducing Vancouver Scar Scale score (surface under the cumulative ranking curve [SUCRA] = 89.0%), CO2 laser therapy exhibited the highest efficacy in decreasing scar thickness (SUCRA = 96.8%), and extracorporeal shock wave therapy + routine treatment showed the most significant efficacy in reducing Visual Analogue Scale score (SUCRA = 58.6%). Conclusion: This network meta-analysis illustrates that massage therapy, CO2 laser therapy and extracorporeal shock wave therapy + routine treatment are the most effective non-surgical treatments for reducing Vancouver Scar Scale score, scar thickness and Visual Analogue Scale score for post-burn scars, respectively. However, the findings reflect outcomes at a specific stage of scar maturation. And our conclusions must be interpreted with caution due to the limited number of studies included. In the future, well-designed randomized controlled trials with a large sample size are needed to validate these findings.