Study selection
Initially, we identified 318 articles. After deduplication (n=93) based on title and abstracts, we excluded 193 irrelevant papers and two studies without full articles available.
Full texts of 30 articles were selected for detailed evaluation, of which 17 randomized trials were eligible for analysis. Three studies were excluded because we could not obtain sufficient information for comparison, and two had an overlap population. Thirteen papers that met the inclusion criteria were left 2-4, 16-25. The flow diagram illustrating the information extracted from a review of potentially relevant articles can be observed in Figure 1.