近位上腕骨骨折管理におけるランダム化比較試験の脆弱性と頑健性の評価

Despite multiple surgical techniques for proximal humerus fractures (PHFs), no method has demonstrated superiority, and randomized controlled trials (RCTs) on PHF treatments have yielded conflicting results. The purpose of this study is to analyze the fragility indices for treatment of PHF by calculating fragility index (FI) and continuous fragility index (CFI) for PHF trials outcomes.

A systematic review was conducted following Preferred Reporting Items for Systematic Reviews and Meta-Analyses to evaluate PHF RCTs. RCTs reporting at least one statistically significant dichotomous outcome published between 2000 and 2024 were included. The FI and CFI for continuous outcomes were calculated for each trial.

Seventeen RCTs were included in FI analysis. The FI was 1 across 21 outcomes, meaning it would take 1 patient reversal to alter significance (FI: 1, interquartile range 0-2.5). Ten RCTs were included in the CFI analysis, which had a median of 4, indicating that across these 10 RCTs, changing 4 points on patient outcomes could nullify statistical significance (CFI: 4, interquartile range 0-8.0). After adjustment, the reported P value remained an independent predictor of fragility, with an effect size of β = -166.7 (standard error = 51.6, P = .0103) for CFI.

The observed median FI and CFI suggest significant findings of PHF RCTs could be modified with a single patient switching groups or patient-reported outcomes by changing just 4 points. Surgeons need to be skeptical of the results and conclusions drawn from high-level studies and consider the fragility and significant loss of follow-up impacting conclusions.