The fragility index applied to liver-related trials.

The goal of this manuscript was to apply the fragility index (FI), which is a statistically sound method to evaluate robustness of test results, to liver-related randomized clinical trials. The authors searched the ClinicalTrials.gov database with the following limitations: term "liver," recruitment completed, with results, interventional study type, last updated May 01, 2016, to May 01, 2017. Forty-eight trials were included and four had FI of 0. The median FI for trials moving from significance to non-significance was 6 (IQR 18; 2 to 20), while the median for trials moving from non-significance to significance was 5 (IQR 5; 4 to 9). The median number lost to follow up was 17 (IQR 42; 3 to 45). Of the 21 trials that showed statistical significance, the number lost to follow up was greater than the FI in 13 (61.90%) trials. Investigators of liver-related studies should consider adding the FI to evaluate their work.