Fragility Index in Randomized Controlled Trials on Noninvasive Ventilation as a Weaning Strategy in Subjects With Acute Hypoxemic Respiratory Failure

To the Editor:

Optimal ventilation and weaning strategies in patients with acute hypoxemic respiratory failure are far to be assessed.¹ We applaud the systematic review by Shan et al² aiming to evaluate the efficacy of noninvasive ventilation (NIV) weaning on hospital and ICU mortalities. In a review of 6 randomized controlled trials (RCTs) with moderate-to-high risk of bias, the authors stated that there was no effect of NIV weaning on hospital and ICU mortality even if it reduced the length of ICU stay and adverse events compared with invasive weaning in acute hypoxemic respiratory failure.²

The fragility index (FI), an intuitive measure of the robustness of RCTs, was recently introduced in critical care medicine and has been used in several different systematic reviews.^3-5 The FI is achieved by using a 2-by-2 contingency table and P values produced with the Fisher exact test.³ We calculated the FI of RCTs included in the systematic review by Shan et al² and that all of the included studies had a FI of zero (FI = 0 and P > .05). This FI score means that the RCTs evaluating the use of NIV weaning on mortality are very fragile and the evidence from these studies is very weak. The FI may be an easy additional index to aid the interpretation of studies and may assist clinicians in appropriate and optimal decision-making on critically ill patients.⁶ Our findings support the author’s conclusion that stronger evidence is needed to definitively assess whether NIV weaning may reduce hospital and ICU mortality rates. We further suggest that Shan et al² include the FI of zero for the included RCTs as a fourth limitation of their systematic review.

• Copyright © 2021 by Daedalus Enterprises