Neuromuscular Blocking Agents for ARDS: Firm Evidence for ICU Mortality but Not for Long-Term Mortality

To the Editor:

Optimal ventilation and weaning strategies in patients with ARDS should be carefully assessed.^1,2 We read with interest the systematic review by Torbic et al,³ which sought to evaluate mortality when using neuromuscular blocking agents (NMBAs) in early, moderate-to severe ARDS compared to usual care or placebo. The authors included 6 randomized controlled trials (RCTs), but only 2 of the 6 RCTs had a low risk of bias.³ We applaud the effort of the authors, but we have some concerns.

First, the fragility index, which is an intuitive measure of the robustness of RCTs, was introduced in critical care medicine and has been utilized in different systematic reviews.^4–6 The fragility index is achieved by using a two-by-two contingency table and P value calculated with the Fisher exact test.⁴ When we calculated the fragility index of RCTs included in the systematic review by Torbic et al,³ we found that all of the included studies had a fragility index of zero (P > .05). According to this, the RCTs evaluating mortality when using NMBAs in early and moderate-to severe ARDS are very fragile and the evidence from these studies is very weak.

Second, the authors reported that the use of NMBAs reduced the risk for ICU and 21–28-d mortality but not 90-d mortality. According to these results, we further performed a trial sequential analysis to evaluate whether this meta-analysis had sufficient statistical power to detect or reject the intervention effects.⁷ For ICU mortality, the 95% CI adjusted with the trial sequential analysis ranged from 0.41 to 0.8 and indicated that 431 of 433 subjects were enough to reach the required information size. According to this, firm evidence existed in favor of the use of the NMBAs to reduce ICU mortality. For 21–28-d mortality, the trial sequential analysis was not conclusive because the inclusion of 1,485 subjects was far short of the required sample size of 18,618 subjects to ensure conclusive evidence. By evaluating the robustness of RCTs and trial sequential analysis, our analysis supports the reported effects of NMBAs to reduce ICU mortality but not the 21–28-d and 90-d mortality. We believe we would need 18,470 subjects with moderate-to-severe ARDS to be treated with NMBAs to achieve firm evidence on long-term mortality.

• Copyright © 2021 by Daedalus Enterprises