Reply to Cortegiani and Giarratano, “Untargeted Antifungal Treatment in Nonneutropenic Critically Ill Patients: Should Further Studies Be Performed Based on Trial Sequential Analysis Results?”

REPLY

We thank Cortegiani and Giarratano (1) very much for the attention. The issue that they raised is critical for the use of trial sequential analysis. The minimum selected desired effect size indeed is an important parameter that might have substantial impact on results. Actually, we reported this parameter as 20% (we named it effect measure reduction) in Appendix 2 of our supplemental material. The value that we chose was consistent with those of several studies in different areas (2, 3). The sensitivity analyses of this parameter were not performed in our study, so that it is unknown whether the results would change with other values of the minimum selected desired effect size (4).

Furthermore, the conclusion that we drew, namely, that since Z-curves crossed the futility boundary (for all-cause mortality) and the trial sequential monitoring boundary for benefit (for the incidence of IFI), the evidence should be considered sufficient and conclusive and hence indicated that further trials would not be required, was only about untargeted treatment, which integrates three different treatment strategies (i.e., prophylaxis, empirical treatment, and preemptive treatment). However, only one trial was described as applying preemptive treatment, and five trials were defined as applying empirical treatment in our included randomized controlled trials. Therefore, once these strategies were treated separately, the above conclusion should have been reconsidered. As we mentioned in limitations of our original article, large-scale, high-quality research on this topic (covering different treatment strategies) should be performed to further reduce the bias and verify the conclusion.