We thank Nelson et al.,  Goodwin et al., and Hughes et al.  for their responses to our article.

We would like to express our gratitude to Nelson et al.  for raising important concerns. The fragility index is a novel method for measuring the robustness of results in randomized controlled trials that measure dichotomous outcomes. Although we agree that the fragility index provides information about the robustness of randomized controlled trials, the results should not be interpreted solely on the basis of the fragility index. One weakness of the fragility index is that the specific fragility index cutoff value is not known and larger fragility indexes can be obtained simply by increasing the sample size.  In our study, the fragility index was calculated to be 1, indicating a significant fragility. However, the low fragility index does not necessarily mean that the evidence is insufficient. We believe that the 58% reduction in postoperative delirium observed in our study is a clinically important finding with respect to patient safety. The application of these research results to clinical practice requires clinical judgment that considers both the clinical evidence and the possibility of errors. The fragility index can be used as a guide to confirm the robustness of the research evidence, and readers should take the low fragility index into account when interpreting the present results.

We thank Goodwin et al.  for their insightful commentary and thorough review of our study. We agree with the authors’ concerns regarding the potential impact of opioids and benzodiazepines on the occurrence of postoperative delirium.  To minimize this risk, participants in our study did not receive benzodiazepines for premedication. However, the use of opioids and pain scores were not recorded in the present study. Unfortunately, due to the heterogeneity of the types of orthopaedic surgery, perioperative pain control protocol (drugs and regimens), and postoperative hospital stay plan, it was impossible to control all of these variables effectively. We acknowledge that the strong association between opioids, pain intensity, and delirium should not be overlooked in the interpretation of our results.  This drawback of the present study was noted in the original article. 

We appreciate Hughes et al.  for their interest and thoughtful comments on our recent study. The main objective of our work was to investigate the effect of sedatives on the occurrence of postoperative delirium, particularly in healthy older adults, rather than frail individuals. As the authors comment, there is a strong association between the risk of delirium and frailty in older adults.  In addition to the change in pharmacokinetics and pharmacodynamics in the geriatric population, frail older adults have more severe comorbidities compared to nonfrail older adults. As a result, great caution is required when administering sedatives to frail older adults. Based on our findings, the results of our study should not be generalized to all geriatric populations, only to healthy older adults.