Document resource
We thank Lee and Gau for their methodological comments regarding our observational study, which prompted a careful discussion.1 2 The authors raise three main points: variable dichotomization, residual confounding due to frailty, and the possible mediating effect of intraoperative fentanyl.