Fig. 10

Trial sequential analysis plot of postoperative ICU length of stay (sensitivity analysis). Trial sequential analysis of 7 RCTs that did not use propofol anesthesia on postoperative ICU length of stay. The cumulative z curve crossed the conventional boundary but not the trial sequential monitoring boundary for benefit, indicating the current evidence is inconclusive. A required information size of 1,417 patients was calculated using α = 0.05 (two sided), β = 0.20 (power 80%) and the mean difference generated in the conventional meta-analysis