8.

Trial sequential analysis on cardiac mortality in 12 lidocaine vs placebo or no intervention trials

Trial sequential analysis of lidocaine vs placebo or no intervention on cardiac mortality in participants with or without proven myocardial infarction based on the diversity‐adjusted required information size (DARIS) of 22,745 participants. This DARIS was calculated on the basis of a proportion of participants with cardiac mortality among those with suspected myocardial infarction of 1.51% in the control group; post hoc selected RRR of 30% in the experimental intervention group; alpha (α) of 5%; beta (β) of 20%; and diversity of 14%. The cumulative Z‐curve (blue line) crossed the conventional alpha of 5% (green line) after 3 trials suggested harm. After 10 trials, however, the cumulative Z‐curve (blue line) crossed the trial sequential monitoring boundary for futility. Accordingly, after only 36.4% (8277/22,745) of the DARIS had been obtained, we were able to reject an intervention effect of 30% or larger. Had we calculated the DARIS on the basis of a more realistic RRR like 20% (as originally planned) or less, the obtained evidence would represent a smaller part of the DARIS. Accordingly, boundaries for futility would not have been crossed in such scenarios. Therefore, risk reductions of 20% or less may require additional trials with larger sample sizes.