RESTORE Trial

The RESTORE trial uses 1:1 randomisation stratified by surgical technique, age group, and baseline CPET result. No statistically significant differences between melatonin and placebo emerged in any secondary outcome, including sleep duration, length of stay, readmission, and mortality. RESTORE is a pragmatic, multicentre, superiority randomised controlled trial with an embedded observational cohort. There was no statistically significant difference in delirium incidence between the melatonin group (3.64%) and the placebo group (1.85%), with a p-value of 1.000. Subgroup analyses of the 5 mg and 8 mg melatonin arms found no clinically meaningful differences and no evidence of a dose-response relationship. The overall five-day delirium incidence in RESTORE was 2.75%, far below the 55% rate assumed in the sample size calculation. No randomised controlled trial has previously compared corrective surgery against conservative management for pectus excavatum. The RESTORE trial was stopped early for futility after 115 patients were enrolled because the observed delirium incidence was far below the anticipated rate and no significant difference between groups was detected. The primary stati…