Khaled L, Godet T, Jaber S, et al. Intraoperative protective mechanical ventilation in patients requiring emergency abdominal surgery: the multicentre prospective randomised IMPROVE-2 study protocol. BMJ Open 2022;12:e054823. doi:10.1136/bmjopen-2021-054823

The definition of the primary outcome measure, as reported in the manuscript, contains an error.

The outcome was mistakenly reported as a composite of postoperative respiratory failure (defined as failure to wean from the ventilator after surgery or requiring unplanned reintubation or curative non-invasive ventilation once extubated postoperatively) and all-cause mortality by day 30 or hospital discharge.

The trial protocol (original version 2.0, dated 12/04/2019, and final version, dated 20/05/2022) indicated that the primary endpoint will be considered present if at least one of the following occurs:

a patient fails weaning from the ventilator within the first 12 hours following the end of surgery,

a patient requires reintubation or curative non-invasive ventilation between the end of surgery and day 30 after surgery,

a patient dies prior to the earlier of day 30 or hospital discharge.

This was also recorded as such at ClinicalTrials.gov study registry (ClinicalTrials.gov identifier: NCT03987789).

Of note, the authors had calculated that 2×340 patients would be needed to have 90% power to show an absolute between-group difference of 10% in the primary outcome measure, under the assumption of a 10% mortality rate and 15% rate of postoperative respiratory failure at 30 days (thus 25% for the composite outcome measure).

Consequently, from a statistical perspective, the primary outcome measure should thus have been reported as the composite of postoperative respiratory failure or all-cause mortality by day 30, and not as postoperative respiratory failure and all-cause mortality by day 30, as currently formulated. Otherwise, the authors’ assumption lapses on the understanding that the additional incidence of the two components of the composite outcome is obviously much less than 25%. The authors concur that this is a fine distinction, but that carries a world of meaning in the way of analysing and reporting the results of the trial.