Initial sample size calculation was based on our primary hypothesis and primary study outcome, and was informed by data collected during a multicenter Spanish trial (ARISCAT) 1) and a single-center, relatively small study reporting the effects of intraoperative higher PEEP and recruitment maneuvers on the incidence of postoperative desaturation, chest infection and bronchospasm in obese patients who underwent laparoscopic bariatric surgery 2). These calculations indicated that 356 patients were required per group, assuming a two–sided significance level of 0.05 and a power of 80%, to detect the expected difference in postoperative pulmonary complications between the higher PEEP group of 30% and the lower PEEP group of 40% (risk ratio of 0.75). Assuming a dropout rate of 5%, a total of 748 patients (n=374 per group) were initially planned to be included into the study.
Sample size was re-estimated based on recommendations of the DSMB after data on the first 618 patients revealed that the incidence of the collapsed composite outcome was considerably lower than expected. In the revised calculations, the control group incidence was assumed to be 0.20 (instead of 0.40), and sample size was calculated to have 80% power at the overall 0.05 significance level to detect a relative risk (as originally planned) of 0.75. This design requires a maximum of 1912 patients, adjusting for the interim monitoring for efficacy and futility. Assuming a dropout rate of 5%, a total of 2013 patients need to be included into the study.
Interim analyses for efficacy and futility will be conducted at 50% (N=956), 75% (N=1434) and 100% (N=1912) of the planned enrollment, as needed, using a non-binding group sequential design with gamma spending functions (detail are available in the study protocol). The decision to terminate the study early will be made by a data and safety monitoring board (DSMB). Except in the case of early study termination, the results of the interim analysis will not be disclosed to the participating study centers.