The Use of Apneic Oxygenation During Prolonged Intubation in Pediatric Patients: a Randomized Clinical Trial

Secondary outcomes: Secondary-1. We will conduct tests for superiority to compare the Truview and DL with oxygen cannula methods on each of the two primary outcomes, and report the treatment effect estimates and adjusted confidence intervals. Notably, the study is not powered to be able to assess equivalence between these two oxygen techniques. Secondary-2. We will compare the three randomized groups on total desaturation slope (rate) using a random slope linear mixed effects model with repeated measures. This model will allow accounting for the variability of the saturation rate for the patients and possible correlation of the saturation measurements within a patient (autoregressive correlation). We only will use saturation data before it reaches 90%. Secondary-3. We will assess the relative efficacy of Truview and DL with oxygen cannula in preventing 90% saturation compared with DL alone using 2-tailed chi-square tests of proportions (for superiority), summarized with a relative risk and confidence interval. Secondary -4: We will assess the correlation between the rate of desaturation and both age and body mass index, independent of randomized group. This will be done in the context of a random slope model as in the primary aim, assessing the interaction between the mean desaturation slope and each of age and body mass index. To restrict overall Type I error of the secondary hypotheses at 5% we will apply Bonferroni correction for multiple inferences and adjustment for interim analyses. Interim analyses will be conducted every 25% of the maximum planned enrollment using a groups sequential design to test for efficacy and futility. We will use a gamma spending function (gamma = -4 for efficacy and -2 for futility) to maintain the significance level for the primary outcome at 2.5% and the power at 90% across the interim analyses. Significance criteria for all tests will be adjusted for interim analyses based on the above-mentioned gamma spending function and the corresponding z-statistic at teach interim look.