Assessment of intensive care unit-free and ventilator-free days as alternative outcomes in the pragmatic airway resuscitation trial

Objective: We sought to evaluate the utility and validity of ICU-free days and ventilator-free days as candidate outcomes for OHCA trials.

Methods: We conducted a secondary analysis of the Pragmatic Airway Resuscitation Trial. We determined ICU-free (days alive and out of ICU during the first 30 days) and ventilator-free days (days alive and without mechanical ventilation). We determined ICU-free and ventilator-free day distributions and correlations with Modified Rankin Scale (MRS). We tested associations with trial interventions (laryngeal tube (LT), endotracheal intubation (ETI)) using continuous (t-test), non-parametric (Wilcoxon Rank-Sum test - WRS), count (negative binomial - NB) and survival models (Cox proportional hazards (CPH) and competing risks regression (CRR)). We conducted bootstrapped simulations to estimate statistical power.

Main results: ICU-free days was skewed; median 0 days (IQR 0, 0), survivors only 24 (18, 27). Ventilator-free days was skewed; median 0 (IQR 0, 0) days, survivors only 27 (IQR 23, 28). ICU-free and ventilator-free days correlated with MRS (Spearman's ρ = -0.95 and -0.97). LT was associated with higher ICU-free days using t-test (p = 0.001), WRS (p = 0.003), CPH (p = 0.02) and CRR (p = 0.04), but not NB (p = 0.13). LT was associated with higher ventilator-free days using t-test (p = 0.001), WRS (p = 0.001) and CRR (p = 0.03), but not NB (p = 0.13) or CPH (p = 0.13). Simulations suggested that t-test and WRS would have had the greatest power to detect the observed ICU- and ventilator-free days differences.

Conclusion: ICU-free and ventilator-free days correlated with MRS and differentiated trial interventions. ICU-free and ventilator-free days may have utility in the design of OHCA trials.