Comparison of cumulative incidence analysis and Kaplan-Meier for analysis of shock reversal in patients with septic shock

Abstract Introduction Kaplan-Meier (KM) has become the most used method to evaluate time-to-event analysis, although it is unsuitable in competing event situations such as death and shock reversal. Despite that the use of this methodology is not widely disseminated, cumulative incidence analysis (CIA) is more appropriate in these situations. We used CIA and KM (with 2 different techniques of censoring) to compare shock reversal in a cohort of patients with septic shock after steroid therapy. Furthermore, we have analyzed shock reversal in responders and nonresponders to high-dose cortrosyn test (250 μ g). Methods Analysis of shock reversal in a cohort of 74 patients with septic shock at a university hospital was done. Results Shock reversal by the 28th day was estimated to be 88% and 72% by KM methods and 59% by CIA. In nonresponders to cortrosyn test ( Δ ≤9 μ g/dL), shock reversal was estimated in 80% and 56% according to KM and 47% according to CIA. As for responders to cortrosyn test, shock reversal was estimated in 90% and 77% according to KM and 64% according to the CIA method. Conclusion Kaplan-Meier overestimates shock reversal. Cumulative incidence analysis seems to be a more appropriate method to analyze shock reversal. Future trials intended to analyze shock reversal should apply CIA.