Fig. 1

Intensive Care Unit Survival in patients with and without occurrence of fever or hypothermia after ECMO decannulation. Cumulative incidence function of ICU discharge and death (competing event) in patients with and without fever (A) and with- and without hypothermia (C) following ECMO decannulation. Cumulative incidence of ICU discharge and death as multistate comparison showing no increased incidence of death in patients with fever (B), but increased incidence of death in patients with hypothermia (D) after ECMO explantation, respectively. The adjusted multivariable competing risk regression model shows hypothermia, increased lactate and need for renal replacement therapy following decannulation as predictive for ICU-mortality (E). Survival is recorded beginning from the time point of ECMO explantation. ICU survival following ECMO decannulation was analyzed for the whole cohort and for the subgroups fever vs. no fever and hypothermia vs. no hypothermia within 72 h following ECMO removal, respectively. Influence of post-ECMO-fever and hypothermia on death on ICU was analyzed by means of a competing risk regression model using death as the primary event and discharge from ICU as a competing event. Post-ECMO-fever or hypothermia was used as a univariable independent risk factor. All reported p-values are two-sided unless indicated otherwise; p values < 0.05 were considered statistically significant. We used GraphPad Prism (Version 10.0, GraphPad Software, La Jolla, CA) and IBM SPSS Statistics (Version 25.0, IBM Corp., Armonk, NY) and for data analysis and graph generation. Competing risk regression was performed with the cmprsk R package within the R environment for statistical computing version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria). SHR subdistribution hazard ratio, CI confidence interval