Abstract
Background/Objectives: Extracorporeal cardiopulmonary resuscitation (ECPR) is increasingly used for refractory cardiac arrest, yet overall survival and neurologic recovery remain poor. We examined whether creatine kinase (CK) levels measured at predefined times after ECPR were associated with in-hospital mortality and evaluated their discriminatory performance. Methods: We retrospectively analyzed adults (≥18 years) who underwent ECPR at a single tertiary center between January 2015 and December 2022. CK was measured at 4, 12, 24, and 48 h after ECPR initiation; lactate at 24 and 48 h. The primary outcome was in-hospital mortality. For each CK time point, we built multivariable logistic regression models adjusted for age, sex, initial rhythm, and total arrest time. Discrimination was assessed using receiver operating characteristic curves. Results: Of 183 patients screened, 102 met the inclusion criteria; 29 (28.4%) survived to discharge. Median total arrest time was longer in non-survivors than in survivors, 40.0 min (28.0-58.0) vs. 30.0 min (20.0-44.0; p = 0.008). CK at 4 h showed limited discrimination (area under the curve = 0.567). CK at 24 and 48 h was higher in non-survivors (24 h: 5444 vs. 3954 U/L, p = 0.045; 48 h: 4793 vs. 2234 U/L, p = 0.003), with the highest predictive value at 48 h (area under the curve = 0.69; optimal cutoff = 2001 U/L). Conclusions: CK levels after ECPR were independently associated with in-hospital mortality, with the moderate predictive performance at 48 h after initiation. CK may serve as an adjunct to risk assessment and management in patients undergoing ECPR.