Ten-year outcomes of extracorporeal life support for in-hospital cardiac arrest at a tertiary center

Abstract

Extracorporeal cardiopulmonary resuscitation (ECPR) is controversial, given both the lack of evidence for improved outcomes and clarity on appropriate candidacy during time-sensitive cardiac arrest situations. The primary objective of our study was to identify factors predicting successful outcomes in ECPR patients.Between March 2007 and November 2018, 112 patients were placed on extracorporeal life support (ECLS) during active CPR (ECPR) at our institution. The primary outcome was survival to hospital discharge. Survivors and non-survivors were compared in terms of pre-cannulation comorbidities, laboratory values, and overall outcomes. Multivariable logistic regression was used to identify pre-cannulation predictors of in-hospital mortality. Among 112 patients, 44 (39%) patients survived to decannulation and 31 (28%) survived to hospital discharge. The median age was 60 years (IQR 45–72) with a median ECLS duration of 2.2 days (IQR 0.6–5.1). Patients who survived to discharge had lower rates of chronic kidney disease than non-survivors (19% vs. 41%, p = 0.046) and lower baseline creatinine values [median 1.2 mg/dL (IQR 0.8–1.7) vs. 1.7 (0.7–2.7), p = 0.008]. Median duration from CPR initiation to cannulation was 40 min (IQR 30–50) with no difference between survivors and non-survivors (p = 0.453). When controlling for age and CPR duration, multivariable logistic regression with pre-procedural risk factors identified pre-arrest serum creatinine as an independent predictor of mortality [OR 3.25 (95% CI 1.22–8.70), p = 0.019] and higher pre-arrest serum albumin as protective [OR 0.32 (95% CI 0.14–0.74), p = 0.007]. In our cohort, pre-arrest creatinine and albumin were independently predictive of in-hospital mortality during ECPR, while age and CPR duration were not.