Abstract Background Extracorporeal cardiopulmonary resuscitation (ECPR) can improve survival in patients with refractory cardiac arrest (CA). However, defining optimal selection criteria for ECPR remains a major challenge. Methods We retrospectively analyzed all ECPR treatments for refractory in-hospital CA (IHCA) and out-of-hospital CA (OHCA) in adult patients from January 1, 2010 through December 31, 2024 at our tertiary 35-bed Intensive Care Unit. Before July 2017 (Period 1), ECPR was implemented at physician discretion. From July 2017 (Period 2), a dedicated protocol recommended physicians to implement ECPR based on four criteria: age < 70 years, shockable rhythm, no-flow duration < 5 min, and total low-flow duration < 80 min. The primary outcome was hospital mortality. The secondary outcome was good neurological outcome at 3 months, defined by a cerebral performance category (CPC) score of 1 or 2. Results A total of 166 patients (45 in period 1, 121 in period 2), including 80 IHCAs and 86 OHCAs, were included. The proportion of patients fulfilling the 4 criteria was low yet significantly greater in period 2 than in period 1 (35.0 vs. 17.8%, p = 0.027). Hospital survival was improved in period 2 (26.5% vs. 8.9%, p = 0.015), whereas good neurological outcome was not (14.9 vs. 6.7%, p = 0.157). When evaluating the impact of the 4 criteria over the whole study period, patients with 4 criteria vs. those with < 4 criteria displayed marked improvements in survival (48.0 vs. 9.6%, p < 0.001) and good neurological outcome (30.0 vs. 5.2%, p < 0.001). In multivariable analysis, only the simultaneous presence of the 4 criteria was independently associated with a decreased risk of death (OR = 0.11, 95% CI 0.01–0.87, p = 0.037), whereas no single criterion alone was significantly predictive. Conclusion Implementing a clinical ECPR protocol in our institutional practice improved meaningful survival in patients with refractory IHCA and OHCA fulfilling four predefined criteria including an age < 70 years, a shockable rhythm, a no-flow < 5 min, and a low-flow < 80 min.
Ltaief et al. (Mon,) studied this question.