Candidacy Decision Making for Extracorporeal Cardiopulmonary Resuscitation (ECPR): Lessons from a Single-Center Retrospective Analysis.

Objective: Extracorporeal cardiopulmonary resuscitation (ECPR) is increasingly used to rescue patients in cardiac arrest. However, minimal data guide candidacy decisions, and centers must develop their own initiation criteria, raising concern for inconsistent application between and even within centers. This single-center analysis of ECPR decisions was conducted to demonstrate an internal review process, identify patterns of inconsistency, and generate hypotheses for potential sources of inappropriate inconsistency and means of mitigation.

Design: Retrospective cohort study.

Setting: Single quaternary academic center.

Participants: Seventy-three patients for whom ECPR was considered between 2021 and 2024.

Interventions: None.

Measurements and main results: Seventy-three consultations resulted in 14 candidates who underwent ECPR, 53 noncandidates, and 6 patients who achieved return of spontaneous circulation before a decision. Twenty unique contraindications were invoked across all noncandidates; the 5 most common were duration of CPR (n = 21), age (n = 17), nonshockable rhythm (n = 16), comorbidities (n = 15), and acidemia (n = 11). We identified 5 patterns of inconsistency: in (1) application of contraindications between candidates and noncandidates, (2) invoked contraindications between noncandidates, (3) application of contraindications in young and peri- and postoperative patients, (4) documentation, and (5) terminology use. We propose Domain-Based Decision-Making invoking contraindications to inform whether the patient belongs to 1 of 3 prognostic domains: (1) inability to achieve cardiovascular recovery/destination therapy or (2) meaningful neurologic recovery, or (3) ECPR technically/practically infeasible.

Conclusions: We demonstrate an effective process for assessing internal candidacy decision making processes for centers performing ECPR. We identify 5 patterns of inconsistency, propose a Domain-Based Decision-Making model, and share lessons likely applicable to other centers.