Bembea MM, Ng DK, Rizkalla N, Rycus P, Lasa JJ, Dalton H, Topjian AA, Thiagarajan RR, Nadkarni VM, Hunt EA; American Heart Association’s Get With The Guidelines – Resuscitation Investigators.
Crit Care Med. 2019 Feb 8. doi: 10.1097/CCM.0000000000003622. [Epub ahead of print]
PMID: 30747771
Select Item 30753393
Take-Home Points:
- Extracorporeal membrane oxygenation (ECMO) in the setting of cardiopulmonary resuscitation is a necessary and useful adjunct salvage therapy, but is characterized by high peri-procedural and in-hospital mortality.
- Several pre-existing patient conditions, procedural considerations, and ECMO-related complications predispose to an increased risk of mortality, some of which may be improved upon to afford better outcomes into the future.
Commentary from Dr. Michael Ma (Stanford, CA), section editor of Congenital Heart Surgery Journal Watch: Extracorporeal membrane oxygenation (ECMO) cardiopulmonary resuscitation (ECPR) is a critical salvage therapy for patients that fail return of spontaneous circulation (ROSC) despite more conventional Pediatric Advanced Life Support (PALS) algorithms. This study, designed as a multi-center, multi-registry, observational cohort, provides a high-level glimpse at the use of ECPR across the pediatric population (<18 years of age) spanning 32 hospitals over a 15 year period (2000-2014), totaling 593 cases. Registry data from Extracorporeal Life Support Organization (ELSO) and American Heart Association Get With the Guidelines-Resuscitation were linked and de-identified prior to analysis, which was performed using a variety of established statistical methods.
Of these 593 cases, median age 2.9mo, with 70% (414) infants, and 59% (351) were male. 59% (349) were cardiac surgical patients. Median duration of CPR prior to initiation of ECMO was 48 minutes (IQR 28-70 min), and median duration of ECMO support 3.9 days (IQR 2.0-6.7 days). Primary outcome (mortality) was characterized by 40.5% (240) death prior to ECMO decannulation, and 59.4% (352) death prior to hospital discharge. Increasing risk of mortality was associated with non-cardiac diagnosis, prearrest sepsis, respiratory insufficiency, renal insufficiency, longer duration from CPR activation to ECMO initiation, and adverse events reported while on ECMO. In multivariate regression (Figure 3, included below), a noncardiac diagnosis (adjusted OR 1.85 (95% CI 1.19-2.89)) and renal insufficiency (aOR 4.74 (95% CI 12.06-10.9)) were the strongest predictors of death. Each 5 minute incremental delay in the initiation of ECMO increased risk of death by 4% (aOR 1.04 (95% CI 1.01-1.07)). Adverse events on ECMO increased the risk of death, with neurologic and pulmonary complications being the most striking (aOR 2.76 and 2.55 respectively). Multiple adverse events compounded the risk of death further.
These findings corroborate existing, primarily single-center studies that examine ECPR in the pediatric population. The trends discovered are not altogether surprising; patients that require ECPR are at high risk of mortality, and those patients that have pre-existing conditions (i.e. non-cardiac diagnosis, renal insufficiency), peri-procedural difficulty (i.e. time to ECMO support), and post-procedural complications (i.e. adverse events on ECMO), are at the highest risk. The study does add substantially to our specific understanding and historical use of ECPR however, in that it aggregates two large and comprehensive datasets to provide an intimate and quantified understanding of how we, as a community, have performed with this therapy, and how we might go about improving that performance (i.e. patient selection, procedural efficiency, etc.) and counseling patients and families regarding expected prognosis.
