Eric Etchill, MD, MPH
Surgery resident
The Johns Hopkins Hospital
Baltimore, MD
Disclosure information not submitted.
Glenn Whitman, MD
Professor
Johns Hopkins Hospital, United States
Disclosure information not submitted.
Bo Kim, MD
Assistant Professor
Johns Hopkins Pulmonary and Critical Care Medicine, United States
Disclosure information not submitted.
Chun (Dan) Woo Choi, MD
Physician
Johns Hopkins Hospital, United States
Disclosure information not submitted.
Romergryko Geocadin, MD
Professor
Johns Hopkins Hospital, United States
Disclosure information not submitted.
Sung-Min Cho, DO, MHS
Assistant Professor
Johns Hopkins Hospital, United States
Disclosure information not submitted.
Title: A Study of Withdrawal of Life Sustaining Therapy in Extracorporeal Membrane Oxygenation
Introduction: Withdrawal of life sustaining therapy (WLST) is an important cause of mortality after initiation of extracorporeal membrane oxygenation (ECMO). There are limited data on WLST after ECMO. Understanding the patient and ECMO factors associated with WLST can aid in avoiding premature WLST. We describe the patient population in an international registry who experienced WLST.
Methods: The Extracorporeal Life Support Organization (ELSO) database was queried from 2007-2017 for patients > 18 years who died on ECMO due to family request. Data after 2017 were excluded due to a form change. Patients were stratified by early WLST < 1 week and late WLST > 1 week. One week was selected as an adequate time trial for venoarterial (VA) and venovenous (VV) ECMO. Continuous variables were compared using t-test and categorical variables were compared using Chi-square test.
Results: 2911 patients (63.6% VA; 30% VV; 6.4% other) experienced WLST during the study period (median age 58; interquartile range 24-77; 36.2% female) with an average ECMO duration of 211.6±293.9 hours. Cardiac diagnoses (48.6%) were the most frequent ECMO indication. Early WLST occurred in 61.4% of patients with an average duration of 67.5± 46.48 hrs. Race was also a significant factor in comparing timing of WLST (p = 0.032), with Black and Hispanic patients more frequently undergoing late WLST. Early WLST patients were older (56.4±15.6 vs 52.8±15.3 years; p< 0.001) but did not vary by sex or body mass index. Early WLST varied significantly by ECMO primary indication (p < 0.001): early WLST most frequently had a cardiac indication (54.9%) compared to late WLST who most commonly had a pulmonary indication (53.9%). This was confirmed when evaluated by ECMO mode: 74.5% of early WLST compared to late WLST were on a VA mode (p < 0.001). Of the hemodynamic, ventilator, and laboratory data, baseline pH was significantly lower in early WLST (7.22 ±0.167 vs. 7.26 ±0.143; p< 0.001).
Conclusions: Patients with early WLST were older most frequently had cardiac indications and VA mode. Patients with pulmonary indications, VV mode, and Black or Hispanic were more frequently observed in the WLST > 1 week. Future studies of WLST in ECMO should be prospective and consider qualitative description of WLST.