Impact of Center of Admission on Receipt of Extracorporeal Membrane Oxygenation Among Patients With Hypoxemic Respiratory Failure

Background

Given the resources and specialized training required to deliver extracorporeal membrane oxygenation (ECMO), the provision of ECMO often is centralized within expert centers. Spurred by recent evidence, the use of ECMO has increased dramatically. However, given the centralized nature of ECMO, it is possible that inequities in access exist.

Research Question

Does center of admission impact the likelihood of receiving ECMO among adults with moderate or severe acute hypoxemic respiratory failure (Pao₂ to Fio₂ ratio ≤ 200 mm Hg within 48 h of ventilation).

Study Design and Methods

We performed a retrospective cohort study using data from the Large Observational Study to Understand the Global Impact of Severe Acute Respiratory Failure (LUNG SAFE) and Reseau Europeen de Recherche en Ventilation Artificielle (REVA) Influenza A(H1N1) Registry databases. Using modified log-Poisson analysis, we estimated the likelihood of receiving extracorporeal membrane oxygenation (ECMO) (initiation at the admitting hospital or transfer for initiation), adjusting for disease severity over time. To explore unmeasured confounding, we evaluated the association between center of admission on three negative controls: neuromuscular blockade, prone positioning, and dialysis.

Results

Among 1,581 patients (37.7% female patients; mean age, 60.7 years), 76 patients (4.8%) received ECMO. Longitudinal analysis, adjusted for trends in disease severity, demonstrated that patients admitted to ECMO centers were two times more likely to receive ECMO than those admitted to non-ECMO centers (relative risk [RR], 2.00; 95% CI, 1.17-3.41). Patients at ECMO centers received ECMO 2 days earlier than those at non-ECMO centers: median time to initiation was 1 day (interquartile range, 1-5 days) vs 3 days (interquartile range, 2-5 days; P = .04). Center of admission was not associated with neuromuscular blockade (RR, 1.08; 95% CI, 0.90-1.30), prone positioning (RR, 0.93; 95% CI, 0.68-1.28), or dialysis (RR, 1.04; 95% CI, 0.84-1.27).

Interpretation

Adults with hypoxemic respiratory failure admitted to ECMO centers were twice as likely to receive ECMO as those admitted to non-ECMO centers. These finding raise concerns regarding equity in access to care and suggest a potential lower threshold among clinicians at ECMO centers for initiation of ECMO.