Abstract
BACKGROUND: The aim of this study was to evaluate institutional volume-outcome relationships in extracorporeal membrane oxygenation (ECMO) with subanalyses of ECMO in patients with a primary diagnosis of respiratory failure. METHODS: All institutions with adult ECMO discharges in the Nationwide Inpatient Sample from 2002 to 2011 were evaluated. International Classification of Diseases (ninth revision) codes were used to identify ECMO-treated patients, indications, and concurrent procedures. Patients who were treated with ECMO after cardiotomy were excluded. Annual institutional and national volume of ECMO hospitalizations varied widely, hence the number of ECMO cases performed at an institution was calculated for each year independently. Institutions were grouped into high-, medium-, and low-volume terciles by year. Statistical analysis included hierarchical, multivariable logistic regression. RESULTS: The in-hospital mortality rates for ECMO admissions at low-, medium-, and high-volume ECMO centers were 48% (n = 467), 60% (n = 285), and 57% (n = 445), respectively (p = 0.001). In post hoc pairwise comparisons, patients in low-volume hospitals were more likely to survive to discharge compared with patients in medium-volume (p = 0.001) and high-volume (p = 0.005) hospitals. There was no significant difference in survival between medium-volume and high-volume hospitals (p = 0.81). In a subanalysis of patients with respiratory failure, low-volume ECMO centers maintained the lowest rates of in-hospital mortality (47%), versus 61% in medium-volume institutions (p = 0.045) and 56% in high-volume institutions (p = 0.15). Multivariable logistical regression produced similar results in the entire study sample and in patients with respiratory failure. CONCLUSIONS: ECMO outcomes in the Nationwide Inpatient Sample do not follow a traditional volume-outcome relationship, and these results suggest that, in properly selected patients, ECMO can be performed with acceptable results in U.S. centers that do not perform a high volume of ECMO.