Extracorporeal membrane oxygenation for cardiogenic shock in the United States-association between hospital volume and outcomes

Background: There has been a rapid increase in the use of extracorporeal membrane oxygenation (ECMO) in patients who present with cardiogenic shock (CS). We sought to examine the relationship between hospital ECMO volumes and clinical outcomes in such patients. Methods: Using the Nationwide Readmission Database we identified adult patients in CS who received ECMO from 2010-2015. Institutions were grouped into tertiles based on annual ECMO volume: low (≤6/year), medium (7-18/year) and high (≥19/year) volume. Results: Of the 763 hospitals included, 611 (80%) were categorized as low-volume, 109 (14.2%) as medium-volume and 43 (5.6%) as highvolume hospitals. Of the total 8265 patients, 3008 (36.3%), 2812 (34%), and 2445 (29.5%) patients were admitted to low-, medium-, and high-volume centers, respectively. In-hospital mortality was significantly higher in low-volume hospitals compared to high-volume hospitals (62% vs 52%, p = 0.001). There was no difference in 30-day mortality and vascular complications between low-volume and highvolume hospitals (0.7% vs 0.2%, p = 0.45 and 58.9% vs 65.9%, p = 1.28) respectively. Median length of stay was significantly lower in low-volume hospitals compared to both medium and high-volume hospitals (9 days vs 14 days vs 20 days, p \u3c 0.001). Hospitalization cost at low-volume hospitals was less compared to both medium- and high-volume hospitals ($121,437 vs $ 153,029 vs $172,947, p \u3c 0.05). Conclusions: High-volume ECMO hospitals had lower in-hospital mortality in patients who presented with CS when compared to low volume hospitals. (Table Presented)