The impact of hospital volume on the prognosis of patients with cardiogenic shock requiring mechanical circulatory support

Abstract Background Short-term mechanical circulatory support (MCS) is indicated in refractory cardiogenic shock (CS). Each hospital's volume of MCS relates to the many factors including population served, clinician experience and availability of MCS. To establish regional integrated Hub-and-spoke care CS systems for improving the prognosis of patients with CS, detailed data on relationship between the volume of MCS and prognosis are required but are lacking. We evaluated the relationship between outcomes, including in-hospital mortality and case volume, and explored whether there is an upper limit to this relationship. Methods We used the Japanese nationwide database to identify patients receiving short-term MCS for CS between April 2012 and March 2020. Of 65,837 patients receiving short-term MCS, three sub-cohorts were created based on the MCS type used; the intra-aortic balloon pump (IABP) alone (n=48,643), the extracorporeal membrane oxygenation (ECMO) (n=16,871), and the Impella cohorts (n=696). Results The median volume of annual cases in each hospital was 13.5 (7.4–22.1) in the IABP alone cohort, 6.4 (3.4-11.0) in the ECMO cohort, and 7.5 (4.0–10.7) in the Impella cohort. In the IABP alone and ECMO cohorts, the group with the higher case volume had less acute myocardial infarctions and received right heart catheterization more frequently. The highest quintile for the volume of cases in the IABP alone and ECMO cohorts had the lowest in-hospital mortality (IABP alone cohort, 25.1% in quintile 1 vs. 15.2% in quintile 5; ECMO cohort, 73.7% in quintile 1 in 67.4% in quintile 5), while it had the highest cost. Adjusted odds ratios for in-hospital mortality increased as case volume increased (IABP alone cohort, 0.63 [0.58–0.68] in quintile 5; ECMO cohort, 0.73 [0.65–0.82] in quintile 5, with the lowest quintile as reference) but did not increase significantly in the Impella cohort (0.90 [0.58–1.39] in tertile 3, with the lowest tertile as reference) (Figure 1). In continuous models, in-hospital mortality decreased to 28 cases/year in the IABP alone cohort and 12 cases/year in the ECMO cohort, and the mortality became almost flat above that (Figure 2). Hospitals with 28 or more cases/year in the IABP cohort were 3.6 % including 15.5 % of patients, and hospitals with 12 or more cases/year in the ECMO cohort were 4.7% including 20.9% of patients. Conclusions Higher volumes of IABP and ECMO are associated with a lower mortality. There is an upper limit to the decline, above which in-hospital mortality remains constant or increases slightly. Centralizing patients with refractory cardiogenic shock in a particular hospital might improve patient outcomes in each region.