NYMC Faculty Publications


Cardiogenic Shock During Heart Failure Hospitalizations: Age-, Sex-, and Race-Stratified Trends in Incidence and Outcomes

First Page


Last Page


Document Type


Publication Date

April 2019




The objectives were to study the overall and age-, sex-, and race-stratified incidence of cardiogenic shock (CS) during heart failure hospitalizations (HFHs) not complicated by acute coronary syndromes (ACS), utilization of short-term mechanical circulatory support (MCS) and in-hospital mortality with non-ACS-related CS, and respective temporal trends. Data are lacking regarding the epidemiology of non-ACS-related CS during HFHs. METHODS: Retrospective observational analysis of the National Inpatient Sample 2005-2014 to identify all HFHs in adult patients without concomitant ACS. RESULTS: Of 8,333,752 HFHs, incidence rate of non-ACS-related CS was 8.7 per thousand HFHs (N=72,668), a 4-fold increase from 4.1 to 15.6 per thousand HFHs between 2005 and 2014 (Ptrend <.001). Among those with non-ACS-related CS, utilization rates of intra-aortic balloon pump, extracorporeal membrane oxygenation, and temporary ventricular assist devices were 12.8%, 1.4%, and 2.5%, respectively. Respective 2005 to 2014 trends were 14.2% to 10.7%, 0.6% to 1.8%, and 0.8% to 2.7% (Ptrend for all, <.001). In-hospital mortality rate was 27.1%, with a substantial decrease from 42.4% in 2005 to 23.3% in 2014 (Ptrend <.001). These temporal trends were largely consistent across age, sex, and race subgroups. CONCLUSION: During HFHs in the United States, non-ACS-related CS occurred infrequently but was associated with substantial mortality. Non-ACS-related CS incidence and certain MCS utilization rates increased, and in-hospital mortality rate decreased between 2005 and 2014. These trends were generally homogenous across the age, sex, and race groups. The observed trends in incidence and mortality may be a reflection of increased identification of CS during HFHs, although further study is needed to assess whether temporal changes in care may have influenced outcomes.