136 - Baseline Chronic Kidney Disease Predicts In-Hospital and Long-Term Mortality in Cardiogenic Shock

Abstract

Background: Despite advances in care, mortality in cardiogenic shock (CS) remains high. Further, there is limited data to stratify patients in CS based on patient characteristics. Methods: Patients admitted to the University of Virginia Coronary Care Unit from 2011 to 2013 were retrospectively analyzed to identify patients with CS as defined by sustained hypotension, reduced cardiac index <2.2 L/min/m2, and elevated pulmonary capillary wedge pressure >15 mmHg. Patients with primary cardiac pathology, pressor requirement, and signs of end-organ damage were also included. Using multivariable logistic regression, baseline patient characteristics and admission data including age, sex, presence of CS on admission, and co-morbidities prior to admission (congestive heart failure, coronary artery disease, diabetes mellitus, chronic kidney disease (CKD), chronic obstructive pulmonary disease, peripheral vascular disease) were evaluated for association with our primary outcome, in-hospital mortality. Three-year survival was assessed with Cox proportional hazards regression. Results: 248 patients were identified with CS and overall in-hospital mortality was 46%. Of all characteristics evaluated, only CKD (OR 3.13, 95% CI 1.72–5.68, P < .0002) was significantly associated with mortality. With adjustment for number of vasopressors/inotropes used, the association of CKD remained significant (OR 2.74, 95% CI 1.43–5.27, P = .0025). CKD was more strongly associated with mortality in CS secondary to acute coronary syndrome (ACS) (n = 93; OR 8.98, 95% CI 2.33–34.61, P = .0014) than in CS secondary to decompensated heart failure (n = 100; OR 2.82, 95% CI 0.93–8.58, P = .068). Patients with CKD also demonstrated lower three-year survival (HR 1.56 for mortality, 95% CI 1.15–2.12, P = .004, Fig. 1). Conclusions: This data demonstrates that the presence of CKD prior to admission is a strong, independent predictor of in-hospital and long-term mortality in CS, particularly in CS due to ACS, while no other factors showed predictive utility. Further investigation is warranted as predictive factors in CS could prove useful in risk stratification, patient counseling, and early therapeutic interventions.