Antithrombin III predicts the hepatic reserve in terminal heart failure patients with cardiogenic shock

Abstract

Objective: Acute liver failure (ALF) developing from passive backward congestion and foreward ischemia in heart failure (HF) with cardiogenic shock is associated with high mortality. Aim of our study was to investigate independent predictors for ALF in HF patients undergoing mechanical bridging in cardiogenic shock. Methods: A consecutive series of 164 terminal HF patients in NYHA functional class IV and ACC/AHA stage D undergoing mechanical bridging by extracorporeal membrane oxygenation (ECMO) and ventricular assist device (VAD) was studied to investigate independent predictors for the development of subsequent ALF. Development of ALF was defined by the Kings College for ALF. Clinical conditions of HF, hemodynamic and laboratory parameters at the timepoint of device implantation were assessed by means of multivariate logistic regression analysis. The study was approved by the local Institutional Review Board. Results: A total of 45 HF patients (27.4%) developed subsequent ALF with a hospital mortality of 88.4% (40 patients). In uni-variate analysis, neither duration of HF (new-onset, transient, or chronic, p=0.84), nor the ethiology of HF (ischemic or non-ischemic, p=0.82) were associated with subsequent ALF. Previous cardiopulmonary resuscitation (CPR) or implementation under CPR were also not predictive for ALF. In contrast, the number of previous cardiac decompensations, the need for hemofiltration at bridging were highly associated with subsequent ALF (p<0.001). Hemodynamic parameters, such as MAP (p=0.005), mPAP (p=0.042), but not CVP, PCWP or cardiac index were associated with later ALF. Liver specific laboratory parameters such as cholinesterase, INR, total bilirubine, antithrombin III, lactate and pH (all p<0.001) but not transaminases were predictive for ALF. In multivariate analysis, however, antithrombin III was the most relevant parameter indicating onset of profound ALF (Wald: 11.8; Relative Risk (RR): 0.84, 95% Confidence Interval (CI): 0.77–0.93, p=0.001). Additionally, the need for hemofiltration at bridging (Wald: 4.7, RR: 8.6, 95% CI: 1.2–61.2, p= 0.031) and repeated decompensations (Wald: 4.0, RR: 6.7, 95% CI: 1.1–43.0, p=0.046) were independent predictors for ALF. Conclusion: Beside preexisting renal failure and repeated decompensations, decreased antithrombin III revealed to be the most relevant predictor for hepatic reserve in HF patients with cardiogenic shock resistant to conservative treatment. As antithrombin III is a not routinely used marker in clinical practice, it should be implemented into cardiologic practice to prevent HF patients from fatal ALF.