Risk stratification in patients with chronic heart failure based on metabolic-immunological, functional and haemodynamic parameters

BackgroundA vast array of parameters has been proposed to predict mortality in chronic heart failure (CHF). Their applicability into clinical practice remains challenging due to economical and availability considerations.Methods and resultsWe studied serum uric acid, total cholesterol, and soluble tumour necrosis factor receptor 1 (sTNF-R1) in 114 CHF patients (63.0 ± 1.0 years, NYHA functional class I/II/III/IV: 11/34/54/15) recruited prospectively into a metabolic study program. All patients underwent assessment of left ventricular ejection fraction and measurement of peak oxygen consumption (pVO2). Patients were followed for 24 months or until death. A total of 31 patients died; cumulative survival was 78% (95% confidence interval [CI] 70-86%) and 73% (65-81%) at 12 and 24 months, respectively. In single predictor Cox proportional hazard analysis, uric acid, pVO2, sTNFR-1, LVEF (all p < 0.0001) and cholesterol (p < 0.02) all predicted survival. All parameters remained significant predictors of death after multivariable adjustment (all p < 0.02). Receiver-operator characteristic (ROC) curve analyses showed that uric acid and sTNF-R1 are equally strong with regards to their prognostic performance in CHF like pVO2, but even better than LVEF. The combination of pVO2, LVEF, uric acid, and sTNF-R1 in ROC statistics turned out as the best model with the highest prognostic value in CHF (AUC: 0.91, sensitivity: 90.4, specificity: 74.2, p = 0.0001).ConclusionIncluding metabolic-immunological parameters into risk assessment might result in a better risk stratification than modeling based on clinical parameters alone, probably due to a better reflection of CHF as multisystem disease. We suggest metabolic-immunological parameters to be tested in larger populations.