engleski

A simple prognostic model of in-hospital mortality risk in acutely decompensated chronic heart failure

Purpose: To develop a simple formula for assessing the risk of in-hospital mortality in patients with acutely decompensated chronic heart failure (ADCHF) using already known risk indicators. Methods:We performed single-centre retrospective study of patients consecutively hospitalised for ADCHF. The study was approved by the institution’s Ethics Committee. The retrospective analysis of a derivation group enrolled patients from 2006–2009, (533 survivors, 121 deceased) and identified six risk indicators (age, admission heart rate [AHR], systolic blood pressure [SBP] at admission, blood urea [UR] concentration, serum sodium [Na], and uric acid [UA]). Age, HR and SBP were combined in a formula ([age/10]2×HR/SBP), that was previously used for risk assessment in patients with acute myocardial infarction and proved to be useful in predicting post-discharge mortality and risk for developing HF. However, it has never been used to assess the risk in patients with HF. Since in HF there is an increase of the UR level and decrease of the Na level in the serum, it was assumed that the difference between the UR and Na values would have a greater predictive value than each individual component, thus obtained indicator was (UR - Na/10). UA was a sole parameter. The final formula ([age/10]2×AHR/SBP)+(UR-Na/10)+ UA/100 was tested in a validation group of 591 patients (527 survived, 64 died), enrolled from 2010 – 2011. Results: In the derivation group the value of the area under the ROC curve (AUC) for the final formula was 0.741 (95% confidence interval 0.701 - 0.776). At the cut-off point of 53.17, sensitivity was 53.7% and specificity 83.3%. The discriminative capacity of the formula was significantly higher in relation to each of its components. In the validation group AUC was 0.741 also, (95% confidence interval 0.706 - 0.774). Using the proposed formula, with the cut-off point of 53, sensitivity was 65.6%, specificity 75.7%, the positive predictive value 34.4% and the negative predictive value 94.8%. Conclusions: The identification of individuals at high risk of death among hospitalized patients with ADCHF presents a continuous challenge. In this study, by combining six significant indicators of increased risk of in-hospital mortality (age, AHR, SBP at admission, UR, Na and UA) in a common formula, the predictive power with respect to each individual component was increased. The resulting formula makes possible a simple, rapid and inexpensive assessment of the risk for in-hospital mortality in patients with ADCHF.

chronic heart failure; prognosis; mortality

nije evidentirano