Biomarkers and their combination in a prediction of decompensation after an index hospitalization for acute heart failure.

ABSTRACT

Introduction: Heart failure (HF) still remains as one of the most common causes of hospital admission with a high mortality rate. Aim: To investigate the possible prognostic role of brain natriuretic peptide (BNP), high-sensitivity (hs) cardiac troponin (cTn) I, cystatin C, and cancer antigen 125 (CA125) in the prediction of decompensation after an index hospitalization and to investigate their possible additive prognostic value. Patients and Methods: Two hundred twenty-two patients hospitalized with acute HF were monitored and followed for 18 months. Results: BNP at discharge has the highest sensitivity and specificity in the prediction of decompensation. For a cutoff value of 423.3 pg/ml, sensitivity was 64.3% and specificity was 64.5%, with a positive predictive value of 71.6% and an area under the curve (AUC) of 0.69 (P < 0.001). The hazard risk (HR) for decompensation when the discharge BNP was above the cutoff value was 2.18. Cystatin C, at a cutoff value of 1.46 mg/L, had a sensitivity of 57% and specificity of 57.8%, with a positive predictive value of 65.8% and an AUC of 0.59 (P = 0.028). CA125, in the prediction of decompensation in patients with acute heart failure (AHF) and at a cutoff value of 80.5 IU/L, had a sensitivity of 60.5% and specificity of 53.3%, with a positive predictive value of 64.5% and an AUC of 0.59 (P = 0.022). The time till onset of decompensation was significantly shorter in patients with four versus three elevated biomarkers (P = 0.047), with five versus three elevated biomarkers (P = 0.026), and in patients with four versus two elevated biomarkers (P = 0.026). The HR for decompensation in patients with five positive biomarkers was 3.7 (P = 0.001) and in patients with four positive biomarkers was 2.5 (P = 0.014), compared to patients who had fewer positive biomarkers. Conclusion: BNP, cystatin C, and CA125 are predictors of decompensation, and their combined usage leads to better prediction of new decompensation.