Multiparametric Measurements Predict Long-Term Survival in Acute Heart Failure

A dependable method for prediction of long-term mortality risk in patients with acute decompensated heart failure (ADHF) could be based on multiparametric measurements of brain natriuretic peptide (BNP), partial pressure in oxygen (PaO₂), blood urea nitrogen (BUN), and phase angle according to study findings published in the Scandinavian Cardiovascular Journal.

Researchers sought to evaluate the effect of congestion and arterial blood gases (ABG) on long-term mortality in patients with ADHF. The primary endpoint was all-cause death.

They conducted a retrospective study of 252 consecutive patients with acute heart failure referred between January 2010 and November 2013 to the Cardiology Unit of Altamura Hospital, Bari, Italy. Among these patients (77±10 years of age; 50% women), ABG profiles showed 9% acidosis, 77% alkalosis, and 14% had a normal ABG profile. The researchers reported that 39% of patients had hypoxemia, 58% of patients had hypercapnia, and 10% of patients had low bicarbonate.

At admission, BUN, BNP, phase angle (assessed by bioimpedance vector analysis [BIVA]), and PaO₂ analysis were performed. The researchers found that in a median follow-up of 447 days (IQR, 248-667), 29% of patients died within 840 days (median, 106 [IQR, 29-233]) following discharge. They observed respiratory failure type I (n=78) and type II (n=53).

BNP, BUN, PaO₂, phase angle, creatine, plasma levels of hemoglobin, and age associated with higher mortality rates according to univariate Cox regression analysis. This analysis also revealed de-novo/acute decompensation of chronic HF and ischemic dilated cardiomyopathy showed no statistically significant association to mortality rate.

BNP greater than 441 pg/mL, BUN greater than 1.67 mmol/L, PaO₂ equal to or less than 69.7 mm Hg, and phase angle equal to or less than 4.9° were the cut-off points for predicting death according to receiver-operating characteristic curve analysis. When parameter values were dichotomized according to previous calculated cut-offs, a significant increase in mortality rate was noted with Kaplan-Meier analysis. These 4 variables taken together predicted long-term mortality (60% of deaths) in ADHF (area under the curve, 0.78; 95% CI, 0.72–0.78) and produced a multiparametric score that with each single-unit increase indicated a 2.2-fold increase in the risk of death (hazard ratio, 2.2; 95% CI, 1.8–2.8; P <.0001).

The retrospective design, lack of evaluation of other biomarkers, and lack of data concerning therapies at discharge represent significant limitations of this study.

“PaO₂, BUN, BNP, and PhA [phase angle] as assessed by BIVA predicts long-term mortality in ADHF patients,” the study authors wrote. “A multiparametric approach based on these markers could explain the 60% of long-term death of patients admitted for ADHF.”