Head-to-head comparison of diagnostic scores for acute heart failure in the emergency department: results from the PARADISE cohort

Abstract

BREST and PREDICA scores have recently emerged for the diagnosis of acute heart failure (AHF) in the emergency department (ED). This study aimed to perform a head-to-head comparison in a large contemporary cohort. BREST and PREDICA scores were calculated from, respectively, 11 and 8 routine clinical variables recorded in the ED in 1386 patients from the PArADIsE cohort. The diagnostic performance of the scores for adjudicated AHF diagnosis was assessed by the area under the ROC curve (AUC). Acute HF diagnosis was adjudicated according to the European Society of Cardiology criteria and BNP levels. A BREST score ≤ 3 or PREDICA score ≤ 1 was associated with low probabilities of AHF (5.7% and 2.6%, respectively). Conversely, a BREST score ≥ 9 or PREDICA score ≥ 5 was associated with a high risk of AHF diagnosis (77.3% and 66.9%, respectively) although more than half of the population was within the "gray zone" (4-8 and 2-4 for the BREST and PREDICA scores, respectively). Diagnostic performances of both scores were good (AUC 79.1%, [66.1-82.1] for the BREST score and 82.4%, [79.8-85.0] for the PREDICA score). PREDICA score had significantly higher diagnostic performance than BREST score (increase in AUC 3.3 [0.8-5.8], p = 0.009). Our study emphasizes the good diagnostic performance of both BREST and PREDICA scores, albeit with a significantly higher diagnostic performance of the PREDICA score. Yet, more than half of the population was classified within the "gray zone" by these scores; additional diagnostic tools are needed to ascertain AHF diagnosis in the ED in a majority of patients. Clinical trial registration: NCT02800122.

Keywords: Cardiac edema; Cardiovascular diseases; Diagnostic; Dyspnea; Heart failure.