Evaluation of eight biomarkers to predict short-term mortality in patients with acute severe dyspnea

BACKGROUND: Being able to better predict risk and optimal care for patients presenting with acute dyspnea is critical. Prognostic biomarkers are well known: amino-terminal pro-B-type Natriuretic Peptide, troponin, C-reactive protein, procalcitonin. Some were more recently developed: mid-regional pro-A-type natriuretic peptide (Mid Pro-ANP), midregional- pro-adrenomedullin (MR-proADM), pro-endothelin, copeptin. The aim of the paper was to evaluate prognostic value of clinical findings and 8 biomarkers in patients with severe acute dyspnea. METHODS: We designed a prospective cohort study targeting patients admitted in the Emergency Department and in Intensive Care Unit of a University Hospital. Inclusion criteria were acute dyspnea with SpO2 less than 92% and/or respiratory rate (RR) greater than or equal to 25 bpm. Clinical and biological data, including biomarker levels, were recorded. The contribution of the biomarkers in the prognosis was assessed using AUC-ROC curves and by multiple logistic regression. RESULTS: Three hundred and eighty four patients (median age 74 years, 28-day mortality 17%) were enrolled. All biomarkers were available for 317 patients. Main diagnoses were sepsis in 141 cases (36.7%), and acute heart failure in 84 (21.9%) cases. All biomarkers were correlated with prognosis. Pro-ADM (AUC-ROC=0.731; 95% CI: 0.658-0.804) showed the best accuracy. The parameters independently associated with prognosis led to a clinical/biological model with an AUC=0.809 and a good calibration (P (HLchi2)=0.9). Three biomarkers added prognostic information to the model: MR-proADM (P=0.005), copeptin (P=0.006) and troponin (P= 0.05). CONCLUSIONS: Biomarkers can contribute to determine the day-28 outcome of patients with acute severe dyspnea.