Use of Biomarkers to Predict Readmission for Congestive Heart Failure

Abstract

Acute decompensated heart failure (ADHF) is a major reason for repeated hospitalizations. Identifying those patients with ADHF at risk for readmission is critical so that preventive interventions can be implemented. Biomarkers such as B-type natriuretic peptide (BNP), high-sensitivity troponin I, and galectin-3 (Gal-3) assessed at discharge may be useful, although their role in predicting short-term readmission is not well defined in the literature. We enrolled and had follow-up data for 101 participants admitted to our facility from April 2013 to March 2015 with a primary diagnosis of ADHF. Gal-3, high-sensitivity troponin I, and BNP were obtained within 48 hours before hospital discharge after management of ADHF. Gal-3 was assessed using 2 commercially available assays. We compared subjects who were and were not readmitted. Discharge BNP was found to be a significant predictor of 30- and 60-day readmission (area under the curve [AUC] 0.69 [p = 0.046], AUC 0.7 [p = 0.005], respectively). The addition of Gal-3 to discharge BNP provided significantly improved prediction of 60-day readmission. Gal-3 alone was found to be a significant predictor of 60-day readmission in patients with preserved ejection fraction (AUC 0.85, p <0.001). The net reclassification improvement was 55.2 (p = 0.037). Using multivariate analysis, for every 100 pg/L BNP increase, the probability of readmission increased by approximately 10%, and for every 1-ng/ml Gal-3 increase, the probability further increased 8%. A statistically significant net reclassification improvement was not found on examination of 30-day readmission. In conclusion, measurement of both Gal-3 and BNP at hospital discharge provides significant prediction of hospital readmission within 60 days. When combined, the prediction of readmission is significantly improved.