Growth-differentiation factor-15 improves reclassification for the diagnosis of heart failure with normal ejection fraction in morbid obesity

• Published in: European journal of heart failure Vol. 14; no. 11; pp. 1240 - 1248
• Main Authors: Baessler, Andrea, Strack, Christina, Rousseva, Elena, Wagner, Florian, Bruxmeier, Janine, Schmiedel, Martin, Riegger, Günter, Lahmann, Claas, Loew, Thomas, Schmitz, Gerd, Fischer, Marcus
• Format: Journal Article
• Language: English
• Published: England Blackwell Publishing Ltd 01.11.2012
• Subjects: Adolescent; Adult; Analysis of Variance; Biomarkers; Female; GDF-15; Growth Differentiation Factor 15; Health Status Indicators; Heart failure with normal ejection fraction; Heart Failure, Diastolic - diagnosis; Heart Failure, Diastolic - diagnostic imaging; Heart Failure, Diastolic - pathology; Humans; Male; Middle Aged; Natriuretic Peptide, Brain; Natriuretic peptides; Obesity; Obesity, Morbid - complications; Obesity, Morbid - pathology; Peptide Fragments; Predictive Value of Tests; Prognosis; Statistics as Topic; Ultrasonography; Ventricular Dysfunction, Left - diagnosis; Ventricular Dysfunction, Left - diagnostic imaging; Ventricular Dysfunction, Left - pathology; Young Adult
• ISSN: 1388-9842; 1879-0844; 1879-0844
• DOI: 10.1093/eurjhf/hfs116

Aims This study aimed to examine the incremental value of growth‐differentiation factor‐15 (GDF‐15) to N‐terminal pro brain natriuretic hormone (NT‐proBNP) levels for the diagnosis of left ventricular diastolic dysfunction (LVDD) and possible heart failure (HF) in morbidly obese patients. Method and results We analysed data from 207 obese subjects [body mass index (BMI) 41 ± 8 kg/m2] with normal ejection fraction, LVDD, and symptoms and/or signs of HF (referred to as ‘LVDD with possible HF’, n = 88) and with normal left ventricular function (n = 119) before participating in a medical weight loss programme, in addition to the study of healthy lean subjects (n = 51). Median NT‐proBNP (interquartile range) for obese subjects with ‘LVDD and possibe HF’ and with normal LV function was 52 (29–96) and 42 (25–66) pg/mL, respectively (P = 0.12). There was no correlation of NT‐proBNP with parameters of left ventricular filling pressure, i.e. E/E' (r2 = 0.002, P = 0.63) or E' velocity (r2 = 0.02, P = 0.24). In contrast, GDF‐15 was 665 (496–926) with ‘LVDD and possible HF’ and 451 (392– 679) pg/mL without (P < 0.0001). GDF‐15 was significantly correlated to E/E', E' velocity, E/A ratio, isovolumetric relaxation time, duration of reversed pulmonary vein atrial systolic flow, and left atrial size. The area under the receiver operating characteristic curve that defines LVDD with possible HF was 0.56 for NT‐proBNP and 0.74 for GDF‐15 (P < 0.0001). The addition of GDF‐15 to a multivariate predicition model increased the net reclassification improvement (NRI) by 9% (P= 0.022). Conclusion In morbidly obese individuals, GDF‐15 levels seem to better correlate with diastolic dysfunction than NT‐proBNP levels. GDF‐15 significantly improves reclassification for the diagnosis of ‘LVDD with possible HF’ and, thus, adds incremental value to NT‐proBNP.