Background-Growth-differentiation factor-15 (GDF-15) is emerging as a prognostic biomarker in patients with coronary artery disease. Little is known about GDF-15 as a biomarker in patients with heart failure. Methods and Results-The circulating concentration of GDF-15 was measured at baseline (nϭ1734) and at 12 months (nϭ1517) in patients randomized in the Valsartan Heart Failure Trial (Val-HeFT). GDF-15 levels at baseline ranged from 259 to 25 637 ng/L and were abnormally high (Ͼ1200 ng/L) in 85% of patients. Higher levels were associated with features of worse heart failure and biomarkers of neurohormonal activation, inflammation, myocyte injury, and renal dysfunction. Baseline GDF-15 levels (per 100 ng/L) were associated with the risks of mortality (hazard ratio, 1.017; 95% confidence interval, 1.014 to 1.019; PϽ0.001) and first morbid event (hazard ratio, 1.020; 95% confidence interval, 1.017 to 1.023; PϽ0.001). In a comprehensive multiple-variable Cox regression model that included clinical prognostic variables, B-type natriuretic peptide, high-sensitivity C-reactive protein, and high-sensitivity troponin T, GDF-15 remained independently associated with mortality (hazard ratio, 1.007; 95% confidence interval, 1.001 to 1.014; Pϭ0.02) but not first morbid event. At 12 months, the GDF-15 levels had increased by a similar amount in the placebo and valsartan groups (Pϭ0.94). Increases in GDF-15 over 12 months were independently associated with the risks of future mortality and first morbid event also after adjustment for clinical prognostic variables, B-type natriuretic peptide, high-sensitivity C-reactive protein, and high-sensitivity troponin T and their changes. Conclusions-GDF-15 reflects information from several pathological pathways and provides independent prognostic information in heart failure. GDF-15 levels increase over time, suggesting that GDF-15 reflects a pathophysiological axis that is not completely addressed by the therapies prescribed in Val-HeFT. (Circulation. 2010;122:1387-1395.)