Prognostic value of soluble ST2 in the Valsartan heart failure trial

Background: Soluble ST2 (sST2), a biomarker related to inflammation, is associated with outcomes of patients with heart failure. In-depth analyses of the relationship among sST2, changes in sST2, and patient outcomes are reported. Methods and Results: sST2 was measured at baseline (n=1650), 4 months (n=1345), and 12 months (n=1094) in Valsartan Heart Failure Trial. Baseline sST2 averaged 28.7&plusmn;16.2 ng/mL, significantly (P<0.001) higher in men than women but supranormal in only 9% and 15%, respectively. A continuous relationship between sST2 and the log hazard ratio for outcomes was modeled as 2 linear segments with a significant decrease in the rate of increase in hazard ratios >33.2 ng/mL. Each segment of the sST2 distribution was significantly (P<0.0001) associated with the risks of morbid event, mortality, and hospitalization for heart failure. Only sST2 values <33.2 ng/mL were significantly related to the outcomes when 23 readily available clinical variables including N-terminal probrain natriuretic peptide were included in the Cox regression model. sST2 did not improve discrimination of patient outcomes. Compared with placebo, valsartan significantly (P<0.001) reduced the rate of increase in sST2. Increases in sST2 for 12 months, but not decreases, were significantly associated with subsequent outcomes, independent of clinical variables, sST2, and valsartan treatment. Conclusions: In this study, baseline sST2 was nonlinearly associated with patient outcomes but did not provide substantial prognostic information when added to a clinical prediction model that included N-terminal probrain natriuretic peptide. An increase but not decrease in sST2 was independently associated with outcomes. Additional research is needed to determine whether monitoring ST2 levels can improve patient outcomes.