Growth differentiation factor-15 and 5-years cardiovascular risk in chest pain patients presenting to the emergency department: a multicenter cohort study

Abstract

Abstract Background Growth differentiation factor 15 (GDF-15) has been explored as a prognostic biomarker in multiple disease entities, including ischemic heart disease and heart failure (HF), but its role as a prognostic marker in chest pain patients presenting to the emergency department (ED) is unknown. Purpose To investigate the association between GDF-15 and 5-year all-cause mortality, cardiovascular mortality, and major adverse cardiac events (MACE), and to evaluate its prognostic incremental value. Methods Patients presenting to the ED with acute chest pain were prospectively enrolled in an international multicenter diagnostic study. GDF-15 concentrations were measured in a blinded fashion upon ED presentation. Follow-ups were performed at 3 months, 1-, 2- and 5-years. MACE was defined as the composite of cardiovascular mortality, acute myocardial infarction, stroke, and HF rehospitalization. Associations between GDF-15 and each prognostic outcome were modeled using multivariable Cox proportional hazard regression fitting restricted cubic splines to continuous variables. The magnitude of the effect of each GDF-15 unit change on multivariable-adjusted models was assessed graphically using dose-response plots. To compute an HR for the dose-response plots we chose a GDF-15 value of 1200 pg/mL as a reference value as this has been chosen in previous literature1. For the restricted cubic splines three knots were used. The added prognostic value of GDF-15 was evaluated by the increase in the C-statistic between a base and an extended model and the fraction of new information (FNI) provided by GDF-15. Results A total of 4779 chest pain patients had baseline GDF-15 values available. The median age was 61 [48-74] and 33% were female. Higher GDF-15 levels were more strongly associated with heart failure, kidney dysfunction, older age, hypertension, diabetes and coronary artery disease. GDF-15 was independently associated with 5-years all-cause mortality, cardiovascular mortality and MACE (Figure). When adding GDF-15 to a rich-for-prior-information base model which included age, sex, cardiovascular risk factors, creatinine, cardiac troponin and NT-proBNP, GDF-15 provided relevant incremental prognostic discrimination for 5-years all-cause death and cardiovascular death. (Table). Conclusions In chest pain patients presenting to the emergency department, GDF-15 concentration on admission was associated with 5-years all-cause mortality, cardiovascular mortality, and MACE. In addition, it provided relevant incremental prognostic information when added on top of troponin and NT-proBNP, the two current prognostic gold standard biomarkers.FigureTable