RESUMO
The prognostic value of growth differentiation factor-15 (GDF-15) in predicting long-term adverse outcomes in coronary heart disease (CHD) patients remains limited. Our study examines the association between GDF-15 and adverse outcomes over an extended period in CHD patients and firstly assesses the incremental prognostic effect of incorporating GDF-15 into the Framingham risk score (FRS)-based model. This single-center prospective cohort study included 3,321 patients with CHD categorized into 2,479 acute coronary syndrome (ACS) (74.6%) and 842 non-ACS (25.4%) groups. The median age was 61.0 years (range: 53.0-70.0), and 917 (27.6%) were females. Mortality and major adverse cardiovascular events (MACEs) included cardiovascular mortality, myocardial infarction (MI), stroke, and heart failure (HF) (inclusive of HF episodes requiring outpatient treatment and/or hospital admission). Cox regression models assessed the associations between GDF-15 and the incidence of all-cause mortality and MACEs. Patients were stratified into three groups based on GDF-15 levels: the first tertile group (< 1,370 ng/L), the second tertile group (1,370-2,556 ng/L), and the third tertile group (> 2,556 ng/L). The C-index, integrated discrimination improvement (IDI), net reclassification improvement (NRI), and decision curve analysis (DCA) were used to assess incremental value. Over a median 9.4-year follow-up, 759 patients (22.9%) died, and 1,291 (38.9%) experienced MACEs. The multivariate Cox model indicated that GDF-15 was significantly associated with all-cause mortality (per ln unit increase, HR = 1.49, 95% CI: 1.36-1.64) and MACEs (per ln unit increase, HR = 1.29, 95% CI: 1.20-1.38). These associations persisted when GDF-15 was analyzed as an ordinal variable (p for trend < 0.05). Subgroup analysis of ACS and non-ACS for the components of MACEs separately showed a significant association between GDF-15 and both cardiovascular mortality and HF, but no association was observed between GDF-15 and MI /stroke in both ACS and non-ACS patients. The addition of GDF-15 to the FRS-based model enhanced the discrimination for both all-cause mortality (∆ C-index = 0.009, 95% CI: 0.005-0.014; IDI = 0.030, 95% CI: 0.015-0.047; continuous NRI = 0.631, 95% CI: 0.569-0.652) and MACEs (∆ C-index = 0.009, 95% CI: 0.006-0.012; IDI = 0.026, 95% CI: 0.009-0.042; continuous NRI = 0.593, 95% CI: 0.478-0.682). DCA suggested that incorporating GDF-15 into the FRS-based model demonstrated higher net benefits compared to FRS-based models alone (All-cause mortality: FRS-based model: area under the curve of DCA (AUDC) = 0.0903, FRS-based model + GDF-15: AUDC = 0.0908; MACEs: FRS-based model: AUDC = 0.1806, FRS-based model + GDF-15: AUDC = 0.1833). GDF-15 significantly associates with the long-term prognosis of all-cause mortality and MACEs in CHD patients and significantly improves the prognostic accuracy of the FRS-based model for both outcomes.
RESUMO
AIMS: The utility of growth differentiation factor-15 (GDF-15) in predicting long-term adverse outcomes in heart failure (HF) patients is not well established. This study explored the relationship between GDF-15 levels and adverse outcomes in HF patients across various ejection fraction (EF) phenotypes associated with coronary heart disease (CHD) and evaluated the added prognostic value of incorporating GDF-15 into the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) risk score-based model. METHODS AND RESULTS: This single-centre cohort study included 823 HF patients, categorized into 230 (27.9%) reduced EF (HFrEF), 271 (32.9%) mid-range EF (HFmrEF), and 322 (39.1%) preserved EF (HFpEF) groups. The median age was 68.0 years (range: 56.0-77.0), and 245 (29.8%) were females. Compared with the HFrEF and HFmrEF groups, the HFpEF group had a higher GDF-15 concentration (P = 0.002) and a higher MAGGIC risk score (P < 0.001). We examined the associations between GDF-15 levels and the risks of all-cause mortality and HF rehospitalization using Cox regression models. The C-index, integrated discrimination improvement (IDI), and net reclassification improvement (NRI) metrics were employed to assess the incremental prognostic value. During the 9.4 year follow-up period, 425 patients died, and 484 were rehospitalized due to HF. Multivariate Cox regression analysis revealed that elevated GDF-15 levels were significantly associated with an increased risk of all-cause mortality [hazard ratio (HR) = 1.36, 95% confidence interval (CI): 1.20-1.54; P < 0.001] and HF rehospitalization (HR = 1.75, 95% CI: 1.57-1.95; P < 0.001) across all HF phenotypes. This association remained significant when GDF-15 was treated as a categorical variable (high GDF-15 group: all-cause death: HR = 1.73, 95% CI: 1.40-2.14; P < 0.001; HF rehospitalization: HR = 3.37, 95% CI: 2.73-4.15; P < 0.001). Inclusion of GDF-15 in the MAGGIC risk score-based model provided additional prognostic value for all HF patients (Δ C-index = 0.021, 95% CI: 0.002-0.041; IDI = 0.011, 95% CI: 0.001-0.025; continuous NRI = 0.489, 95% CI: 0.174-0.629) and HF rehospitalization (Δ C-index = 0.034, 95% CI: 0.005-0.063; IDI = 0.021, 95% CI: 0.007-0.032; continuous NRI = 0.307, 95% CI: 0.147-0.548), particularly in the HFpEF subgroup. CONCLUSIONS: GDF-15 is identified as an independent risk factor for adverse outcomes in HF patients across the entire EF spectrum in the context of CHD. Integrating GDF-15 into the MAGGIC risk score-based model enhances its prognostic capability for adverse outcomes in the general HF population. This incremental prognostic effect was observed specifically in the HFpEF subgroup and not in other subgroups.