Introduction: Prognostication for heart failure (HF) outcomes remains challenging in patients with type 2 diabetes mellitus (T2DM) post acute coronary syndrome (ACS). This study evaluated both a clinical model and a biomarker based approach [using growth-differentiation-factor-15 (GDF-15), high-sensitivity troponin I (hsTnI), N-terminal proBNP (NT-proBNP), adiponectin, galectin-3 (GAL3), fibroblast growth factor 23 (FGF 23), ST2, and high sensitivity C-reactive protein (hsCRP)] to discriminate for expanded HF outcomes.
Methods: The EXAMINE trial randomized 5,380 patients who were 15-90 days post ACS to alogliptin or placebo; mean follow-up was 18 months. At baseline, 5,154 patients had biomarker measurements. We evaluated the prognostic utility of a clinical model (table) and a biomarker based approach (dictomized by elevated versus not elevated) through changes in c-statistic. The primary HF outcome of interest was cardiovascular (CV) death, HF hospitalization (HHF), loop diuretic initiation, or elevated 6-month NTproBNP. Other outcomes included CV death, HHF, or loop diuretic initiation and CV death, HHF, or elevated 6-month NTproBNP.
Results: Patients who experienced the primary outcome (vs. those who did not) were older (63.2 vs 60.5 years of age), less likely to be male (58.8% vs. 69.5%), and had greater baseline HF history (41.1 vs. 25.5%). The clinical model modestly discriminated for expanded HF outcomes (table). When added to the clinical model, biomarkers increased the discrimination of CV death, HHF, loop diuretic initiation, or elevated 6-month NTproBNP (0.661 to 0.702; p<0.001); CV death, HHF, or loop diuretic initiation (0.661 to 0.717; p<0.001); and CV death, HHF, or elevated 6-month NTproBNP (0.727 to 0.764; p<0.001).
Conclusion: A clinical model modestly discriminates for expanded HF outcomes in patients with T2DM post ACS. When added to the clinical model, biomarkers significantly improved risk stratification and discrimination.