Prognostic significance of metabolomic biomarkers in patients with diabetes mellitus and coronary artery disease

Diabetes mellitus (DM) and coronary artery disease (CAD) constitute inter-related clinical entities. Biomarker profiling emerges as a promising tool for the early diagnosis and risk stratification of either DM or CAD. However, studies assessing the predictive capacity of novel metabolomics biomarkers in coexistent CAD and DM are scarce. This post-hoc analysis of the CorLipid trial (NCT04580173) included 316 patients with CAD and comorbid DM who underwent emergency or elective coronary angiography due to acute or chronic coronary syndrome. Cox regression analyses were performed to identify metabolomic predictors of the primary outcome, which was defined as the composite of major adverse cardiovascular or cerebrovascular events (MACCE: cardiovascular death, myocardial infarction, stroke, major bleeding), repeat unplanned revascularizations and cardiovascular hospitalizations. Linear regression analyses were also performed to detect significant predictors of CAD complexity, as assessed by the SYNTAX score. After a median 2-year follow up period (IQR = 0.7 years), the primary outcome occurred in 69 (21.8%) of patients. Acylcarnitine ratio C4/C18:2, apolipoprotein (apo) B, history of heart failure (HF), age > 65 years and presence of acute coronary syndrome were independent predictors of the primary outcome in diabetic patients with CAD (aHR = 1.89 [1.09, 3.29]; 1.02 [1.01, 1.04]; 1.28 [1.01, 1.41]; 1.04 [1.01, 1.05]; and 1.12 [1.05–1.21], respectively). Higher levels of ceramide ratio C24:1/C24:0, acylcarnitine ratio C4/C18:2, age > 65 and peripheral artery disease were independent predictors of higher CAD complexity (adjusted β = 7.36 [5.74, 20.47]; 3.02 [0.09 to 6.06]; 3.02 [0.09, 6.06], respectively), while higher levels of apoA1 were independent predictors of lower complexity (adjusted β= − 0.65 [− 1.31, − 0.02]). In patients with comorbid DM and CAD, novel metabolomic biomarkers and metabolomics-based prediction models could be recruited to predict clinical outcomes and assess the complexity of CAD, thereby enabling the integration of personalized medicine into routine clinical practice. These associations should be interpreted taking into account the observational nature of this study, and thus, larger trials are needed to confirm its results and validate them in different and larger diabetic populations.