Biomarker-based prediction of fatal and non-fatal cardiovascular outcomes in individuals of the general population with and without diabetes mellitus

Cardiovascular biomarkers may reflect different aspects of cardiovascular disease, including myocardial tissue damage (high-sensitive cardiac troponin [hs-cTn]), hemodynamic stress (N-terminal prohormone of brain natriuretic peptide [NT-proBNP)), or inflammation (high-sensitivity C-reactive protein [hs-CRP]).

To determine the risk for fatal and non-fatal cardiovascular events in patients with diabetes mellitus (DM), a high-risk group for cardiovascular complications, after accounting for these biomarkers and to determine the risk associated with these biomarkers.

Harmonized data of population-based studies from the Biomarkers for Cardiovascular Risk Assessment in Europe (BiomaCaRE) and MOnica Risk, Genetics, Archiving and Monograph (MORGAM) consortia were used to calculate hazard ratios (HRs, 95% confidence intervals [CI] per standard deviation) for these biomarkers adjusted for diabetes, patient characteristics and biomarkers for their association with the primary endpoint of fatal and non-fatal cardiovascular events during a median follow-up of 9.6 years (maximum 28 years). Additionally, a years-of-life-lost analysis was conducted stratified by prevalent diabetes and specific biomarker cut-offs known to be associated with increased risk for events (hs-cTnI >5 ng/L, NTproBNP >125 ng/L, hs-CRP >5mg/L).

We included 95,302 individuals, of whom 6,501 had DM (6.8%). Cox-regression analysis revealed DM to be independently associated with the primary endpoint (2.1 [95% CI 1.9, 2.3], p<0.001) despite adjustment for clinical characteristics and biomarkers. Also, all three biomarkers were independent predictors themselves: log-transformed NT-proBNP 1.3 [95% CI 1.3, 1.4] p<0.001; log-transformed hs-CRP 1.2 [95% CI 1.1, 1.2] p<0.001; third-root-transformed hs-cTnI 1.1 [95% CI 1.0, 1.1] p=0.0038). The sex-, age- and cohort-adjusted HR for the primary endpoint according to absolute biomarker concentrations derived by cox-regression models using cubic splines is provided for the three biomarkers in Figure 1. Upon dichotomization of biomarkers, individuals with diabetes and at least one elevated biomarker lost a median of 15.5 healthy years because of the primary endpoint (Kaplan-Meier plot in Figure 2, with age on the x-axis).

Our findings confirm that diabetes confers a residual cardiovascular risk beyond adjustment for clinical characteristics and cardiovascular biomarker. Furthermore, biomarkers may aid in the identification of patients at highest risk, which should be considered in future models of risk prediction.

Type of funding sources: None.