Prognostic and comparative performance of cardiovascular risk markers in patients with type 2 diabetes

Diagnostic tests including echocardiography, albuminuria, electrocardiogram (ECG), high-sensitivity troponin I (hs-TnI), and N-terminal prohormone brain natriuretic peptide (NT-proBNP) have been suggested as cardiovascular (CV) risk predictors in type 2 diabetes. We studied the separate and combined prognostic yield of these risk markers.

In all, 1030 patients with type 2 diabetes were recruited from specialized clinics in this prospective cohort study. Full echocardiographic evaluation was feasible in 886 patients in sinus rhythm with adequate image quality. ECG was performed in 998 patients. Albuminuria was measured in 1009 and NT-proBNP/hs-TnI in 933 patients. The end point was a composite of CV events.

The median follow-up was 4.7 years (interquartile range: 4.0-5.3), and 174 patients experienced a CV disease event. All considered markers, except hs-TnI, were significantly (P < .001) associated with the outcome: abnormal echocardiogram (hazard ratio 2.40 [1.70-3.39]), albuminuria 2.01 (1.47-2.76), abnormal ECG (2.27 [1.66-3.08]), high NT-proBNP (>150 pg/mL) 3.05 (2.11-4.40), and hs-TnI 1.12 (0.79-1.59). After adjusting for clinical variables, all remained significantly associated with the end point. However, after adjusting for each other, only NT-proBNP >150 pg/mL remained significantly associated with the end point (2.07 [1.28-3.34], P < .001). Measured by C-statistics, model performance was highest with log₂ (NT-proBNP) (0.70 [0.65-0.75]) and similar to clinical variables alone (0.71 [0.67-0.76]). Combining all risk markers only resulted in a very limited increase in C-statistics (0.69 [0.64-0.74]).

This study identified NT-proBNP over echocardiography, ECG, and albuminuria in risk prediction in patients with type 2 diabetes. The diagnostic yield in considering more than one risk marker was limited in this population.