Eur J Clin Invest 2011; 41 (12): 1292–1298
Background Patients with diabetes mellitus have a substantially increased risk of developing cardiovascular disease. However, the absolute risk greatly varies not only among patients, but the risk profile for an individual patient may also change over time. We investigated the prognostic role of repetitive measurements of Glycated haemoglobin A1c (HbA1c) and N-terminal pro-B-type natriuretic peptide (NT-proBNP) in patients with longstanding diabetes.
Materials and methods For this prospective, observational study data from 544 consecutive patients were collected between 2005 and 2008. HbA1c and NT-proBNP were measured at baseline and after 1 year. The median observation period was 40 months. Endpoints were all-cause mortality, cardiac, cardiovascular and all-cause hospitalizations.
Results N-terminal pro-B-type natriuretic peptide concentrations significantly increased from 230 ± 385 to 280 ± 449 pg mL−1 (P < 0·001); during the same time, HbA1c significantly decreased from 7·6 ± 1·5 to 7·3 ± 1·2 (P < 0·001). NT-proBNP was the best baseline predictor in a Cox regression model consisting of NT-proBNP, HbA1c, age, gender and duration of diabetes for all endpoints (P < 0·001). NT-proBNP at follow-up was the best predictor for the remaining period (P < 0·001, all endpoints). HbA1c at baseline and follow-up was predictive for all-cause hospitalizations (P = 0·005 both). In a third model that investigated the plasticity of both markers, changes in HbA1c concentration had no predictive value, but a change of NT-proBNP concentration was highly predictive (P = 0·025 all-cause mortality, P < 0·001 all other endpoints).
Conclusions N-terminal pro-B-type natriuretic peptide and HbA1c concentrations significantly diverged over a 1-year period. NT-proBNP was the most potent predictor of outcome at baseline and follow-up, and changes in NT-proBNP concentrations were linked to an altered risk profile, unlike changes in HbA1c levels.