Change in HbA1c over 3 years does not improve the prediction of cardiovascular disease over and above HbA1c measured at a single time point

Abstract

Aims/hypothesisHbA1c is an important risk factor for cardiovascular disease (CVD), with 1% higher HbA1c levels associated with a 10–20% increased risk of CVD. Little is known about the association between change in HbA1c over time and cardiovascular risk in non-diabetic populations. This study examined the association between change in HbA1c over time and cardiovascular risk in a non-diabetic British population.MethodsWe used data on HbA1c collected at baseline and at a second health examination 3 years later among a population of 5,790 non-diabetic men and women who participated in the European Prospective Investigation of Cancer (EPIC)–Norfolk. The association between change in HbA1c over 3 years and incident cardiovascular events over the following 8 years was examined using multivariate Cox regression. We also examined whether information on change in HbA1c over time improved prediction of cardiovascular events over a single measure of HbA1c by comparing the area under the receiver operating characteristic curves (aROC) and computing the net reclassification improvement.ResultsThe mean change (SD) in HbA1c over 3 years was 0.13% (0.52). During 44,596 person-years of follow-up, 529 cardiovascular events occurred (incidence 11.9 per 1,000 person-years). Each 0.5% rise in HbA1c over 3 years was associated with a 9% increase in risk of a cardiovascular event (HR 1.09; 95% CI 1.01, 1.18) after adjustment for baseline HbA1c and other major cardiovascular risk factors. However, change in HbA1c was not associated with cardiovascular risk after adjustment for HbA1c at follow-up. Multivariate models with and without information on change in HbA1c over time showed a similar aROC of 0.78. Adding change in HbA1c to the model with HbA1c at follow-up did not improve risk classification.Conclusions/interpretationAddition of information on change in HbA1c over 3 years did not improve the prediction of CVD over and above information on HbA1c and other major cardiovascular risk factors from a single time point.Electronic supplementary materialThe online version of this article (doi:10.1007/s00125-013-2854-8) contains peer-reviewed but unedited supplementary material, which is available to authorised users.