Haemoglobin A1c even within non-diabetic level is a predictor of cardiovascular disease in a general Japanese population: the Hisayama Study
1 Department of Environmental Medicine, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka City 812-8582, Japan
2 Department of Medicine and Clinical Science, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan
Cardiovascular Diabetology 2013, 12:164 doi:10.1186/1475-2840-12-164Published: 7 November 2013
There is little information about predictive ability of haemoglobin A1c (HbA1c) for cardiovascular disease (CVD) in Asians. To investigate the discriminatory ability of HbA1c to identify subjects who are at greater risk of developing CVD in a prospective study of a defined community-dwelling Japanese population.
A total of 2,851 subjects aged 40–79 years were stratified into five groups (HbA1c levels with ≤ 5.0, 5.1–5.4, 5.5–6.4, and ≥ 6.5% and a group with antidiabetic medication) and followed up prospectively for 7 years (2002–2009).
During the follow-up, 119 subjects developed CVD. The multivariable-adjusted risk of CVD was significantly increased in subjects with HbA1c levels of 5.5–6.4 and ≥ 6.5% and diabetic medication compared to HbA1c level with ≤ 5.0% (hazard ratio, 2.26 [95% confidence interval, 1.29–3.95] for the 5.5–6.4%; 4.43 [2.09–9.37] for the ≥ 6.5%; and 5.15 [2.65–10.0] for the antidiabetic medication group). With regard to CVD subtype, the positive associations between HbA1c levels and the risk of coronary heart disease (CHD) and ischaemic stroke were also significant, but no such associations were seen for haemorrhagic stroke. The C statistic for developing CVD was significantly increased by adding HbA1c values to the model including other risk factors (0.789 vs. 0762, p = 0.006), and the net reclassification improvement was 0.105 (p = 0.004).
Our findings suggest that elevated HbA1c levels are an independent risk factor for CVD, especially CHD and ischaemic stroke, and that the addition of HbA1c to the model with traditional risk factors significantly improves the predictive ability of CVD.