Increased plasma glucose concentration is a predictive factor for mortality in both diabetic and non-diabetic subjects. Although glycated haemoglobin (HbA 1c ) is a useful index of mean blood glucose concentrations over the preceding 1 to 3 months, there are few data regarding its relationship to cardiovascular risk. We have examined the relationship between HbA 1c and cardiovascular risk factors in 1280 subjects with normal glucose tolerance. Based on HbA 1c tertiles (tertile 1: n = 427, 262 men and 165 women, HbA 1c level: 2.9-4.7 % in men and 3.2-4.2 % in women; tertile 2: n = 426, 261 men and 165 women, HbA 1c level: 4.7-5.1 % in men and 4.2-4.6 % in women; tertile 3: n = 427, 262 men and 165 women, HbA 1c level: 5.1-6.7 % in men and 4.6-6.9 % in women), increasing HbA 1c was associated with increasing age, blood pressure, waist-hip ratio, fasting and 2-h plasma glucose, 2-h insulin, cholesterol, low-density lipoprotein cholesterol, apolipoprotein B and urate concentrations. When age and sex were included as covariates, increasing HbA 1c remained associated with increasing fasting and 2-h plasma glucose, 2-h insulin, total cholesterol, and low-density lipoprotein cholesterol concentrations. These findings emphasize the importance of hyperglycaemia, as reflected by HbA 1c , as a continuum in the evaluation of cardiovascular risk. Furthermore, these findings support the hypothesis that cardiovascular disease risk commences with rising glucose concentrations before 'conventionally-defined' glucose intolerance occurs.