The relation of body fat distribution and body mass with haemoglobin A1c, blood pressure and blood lipids were examined in 874 men aged 40 to 59 not taking medication for diabetes mellitus and who worked for an urban company in Japan. Body fat distribution was measured by the waist hip circumference ratio. Body mass was estimated by Quetelet index. Haemoglobin A1c was measured from casual venous blood samples by high-pressure liquid chromatography. Measurement of haemoglobin A1c was validated by a 75 g oral glucose tolerance test conducted in a 7% sample. There was a dose-response relation between waist-hip ratio and haemoglobin A1c concentration while the relation between body mass index and haemoglobin A1c was not evident. Using linear regression to control for age, serum total cholesterol, usual alcohol consumption, cigarette smoking and body mass index, the positive association between waist-hip ratio and haemoglobin A1c remained significant (p = 0.02). This was not true for the positive association between body mass index and haemoglobin A1c (p = 0.32). Both waist-hip ratio and body mass index were positively associated with blood pressure and serum total cholesterol, and inversely associated with HDL-cholesterol. The associations of waist-hip ratio with blood pressure and blood lipids were significant after controlling for body mass index. Therefore, the waist-hip ratio is a correlate of both glucose abnormalities and known coronary risk factors in urban Japanese men even when body mass is controlled for.