Impaired left ventricular diastolic function is not uncommon in patients with either diabetes mellitus or hypertension. This study was carried out to assess the contribution of left ventricular hypertrophy, high blood pressure, preclinical impaired glucose tolerance and diabetes mellitus to left ventricular diastolic function in patients attending a hypertension clinic. Echocardiography, 24-h ambulatory blood pressure monitoring and oral glucose tolerance tests were carried out in 152 consecutive hypertensive patients who had no evidence of ischaemic heart disease and were not known to be diabetic. From echocardiography, E/A (peak velocity of early/atrial filling waves of the transmitral flow) at rest and at peak standardized isometric exercise using handgrip, left ventricular mass index and deceleration time of the E wave were derived. Patients with impaired glucose tolerance and diabetes mellitus had lower E/A than the euglycaemic subjects both at rest (P=0.0073) and during isometric exercise (P<0.0001). E/A significantly reduced during isometric exercise in patients with impaired glucose tolerance and diabetes but not in euglycaemic patients. Deceleration time was shortened with a worsening degree of glucose intolerance in all the patients (P=0.0005), in those with left ventricular hypertrophy (P=0.0006) and in those without left ventricular hypertrophy (P=0.033). When adjusted for age, gender, race, body mass index, smoking history, ambulatory blood pressure findings, cholesterol and triglyceride levels and antihypertensive medications taken, E/A at isometric exercise was related to results of glucose tolerance tests and was inversely proportional to left ventricular mass index (P<0.0001). No significant differences were found whether patients were taking antihypertensive medications or not. In hypertensive patients, left ventricular diastolic function is determined by left ventricular mass index and the status of preclinical glucose intolerance, independent of age, gender, race, body mass index, blood pressure level, nocturnal drop in blood pressure or lipid level. These findings were not prejudiced by antihypertensive medications.