Eighteen selected diabetic patients, without symptoms or signs of cardiac diseases and a control group of 15 normal subjects, age and sex matched, underwent polycardiographic and echocardiographic study, with the aim to evaluate their systolic and diastolic function of the left ventricle. PEP, LVET and PEP/LVET ratio were determined on the polycardiographic tracings. In eleven diabetic patients these data were also determined after amyl nitrite inhalation. Left ventricular diastolic dimension (DD), LV systolic dimension (DS), ejection fraction (EF), VCF, fractional shortening (FS), mitral opening delay (MOD) and the relaxation speed of the left ventricular posterior wall were determined on the echocardiographic tracings. All these data underwent statistical analysis by Student test, and some significant differences were found between diabetic patients and normal subjects. The mean value of PEP was augmented (diabetic patients: 100 +/- 18 msec; normal subjects: 88 +/- 9 msec; p 0,05), while the mean value of LVET was diminished (diabetic patients: 275 +/- 22 msec; normal subjects: 300 +/- 23 msec; p less than 0,005); the PEP/LVET ratio was, therefore, increased in diabetic patients (diabetic patients: 0,37 +/- 0,08; normal subjects: 0,29 +/- 0,04; p less than 0,005). Of all echocardiographic data only MOD (diabetic patients: 54 +/- 31 msec; normal subjects: 14 +/- 21 msec; p less than 0,001) was significantly increased in diabetic patients. The feasibility and reliability of polycardiographic and echocardiographic techniques in detecting early myocardial disfunction is discussed and the accurate selection of patients in order to exclude other myocardiopathies in pointed out. It is then concluded that diabetic patients, in comparison with normal subjects, show an impairment of both systolic and diastolic function of the left ventricle, exclusively depending on diabetes.