21 patients accusing previous syncopal attacks, without myocardial infarction, preexcitation syndrome or severe cardiac failure were submitted to a 24 hrs Holter monitoring with 10 min. of esophageal electrocardiogram. Nine patients were considered as cases of sinus bradycardia (mean HR = 49.5 +/- 4.4 b/min). Other 12 patients with normal basic sinus rhythm were admitted for ventricular premature beats (HR = 65.5 +/- 8.7 b/min.). Fourier transformation analysis (mean method, linear interpolation) was used for determination of the heart rate variability (HRV) and other two variables: variation range (VR)--(longest-shortest cycle/mean basic sinus cycle, %) and differences of sinus cycle length (DSCL = maximal difference of any two consecutive sinus cycle, ms). In all the 9 patients with low sinus rhythm and in 7 patients with normal basic sinus rhythm, at least one episode of critical bradycardia (under 45 b/m) was observed. In 3 of the 9 patients with sinus bradycardia at least two episodes of nonsustained ventricular tachycardia were observed. In all the 12 patients with ventricular premature beats, many episodes of nonsustained ventricular tachycardia were registered. In all the 21 studied patients low and very low frequency HRV was observed. The VR and DSCL presented a significant dependence on age, but this relationship was weaker than the established limits of normality. In the cases of sick sinus syndrome with sinus bradycardia and/or severe ventricular tachyarrhythmias, the noninvasive Holter analysis of HRV can determine the correct diagnosis and indications for permanent cardiac pacing and antiarrhythmic drug therapy.