METHODS: Eighty patients with recurrent vasovagal syncope (VVS) and at least two falls during the last year were enrolled into the study. The first group consisted of 29 patients with positive HUT by Westminster protocol. The second group included 22 patients with negative HUT by Westminster protocol. The third group included 29 patients with positive HUT by Italian protocol. Fourteen healthy age- and gender-matched volunteers with no history of syncope comprised the control group. Spectral indices of HRV variability were analysed for three short-term intervals: last 5 min before HUT, first 5 min of HUT, and last 5 min before syncope or end of HUT. RESULTS: Patients with VVS showed significantly lower initial values of LF1 (n.u.) and higher HF1 (n.u.) comparing with healthy persons. During the 2-day period, Westminster-positive patients had the most considerable increase of LF2 (n.u.) and more prominent decrease of HF2 (n.u.); therefore, they had the highest level of LF2/HF2 ratio (P < 0.01). Before syncope development (the third period), LF3 (n.u.) reached the maximal level in all groups, but without significant difference between them. The HF3 (n.u.) level decreased in all patients, but only Westminster-positive patients showed a significant decrease in parasympathical tone. The LF3/HF3 ratio increased in all three groups without significant difference between them. CONCLUSION: In patients susceptible to VVS, development of syncopal attack begins during the first 5 min of the test, but it becomes clinically evident later on. There were no vegetative abnormalities even before syncope in patients with positive results by Italian protocol.