To assess the variability of head-up tilt table testing, 35 patients with vasovagal syncope, shown by a positive tilt table test, underwent a second test 1 week later. Also, on the day before each test, spectral and time-domain indexes of heart rate variability were derived from Holter recordings to examine the stability of autonomous nervous system activity in these patients. Fifteen healthy volunteers served as a control group and also underwent two tilt table tests with prior Holter monitoring. Twenty-one (60%) of the 35 patients had a second positive test. None of the patients in the control group experienced syncope during either of the tests. The heart rate variability measures in the control group varied slightly from 1 day to the other, in contrast to the syncopal patients, where only low frequency spectral power and the mean of all 5-minute standard deviations of RR intervals showed comparable behavior. The indexes which reflect parasympathetic activity exhibited significant fluctuations in the syncopal patients. These fluctuations were due entirely to the patients who did not reproduce the outcome of the tilt table test, where high parasympathetic tone was associated with the positive test and normal parasympathetic tone with the negative test. In contrast, the patients with two positive tests had high parasympathetic tone during both test periods, with low individual variability. In conclusion, patients with vasovagal syncope show variations in vagal autonomic tone and appear to be more prone to syncope when their parasympathetic tone is elevated.