To evaluate the day-to-day reproducibility of upright tilt-table testing, 109 patients with unexplained syncope prospectively underwent testing on 2 consecutive days using a uniform protocol. Results of testing on 2 separate days were concordant in 69 of 109 patients (63%), and discordant in 40 of 109 patients (37%). Thirty-six of 109 patients (33%) had vasodepressor syncope on 1 or both days of testing. Nineteen of 30 patients (63%) with vasodepressor responses on the first day did not reproduce this response during the second day of testing. An additional 6 patients with an initial negative tilt test had a vasodepressor response on the second day. Only 11 of 36 patients (31%) had reproducible vasodepressor responses on both days of testing. Patients with reproducible vasodepressor responses had a significantly higher mean number of preceding clinical syncopal events than patients with 2 normal tests (p < 0.02) or nonreproducible results (p < 0.04). In addition, these patients had a significantly longer duration of clinical symptoms relative to patients with 2 tests that yielded negative results (p < 0.008) and nonreproducible results (p < 0.01). The elapsed time between the most recent clinical event and the performance of tilt-table testing was not significantly different among the 3 groups, and did not appear to influence the outcome of testing. These data show that vasodepressor responses elicited by upright tilt-table testing show day-to-day variability in many patients, a finding that may limit the interpretation of initial and follow-up test results.