To assess the reproducibility of head-up tilt-table testing 46 patients each underwent 2 isoproterenol tilt-table tests 1 to 6 weeks apart. Of 20 patients with an initially asymptomatic negative test result, 17 (85%) had a second negative test result. Of 20 patients whose initial test ended in syncope, 18 had a second test ending in syncope (n = 12) or presyncope (n = 6). Five of 6 patients whose first test ended in presyncope had a second test that ended in presyncope, and 1 had a second test that was asymptomatic. Finally, a total of 14 patients had at least 1 test (either the first or second) ending in presyncope, and 11 of these (79%) had another test ending in presyncope or syncope. The dose-dependence of isoproterenol was irreproducible. The reproducibility of the duration of head-up tilt necessary to elicit symptoms of presyncope or syncope was determined in the 23 patients who had these symptoms on both tests. Symptoms developed monoexponentially with duration of head-up tilt with half-times of 1.4 to 2.6 minutes, but these times did not correlate significantly between tests. In a selected subgroup of 12 patients who developed syncope in the first test and either presyncope or syncope in the second test during administration of isoproterenol (5 micrograms/min), the time of onset of presyncope correlated well (r = 0.73, p = 0.007) as did that of syncope (r = 0.86, p = 0.012). Finally, hemodynamic changes during symptoms were compared between the 2 tests.(ABSTRACT TRUNCATED AT 250 WORDS)