OBJECTIVE To review our experience with the upright tilt table test for the diagnosis of vasovagal syncope in a group of unselected patients with a history of syncope or presyncope. METHODS 179 patients with a history of syncope or presyncope were subjected to upright tilt test. After carotid sinus massage to exclude carotid sinus hypersensitivity, the patients were tilted on a motorised tilt table with footplate support to an angle of sixty to seventy degrees for thirty minutes. If syncope was not induced, isoprenaline was then infused for a further fifteen minutes. A positive response was defined as fulfilling at least two out of three criteria: (i) syncope or presyncope similar to the spontaneous episodes of syncope, (ii) relative slowing of the heart rate at the onset of symptoms, (iii) drop of systolic pressure to less than 90 mmHg or by more than 50 mmHg. METHODS Continuous variables are expressed as mean values +/- one standard deviation and analysed for statistical significance by the unpaired Student's t-test. Chi-squared test with continuity correction was used for dichotomous variables. RESULTS Ninety-four patients (53%) were positive for vasovagal syncope. Fourteen patients (8%) were positive at baseline tilt. An additional 80 patients (45%) were positive with the use of isoprenaline. Ten percent of the positive responses were purely cardioinhibitory, 10% purely vasodepressor and 80% mixed. The commonest cardiac rhythm during a positive response was junctional rhythm (46%) followed by sinus rhythm (44%). Sinus arrest with ventricular standstill occurred in only 5%. Accelerated idioventricular rhythm, 2:1 atrioventricular block and ventricular bigeminy accounted for the remaining 5%. CONCLUSIONS The upright tilt table test is useful for the diagnosis of vasovagal syncope.