Head-up tilt test was introduced in clinical practice to assess vasovagal syncope and its use has further been extended to evaluate the efficacy of drug administration in these patients. Nevertheless, the effects of tilt test on vasovagal syncope have never been compared with those obtained by ethylephrine or propranolol administration. One hundred and sixty-nine consecutive patients with vasovagal syncope and positive baseline or nitrate-potentiated tilt test (60 degrees upright position for 45 min, or until syncope occurred; 5 mg sublingual isosorbide dinitrate administration if no symptoms occurred) were randomly distributed among three groups: Group A (57 control patients discharged without medical therapy); Group B (56 patients discharged with 75 mg/die ethylephrine); Group C (56 patients discharged with 80 mg/die propranolol). Tilt test was repeated after 1 month, while clinical outcome was evaluated monthly for a mean follow-up of 37.1 +/- 15.6 months. No significant differences in acute tilt-induced syncope recurrence rates were obtained among groups at test repetition since 70.2% of Group A, 69.6% of Group B and 62.5% of Group C experienced syncope. At 3-year follow-up 82.4% of Group A, 83.9% of Group B and 87.5% of Group C (NS among groups) remained symptom free, the most important clinical result being obtained in untreated patients. These data suggest that tilt test execution may prevent syncope recurrence as ethylephrine or propranolol administration. Irrespective of the therapeutical choice, the "controlled reproduction" of symptoms and some psychophysical training of patients to avoid precipitating circumstances, to recognize early symptoms promptly to be reverted by Trendelemburg position, may produce the same clinical improvement as (empiric) ethylephrine or propranolol therapy.