Head-up tilt testing demonstrates vasovagal mechanisms as a cause for syncope, but the pathophysiology underlying this condition remains unclear. The aim of this study was (i) to measure plasma beta-endorphins, adrenocorticotrophic hormone, cortisol, catecholamines, and brain natriuretic peptide during head-up tilt, and (ii) to assess the effect of naloxone infusion during head-up tilt in subjects with reproducible vasovagal syncope. During the assessment of unexplained syncope, 71 subjects underwent a total of 93 tilt tests (60-70 degrees head upwards for 40-45 min or until syncope occurred) during which frequent blood sampling was performed. Subjects with a positive tilt test (n = 56) (mean duration to syncope 23.6 min) showed a larger rise in beta-endorphin levels prior to syncope (baseline 4.7 +/- 2.2 vs syncope onset 6.9 +/- 3.2 pmol.l-1, P = 0.0001) than those with a negative test (n = 37) (baseline 3.9 +/- 3.9 vs end of test 4.9 +/- 2.3 pmol.l-1, P = 0.03). During tilting, adrenocorticotrophic hormone, cortisol, and noradrenaline increased; adrenaline and brain natriuretic peptide remained unchanged; and these responses were similar in positive and negative test groups. Naloxone (2.6 mg.kg-1 i.v. bolus followed by 20 micrograms.kg-1.min-1 infusion), administered in a double-blind fashion during head-up tilt in nine subjects, failed to modify either the time to syncope or the vasodepressor response. Thus, endogenous opioids appear not to be an important trigger for vasovagal syncope, and other pathophysiological mechanisms should be considered.