Since the pharmacological challenge with nitroglycerin (NTG) follows the initial drug-free phase in current tilt testing protocols, the effects of nitroglycerin alone and the appropriate duration of the basal phase are unknown.To optimize the test, a randomized intra-patient comparison of two protocols was undertaken: a conventional nitroglycerin test (cHUT) consisting of passive upright posture at 60 degrees for 45 min followed, if negative, by sublingual NTG 0.4 mg spray, with the test continued for 20 min; and, accelerated nitroglycerin test (aHUT) consisting of passive upright posture at 60 degrees for 5 min--to rule out orthostatic hypotension--followed by sublingual NTG 0.4 mg spray, with the test continued for 20 min. Eighty-four consecutive patients (33 males; mean age 55+/-22) with unexplained syncope underwent both cHUT and aHUT in a randomized sequence with a 24-72 h interval between them. Additionally, 25 age-matched control subjects underwent aHUT.In the drug-free phase, cHUT was positive in 15/84 patients (18%) and aHUT in 1/84 patients (1%). After NTG, cHUT and aHUT showed the same positivity rate of 33% (28/84 patients). The overall positivity rate was therefore higher with cHUT than with aHUT (51% vs 35%, P=0.04). Times to syncope were 29+/-12 min, (range 2-44) for cHUT drug-free phase, 5+/-2 min (range 2-9) for cHUT NTG phase, and 5+/-2 min (range 2-9) for aHUT. Only one (4%) of the control subjects had a positive response to aHUT.The contribution of NTG to the diagnosis is independent of the presence of an unmedicated phase. The appropriate duration of the NTG phase is 10 min. aHUT has good specificity, but a positivity rate lower than cHUT; thus a drug-free phase is necessary to increase the sensitivity of the test.