The tilt test is a widely used diagnostic tool for the investigation of syncope of suspected vasovagal origin. Some workers suggest that the sensitivity should be increased at the price of the loss of some specificity. Passive (non-sensitised) or isoproterenol (sensitised) protocols have not been widely studied on a large scale. The authors report the respective results of two protocols (passive and sensitised with isoproterenol proposed by Benditt et al.) applied to the same subject at 24 hours' interval in a random order in a series of 108 patients (age 56 +/- 19.3 years, 72 men) with unexplained syncope or malaise. After 30 minutes in decubitus, an 80 degrees tilt was maintained for 45 minutes for the passive test and for 25 minutes for the sensitised test, followed by a continuous infusion of isoproterenol in 10 minutes intervals of decubitus and inclination at successive doses of 1.3 and 5 micrograms/min. One or both tests were positive in 62 patients (57.4%). The overall concordance of the two tests (both positive or both negative) was 53.7%. In a subgroup of 60 subjects with clinically suggestive vasovagal attacks, the sensitivity of the passive test was 25% compared with 73.3% with the isoproterenol test (p < 0.001). Based on this difference of sensitivity between the two protocols, and knowing the good specificity of the isoproterenol test, the authors recommended the isoproterenol test as being of great practical value.