Simultaneous intra-radial and non-invasive (Finapres, Ohmeda) blood pressures were compared during prolonged head-up tilt, in eight patients (mean age 49 years) with malignant vasovagal syncope. Twelve tilts were performed, of which eight resulted in vasovagal syncope. The mean bias (difference between Finapres and intra-arterial pressures) for systolic pressure was +0.7 mmHg (standard deviation 11.3 mmHg) and for diastolic pressure was +5.4 mmHg (standard deviation 7 mmHg). The within-tilt precision (standard deviation of the bias) of the non-invasive measurements varied between 2.9-12.4 mmHg (median 4.5 mmHg) for systolic comparisons, and 1.6-8.4 mmHg (median 4.4 mmHg) for diastolic comparisons. In all but one tilt highly significant positive increases in both systolic (median 7.1 mmHg) and diastolic bias (median 8.1 mmHg) occurred on tilt with respect to resting pre-tilt levels. Independent of the absolute level of agreement, the non-invasive measurements followed changes in intra-arterial pressure closely, with 89% of beat-to-beat changes in systolic pressure, and 95% of beat-to-beat changes in diastolic pressure followed to within +/- 2 mmHg. This study suggests that the Finapres is well suited for use during diagnostic tilt testing, demonstrating an acceptable within-tilt precision and closely following pressure changes during vasovagal syncope.