Arterial stiffness, estimated by pulse wave velocity (PWV), is an independent predictor of cardiovascular mortality and morbidity. However, the clinical applicability of these measurements and the elaboration of reference PWV values are difficult due to differences between the various devices used. In a population of 50 subjects aged 20-84 years, we compared PWV measurements with three frequently used devices: the Complior and the PulsePen, both of which determine aortic PWV as the delay between carotid and femoral pressure wave and the PulseTrace, which estimates the Stiffness Index (SI) by analyzing photoplethysmographic waves acquired on the fingertip. PWV was measured twice by each device. Coefficient of variation of PWV was 12.3, 12.4 and 14.5% for PulsePen, Complior and PulseTrace, respectively. These measurements were compared with the reference method, that is, a simultaneous acquisition of pressure waves using two tonometers. High correlation coefficients with the reference method were observed for PulsePen (r = 0.99) and Complior (r = 0.83), whereas for PulseTrace correlation with the reference method was much lower (r = 0.55). Upon Bland-Altman analysis, mean differences of values +/- 2s.d. versus the reference method were -0.15 +/- 0.62 m/s, 2.09 +/- 2.68 m/s and -1.12 +/- 4.92 m/s, for PulsePen, Complior and Pulse-Trace, respectively. This study confirms the reliability of Complior and PulsePen devices in estimating PWV, while the SI determined by the PulseTrace device was found to be inappropriate as a surrogate of PWV. The present results indicate the urgent need for evaluation and comparison of the different devices to standardize PWV measurements and establish reference values.