Background: Carotid-femoral pulse wave velocity (PWV) is calculated on the basis of body surface distance (determined by either direct or indirect paths) and the corresponding transit time. The impact of distance vs time measurement on PWV reproducibility has not been quantified.

Methods: In 34 volunteers (age 47±19 years), carotid-femoral distance and transit time were measured twice by each of 2 trained observers, 2 hours apart, using a tape and a caliper. Two commonly used estimates of the travelled distance were calculated, namely "subtracted" (suprasternal notch to femoral artery minus suprasternal notch to carotid artery) and "direct" distance (carotid to femoral artery, multiplied by 0.8). Transit time was measured by high-fidelity tonometry (SphygmoCor, average of 3 readings for each of the 2 sessions). Variability was expressed as interobserver coefficient of variation (CV) and intra-class correlation (ICC).

Results: The CV was lowest for transit time (3.0%; interobserver difference ± SD, 0.0±1.8 ms), highest for subtracted distance (6.8%; 3.1±29 mm), and intermediate for direct distance (4.2%; 12.5±20 mm). The resulting interobserver differences in PWV were +0.0±0.2 m/s, +0.0±0.5 m/s, and +0.2±0.3 m/s, respectively. ICC was 0.98 for transit time (95% confidence interval [CI], 0.97–0.99), 0.73 for subtracted distance (95% CI, 0.53–0.86), and 0.81 for direct distance (95% CI, 0.66–0.90).

Conclusion. Interobserver variability of aortic PWV depends more on the measurement of body surface distance than on transit time. Estimates of the distance based on direct paths may generate a lower interobserver variability than those resulting from the combination of 2 paths ("subtracted" distance).