The effects of experimental, meteorological, and physiological factors on short-term repeated pulse wave velocity measurements, and measurement difficulties: A randomized crossover study with two devices

Background: Large longitudinal studies with repeated pulse wave velocity (PWV) measurements, a direct measure of arterial stiffness, are required to realize the full potential of arterial stiffness in clinical practice. To facilitate such studies it is important to increase the power of a study by reducing within-subject variability of PWV, and to ease the use of a PWV device in clinical settings by minimizing PWV measurement difficulties. Methods: We systematically investigated experimental setting and meteorological conditions, as well as physiological factors and participant characteristics, to determine whether and to what extent they affected: between- and within-subjects variability of PWV recordings, and measurement difficulties of a particular device. We conducted a 2-week longitudinal block-randomized cross-over study with two blinded observers and two commonly used devices: applanation tonometry SphygmoCor CvMS and oscillometric Arteriograph to assess carotid-femoral (cfPWV) or aortic (PWVao) PWV, respectively. Our sample had uniform and wide-spread distribution of age, blood pressures, hypertensive status and BMI. Each participant (N = 35) was recorded 12 times over 3 visiting days, 7 days apart. On each day, recordings were made twice in the morning (7-10 a.m.) and afternoon (16-18 p.m.). Data were analyzed using multilevel mixed-effects models, separately for each device. Results: In addition to age and mean arterial pressure (MAP) that strongly affected both cfPWV and PWVao, other significant factors appeared to indicate a measurement approach. cfPWV as a more direct measure of arterial stiffness was additionally affected by hypertension status, outdoor temperature, interaction of MAP with outdoor temperature and the order of visit, with MAP within-subject variability contributing on average 0.27 m/s to difference in repeated measurements at 5 degrees C and 0.004 m/s at 25 degrees C. PWVao measurements derived at a single brachial site were more dependent on age than cfPWV and also depended on personal characteristics such as height and sex, and heart rate; with within-subject MAP variability adding on average 0.23 m/s to the difference in repeated measures. We also found that female sex significantly increased, and recording in afternoon vs. morning significantly decreased measurement difficulties of both devices. Conclusion: We identified factors affecting PWV recordings and measurement-difficulties and propose how to improve PWV measuring protocols.