THE VARIABILITY OF REPEATED MEASUREMENTS OF ARTERIAL STIFFNESS BIOMARKERS: A SHORT-TERM LONGITUDINAL STUDY IN A WIDE RANGE OF NORMOTENSIVE AND HYPERTENSIVE PARTICIPANTS

Objective: Arterial stiffness (AS) biomarkers, such as pulse wave velocity and augmentation indices, are predictive of future cardiovascular events in subjects free of overt cardiovascular disease and are regarded as indicators of atherosclerosis and subclinical organ damage. As the surrogate endpoints of future cardiovascular events, they are emerging as an appealing therapeutic target. However, the variability of intrasubject measurements might hinder their use in clinical practice. Design and method: In this study, we estimate the reproducibility of intrasubject measurements of AS biomarkers: pulse wave velocity and central augmentation index, measured by oscillometry (Arteriograph) and applanation tonometry (SphygmoCor) devices. Thirty-six participants evenly distributed across different genders, BMI categories, ages, and hypertensive status were enrolled in the study. Participants visited the laboratory for three days, one week apart. On each visiting day, they came in the morning and afternoon (6 visits in total). All measurements were performed in accordance with the ARTERY society recommendations and in double by two operators who received the necessary training and performed measurements in a randomized order. Linear mixed models were used to analyze the nested data grouped by a participant and a visiting day. Results: Based on the interim analysis of 451 recordings that passed the quality control, we found that a device type and BMI category significantly affected pulse wave velocity in intrasubject measurements. As for the central augmentation index, it was a device and a time of the day (morning/afternoon) that affected its variability within a participant. An operator or interaction between an operator and a device did not affect measurements of these biomarkers. Conclusions: Since expected intersubject PWV variability due to different devices or BMI allows for differences to extend 10% of the observed population-based range, the identification of weak/moderately effective agents used in de-stiffening therapies might be hampered. For cAIx, this variability was somewhat less pronounced.