Reproducibility and agreement of pulse wave velocity and augmentation index over repeated assessments using two different devices in adolescents

TL;DR

This study compared the reliability of two devices for measuring pulse wave velocity and augmentation index in adolescents under different conditions.

Contribution

The study provides new insights into the reproducibility and agreement of pulse wave measurements in adolescents using two devices.

Findings

PulsePen showed better reproducibility for pulse wave velocity than Arteriograph.

Both devices had limited reliability for augmentation index measurements.

Device agreement was poor, suggesting results should not be directly compared.

Abstract

We investigated the agreement between pulse wave velocity (PWV) and augmentation index (AIx%) obtained in a controlled fasted condition versus a non‐fasted, uncontrolled, “real‐world” condition. Thereafter, we assessed the reproducibility of PWV and AIx% over three repeated visits under controlled fasted conditions.

PWV and AIx% were assessed in one uncontrolled visit, and three controlled fasted visits after a 10‐min supine rest in 28 adolescents (61% girls) aged 12–14 years.

Intraclass correlation coefficients (ICCs) between controlled visits were 0.5–0.6 (PWV) and −0.3–0.6 (AIx%) for PulsePen, and 0.5–0.6 (PWV) and 0.4–0.7 (AIx%) for Arteriograph. ICCs between a single uncontrolled visit and controlled visits were lower, especially for the Arteriograph (PWV: 0.0–0.1; AIx%: −0.4–0.0). Device agreement was poor (PWV ICC: −0.2–0.1; AIx% ICC: −0.5–0.3). Reproducibility using PulsePen was (PWV: 0.2–0.8; AIx%: −0.3–0.6) and Arteriograph (PWV: 0.0–0.1; AIx%: −0.4–0.0).

The PulsePen had better reproducibility of PWV than the Arteriograph. Both devices had limited reliability and high variability in AIx% measurements. PWV from the PulsePen may be applicable in healthcare, while results from the Arteriograph should be interpreted with caution. Because of the poor agreement between the devices, the data from these devices should not be directly compared. These results may not be generalizable based on the modest sample size. In addition, the variability in real‐world situations should not be overlooked.

Figures