Hour-to-hour and week-to-week variability and reproducibility of wave reflection indices derived by aortic pulse wave analysis: implications for studies with repeated measurements.

BACKGROUND: Wave reflections are implicated increasingly in clinical research. AIMS: The purpose of the present study was to investigate whether wave reflection indices are reproducible when measured repeatedly (more than twice) at longer time intervals, namely hour-to-hour and week-to-week, in healthy subjects; something that has not yet been examined.
METHODS: Bland-Altman plots, the interclass correlation coefficients (ICC) and coefficient of variation were used for this purpose. Two series, with measurements repeated in triplicate, were performed in 22 healthy subjects: the first at intervals of 1 h and the second at 1-week time intervals. Augmentation index (AIx), heart rate-corrected AIx (AI@75) and arrival time of reflected waves at the central aorta (tr) were calculated by aortic pulse wave analysis.
RESULTS: AIx and AI@75 presented very good to excellent reproducibility (ICC = 0.86) for hour-to-hour repeated measurements, while tr was also highly reproducible (ICC = 0.79). AIx, AI@75 and tr were substantially reproducible when measured repeatedly with 1-week intervals, providing ICCs greater than 0.70. Bland-Altman plots confirmed these results, indicating that more than 90% of AIx, AI@75 and tr measurements fell within two standard deviations of the mean difference.
CONCLUSIONS: Wave reflections are substantially reproducible even when measurements repeated in triplicate are performed at longer time intervals (hours and weeks). A quantifiable amount of variation was reported, which should be taken carefully into consideration in interventional studies with repeated measurements and in observational studies investigating differences or correlations of these indices.