Influence of Epoch Length and Recording Site on the Relationship Between Tri-Axial Accelerometry-Derived Physical Activity Levels and Structural, Functional, and Hemodynamic Properties of Central and Peripheral Arteries

Abstract

Background: It remains to be established to what extent physical activity (PA) levels among individuals are independently associated with deviations from the "optimal" state of the arterial system. Accelerometers have been proposed as means to obtain reliable, objective, and more comprehensive data of PA. Decisions at the time of data collection/processing could influence the association between accelerometry-derived indices and arterial properties.

Objectives: (i) To identify to what extent the strength of association between arterial properties and accelerometer-derived indices depend on the recording site and/or the epoch length; (ii) to determine whether some arterial characteristics (hemodynamic vs. structural vs. functional) or regions (elastic vs. transitional vs. muscular arteries; central vs. peripheral) have higher levels of association with accelerometry-derived indices.

Methods: Physical activity (PA), cardiovascular risk factors (CRFs), and cardiovascular properties were evaluated in 60 volunteers (general population; age: 23-62 years; women: 43%). PA was measured daily for 7 days (free-living situation; triaxial-accelerometers ActiGraph-GT3X+; hip and wrist; "Worn-to-wrist" option) and raw data was converted at epoch lengths of 1, 5, 10, 30, and 60-s. PA-related energy expenditure, daily time in moderate-to-vigorous PA, steps/minute, and counts-per-minute for vector magnitude were calculated. The cardiovascular evaluation included hemodynamic (central and peripheral pressure), structural (diameters and intima-media thickness), and functional (local and regional stiffness) parameters of carotids, femoral, and brachial arteries, and carotid-femoral and carotid-radial pathways. Arterial z-scores were obtained using age-related equations derived from healthy participants not exposed to CRFs (n = 1,688; age: 2-84 years; female: 51.2%) to evaluate at which degree each parameter deviates from the "optimal" value.

Methods: In general, hip recordings outperformed those obtained on the wrist regarding the strength of association with arterial parameters. Accelerometer-derived indices and their association with arterial properties vary depending on the recording site and epoch length. PA indices are stronger associated with functional (local) than structural variables and with central than peripheral arteries.

Conclusions: Regardless of the PA index, there were independent associations with central artery characteristics, which reinforces that these territories would be the most related to PA levels. Differences in data acquisition and processing could lead to differences in conclusions when addressing the association between accelerometer-derived indices and the cardiovascular system.

Keywords: accelerometry; arterial system; cardiovascular; epoch lengths; hip; physical activity; recording site; worn on wrist option (ActiLife software).

Figure 1

Summary of the non-invasive cardiovascular and accelerometry recordings approach performed on each participant. (A) Central aortic pressure measurements using SphygmoCor device and software (Applanation tonometry). GTF, general transfer function. (B) Regional arterial stiffness measurements using SphygmoCor device and software (Applanation tonometry). ΔxCR and ΔxCF: carotid-radial and carotid-femoral distances. Δt1 and Δt2: time between R-wave (QRS) and carotid and femoral “foot wave,” respectively. crPWV, cfPWV: carotid-radial and carotid-femoral pulse wave velocity. PWV, pulse wave velocity. (C) Arterial diameter, local stiffness, and intima-media thickness (IMT) measurements using echography and/or blood pressure measurement using echography and HemoDyn4M software. CCA, CFA, and BA, common carotid artery, common femoral artery, and brachial artery, respectively; EM, elastic modulus; β, beta or stiffness index; SBP and DBP, systolic and diastolic blood pressure; SysD and DD, peak systolic and end-diastolic arterial diameter; (D) Diagram of the placement site of the hip and wrist accelerometers (ActiGraph devices, model GT3X+, and ActiLife software).

Figure 2

Levels of energy expenditure, daily time in moderate-to-vigorous physical activity (MVPA%), steps per minute and vector magnitude (VM) counts per minute (CPM), found depending on the registration site (hip or wrist) and the epoch length: 1, 5, 10, 30, and 60 s. All comparisons between hip and wrist records were statistically significant (p < 0.05). For the same recording site (hip or wrist), the levels of Energy Expenditure and daily time in MVPA% were different (p < 0.05) when comparing the values obtained with different epoch lengths.

Figure 3

Relative differences (%), median levels, between Energy Expenditure, daily time in Moderate-to-Vigorous Physical Activity (MVPA, %), steps/min, and VM CPM found depending on the registration site (hip or wrist) and the epoch length: 1, 5, 10, 30, and 60 s. Levels of statistical significance (p-values) for each paired comparison are detailed in Supplementary Tables 15–18 in Supplementary Material 2.

Figure 4

Strength of association (r coefficients; partial correlations) between accelerometry-derived physical activity indices and cardiovascular properties (measured levels and/or z-scores), for hip and/or wrist records, as a function of epoch length (1, 5, 10, 30, and 60 s). Only the associations in which statistical significance was reached for at least one epoch are plotted. A more extensive description can be found in Supplementary Tables 25–28 in Supplementary Material 2.

Figure 5

Strength of association (r coefficients absolute value; partial correlations), ranked from highest to lowest “r,” between accelerometry-derived physical activity indices and cardiovascular properties (measured levels and/or z-scores). Separate charts are presented for wrist and hip, and 1- and 60-s epoch. Only the associations in which statistical significance was reached are plotted. Black and gray bars: central (e.g., common carotid artery, CCA) and peripheral [e.g., common femoral artery (CFA) or brachial artery (BA)] parameters, respectively. The letters above each bar indicate whether it corresponds to a functional (F), structural (S), or hemodynamic (H) parameter. A more extensive description can be found in Supplementary Tables 25–28 in Supplementary Material 2.