Accelerometer Metrics: Healthy Adult Reference Values, Associations with Cardiorespiratory Fitness, and Clinical Implications

Abstract

Purpose: Accelerometer-assessed physical activity (PA) can be summarized using cut-point-free or population-specific cut-point-based outcomes. We aimed to 1) examine the interrelationship between cut-point-free (intensity gradient (IG) and average acceleration (AvAcc)) and cut-point-based accelerometer metrics, 2) compare the association between cardiorespiratory fitness (CRF) and cut-point-free metrics to that with cut-point-based metrics in healthy adults aged 20 to 89 yr and patients with heart failure, and 3) provide age-, sex-, and CRF-related reference values for healthy adults.

Methods: In the COmPLETE study, 463 healthy adults and 67 patients with heart failure wore GENEActiv accelerometers on their nondominant wrist and underwent cardiopulmonary exercise testing. Cut-point-free (IG: distribution of intensity of activity across the day; AvAcc: proxy of volume of activity) and traditional (moderate-to-vigorous and vigorous activity) metrics were generated. The "interpretablePA" R-package was developed to translate findings into clinical practice.

Results: IG and AvAcc yield complementary information on PA with both IG ( P = 0.009) and AvAcc ( P < 0.001) independently associated with CRF in healthy individuals (adjusted R2 = 73.9%). Only IG was independently associated with CRF in patients with heart failure ( P = 0.043, adjusted R2 = 38.4%). The best cut-point-free and cut-point-based model had similar predictive value for CRF in both cohorts. We produced age- and sex-specific reference values and percentile curves for IG, AvAcc, moderate-to-vigorous PA, and vigorous PA for healthy adults.

Conclusions: IG and AvAcc are strongly associated with CRF and thus indirectly with the risk of noncommunicable diseases and mortality, in healthy adults and patients with heart failure. However, unlike cut-point-based metrics, IG and AvAcc are comparable across populations. Our reference values provide a healthy age- and sex-specific comparison that may enhance the translation and utility of cut-point-free metrics in clinical practice.

Trial registration: ClinicalTrials.gov NCT03986892.