Actigraphy can provide useful patient-centered outcomes for quantification of physical activity in the “real-world” setting.

To characterize the relationship of actigraphy outputs with “in-laboratory” measures of cardiopulmonary function and respiratory symptoms in pre-COPD, we obtained actigraphy data for 8 h/day for 5 consecutive days a week before in-laboratory administration of respiratory questionnaires, PFT, and CPET to a subgroup of subjects participating in the larger study of the health effects of exposure to secondhand tobacco smoke who had air trapping but no spirometric obstruction (pre-COPD). Using machine learning approaches, we identified the most relevant actigraphy predictors and examined their associations with symptoms, lung function, and exercise outcomes.

Sixty-one subjects (age = 66±7 years; BMI = 24±3 kg/m²; FEV₁/FVC = 0.75 ± 0.05; FEV₁ = 103 ± 17 %predicted) completed the nested study. In the hierarchical cluster analysis, the activity, distance, and energy domains of actigraphy, including moderate to vigorous physical activity, were closely correlated with each other, but were only loosely associated with spirometric and peak exercise measures of oxygen consumption, ventilation, oxygen-pulse, and anaerobic threshold (VO_2AT), and were divergent from symptom measures. Conversely, the sedentary domain clustered with respiratory symptoms, air trapping, airflow indices, and ventilatory efficiency. In Regression modeling, sedentary domain was inversely associated with baseline lung volumes and tidal breathing at peak exercise, while the activity domains were associated with VO_2AT. Respiratory symptoms and PFT data were not associated with actigraphy outcomes.

Outpatient actigraphy can provide information for “real-world” patient-centered outcomes that are not captured by standardized respiratory questionnaires, lung function, or exercise testing. Actigraphy activity and sedentary domains inform of distinct outcomes.