Physical activity monitoring in COPD: Compliance and associations with clinical characteristics in a multicenter study

BACKGROUND: Little is known about COPD patients' compliance with physical activity monitoring and how activity relates to disease characteristics in a multi-center setting. METHODS: In a prospective study at three Northern European sites physical activity and clinical disease characteristics were measured in 134 COPD patients (GOLD-stage II-IV; BODE index 0-9) and 46 controls. Wearing time, steps per day, and the physical activity level (PAL) were measured by a multisensory armband over a period of 6 consecutive days (in total, 144h). A valid measurement period was defined as >/=22 h wearing time a day on at least 5 days. RESULTS: The median wearing time was 142 h:17 min (99%), 141 h:1min (98%), and 142 h:24 min (99%), respectively in the three centres. A valid measurement period was reached in 94%, 97%, and 94% of the patients and did not differ across sites (P = 0.53). The amount of physical activity did not differ across sites (mean steps per day, 4725 +/- 3212, P = 0.58; mean PAL, 1.45 +/- 0.20, P = 0.48). Multivariate linear regression analyses revealed significant associations of FEV(1), 6-min walk distance, quadriceps strength, fibrinogen, health status, and dyspnoea with both steps per day and PAL. Previously unrecognized correlates of activity were grade of fatigue, degree of emphysema, and exacerbation rate. CONCLUSIONS: The excellent compliance with wearing a physical activity monitor irrespective of study site and consistent associations with relevant disease characteristics support the use of activity monitoring as a valid outcome in multi-center studies.