Patterns of daytime physical activity in patients with chronic fatigue syndrome.

Objectives: To classify patients with chronic fatigue syndrome (CFS) by pattern of physical activity and determine the clinical associations of each type.

Methods: 579 out of 641 participants with CFS from the PACE (Pacing, graded Activity, Cognitive behavioural therapy: a randomised Evaluation) trial wore an Actiwatch (accelerometer) for between 3 and 7 days before any trial treatments, which provided a measure of physical activity. Participants' activity was categorised into one of four patterns (pervasively inactive, pervasively active, boom and bust, or indeterminate) primarily using a priori definitions of activity. Clinical associations were sought with each group using an exploratory logistic regression with the indeterminate activity group being the reference group.

Results: 124 (21%) of the participants were classified as pervasively inactive, 65 (11%) as pervasively active, 172 (30%) showed a 'boom and bust' pattern of activity, and 218 (38%) had an indeterminate pattern. Pervasively inactive patients were more physically disabled, those in the pervasively active group were more anxious, and those in the boom and bust group had more sleep disturbance.

Conclusion: We were able to classify patients with CFS into groups by their daytime activity pattern. The different patterns of activity were associated with important clinical variables, suggesting that they might be helpful in determining prognosis and targeting treatments. These associations need replication.