Chronic pain predicts reduced physical activity in a large population cohort study

Background/Purpose:
Chronic musculoskeletal pain (CP) is associated with reduced levels of physical activity (PA), however few studies have
examined the prospective nature of CP on PA at the population level. The aim of this study is to examine whether CP independently predicts a reduction inPA.
Methods:
Data are from the UK Biobank, a large prospective cohort
designed to support investigation of risk factors for major diseases of middle and old age. More than 500,000 men and women aged 40–69 were recruited between 2006–2010 (baseline); approximately 20,000 repeat assessments
were conducted between 2012–2013 (follow-up). Participants answered questions on health and lifestyle by touch-screen questionnaire. CP is defined as self-reported pain lasting at least three months either all over the body or related to the neck or shoulder, back, abdomen, hip or knee, as well as facial
pain or headache. Questions on PA and sedentary behaviour were asked using a modified short-form of the International Physical Activity Questionnaire at baseline and follow-up. PA was categorized as low, moderate or high according to frequency/duration of moderate and vigorous activity; PA
change was calculated and dichotomized into same or improved activity category or worsened activity category. Hours spent watching TV and in non-work computer use were combined to estimate time in sedentary behaviour. Descriptive, linear and logistic regression analyses were conducted on 14,391 men and women with complete data on a priori covariates (age, gender, ethnicity, highest educational attainment, work status at follow-up, change in BMI status, baseline self-rated health, self-reported depression at follow-up and time-to-follow-up).
Results:
Mean length of follow-up was 4 years (range 2.1–6.5). 5,567
participants (39%) had CP at baseline. Of those without CP at baseline, 20% (n=1,793) had developed it at follow-up. At baseline, fewer individuals with CP were categorized with moderate or high levels of PA (chi-squared=12.5, p=0.002). Overall, the majority of participants maintained the same PA
category, while 21% had an improved activity classification and 20% had a worsened activity classification. Individuals with CP at baseline were more likely to deteriorate in activity status, than those without, who were more likely to stay the same or improve activity status (chi-squared=9.2, p=0.002). This relationship remained in multivariate logistic regression after adjustment for
covariates, with baseline CP predicting a negative change in PA category (OR: 1.13; 95% CI: 1.04–1.23). Individuals with CP spent more time in sedentary behaviour than those without CP (B=0.07, 95% CI: 0.01–0.13) in linear regression analysis adjusted for covariates and baseline sedentary behaviour.
Conclusion:
CP predicts a negative change in PA over an average four-year follow-up period. Given the prevalence and persistence of CP, attention should be given to supporting and encouraging physical activity so as to prevent deterioration of PA levels and the spiral of pain-related inactivity within middle-age and early old age adults with CP.