Predictors of New-Onset Widespread Pain in Older Adults: Results From a Population-Based Prospective Cohort Study in the UK


         McBeth, J., Lacey, R. J. and Wilkie, R. (2014), Predictors of New‐Onset Widespread Pain in Older Adults: Results From a Population‐Based Prospective Cohort Study in the UK. Arthritis & Rheumatology, 66: 757–767. doi:10.1002/art.38284


In older adults, widespread pain (WP) is common, although its etiology is unclear. This study sought to identify factors associated with an increased risk of developing WP in adults age ≥50 years.


A population-based prospective study was conducted. A baseline questionnaire was administered to subjects to collect data on pain, psychological status, lifestyle and health behaviors, and sociodemographic and clinical factors. Participants free of WP (as defined by the American College of Rheumatology 1990 criteria for fibromyalgia) were followed up for 3 years, and those with new-onset WP at followup were identified. Logistic regression analyses were used to test the relationship between baseline factors and new-onset WP. Multiple imputation was used to test the results for sensitivity to missing data.


In this population-based study, 4,326 subjects (1,562 reporting no pain at baseline and 2,764 reporting some pain at baseline) participated at followup. Of these participants, 800 (18.5%) reported a status of new WP at followup (of whom, 121 [7.7%] had reported no pain at baseline and 679 [24.6%] had reported some pain at baseline). The majority of the study factors were associated with new-onset WP. However, only a few factors showed a persistent association with new-onset WP in the multivariate analysis, including age (odds ratio [OR] 0.97, 95% confidence interval [95% CI] 0.96–0.99), baseline pain status (OR 1.1, 95% CI 1.08–1.2), anxiety (OR 1.5, 95% CI 1.01–2.1), physical health-related quality of life (OR 1.3, 95% CI 1.1–1.5), cognitive complaint (OR 1.3, 95% CI 1.04–1.6), and nonrestorative sleep (OR 1.9, 95% CI 1.2–2.8). These associations persisted after adjustment for the presence of diffuse osteoarthritis (OA), which led to a modest increase in model fit (C-statistic 0.738, compared with 0.731 in the model excluding diffuse OA). The results were not sensitive to missing data.

Of the factors measured in this study, nonrestorative sleep was the strongest independent predictor of new-onset WP.