Age-Dependent Patterns of Short Sleep Among U.S. Adults with Arthritis: A Nationally Representative Analysis
TL;DR
Arthritis was not independently associated with short sleep after full adjustment but age meaningfully modified this relationship, suggesting sleep health interventions may require age-specific strategies in adults with arthritis.
Abstract
Background: Sleep duration is a critical determinant of metabolic, psychological and physical health. Arthritis is highly prevalent among adults and frequently coexists with sleep disturbances; however, evidence examining whether arthritis independently increases the risk of short sleep duration remains inconsistent. Objective: To evaluate whether adults with arthritis (including osteoarthritis [OA] and Rheumatoid Arthritis [RA]) have higher odds of short sleep (<7 hours/night) and whether age modifies this association. Methods: We analyzed 9,160 adults aged ≥20 years from NHANES 2015-2018, applying the complex survey design and 4-year MEC examination weights. Arthritis status (none, OA, RA) was self-reported. Sleep duration was categorized as short sleep (<7 hours). Survey-weighted descriptive statistics and logistic regression models were performed (unadjusted, demographic-adjusted and fully adjusted). Age was mean-centered at 50.22 years. Subtype-specific and sex-stratified analyses were conducted and interaction terms tested effect modification by age. Results: Adults with arthritis were older and had higher BMI and smoking prevalence than those without arthritis. Overall, 22.0% of adults reported short sleep (<7 hours/night), with comparable prevalence in those with (21.8%) and without (22.1%) arthritis. In fully adjusted models, arthritis did not independently predict short sleep (adjusted odds ratio [AOR] 1.14, 95% CI 0.98-1.32, p = 0.112). However, the arthritis×age interaction was significant (AOR 0.97, 95% CI 0.97-0.98, p<0.001), indicating age-dependent differences in predicted probability of short sleep. OA but not RA demonstrated a significant age interaction (OA×age AOR 0.97, p = 0.001; RA×age AOR 1.00, p = 0.897). Sex-stratified models showed similar patterns. Conclusions: Arthritis was not independently associated with short sleep after full adjustment but age meaningfully modified this relationship. Sleep health interventions may require age-specific strategies in adults with arthritis.