Abstract
INTRODUCTION: Real-world data are sparse on longitudinal associations of opioid use with pain interference with activities (PIA) and daily function with osteoarthritis (OA) in the USA. METHODS: Data from the 2010-2015 Medical Expenditure Panel Surveys were analyzed for community-dwelling adults with OA. Opioid use patterns were defined as persistent, intermittent, and no use. Evaluated outcomes were a change in PIA and functional limitations (activities of daily living [ADL], instrumental ADL [IADL], social and work activities, and cognitive function). Multivariable regression analyses explored the association of persistent/intermittent versus no opioid use with PIA and functional limitations. Results were weighted for the US population. RESULTS: Among 4172 patients (66.2% female, 80.8% white, mean age 61.7 years), 62.1% reported no PIA change at follow-up, 17.9% worsened, and 20.0% improved. Although 51.0-93.1% of patients reported no functional limitations, 3.8-13.1% worsened (1.1-11.3% improved). Relative to no opioid use, persistent users had higher odds of severe/extreme PIA (adjusted odds ratio [AOR], 2.91; 95% confidence interval [CI], 1.95-4.32; P < 0.001) and moderate PIA (AOR, 2.04; 95% CI, 1.31-3.20; P < 0.01) at follow-up. For patients with baseline functional limitations, persistent opioid users were more likely to report physical and work limitations at follow-up (both P < 0.05). For patients without baseline functional limitations, persistent opioid users had higher odds than those without use of reporting IADL, physical, social, and cognitive limitations at follow-up (all P < 0.05); intermittent users were more likely to report physical and social limitations (both P < 0.05). CONCLUSIONS: Persistent opioid use for pain in patients with OA appeared to be associated with poorer PIA and functional outcomes, regardless of baseline functional status. These findings highlight the importance of patient-reported outcomes for opioid benefit-risk assessment and suggest the need for alternative analgesic approaches.