Factors associated with commencing and ceasing opioid therapy in patients with rheumatoid arthritis

Abstract

Objective: To determine factors associated with opioid use in rheumatoid arthritis (RA) patients. Methods: Adult RA patients (n = 3225, 73% female, mean age 57 years, median follow-up 54 months) were recruited into the Australian Rheumatology Association Database (ARAD) between 2001-2015. A logistic regression examining both within- and between-patient effect sizes for time-varying covariates, and transition-state analysis for covariates associated with opioid commencement or cessation were used to examine determinants of current opioid use. Results: The population-averaged prevalence of any opioid use was 33% (95%CI 32-34), 9% (95% CI 8, 10) for high potency opioid use, and 62% (95% 60, 64) of patients reported opioid ever-use after five years of follow-up. Opioid use was higher in females and decreased with older baseline age. Within-patients opioid use was associated with higher self-reported pain and HAQ scores (p < 0.001), and NSAID (OR 1.88; 1.67-2.10), oral glucocorticoid (2.23;1.93-2.58), csDMARD (2.08;1.78-2.44) and bDMARD (1.22;1.06-1.40) treatment. Younger baseline age, higher pain scores, HAQ scores and oral GC use were important determinants of change in opioid use, associated with both a higher probability of commencing opioid use, and a lower probability of cessation. Paradoxically, NSAID and DMARD treatments were associated with both a lower probability of commencing opioids, and a lower probability of cessation. Conclusions: There was a high prevalence of opioid use among RA patients, which was associated with pain, function and GC treatment. NSAID, and DMARD treatments obviate the need for opioids in some, but not all, patients.