(B21) Joint pain in patients with atopic dermatitis receiving treatment with dupilumab in a nation-wide cohort study

Publication Number: 633

Background: Case reports and small cross-sectional studies of atopic dermatitis (AD) patients have reported an association between dupilumab and joint pain and arthritis, though this was not observed in randomized trials. Recently, the FDA label for dupilumab added “arthralgia” as a possible side effect of dupilumab based on these case reports, noting that the onset of arthralgia ranges from days to months after the first dose.

Objectives: We compared rates of joint pain/arthralgia and arthritis between patients with AD who started dupilumab vs. mycophenolate or cyclosporine.

Methods: Separately in each of the two national claims databases, we identified patients with AD newly treated with dupilumab, mycophenolate, or cyclosporine. New use was defined as no use of the exposure or referent agent within the past 180 days. We excluded patients with existing joint pain (ICD-10 code for arthralgia is ‘joint pain’) or arthritis. During 6 months of treatment, we compared the frequency of newly recorded diagnoses of joint pain (primary outcome), inflammatory arthritis, lower back pain, tendinitis/enthesitis and myalgia. Incidence rates and hazard ratios (HR) were estimated before and after 2:1 propensity-score matching. In trials and multiple cohort studies, dupilumab has been associated with an increased risk of conjunctivitis, while mycophenolate and cyclosporine are not associated with conjunctivitis. Thus, conjunctivitis was included as a positive tracer outcome. Using an as-treated analysis, patients were censored at the date of outcome, medication discontinuation, medication switch, maximum follow-up (180 days), death, disenrollment, or end of data stream, whichever came first. After separate estimation in each database, findings were pooled using a fixed-effects meta-analysis.

Results: Overall, 4,249 patients started dupilumab and 2,167 started mycophenolate or cyclosporine. The incidence rate (per 1,000 person-years) of ‘joint pain’ was 100 in dupilumab vs. 65 among mycophenolate/cyclosporine initiators, yielding a PS-matched hazard ratio of 1.27 (0.93-1.72) for dupilumab compared to mycophenolate/cyclosporine. There was no consistent association with other endpoints: HR=0.52 (0.15-1.79) for arthritis; 1.75 (0.84-3.70) for tendinitis/tenosynovitis; 1.25 (0.86-1.79) for lower back pain; 0.53 (0.29-0.99) for myalgia. We confirmed the known increased risk of conjunctivitis in dupilumab (HR=2.56;1.79- 3.70).

Conclusions: In this nationwide cohort study, we did not observe meaningfully increased rates of newly recorded joint pain or arthritis in patients with AD who started dupilumab compared to mycophenolate or cyclosporine.