Background/Purpose:

Biologic agents are highly effective for the treatment of juvenile idiopathic arthritis (JIA) but have the potential risk of increased serious infections. Using observational administrative claims data, we compared the rate of hospitalized infections among patients with JIA newly starting biologic agents versus newly starting methotrexate (MTX) without concurrent biologic use.

Methods:

We used national U.S. Medicaid analytic eXtract (MAX) administrative claims files from 2000 – 2010 inclusive. New users of the 5 available tumor necrosis factor inhibitors (TNFi), the interleukin 1 inhibitor anakinra (ANA), and MTX (without concurrent biologic use) were defined by a 6 month clean period of non-use (“baseline”). Patients with a physician diagnosis code for JIA before age 16 years and prior to new use were included. Follow-up began on the day of the first prescription fill and medication exposures were extended for 90 days beyond the days supplied by each fill. The outcome was hospitalization with any infection as the primary discharge diagnosis. Cox proportional hazards models were used to generate hazard ratios adjusted (aHR) for age, sex, race, comorbidities, baseline mean daily oral systemic glucocorticoid (GC) dose, and infections during baseline. Among biologic users, concurrent MTX use during follow-up was treated as a time-varying covariate.  Sensitivity analyses included shortening the medication exposure risk window and including all hospital discharge diagnoses in the outcome assessment.  

Because nearly all patients who were new users of ANA were expected to have systemic JIA (sJIA) and there is no reliable physician diagnosis code to identify sJIA patients in the MTX comparator cohort, we attempted to identify patients with a high likelihood of having sJIA based upon any physician diagnoses for macrophage activation syndrome or any use of relatively sJIA-specific medications (cyclosporine or tacrolimus in the absence of uveitis, anakinra, canakinumab, rilonacept, thalidomide, lenalidomide) at any time.

Results:

We identified 3075 new MTX users (204 with sJIA), 3471 new TNFi users (60% etanercept, 28% adalimumab, 11% infliximab), and 247 new ANA users. Crude infection rates per 100 person-years were: MTX 1.46 (39/2668); TNFi 1.70 (66/3887); ANA 8.41 (19/226); MTX with sJIA 2.64 (3/114). The aHR for TNFi vs MTX was 1.11 [0.69-1.79]. Among TNFi users, concurrent MTX use suggested an increased risk of infection that was not statistically significant (aHR 1.44 [0.89-2.34]).  The aHR for ANA vs MTX was 2.92 [1.77-7.71] and was 2.91 [0.62-13.5] vs MTX users with sJIA. Among all patients in the study, baseline high dose oral GC (>10mg/day) was associated with increased infection risk compared to no GC use (aHR 1.86 [1.16-2.97]). The results were robust to sensitivity analyses.

Conclusion:

There was no significant increased risk of hospitalized infection among new TNFi users compared to new MTX users. New ANA users had an increased risk of infection compared to MTX, even when restricted to MTX users who likely had sJIA. High dose oral GC was associated with a significantly increased risk of infection among all children with JIA. Further investigation about the risk of infection in children with sJIA is needed.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/the-risk-of-hospitalized-infection-following-initiation-of-biologic-agents-versus-methotrexate-in-the-treatment-of-juvenile-idiopathic-arthritis/