Real world population-based assessment of COVID-19 outcomes among rheumatoid arthritis patients using biologic or synthetic DMARDs

ABSTRACT

Background: While some risk factors for severe COVID have been identified for patients with rheumatic diseases,1 few studies have investigated whether outcomes differ based on the type of rheumatoid arthritis (RA) treatment. Most existing reports have been limited to individual centers or voluntary reporting registries.2,3 Objectives: To compare the occurrence of hospitalizations following COVID-19 diagnosis among patients with RA treated with various classes of DMARDs. Methods: A cohort of patients with confirmed COVID-19 (ICD10 diagnosis code or positive PCR or antigen test result) were identified within a large US electronic health record (EHR) dataset (Optum, Inc.) during the time period Feb 1, 2020 through Oct 14, 2020. From these, we identified RA patients (ICD10 RA diagnosis code) with treatment (most recent of JAK inhibitor [JAKi], biologic [bDMARD] or conventional synthetic [csDMARD] only) within the 12 months prior to COVID-19 diagnosis (i.e., index). The primary outcome was any hospitalization on or within 30 days after COVID-19 diagnosis. Multivariable logistic regression models compared users of JAKi's to non-TNFi bDMARDs and csDMARDs (separately), as well as users of TNFi's to non-TNFi bDMARDS and csDMARDs (separately), and were adjusted for age, gender, index month and baseline corticosteroid use. Sensitivity analyses included restriction of prevalent treatment use to within 180 days prior to COVID-19 diagnosis and restriction of csDMARDs to a group without hydroxychloroquine or chloroquine. Results: The study included 910 RA patients on DMARD treatment who were diagnosed with COVID-19 (mean age ± SD 61±15, 80% female, 62% white. Of those, 26% (n=240) were hospitalized on or within 30 days after COVID-19 diagnosis. The proportion of patients hospitalized was highest in non-TNFi bDMARD users (37/87;43%), followed by csDMARDs users (161/581;28%) and lowest in JAKi (13/68;19%) and TNFi users (29/174;17%). In multivariable-adjusted models, no differences in risk of hospitalization were found comparing JAKi users to csDMARD users (aOR=0.71;95% CI 0.37-1.36) or TNFi users to csDMARD users (aOR=0.67;95%CI 0.43-1.06). Compared to non-TNFi bDMARD users, JAKi use and TNFi use was associated with reduced risk of hospitalization (JAKi aOR=0.32;95%CI 0.14-0.71;TNFi aOR=0.34;95%CI 0.18-0.62). Age and corticosteroid use were positively associated with 30-day hospitalization in all models. Results of sensitivity analyses were consistent with the main findings. Conclusion: In this study, roughly a quarter of RA patients with recent DMARD treatment were hospitalized within 30 days after COVID diagnosis. Patients treated with JAKi and TNFi therapies experienced the lowest risk of hospitalization, with risk of hospitalization significantly lower than non-TNFi bDMARDs. However, recent therapy recorded in the EHR may not reflect exposure at time of COVID-19 diagnosis and small sample size per treatment may limit interpretation.