Real-world treatment patterns of immunosuppressants in Adults with systemic lupus erythematosus: A claims database analysis in the United States

Objective To explore initiation, persistence, and adherence to second-line prescribed treatments for SLE, specifically regarding the immunosuppressants azathioprine, methotrexate, and mycophenolate (conventional DMARDs), and belimumab (a biologic). Methods Clinical and insurance records were obtained for 801 patients with SLE who initiated treatment with azathioprine, belimumab, methotrexate, or mycophenolate between July 2015 and June 2019. The date of initiation defined the index date, with a 6-month pre-index and 12-month post-index period. Patient characteristics (age, gender, race, sex, ethnicity, geographic region of the US, diagnosing specialty, and type of insurance) and treatment patterns were tabulated overall and by each index medication. Logistic regression was used to model predictors of persistence for the entire sample and for each treatment cohort. Findings Approximately one-third of patients initiated methotrexate (n = 282, 35.2%) or mycophenolate (n = 258, 32.2%), with the remaining receiving azathioprine (n = 173, 21.6%) or belimumab (n = 88, 11.0%). 30% of patients were persistent with their index immunosuppressant therapy over the 12-month follow-up. The most common non-persistent treatment pattern was discontinuation which occurred in 55% of patients and was highest in the mycophenolate (58%) and lowest in the azathioprine (47%) groups. In total, 17% of patients switched to a different immunosuppressant, which was highest for the belimumab (25%) group. The average time to discontinuation was over 3 months and average time to switch was about 5 months, with patients receiving azathioprine tending to have shorter and belimumab having longer times to discontinuation or switch. Predictors of persistence were limited. Patients under the care of rheumatologists versus primary care and having higher co-morbidity assessed by CCI were associated with non-persistence for the overall sample. Race, number of SLE-related medications, census region, sex, and age were not found to be significantly related to non-persistence of immunosuppressants in this study.