Background/Purpose: Among patients with rheumatoid arthritis (RA) who have an inadequate response to conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), the addition of a biologic DMARD (bDMARD), is recommended. If treatment with a first bDMARD fails, patients may move to subsequent bDMARD lines of therapy (LOT), either cycling to a bDMARD with the same mode of action (MOA) or switching to a targeted therapy (bDMARD or janus kinase inhibitor) with a different MOA. The objective of this study was to evaluate LOT among RA patients receiving bDMARDs in the US.

Methods: Data were taken from the MarketScan Commercial claims database between Jan 01, 2011 and Sep 30, 2015. Patients were included if they had: ≥1 inpatient or ≥2 outpatient claims at least 30 days apart for RA; ≥1 bDMARD (tumor necrosis factor inhibitor [TNFi] or non-TNFi [date of first bDMARD initiation was the index date]); and 12 months continuous medical and prescription insurance coverage prior to the index date. The bDMARD at the index date was defined as LOT1 and could be used as either monotherapy or in combination with a csDMARD. Patients were considered to have a new LOT (e.g. LOT2, LOT3) when a prescription or administration for a different RA therapy was filled, or the patient resumed the same therapy after a gap of >90 days. Counts and percentages were reported for categorical variables. Means (SD) were reported for continuous variables. Kaplan-Meier (KM) analysis was used to analyze time to discontinuation of each LOT.

Results: 17,525 RA patients were included in the analysis: mean (SD) age 50.9 (9.6) years; 79.0% women; mean (SD) number of LOT 1.8 (1.1); and median follow-up 27 months. The most commonly used medications for LOT1 were etanercept (36.6%) and adalimumab (31.4%). Across all LOT and individual medications, the mean proportion of days covered in the 12 months post-index was 0.63. Approximately half of patients (54.5%) had 1 LOT; 25.8% had 2 LOT; 11.7% had 3 LOT; 8% had ≥4 LOT. In patients who started with a TNFi on the index date, 20.6% had ≥3 LOT, and in patients who had a non-TNFi on the index date, 15.2% had ≥3 LOT. Among patients who had a non-TNFi on the index date, 75.6% and 75.6% of patients who moved to LOT 2 and 3, respectively, had subsequent non-TNFi’s. Among patients who had a TNFi on the index date, 76.0% and 51.5% of patients who moved to LOT2 and 3, respectively, had subsequent TNFi’s. Overall, the use of concomitant methotrexate was greater (47.8%) in LOT1 than in LOT2 (38.4%) or LOT3 (35.4%). KM estimates showed that patients spent the longest time on LOT1 (mean: 726.9 days), followed by LOT2 and LOT3 (mean: 567.6 days and 487.2 days, respectively).

Conclusion: In this study, the longest duration of treatment occurred for LOT1, and shortened with each subsequent LOT. These results suggest frequent cycling occurs within the non-TNFi and TNFi categories. Future research is needed to understand the duration of LOT by non-TNFi and TNFi agents and the impact on patient outcomes.