Background Biologic disease modifying antirheumatic drugs (DMARD) are a mainstay of Rheumatoid Arthritis (RA) treatment. A validated claims-based RA biologic effectiveness algorithm was developed1, allowing administrative claims data to be used for comparative effectiveness research.
Objectives To estimate time while effectively treated with biologics for RA over 2-years following biologic initiation using a claims based algorithm.
Methods Adults (18-63) newly initiating a biologic for RA (etanercept, abatacept, adalimumab, certolizumab, golimumab, or infliximab) from Jan 2009 to July 2011 were identified in the MarketScan® U.S. Commercial Database. The patient's first biologic claim (index claim) defined their treatment cohort. Patients were required to be continuously enrolled 6-months pre- and 24-months after their index claim, have a pre-index RA diagnosis and no pre-index biologic use or other conditions for which the study biologics are FDA-approved. Effectiveness was estimated using a claims-based algorithm1. Medications were classified as effective until subjects failed any of the algorithm's 6 criteria: biologic dose escalation, switching biologics, adding a new non-biologic DMARD, receiving >1 intra-articular injection per year, adding a glucocorticoid, increasing glucocorticoid dose, or low treatment compliance (<80%). Each failure event was associated with a 90-day non-effective period and patients could experience multiple failure events. Sensitivity analyses varied the non-effective time from 30 to 180 days. Total time while the patients' treatment was classified as effective by the algorithm over 2-years was calculated based on their index biologic.
Results 8,193 patients met the study selection criteria (mean age, 49, 78% female). More patients initiated etanercept (43%) or adalimumab (32%) than infliximab (12%), abatacept (7%), golimumab (4%) or certolizumab (2%). On average, patients had 526 days while effectively treated over the 2-year follow-up. The time while the medications were effective according to the algorithm was significantly higher for etanercept (538 days) than abatacept (482 days) or infliximab (480 days). Time while effectively treated was not statistically significantly different among subcutaneous therapies, etanercept (538 days), adalimumab (535 days), golimumab (537 days), or certolizumab (524 days). Sensitivity analyses varying the time not effective assigned to each failure event changed the overall effectiveness time but not treatment rankings. The majority of patients (85%) had at least 1 failure event (mean 3.2). Low compliance was the most common failure type for subcutaneous (76%) and infused biologics (80%), followed by switching biologics (26%), adding a non-biologic DMARD (17%), increasing glucocorticoid dose (17%) and increasing biologic dose (14%). Etanercept had the lowest failure rate (82%) and number of failures per patient (3.0) followed by adalimumab (84%*; 3.1*), certolizumab (85%; 3.3*), golimumab (86%*; 3.0), abatacept (93%*; 4.4*), and infliximab (95%*; 4.2*) (*: p<0.05 vs. etanercept).
Conclusions Patients initiating biologic treatment for RA had, on average, 526 days while effectively treated according to the algorithm during the 2-years following initiation. This varied from 480 (infliximab) to 538 days (etanercept).
Curtis et al. Arthritis Res Ther, 2011, 13(5), R155
Acknowledgements Research funded by Immunex Corporation, a wholly owned subsidiary of Amgen Inc.
Disclosure of Interest M. Bonafede: None declared, B. Johnson: None declared, N. Princic: None declared, N. Shah Shareholder of: Amgen, Employee of: Amgen, D. Harrison Shareholder of: Amgen, Employee of: Amgen