Article Text

SAT0566 Real World Study of Biologic Disease Modifying Antirheumatic Drug Persistence among Patients with Rheumatoid Arthritis who have Previously used at least One Other Biologic Agent: a U.S. Administrative Claims Database Analysis
  1. S. Johnston1,
  2. D. McMorrow1,
  3. A. M. Farr1,
  4. P. Juneau1,
  5. S. Ogale2
  1. 1Truven Health Analytics, Washington
  2. 2Genentech, South San Francisco, United States


Background Switching and discontinuation of biologic disease modifying antirheumatic drugs (DMARDs) may occur due to inadequate response to the current biologic, side effects, or other factors.

Objectives To compare, between biologic agents, persistence (time-to biologic DMARD switching and time-to switching/discontinuation) among patients with rheumatoid arthritis (RA) who have previously used at least one other biologic agent.

Methods Using a large U.S. administrative claims dataset, we conducted a retrospective study of adult RA patients, starting a biologic (abatacept [ABA], adalimumab [ADA], certolizumab [CZP], etanercept [ETA], golimumab [GOL], infliximab [INF], rituximab [RTX], or tocilizumab [TCZ]) between Jan 1, 2010 and Jun 30, 2011. The date of initiation for the biologic used during this period was designated the index, and patients who had used at least one other biologic at any time prior to the index were included. A 6-month pre-index period was used to measure patient characteristics and a variable-length post-index period was used to measure persistence defined in two ways: 1) time-to switch (days from index until switch to a different biologic agent) and 2) time-to switch/discontinuation (days from index until the first occurrence of a 90-day gap in treatment with the initiated biologic agent or a switch to a different biologic agent); for patients who had not switched or discontinued, follow-up was censored at the end of the study period (Jun 30, 2011) or disenrollment from insurance. Patients could contribute multiple observations to the dataset; one for each biologic they initiated sequentially during the study period.

Multivariable Cox proportional hazards models with the Huber-White “sandwich” variance estimator were used to compare persistence between the biologic agents, adjusting for patients’ demographics and pre-index health status, comorbidities, and medication use. Rituximab was excluded from analyses of time-to switch/discontinuation because its long retreatment intervals complicate the definition of discontinuation based on gaps in therapy.

Results The sample comprised 7,515 observations; mean patient age 54 years, 81% female. During the pre-index period 75% of patients used corticosteroids and 53% used methotrexate. Table displays multivariable-adjusted hazard ratios (HRs), with TCZ as the reference category, for the two persistence definitions.

Conclusions Among RA patients initiating biologic agents after previously failing a different biologic, differences in persistence exist between agents. Using time-to switch as a measure of persistence, TCZ had significantly better persistence compared with all studied biologics, except RTX, which was similar; using time-to switch/discontinuation as a measure of persistence, TCZ had significantly better persistence compared with ADA, but similar to the other biologics.

Disclosure of Interest S. Johnston Employee of: Truven Health Analytics, D. McMorrow Employee of: Truven Health Analytics, A. Farr Employee of: Truven Health Analytics, P. Juneau Employee of: Truven Health Analytics, S. Ogale Shareholder of: Genentech, Employee of: Genentech

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.