Background/Purpose: Anti-TNFs have been used to treat RA since 2000. Since the availability of new agents with different mechanisms of action around 2007, the pattern of use has evolved. The long-term biologic retention rate is a good surrogate marker for effectiveness in the clinical setting. Randomized controlled trials and their open-label extension studies provide information on drug retention in a highly selected patient population. This information, however, is of limited use when prescribing biologics in a real-world setting. Real-world databases, such as RHUMADATA®, provides an excellent opportunity to examine the changes in retention rates over time.

Objectives: To assess the retention of the first-line anti-TNF and evaluate possible differences between two eras, early (2000-2006) and late (2007-2015).

Methods: Data of biologic naïve RA patients starting an anti-TNF [adalimumab (ADA), etanercept (ETA), golimumab (GOL) or infliximab (INF)] between Jan 1, 2000, and Dec 31, 2006 (early era), and Jan 1, 2007, and Dec 2015 (late era) was extracted. Patients were followed until they stopped, changed medication, or were lost to follow-up. Collected data included demographics, concomitant medication, comorbidities, laboratory variables, PROs, and disease activity measures. Infections occurring while on treatment, biologic status (ongoing or stopped) and the reasons for biologic cessation were also extracted. Overall retention was compared using Kaplan-Meier estimates (SAS 9.4). Cox proportional hazard models were used to adjust for differences between groups at baseline.

Results:  Baseline variables were similar between the two eras, except that patients in the second era were older and had been diagnosed ~1 year sooner than first era patients. These patients were less exposed to corticosteroids, more exposed to HCQ, and had lower ESR and CRP at treatment initiation. Over the entire period, ETN had a longest mean retention time of 6.32 years versus ADA (6.24y) and INF (5.6y). This difference was statistically significant between ETN and INF (p=0.026). When analyzed in the two eras, the mean retention times of all biologics were longer in the first era (6.72y) than the second (4.7y).

Conclusion: Patients treated in the first era appear to have more severe disease, greater exposure to corticosteroids and HCQ, and better retention rate compared to patients starting an anti-TNFi after 2006. More analysis should be done to understand these differences.