Background/Purpose:
The objective of this analysis was to examine persistence with intravenous (IV)
and subcutaneous (SC) anti-TNF therapies among rheumatoid arthritis (RA)
patients (pts) within the US CORRONA RA registry.

Methods: A
retrospective analysis from the CORRONA RA registry was used for analyses. 
Adult RA pts (biologic-naïve and biologic-experienced) with moderate-to-severe
RA (CDAI>10) who initiated anti-TNF biologics within the CORRONA registry after
June 1, 2009 and had ≥6 months of follow up were eligible. Pts receiving IV
anti-TNF therapy were compared with those receiving SC anti-TNF therapy. Propensity
scores (PS) were generated to match IV anti-TNF initiators to SC anti-TNF
initiators, stratified by line of therapy. Variables used for PS matching were
based on characteristics with standardized differences of ≥0.1 including age,
gender, duration of RA, work status, smoking status, insurance, prednisone use,
monotherapy, prior serious infections, patient global, morning stiffness, year
of initiation, time from initiation to 6 month visit, and prior number of biologics
to control for population imbalances. Time-varying hazard ratios (HR) comparing
persistence in matched IV vs SC through 48 months were estimated and stratified
by line of therapy.

Results: A
total of 1544 pts were included (772 each for IV and SC therapy). In each
treatment group, 387 pts were bio-naïve, 220 had received 1 prior biologic, and
165 had received ≥2 prior biologics. Among all pts, significantly greater
proportions of patients receiving IV therapy remained on that agent compared
with pts receiving SC therapy (p=0.002) during the first 12 months (HR=0.76
[0.63-0.90]). No significant differences in persistence were observed from 24 –
48 months (HR=1.09 [0.88-1.35], p=0.45) (Figure 1). Similar results were
observed for pts who were biologic-naïve (0-12 mons: HR=0.69 [0.52-0.90],
p=0.007) and who had received 1 prior biologic (0-12mons: HR=0.80 [0.59-1.10],
p=0.17). For pts who had received ≥2 biologics, there was no significant
difference in persistence between patients receiving IV or SC anti-TNF therapy.

Conclusion:
Within the first 12 months of initiating anti-TNF therapy, pts receiving IV
therapy who were biologic-naïve or had received only one prior biologic were
more likely to remain on treatment compared with those receiving SC anti-TNF
agents. These results may have implications on cost savings, in the first year
of treatment, that need to be investigated in further studies.

Figure1.
Persistence estimates for matched populations.