Background/Purpose:  In November 2012, the first oral Janus kinase (JAK) inhibitor
tofacitinib was approved in the US for the treatment of RA with or without
methotrexate. Given its recent introduction to the market, limited real world data
are available in patient (pts) with RA receiving tofacitinib vs. TNFi. Here, we
compare characteristics of pts with RA initiating tofacitinib vs. adalimumab
(ADA) and etanercept (ETN) in a US administrative claims database and describe
the feasibility of matching treatment cohorts to yield comparable pts for an
outcomes evaluation. 

Methods: This
was a retrospective cohort study of pts aged ≥18 years at index with an
RA diagnosis (ICD9: 714.0x-714.4x and 714.81) newly initiating tofacitinib in
the Optum Research Database (11/12-7/14). Pts starting TNFi were also selected.
Pts were continuously enrolled for ≥12 months (mos) pre-index and
≥6 mos post-index (tofacitinib or TNFi start). Pts with index medication
use in prior 12 mos were excluded. Initial propensity score matching (1:1) was
conducted among biologic-naïve and biologic-experienced subgroups (using a
variable-length [VL] baseline starting 12/06) based on demographics, 12-mo
pre-index total RA-related costs and opioid use, and number of prior biologics
in the VL baseline (for biologic-experienced subgroup).

Results: Prior
to matching, 647 tofacitinib pts, 703 ADA pts, and 1864 ETN pts met the cohort
selection criteria. Tofacitinib pts had significantly (p<0.001 all
comparisons) higher 12-mo pre-index mean RA-related costs ($20,067 vs. $4,405 and $4,305) and opioid use (72% vs. 58% and 60%), VL baseline biologic experience (81%
vs. 22% and 14%) and index monotherapy regimen use (55% vs. 41% and 41%). After
matching (Table) among biologic-naïve pts, those initiating tofacitinib had
significantly higher VL baseline RA-related out-of-pocket costs, more rheumatologist
visits (vs. ADA), more Medicare Advantage pts (vs. ETN), and greater use of monotherapy
(vs. ADA) and leflunomide. In matched biologic-experienced cohorts, 49% (ADA) –
53% (ETN) of pts had prior experience with the respective index treatment.
Significantly more ADA and ETN pts had prior ETN and ADA use, respectively,
while more tofacitinib pts had abatacept, certolizumab pegol (vs. ETN) and
tocilizumab (vs. ETN) use.         

Conclusion: Tofacitinib
was more likely to be used as monotherapy and in biologic-experienced pts vs.
ETN and ADA. Initial attempts to match treatment cohorts were not feasible
given prior treatment history with respective TNFi comparators and channeling
bias with tofacitinib.