Background/Purpose: Anterior uveitis is a relatively frequent extra-articular
manifestation in ankylosing spondylitis (AS), with a
prevalence of 25.8% in a recent meta-analysis. TNF
inhibitor (TNFi) treatment has been shown to reduce
the rates of uveitis in this patient group. Some studies have suggested a
greater protective effect with the anti-TNF
monoclonal antibodies infliximab (IFX) and adalimumab (ADA) than with etanercept
(ETN). Our objective was to study the effect of ADA, ETN or IFX treatment on occurrence
of anterior uveitis in a cohort of patients (pts)
with AS.

Methods: Data on AS pts starting ADA, ETN
or IFX as their first TNFi
from Jan 2003 through Dec 2010 were extracted from the Swedish Biologics
Register ARTIS (follow-up data available through Dec
2011). Data on anterior uveitis, available from Jan 2001, were obtained by
linkage to the National Patient Register as visits to an ophthalmologist with
ICD-10 codes indicating anterior uveitis (H20, H22.1). We applied two main
analytical approaches: (1) Comparison of uveitis rates (per 100 patient-years)
before TNFi start (follow-up time pre-Rx 2-10 yrs) and during TNFi treatment
for each TNFi, analysing (by Poisson regression): (a)
total numbers of uveitis visits, (b) uveitis flares defined by a 60-day penalty
from the index visit of one uveitis flare for a new flare to be counted, and
(c) uveitis flares defined by a 90 day penalty from last uveitis flare
follow-up visit for a new flare to be counted. (2) Cox regression analysis
comparing hazard of first uveitis flare for ADA, ETN
or IFX in a subgroup of pts
who were uveitis-free during the 2 yrs before TNFi start, +/- adjustment for potential confounders (age,
sex, prescription year, level of education, disease duration, presence of IBD, baseline CRP and use of DMARD
co-medication, prednisolone and NSAIDs).

Results: 1365 AS pts
were included (406 ADA, 354 ETN, 605 IFX). IFX pts
more often used DMARD co-medication; ADA pts had lower baseline CRP. ADA was more frequently used
towards the end of the study period. IFX doses tended
to be lower than 5 mg/kg (at baseline/last registration 62%/50% used <=200 mg). A clear reduction in uveitis rates was observed for ADA, a
slight reduction for IFX and a marked increase for ETN, irrespective of the method for counting flares
(Table). 1127 pts were uveitis-free 2 yrs prior to TNFi start and
included in the Cox regression analyses. In the adjusted analysis ADA and IFX were superior vs. ETN
(Table).

Conclusion: In this large register-based study
of AS pts we found reduction in uveitis rates with
ADA treatment, a slight reduction with IFX and
increased rates with ETN. In a subgroup free of
recent uveitis, ADA and IFX were associated with
significantly lower hazard for first uveitis flare than ETN.
Potential caveats of this study include residual confounding, the relatively
low doses of IFX used, reporting bias, and the
possibility that the required uveitis-free period for pts
included in the Cox regression analysis was too short.