Background/Purpose: An analysis of radiographic
progression in an observational cohort has suggested that anti-TNF therapy lessens
radiographic progression, especially when treatment is introduced within 5
years of AS onset, although development of new syndesmophytes
was not reported¹. We aimed to assess the impact of anti-TNF therapy
on radiographic progression in an observational cohort of AS by analyzing
propensity-matched patients.

Methods: In the FOllow-up
Research Cohort in AS (FORCAST), 374 patients with AS from Northern Alberta
attending community and academic rheumatology practices have been assessed over
a mean follow up of 3.1 years for clinical and laboratory outcomes every 6
months, and had spinal radiography at baseline and follow up. Patients received
standard (n=147) or anti-TNF therapy (n=227). Radiographs
were scored by two readers using the mSASSS with
adjudication by a third reader according to pre-specified rules. The
impact of anti-TNF therapy was assessed using univariate
and multivariate Tobit regression for mSASSS data and Poisson regression for development of new syndesmophytes. Propensity matching was based on a caliper of 0.1 in a logistic model using age,
gender, disease duration, smoking, mSASSS, and
pre-treatment BASDAI and CRP. Multiple propensity score analysis based on multinomial logistic
regression was used to compare treatment groups according to disease duration
before start of anti-TNF. Missing values were imputed using the multiple imputation by chained equations (MICE) method.

Results: Subjects were predominantly males (74.9%)
of mean (SD) age 43.0 (11.9) years, mean (SD) disease duration 21.4 (11.6) years,
mean (SD) baseline mSASSS of 17.1 (18.4), mean (SD)
radiographic progression of 0.97 (1.31) mSASSS unit/year,
and 38.4% developed a new syndesmophyte. After
propensity matching using baseline covariates 218 and 230 patients could be
included in the post-match analysis of mSASSS
progression and new syndesmophytes, respectively.
These numbers were 158 and 190, respectively, after tighter matching using
pre-treatment covariates. In univariate analyses of
propensity-matched cases, significant predictors were baseline mSASSS and pre-treatment,ASDAS for mSASSS
progression, and baseline mSASSS and age for new syndesmophyte.  In
the final multivariable Tobit and Poisson regression
models using propensity-matched data, only pre-treatment ASDAS was a
significant predictor of mSASSS progression (β=0.40 [95%CI: 0.08-0.72]; p=0.014). Multiple propensity
score adjusted analyses demonstrated significantly less radiographic
progression in patients who received anti-TNF within 5 years of disease onset versus
those on standard therapy or treated with anti-TNF after >10 years of
disease (Table).

Conclusion: A beneficial effect of anti-TNF
treatment may be evident in AS patients receiving treatment within 5 years of
disease onset.

1.    
Haroon et al. Arthritis Rheum 2013; 65:
2645-54.