Background/Purpose:   Smoking has been
associated with radiographic severity and progression in SpA
although there is little understanding of the mechanism. Prospective MRI data
indicates that ankylosis develops following an
intermediary phase of fat metaplasia, which follows resolution of inflammation
in subchondral bone and at sites of erosion, when it
is termed backfill¹. We aimed to determine whether smoking
influences the propensity to develop fat metaplasia as a potential mechanism
for its association with progression in SpA.

Methods: In the FOllow-up Research Cohort in AS
(FORCAST), AS patients from Northern Alberta attending community and academic
practices are assessed for clinical and laboratory outcomes every 6 months, MRI
is performed at baseline, at 3-6 months for patients starting anti-tumor
necrosis factor alpha (anti-TNFα), and annually. MRI scans are scored
independently by 2 readers and adjudicated by a third reader according to pre-specified
rules. MRI inflammation is assessed on short tau inversion recovery (STIR)
scans using the Spondyloarthritis Research Consortium
of Canada (SPARCC) Sacroiliac Joint (SIJ) and 23-DVU Spine scores while
structural change is assessed independently on T1-weighted (T1W) scans using
the SPARCC SIJ Structural Score (SSS) score for fat metaplasia, erosion,
backfill, ankylosis, and the Fat AS Spine Score
(FASSS). We used univariate and multivariate regression
to assess associations between smoking (current (yes/no, <10/ ³10 years),
past, never, pack per day, pack year) and MRI parameters.

Results: The cohort includes 730 patients: mean age 41.3 years, 72.7%
males, 78.6% B27 positive, mean disease duration of 17.5 years. Of 517 patients
reporting smoking history, 105 (20.2%) were current smokers, 148 (29.0%) past
smokers, and 264 (51.8%) never smoked. MRI scans were available on 250 cases in
the cohort. In univariate analyses, current but not
previous smoking, especially intensity (from 0.25 to 1 pack/day) and duration of
current smoking (³10 years vs <10), was associated
with spinal (FASSS: β=11.1, p=0.03)) and SIJ fat metaplasia (SSS backfill:
β=2.4, p=0.01; SSS fat ³2: OR
2.4, p=0.03), SIJ ankylosis (β=4.5, p=0.01), and
spinal inflammation (SPARCC 23-DVU (β=8.3, p=0.02). In multivariate models
that included age, sex, B27, smoking, ASDAS, and selected according to the best
goodness of fit Akaike Information Criterion (AIC), current smoking
(intensity and/or duration) was independently associated with several MRI
parameters (Table).

Conclusion: Current, but not past smoking, and its intensity and duration is
associated with the degree of fat metaplasia and ankylosis
on MRI of the SIJ suggesting that it may influence the tissue response to
inflammation.

1. Maksymowych et al. Arthritis Rheum 2014;
66: 2958-67.