Background/Purpose: Previous studies evaluating predictors of major clinical response in patients with non-radiographic axial SpA (nr-axSpA) receiving treatment with anti-TNF agents have been limited by the heterogeneity of patients recruited according to classification criteria and/or disease duration. We aimed to assess the predictive capacity of active and structural lesions on MRI of the sacroiliac joints (SIJ) in a cohort of patients selected according to objective measures of inflammation and limited duration of disease.

Methods: Patients had axial SpA per the Assessment of SpondyloArthritis (ASAS) classification criteria, but did not meet modified NY radiographic criteria. Patients had symptoms for >3 months and <5 years, Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) ≥4, and failed ≥2 NSAIDs. Patients were randomly assigned to etanercept 50 mg/week or placebo, then after 12 weeks, all patients received open-label etanercept 50 mg/week. Clinical and health outcomes were assessed throughout the study, and MRI of the SIJ and spine was performed by two central readers at baseline, weeks 12 and 48 to assess bone marrow edema (BME) using the Spondyloarthritis Research Consortium of Canada (SPARCC) score. Additionally, a post-hoc analysis was conducted to score structural lesions using the SPARCC SIJ structural method (SSS), which assesses fat metaplasia, erosion, backfill, and ankylosis on T1-weighted spin echo (T1WSE) MRI. Two independent readers scored baseline and 48 week T1WSE MRI scans from 187 cases blinded to patients and short tau inversion recovery (STIR) MRI scans. Mean scores of the readers were used. Baseline high sensitivity CRP (hsCRP) levels, SPARCC MRI inflammation and SSS erosion scores were analyzed using logistic models of week 48 ASAS40 and ASDAS major improvement (ASDAS MI change ≥2.0), adjusted for treatment.

Results: Mean (SD) age was 32 (7.8) years, 60.5% were male, and mean (SD) duration of disease symptoms was 2.5 (1.8) years. A total of 73% of patients were human leukocyte antigen B27 (HLA-B27) positive and 81% met the ASAS MRI imaging criteria at baseline. Baseline CRP, SPARCC SIJ inflammation, and SSS erosion scores, but not fat metaplasia, backfill, or ankylosis were significant predictors of both ASAS40 and ASDAS MI responses at week 48 in both last observation carried forward and observational data analyses (see table). The higher the baseline value the greater the likelihood of response.

Table: Logistic Models of week 48 ASAS40 or ASDAS MI for each of the baseline SSS components, adjusted for treatment
Week 48 Outcome	Baseline Predictor	Adjusted Odds Ratio (95% CI)	P-value
ASAS40, LOCF	CRP	1.05 (1.02,1.09)	0.0042
ASAS40, LOCF	SSS Erosion Score	1.09 (1.00,1.19)	0.0432
ASAS40, LOCF	SPARCC SIJ	1.06 (1.02,1.09)	0.0006
ASAS40, OC	CRP	1.05 (1.01,1.09	0.0110
ASAS40, OC	SSS Erosion Score	1.09 (1.00,1.19)	0.0412
ASAS40, OC	SPARCC SIJ	1.05 (1.01,1.08)	0.0057
ASDAS MI, LOCF	CRP	1.17 (1.11,1.24)	<0.0001
ASDAS MI, LOCF	SSS Erosion Score	1.11 (1.03,1.21)	0.0099
ASDAS MI, LOCF	SPARCC SIJ	1.07 (1.04,1.10)	<0.0001
ASDAS MI, OC	CRP	1.15 (1.09,1.22)	<0.0001
ASDAS MI, OC	SSS Erosion Score	1.12 (1.03,1.21)	0.0088
ASDAS MI, OC	SPARCC SIJ	1.07 (1.04,1.11)	<0.0001
LOCF=last observation carried forward, OC=observed case, CI=confidence interval

Conclusion: The presence of objective manifestations of active disease as indicated by CRP and inflammatory or erosive lesions on MRI prior to the start of anti-TNF therapy has predictive capacity for major treatment responses.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/active-and-structural-lesions-on-mri-of-the-sacroiliac-joints-predict-major-clinical-responses-in-patients-with-non-radiographic-axial-spondyloarthritis-treated-with-etanercept/