SAT0378 THE RELATIVE DIAGNOSTIC UTILITY OF INFLAMMATORY BACK PAIN CRITERIA IN AN INCEPTION COHORT OF PATIENTS WITH PSORIASIS, IRITIS, AND COLITIS PRESENTING WITH UNDIAGNOSED BACK PAIN

Background:

Clinicians rely on the elicitation of features of inflammatory back pain (IBP) for diagnosis of axial spondyloarthritis (axSpA) but the utility of IBP criteria in patients presenting with extra-articular features of axSpA remains unclear. Assessment of utility should include not only rheumatologist diagnosis as benchmark but imaging to address the circularity between elicitation of IBP and clinical diagnosis.

Objectives:

To assess the diagnostic utility of all criteria for IBP in patients with psoriasis, iritis, or colitis and undiagnosed back pain using the rheumatologist diagnosis and imaging as benchmarks.

Methods:

Consecutive patients (n=246) with undiagnosed back pain ≤45 years of age, ≥3 months, with any one of psoriasis (n=46), acute anterior uveitis (AAU)(n=73), or colitis (n=127) had diagnostic evaluation by a rheumatologist. Majority central reader assessment of MRI indicative of axSpA and diagnosis by the rheumatologist were external standards for testing the utility of these IBP criteria: ASAS, Berlin, Calin, rheumatologist global for IBP >5 (0-10 scale).

Results:

AxSpA was diagnosed in 44.4%, 61.6%, and 41.8% of patients with psoriasis, iritis, and IBD, respectively. Diagnostic utility for all IBP criteria was comparably poor (Table 1). MRI was indicative of axSpA in 21.2%, 43.5%, and 19.7% of patients with psoriasis, iritis, and IBD. The utility of the IBP criteria was even worse using MRI as the external reference (Table 2), especially in patients with psoriasis. Only 14% of psoriasis patients with a positive MRI reported “improvement with exercise but not rest” as compared to 70% and 62% of patients with iritis and IBD, respectively.

Table 1.

Rheumatologist diagnosis as external reference.

SensitivitySpecificityLR+LR-PsoriasisASAS IBP65.00%52.00%1.350.67Berlin IBP80.00%36.00%1.250.56Calin IBP80.00%28.00%1.110.71All 3 criteria sets60.00%56.00%1.360.71IBP global >585.00%36.00%1.330.42AAUASAS IBP84.44%42.86%1.480.36Berlin IBP80.00%57.14%1.870.35Calin IBP93.33%17.86%1.140.37All 3 criteria sets77.78%60.71%1.980.37IBP global >586.67%57.14%2.020.23IBDASAS IBP78.43%45.07%1.430.48Berlin IBP82.35%52.11%1.720.34Calin IBP84.31%19.72%1.050.80All 3 criteria sets70.59%57.75%1.670.51IBP global >580.39%66.20%2.380.30Table 2.

Central assessment that MRI is indicative of axSpA as external reference.

SensitivitySpecificityLR+LR-PsoriasisASAS IBP28.57%38.46%0.461.86Berlin IBP42.86%15.38%0.513.71Calin IBP71.43%23.08%0.931.24All 3 criteria sets14.29%42.31%0.252.03IBP global >585.71%23.08%1.110.62AAUASAS IBP75.00%26.92%1.030.93Berlin IBP70.00%38.46%1.140.78Calin IBP90.00%15.38%1.060.65All 3 criteria sets65.00%38.46%1.060.91IBP global >575.00%38.46%1.220.65IBDASAS IBP92.31%37.74%1.480.20Berlin IBP76.92%39.62%1.270.58Calin IBP92.31%16.98%1.110.45All 3 criteria sets76.92%45.28%1.410.51IBP global >592.31%47.17%1.750.16

Conclusion:

All IBP criteria have poor diagnostic utility for diagnosis of axSpA, especially in patients with psoriasis. This reinforces the desirability of less subjective assessment tools, especially imaging.

Disclosure of Interests:

Georg Kröber: None declared, Ulrich Weber: None declared, Raj Carmona: None declared, James Yeung: None declared, Jon Chan: None declared, Sibel Aydin: None declared, Liam Martin: None declared, Ariel Masetto: None declared, Stephanie Keeling: None declared, Olga Ziouzina: None declared, Sherry Rohekar: None declared, Rana Dadashova: None declared, Joel Paschke: None declared, Amanda Carapellucci: None declared, Robert G Lambert: None declared, Walter P. Maksymowych Grant/research support from: AbbVie, Novartis, Pfizer, and UCB, Consultant of: AbbVie, Boehringer Ingelheim, Celgene, Eli Lilly, Galapagos, Janssen, Novartis, Pfizer, and UCB, Employee of: Chief Medical Officer of CARE Arthritis Limited, Speakers bureau: AbbVie, Janssen, Novartis, Pfizer, and UCB

FRI0298 ASAS MODIFICATION OF THE BERLIN ALGORITHM AND THE DUET ALGORITHM FOR DIAGNOSING AXIAL SPONDYLOARTHRITIS: RESULTS FROM THE SCREENING IN AXIAL SPONDYLOARTHRITIS FOR PSORIASIS, IRITIS, AND COLITIS COHORT

Background:

Patients presenting with back pain and psoriasis, iritis, or colitis, represent a high-risk population for the presence of axial spondyloarthritis (axSpA). The Dublin Evaluation Tool (DUET)1, the Berlin algorithm2, and the ASAS modification of this algorithm3are recommended referral strategies aimed at early diagnosis of axSpA. DUET was developed for patients presenting with AAU. Validation of these algorithms in inception cohorts is limited.

Objectives:

1. To assess the performance of referral algorithms for diagnosis of axSpA when tested against the final local rheumatologist diagnosis in an inception cohort of patients presenting with undiagnosed back pain and extra-articular manifestations. 2. To determine whether different criteria for inflammatory back pain (IBP) impact the performance of the algorithms.

Methods:

The multicenter Screening for Axial Spondyloarthritis in Psoriasis, Iritis, and Colitis (SASPIC) Study at 11 sites is aimed at early detection of axial SpA in patients presenting with undiagnosed back pain to the rheumatologist. Consecutive patients ≤45 years of age with ≥3 months undiagnosed back pain with any one of psoriasis, acute anterior uveitis (AAU), or colitis diagnosed by the relevant specialist undergo routine clinical evaluation by a rheumatologist for axial SpA. The rheumatologist determines the presence or absence of axial SpA at 3 consecutive stages: 1. After the clinical evaluation; 2. After the results of labs (B27, CRP) and radiography; 3. After the results of MRI evaluation. Final diagnosis by the rheumatologist was used as external standard to test the performance of the algorithms. We tested the following criteria for IBP in the algorithm: ASAS, Berlin, rheumatologist global for likelihood of IBP >5 (0-10 scale), and DUET algorithm in AAU patients.

Results:

A total of 246 patients were recruited, 73 presented with iritis, 46 with psoriasis, and 127 with colitis, 47.6% were diagnosed with axSpA. The diagnosis of axSpA was established in 45.7%, 61.6%, and 40.2% of patients with psoriasis, AAU, and IBD, respectively. The performance of the ASAS-modification of the Berlin algorithm was superior to the original algorithm as reported previously3, primarily for enhanced sensitivity, and this was observed irrespective of the criteria used to define IBP (Table 1). Conversely, the performance of the Duet algorithm in the subset of patients with AAU was substantially worse than previously reported1.

Conclusion:

The ASAS modification of the Berlin algorithm is the preferred referral strategy for patients presenting with undiagnosed back pain to the rheumatologist.

References:

[1]Haroon M, et al. Ann Rheum Dis 2015; 74: 1990-5

[2]Poddubnyy D, et al. J Rheumatol 2011; 38: 2452–60

[3]Van den Berg R, et al. Ann Rheum Dis 2013;72:1646–53

AlgorithmSensitivity (%)Specificity (%)Correct diagnosis (%)False negative (%)False positive (%)Original Berlin(ASAS criteria for IBP)65.376.671.116.712.2Original Berlin(Berlin criteria for IBP)64.476.670.717.112.2Original Berlin(IBP global >5)67.878.173.215.411.4ASAS Modification of Berlin algorithm (ASAS criteria for IBP)73.775.874.812.612.6ASAS Modification of Berlin algorithm (Berlin criteria for IBP)73.775.074.412.613.0ASAS Modification of Berlin algorithm(IBP global >5)76.377.376.811.411.8DUET84.450.071.29.619.2

Disclosure of Interests:

Ulrich Weber: None declared, Georg Kröber: None declared, Raj Carmona: None declared, James Yeung: None declared, Jon Chan: None declared, Sibel Aydin: None declared, Liam Martin: None declared, Ariel Masetto: None declared, Stephanie Keeling: None declared, Olga Ziouzina: None declared, Sherry Rohekar: None declared, Rana Dadashova: None declared, Amanda Carapellucci: None declared, Joel Paschke: None declared, Robert G Lambert: None declared, Walter P. Maksymowych Grant/research support from: AbbVie, Novartis, Pfizer, and UCB, Consultant of: AbbVie, Boehringer Ingelheim, Celgene, Eli Lilly, Galapagos, Janssen, Novartis, Pfizer, and UCB, Employee of: Chief Medical Officer of CARE Arthritis Limited, Speakers bureau: AbbVie, Janssen, Novartis, Pfizer, and UCB

Scoring magnetic resonance imaging (MRI) inflammation and structural lesions in sacroiliac joints of patients with axial spondyloarthritis: assessment of all MRI slices of the cartilaginous compartment versus standardized six or five slices

Objectives: The Spondyloarthritis Research Consortium of Canada (SPARCC) sacroiliac joint (SIJ) scoring system assesses six or five (6/5) semicoronal magnetic resonance imaging (MRI) slices for inflammation/structural lesions in patients with axial spondyloarthritis (axSpA). However, the cartilaginous SIJ compartment may be visible in a few additional slices. The objective was to investigate interreader reliability, sensitivity to change, and classification of MRI scans as positive or negative for various lesion types using an 'all slices' approach versus standard SPARCC scoring of 6/5 slices.Method: Fifty-three axSpA patients were treated with the tumour necrosis factor inhibitor golimumab and followed with serial MRI scans at weeks 0, 4, 16, and 52. The most anterior and posterior slices covering the cartilaginous compartment and the transitional slice were identified. Scores for inflammation, fat metaplasia, erosion, backfill, and ankylosis in the cartilaginous SIJ compartment were calculated for the 'all slices' approach and the 6/5 slices standard.Results: By the 'all slices' approach, three readers scored mean 7.2, 7.7, and 7.0 slices per MRI scan. Baseline and change scores for the various lesion types closely correlated between the two approaches (Pearson's rho ≥ 0.95). Inflammation score was median 13 (interquartile range 6-21, range 0-49) for 6/5 slices versus 14 (interquartile range 6-23, range 0-69) for all slices at baseline. Interreader reliability, sensitivity to change, and classification of MRI scans as positive or negative for various lesion types were similar.Conclusion: The standardized 6/5 slices approach showed no relevant differences from the 'all slices' approach and, therefore, is equally suited for monitoring purposes.