Background/Purpose: The value of a combination of items defining inflammatory back pain (IBP) to screen for axial spondyloarthritis (axSpA) in primary care has recently been studied. However, whether and how measurement of HLA B27 may contribute to that is unclear. To study the value of HLA B27 determination for early identification of patients with axSpA in primary care.

Methods: Consecutive patients <45 years (n=950) with back pain >2 months who presented to orthopaedists (n=143) were randomized based on 4 key questions related to inflammatory back pain (IBP) for referral to rheumatologists (n=36) who made the diagnosis. HLA B27 was either assessed in primary or in secondary care in 298 patients. The predictive value of HLA B27 alone and in combination with other items for a diagnosis of axSpA was calculated. For variable selection logistic regression was applied, optimizing sensitivity, specifity and likelihood ratios using different strategies to predict axSpA.

Results: Rheumatologists saw 325 randomly selected patients. Due to missing HLA B27 values the main analysis is based on 298 patients, mean age 36 years (y), 52% female, median duration of back pain 32 months: 107 patients were diagnosed as axSpA (36%), 46 ankylosing spondylitis and 61 axial non-radiographic axSpA. A simple model with only HLA B27 as independent variable, indicates that HLA B27+ patients have an odds ratio (OR) of 12.24  to have axSpA in comparison with HLA B27- patients (sensitivity 62.6%, specifity 88.0%). The positive likelihood ratio (LR+) was 5.2 and the negative LR- 0.43. Thus, HLA B27 alone performed better than our recent 3/5 items proposal, which had a LR+ of 1.47 and LR- 0.53.  Simply adding HLA B27 to those 5 criteria did not improve the LR+ substantially: for 3/6 items it was 1.64 (LR-: 0.33) and for 4/6 LR + 3.26 (LR-: 0.51). However, the simple combination of HLA B27 and/or buttock pain was 89.7% sensitive and 40.3% specific.A model based on determination of HLA B27 in primary care enabled us to analyse the performance of a two-phase strategy by dividing the patients into two groups according to their B27 status. Including additional variables to HLA B27+ patients did not improve the overall performance. However, in the HLA B27- group the following items were most relevant: improvement by movement, buttock pain and psoriasis. Moreover, a combination of this information revealed that a young patient with chronic back pain is likely to have axSpA if HLA B27 is positive and/or if 2 or 3 of the following symptoms are present: improvement by movement, buttock pain (both sided) and history of psoriasis. This combination was 80.4% sensitive and 75.4% specific (LR+: 3.27 and LR-: 0.26).

Conclusion: This study shows that patients with axSpA can be identified with or without knowledge of HLA B27 based on questions specific for IBP in primary care. However, since models including HLA B27 had better predictive results in this study, it seems to be more useful to generally determine HLA B27 in all patients with chronic back pain of young age in primary care to further reduce the delay in diagnosing axSpA.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/identification-of-axial-spondyloarthritis-among-patients-with-chronic-back-pain-in-primary-care-how-does-determination-of-hla-b27-influence-the-performance-of-clinical-assessments-of-inflamm/