Background/Purpose: Classification criteria lacking specificity cause inclusion of many false positives in settings with low prevalence of disease. The ASAS axial spondyloarthritis (axSpA) criteria have 84% specificity. If the prevalence of axSpA among chronic back pain (CBP) patients is 5% one expects 3.6 false positively labeled persons for each true axSpA patient. Back pain is a key issue in the classification of axSpA. Our aim is to assess the specificity of spinal pain features to obtain better performance of the ASAS criteria.

Methods: We assessed (1) low back pain or stiffness; (2) thoracic inter-scapular back pain or stiffness; (3) frontal chest pain or discomfort by analyzing data from two earlier studies.^1,2 Inflammatory back pain (IBP) is defined as 4 or 5 Calin criteria: onset of back discomfort ≤ age 40; insidious onset; persistence for ≥ 3 months; morning stiffness; improvement with exercises. In the first study¹, 739 apparently healthy leisure time sportsmen (orienteers) completed a detailed questionnaire on spinal and chest pain features. AxSpA is assumed to be absent in this group. The second study² (a Swiss family study on AS) enabled assessing specificity of chest and back pain features in B27- first degree relatives of 275 B27+ AS patients. All (both B27+ and B27-) participants underwent clinical exam, full HLA-typing, radiography of SI joints and completed questionnaires, including chest and spinal pain features. Pelvic x-rays were blindly read, twice by each of 2 investigators. They agreed in 95.6% (kappa 66.4).

Results: The table shows the specificities for the sportsmen. The 3 pain syndromes were highly associated (p<0.001). Low back pain ever is age related. The specificity in the age groups 10-19, 20-29, 30-39, 40-49, 50-59 year are 70.1%, 46.3%, 52.1%, 42.8%, and 37.1% respectively. Altogether 24.1% of back pain ever can be classified as chronic IBP. The specificity of diagnostic spinal x-rays is 67.5%. Radiographs were done in 25% (6/24) in people with IBP < 3 months duration, but in 67 % (40/60) of people with ≥ 3 months IBP. Specificity was 99.1% for not using analgesics because of back pain ≤ last 6 months. The likelihood ratio for having had a diagnostic spinal radiograph is 2.2 (46/84:67/264) if chronic IBP is present compared to absent. Specificity was very high (99.1%) for not using analgesics because of back pain ≤ last 6 months. AS or sacroiliitis was found in 20/274 (7.3%) B27+ relatives. Sacroiliitis was never observed in B27-  relatives with IBP. Of note B27+ relatives without sacroiliitis have more often chest pain and IBP than B27- relatives suggesting axSpA or spondylitic disease in B27+ relatives.² The specificities for the B27- relatives are shown in the table.

Conclusion: Better knowledge of specificity of spinal pain syndromes may help improving performance of classification criteria for axSpA. Chest pain and use of analgesics seem promising candidate variables.