• spondyloarthritis
• low back pain
• inflammation

Inflammatory back pain (IBP) is considered so essential in the axial spondyloarthritis (axSpA) diagnostic process that it is recommended as referral parameter in primary care.1–3 However, axSpA patients without IBP do exist as well as patients with IBP that do not have an axSpA diagnosis, leaving IBP to be a strong and useful feature for recognising axSpA but not pathognomonic.4 5 A recent study in German chronic back pain patients with a suspicion of axSpA report on the performance of, among others, the Assessment of Spondyloarthritis international Society (ASAS) IBP criteria and individual IBP parameters. Data showed high sensitivity of the criteria and individual IBP parameters, however specificity was low.6 Important findings like low specificity indicate that in the German rheumatology setting diagnostic utility of IBP criteria was lower than expected. In addition, they commonly find IBP in patients in routine rheumatology care, suggesting that IBP is used as a reference parameter in primary care and that the diagnostic value is almost completely used up in that setting.

The performance of ASAS IBP criteria in other countries is less known. Therefore, this study reports on the diagnostic utility of IBP according to the ASAS criteria and the individual IBP parameters in several rheumatology settings throughout Europe.

Data of the multicentre, observational BelGian inflammatory arthritis and spondylitis (Be-Giant) and international SPondyloArthritis Cause Early (SPACE) cohorts were used. The Be-Giant cohort included ≥18 years old, newly diagnosed axSpA patients who were diagnosed by a team of expert rheumatologists and fulfilled the ASAS axSpA criteria. The SPACE cohort included patients, ≥16 years, with (almost) daily chronic back pain for ≥3 months and ≤2 years, with a symptom onset <45 years. All patients underwent a diagnostic workup according to a predefined protocol of the corresponding cohort, including data on the individual parameters of the ASAS IBP criteria.

AxSpA patients from the SPACE cohort diagnosed by a rheumatologist with a level of confidence (LOC) on the diagnosis of ≥7 on a 0–10 scale were included as ‘axSpA’. All patients in SPACE without axSpA diagnosis (LOC ≥7 for no axSpA) were categorised as ‘no axSpA’. For patients with a LOC <7, diagnosis was considered not sufficiently certain and these patients were excluded from primary analyses. Sensitivity, specificity and positive likelihood ratios (LR+) were calculated using axSpA patients from each cohort separately and no axSpA patients as control group.

Information on IBP parameters was available in 228 patients (Be-Giant) and 559 patients (SPACE) of whom 49.6% and 39.5% were male and mean age was 34.7 (SD 9.7) and 30.7 (SD 8.0) years, respectively. Few axSpA patients had <2 IBP parameters (table 1A; Be-Giant 3.9%, SPACE 9.1%). The individual IBP parameters show high sensitivity (>80%) but low specificity (<36%) with ‘pain at night’ as exception with slightly better specificity. LR+ of IBP were in both cohorts much lower (table 1B) than the LR+ for IBP of 3.1 previously reported as diagnostic estimate in routine practice.7 When including patients with an uncertain diagnosis (LOC <7) as control group, we see similar low specificity as with the stricter diagnosis of no axSpA patients (LOC ≥7).

IBP and individual IBP parameters are commonly seen in patients with and without an axSpA diagnosis. Our data confirm results from Germany, in which similar sensitivity (73.9–88.9) and specificity (22.7–39.5) for IBP according to the ASAS criteria are described.6 These studies show that approximately every second patient seen by a rheumatologist fulfils the IBP criteria but does not have axSpA. The diagnostic utility of IBP criteria in a rheumatology setting in several European countries is lower than previously assumed with a (very) low specificity and LR+. This suggests that the distinctive impact of IBP is almost fully expressed when physicians refer their patient to the rheumatologist.