Objectives: To analyse the predictive values of inflammatory back pain (IBP), positive HLA B27 antigen, increased C-reactive protein (CRP), Spondyloarthritis (SpA) features, familial history (FH), magnetic resonance sacroiliac joints (MRI-SIJ) imaging and its weight in early SpA diagnosis.
Methods: 133 patients with back pain, aged <50, duration of the pain <2 years were included. Data such as IBP, HLA B27, increased CRP, SpA features, FH, SIJ´s radiography and MRI were collected for each patient. STIR sequences were classified as strongly positive bone morrow oedema (SPBME ≥2), clearly present and easily recognisable as positive according to the ASAS criterion, weakly positive (WPBME ≥2), suggestive, but not easily recognisable and, clearly negative none of those features. T1-weighted sequences were assessed as positive/negative for erosion, fat metaplasia, backfill and sclerosis, if ≥1, for each lesion was present. MRI images were read by three blinded readers.
Results: The average age was 38.9 years. 47 (35.3%) patients received SpA diagnosis according to the clinical opinion. IBP was highly specific, 0.81 and sensitive, 0.83. HLA B27 was positive in a half of the SpA patients. SPBME ≥2 provided a great specificity, 0.94 and an acceptable sensitivity, 0.79. Erosion was significantly more frequent in SpA patients (72% vs 7%), specificity 0.93. The addition of erosion ≥1 to the WPBME ≥2 noticeably improved specificity, 0.98, although slightly decreased sensitivity, 0.64. Fat metaplasia and backfill were highly specific, but poorly sensitive. Factors forecasting positive diagnosis were IBP, followed by SpA features and increased CRP.
Conclusions: At the onset, IBP might be a good marker for selecting patients with suspicion of SpA. The addition of erosion to the ASAS criterion might be helpful for early diagnosis, especially in patients with doubtful STIR imaging where BME is present but it is hard to determinate whether the ASAS "highly suggestive" criterion is met.