Evaluation of the predictive validity of the ASAS axial spondyloarthritis criteria in the DESIR cohort

Objective To evaluate the predictive validity of the Assessment of SpondyloArthritis international Society (ASAS) axial spondyloarthritis (axSpA), Amor, European Spondylarthropathy Study Group (ESSG) and modified New York (mNY) classification criteria. Methods Patients from the DESIR cohort (inflammatory back pain suggestive of axSpA for >3 months but <3 years duration), followed for up to 5 years. Positive predictive value (PPV) of the set of criteria collected at baseline (ASAS, and its arms, Amor, ESSG and mNY: fulfilled/not fulfilled) were tested against the rheumatologist's axSpA diagnosis (fulfilled/not fulfilled) after 5 years of follow-up. Results In total, among the 708 patients included in the DESIR cohort at baseline, data on rheumatologist's diagnosis at 5 years was available in 411 patients; amongst them, 352 (85.6%) had an axSpA diagnosis according to the rheumatologist; 268 patients fulfilled the ASAS axial SpA (axSpA) criteria at baseline and of these, 245 were diagnosed as SpA after 5 years follow-up (PPV: 91%). The PPV of the ASAS "imaging" arm and "clinical" arm was 97% and 82%, respectively. Other criteria also showed similar PPV - Amor (91%), ESSG (90%) and mNY (99%). Conclusion Positive predictive validity of the ASAS criteria for axSpA (including both arms) at 5 years was excellent; it is worth noting that the performances of the other criteria were also very good in the DESIR cohort.