• spondylitis, ankylosing
• magnetic resonance imaging
• low back pain

We thank Dr Zanwar for his interest in our study. We appreciate that he took the time to write down his remarks1 regarding our work.2 We truly believe that overdiagnosis of axial spondyloarthritis (axSpA) is an issue, since the specificity of MRI for sacroiliitis seems to be overestimated.3–5 By including asymptomatic subjects with an uncomplicated pregnancy and childbirth, we demonstrated that pregnancy and giving birth are associated with the occurrence of spondyloarthritis (SpA)-like sacroiliac joint lesions. By using this strategy, we limited the possibility of erroneously including patients with SpA, whereas if all study subjects would have had postpartum chronic back pain, the distinction with patients with SpA would have been difficult to make. The present study design limited the possibility that the detected sacroiliac joint lesions can be attributed to a back pain-related pathology. By demonstrating important sacroiliitis-like images even in healthy, asymptomatic postpartum subjects, we showed that caution is truly warranted in interpreting MRI images of the sacroiliac joints in the postpartum period. Consequently, performing an MRI in the postpartum period should be a well-considered decision. Therefore, we want to underscore that if an MRI of the sacroiliac joints is performed during the first 6 months postpartum, the possibility of pregnancy/childbirth-associated sacroiliac joint lesions should be seriously considered, as these are difficult to distinguish from sacroiliitis in the context of SpA. When in doubt, a treatment with non-steroidal anti-inflammatory drugs can be attempted to evaluate the treatment response and to buy some time before performing an additional MRI scan.

The link of the imaging findings with the presence or absence of back pain was another point touched on. For clarity, Assessment of Spondyloarthritis International Society classification criteria for axSpA do not mandate presence of inflammatory back pain (IBP) for more than 3 months before proceeding to MRI of the sacroiliac joints.6 7 Furthermore, IBP has been shown to have high sensitivity but low specificity.8 As in our study only four subjects had IBP, it was not possible to make hard statements regarding this matter. Three out of four subjects with IBP had sacroiliitis at baseline, persisting in two out of four at month 6. The residual sacroiliitis at month 12 was limited in both subjects. The decline in Spondyloarthritis Research Consortium of Canada (SPARCC) scores over time in these subjects suggests that these lesions are pregnancy/childbirth-related and are not attributed to SpA. Inflammatory serum markers were not regarded in this study since they were not available for all subjects and they are heavily influenced by the process of childbirth itself and are therefore presumably not directly linked to the presence of sacroiliac joint MRI lesions.