Op0222 Sacroiliac Bone Marrow Oedema on Postpartum Mri Does Not Develop Into Structural Lesions Over a 5-Year Follow-Up Period

Background: Sacroiliac bone marrow oedema (BMO) on MRI, immediately after childbirth, has been observed in a large proportion of postpartum women and appears to be a transient phenomenon. Although generally attributed to mechanical stress, these postpartum BMO lesions bear a striking resemblance to the inflammatory lesions seen in axial spondyloarthritis (axSpA). In axSpA, active inflammatory lesions are likely to progress to structural lesions over time, particularly fat lesions and erosions. It is not known whether the same is true for mechanical stress-induced BMO. Objectives: To assess the long-term evolution of postpartum BMO and the potential development of structural lesions, which are also often reported in patients with axSpA. Methods: A 5-year follow-up evaluation of the prospective study conducted by Renson et al was performed on sacroiliac MRI lesions in healthy women following an uncomplicated vaginal delivery. 1 Nineteen (54.3%) of the original 35 participants underwent a new MRI of the sacroiliac joints (MRI-SIJ) approximately 5 years after initial inclusion in the study. Structural lesions (fat lesions, erosions, sclerosis, (partial) ankylosis) and inflammatory lesions were assessed by three well-trained, calibrated readers using a method similar to the Spondyloarthritis Research Consortium of Canada (SPARCC) scoring system (NH, MdH, GV). With this scoring method individual lesions were assessed on their presence/absence in four quadrants per SI joint on 6 consecutive slices covering the cartilaginous part of the joint. Baseline MRI images were re-evaluated together with the 5-year follow-up MRI data. Demographic, clinical, and obstetric data were reported by mean with standard deviation or n with percentages. Imaging data was reported as minimum, first quartile, median; third quartile and maximum number of lesions in women with at least one structural lesion on MRI. Results: The baseline characteristics of the 5-year follow-up subjects and the dropouts were similar between groups except for weight, body mass index (BMI) and presence of inflammatory back pain according to the ASAS criteria (Table 1). No significant association was found between inflammatory back pain and the total SPARCC score (p = 0.24) nor the prevalence of a 'positive MRI' according to the ASAS definition at baseline (p = 0.64). The overall prevalence of fat lesions, erosions and sclerosis on baseline and 5-year MRI images was low. No ankylosis was observed. Scores for fat lesions and erosions did not change significantly between the two time points. A numerical increase in sclerosis at year 5 was seen in only 3 of the 19 participants, all of whom had BMO after delivery (Table 2). As expected, there was a significant decrease in BMO from baseline to year 5. Six of the 19 participants (32%) still had BMO at the 5-year follow-up MRI-SIJ, 3 of whom met the ASAS definition of sacroiliitis. Having another delivery during the follow-up period (3 out of 6 subjects) did not affect the presence of BMO at 5-year follow-up. No deep or intense BMO lesions were observed. Conclusion: In postpartum women, no significant development of structural MRI lesions was observed 5 years after delivery. Nevertheless, BMO remained prevalent. These results support the hypothesis that transient mechanical stress-induced sacroiliac BMO on MRI does not seem to evolve into structural lesions over a follow-up time of 5 years. REFERENCES: 1 Renson T., et al. Ann Rheum Dis. 2020;79(7):929-34. Acknowledgements: NIL. Disclosure of Interests: Liesbet De Meester: None declared, Gaëlle Varkas No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract, Manouk de Hooge: None declared, Ann-Sophie De Craemer: None declared, Nele Herregods: None declared, Lennart Jans: None declared, Phillippe Carron: None declared, Dirk Elewaut No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract, Filip van den Bosch No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract, No relevant disclosures for the purpose of this abstract.