Questioning the role of lumbar BMD in vertebral fragility among children with chronic non-bacterial osteomyelitis

To the Editor,

We thank the authors for their valuable contribution and Therapeutic Advances in Musculoskeletal Disease for publishing this important study on vertebral fragility in chronic non-bacterial osteomyelitis (CNO). The work by Chighizola et al. ¹ highlights an underexplored aspect of pediatric inflammatory bone disease. We would, however, like to offer a methodological observation regarding the interpretation of bone mineral density (BMD) data that may affect conclusions about the relationship between BMD and vertebral fractures (VFs).

In the study, lumbar-spine BMD was assessed by dual-energy X-ray absorptiometry (DXA) and automatically converted to age- and sex-matched Z-scores. While this approach follows standard manufacturer algorithms, no adjustment for height or body size is described. In pediatric patients, DXA measures areal BMD (g/cm²), which is strongly influenced by bone size. Children with shorter stature or growth delay—features frequently observed in chronic inflammatory conditions—may therefore show artificially low BMD values despite normal volumetric bone density and bone strength. Current ISCD Pediatric Official Positions explicitly recommend height-adjusted Z-scores or alternative size-corrected measures, such as bone mineral apparent density, to avoid this well-recognized artifact.^2–4

Without appropriate height adjustment, the reported lack of difference in lumbar BMD between patients with and without VFs (−1.7 vs −0.5; p = 0.228) cannot be reliably interpreted as evidence against an association between bone density and fracture risk. Moreover, ISCD guidance emphasizes that the presence of vertebral compression fractures in children establishes osteoporosis irrespective of BMD values, indicating that BMD should not be used as an explanatory determinant of fracture occurrence. Although the authors appropriately acknowledge several limitations, the absence of size-adjusted BMD remains an important methodological consideration.

These observations are further supported by broader evidence. Multicenter and systematic studies have not demonstrated a consistent association between BMD and vertebral involvement in CNO. Roderick et al. ⁵ identified VFs in 8% of patients without correlation to BMD, while Hofmann et al. ⁶ reported vertebral lesions in 10%–15% of cases and suggested that fractures may arise from localized inflammatory bone remodeling rather than systemic osteopenia.

Finally, while bisphosphonates are suggested as potentially beneficial for vertebral outcomes, current evidence remains inconclusive. Schnabel et al. ⁷ showed improvements in pain and MRI inflammation with bisphosphonate therapy, but did not demonstrate a reduction in fracture risk.

These comments are intended not to diminish the value of the study, but to encourage careful adherence to pediatric densitometry standards when interpreting DXA findings in CNO. Addressing these methodological issues in ISCD-concordant, multicenter studies will be essential to clarify whether vertebral fragility in CNO reflects systemic bone loss or localized inflammatory processes.