Vertebral Fractures in children with chronic inflammatory and/or disabling conditions: the SNAP study

Nicola Crabtree; Wolfgang Hogler; Dee Chapman; Jacky Walford; Nicholas Shaw

doi:10.1530/boneabs.6.OC17

Objectives: The SNAP study is a prospective fracture study of children with chronic inflammatory and/or disabling conditions. The overall study aim is to assess causal links between body-size adjusted bone density and low trauma fracture.

Methods: 330 children aged 5–18 years were recruited from seven disease groups namely; acute lymphoblastic leukaemia (ALL), rheumatological disease, inflammatory bowel disease, cystic fibrosis, coeliac disease, Duchenne muscular dystrophy (DMD) and cerebral palsy. At baseline, bone density by DXA (lumbar spine (LS BMAD) and total body less head (TBLH BMD)), forearm pQCT (trabecular density at the 4% site (Trab vBMD), and ratio of bone/muscle area at the 66% site [Radius BA/MA], hand radiographs (Bone health index (BHI), BoneXpert), lateral spinal radiographs and medical history were assessed. A threshold of Z-score<−2.0 was set to dichotomise the bone density Z-scores and used in conjunction with a binary prediction model to assess diagnostic accuracy.

Results: Spinal radiographs identified 71 children (21.5%) with vertebral fractures, with highest incidence for children with ALL (26/51) and DMD (14/42) (P<0.001). Steroid exposure, back pain and immobility were reported in 50, 37 and 14% of patients, respectively. Bone density Z-scores were significantly lower in the fracture group for LS BMAD, Trab vBMD, Radius BA/MA, and BHI. The variables most predictive of vertebral fracture were Trab vBMD, BHI and BA:MA (P<0.05), with corticosteroid exposure and back pain also significant (Table 1).

Table 1
	Odds Radio (96%CI)	Sensitivity (96%CI)	Specificity (96%CI)
Trab vBMD	3.4 (1.6–7.0)	0.29 (0.17–0.41)	0.89 (0.85–0.93)
Hand BHI	2.8 (1.6–4.9)	0.45 (0.33–0.57)	0.77 (0.71–0.82)
Radius BA/MA	2.3 (1.0–5.3)	0.20 (0.09–0.32)	0.90 0.86–0.94)
Corticosteroids	1.7 (1.0–2.9)	0.61 (0.49–0.72)	0.53 (0.47–0.59)
Back Pain	1.7 (1.0–2.8)	0.46 (0.35–0.58)	0.66 (0.60–0.71)
TBLH BMD	1.4 (0.8–2.6)	0.27 (0.17–0.38)	0.79 (0.75–0.84)
LS BMAD	1.3 (0.5–3.2)	0.10 (0.03–0.17)	0.92 (0.89–0.95)
Immobility	1.0 (0.5–2.1)	0.14 (0.06–0.22)	0.86 (0.81–0.90)

Conclusion: Disease itself, back pain and corticosteroid exposure are significantly associated with risk of vertebral fracture. However, the variables most predictive of vertebral fracture were low trabecular density measured by pQCT and BHI by BoneXpert. Evidence of the predictive power of these measurements will only be confirmed with future follow-up of this group. Funded by NIHR Clinical Development Fellowship (HCS/P10/009).

Disclosure: The authors declared no competing interests.

Bone Abstracts

OC17