Vertebral fractures assessment in children: Evaluation of DXA imaging versus conventional spine radiography.

Vertebral fracture assessment (VFA) by DXA is an accepted tool in adults. However, its use in children has not been assessed. The aim of this study was to evaluate DXA VFA and morphometric analysis (MXA) using a GE Lunar iDXA bone densitometer against spinal radiographic assessment (RA) for the identification of vertebral fractures in children. Spine RA and VFA (T3-L5) were acquired on the same day in 80 children. Forty children considered high risk for fracture by their metabolic bone specialist were referred for spinal RA. Another 40 children were recruited as part of a prospective fracture study and were considered low risk for vertebral fracture. Agreement between RA and VFA was assessed by an expert paediatric radiologist and two paediatricians with expertise in bone pathology. Agreement between RA and MXA was assessed by an expert paediatric radiologist, two clinical scientists and an experienced paediatric radiographer. Vertebrae were ranked as normal, mild, moderate or severe if they had <10%, 11-25%, 26-50% and >50% deformity, respectively. Levels of agreement were calculated using the Cohen kappa score. Evaluating the data from all readable vertebrae, 121 mild, 44 moderate and 16 severe vertebral fractures were identified; with 26, 8, and 5 subjects having at least one mild, moderate or severe fracture, respectively. Depending on rater, 92.8-94.8% of the vertebrae were evaluable by RA. In contrast, 98.4% were evaluable by VFA and only 83.6% were evaluable by MXA. Moderate agreement was found between raters for RA [kappa 0.526-0.592], and VFA [kappa 0.601-0.658] and between RA and VFA [kappa 0.630-0.687]. In contrast, only slight agreement was noted between raters for MXA [kappa 0.361-0.406] and between VFA and MXA [kappa 0.137-0.325]. Agreement substantially improved if the deformities were dichotomised as normal or mild versus moderate or severe [kappa 0.826-0.834]. For the detection of moderate and/or severe fractures the sensitivities & specificities were 81.3% & 99.3%, and 62.5% & 99.2% for VFA and MXA, respectively. This study demonstrates that VFA is as good as RA for detecting moderate and severe vertebral fractures. Given the significant radiation dose saving of VFA compared with RA, VFA is recommended as a diagnostic tool for the assessment of moderate or severe vertebral fracture in children.

Impact of UK National Guidelines based on FRAX®--comparison with current clinical practice.

OBJECTIVE: To assess whether clinician-determined treatment intervention thresholds are in line with the assessment of fracture risk provided by FRAX® and treatment recommendations provided by UK guidelines produced by the National Osteoporosis Guidelines Group (NOGG). DESIGN, PATIENTS AND MEASUREMENTS: This was a retrospective cohort analysis of 288 patients consecutively referred for dual-energy X-ray absorptiometry (DXA) scanning from primary care immediately prior to the introduction of the FRAX® algorithm. In addition to DXA assessment, patients completed a clinical risk factor questionnaire which included risk factors used in the FRAX® algorithm. Initial risk assessment and treatment decisions were performed after DXA. FRAX® was used, retrospectively, with femoral neck T-score, to estimate fracture risk which was applied to NOGG to generate guidance on treatment intervention. Clinician- and NOGG-determined outcomes were audited for concordance. RESULTS: There was concordance between clinician and NOGG treatment decisions in 215 (74.6%) subjects. Discordance was observed in 73 (25.3%) subjects. In the discordant group, seven subjects were given lifestyle advice when NOGG recommended treatment, 42 given treatment when NOGG recommended lifestyle advice only, and 24 were referred to a metabolic bone clinic for further evaluation. The reasons for treatment differences in subjects recommended treatment by clinician but not NOGG were largely (90.2%) attributed to the use of lumbar spine bone mineral density (BMD). CONCLUSIONS: There is high concordance between clinician-determined and FRAX®-NOGG intervention. The absence of spine BMD from FRAX® is the primary source of discrepancy. This study provides some assurance of the validity of the treatment thresholds generated from FRAX®-NOGG in 'real-world' usage.