Abstract
Introduction
The main challenge in management of adolescent idiopathic scoliosis (AIS) is to predict which curve will progress so that appropriate treatment can be given. We previously reported that low bone mineral density (BMD) was one of the adverse prognostic factors for AIS. With advancement in imaging technology, quantitative ultrasound (QUS) becomes a useful method to assess bone density and bone quality. The objective of this study was to assess the role of QUS as a radiation-free method to predict curve progression in AIS.
Methods
294 girls with AIS were recruited at ages 11–16 years and followed up until skeletal maturity. 269 age-matched healthy girls were recruited as controls. They provided the normal reference for calculation of Z score for QUS parameters. QUS measurements, including BUA (broadband ultrasound attenuation), VOS (velocity of sound) and SI (stiffness index) of the calcaneum, BMD of femoral neck, menarche history, ages, and Cobb angle of the major curve were recorded at baseline as independent variables. The predictive outcome was curve progression defined as an increase of Cobb angle of 6° or more. Logistic regression model and the ROC curve were used for statistical analysis.
Results
Mean follow-up was 3·4 years (SD 1·57). At baseline, mean age was 13·4 years (1·23), 73 (24·8%) patients were premenarchal, and mean Cobb angle was 26·3° (SD 8·2°). 202 (68·7%), 194 (66%), and 202 (68·7%) of patients with AIS had Z score of BUA, VOS, and SI of 0 or less, respectively. Initial univariate analysis indicated all independent variables had p values less than 0.2. Logistic regression analysis indicated that the p values of their regression coefficients were: age (p<0·001), menarchal status (p<0·001), Cobb angle (p=0·008), BMD (p=0·084), BUA (p=0·722), VOS (p=0·112), and SI (p=0·027). SI, age, menarchal status, and Cobb angle were therefore included in the final prediction equation. The adjusted odds ratio for Z score of SI of 0 or less was 2·00 (95% CI 1·08–3·71). The area under the ROC curve was 0·831(95% CI 0·785–0·877). The predictive model had a sensitivity of 0·847 and a specificity of 0·665 at a probability cutoff of 0·368.
Conclusions
We recorded evidence of deranged bone density and bone quality in AIS, as indicated by QUS investigation. SI is an independent and significant prognostic factor for AIS. It can be used as a radiation-free parameter to predict curve progression in combination with initial Cobb angle, age, and menarchal status, especially when DXA is not available.
Acknowledgments
This study is supported by Research Grant Council—The government of HKSAR (project number CUHK4498/06M).
