Impact of the Change in the Bone Densitometer Reference Database on the Prevalence of Osteoporosis and the Identification of Patients with Vertebral Fractures.

Objective: The aim of our study was to evaluate the impact of the change in the bone densitometerreference database on the prevalence of osteoporosis and the identification of vertebral fractures (VFs) in postmenopausal Moroccan women. Methods: 328 post-menopausal women were recruited. Bone mineral density (BMD) and lateral Vertebral Fracture Assessment (VFA) images were obtained using a Bone densitometer. VFs were defined using a combination of Genant semi quantitative approach and morphometry. The discriminative performance of Morocco, Europe, Middle East and USA reference database to identify osteoporosis and prevalent VFs was assessed by evaluating the areas under the ROC curve (AUC), and calculating sensitivity, specificity, positive, negative predictive value and positive likelihood ratio. A stepwise conditional binary regression analysis was reported in the order to assess the risk of occurrence of vertebral fractures for all bone reference databases. Results: the mean age was 65 ± 6.5 (50 to 84) years. Vertebral fractures were identified in 84 (25.6%). Using the Moroccan database, 144 women (43.9%) were identified with osteoporosis at any site. According the European, Middle Eastern and American reference databases, osteoporosis prevalence at any site was 138 (42%), 137 (41.7%) and 132 (40.2%); respectively (p=0.01). The highest AUC was observed for Moroccan lumbar spine reference database which had also the best sensitivity and specificity balance (65.5% - 63.8%) and the highest positive predictive value and the positive likelihood ratio (38.5-1.8). The stepwise conditional binary regression analysis showed that Moroccan (OR [CI 95%] = 5.8 [1.2 – 27.8]) and, at a lower odds, European (OR [CI 95%] = 2.7 [1.3 – 5.7]) database were independently able to predict prevalent vertebral fractures. Conclusion: The Moroccan reference database with the elevated prevalence of osteoporosis noted can better identify subjects at highest risk of osteoporotic VFs.