Development and validation of osteoporosis risk assessment model for Koreans (KORAM)

Abstract

Objective: Currently, dual energy X-ray absorptiometry (DXA) is the gold standard for detecting osteoporosis. However, it is not recommended for screening purpose because of its high cost. Thus, several osteoporosis risk-assessment models have been developed for pre-screening without the added expenses, but their feasibility in the Korean population has not yet been proved. Therefore, this study aims to develop and validate a new model, named the Korean osteoporosis risk-assessment model for postmenopausal women (KORAM-F) and men (KORAM-M) using a nationally representative dataset. Methods: Data from the 2009 Korean National Health and Nutrition Examination Survey (KNHANES) with 1,209 postmenopausal women and 1,340 men 50 years or older were used for the development of KORAM. For validation of KORAM, data including 1,046 postmenopausal women and 1,110 men from the 2010 KNHANES were used. Osteoporosis was defined as T-score ≤ −2.5 at either femoral neck or lumbar spine. To select variables for the model, a 10-fold cross validation was performed. Performance of the candidate models and the Osteoporosis Self assessment Tool for Asian (OSTA) were compared by sensitivity, specificity, and area under the receiver operating characteristics curve (AUC). Model fitness was confirmed using the Hosmer-Lemeshow goodness-of-fit test. To compare KORAM with OSTA, a net reclassification improvement was further calculated. Results: In the development dataset, the prevalence of osteoporosis was 33.9% in women and 8.1% in men. KORAM-F, which consists of age, weight, and hormone therapy, had a sensitivity of 91.2%, a specificity of 50.6%, and an AUC of 0.709 with the specific cut-off score of −9. KORAM-M, based on only age and weight, demonstrated a sensitivity of 90.8%, a specificity of 42.4%, and an AUC of 0.666 with the same cut-off. Additionally, the risk category with KORAM showed an improved reclassification over that of OSTA from 7.4% to 41.8% in women and up to 22.8% in men. Conclusion: KORAM can be easily used as a pre-screening tool to decide who needs a DXA test to detect osteoporosis. Before using KORAM in clinical practice, the studies investigating cost-effectiveness and replicability in other dataset are necessary.