FRI0503 USING BONE MINERAL DENSITY VERSUS THE RATIO OF BODY MASS INDEX TO BONE MINERAL DENSITY TO PREDICT FRACTURE RISK IN HYPERTHYROIDISM

Abstract

Background Euthyroidism is important in the development of a normal skeletal system, with thyroid hormones acting as important regulators of bone homeostasis in adults1. Hyperthyroidism, whether current or previous, increases the risk of developing osteoporotic fractures by stimulating osteoclastic bone resorption and hence bone remodelling, which overall results in decreased bone mineral density (BMD)2. Generally, BMD is used as a predictor of fracture risk; however there has been recent research that suggests using the ratio of BMD to Body Mass Index (BMI) is a better marker of predicting fracture risk in obese patients than BMD alone3. Objectives Our research set out to find whether BMD alone or the ratio of BMI to BMD is a better predictor of fracture risk in patients with current or previous hyperthyroidism. Methods Data were used from a cohort of patients with current or previous hyperthyroidism, referred for DEXA scan to a District General Hospital between June 2004 and October 2010. The following were recorded: age, sex, whether a fracture was sustained, whether they had had steroid therapy at any point, BMI, BMD at L1-L4, BMD at femoral neck (left and right) and BMD at hip (left and right). Logistic regression models were fitted using fracture as the dependent variable. The independent variables for the first set of logistic regression models were BMD at each level and for the second set BMI:BMD ratio at the same levels. Data were adjusted for sex and age at scan. The fit of logistic models were compared using area under the ROC curves (AUC). Results 720 patients were used in the study, of whom 643 (89.3%) were female. Mean age was 63.6 years (SD 11.6) with age range of 28.4 to 89.6 years. 120 (16.7%) were recorded to have had steroid therapy at any point. Mean BMI was 25.9 kg/cm2 (SD 4.77). 274 (38.1%) had sustained a fracture. Odds ratios and AUC values for each level were as shown in the table. The fit of the models using the ratio was generally superior to the fit of the models using BMD alone at the hips and femurs, as AUC values were generally greater for the ratio. At each individual level of the lumbar spine the BMD alone provided a better fit, however overall the ratio gave a slightly better fit. Conclusion This study identifies that the BMI:BMD ratio does not provide better indication of fracture risk than BMD alone in our cohort of patients with current or previous hyperthyroidism. We have previously shown that the same is true for patients with rheumatoid arthritis. A limitation of this study is not stratifying by presence of other diseases or steroid use. Further work will be done to study the role of the ratio in predicting fracture risk in patients with other conditions.