The Impact of Glucose Tolerance States on Bone Mineral Density and Fracture Rate

It is generally acknowledged that fracture rate is higher in diabetic subjects than non-diabetic subjects. However, the impact of diabetes on bone is less clear due to contradictory results of bone mineral density (BMD) and fracture rate. To date, most of reports were based on the studies from relatively small sample sizes. To clarify the issues, we examined the fracture rates and BMD across a spectrum of glucose tolerance in a representative US population. The participants of the National Health and Nutrition Survey 2005–2010 were used in this study. Among adult subjects (age≥20 years old) with reported BMI, we were able to define the states of glucose tolerance in 31,073 subjects cording to the diagnostic criteria based on HbA1c, fasting glucose, and/or 2-h post-changed glucose with established diabetes and using diabetes medications, into normal glucose tolerance (NGT), abnormal glucose tolerance (AGT), and diabetes mellitus (DM). Those who received osteoporosis medications were excluded from BMD analysis. Fracture information was available in 15,547 subjects; validated hip BMD was available in 12,317 subjects; and validated lumbar spine BMD was available in 10,329 subjects. Fracture rates were compared among 3 groups of glucose tolerance states and odds ratio (OR) with 95% confidence intervals (95% CI) were calculated in reference to the NGT group with sample weighting. BMD was compared among 3 groups of glucose tolerance with consideration of covariates. The reported osteoporosis diagnosed rate differed among 3 groups of glucose tolerances (3.99%, 5.77%, and 8.41%, P<0.001, for NGT, AGT, and DM respectively). Worsening states of glucose tolerance were associated increased fracture OR at Hip [AGT, 2.1770 (95% CI: 2.1732–2.1807) and DM, 2.7369 (95% CI: 2.7315–2.7423)], spine [AGT, 0.9924 (95% CI: 0.9912–0.9936); DM, 1.2405 (95% CI: 1.2387–1.2423)]. In contrast, a different trend was observed on the wrist fracture rate [AGT, 0.9556 (95% CI: 0.9551–0.9562); DM, 0.9053 (95% CI: 0.9045–0.9060)]. After adjustment for covariates, higher BMD was noted in AGT and DM when compared to NGT at total femur (NGT, 0.9760±0.0015 gm/cm(2); AGT, 0.9853±0.0021 gm/cm(2); DM 0.9847±0.0034 gm/cm(2), mean±SE, P=0.001) and femoral neck (NGT, 0.8388±0.0015 gm/cm(2); AGT, 0.8474±0.0020 gm/cm(2); DM, 0.8496±0.0032 gm/cm(2), P=0.0007) while no difference was found in lumbar spine BMD (NGT, 0.1.0441±0.0018 gm/cm(2); AGT, 1.0406±0.0025 gm/cm(2); DM, 1.0464±0.0041 gm/cm(2), P=0.35). Our observed significant increased fracture risk at hip (OR: 2.7369) and lumbar spine (OR: 1.2405) in DM subjects when compared to NGT subjects. DM subjects had higher BMD at total femur and femoral neck than NGT subjects while no difference was noted at lumbar spine BMD when compared to NGT subjects. Further studies are required to explore the discrepancy between the increased fracture risk with higher BMDs in diabetes.