Diabetes Mellitus and Trabecular Bone Score (TBS) in Caucasians and African Americans Referred for Bone Densitometry

Presentation Number: SUN 298
Date of Presentation: April 2nd, 2017

Rajesh K Jain*1, Meltem Zeytinoglu1 and Tamara J. Vokes2
1University of Chicago, Chicago, IL, 2University of Chicago Medicine, Chicago, IL

Abstract

Background: Patients with diabetes mellitus (DM) have been shown to have higher fracture rates than expected from their usually normal bone mineral density (BMD). Trabecular bone score (TBS), a textural analysis of the lumbar spine DEXA image, is an indirect measure of bone microarchitecture that has been shown to predict fractures independently of BMD. Interestingly, patients with DM have lower TBS than those without DM, suggesting that TBS may be capturing the increased fracture risk in DM. However, studies to date have been done in Caucasians (CA) and Asians, while the effect of DM on TBS is presently unknown in African Americans (AA), a population with a significant burden of DM. We have previously reported that among all patients referred for densitometry, TBS was significantly lower in AA than CA. One possibility is that AA have lower TBS than CA due to higher rates of DM and/or a stronger effect of DM on TBS in AA. Thus, we sought to examine whether the effect of DM on TBS differed between AA and CA and whether poor glycemic control resulted in worse TBS.

Methods: This is a retrospective review of all women who had undergone DEXA scan at our facility as part of their clinical care during years for which TBS can be derived (2011-2016). DM was defined as having an A1c of at least 6.5% within 2 years before or 90 days after DEXA scan. Only patients with BMI 15-37 were included as that is the working range for TBS.

Results: There were 341 women with DM out of 3187 (10.7%) who had undergone DEXA. Prevalence of diabetes was higher in AA (293 out of 1623, 18.1%) than in CA (only 48 out of 1564, 3.1%, p<0.001). BMD T-scores (lowest of lumbar spine, femoral neck or total hip) were higher in AA than CA (-1.5 ± 1.2 vs. -1.9 ± 1.0, p<0.001) and higher in those with DM than those without DM (-1.4 ± 1.2 vs. -1.7 ± 1.1, p<0.001). In both ethnicities, TBS was approximately ¼ standard deviation lower in those with DM than in those without DM (1.231 ± 0.149 vs. 1.273 ± 0.125, p=0.02 for CA and 1.188 ± 0.151 vs. 1.216 ± 0.140, p=0.002 for AA). In a multivariate model that included age, BMI, and spine BMD, the effect of DM was similar in AA and CA (p=0.84). Regardless of DM status, AA had lower TBS than CA (p<0.001 without DM, p=0.07 in DM).

We then examined whether the level of glycemic control was associated with TBS. The average A1c in the AA and CA patients with diabetes was similar (7.5% ± 1.1% vs. 7.4% ± 1.1%, p=0.58). In a multivariate model that included age, BMI, and spine BMD, TBS was negatively associated with A1c (reduction in TBS of 0.016 for each percent increase in A1c, p=0.02) without significant ethnic differences (p=0.67).

Conclusion: Though AA had lower TBS regardless of DM status, both AA and CA patients with DM and poor glycemic control had similar reductions in TBS. Thus, TBS may be an indicator of compromised bone quality in all patients with DM, regardless of ethnicity, and could serve as an important tool for assessing fracture risk in diabetes.

 

Disclosure: TJV: Advisory Group Member, NPS-Shire. Nothing to Disclose: RKJ, MZ