This case-control study shows higher values of femoral neck BMD and lower values of LS-TBS in the total population and in women, as well as higher LS-BMD in male T2DM patients, compared to controls. Moreover, LS-TBS was found to have a significant discriminating ability similar to the femoral neck and LS-BMD in differentiating between diabetics and control individuals in the total population and in women.
Previous studies reported higher BMD values in T2DM patients in comparison with healthy individuals, which is in accordance with the results of the current study [6,7,8, 18]. A meta-analysis on 3437 diabetics and 19,139 controls showed that BMD in diabetics was significantly higher at the femoral neck, hip, and LS in both genders [7]. In the present study, femoral neck BMD but not LS-BMD was significantly higher in the total population as well as in women; however, only LS-BMD showed significantly higher values in male T2DM patients. Increased BMD in T2DM patients might be due to the positive correlation between BMD and BMI and the protective effect of obesity against osteoporosis [19]; however, there is still controversy in this regard; in some cohort studies, higher BMD values in T2DM patients persist even after adjustment for BMI [7].
Despite higher BMD values, many studies reported an increased risk of fracture among T2DM patients [2, 3]. Two meta-analyses including 7,832,213 subjects demonstrated a higher risk of hip fractures in diabetic patients in comparison to the general population [1, 2]. Taken together, these findings suggest that diabetes through different mechanisms may have a direct adverse effects on bone strength, which cannot be reflected by BMD alone [11, 13]. A context of low bone turnover, low bone formation, and accumulation of advanced glycation end-products (AGEs) in the bone matrix, inflammation response, oxidative stress, adipokine alterations, Wnt dysregulation, and increased marrow fat and increased bone porosity are reported as plausible causes for bone fragility in T2DM patients [11, 13]. Moreover, low insulin growth factor-1 (IGF-1) levels, a substance responsible for promoting osteoblasts and collagen synthesis, reported in T2DM patients are probably contributing to higher fracture rates [11]. Hence, it remains to be evaluated if and how these factors may interact with bone microarchitecture and bone quality in T2DM patients and whether the effects of these mechanisms on the bone strength could be detected by BMD and TBS.
Recent evidence suggests that TBS determined by using *gray textures of 2D DXA images and evaluating bone trabecular microarchitecture, a component of bone quality, might be a more accurate determining factor of bone strength and fracture risk, independent of BMD, in T2DM patients [6]. There are some studies that reported significantly lower TBS levels, despite higher mean BMD levels, in T2DM patients than non-diabetics [16], a finding in line with the results of the present study and also confirmed in larger studies such as a study on 29,407 postmenopausal women, including 2356 DM patients in Manitoba, Canada [6], and the Ansung study on 1229 men and 1529 postmenopausal women in Korea (325 men and 370 women with T2DM) [8].
However, there is evidence showing higher TBS values among T2DM patients than controls [20] as well as evidence showing no significant difference in this regard, despite significant higher values of BMD among diabetics, for example, a case-control study by Zhukouskaya et al. on 99 diabetic patients and 107 controls showing that TBS values were not significantly different among DM patients and controls; however, it was lower in patients with DM who suffered fractures [21]. Another population-based study among the Iranian population consisting of 2263 participants aged 60 years and above also showed that mean TBS values were not significantly different among diabetics and normoglycemics [17].
Moreover, we found a gender dissimilarity regarding LS-TBS differences between T2DM patients and controls. A significant difference between LS-TBS values of the two groups was observed only among women, similar to a study by Rianon et al., which was conducted on 153 men and women [22]. The observed gender difference may be due to the relatively small number of men in our sample. However, a recent meta-analysis conducted on 35,546 women and 4962 men reported that patients with diabetes and prediabetes had significantly lower TBS than non-diabetics; however, the difference was greater in women than in men [23]. In addition, a large-scale study among DM and non-diabetic men aged 65 years and older showed no significant association between DM and increased prevalence or incidence of vertebral fracture [24]. This suggests that the relation between DM and fracture risk might be more prominent in women than in men.
Despite the growing body of evidence assessing TBS values in diabetic patients, there are limited studies regarding the ability of TBS vs. BMD in differentiating diabetic patients and control individuals. A study by Xue et al. on 477 diabetic and non-diabetic subjects showed that LS-TBS was more effective in discriminating diabetic patients from control subjects than BMD [16]. In the present study, LS-TBS has been shown to have a significant ability in differentiating between T2DM patients and controls in the total population as well as in women based on AUC and ROC analysis, similar to the femoral neck and LS-BMD in this group (adjusted for age and BMI). In contrast to LS-BMD, the ability of LS-TBS in distinguishing between T2DM patients and controls remains significant after adjustment for age and BMI, which might show that TBS is less affected by age and BMI, which are possible confounding factors in assessing bone characteristics [25,26,27], although more studies are needed to investigate this notion. A possible explanation for the lower ability of LS-BMD in distinguishing T2DM women and controls before adjustment for age and BMI might be the effect of degenerative effects of age on the LS, which had been shown previously to affect osteoporosis diagnosis in elderly women [28]. In men, LS-BMD was the only acceptable method with a relatively good ability in distinguishing between T2DM patients and controls. The comparatively smaller sample size in the male population may be responsible for the observed differences between the two genders. These findings are consistent with studies that considered BMI as a potential confounder in the relationship between T2DM and microarchitectural abnormalities through different mechanisms, including lower bone formation, elevated serum sclerostin, increased adipocyte markers, and abnormal bone marrow fat composition [25,26,27].
There are some strengths of the present study, including having a population-based design with a relatively large number of study participants and assessment of LS-TBS and femoral neck and LS-BMD at the same time. Moreover, in the present study, the type of diabetes that can affect bone health through different pathophysiology and therapeutic interventions [13] was considered, and all the diabetic participants were suffering from T2DM. Another strength of the study was adjustments for age and BMI, which are important confounding factors in the bone assessment. However, the present study has several noteworthy limitations. First, the small number of men included in the study limited the power of the analysis. Second, we did not have information regarding prevalent fracture, glucocorticoid consumption, and the duration and severity of diabetes of the participants or medication used by them, which have been demonstrated to affect bone strength in such patients [13]. Third, the diagnosis of DM in the study participants was based on their history, and no laboratory assessment was done confirming the DM diagnosis. Fourth, while it is believed that menopausal status is a possible confounder in assessing bone characteristics, due to lack of data, we were unable to include it in our adjusted model. Finally, considering the cross-sectional design of the study, we did not test the association of LS-TBS and BMD values in T2DM patients and non-diabetic subjects with further outcomes such as fracture, and more longitudinal studies are needed in this area.