Introduction

The prevalence of type 2 diabetes mellitus (T2DM) is increasing worldwide. While traditional diabetic complications, such as retinopathy, nephropathy, neuropathy, and cardiovascular diseases, are well established, skeletal fragility is becoming an emergent diabetic complication1. Notably, patients with T2DM have higher rates of morbidity following fracture than Non-DM patients with fractures2. As a result of the increased life expectancy of T2DM patients, fragility fractures in these patients have become important in elderly populations.

Despite an apparent increase in fracture risk in patients with T2DM, most studies have shown that, compared with Non-DM controls, patients with T2DM have preserved, or even increased, areal bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA)3,4,5. These findings indicate that bone quality plays a crucial role in affecting skeletal fragility in T2DM patients. Although the etiology of fracture risk in T2DM patients is multifactorial, increased accumulation of advanced glycation end products (AGEs) contributes to impaired bone material properties and microvascular disease, resulting in increased cortical porosity6, has emerged as a key skeletal abnormality that contributes to skeletal fragility in T2DM patients. Furthermore, T2DM patients have lower bone formation indices, increasing their risk of fracture. Additionally, peripheral neuropathy, hypoglycemia and a greater risk of falls contribute to these individuals’ greater fracture risk, despite their relatively preserved BMD1. However, BMD continues to predict fracture risk in T2DM patients, although the fracture risk assessment tool (FRAX) algorithm frequently underestimates fracture risk in these patients5.

The majority of this evidence is obtained from the Caucasian population. Recently, Asian countries have presented contradictory information indicating that BMD is significantly variable in individuals with T2DM. Several studies have shown that individuals with T2DM have lower BMD than healthy individuals7,8. The data in Southeast Asia, particularly in Thailand, have the potential to be more comprehensive. The aim of this study was to determine and compare the prevalence of vertebral fractures and osteoporosis, as well as BMD and the FRAX score, between individuals with and without T2DM and assess the association of these factors with vertebral fractures.

Results

Prevalence of vertebral fractures

A total of 435 participants were enrolled, including 145 postmenopausal women with T2DM and 290 non-DM counterparts. The mean duration of diabetes was 8.7 ± 5.4 years. The mean age and body weight were comparable between those with and without T2DM; however, T2DM patients had a higher mean BMI (25.2 kg/m² ± 4.1) than non-DM individuals (24.1 kg/m² ± 3.4, p = 0.006). BMD at the FN and TH was significantly greater in T2DM patients than in non-DM controls (Table 1), but there was no significant difference in the FRAX score, including the 10-year probability of hip fracture or MOF, either with or without BMD. The 10-year probabilities of hip fracture and major osteoporotic fracture (MOF), both with and without bone mineral density (BMD) assessment via FRAX, did not significantly differ between the DM and non-DM groups; however, lower FRAX scores for MOF with and without BMD, as well as for hip fracture with and without BMD, were observed.

Table 1 Characteristics of participants stratified by diabetes status.
Full size table

When specific bone regions were assessed, individuals with T2DM had a lower prevalence of osteoporosis at the FN than non-DM individuals (OR = 0.56, 95% CI: 0.34 to 0.93, p = 0.029); however, there were no significant differences in the prevalence of osteoporosis at the lumbar spine or total hip. The presence of vertebral fractures, especially those involving two or more levels, was significantly greater in individuals with T2DM than those without T2DM (OR = 3.83, 95% CI: 1.77 to 8.28, p = 0.001) (Table 2).

Table 2 Comparison of the prevalence of FRAX scores above the cutoff threshold, osteoporosis, and vertebral fractures between T2DM patients and non-DM patients.
Full size table

Risk factors of vertebral fractures

According to the univariate analysis, there was a significant association between age and vertebral fracture (OR: 1.03, 95% CI: 1.01–1.05, p = 0.022). However, no significant associations were found between BMI, LSBMD, FNBMD, FRAX score for MOF and hip fracture, diabetes, and vertebral fractures. However, diabetes was significantly associated with vertebral fracture \(\:\ge\:\)2 levels, whereas other factors, including age, BMI, LSBMD or FNBMD, and the FRAX score for both MOF and hip fractures, were not significantly associated with vertebral fracture \(\:\ge\:\)2 levels. In the multivariate analysis, we observed that the association between diabetes and vertebral fracture \(\:\ge\:\)2 levels was unchanged after adjusting for other clinical variables (adjusted OR = 3.72 (1.70 to 8.15), p = 0.001) (Table 3).

Table 3 Univariate and multivariate analyses of clinical factors associated with vertebral fractures.
Full size table

Discussion

Disparities in BMD between postmenopausal women with T2DM and non-DM controls were observed in this study. We found that compared with non-DM individuals, postmenopausal women with T2DM have significantly greater femoral neck and total hip BMD. T2DM patients had modestly greater lumbar spine BMD than non-DM controls. However, the FRAX score, including the 10-year probability of hip fracture or major osteoporotic fracture (MOF), was not significantly different between individuals with and without T2DM. Non-DM women have a higher prevalence of osteoporosis at the femoral neck than at the lumbar spine. Individuals with T2DM have a higher rate of vertebral fractures, particularly those involving two or more levels. Diabetes is significantly associated with two or more levels of vertebral fracture, even after controlling for other clinical factors.

The findings of this study are consistent with those of previous reports from the Women’s Health Initiative; women with T2DM had higher BMD at the lumbar spine and hip at baseline and over nine years of follow-up9. Furthermore, a meta-analysis of women and men with T2DM revealed higher Z-scores of 0.41 at the spine and 0.27 at the hip in patients with T2DM than non-DM individuals4. A recent Thai study including 97 individuals with T2DM (37 men and 60 women) and 342 individuals without T2DM (174 men and 168 women) revealed that T2DM was associated with increased LSBMD after adjusting for BMI in men but not women10. However, our findings differ from those of a prior study from India, which reported lower lumbar spine and hip BMD in T2DM patients than in controls7,8. We cannot conclude the reasons for the discrepancy in the results. Different stages of diabetes, on the other hand, might be one possible explanation. In the early stages of T2DM, hyperinsulinemia is likely responsible for the preserved or improved BMD and trabecular bone parameters11,12,13. In addition, increased BMD in T2DM patients has been reported in several studies because of the osteoanabolic effect of insulin14,15. Notably, the effect of hyperinsulinemia on bone is independent of adiposity16. However, the early stage advantage in trabecular bone properties is not maintained in the late stage of T2DM, when β-cell failure develops, and hyperinsulinemia decreases13. Furthermore, T2DM may have a protective effect on bone in individuals with a high BMI and excess body fat due to the increased mechanical load, resulting in higher BMD in these individuals than in non-DM individuals. The presence of hyperinsulinemia and obesity may partially contribute to the increase in BMD observed in individuals with T2DM. Nevertheless, the diversity of individuals, including age, sex, BMI, and diabetes characteristics, leads to different BMD results in diabetic populations.

There was no significant difference in the 10-year probability of hip fracture or MOF between individuals with and without T2DM in this study, regardless of whether BMD was incorporated. However, lower FRAX scores were observed for MOF with and without BMD and hip fractures with and without BMD. This finding was consistent with a prior study that reported that the FRAX algorithm underestimated fracture risk in these patients5. Notably, a prior study conducted in Thailand included 144 patients with osteoporotic hip fractures and 139 controls without diabetes. A significantly lower FRAX score between the diabetic and control groups was observed for both MOF and hip fractures17. While a statistically significant difference in FRAX between individuals with T2DM and controls was not observed in our study, this important finding suggests that the data align with a similar pattern. However, the lack of significance may be attributed to ethnic variations or varying degrees of diabetes severity among the populations.

Identifying osteoporotic VFs is very important because it is a strong predictor of subsequent fractures and the need for pharmacological treatment18,19. Interestingly, those with T2DM have considerably more vertebral fractures, especially those involving two or more levels, than those without T2DM. In the multivariate analysis, the association between diabetes and vertebral fracture \(\:\ge\:\)2 levels remained unchanged after adjusting for other factors. Previously available data have shown that T2DM patients have a greater incidence2and prevalence20,21of vertebral fracture (VF) than non-DM controls. Furthermore, diabetic patients with VFs have a higher mortality rate than people without T2DM or VFs (HR, 2.11; 95% CI, 1.72–2.59)2. Our study demonstrated a higher prevalence of multiple (\(\:\ge\:\) two levels) VFs in women with T2DM than non-DM controls. Routine radiographic assessments, particularly in the lumbar spine, are indicated for diabetic patients who are at high risk of fracture or have recently lost significant height. The early detection of silent multiple VFs in diabetic individuals eventually promotes rapid anti-osteoporosis treatment to decrease morbidity and subsequently fracture risk in these populations.

The present findings must be interpreted in the context of several potential strengths and weaknesses. Our study’s main strength is its originality since it is the most extensive study in Thailand that compared BMD and FRAX between T2DM patients and non-DM controls. In addition, this study examines the prevalence of VFs using semiquantitative measurement via Genant’s standard method, which is more objective and reproducible22. However, the lack of availability of spine trabecular bone score (TBS) data derived from the texture of spine DXA images is a limitation of our study. Nonetheless, its application in diabetes patients is not without limitations. Several studies have demonstrated the reduction in TBS in T2DM patients as well as the prediction of fracture risk independent of BMD23. After controlling for age and BMI in either sex, a recent Thai study found no association between T2DM and TBS10. Furthermore, numerous studies that evaluated the use of high-resolution-peripheral quantitative computed tomography (HR-pQCT) to assess bone microarchitecture parameters in diabetics found that trabecular bone microarchitecture was not only preserved but also improved in T2DM patients compared to non-DM individuals11,24,25. As a result, the TBS may not be a generally applicable surrogate measure for predicting fracture risk in T2DM patients. Furthermore, the lack of data on glycemic control, duration of diabetes and antidiabetic medication are the limitations of our study. The future cohort might be useful in assessing these components as possible risk factors for multiple vertebral fractures.

In conclusion, patients with T2DM had significantly greater femoral neck and total hip BMD than non-DM individuals. There was no statistically significant difference in the 10-year probability of hip fracture or MOF between individuals with and without T2DM. Compared with non-DM controls, postmenopausal women with T2DM had a higher prevalence of multiple (\(\:\ge\:\) 2 levels) VFs. We found a significant association between diabetes and multiple VFs that was independent of other clinical variables.

Methods

Study design and participants

This cross-sectional study was conducted in Thai postmenopausal women attending a diabetic and health check-up clinic at Srinagarind Hospital, Khon Kaen University. Individuals with a history of premature or surgical menopause, a history of previous low-trauma fracture, known underlying diseases (i.e., type 1 diabetes, rheumatoid arthritis, uncontrolled thyrotoxicosis), current use of medications affecting bone metabolism (i.e., corticosteroids, thyroid hormones, proton pump inhibitors, and antiepileptic and antiosteoporotic drugs) were excluded.

Bone mineral density (BMD) and the fracture risk assessment tool (FRAX)

BMD at the lumbar spine (LS), total hip (TH), and femoral neck (FN) was measured via DXA (GE Healthcare, Madison, WI). BMD T scores were analyzed using Asian population reference databases supplied by the manufacturer. The coefficient of variation for BMD for normal participants ranged from 1.5 to 2.0% for LS, 1.5 to 1.9% for TH, and 1.3–1.5% for FN. FRAX with and without BMD were used to calculate the 10-year probabilities of hip and major osteoporotic fracture (MOF), which were calculated on the basis of the Thai FRAX model, whereas the optimal FRAX threshold for MOF for the Thai population was 10%26.

Radiography and vertebral fracture assessment

A lateral thoracolumbar (T-L) X-ray radiograph was taken with a 101.6 cm tube-to-film distance per the standard protocol, which included details regarding the positioning of the participants and the radiographic technique used. Radiographs were taken in the left lateral position centered at the L1 level. Vertebral bodies from the T4 to L4 levels were assessed by three blinded and independent radiologists to define morphometric VFs via the Genant semiquantitative method22. Multiple Fracture is defined as \(\:\ge\:\)2 levels of vertebral fracture from the T4 to L4 levels. The kappa coefficient among the radiologists was 0.64 (95% CI 0.54–0.76), which was considered moderate agreement27. Any difference in the assessment of a joint (among the readers) was resolved by consensus.

Data analysis

Continuous variables are represented by the mean and standard deviation, whereas categorical variables are represented by the observed number and percentage. The prevalence rates of osteoporosis and vertebral fracture are reported as percentages. When comparing dependent variables between study groups, the independent t test or ANOVA was used where appropriate. Univariate and multivariate logistic regression analyses were used to determine the associations between clinical factors (age, BMI, BMD, and FRAX score) and prevalent vertebral fractures. All the statistical analyses were performed via SPSS, version 17 for Windows Evaluation Software (SPSS Inc. Chicago, IL). Statistical significance was considered at a p-value < 0.05.