Comparison of Bone Microarchitecture and Clinical Biomarkers between Diabetic and Non-Diabetic Postmenopausal Osteoporotic Women

Abstract

Background

To characterize skeletal and metabolic profiles in postmenopausal women with osteoporosis, type 2 diabetes mellitus (T2DM), their coexistence (T2DM+osteoporosis), and healthy controls using biochemical markers, dual energy X-ray absorptiometry (DXA), and high-resolution peripheral quantitative computed tomography (HR-pQCT).

Methods

This cross-sectional study included 220 postmenopausal women (60 each in osteoporosis, T2DM, and T2DM+osteoporosis groups; 40 controls). Participants underwent biochemical, endocrine, and hematological profiling, DXA of the spine and hip, and HR-pQCT imaging at the distal radius and tibia. Bone turnover markers (propeptide of type I collagen, C-terminal telopeptide), parathyroid hormone, and 25-hydroxy-vitamin D were assessed. Data were analyzed using analysis of variance and correlation analyses.

Results

The osteoporosis and T2DM+osteoporosis groups demonstrated significantly lower areal bone mineral density (BMD), elevated bone turnover markers, and marked deterioration in cortical and trabecular microarchitecture. In contrast, T2DM participants exhibited preserved or higher BMD but showed microarchitectural alterations on HR-pQCT. Significant inverse correlations between DXA and HR-pQCT parameters were observed in osteoporosis and T2DM+osteoporosis groups, whereas such associations were not evident in T2DM.

Conclusions

Distinct skeletal phenotypes were observed across osteoporosis, T2DM, and T2DM+osteoporosis. HR-pQCT identified structural deficits not captured by DXA, highlighting that microarchitectural assessment may provide additional insights beyond areal BMD; however, these findings are associative and hypothesis-generating.

GRAPHICAL ABSTRACT

INTRODUCTION

Osteoporosis is characterized by reduced bone mass and compromised bone microarchitecture and is a systemic skeletal disorder that increases fracture risk, particularly among postmenopausal women, in whom estrogen deficiency accelerates bone loss.[1] The resulting skeletal fragility significantly affects mobility, quality of life, and mortality.[2] In individuals with type 2 diabetes mellitus (T2DM), chronic hyperglycemia and poor glycemic control adversely affect bone metabolism, contributing to skeletal fragility and complicating bone health assessment.[3,4] Diabetes-related bone disease is often characterized by low bone turnover, mediated in part by increased sclerostin expression and reduced osteoblast activity; however, this phenotype may vary with disease duration, glycemic control, and the coexistence of osteoporosis.

Women may spend nearly one-third of their lives in a state of progressive bone loss, a process further exacerbated by T2DM. Despite having normal or even higher areal bone mineral density (aBMD), postmenopausal women with T2DM exhibit a 40% to 70% increased risk of fragility fractures compared with non-diabetic women.[5,6] This apparent paradox is attributed to impaired bone quality, driven by oxidative stress, chronic hyperglycemia, and the accumulation of advanced glycation end-products (AGEs), which disrupt collagen integrity and reduce bone strength. [7] Large population-based studies, including analyses incorporating trabecular bone score (TBS) in the Manitoba cohort, have consistently demonstrated increased fracture risk in T2DM despite preserved or elevated BMD, reinforcing the concept that bone quality, rather than bone quantity alone, determines skeletal fragility.[8]

BMD is commonly assessed using T-scores and Z-scores, where osteoporosis is defined by a T-score≤ −2.5 and osteopenia by T-scores between −2.5 and −1. Although dual energy X-ray absorptiometry (DXA) remains the gold standard for measuring aBMD, it provides a two-dimensional assessment and does not capture three-dimensional (3D) bone microarchitecture or material properties that contribute to bone strength.[9] In T2DM, BMD findings are often inconsistent, with spinal and femoral neck BMD typically preserved or increased, while forearm BMD may vary. [10–12] These site-specific discrepancies complicate fracture risk assessment. Some studies report a similar prevalence of osteoporosis in T2DM as in controls,[13] whereas others report increased fracture susceptibility.[14]

High-resolution peripheral quantitative computed tomography (HR-pQCT) provides a detailed evaluation of trabecular and cortical bone microarchitecture.[15] It enables assessment of parameters such as cortical thickness (Ct.Th) and porosity, which are critical determinants of bone strength. Several studies have reported cortical deficits in T2DM despite preserved trabecular structure,[16,17] although findings are not entirely consistent across populations.[18] HR-pQCT has also demonstrated strong associations with bone strength and fracture risk; for instance, cortical pore volume at the distal radius has been reported to be markedly higher in postmenopausal women with T2DM and fractures compared with those without fractures, and increased tibial porosity has been linked to reduced mechanical competence.[19]

Given the frequent coexistence of postmenopausal status and T2DM, this represents a high-risk phenotype for skeletal fragility in which conventional diagnostic tools may not adequately capture alterations in bone quality. Building on these findings, the present study aimed to comprehensively evaluate bone health in postmenopausal women with varying skeletal and metabolic profiles. We analyzed a broad range of systemic parameters, including biochemical, endocrine, and hematological markers, alongside DXA-derived areal BMD and HR-pQCT–based volumetric and microarchitectural indices, with particular emphasis on cortical bone. This integrative approach was designed to examine the relationship between systemic metabolic factors and skeletal integrity, providing exploratory insights into associations between metabolic factors and bone microarchitecture beyond conventional BMD measurements.

METHODS

1. Study design

This cross-sectional study was conducted between January 2023 and December 2025 at the Department of Endocrinology, Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, India. Ethical approval was obtained from the Institute Ethics Committee (IEC No. PGI/IEC/2020/000610), and the study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment.

2. Study population

Asian-Indian postmenopausal women with a history of menopause for at least 5 years were approached. Patients were further evaluated for the presence of T2DM (duration of more than 5 years). Based on the presence or absence of osteoporosis and T2DM, the participants were categorized into three major groups: (1) Osteoporosis without diabetes; (2) Diabetes without osteoporosis; (3) Osteoporosis with diabetes. 40 healthy female controls within the same age range were also recruited for baseline comparison. Participants were recruited consecutively from the outpatient endocrine and metabolic bone disease clinics of a tertiary care referral center. Referrals were primarily for evaluation of osteoporosis, poor glycemic control, or metabolic bone health assessment. Healthy controls were recruited through hospital staff, relatives, and community volunteers after screening to exclude diabetes, osteoporosis, or other metabolic bone disorders. A proportion of participants were receiving calcium (Ca) and vitamin D supplementation as part of standard clinical care prior to recruitment. Participants in the T2DM and T2DM+osteoporosis groups were already on vitamin D supplementation (60,000 IU once weekly) and Ca, as prescribed by their treating physicians. In the osteoporosis group, some participants were receiving similar supplementation, while others were not yet initiated on therapy at the time of enrolment. The control group was not receiving any routine vitamin D or Ca supplementation prior to recruitment.

3. Inclusion and exclusion

Participants eligible for inclusion were postmenopausal women aged ≥18 years with a body mass index (BMI) ≥23 kg/m² and available BMD assessment by DXA. Osteoporosis was defined as a T-score≤−2.5 at the lumbar spine or hip. Individuals with T2DM were required to have a documented disease duration of more than 5 years with glycated hemoglobin (HbA1c) levels between 6.5% and 10%. Based on these criteria, participants were categorized into four groups: controls (normal BMD and non-diabetic), osteoporosis (T-score≤−2.5 on DXA without diabetes), T2DM (HbA1c 6.5%–10% with diabetes duration >5 years and normal BMD), and T2DM+osteoporosis (HbA1c 6.5%–10% with diabetes duration >5 years and T-score≤−2.5 on DXA).[20] History of fragility fractures was recorded but was not used as a primary grouping variable.

Exclusion criteria included individuals with type 1 diabetes or latent autoimmune diabetes, as well as those with prior exposure to osteoporosis treatments such as bisphosphonates, teriparatide, denosumab, zoledronic acid, hormone replacement therapy, or calcitonin. Participants with a history of glucocorticoid use (≥5 mg prednisolone or equivalent for ≥3 months), active malignancy, or implanted medical devices were also excluded. Additionally, individuals with renal dysfunction (serum creatinine >1.5 mg/dL). Participants with prior spinal fusion or artifacts affecting DXA measurements were excluded. We also excluded individuals with infectious diseases or medical conditions that affect bone metabolism, including juvenile or perimenopausal idiopathic osteoporosis. The history of fractures was recorded but not used as a primary grouping variable.

4. Investigations

A questionnaire collected data on the age of menopause, falls in the past 12 months, medication use, and lifestyle habits such as smoking, alcohol consumption, and exercise. The self-reported diagnosis and duration of type 2 diabetes were verified with medical records, which also provide details of diabetic complications. Height and weight were measured to calculate BMI.

Peripheral blood samples were drawn from the median cubital vein using both plain and ethylenediaminetetraacetic acid (EDTA) collection tubes. Serum from the plain tube was used to measure biochemical parameters on a Roche COBAS c501. which included liver function tests (total protein [TP] total bilirubin, alanine aminotransferase [ALT], aspartate aminotransferase, alkaline phosphatase [ALP] and Albumin), fasting plasma glucose, postprandial plasma glucose, renal function tests (urea, creatinine, electrolyte profiling (potassium, sodium [Na], chloride), lipid profile (cholesterol [CHOL], triglycerides [TG], high-density lipoprotein, low-density lipoprotein [LDL]), serum total phosphorus) and Ca. Meanwhile, plasma from the EDTA tube was used for endocrinology profiling, including propeptide of type I collagen (P1NP), β-C-terminal telopeptide (CTX), plasma intact parathyroid hormone (iPTH), HbA1c, and 25-hydroxy-cholecalciferol (vitamin D), assessed using a COBAS-8000 instrument. Complete blood count performed using whole blood by XN-1000TM automated haematology analyser -Sysmex, Japan.

5. DXA and HR-pQCT

The aBMD of the lumbar spine and femoral neck will be assessed using DXA (HOLOGIC Discovery A, QDR 4500; Hologic Inc., Bedford, MA, USA). In addition to aBMD, eligible subjects also underwent high-resolution Microarchitecture HR-pQCT scans. Imaging of the non-dominant distal radius and distal tibia was performed using a second-generation XtremeCT II scanner (Scanco Medical, Brüttisellen, Switzerland). Bone structural and density parameters assessed included total (integral) volumetric BMD (Tt.vBMD; mg HA/cm³) and total cross-sectional area (Tt.CSA; mm²). Trabecular bone parameters measured were trabecular volumetric BMD (Tb.vBMD; mg HA/cm³), trabecular thickness (Tb.Th; mm), trabecular number (Tb.N; 1/mm), trabecular separation (Tb.Sp; mm), and the standard deviation of Tb.Sp (Tb. Sp.SD; mm). Cortical bone evaluation was conducted using a dedicated analysis algorithm, which provided cortical volumetric BMD (Ct.vBMD; mg HA/cm³), cortical porosity (Ct. Po; %), Ct.Th (mm), cortical tissue mineral density (Ct.TMD; mg HA/cm³), and cortical area fraction (Ct.Ar/Tt.Ar; %).

6. Statistical analysis

Descriptive statistics for continuous variables, including mean, standard deviation, and range, were computed using SPSS version 21.0 for Windows (SPSS Inc., Chicago, IL, USA). The four groups were compared using an appropriate parametric test, i.e., one-way analysis of variance (ANOVA), on various biochemical and clinical variables. Differences in DXA and HR-pQCT parameters among the four groups were previously examined by first testing for normality using the Kolmogorov-Smirnov test, followed by ANOVA to compare groups, with the Scheffé test for post hoc analysis. A P value of less than 0.05 is considered statistically significant. Correlation statistics were used to identify any association between HR-pQCT variables and DXA parameters across all four groups. A statistically significant ANOVA was interpreted as evidence of overall group differences; however, pairwise comparisons using the Scheffé post-hoc test were considered significant only when adjusted P values met the threshold, recognizing that conservative correction may yield non-significant pairwise results despite a significant overall ANOVA. Given the large number of comparisons, analyses should be considered exploratory, and no formal correction for multiple testing was applied. Therefore, the risk of type I error cannot be excluded. Further, no multivariable regression models were applied; therefore, potential confounding by variables such as BMI, glycemic control, disease duration, and supplementation status cannot be excluded.

RESULTS

Based on the presence and absence of osteoporosis and T2DM, the participants were categorized into four major groups: (1) Osteoporosis without diabetes (N=60); (2) Diabetes without osteoporosis (N=60); (3) Osteoporosis with diabetes (N=60); and (4) Healthy controls (N=40). Among the study participants, none reported a history of alcohol consumption or smoking. The majority were women with a mean menopausal age ranging from 44 to 47 years. Out of 220 postmenopausal individuals, 137 reported engaging in regular walking as a form of physical exercise, and out of 220 participants, 27 reported a history of fractures. Among the 27 participants with a history of fractures, fractures were most prevalent in the osteoporosis and T2DM+ osteoporosis groups. In some variables, a significant overall ANOVA was observed without statistically significant pairwise differences on post-hoc testing, likely reflecting modest intergroup differences and the conservative nature of the Scheffé correction.

1. Comparison of 4 groups on age, height, weight and BMI

As shown in Table 1, no statistically significant differences in age were found among the groups, indicating that the groups were comparable in age. For height, a significant difference was observed between the T2DM osteoporosis and control groups (P=0.006), with the T2DM osteoporosis group being significantly shorter. Similarly, significant differences in weight were observed in the T2DM group, which had a higher weight than both the osteoporosis (P=0.001) and T2DM osteoporosis (P=0.002) groups. The osteoporosis and T2DM osteoporosis groups had significantly lower weight than controls (P<0.001). BMI also varied significantly across groups. The T2DM group had a higher BMI than the osteoporosis (P<0.001), T2DM osteoporosis (P=0.003), and control (not significant) groups. Both the osteoporosis and T2DM osteoporosis groups had lower BMI compared to controls (P=0.001 and P=0.035, respectively).

Table 1.

Comparison of 4 groups on age, height, weight, and body mass index

Variables	Group 1: OP (N=60)	Group 2: T2DM (N=60)	Group 3: T2DM+OP (N=60)	Group 4: control (N=40)	ANOVA F-value (P-value)	Group ranking (low to high)d)	Significant comparisonsd)
Age (yr)	58.65±7.3	57.73±7.5	59.2±12.39	55.94±9.1	1.09 (0.351)	4<2<1<3	No significant pairwise differences on Scheffé test
Height (cm)	155.9±6.6	154.5±6.69	152.5±7.75	157.9±8.1	4.75 (0.003a))	3<2<1<4	3<4b)
Weight (kg)	58.52±9.1	66.9±11.1	59.13±12.77	68.94±9.9	12.28 (<0.001c))	1<3<2<4	1<2b), 1<4c), 2>3c), 3<4c)
Body mass index (kg/m2)	24.25±3.5	28.21±4.42	25.1±9.87	27.7±4.4	11.11 (<0.001c))	1<3<4<2	1<2c), 1<4b), 2>3b), 3<4a)

A statistically significant analysis of variance (ANOVA) indicates overall group differences. However, post-hoc pairwise comparisons (Scheffé test) may not reach statistical significance due to the conservative nature of the test and variability within groups.

^a)P<0.05,

^b)P<0.01,

^c)P<0.001.

^d)Scheffé post-hoc test.

OP, osteoporosis; T2DM, type 2 diabetes mellitus.

2. Comparison of 4 groups on biochemical, glycemic and haematological parameters

Table 2 summarizes systemic metabolic and inflammatory parameters that provide important contextual insights into disease-specific skeletal alterations. Significant intergroup differences were observed across metabolic and inflammatory parameters. Glycemic indices (fasting glucose, postprandial glucose, and HbA1c) were markedly elevated in both T2DM and T2DM+osteoporosis groups compared to osteoporosis and controls (all P<0.001). Lipid profiling revealed higher TG levels in diabetic groups, whereas total CHOL and LDL levels were significantly higher in the osteoporosis group (P<0.01). Serum uric acid was elevated in T2DM, while ALP levels were highest in the T2DM+osteoporosis group, reflecting increased bone turnover. Total leukocyte counts were significantly higher in osteoporosis and T2DM+osteoporosis, suggesting low-grade systemic inflammation.

Table 2.

Comparison of 4 groups on biochemical, glycemic, and haematological parameters

Variables (normal range)	Group 1: OP (N=60)	Group 2: T2DM (N=60)	Group 3: T2DM+OP (N=60)	Group 4: control (N=40)	ANOVA F-value (P-value)	Group ranking (low to high)d)	Significant comparisonsd)
Serum electrolytes
Sodium (135–145 mmol/L)	139.92±3.1	135.58±14.67	138.27±3.8	140.16±3.1	3.65 (0.01a))	2<3<1<4	OP>T2DMa)
Potassium (3.5–5.5 mmol/L)	4.65±0.5	4.66±0.54	5.46±5.6	4.73±1.0	0.98 (0.40)	1<2<4<3	No significant pairwise differences
Chloride (90–107 mmol/L)	102.0±2.9	99.61±12.86	100.6±4.31	103.93±6.2	2.82 (0.04a))	2<3<1<4	ANOVA significant; no significant pairwise differences on Scheffé test
Calcium (8.8–10.2 mg/dL)	9.42±0.58	9.56±1.2	9.4±0.5	9.47±0.42	0.39 (0.75)	3<1<4<2	No pairwise differences
Phosphorus (2.5–4.5 mg/dL)	3.8±0.8	3.49±0.66	3.68±0.70	3.48±0.46	2.77 (0.04a))	3<2<4<1	ANOVA significant; no significant pairwise differences on Scheffé test
Renal profile
Serum urea (10–50 mL/dL)	27.64±8.6	29.25±14.13	29.5±8.28	24.20±6.2	2.70 (0.04a))	4<1<2<3	ANOVA significant; no significant pairwise differences on Scheffé test
Serum creatinine (0.5–1.2 mg/dL)	0.78±0.14	0.79±0.22	0.81±0.23	0.72±0.12	1.79 (0.14)	4<1<2<3	No significant pairwise differences
Serum uric acid	4.7±1.15	5.65±2.1	5.42±2.8	5.44±2.8	3.52 (0.01a))	1<3<4<2	T2DM>OPa)
Liver function
Serum albumin (3.4–4.8 g/dL)	4.5±0.5	4.45±0.41	4.49±0.64	4.49±0.30	0.11 (0.95)	4=3<2<1	No significant pairwise differences
Total protein (6.4–8.3 g/dL)	7.2±0.7	7.3±0.5	7.95±3.3	7.38±0.49	2.44 (0.06)	1<2<4<3	No significant pairwise differences
Total bilirubin (0.2–1.2 mg/dL)	0.55±0.20	0.66±0.92	0.56±0.31	0.66±0.82	0.53 (0.65)	1<3<2<4	No significant pairwise differences
AST (2–40 U/L)	20.91±5.9	25.12±11.2	27.1±21.4	26.22±10.7	2.32 (0.07)	1<2<4<3	No significant pairwise differences
ALT (2–41 U/L)	22.63±7.13	25.69±12.42	27.6±18.9	34.23±34.9	2.99 (0.03a))	1<2<3<4	Control>OPa)
ALP (42–128 U/L)	106.37±27.32	100.7±27.8	113.62±53.13	91.34±25.01	3.31 (0.02a))	2<4<1<3	T2DM+OP>controla)
Glycemic parameters
FBG (70–100 mg/dL)	92.98±13.0	147.5±53.5	152.8±67.4	92.01±8.05	29.07 (<0.001c))	4<1<3<2	All diabetic groups>OPc), all diabetic groups>controlc)
PPBG (less than 140 mg/dL)	115.92±15.9	210.73±75.09	226.5±79.0	110.21±9.7	59.01 (<0.001c))	4<1<3<2	All diabetic groups>OPc), all diabetic groups>controlc)
HbA1c (3.8%–5.7%)	5.35±0.32	7.83±1.91	8.1±1.8	5.23±0.42	68.73 (<0.001c))	4<1<3<2	All diabetic groups>OPc), all diabetic groups>controlc)
Lipid profile
Cholesterol (150–200 mg/dL)	185.66±42.30	157.32±51.6	155.9±49.7	169.78±44.2	4.99 (0.002b))	3<2<4<1	OP>T2DMa), OP>T2DM+OPa)
TG (50–200 mg/dL)	122.9±41.5	159.56±93.92	136.9±57.2	113.12±43.34	5.18 (0.002b))	4<1<3<2	T2DM>OPa), T2DM>controla)
LDL (100–129 mg/dL)	115.18±33.4	79.06±41.1	80.0±39.9	94.31±34.65	11.84 (<0.001c))	2<3<4<1	OP>T2DMc), OP>T2DM+OPc)
HDL (35–55 mg/dL)	59.42±20.11	50.0±13.56	49.11±12.4	51.00±11.04	5.86 (<0.001c))	3<2<4<1	OP>T2DMa), control>T2DMc)
Hematological parameters
Hb (12–15 g/dL)	11.96±1.2	11.72±1.36	12.01±1.43	12.33±1.5	1.54 (0.20)	2<1<3<4	No significant pairwise differences
TLC (4.0–11.0 × 109/L)	7.34±1.63	7.29±1.6	7.5±1.5	6.37±1.4	4.96 (0.002b))	4<2<1<3	OP>controla), T2DM+OP>controla)
Total platelets count (150–450 × 109/L)	249.54±94.59	210.46±82.4	221.76±105.0	227.82±72.85	1.92 (0.126)	4<3<2<1	No significant pairwise differences
Total RBC count (4.5–5.9 × 1012/L)	4.30±0.54	4.38±0.42	4.56±1.4	4.44±0.42	0.98 (0.401)	1<2<4<3	No significant pairwise differences

Serum electrolytes included sodium, potassium, chloride, total calcium, and phosphorus. The renal profile included serum urea, creatinine, and uric acid. Liver function tests included albumin, total protein, total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP). Glycemic parameters evaluated were fasting blood glucose (FBG), postprandial blood glucose (PPBG), and glycated hemoglobin (HbA1^c). The lipid profile included total cholesterol, triglyceride (TG), low-density lipoprotein (LDL), and high-density lipoprotein (HDL). Hematological parameters analyzed were hemoglobin (H^b), total leukocyte count (TLC), platelet count, and red blood cell (RBC) count. A statistically significant analysis of variance (ANOVA) indicates overall group differences. However, post-hoc pairwise comparisons (Scheffé test) may not reach statistical significance due to the conservative nature of the test and variability within groups.

^a)P<0.05,

^b)P<0.01,

^c)P<0.001.

^d)Scheffé post-hoc test.

OP, osteoporosis; T2DM, type 2 diabetes mellitus.

3. Comparison of 4 groups on different bone health parameters (bone turnover markers, DXA and HR-pQCT radius and tibia)

Table 3 compares bone health across the four groups using DXA and HR-pQCT. Among bone turnover markers, P1NP levels showed significant group differences (P< 0.001), with the osteoporosis group (Group 1) having higher values than T2D (Group 2; P<0.001) and T2DM osteoporosis (Group 3; P<0.05). CTX, a marker of bone resorption, was also significantly elevated in the osteoporosis group compared to all other groups (P<0.001). The iPTH levels were significantly lower in controls compared to osteoporosis and T2DM osteoporosis groups (P=0.001). Serum 25-hydroxy-vitamin D (25[OH]D) levels differed significantly between the groups (P=0.004), with higher levels in the osteoporosis group than in the control group (P<0.01).

Table 3.

Comparison of 4 groups on bone health parameters (blood markers, dual energy X-ray absorptiometry, and high-resolution peripheral quantitative computed tomography parameters)

Variables	Group 1: OP (N=60)	Group 2: T2DM (N=60)	Group 3: T2DM+OP (N=60)	Group 4: control (N=40)	ANOVA F-value (P-value)	Group ranking (low to high)d)	Significant comparisonsd)
Blood markers (normal range)
P1NP (45.47±20.69 ng/mL)	59.01±41.07	34.19±16.35	43.71±20.3	45.14±21.5	8.58 (<0.001c))	2<3<4≈1	OP>T2DMc), OP>T2DM+OPa)
CTX (281.9±95.84 pg/mL)	579.56±280.17	301.53±180.1	395.07±291.9	360.94±185.6	14.04 (<0.001c))	2<3<4<1	OP>all groupsc)
iPTH	60.71±25.30	52.87±36.20	54.17±19.5	38.63±18.9	5.62 (0.001c))	1<3<2<4	OP>controlb), T2DM+OP>controla)
25(OH)D (11.1–42.9 ng/mL)	45.24±23.8	36.89±22.4	36.4±18.1	29.43±20.5	4.51 (0.004b))	1<3<2<4	OP>controlb)
DXA parameters
Hip T-score	−2.42±0.89	0.9±0.7	−1.7±0.9	−1.2±0.7	49.84 (<0.001c))	2<3<1<4	OP<all groupsc)
Hip Z-score	−1.26±0.7	0.6±0.4	−1.1±0.8	0.8±0.5	12.18 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc)
Hip BMD	0.637±0.123	0.834±0.125	0.749±0.196	0.935±0.117	27.99 (<0.001c))	2<3<1<4	OP<all groupsc)
Spine T-score	−3.22±0.78	−1.1±0.6	−2.7±1.0	−1.2±0.7	90.81 (<0.001c))	2<3<1<4	OP<all groupsc)
Spine Z-score	−1.72±0.69	0.8±0.8	−1.3±0.7	0.8±0.5	17.04 (<0.001c))	2<3<1<4	OP<all groupsc)
Spine BMD	0.690±0.09	0.910±0.120	0.714±0.132	0.935±0.117	64.31 (<0.001c))	2<3<1<4	OP<all groupsc)
HR-pQCT parameters
Radius parameters
Tt.vBMD (mg HA/cm3)	214.31±57.71	295.37±101.1	230.12±59.11	317.02±75.57	22.35 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc), T2DM>T2DM+OPc)
Tb.vBMD (mg HA/cm3)	87.17±31.99	111.37±41.01	104.28±132.61	117.89±35.31	1.61 (0.186)	1<3<2<4	No significant pairwise differences
Tb.Meta.vBMD (mg HA/cm3)	142.69±40.73	197.38±136.7	148.81±49.22	174.44±38.30	5.75 (0.001b))	1<3<4<2	OP<T2DMa), T2DM+OP<T2DMa)
Tb.Inn.vBMD (mg HA/cm3)	42.40±30.02	76.58±36.70	48.85±30.69	80.05±36.86	17.42 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc), T2DM>T2DM+OPc)
Ct.vBMD (mg HA/cm3)	784.0±140.0	847.98±181.7	825.7±76.8	904.31±94.92	6.85 (<0.001c))	2<3<1<4	Control>OPc), control>T2DM+OPa)
BV/TV (%)	0.15±0.14	0.18±0.17	0.13±0.04	0.19±0.15	2.28 (0.080)	2<1<3<4	No significant pairwise differences
Tb.N (1/mm)	0.96±0.30	1.20±0.31	0.99±0.221	1.25±0.27	13.73 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc)
Tb.Th (mm)	0.23±0.87	0.23±0.61	0.22±0.01	0.26±0.18	1.24 (0.294)	2<3<1<4	No significant pairwise differences
Tb.Sp (mm)	1.13±0.48	0.85±0.24	1.05±0.29	0.84±0.31	9.28 (<0.001c))	2<3<1<4	OP>T2DMc), OP>controlc)
Tb.1/N.SD (mm)	0.61±0.50	0.67±0.45	0.51±0.29	0.39±0.38	5.52 (<0.001c))	2<3<1<4	OP<T2DMb), OP>controla)
Ct.Th (mm)	0.81±0.23	1.17±0.81	0.85±0.21	1.07±0.27	7.90 (<0.001c))	2<3<1<4	OP<T2DMc), T2DM>T2DM+OPa)
Ct.Po (%)	0.01±0.20	0.006±0.004	0.021±0.106	0.04±0.21	0.92 (0.429)	2<3<1<4	No significant pairwise differences
Ct.Po.Dm (mm)	0.16±0.03	0.18±0.12	0.17±0.30	0.15±0.38	2.16 (0.93)	2<3<1<4	No significant pairwise differences
Tibia parameters
Tt.vBMD (mg HA/cm3)	226.82±54.50	287.23±73.46	225.14±54.12	288.25±54.38	19.05 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc), T2DM>T2DM+OPc)
Tb.vBMD (mg HA/cm3)	96.27±36.29	121.73±34.74	102.09±34.21	141.35±139.4	4.47 (0.005b))	2<3<1<4	OP<controla)
Tb.Meta.vBMD (mg HA/cm3)	174.6±43.9	195.56±36.83	180.9±41.1	188.124±40.4	2.90 (0.036a))	2<3<1<4	ANOVA significant; no significant pairwise differences on Scheffé test
Tb.Inn.vBMD (mg HA/cm3)	174.6±43.90	72.90±33.83	49.44±30.7	71.73±30.70	0.69 (0.558)	2<3<1<4	No significant pairwise differences
Ct.vBMD (mg HA/cm3)	830.96±132.12	878.49±145.1	810.1±132.8	933.14±73.95	8.81 (<0.001c))	2<3<1<4	Control>OPb), control>T2DM+OPb)
BV/TV (%)	0.17±0.12	0.19±0.46	0.18±0.126	0.20±0.11	0.50 (0.680)	2<3<1<4	No significant pairwise differences
Tb.N (1/mm)	0.83±0.28	1.03±0.25	0.89±0.201	1.04±0.180	10.25 (<0.001c))	2<3<1<4	OP<T2DMc), OP<controlc), T2DM>T2DM+OPc)
Tb.Th (mm)	0.26±0.08	0.25±0.22	0.25±0.02	0.25±0.26	0.74 (0.52)	2<3<1<4	No significant pairwise differences
Tb.Sp (mm)	1.38±0.74	0.99±0.28	1.18±0.36	0.96±0.18	9.79 (<0.001c))	2<3<1<4	OP>T2DMc), OP>controlc)
Tb.1/N.SD (mm)	0.78±0.62	0.45±0.254	0.63±0.45	0.40±0.14	8.46 (<0.001c))	2<3<1<4	OP>T2DMc), OP>controlb)
Ct.Th (mm)	0.78±0.62	0.24±0.02	1.33±0.60	1.48±0.26	11.36 (<0.001c))	2<3<1<4	OP>T2DMc), OP<controlb), T2DM+OP<controlb)
Ct.Po (%)	1.25±0.30	0.22±0.04	0.03±0.17	0.01±0.01	4.74 (0.003b))	2<3<1<4	Control<OPa), control<T2DM+OPa)
Ct.Po.Dm (mm)	0.21±0.02	0.22±0.049	0.23±0.32	0.21±0.03	2.10 (0.101)	2<3<1<4	No significant pairwise differences

These parameters collectively assess bone density and microarchitecture to evaluate bone strength and fracture risk (tibia and radius). A statistically significant analysis of variance (ANOVA) indicates overall group differences. However, post-hoc pairwise comparisons (Scheffé test) may not reach statistical significance due to the conservative nature of the test and variability within groups.

^a)P<0.05,

^b)P<0.01,

^c)P<0.001.

^d)Scheffé post-hoc test.

OP, osteoporosis; T2DM, type 2 diabetes mellitus; P1NP, procollagen type N-terminal propeptide of type I collagen; CTX, C-terminal telopeptide; iPTH, intact parathyroid hormone; 25(OH)D, 25-hydroxy-vitamin D; DXA, dual energy X-ray absorptiometry; BMD, bone mineral density; HR-pQCT, high-resolution peripheral quantitative computed tomography; Tt.vBMD, total volumetric BMD; Tb.vBMD, trabecular volumetric BMD; Tb.Meta.vBMD, meta-trabecular volumetric BMD; Tb.Inn.vBMD, inner trabecular volumetric BMD; Ct.vBMD, cortical volumetric BMD; BV/TV, bone volume/total volume; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tb.Sp, trabecular separation; Tb.1/N.SD, trabecular connectivity; Ct.Th, cortical thickness; Ct.Po, cortical porosity; Ct.Po.Dm, cortical pore diameter.

DXA results showed highly significant group differences in hip and spine T-scores, Z-scores, and bone mineral densities (all P<0.001). The osteoporosis group had the lowest values across all DXA metrics, indicating the most severe bone loss. Post-hoc comparisons confirmed that the osteoporosis group had significantly lower hip and spine T-scores and BMD than all other groups (P<0.001), including those with T2D and T2DM osteoporosis. Even between T2DM osteoporosis and controls, significant differences were noted in several bone density parameters (e.g., spine BMD; P<0.001).

HR-pQCT data from the radius revealed significant deficits in volumetric BMD and cortical architecture among osteoporosis participants. Tt.vBMD, Ct.vBMD, and inner trabecular volumetric BMD (Tb.Inn.vBMD) were significantly lower in Group 1 than in Group 2 (T2D) and controls (all P<0.001). The osteoporosis group had significantly lower Tb.N and higher Tb.Sp than both T2D and controls (P< 0.001), indicating deterioration in trabecular structure. Trabecular connectivity (Tb.1/N.SD) and Ct.Th were also significantly impaired in osteoporosis compared to other groups (P<0.001), while Ct.Po and pore diameter did not differ significantly.

In the tibia, similar patterns were observed. Tt.vBMD, Ct.vBMD, and trabecular structure parameters like Tb.Sp and Tb.1/N.SD were significantly worse in osteoporosis compared to other groups (all P<0.001). Notably, Ct.Th was significantly lower in the osteoporosis group compared to T2D and controls (P<0.001), while controls had significantly better cortical parameters overall. Ct.Po was also significantly greater in osteoporosis compared to controls and T2DM osteoporosis (P=0.003), though no significant differences were seen in pore diameter. Overall, both DXA and HR-pQCT analyses confirmed that bone microarchitecture and density are significantly more compromised in patients with osteoporosis-especially when not associated with diabetes-compared to diabetic and healthy controls. The least significant change at our center was 1.5% for lumbar spine BMD and 2.0% for hip BMD, in accordance with the International Society for Clinical Densitometry recommendations.

Representative 3D HR-pQCT images of tibia and radius of all groups are shown in Figure 1.

Fig. 1.

Three-dimensional representative images of high-resolution peripheral quantitative computed tomography of tibia and radius. Representative images from postmenopausal groups. (A–D) Type 2 diabetes mellitus (T2DM). (E–H) Healthy controls. (I–L) T2DM with osteoporosis (OP). (M–P) OP only. In each group, panels represent tibia and radius regions. (A, B, E, F, I, J, M, N) Tibia cortical and trabecular regions. (C, D, G, H, K, L, O, P) Radius cortical and trabecular region.

4. Association between DXA and HR-pQCT parameters in the individual group

In the osteoporosis group, several HR-pQCT parameters showed significant negative correlations with DXA measures. The T-score of the hip was negatively correlated with tibia Tt.vBMD (r=−0.406, P<0.001) and tibia Tb.N (r=−0.405, P<0.001), suggesting that poorer DXA hip scores are associated with reduced trabecular microarchitecture and bone density in the tibia. Additionally, spine T-scores were inversely correlated with radius Tb.vBMD (r=−0.466, P<0.001) and radius inner Tb.vBMD (r=−0.477, P<0.001). The Z-score of the spine similarly showed strong negative correlations with radius Tb.vBMD (r=−0.467, P<0.001) and radius inner Tb.vBMD (r=−0.534, P<0.001), indicating that more compromised bone status as measured by DXA is associated with deteriorated trabecular integrity in the radius.

In the T2DM osteoporosis group, the T-score of the hip was negatively correlated with multiple HR-pQCT markers including radius Tt.vBMD (r=−0.525, P<0.001), tibia Tt. vBMD (r=−0.587, P<0.001), tibia Tb.vBMD (r=−0.407, P< 0.001), and tibia Ct.Th (r=−0.461, P<0.001). This implies that as DXA hip scores decrease, corresponding deterioration in both cortical and trabecular architecture in the tibia and radius is evident. Hip BMD was positively correlated with radius Tt.vBMD (r=0.404, P=0.001), suggesting that individuals with better hip BMD also tend to have superior volumetric density at the radius. The T-score of the spine in this group was inversely correlated with several microarchitectural indices, including radius bone volume/total volume (BV/TV; r=−0.408, P=0.001), radius Tb.N (r=−0.475, P<0.001), and positively with Tb.Sp (r=0.442, P<0.001). These correlations indicate that worsening DXA spine scores in T2DM osteoporosis are associated with reduced Tb.N and increased Tb.Sp.

In the control group, a different pattern emerged. Hip BMD was positively correlated with tibia Tt.vBMD (r=0.532, P<0.001) and tibia Ct.Th (r=0.615, P<0.001), while spine BMD correlated positively with radius Tt.vBMD (r=0.532, P<0.001). These positive correlations suggest that in healthy individuals, higher DXA-based BMD values reflect stronger volumetric and structural integrity in both cortical and trabecular compartments as assessed by HR-pQCT.

No association between any of the DXA variables with HR-pQCT variables were noted in the T2DM group. It is important to note that these correlation analyses are descriptive in nature and do not establish clinical discordance or predictive inadequacy of DXA-derived BMD. These findings should be interpreted as exploratory associations between areal bone density and microarchitectural parameters.

All the details are mentioned in Table 4.

Table 4.

Correlation between dual energy X-ray absorptiometry and high-resolution peripheral quantitative computed tomography in the individual groups^a)

HR-pQCT variables	Hip T-score	Hip BMD	Spine T-score	Spine BMD	Spine Z-score
DXA variables (OP group)
Tibia Tt.vBMD	−0.406 (<0.001c))	-	-	-	-
Tibia Tb.N	−0.405 (<0.001c))	-	-	-	-
Radius Tb.vBMD	-	-	−0.466 (<0.001c))	-	−0.467 (<0.001c))
Radius inner Tb.vBMD	-	-	−0.477 (<0.001c))	-	−0.534 (<0.001c))
DXA variables (T2DM+OP group)
Radius Tt.vBMD	−0.525 (<0.001c))	0.404 (0.001b))	-	-	-
Tibia Tt.vBMD	−0.587 (<0.001c))	-	−0.506 (<0.001c))	-	-
Tibia Tb.vBMD	−0.407 (<0.001c))	-	-	-	-
Tibia Ct.Th	−0.461 (<0.001c))	-	-	-	-
Radius BV/TV	-	-	−0.408 (0.001b))	-	-
Radius Tb.N	-	-	−0.475 (<0.001c))	-	-
Radius Tb.Sp	-	-	0.442 (<0.001c))	-	-
DXA variables (control group)
Tibia Tt.vBMD	-	0.532 (<0.001c))	-	-	-
Tibia Ct.Th	-	0.615 (<0.001c))	-	-	-
Radius Tt.vBMD	-	-	-	0.532 (<0.001c))	-

^a)Only significant correlations are presented, with no correlations observed in type 2 diabetes mellitus (T2DM) group.

^b)P<0.01,

^c)P<0.001.

HR-pQCT, high-resolution peripheral quantitative computed tomography; BMD, bone mineral density; DXA, dual energy X-ray absorptiometry; OP, osteoporosis; Tt.vBMD, total volumetric BMD; Tb.N, trabecular number; Tb.vBMD, trabecular volumetric BMD; Ct.Th, tibia cortical thickness; BV/TV, bone volume/total volume; Tb.Sp, trabecular separation.

DISCUSSION

In this study, we extensively compared and characterized biochemical, hematological, bone turnover markers, DXA (hip and spine), and HR-pOCT at peripheral sites (tibia and radius) in postmenopausal women with and without osteoporosis and T2DM. To the best of our knowledge, no comparative study has explored a wide range of bone turnover markers, DXA, and HR-pOCT in postmenopausal women before.

1. Comparison of age, anthropometrics, and BMI across groups

While no significant age differences were observed among the groups, individuals with T2DM, both with and without osteoporosis (T2DM+osteoporosis), had significantly higher weight and BMI compared with those in the osteoporosis and control groups. In contrast, the osteoporosis and T2DM+osteoporosis groups exhibited lower weight and BMI compared with controls. The relatively lower BMI observed in diabetic participants may reflect referral bias to a tertiary center, lifestyle modifications after diabetes diagnosis, and regional differences in body composition among Asian-Indian women. Importantly, the BMI threshold (>23 kg/m²) used in this study follows recommendations for Asian populations, where metabolic risk, including type 2 diabetes, occurs at lower BMI levels due to higher body fat percentage and visceral adiposity compared with Western populations.[21,22] Consequently, although this cutoff is appropriate for the studied population, it may limit the generalizability of these findings to other ethnic groups with different body composition profiles.[23] Findings further highlight the complex interaction between adiposity, glycemic status, and skeletal health. Although higher BMI is generally associated with greater BMD due to increased mechanical loading, this protective effect appears attenuated in T2DM because bone fragility is largely driven by impaired bone quality rather than reduced bone mass. Chronic hyperglycemia promotes the accumulation of AGEs in bone collagen, leading to abnormal collagen cross-linking and reduced bone strength.[24,25]

Clinically, deterioration of trabecular microarchitecture and increased Ct.Po further may contribute to the increased fracture risk observed in patients with T2DM despite normal or higher BMD.[26] Conversely, the low BMI observed in the osteoporosis and T2DM+osteoporosis groups aligns with well-established risk factors for osteoporotic fractures, including reduced mechanical loading, lower lean body mass, hormonal alterations, and possible nutritional deficiencies (e.g., calcium and vitamin D), all of which contribute to reduced BMD and skeletal fragility.[27]

2. Comparison of biochemical, glycemic, and hematological parameter

Intergroup comparisons revealed significant alterations in metabolic and inflammatory markers, highlighting the complex pathophysiological interplay between osteoporosis, T2DM, and their coexistence. Elevated serum Na in the osteoporosis group may reflect subtle alterations in renal electrolyte regulation often observed in aging populations with compromised skeletal health. In contrast, the higher serum uric acid observed in the T2DM group is consistent with insulin resistance and reduced renal urate excretion associated with metabolic syndrome and hyperinsulinemia. [28] Notably, uric acid exhibits a recognized “uric acid paradox” in skeletal biology. As a potent endogenous antioxidant, uric acid can neutralize reactive oxygen species and may protect osteoblasts from oxidative stress, thereby contributing to higher BMD in several epidemiological studies. [29] However, hyperuricemia has also been associated with an increased risk of vertebral fractures in individuals with T2DM, likely due to accompanying metabolic dysfunction, inflammation, and impaired bone material properties despite preserved BMD.[30] These findings suggest that uric acid may exert protective effects on bone mass, yet may fail to prevent diabetes-related deterioration in bone microarchitecture and strength.

Liver enzyme profiles further differentiated the groups. ALT levels were reduced in the osteoporosis group, which has been associated with frailty, reduced muscle mass, and poorer nutritional status in elderly populations.[31] In contrast, ALP levels were elevated in the T2DM+osteoporosis group, consistent with increased osteoblastic activity and heightened bone turnover observed in osteoporotic states. [32] As expected, both diabetic groups exhibited significantly elevated fasting glucose, postprandial glucose, and HbA1c levels, reflecting persistent hyperglycemia and metabolic dysregulation characteristic of T2DM.

Lipid profile variations also revealed disease-specific patterns. Participants with osteoporosis demonstrated higher total CHOL levels than diabetic participants, supporting the notion that hypercholesterolemia adversely affects bone metabolism by impairing osteoblast differentiation and promoting osteoclast-mediated bone resorption via oxidative and inflammatory pathways.[33] TG were markedly elevated in the T2DM and T2DM+osteoporosis groups, reflecting classic diabetic dyslipidemia associated with insulin resistance.[34] Large population-based datasets, including analyses from the Korean National Health and Nutrition Examination Survey (KNHANES) and the National Health and Nutrition Examination Survey (NHANES), have also reported inverse associations between TG levels and BMD, suggesting that lipid abnormalities may contribute to skeletal fragility through metabolic and inflammatory mechanisms.[35,36]

Markers of systemic inflammation further distinguished the groups. Increased leukocyte counts in the osteoporosis and T2DM+osteoporosis groups suggest the presence of chronic low-grade inflammation, a recognized contributor to both osteoporosis and metabolic syndrome.[37] Pro-inflammatory cytokines such as interleukin-6 and tumor necrosis factor-α stimulate osteoclastogenesis by activating the receptor activator of nuclear factor-κB (RANK) / RANK ligand signaling pathway, thereby enhancing bone resorption and accelerating skeletal deterioration.[38]

Differences in vitamin D and calcium supplementation between groups may have influenced bone turnover markers. Participants in the T2DM and T2DM+osteoporosis groups were already receiving supplementation prior to enrolment, whereas the control group was not supplemented, and only a subset of the osteoporosis group reported prior intake. Vitamin D plays a central role in calcium homeostasis and skeletal remodelling by enhancing intestinal calcium absorption, suppressing PTH secretion, and regulating osteoblast and osteoclast activity through vitamin D receptor signalling pathways.[39,40] Consequently, prior supplementation in the diabetic groups may have contributed to the relatively higher circulating 25(OH) D levels and may have partially influenced the bone turnover markers observed in this study.

3. Bone health parameters: Serum markers, DXA, and HR-pQCT (radius and tibia)

Marked intergroup differences were also observed in bone turnover markers. Serum P1NP and CTX levels were significantly higher in the osteoporosis group compared with the T2DM and T2DM+osteoporosis groups, consistent with current recommendations from the International Osteoporosis Foundation and the International Federation of Clinical Chemistry, which recognize these markers as reference standards for evaluating bone turnover and fracture risk. The lower P1NP levels observed in the T2DM group align with the well-recognized low-bone-turnover phenotype of diabetic bone disease, where chronic hyperglycemia, accumulation of advanced glycation end products, and impaired osteoblast function suppress bone remodelling.[41]

Interestingly, the T2DM+osteoporosis group did not demonstrate the same degree of suppressed turnover typically observed in isolated T2DM. This may reflect the coexistence of two opposing skeletal processes: diabetes-related suppression of bone formation and osteoporosis-related acceleration of bone remodelling. T2DM is generally characterized by a low bone turnover state with reduced levels of bone formation and resorption markers such as P1NP and CTX.[42] In contrast, osteoporosis is associated with increased bone remodeling, driven by enhanced osteoclastic activity along with a compensatory osteoblastic response.[43] In such cases, the osteoporotic component may partially counterbalance the diabetes-induced suppression of turnover, resulting in intermediate or relatively elevated levels of P1NP and CTX.[42,44] Additionally, factors such as vitamin D status, secondary hyperparathyroidism, and disease duration may further modulate bone remodelling dynamics in patients with combined T2DM and osteoporosis.[44]

Finally, higher circulating 25(OH)D levels in the osteoporosis group likely reflect prior supplementation initiated following clinical recognition of bone loss rather than a physiological compensatory response. Elevated iPTH levels observed in osteoporosis and T2DM+osteoporosis groups may indicate secondary hyperparathyroidism associated with aging, vitamin D insufficiency, or altered calcium homeostasis.

DXA findings demonstrated significantly reduced hip and spine T-scores, Z-scores, and aBMD in the osteoporosis group compared with all other groups, consistent with World Health Organization diagnostic criteria and indicative of an increased fracture risk.[45] The T2DM+osteoporosis group also exhibited a significant reduction in spine BMD relative to controls, aligning with meta-analytic evidence suggesting additive skeletal deterioration when osteoporosis coexists with T2DM.[46] In contrast, individuals with T2DM alone showed preserved or even elevated BMD, reinforcing the well-recognized paradox wherein greater bone mass coexists with increased fracture susceptibility, likely due to compromised bone quality rather than density alone.[47]

Our HR-pQCT analysis reinforces that osteoporosis profoundly compromises bone microarchitecture at both the radius and tibia. In the radius, osteoporosis patients demonstrated significantly lower Tt.vBMD, cortical BMD, and inner trabecular BMD compared with T2DM and controls, consistent with studies showing that reductions in cortical and Tb.vBMD are strong predictors of fragility fractures independent of DXA-derived BMD.[48,49]

The osteoporosis group also showed reduced Tb.N and increased Tb.Sp, reflecting classical trabecular network deterioration associated with fracture risk in large cohort studies such as OFELY and CaMos.[50,51] Trabecular connectivity and Ct.Th were also significantly impaired, further contributing to skeletal fragility through reduced structural integrity and load distribution.[51]

Although Ct.Po and pore diameter did not differ significantly at the radial site, this may reflect technical limitations of HR-pQCT, as smaller intracortical pores often fall below the resolution threshold (~82 μm).[52,53] At the tibia, osteoporosis participants exhibited significantly lower total and cortical vBMD, greater Tb.Sp, reduced connectivity, and thinner cortices compared with T2DM and controls. Similar tibial microstructural deficits have been associated with reduced mechanical competence in HR-pQCT–based finite element analyses.[16] Ct.Po was also significantly higher in osteoporosis, consistent with evidence that increased porosity markedly weakens cortical bone and may contribute to fragility fractures.[54] The elevated bone turnover markers observed in the osteoporosis and T2DM+osteoporosis groups likely reflect high-turnover osteoporosis rather than isolated diabetic bone disease, explaining the divergence from studies reporting suppressed turnover in diabetes alone.

Collectively, these findings highlight that microarchitectural deterioration including trabecular disruption, cortical thinning, and increased porosity plays a critical role in skeletal fragility in osteoporosis, changes that may not be fully captured by conventional DXA measurements.

4. Association between DXA and HR-pQCT parameters

In the osteoporosis group, hip T scores inversely correlated with tibial Tt.vBMD and Tb.N, indicating that lower hip aBMD corresponds to diminished trabecular density and connectivity in the distal tibia. While spine T and Z scores showed strong negative relationships with Tb.vBMD in the radius, including the inner trabecular zone. These results are consistent with earlier studies in osteoporotic populations: HR pQCT parameters, such as trabecular density and number, correlate with DXA aBMD and TBS and serve as independent predictors of fracture risk.[55,56] For instance, a recent meta-analysis of 40 HR-pQCT studies confirmed that HR-pQCT trabecular measures may reflect underlying structural differences independently of aBMD.[49]

In the T2DM osteoporosis cohort, DXA hip T-scores inversely correlated with radius total, tibia total, tibia trabecular, and tibia Ct.Th. These results indicate that diminishing hip bone density is paralleled by structural degradation in both cortical and trabecular compartments at peripheral sites.[16,57] Notably, hip BMD correlated positively with radius Tt.vBMD, suggesting that better central bone density reflects improved volumetric bone quality at the radius. Furthermore, spine T-scores in the T2DM osteoporosis group were inversely related to radius BV/TV and Tb.N, and positively related to Tb.Sp. This emphasizes that deteriorating spinal bone density is accompanied by a decline in trabecular integrity and increased porosity in the distal radius bone.[58]

Spine T scores were inversely correlated with radius BV/TV and Tb.N and positively with Tb.Sp.[57] Among healthy controls, central aBMD positively correlated with peripheral cortical vBMD and thickness, consistent with OFELY and Japanese cohort data linking postmenopausal cortical thinning with reduced bone strength.[50,59] No significant correlation was observed between DXA-derived areal BMD and HR-pQCT–based microarchitectural parameters in the T2DM group. This lack of association, despite preserved or elevated aBMD, highlights the limitation of areal BMD in capturing diabetes-related skeletal fragility. The findings are consistent with the role of AGEs, reduced bone material strength, and increased Ct.Po, which may compromise bone quality independent of BMD and may contribute to fracture susceptibility in T2DM.[60]

Given the cross-sectional design, these findings should be interpreted as associative rather than causal. The observed differences in bone microarchitecture and their relationship with metabolic parameters are hypothesis-generating and require longitudinal validation with fracture outcomes. Further, in the present study, no significant correlations were observed between DXA-derived BMD and HR-pQCT parameters in the T2DM group. This may suggest a potential dissociation between areal bone density and bone microarchitecture in individuals with diabetes. However, it is important to emphasize that correlation analyses are descriptive and do not establish clinical discordance or predictive inadequacy of BMD for fracture risk assessment. Therefore, these findings should be interpreted cautiously as exploratory associations rather than as evidence supporting changes in clinical practice.

5. Strengths and limitations

This study provides a comprehensive assessment of skeletal health by integrating biochemical, densitometric, and HR-pQCT–derived microarchitectural parameters in postmenopausal women with osteoporosis, T2DM, and their coexistence. The combined use of DXA and HR-pQCT offers valuable insight into bone quality beyond areal BMD.

However, the cross-sectional design limits causal inference, and findings should be interpreted as associative. Due to sample size constraints, multivariable adjustment was not feasible. Given the exploratory nature and multiple comparisons performed, the possibility of type I error should be considered when interpreting statistically significant findings and the results are therefore exploratory. The single-center design and use of Asian-specific BMI thresholds may limit generalizability. Additionally, heterogeneity within the T2DM group and the absence of fracture outcomes restrict direct clinical interpretation.

A key limitation of this study is the lack of multivariable adjustment for potential confounders, including BMI, glycemic control (HbA1c), diabetes duration, and prior supplementation. These factors are known to influence both bone density and microarchitecture and may partially account for the observed between-group differences. Therefore, the findings should be interpreted as unadjusted associations.

Future longitudinal studies will help further clarify the role of bone microarchitecture in refining fracture risk assessment in metabolic bone disease.

To conclude, the present study delineates distinct skeletal and metabolic phenotypes in osteoporosis, T2DM, and their coexistence (T2DM+osteoporosis). Despite higher BMI and preserved or even elevated areal BMD in T2DM, HR-pQCT revealed significant alterations in bone microarchitecture, including trabecular and cortical deficits that are not captured by conventional DXA measurements. These findings highlight the limitations of relying solely on BMD for skeletal assessment in T2DM and support the concept that diabetes-related bone fragility is primarily driven by compromised bone quality rather than reduced bone mass. In contrast, osteoporosis and T2DM+osteoporosis groups demonstrated both reduced BMD and microarchitectural deterioration, accompanied by elevated bone turnover markers consistent with high-turnover osteoporosis. Although fracture outcomes were not a primary endpoint and only a small proportion of participants had prior fractures, the structural abnormalities identified by HR-pQCT provide important insights into mechanisms that may contribute to increased fracture susceptibility beyond what can be predicted by DXA alone. While HR-pQCT is not yet feasible for routine clinical practice, it remains a valuable research tool for characterizing skeletal microarchitecture and improving future strategies for fracture risk evaluation in metabolic bone disease. However, these observations should be interpreted as exploratory, and do not imply clinical inadequacy of DXA in fracture risk prediction.