1. Top of page
  2. Abstract
  8. Acknowledgements


Adults with type 2 diabetes mellitus (DM) have an elevated fracture risk despite normal areal bone mineral density (aBMD). The study objective was to compare trabecular bone microarchitecture of postmenopausal women with type 2 DM and women without type 2 DM.


An extremity 1T magnetic resonance imaging system was used to acquire axial images (195 × 195 × 1,000 μm3 voxel size) of the distal radius of women recruited from outpatient clinics or by community advertisement. Image segmentation yielded geometric, topologic, and stereologic outcomes, i.e., number and size of trabecular bone network holes (marrow spaces), endosteal area, trabecular bone volume fraction, nodal and branch density, and apparent trabecular thickness, separation, and number. Lumbar spine (LS) and proximal femur BMD were measured with dual x-ray absorptiometry. Microarchitectural differences were assessed using linear regression and adjusted for percent body fat, ethnicity, timed up-and-go test, Charlson Index, and calcium and vitamin D intake; aBMD differences were adjusted for body mass index (BMI).


Women with type 2 DM (n = 30, mean ± SD age 71.0 ± 4.8 years) had larger holes (+13.3%; P = 0.001) within the trabecular bone network than women without type 2 DM (n = 30, mean ± SD age 70.7 ± 4.9 years). LS aBMD was greater in women with type 2 DM; however, after adjustment for BMI, LS aBMD did not differ between groups.


In women with type 2 DM, the average hole size within the trabecular bone network at the distal radius is greater compared to controls. This may explain the elevated fracture risk in this population.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Individuals with type 2 diabetes mellitus (DM) have an elevated risk of hip, vertebral, proximal humerus, wrist, ankle, and foot fractures despite normal or elevated areal bone mineral density (aBMD) (1, 2). Potential factors contributing to elevated aBMD in adults with type 2 DM include greater body mass index (BMI), hyperinsulinemia, higher fat mass leading to higher estrogen levels, and the presence of diffuse idiopathic skeletal hyperostosis (3–6). Higher than normal aBMD in the presence of elevated fracture risk may limit fracture risk stratification based on aBMD. It also implies that bone strength in those with type 2 DM is altered in ways not captured by dual x-ray absorptiometry (DXA) (7).

Bone strength is dependent on material and structural properties, in addition to aBMD and volumetric BMD (8–10). These include bone geometry, morphology, trabecular bone microarchitecture (i.e., trabecular bone network hole size and trabecular separation, thickness, and number), and cortical porosity (i.e., number and size of cortical bone pores) (8, 10, 11). Microarchitecture is typically assessed invasively by bone biopsy and histomorphometry, and noninvasively by in vivo imaging with quantitative computed tomography (QCT) and magnetic resonance imaging (MRI) (9, 12, 13). Analyses of MRI and QCT scans can yield information about bone microarchitecture, such as trabecular bone volume fraction (BVTV), trabecular number (Tb.N), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th) (14). While bone microarchitecture measurements obtained from MRI and high-resolution peripheral QCT (pQCT) are highly related, the advantage of utilizing MRI for the image-based assessment of trabecular bone microarchitecture is the superior signal-to-noise ratio between bone and bone marrow, as well as the lack of radiation associated with scans (15, 16). In addition, small dedicated radio frequency coils are used in MRI scanning for enhanced image resolution.

A recent finite-element modeling study demonstrated that trabecular bone components contribute 16% to bone strength at the radius (17). MRI-based measures of trabecular bone microarchitecture may provide insight on alterations in bone quality that occur in adults with type 2 DM. It has been effective in detecting differences in BVTV, Tb.N, Th.Th, and Tb.Sp in groups differing in osteoporosis diagnosis and fracture history, suggesting a relationship between 2-dimensional (2-D) variables of trabecular bone microarchitecture derived from MRIs and fracture (18, 19). It has been argued, however, that the resolution of MRIs limits the accurate assessment of such 2-D variables because the images are acquired at the current limits of clinical spatial resolution and can be confounded by partial volume effects (20). However, MRI-based assessment of microarchitecture has proven to be promising because these “apparent” 2-D measures derived using MRI correlate well with measures derived by direct histomorphometry and higher resolution imaging (9, 14, 21).

Trabecular bone network hole size may be less affected by image resolution and still provide an assessment of the structural integrity of the trabecular bone network (22, 23). MacIntyre and colleagues demonstrated that postmenopausal women with a prior wrist fracture had a greater mean hole size at the distal radius compared to nonfracture controls, despite no difference in radial aBMD and other apparent measures of bone strength (24). Therefore, assessing trabecular bone network hole size in women with type 2 DM may yield important information about the trabecular bone network.

The purpose of this study was to compare hole size at the distal radius in postmenopausal women with type 2 DM to postmenopausal women without type 2 DM using MRI. We hypothesized that mean hole size would be greater in patients with type 2 DM, despite elevated aBMD, suggesting a more perforated trabecular network at the distal radius. We also explored the impact of type 2 DM on other geometric, stereologic, and topologic variables, including endosteal area, number of holes within the trabecular network, branch density, nodal density, BVTV, and apparent Tb.N, Tb.Sp, and Tb.Th.

Significance & Innovations

  • The trabecular bone network at the distal radius in women with type 2 diabetes mellitus (DM) is more porous than in women without type 2 DM, as evidenced by greater hole size obtained from magnetic resonance imaging analysis.

  • Trabecular bone microarchitecture contributes to overall bone strength; therefore, these findings may further elucidate the elevated fracture risk in adults with type 2 DM, despite normal or elevated bone mineral density.

  • With the rise in the number of adults with type 2 DM, describing the differences in bone quality is important as findings may guide physicians when assessing skeletal health and fracture risk in adults with type 2 DM.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Study design.

In this cross-sectional study, 2 groups of postmenopausal women were recruited: a group with a diagnosis of type 2 DM and a control group without type 2 DM. Individuals with type 2 DM were recruited from Hamilton Health Sciences Diabetes Clinics in Hamilton, Ontario, Canada, in 2008. Women without type 2 DM were recruited through poster advertisements at local hospitals, clinics, and community centers in 2009. All study participants were age ≥65 years and postmenopausal for >5 years, where menopause was defined as 12 months after the cessation of the menstrual cycle. To ensure that women with type 2 DM had longstanding disease, we included only those who had a diagnosis ≥5 years, by applying the Canadian Diabetes Association diagnostic criteria (25).

Study participants were excluded if they had any one of the following: 1) use of medication in the previous 24 months known to affect bone, including hormone therapy, calcitonin, selective estrogen receptor modulator, parathyroid hormone, and bisphosphonates, 2) chronic systemic glucocorticoid exposure (≥3 months, dosage ≥2.5 mg/day), 3) history of metastatic cancer in the past 5 years, 4) diagnosis of Paget's disease, 5) untreated malabsorption syndrome, 6) hyperparathyroidism or hypoparathyroidism, or 7) renal impairment (Figure 1). Participants with ferromagnetic implants or pacemakers were excluded in accordance with MRI safety standards. This study was approved by the McMaster University Faculty of Health Sciences/Hamilton Health Sciences Research Ethics Board.

thumbnail image

Figure 1. Path outlining participant recruitment and enrollment in the study. DM = diabetes mellitus; MRI = magnetic resonance imaging.

Download figure to PowerPoint

Medical history and lifestyle data.

Participants completed a series of interviewer-administered questionnaires to capture current health status, current medication use (including multivitamins and supplements), and major osteoporotic fracture history (hip, wrist, vertebral, or proximal humerus). Participants were classified as white or nonwhite. The age-adjusted Charlson Index served as a comorbidity index, reflecting the presence of weighted comorbid conditions (26). Physical activity levels were assessed using a modified Paffenbarger Physical Activity Questionnaire, which quantifies the number of kilocalories expended per week based on the number of stairs climbed up, miles walked, and participation in recreational activities during a usual week. Each participant's average dietary intake of calcium and vitamin D was estimated using a food frequency questionnaire previously validated by us for use in postmenopausal community-dwelling women (27). Laboratory biochemistry (random glucose, glycosylated hemoglobin, creatinine) was abstracted from the medical charts of the participants with type 2 DM. The Cockcroft-Gault equation was used to estimate glomerular filtration rate.

The participants' height was captured to the nearest 0.1 cm using a wall-mounted stadiometer and weight was obtained from the whole-body DXA scan to the nearest 0.1 kg, from which BMI was calculated. Waist and hip circumference were also measured. Grip strength was assessed using an isometric dynamometer (Takei T.K.K.5001 Grip A Dynamometer, Takei Scientific Instruments) and average grip strength was calculated from 3 assessments with the dominant hand. A timed up-and-go (TUG) test was used to assess the participant's physical mobility. A normative cutoff point of 12.0 seconds was used for TUG test performance (28).


The nondominant distal radius was imaged with a 1T extremity MRI system (OrthOne, GE Healthcare) by the same operator. Participants were seated in a chair with their wrist in a prone position in a 100-mm diameter transmit/receive coil. Bracing and padding were applied to enhance patient comfort and lessen the potential for motion artifact. A fast spin-echo (FSE) sagittal localizer was performed followed by a FSE coronal localizer, wherein reference lines were placed at the most distal articular surface on the medial aspect of the radius and at 20 mm proximal to this line (Figure 2A). A spoiled 3-D gradient-echo sequence was used to acquire 20 axial images of the wrist at a 1.0-mm slice thickness with the following sequence parameters: repetition time 47 msec, echo time 23.8 msec, flip angle 40°, 15-kHz bandwidth, number of excitations = 1, field of view 100 mm × 100 mm, voxel size 195 × 195 × 1,000 mm3, and 12:09 minute scan time. An anthropometric phantom was scanned on a daily basis to ensure system quality assurance.

thumbnail image

Figure 2. Representative coronal image of the nondominant distal radius (A) showing the placement of the reference line and the region of interest captured in the axial images (B and C [inverted image]). The most proximal edge of the radius at the articular surface was used as a landmark for all participants to set the horizontal reference line. Semiautomatic segmentation delineates the trabecular bone compartment (D) and thresholding separates the marrow and bone (E).

Download figure to PowerPoint

Image analysis.

Axial images were uploaded for slice-by-slice segmentation performed using inhouse software developed at our institution. The software uses a graph-based technique to identify the endosteal boundary of the radius (29). The area inside the endosteal perimeter (i.e., cortical bone excluded) served as the region of interest for each slice. A local thresholding technique was applied to the region of interest to separate the trabecular bone (no signal) and marrow (signal) phases (Figure 2) (30). Finally, a skeletonization algorithm was applied to the segmented trabecular bone images to enable topographic analyses (31). Similar software programs have been used by groups at our institution and others to derive variables of bone microarchitecture at peripheral sites, such as the radius and tibia (9, 23, 24). The central 6 slices were selected for analysis for each participant.

Apparent geometric, stereologic, and topographic measures were derived, including mean hole size (mm2), number of holes, endosteal area (mm2), trabecular BVTV (%), Tb.Th (mm), Tb.Sp (mm), Tb.N (mm−1), nodal density (number of nodal points/mm2), and branch density (number of branches/mm2) (9, 12, 23, 32). Briefly, BVTV was calculated as the area occupied by pixels corresponding to bone divided by the total area within the endosteal boundary. A region growing technique was utilized to determine hole size within the trabecular bone network and is defined as

  • equation image

where Ai denotes the area of the ith hole, and n denotes the number of holes present in the trabecular network (33). Within the skeletonized trabecular network, branches (single-pixel–wide line segments) and nodes (defined as points at which 3 or more branches join) were identified and used to compute nodal and branch density (34). The parallel plate model was used to estimate Tb.Th, Tb.Sp, and Tb.N from the perimeter lengths of network holes (35). Reproducibility of the image analysis technique was assessed by analyzing 30 sets of scans (15 from the group of women with type 2 DM and 15 from the group of women without type 2 DM) in duplicate during independent analysis sessions (Table 1).

Table 1. Reliability of image analysis technique for trabecular bone microarchitecture variables at distal radius*
  • *

    RMSCV% = root mean square coefficient of variation; ICC = intraclass correlation coefficient; BVTV = bone volume fraction; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; Tb.N = trabecular number.

Hole size, mm24.670.98
Number of holes4.500.97
Endosteal area, mm24.900.95
BVTV, %1.100.83
Tb.Th, mm1.660.94
Tb.Sp, mm3.570.99
Tb.N, mm−12.680.90
Nodal density1.540.97
Branch density4.110.87


A DXA (Hologic, Discovery QDR4500A) scan was acquired to determine total aBMD (gm/cm2) at the lumbar spine (LS; L1–L4) and proximal femur (femoral neck [FN] and total hip). Proximal femur T scores were computed using Third National Health and Nutrition Examination Survey reference data. Total mass, fat free (lean) mass, and fat mass were measured in a whole-body DXA scan from which percent body fat was derived. Standard DXA quality assurance protocols were followed, including use of daily spine phantoms and weekly step phantoms. The short-term in vivo operator precision, expressed as the root mean square coefficient of variation (RMSCV%) for the LS, FN, and total hip aBMD were 0.96%, 1.70%, and 1.47%, respectively, which was determined by performing triplicate LS and left hip scans on 13 young, healthy volunteers. DXA scans were analyzed by a certified DXA technician and radiologist who were blinded to the study purpose and participant allocation.

Statistical analyses.

The mean ± SD was determined for continuous variables, and frequency (%) for categorical variables. Differences between groups for the independent variables are presented in Table 2. The dependent variables in the unadjusted analysis were assessed using an unpaired Student's t-test (for continuous variables) or chi-square test (for categorical variables) and are presented in Table 3. Linear regression was used to explore the association between the dependent variable (variables of microarchitecture) and independent variables. A priori, percent body fat, ethnicity, age-adjusted Charlson Index, TUG test result, total calcium intake, and total vitamin D intake were forced into the statistical model to examine the differences in variables of bone microarchitecture in women with and without type 2 DM (36–38). The other independent variables were collected for descriptive purposes. We also adjusted for BMI when comparing LS and FN aBMD, as BMI is related to LS aBMD (3). We verified that the following assumptions in regression analysis were true for our model: 1) linear relationship between dependent and independent variables, 2) normality of errors, 3) homogeneity of variance, and 4) independence of errors associated with each observation. In addition, we found that the independent variables included in the regression model were not highly colinear. The Holm's procedure for multiple comparisons was performed for the comparison of secondary bone microarchitecture variables between groups (39). The reliability of the MRI analysis technique is expressed as RMSCV% and intraclass correlation coefficient (ICC; type 2,1) for 30 scans analyzed in duplicate. We based our sample size calculation on a previous study investigating bone microarchitecture in women with and without a prior wrist fracture using pQCT (in-plane voxel size 333 μm) (24). The investigators found a mean difference in trabecular bone network hole size between the 2 groups of 1.98 mm2 with an average SD of 2.7 mm2. Using a power of 80% and an alpha level of 0.05, we aimed to enroll 30 participants/group. All analyses were performed with SPSS, version 18.0 for Windows. A P value of less than 0.05 was considered significant for this study.

Table 2. Descriptive characteristics of study participants*
 Women with type 2 DM (n = 30)Women without type 2 DM (n = 30)P
  • *

    Values are mean ± SD unless otherwise indicated. DM = diabetes mellitus; NS = not significant; BMI = body mass index; TUG = timed up-and-go test; LS = lumbar spine; BMD = bone mineral density; FN = femoral neck.

  • P < 0.05 considered significant.

  • Ambulation aid includes single-point cane, 4-point cane, standard walker, and rollator walker.

  • §

    Major osteoporotic fracture defined as a hip, wrist, clinical spine, or humerus fracture that occurred from a fall from standing height or less, or a fall from <4 stairs.

Age, years71.0 ± 4.870.7 ± 4.9NS
White, no. (%)23 (79.3)30 (100.0)NS
Height, cm158.6 ± 6.8160.1 ± 5.0NS
Weight, kg86.9 ± 18.771.3 ± 13.7< 0.001
BMI (kg/m2)34.6 ± 7.627.9 ± 5.5< 0.001
 Normal (20.0–24.9), no. (%)08 (26.7)0.002
 Overweight (25.0–29.9), no. (%)5 (17.2)11 (36.7) 
 Obese ≥30, no. (%)23 (76.7)10 (30.0) 
Waist:hip ratio0.89 ± 0.070.83 ± 0.060.002
Body fat percentage40.3 ± 6.137.2 ± 6.50.057
Time since menopause, years22 ± 722 ± 8NS
Ambulation with an aid, no. (%)7 (24.1)2 (6.7)NS
No. prescription medications6.6 ± 3.41.9 ± 2.2< 0.001
Current smoker, no. (%)2 (6.9)0NS
History of atraumatic major osteoporotic fracture after age 40 years, no. (%)§5 (17.2)6 (20.0)NS
Age-adjusted Charlson Index1.5 ± 2.20.3 ± 1.40.023
Bone-related nutrient intake, physical activity, and functional testing   
 Total calcium intake, mg/day1,594 ± 6962,075 ± 5970.007
  From supplements446 ± 481678 ± 482NS
  From food1,148 ± 5641,397 ± 3350.046
 Total vitamin D intake, IU/day806 ± 6221,197 ± 922NS
  From supplements626 ± 573982 ± 921NS
  From food179 ± 142195 ± 130NS
Daily energy expenditure, kcal/day1,904 ± 2,3642,557 ± 2,170NS
TUG test, seconds12.8 ± 4.09.4 ± 2.7< 0.001
 ≤1214 ± 56.026 ± 86.70.011
 >1211 ± 44.04 ± 13.3 
Grip strength, kg18.8 ± 4.821.7 ± 6.3NS
Bone densitometry   
 LS BMD, gm/cm21.07 ± 0.150.98 ± 0.180.045
 LS T score0.15 ± 1.40−0.61 ± 1.66NS
 FN BMD, gm/cm20.73 ± 0.110.69 ± 0.10NS
 FN T score−1.11 ± 1.02−1.40 ± 0.89NS
 Total hip BMD, gm/cm20.87 ± 0.210.86 ± 0.11NS
 Total hip T score−0.58 ± 0.99−0.70 ± 0.95NS
Table 3. Values for trabecular bone microarchitectural variables at the distal radius in women with type 2 DM and controls without type 2 DM*
 Unadjusted analysisMultivariate-adjusted analysisHolm's adjusted P
Women with type 2 DM (n = 29)Women without type 2 DM (n = 25)PWomen with type 2 DM (n = 29)Women without type 2 DM (n = 25)P
  • *

    Values are the mean ± SD unless indicated otherwise. DM = diabetes mellitus; BVTV = bone volume fraction; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; Tb.N = trabecular number.

  • Indicates statistical significance according to Holm's test for multiple comparisons. Multivariate analysis adjusted for percent body fat, ethnicity, age-adjusted Charlson Index, total daily calcium intake, total daily vitamin D intake, and timed up-and-go test result.

Hole size, mm22.20 ± 0.451.94 ± 0.330.0112.22 ± 0.471.96 ± 0.260.001 
No. of holes72.5 ± 16.982.9 ± 15.50.02571.9 ± 17.681.9 ± 15.80.0220.154
Endosteal area, mm2310.24 ± 54.98323.33 ± 56.340.397308.76 ± 56.73322.83 ± 57.870.3190.957
BVTV, %51.2 ± 2.251.6 ± 1.70.45451.3 ± 2.151.5 ± 1.90.4500.957
Tb.Th, mm0.55 ± 0.040.55 ± 0.020.5910.56 ± 0.040.55 ± 0.020.2280.912
Tb.Sp, mm0.92 ± 0.050.93 ± 0.040.2870.92 ± 0.060.93 ± 0.040.0850.425
Tb.N, mm−10.53 ± 0.050.52 ± 0.040.2500.53 ± 0.050.52 ± 0.040.0670.402
Nodal density0.69 ± 0.080.74 ± 0.050.0280.69 ± 0.080.73 ± 0.050.0140.112
Branch density0.33 ± 0.030.33 ± 0.030.8750.33 ± 0.030.33 ± 0.020.7760.957


  1. Top of page
  2. Abstract
  8. Acknowledgements

The study included 30 women with type 2 DM, and 30 women without type 2 DM (Figure 1). Descriptive characteristics for both study groups are summarized in Table 2. A greater proportion of participants with type 2 DM had a diagnosis of osteoarthritis of at least 1 joint (16 out of 30 [53.3%] versus 5 out of 30 [16.7%]; P = 0.008) (data not shown). Women with type 2 DM had a higher mean BMI (P < 0.001) and fewer of these participants completed the TUG test in under 12 seconds, a normal cutoff point for community-dwelling adults (P = 0.011) (28). The serum biochemistry and antihyperglycemic medications used for participants with type 2 DM are summarized in Table 4.

Table 4. Descriptive data for participants with type 2 DM*
 Women with type 2 DM (n = 30)
  • *

    Values are the number (percentage) unless indicated otherwise. DM = diabetes mellitus.

Years with type 2 DM diagnosis, mean ± SD16.6 ± 11.1
Type of antihyperglycemic medication used 
 Insulin18 (60.0)
 Biguanide12 (40.0)
 Insulin secreting sulfonylurea or nonsulfonylurea4 (13.3)
 Thiazolidinedione1 (3.3)
Serum biochemistry 
 Random glucose, mean ± SD mmoles/liter8.3 ± 3.7
 Glycosylated hemoglobin7.8 (1.7)
 Glomerular filtration rate, mean ± SD ml/minute78.4 ± 28.4

Bone microarchitecture.

The reliability of the image analysis technique was good, as evidenced by high ICC and low RMSCV% values for each microarchitecture variable (Table 1). The comparison between bone microarchitecture variables for women with type 2 DM and controls is presented in Table 3. Six distal radius MRI scans were considered unacceptable for analysis due to motion artifact as assessed by an independent observer blind to subject identifier group, resulting in 29 and 25 analyzable image sets for the type 2 DM and control groups, respectively. The participants with discarded scans did not appear to be different from the rest of the control group, with respect to the descriptive characteristics in Table 2 (data not shown). The unadjusted comparison of bone microarchitecture variables revealed that in women with type 2 DM, trabecular bone network holes were 13.4% larger in area (mean ± SD 2.20 ± 0.45 mm2 versus 1.94 ± 0.33 mm2; P = 0.011). After adjusting for multiple comparisons, no differences were detected for number of holes, endosteal area, BVTV, nodal density, branch density, and apparent Tb.Th, Tb.Sp, and Tb.N in the model (Table 3). After considering multiple comparisons and adjusting for percent body fat, ethnicity, age-adjusted Charlson Index, total calcium and vitamin D intake, and TUG result, only trabecular bone network hole size was 13.3% greater in women with type 2 DM (mean ± SD 2.22 ± 0.47 mm2 versus 1.96 ± 0.26 mm2; P = 0.001) (Table 3).

Bone density measurements.

Mean ± SD LS aBMD was greater in women with type 2 DM (1.07 ± 0.15 gm/cm2) compared to women in the control group (0.98 ± 0.18 gm/cm2; P = 0.045]. After adjustment for BMI, there were no differences detected between groups for LS aBMD (P = 0.572), FN aBMD (P = 0.663), or total hip aBMD (P = 0.224) (data not shown). One LS DXA scan from a participant with type 2 DM was excluded from analysis due to image artifact (contrast). There were more cases of spine scoliosis (4 out of 30 [13.3%] versus 2 out of 29 [6.9%]) and vertebral compression (1 out of 30 [3.3%] versus 0 out of 29) in the control group compared to the women with type 2 DM (data not shown).


  1. Top of page
  2. Abstract
  8. Acknowledgements

This is the first cross-sectional study to demonstrate that there are larger trabecular bone network holes at the distal radius in women with type 2 DM, after adjustment for percent body fat, ethnicity, age-adjusted Charlson Index, calcium and vitamin D intake, and TUG result. This finding provides a possible explanation for bone fragility in this population, given the importance of hole size described in previous research. After adjusting for multiple comparisons, all other variables of trabecular bone microarchitecture were not different between groups. This study may have been underpowered to accurately assess these other mircoarchitecture outcomes and further investigation is needed.

Observational imaging studies in adults with a fracture history have demonstrated that reduced bone strength may be influenced by deficits in the trabecular bone network (10, 19, 24). In particular, the size of the holes in the trabecular network at appendicular skeletal sites can provide information about bone structural competence (40). Trabecular bone network hole size has been shown to be greater in women with a history of wrist fracture compared to BMI- and age-matched women who had not had a prior wrist fracture (24). Other surrogate outcomes for skeletal health, including aBMD, connectivity index, and stress-strain index, were not different between wrist fracture and control participants in this prior study. Furthermore, there was a highly significant correlation between hole size and prior wrist fracture (odds ratio 5.4, 95% confidence interval 1.2–24.3, P = 0.03), whereas no relationship was detected between aBMD and prior fracture (24). Hole size also appears to be a more powerful discriminator of vertebral fracture than Tb.Sp and Tb.Th, and can differentiate those with a fracture history with greater sensitivity and specificity than aBMD alone (33). Trabecular bone network hole size has also been shown to contribute to bone strength in an ex vivo study using radial bone specimens, where a strong association was demonstrated between average hole size and maximum hole area, as well as peak load at fracture (41).

The mechanism causing larger holes in the trabecular bone network at the radius in participants with type 2 DM is not fully understood. However, studies in rodent models of type 2 DM and in rodents fed high-fat diets have demonstrated a reduction in osteoblast recruitment and mineral apposition rate and an increase in osteoclastogenesis, resulting in an imbalance between bone formation and resorption (42, 43). Cross-sectional findings in humans further support these data, as serum markers of bone formation, such as osteocalcin, are lower and markers of bone resorption, such as C-terminal telopeptide of type 1 collagen, are elevated in participants with type 2 DM (44, 45). Increased bone resorption may be mediated by the formation of advanced glycation end products, which stimulate osteoclast activity and may lead to an uncoupling of bone formation and resorption (46).

After the adjustment for multiple comparisons, we did not detect differences in branch density, nodal density, endosteal area, BVTV, Tb.Sp, Tb.Th, or Tb.N between women with type 2 DM and nondiabetic controls. This could be explained by the different methods used in deriving hole size and 2-D variables such as Tb.Sp, Tb.Th, and Tb.N. Hole size analyses do not depend on any stereologic assumptions, whereas the derivation of Tb.Sp, Tb.Th, and Tb.N does indeed depend on sterologic assumptions (i.e., parallel plate model) (35). Moreover, some suggest that comprehensive imaging studies aiming to determine histomorphometric differences between groups should have at least 50 participants per group to provide sufficient power (19). Burghardt and colleagues also found no differences in radius microarchitectural indices, including Tb.N, in a smaller pilot study in women with type 2 DM and age- and height-matched healthy controls using pQCT (47). The authors also reported no difference in indices of bone strength between groups, which might be attributed to the sample size in this study. While variables of trabecular bone microarchitecture and indices of bone strength at the radius were not found to be different in this previous pQCT study, Burghardt and colleagues did report a greater number of cortical pores and greater cortical bone pore volume at the radius (47). Similar findings have been reported in men with type 2 DM (48). Conversely, a smaller pQCT study conducted in a cohort of postmenopausal women found no differences in trabecular bone microarchitecture or cortical porosity in women with type 2 DM and controls (49). In comparison to these prior studies, our study population was composed of older women who had type 2 DM for a greater number of years. The radius site assessed in the present study was also distal to the sites assessed in the prior studies, which may explain discrepancies in trabecular bone microarchitecture results. In addition, the in-plane image resolution of the MRI system that was employed for the present study was poorer (195 μm) compared to the image resolution of the high-resolution pQCT system (82 μm), limiting our ability to resolve the cortical bone. Others have also reported difficulties resolving the cortical bone at the distal radius with MRI due in part to the low-intensity signal of neighboring regions of connective tissue, which is similar to that of bone (15). Although previous work suggests that cortical bone is more porous, our data indicate that changes to the trabecular network may also contribute to increased fracture risk in those with type 2 DM.

There are limitations with this study. First, partial volume effects, produced when imaging trabeculae that are smaller than the spatial resolution, may confound and overestimate measures of bone microarchitecture such as BVTV and Tb.Th (9, 20). However, studies have demonstrated that “apparent” histomorphometric measures derived from MRI correlate well with measures derived through the use of higher-resolution modalities, such as micro–computed tomography, wherein partial volume effects are mitigated (21). Second, the 195 μm in-plane resolution limited our ability to resolve cortical bone, which is more porous in individuals with type 2 DM (47). Third, trabecular bone network hole size is a 2-D measurement, which does not completely account for the anisotropic 3-D nature of trabecular bone in vivo. Fourth, unlike pQCT, MRIs do not provide information pertaining to areal or volumetric bone density, and therefore we cannot compare or adjust for bone density at the distal radius. Fifth, due to our small sample size and few fractures, we did not have the power to investigate the association between hole size and prevalent osteoporotic fractures and the influence of medications on hole size, such as thiazolidinediones, which are associated with elevated fracture risk (50). Finally, we did not use a clinical test to screen for peripheral neuropathy in participants with type 2 DM, which could influence bone quality at the distal radius.

Advances in the use of MRI have afforded insight into the impact of disease on bone microarchitecture. Our results suggest that in women with type 2 DM, the trabecular network at the distal radius is characterized as having larger holes compared to nondiabetic women of similar age. Given the known contribution of bone microarchitecture to overall bone strength, these findings may explain why an elevated fracture risk has been observed in women with type 2 DM despite normal or elevated aBMD. Future research should clarify the independent contribution of DM to fracture risk, to inform risk assessment and stratification, and should focus on understanding other mechanisms behind diabetic bone fragility.


  1. Top of page
  2. Abstract
  8. Acknowledgements

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Papaioannou had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Pritchard, Giangregorio, Atkinson, Beattie, Punthakee, Adachi, Papaioannou.

Acquisition of data. Pritchard, Inglis.

Analysis and interpretation of data. Pritchard, Ioannidis.


  1. Top of page
  2. Abstract
  8. Acknowledgements

We would like to sincerely thank all study participants for their participation. The authors are indebted to the Director of the Diabetes Clinic, Dr. H. Gerstein, and clinical staff at the Hamilton Health Sciences Well-Health Centre, including Janet MacLeod, Marian Wheeler, Jennifer Holterman, Anka Brozik, Brenda Murch, and Cheryl Miller. We also thank Jackie Kinch for assisting in the analysis of the DXA scans, and Dr. Colin Webber for reviewing the DXA scans and providing valuable advice throughout the study.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  • 1
    Janghorbani M, Feskanich D, Willett WC, Hu F. Prospective study of diabetes and risk of hip fracture: the Nurses' Health Study. Diabetes Care 2006; 29: 15738.
  • 2
    Vestergaard P, Rejnmark L, Mosekilde L. Relative fracture risk in patients with diabetes mellitus, and the impact of insulin and oral antidiabetic medication on relative fracture risk. Diabetologia 2005; 48: 12929.
  • 3
    Register TC, Lenchik L, Hsu FC, Lohman KK, Freedman BI, Bowden DW, et al. Type 2 diabetes is not independently associated with spinal trabecular volumetric bone mineral density measured by QCT in the Diabetes Heart Study. Bone 2006; 39: 62833.
  • 4
    Reid IR, Plank LD, Evans MC. Fat mass is an important determinant of whole body bone density in premenopausal women but not in men. J Clin Endocrinol Metab 1992; 75: 77982.
  • 5
    Reid IR, Evans MC, Cooper GJ, Ames RW, Stapleton J. Circulating insulin levels are related to bone density in normal postmenopausal women. Am J Physiol 1993; 265: E6559.
  • 6
    Di Franco M, Mauceri MT, Sili-Scavalli A, Iagnocco A, Ciocci A. Study of peripheral bone mineral density in patients with diffuse idiopathic skeletal hyperostosis. Clin Rheumatol 2000; 19: 18892.
  • 7
    Bouxsein ML. Technology insight: non-invasive assessment of bone strength in osteoporosis. Nat Clin Pract Rheumatol 2008; 4: 3108.
  • 8
    Davison KS, Siminoski K, Adachi JD, Hanley DA, Goltzman D, Hodsman AB, et al. Bone strength: the whole is greater than the sum of its parts. Semin Arthritis Rheum 2006; 36: 2231.
  • 9
    Majumdar S, Newitt D, Mathur A, Osman D, Gies A, Chiu E, et al. Magnetic resonance imaging of trabecular bone structure in the distal radius: relationship with x-ray tomographic microscopy and biomechanics. Osteoporos Int 1996; 6: 37685.
  • 10
    Follet H, Viguet-Carrin S, Burt-Pichat B, Depalle B, Bala Y, Gineyts E, et al. Effects of preexisting microdamage, collagen cross-links, degree of mineralization, age, and architecture on compressive mechanical properties of elderly human vertebral trabecular bone. J Orthop Res 2011; 29: 4818.
  • 11
    Sornay-Rendu E, Boutroy S, Munoz F, Delmas PD. Alterations of cortical and trabecular architecture are associated with fractures in postmenopausal women, partially independent of decreased BMD measured by DXA: the OFELY study. J Bone Miner Res 2007; 22: 42533.
  • 12
    Parfitt AM. Bone histomorphometry: proposed system for standardization of nomenclature, symbols, and units. Calcif Tissue Int 1988; 42: 2846.
  • 13
    Gordon CL, Webber CE, Adachi JD, Christoforou N. In vivo assessment of trabecular bone structure at the distal radius from high-resolution computed tomography images. Phys Med Biol 1996; 41: 495508.
  • 14
    Laib A, Beuf O, Issever A, Newitt DC, Majumdar S. Direct measures of trabecular bone architecture from MR images. Adv Exp Med Biol 2001; 496: 3746.
  • 15
    Kazakia GJ, Hyun B, Burghardt AJ, Krug R, Newitt DC, de Papp AE, et al. In vivo determination of bone structure in postmenopausal women: a comparison of HR-pQCT and high-field MR imaging. J Bone Miner Res 2008; 23: 46374.
  • 16
    Wehrli FW, Saha PK, Gomberg BR, Song HK, Snyder PJ, Benito M, et al. Role of magnetic resonance for assessing structure and function of trabecular bone. Top Magn Reson Imaging 2002; 13: 33555.
  • 17
    Vilayphiou N, Boutroy S, Szulc P, van Rietbergen B, Munoz F, Delmas PD, et al. Finite element analysis performed on radius and tibia HR-pQCT images and fragility fractures at all sites in men. J Bone Miner Res 2011; 26: 96573.
  • 18
    Majumdar S, Genant HK, Grampp S, Newitt DC, Truong VH, Lin JC, et al. Correlation of trabecular bone structure with age, bone mineral density, and osteoporotic status: in vivo studies in the distal radius using high resolution magnetic resonance imaging. J Bone Miner Res 1997; 12: 1118.
  • 19
    Majumdar S, Link TM, Augat P, Lin JC, Newitt D, Lane NE, et al. Trabecular bone architecture in the distal radius using magnetic resonance imaging in subjects with fractures of the proximal femur. Osteoporos Int 1999; 10: 2319.
  • 20
    Kothari M, Keaveny TM, Lin JC, Newitt DC, Genant HK, Majumdar S. Impact of spatial resolution on the prediction of trabecular architecture parameters. Bone 1998; 22: 43743.
  • 21
    Krug R, Carballido-Gamio J, Burghardt AJ, Kazakia G, Hyun BH, Jobke B, et al. Assessment of trabecular bone structure comparing magnetic resonance imaging at 3 Tesla with high-resolution peripheral quantitative computed tomography ex vivo and in vivo. Osteoporos Int 2008; 19: 65361.
  • 22
    Vesterby A, Mosekilde L, Gundersen HJ, Melsen F, Mosekilde L, Holme K, et al. Biologically meaningful determinants of the in vitro strength of lumbar vertebrae. Bone 1991; 12: 21924.
  • 23
    Gordon CL, Webber CE, Christoforou N, Nahmias C. In vivo assessment of trabecular bone structure at the distal radius from high-resolution magnetic resonance images. Med Phys 1997; 24: 58593.
  • 24
    MacIntyre NJ, Adachi JD, Webber CE. In vivo measurement of apparent trabecular bone structure of the radius in women with low bone density discriminates patients with recent wrist fracture from those without fracture. J Clin Densitom 2003; 6: 3543.
  • 25
    Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada. Can J Diabetes 2008; 32: S1201.
  • 26
    Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 37383.
  • 27
    Pritchard JM, Seechurn T, Atkinson SA. A food frequency questionnaire for the assessment of calcium, vitamin D and vitamin K: a pilot validation study. Nutrients 2010; 2: 80519.
  • 28
    Bischoff HA, Stahelin HB, Monsch AU, Iversen MD, Weyh A, von Dechend M, et al. Identifying a cut-off point for normal mobility: a comparison of the timed ‘up and go’ test in community-dwelling and institutionalized elderly women. Age Ageing 2003; 32: 31520.
  • 29
    Falcao AX, Stolfi J, de Alencar Lotufo R. The image foresting transform: theory, algorithms, and applications. IEEE Trans Pattern Anal Mach Intell 2004; 26: 1929.
  • 30
    Vasilic B, Wehrli FW. A novel local thresholding algorithm for trabecular bone volume fraction mapping in the limited spatial resolution regime of in vivo MRI. IEEE Trans Med Imaging 2005; 24: 157485.
  • 31
    Naccache NJ, Shinghal R. SPTA: a proposed algorithm for thinning binary patterns. IEEE Trans Pattern Anal Mach Intell 1984; 14: 40918.
  • 32
    Garrahan NJ, Mellish RW, Compston JE. A new method for the two-dimensional analysis of bone structure in human iliac crest biopsies. J Microsc 1986; 142: 3419.
  • 33
    Gordon CL, Lang TF, Augat P, Genant HK. Image-based assessment of spinal trabecular bone structure from high-resolution CT images. Osteoporos Int 1998; 8: 31725.
  • 34
    Chappard D, Legrand E, Audran M, Basle MF. Histomorphometric measurement of the architecture of the trabecular bone in osteoporosis: comparative study of several methods. Morphologie 1999; 83: 1720.
  • 35
    Parfitt AM, Mathews CH, Villanueva AR, Kleerekoper M, Frame B, Rao DS. Relationships between surface, volume, and thickness of iliac trabecular bone in aging and in osteoporosis: implications for the microanatomic and cellular mechanisms of bone loss. J Clin Invest 1983; 72: 1396409.
  • 36
    Lu H, Fu X, Ma X, Wu Z, He W, Wang Z, et al. Relationships of percent body fat and percent trunk fat with bone mineral density among Chinese, black, and white subjects. Osteoporos Int 2011; 22: 302935.
  • 37
    Looker AC, Melton LJ III, Harris T, Borrud L, Shepherd J, McGowan J. Age, gender, and race/ethnic differences in total body and subregional bone density. Osteoporos Int 2009; 20: 11419.
  • 38
    Bolton-Smith C, McMurdo ME, Paterson CR, Mole PA, Harvey JM, Fenton ST, et al. Two-year randomized controlled trial of vitamin K1 (phylloquinone) and vitamin D3 plus calcium on the bone health of older women. J Bone Miner Res 2007; 22: 50919.
  • 39
    Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat 1979; 6: 6570.
  • 40
    Vesterby A, Gundersen HJ, Melsen F, Mosekilde L. Marrow space star volume in the iliac crest decreases in osteoporotic patients after continuous treatment with fluoride, calcium, and vitamin D2 for five years. Bone 1991; 12: 337.
  • 41
    Gordon CL, Webber CE, Nicholson PS. Relation between image-based assessment of distal radius trabecular structure and compressive strength. Can Assoc Radiol J 1998; 49: 3907.
  • 42
    Kawashima Y, Fritton JC, Yakar S, Epstein S, Schaffler MB, Jepsen KJ, et al. Type 2 diabetic mice demonstrate slender long bones with increased fragility secondary to increased osteoclastogenesis. Bone 2009; 44: 64855.
  • 43
    Patsch JM, Kiefer FW, Varga P, Pail P, Rauner M, Stupphann D, et al. Increased bone resorption and impaired bone microarchitecture in short-term and extended high-fat diet-induced obesity. Metabolism 2011; 60: 2439.
  • 44
    Akin O, Gol K, Akturk M, Erkaya S. Evaluation of bone turnover in postmenopausal patients with type 2 diabetes mellitus using biochemical markers and bone mineral density measurements. Gynecol Endocrinol 2003; 17: 1929.
  • 45
    Isaia GC, Ardissone P, Di Stefano M, Ferrari D, Martina V, Porta M, et al. Bone metabolism in type 2 diabetes mellitus. Acta Diabetol 1999; 36: 358.
  • 46
    Miyata T, Notoya K, Yoshida K, Horie K, Maeda K, Kurokawa K, et al. Advanced glycation end products enhance osteoclast-induced bone resorption in cultured mouse unfractionated bone cells and in rats implanted subcutaneously with devitalized bone particles. J Am Soc Nephrol 1997; 8: 26070.
  • 47
    Burghardt AJ, Issever AS, Schwartz AV, Davis KA, Masharani U, Majumdar S, et al. High-resolution peripheral quantitative computed tomographic imaging of cortical and trabecular bone microarchitecture in patients with type 2 diabetes mellitus. J Clin Endocrinol Metab 2010; 95: 504555.
  • 48
    Petit MA, Paudel ML, Taylor BC, Hughes JM, Strotmeyer ES, Schwartz AV, et al. Bone mass and strength in older men with type 2 diabetes: the osteoporotic fractures in men study. J Bone Miner Res 2010; 25: 28591.
  • 49
    Shu A, Yin MT, Stein E, Cremers S, Dworakowski E, Ives R, et al. Bone structure and turnover in type 2 diabetes mellitus. Osteoporos Int 2011. E-pub ahead of print.
  • 50
    Schwartz AV, Sellmeyer DE, Vittinghoff E, Palermo L, Lecka-Czernik B, Feingold KR, et al. Thiazolidinedione use and bone loss in older diabetic adults. J Clin Endocrinol Metab 2006; 91: 334954.