SEARCH

SEARCH BY CITATION

Keywords:

  • Bone Microarchitecture;
  • Vertebral Fracture;
  • Fragility Fracture;
  • Men

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Areal bone mineral density (aBMD) measured by dual-energy X-ray absorptiometry (DXA) identifies 20% of men who will sustain fragility fractures. Thus we need better fracture predictors in men. We assessed the association between the low-trauma prevalent fractures and bone microarchitecture assessed at the distal radius and tibia by high-resolution peripheral quantitative computed tomography (HR-pQCT) in 920 men aged 50 years of older. Ninety-eight men had vertebral fractures identified on the vertebral fracture assessment software of the Hologic Discovery A device using the semiquantitative criteria, whereas 100 men reported low-trauma peripheral fractures. Men with vertebral fractures had poor bone microarchitecture. However, in the men with vertebral fractures, only cortical volumetric density (D.cort) and cortical thickness (C.Th) remained significantly lower at both the radius and tibia after adjustment for aBMD of ultradistal radius and hip, respectively. Low D.cort and C.Th were associated with higher prevalence of vertebral fractures regardless of aBMD. Severe vertebral fractures also were associated with poor trabecular microarchitecture regardless of aBMD. Men with peripheral fractures had poor bone microarchitecture. However, after adjustment for aBMD, all microarchitectural parameters became nonsignificant. In 15 men with multiple peripheral fractures, trabecular spacing and distribution remained increased after adjustment for aBMD. Thus, in men, vertebral fractures and their severity are associated with impaired cortical bone, even after adjustment for aBMD. The association between peripheral fractures and bone microarchitecture was weaker and nonsignificant after adjustment for aBMD. Thus bone microarchitecture may be a determinant of bone fragility in men, which should be investigated in prospective studies. © 2011 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Osteoporotic fractures in men are a public health problem because of the morbidity, mortality, and cost.1 However, prediction of fractures in men by areal bone mineral density (aBMD) assessed by dual-energy X-ray absorptiometry (DXA) is disappointing. Low aBMD (gender-specific T-score < −2.5) identified 20% of men who later sustained fragility fractures.2, 3 Quantitative ultrasound (QUS) parameters identify men at high fracture risk similarly to aBMD, but DXA and QUS used jointly do not predict more fractures.4 Classic biochemical bone turnover markers do not improve fracture prediction in men.5, 6 Quantitative computed tomography (QCT) assesses trabecular and cortical bone separately. Low-trauma fractures were associated with low trabecular and cortical volumetric BMD (vBMD), but QCT parameters of femoral neck (ie, cortical volume, trabecular vBMD) did not predict more hip fractures in men than DXA alone.7, 8 Peripheral QCT (pQCT) assesses trabecular and cortical bone at the distal radius and tibia. In young and older men, a history of fracture was associated with lower cortical and trabecular vBMD.9, 10 In addition, cortical bone mineral content measured by pQCT predicted incident peripheral fractures after adjustment for femoral neck aBMD.11

Thus we need better fracture predictors in men. Several studies suggest that assessment of bone microarchitecture may improve the prediction of peripheral fractures in men. Bone strength depends on microarchitecture of cortical and trabecular bone.12 In women with vertebral fractures, trabecular parameters were impaired in those with normal cortical bone, whereas cortex was thinner in those with normal trabecular bone.13 Women with vertebral fractures had thinner cortex and poor trabecular connectivity, as assessed by histomorphometry.14 More severe vertebral fractures were associated with poorer bone microarchiteture.15 However, the bone biopsy is invasive, which limits the use of histomorphometry.

High-resolution peripheral QCT (HR-pQCT) permits assessment of trabecular and cortical bone microarchitecture in clinical studies.16 Postmenopausal women with fractures had lower trabecular and cortical thickness at the tibia after adjustment for aBMD compared with age-matched controls.17 The microarchitectural defect was associated with poor mechanical properties, as assessed with micro–finite element analysis.18 The increasing severity of vertebral fractures was associated with a progressive decrease in cortical thickness and density measured by HR-pQCT.19

In men, the association between bone microarchitecture and fracture is poorly studied. In iliac crest biopsy, osteoporotic men with vertebral fractures had decreased trabecular number and poor trabecular connectivity,20 and men with idiopathic osteoporosis and vertebral fractures had increased cortical porosity.21 Men reporting childhood fractures had lower cortical thickness than controls.10 Therefore, the aim of our study was to assess the association between low-trauma prevalent fractures and bone microarchitecture assessed at the distal radius and tibia by HR-pQCT in 920 men aged 50 years and older from the Structure of Aging Man's Bone (STRAMBO) cohort.

Subjects and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Cohort

The STRAMBO study is a single-center prospective cohort study of skeletal fragility and its determinants in men.22 It was carried out as a collaboration between INSERM (National Institute of Health and Medical Research) and MTRL (Mutuelle des Travailleurs de la Région Lyonnaise). MTRL is a complementary health insurance company open to all citizens. Its insured are representative of the French population from the point of view of age groups and of the proportion between white-collar and blue-collar workers. The study obtained authorization from the local ethics committee and was performed in agreement with the Helsinki Declaration of 1975 and 1983. Participants were recruited in 2006–2008 from the MTRL lists in Lyon. Letters inviting participation were sent to a randomly selected sample of men aged 20 to 85 years living in greater Lyon. Informed consent was provided by 1169 men. All men replied to an interviewer-administered epidemiologic questionnaire that covered lifestyle factors and health status. All men able to give informed consent, to answer the questions, and to participate in the diagnostic examinations were included. No specific exclusion criteria were used. This analysis was made in 920 men aged 50 years and older.

Fracture assessment

Vertebral fractures were assessed on the lateral scans of thoracic and lumbar spine obtained in the dorsal decubitus position by the Vertebral Fracture Assessment (VFA) software using the Hologic Discovery A device equipped with the C-arm (Hologic, Bedford, MA, USA). Using this method, vertebral bodies from T7 to L4 were assessable for all the patients, whereas the T5 and T6 vertebrae were not assessable in 10% to 11% of the men and the T4 vertebral body was not assessable in 23% of the men. Vertebral fractures were assessed by one reader (PS) using the semiquantitative method of Genant,23 as modified for men by Szulc and colleagues.24 Vertebral fractures were identified in 98 men who were classified according to the most severe fracture as follows: grade 1 (n = 18), grade 2 (n = 60), and grade 3 (n = 20). Sixty-one men had one fracture, 23 men had 2 fractures, 7 men had 3 fractures, 2 men had 4 fractures, 1 man had 5 fractures, 3 men had 6 fractures, and 1 man had 7 fractures. Vertebral fractures sustained after a major trauma were excluded. Mild vertebral deformities supposedly related to other conditions (eg, arthritis, Scheuerman disease) were excluded, specificity being preferred to sensitivity. The reproducibility of diagnosis of vertebral fracture was assessed using the simple κ score per fracture (yes versus no) and per grade (no fracture, grade 1, grade 2, or grade 3). The intraobserver agreement scores were κ = 0.93 [95% confidence interval (CI) 0.89–0.98] and κ = 0.89 (95% CI 0.84–0.94), respectively. The interobserver agreement scores were κ = 0.88 (95% CI 0.81–0.95) and κ = 0.87 (95% CI 0.82–0.92), respectively.

Peripheral fractures were assessed using an interviewer-assisted questionnaire. Only fractures that occurred after the age of 18 and after a low trauma were retained as fragility fractures. One hundred men reported 119 fractures at the following skeletal sites: clavicle, 1; scapula, 2; proximal humerus, 11; distal radius, 34; other forearm, 5; ribs, 18; pelvis, 1; hip, 6; femur, 2; tibia, 8; fibula, 7; ankle, 16; heel bone, 3; and metatarsal, 5. Fractures of the face, hand, and toes were excluded. Fractures were self-reported and not ascertained further. Twenty-one men had both vertebral and peripheral fractures.

BMD and bone microarchitecture measurement

Volumetric BMD (vBMD) and microarchitecture were assessed at the nondominant distal radius and right distal tibia by HR-pQCT (XtremeCT, Scanco Medical, Brüttisellen, Switzerland). The arm or leg of the patient was immobilized in a carbon-fiber shell. An anteroposterior scout view was used to define the measured volume of interest (VOI).17 At each site, a stack of 110 parallel CT slices with an isotropic voxel size of 82 µm was obtained, thus delivering a 3D representation of approximately 9 mm in the axial direction. The most distal CT slice was placed 9.5 and 22.5 mm proximal to the endplate of the radius and tibia, respectively. Quality control was performed by daily scans of a phantom containing rods of HA (densities of 0 to 800 mg HA/cm3) embedded in a soft-tissue–equivalent resin (QRM, Moehrendorf, Germany). The coefficient of variation (CV) for the phantom densities varied from 0.05% to 0.9%.

The VOI is separated into cortical and trabecular regions using a threshold-based algorithm. This threshold was set to one-third the cortical vBMD (D.cort). Cortical thickness (C.Th) was defined as the mean cortical volume divided by the outer bone surface. Trabecular vBMD (D.trab, mg HA/cm3) was computed as the average vBMD in the trabecular VOI. Trabecular bone volume (BV) fraction [BV/trabecular volume (TV), %] was derived from D.trab assuming fully mineralized bone to have a mineral density of 1200 mg HA/cm3 {ie, BV/TV (%) = 100 × [D.trab (mg HA/cm3)/1.2 g HA/cm3]}. Trabecular elements were identified by the midaxis transformation method, and the distance between them was assessed three-dimensionally by the distance transform method. Trabecular number (Tb.N, mm−1) was defined as the inverse of the mean spacing of the midaxes. Trabecular thickness (Tb.Th, µm) and separation (Tb.Sp, µm) were derived from BV/TV and Tb.N: Tb.Th = (BV/TV)/Tb.N and Tb.Sp = (1 – BV/TV)/Tb.N. Intraindividual distribution of separation (Tb.SpSD, µm) was quantified by standard deviation of Tb.Sp, a parameter reflecting the heterogeneity of the trabecular network. The CVs for the parameters of the radius and tibia, respectively, were as follows: total vBMD (D.tot), 0.9% and 1.3%; D.cort, 0.7% and 0.9%; C.Th, 1.2% and 0.9%; D.trab, 1.0% and 1.5%; Tb.N, 3.0% and 3.8%; Tb.Th, 3.2% and 4.4%; Tb.Sp, 2.8% and 4.3%; and Tb.NSD, 2.5% and 3.3%. Sixty-nine scans of the distal radius and 29 scans of the distal tibia were excluded because of poor quality owing to movement (nonuniform contour of the cortical bone).

Dual-energy X-ray absorptiometry (DXA)

Areal bone mineral density (aBMD) was measured at the lumbar spine, total hip, and ultradistal nondominant forearm by DXA using the Hologic Discovery A device. The long-term stability of the device was assessed by daily measurements of the commercial phantom of the lumbar spine. The long-term CV of the phantom was 0.35%.

Statistical analysis

All calculations were performed using the SAS Version 9.1 software (SAS Institute, Inc., Cary, NC, USA). Data are presented as mean and SD or as median and interquartile range. Analysis of covariance was used for multivariate comparisons of continuous variables. Tb.Sp and Tb.SpSD had skewed distributions and were log-transformed. The odds ratios (ORs) for the presence of fracture were calculated using logistic regression. The analysis of covariance and logistic regression models for all fractures and vertebral fractures were adjusted for age and weight. The models for the assessment of peripheral fractures were adjusted for age, weight, and height. The variable that was most strongly associated with vertebral fractures was assessed using the stepwise logistic regression. In the logistic regression, we had 90% power to detect a 1.4 increase in the fracture risk per 1 SD change as significant at the level of p < .05. In the bivariate comparisons of two groups, we had 90% power to detect a 0.35 SD difference as significant at the level of p < .05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

All prevalent fractures

Men who had low-trauma fractures were older, slightly shorter, and had a lower aBMD (Table 1). At the distal radius and distal tibia, almost all microarchitectural parameters differed between the men who did and did not have fractures. At the radius, all the parameters became nonsignificant after adjustment for the ultradistal radius aBMD. Some parameters of the distal tibia remained weakly significant after adjustment for total-hip aBMD. In the logistic regression models adjusted for age and weight, almost all the microarchitectural parameters of the distal radius and tibia were significantly associated with the presence of fractures (Table 2). However, almost all the odds ratios became nonsignificant after adjustment for aBMD.

Table 1. Comparison of Men Who Did and Did Not Have Prevalent Fractures
 Fracture (−), n = 743Fracture (+), n = 177p*p**
  • *

    p adjusted for age and weight;

  • **

    p adjusted for age, weight, and aBMD [ultradistal radius aBMD (UD aBMD) for the parameters of the distal radius and total hip aBMD for the parameters of the distal tibia].

    D.tot = total vBMD; D.cort = cortical vBMD; C,Th = cortical thickness; D.trab = trabecular vBMD; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb,Sp = trabecular spacing; Tb.SpSD = trabecular distribution.

Age (years)69 ± 973 ± 8<.001 
Weight (kg)79 ± 1278 ± 11.45 
Height (cm)169.3 ± 6.6167.1 ± 6.1<.001 
Areal bone mineral density (g/cm2)
 Lumbar spine1.052 ± 0.1850.965 ± 0.164<.001 
 Total hip0.971 ± 0.1330.892 ± 0.138<.001 
 UD radius0.469 ± 0.0720.423 ± 0.069<.001 
Distal radius
 D.tot (mg/cm3)299.8 ± 63.2265.3 ± 60.4<.001.85
 D.cort (mg/cm3)809.1 ± 70.2775.2 ± 79.4<.001.79
 C.Th (mm)0.721 ± 0.2200.609 ± 0.204<.001.64
 D.trab (mg/cm3)177.0 ± 38.7157.0 ± 37.2<.001.70
 Tb.N (1/mm)1.87 ± 0.251.76 ± 0.26<.001.25
 Tb.Th (µm)78.6 ± 11.873.9 ± 11.7<.001.58
 Tb.Sp (µm)458 [408; 507]494 [440; 546]<.001.25
 Tb.SpSD (µm)194 [168; 223]215 [185; 256]<.001.32
Distal tibia
 D.tot (mg/cm3)294.3 ± 56.5264.4 ± 57.6<.001.22
 D.cort (mg/cm3)840.6 ± 57.0807.0 ± 85.2<.001.08
 C.Th (mm)1.22 ± 0.291.06 ± 0.31<.001.07
 D.trab (mg/cm3)175.0 ± 37.4157.3 ± 36.9<.001.15
 Tb.N (1/mm)1.75 ± 0.291.63 ± 0.31<.001.05
 Tb.Th (µm)83.3 ± 13.080.6 ± 13.2.06.93
 Tb.Sp (µm)487 [429; 555]529 [459; 603]<.001<.05
 Tb.SpSD (µm)228 [191; 267]248 [211; 304]<.001.06
Table 2. Age- and Weight-Adjusted Association of the Microarchitectural Parameters at the Distal Radius and the Distal Tibia With the Presence of Fractures (All, Vertebral, Peripheral) Without and With Additional Adjustment for Areal Bone Mineral Density
 Distal radius (OR, 95% CI)Distal tibia (OR, 95% CI)
 + UD aBMD + Total-hip aBMD
  • UD aBMD = ultradistal radius areal bone mineral density; D.tot = total vBMD; D.cort = cortical vBMD; C.Th = cortical thickness; D.trab = trabecular vBMD; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular spacing; Tb.SpSD = trabecular distribution.

  • a

    p < .05;

  • b

    p < .01;

  • c

    p < .005;

  • d

    p < .001.

All fractures
 D.tot1.61 (1.32, 1.97)d1.03 (0.73, 1.46)1.59 (1.32, 1.92)d1.15 (0.91, 1.46)
 D.cort1.36 (1.13, 1.64)d1.00 (0.79, 1.26)1.44 (1.22, 1.71)d1.15 (0.94, 1.39)
 C.Th1.54 (1.26, 1.89)d1.08 (0.81, 1.43)1.59 (1.31, 1.91)d1.22 (0.98, 1.52)
 D.trab1.57 (1.30, 1.90)d1.06 (0.77, 1.45)1.54 (1.28, 1.85)d1.17 (0.94, 1.46)
 Tb.N1.45 (1.21, 1.72)d1.11 (0.89, 1.38)1.53 (1.27, 1.83)d1.22 (0.99, 1.49)
 Tb.Th1.39 (1.15, 1.68)d0.93 (0.73, 1.19)1.18 (0.99, 1.41)0.99 (0.82, 1.19)
 Tb.Sp1.38 (1.19, 1.61)d1.08 (0.88, 1.31)1.44 (1.24, 1.68)d1.19 (1.01, 1.41)a
 Tb.SpSD1.22 (1.07, 1,39)c1.01 (0.88, 1.17)1.25 (1.10, 1.43)b1.09 (0.97, 1.23)
Vertebral fractures
 D.tot1.79 (1.38, 2.33)d1.46 (0.91, 2.33)1.79 (1.40, 2.29)d1.28 (0.94, 1.75)
 D.cort1.66 (1.32, 2.07)d1.39 (1.06, 1.84)a1.64 (1.34, 2.01)d1.33 (1.05, 1.68)a
 C.Th1.89 (1.44, 2.49)d1.61 (1.12, 2.32)a1.77 (1.40, 2.25)d1.36 (1.02, 1.79)a
 D.trab1.48 (1.17, 1.86)d0.95 (0.64, 1.40)1.57 (1.25, 1.97)d1.13 (0.85, 1.49)
 Tb.N1.37 (1.11, 1.69)c1.04 (0.80, 1.37)1.56 (1.25, 1.95)d1.19 (0.92, 1.54)
 Tb.Th1.35 (1.06, 1.71)a0.91 (0.67, 1.23)1.17 (0.93, 1.46)0.95 (0.74, 1.20)
 Tb.Sp1.34 (1.13, 1.59)d1.08 (0.87, 1.36)1.45 (1.23, 1.72)d1.19 (0.97, 1.26)
 Tb.SpSD1.21 (1.05, 1.39)b1.05 (0.90, 1.23)1.25 (1.09, 1.43)c1.11 (0.97, 1.26)
Peripheral fractures
 D.tot1.43 (1.12, 1.83)c0.91 (0.58, 1.42)1.43 (1.14, 1.79)c1.04 (0.78, 1.40)
 D.cort1.12 (0.89, 1.42)0.79 (0.58, 1.06)1.28 (1.05, 1.55)a0.99 (0.78, 1.25)
 C.Th1.26 (0.98, 1.61)0.81 (0.57, 1.16)1.42 (1.13, 1.78)c1.07 (0.82, 1.39)
 D.trab1.55 (1.23, 1.96)d1.30 (0.87, 1.94)1.45 (1.15, 1.79)c1.05 (0.80, 1.37)
 Tb.N1.47 (1.19, 1.82)d1.25 (0.95, 1.63)1.49 (1.19, 1.87)d1.11 (0.86, 1.44)
 Tb.Th1.32 (1.04, 1.67)a0.97 (0.71, 1.32)1.10 (0.89, 1.37)0.94 (0.78, 1.19)
 Tb.Sp1.33 (1.13, 1.58)d1.14 (0.91, 1.43)1.32 (1.11, 1.56)c1.06 (0.87, 1.30)
 Tb.SpSD1.19 (1.04, 1.37)a1.06 (0.91, 1.24)1.17 (1.03, 1.33)d1.04 (0.90, 1.19)

One hundred and seventeen men who had one fracture had lower D.tot, C.Th, D.trab, and Tb.N (3.0% to 8.5%, 0.20 to 0.36 SD, p < .05 to p < .001) at both skeletal sites compared with men without fracture. Sixty men who had more than one fracture had lower D.tot, D.trab, Tb.N, D.cort, and C.Th (4.1% to 15.7%, 0.47 to 0.70 SD, p < .005 to p < .001). All the differences became nonsignificant after adjustment for aBMD. At both skeletal sites, Tb.Sp and Tb.SpSD were higher in men with multiple fractures (11.6% to 27.0%, 0.77 to 0.87 SD, p < .001), not in men with one fracture. The differences remained significant after ajustment for aBMD (p < .005).

In the age- and weight-adjusted polytomous logistic regression models, all microarchitectural parameters of distal radius (except D.cort) and tibia (except Tb.Th) were associated with the presence of one fracture (OR = 1.16 to 1.47 per 1 SD change, p = .05 to .001). However, ORs lost significance after additional adjustment for aBMD. In similar models, all the parameters (except Tb.Th) of the distal radius and tibia were associated with the multiple fractures (OR = 1.50 to 2.07 per 1 SD change, p < .001). After adjustment for aBMD, only distal radius Tb.N as well as Tb.Sp and Tb.SPSD of both skeletal sites remained significantly associated with the multiple fractures (OR = 1.50 to 1.82 per 1 SD change, p < .05).

Vertebral fractures

Men who had vertebral fractures were older and had lower aBMD values (Table 3). All the microarchitectural parameters except Tb.Th at the distal tibia differed significantly between the men who did and did not have vertebral fractures. After adjustment for aBMD, men with vertebral fractures still had lower D.cort and C.Th at both skeletal sites.

Table 3. Comparison of Bone Microarchitecture Parameters at the Distal Radius and Distal Tibia According to the Presence of Vertebral Fractures and of Peripheral Fractures
 Vertebral fracturesPeripheral fractures
Fx (−) (n = 822)Fx (+) (n = 98)p*Fx (−) (n = 820)Fx (+) (n = 100)p**
  • *

    Adjusted for age and weight.

  • **

    Adjusted for age, weight, and height.

  • a

    p < .05;

  • b

    p < .005—after additional adjustment for areal bone mineral density (ultradistal radius for the parameters of the distal radius and total hip for the parameters of the distal tibia).

    D.tot = total volumetric bone mineral density (vBMD); D.cort = cortical vBMD; C.Th = cortical thickness; D. trab = trabecular vBMD;

    Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular spacing; Tb.SpSD = trabecular distribution.

Age (years)69 ± 974 ± 8<.00170 ± 972 ± 8<.05
Weight (kg)79 ± 1177 ± 11.2679 ± 1178 ± 11.27
Height169 ± 7166 ± 6<.001169 ± 7167 ± 6<.05
Hip aBMD (g/cm2)0.966 ± 0.1330.875 ± 0.140<.0010.964 ± 0.1350.894 ± 0.142<.001
UD aBMD (g/cm2)0.465 ± 0.0720.419 ± 0.075<.0010.464 ± 0.0730.427 ± 0.068<.001
Distal radius
 D.tot (mg/cm3)298.2 ± 62.9255.2 ± 61.6<.001295.9 ± 64.0274.2 ± 62.4<.005
 D.cort (mg/cm3)809.0 ± 69.9755.0 ± 82.6<.001b804.7 ± 72.8792.1 ± 75.3.35
 C.Th (mm)0.718 ± 0.2170.561 ± 0.206<.001a0.706 ± 0.2230.654 ± 0.206.09
 D.trab (mg/cm3)177.3 ± 38.7155.3 ± 38.7<.001175.0 ± 39.1157.3 ± 36.8<.001
 Tb.N (1/mm)1.86 ± 0.251.75 ± 0.29<.0051.86 ± 0.261.75 ± 0.25<.001
 Tb.Th (µm)78.3 ± 11.973.3 ± 11.2<.0178.1 ± 11.974.6 ± 12.3<.05
 Tb.Sp (µm)460 [412; 511]494 [438; 550]<.001459 [412; 511]496 [452; 545]<.001
 Tb.SpSD (µm)195 [169; 225]216 [185; 260]<.001195 [169; 225]216 [186; 263]<.001
Distal tibia
 D.tot (mg/cm3)293.0 ± 56.1255.5 ± 58.5<.001291.3 ± 57.1268.9 ± 59.9<.001
 D.cort (mg/cm3)839.9 ± 57.8791.5 ± 96.3<.001b837.2 ± 61.1814.8 ± 86.5<.005
 C.Th (mm)1.21 ± 0.291.01 ± 0.33<.001a1.20 ± 0.301.08 ± 0.31<.001
 D.trab (mg/cm3)173.9 ± 37.3154.3 ± 36.2<.001173.4 ± 37.6158.2 ± 37.5<.005
 Tb.N (1/mm)1.75 ± 0.291.61 ± 0.33<.0011.75 ± 0.301.63 ± 0.30<.005
 Tb.Th (µm)83.1 ± 13.180.1 ± 12.1.1083.0 ±12.981.3 ± 14.3.18
 Tb.Sp (µm)490 [432; 558]541 [464; 625]<.001490 [433; 558]528 [469; 603]<.005
 Tb.SpSD (µm)229 [193; 270]252 [218; 303]<.001228 [193; 270]253 [211; 308]<.05

In the age- and weight-adjusted logistic regression models, all the microarchitectural parameters except Tb.Th at the distal tibia were associated with the presence of vertebral fractures (Table 2). After adjustment for aBMD, D.cort and C.Th were associated with the presence of vertebral fractures, whereas the ORs for the trabecular parameters became nonsignificant at both sites. In the stepwise logistic regression, D.cort entered the model as the first and strongest parameter (before aBMD) for both the radius and the tibia. When D.cort was removed from the variables, only C.Th and aBMD were retained in the models for both skeletal sites.

Then men were analyzed according to the number of vertebral fractures. Compared with the controls, 61 men with one fracture had lower D.tot, D.cort, and C.Th at both skeletal sites (p < .01 to p < .001; Fig. 1). D.cort at both skeletal sites and C.Th at the radius were lower (p < .005) after adjustment for aBMD. In 37 men who had two or more vertebral fractures, all the microarchitectural parameters except Tb.Th were significantly different (p < .005 to p < .001) from the controls. After adjustment for aBMD, only Tb.Sp and Tb.SpSD at both skeletal sites (p < .05) remained significantly different from the controls.

thumbnail image

Figure 1. Association between the number of vertebral fractures (0 = no fracture, n = 822; 1 = one vertebral fracture, n = 61; ≥2 = 2 or more vertebral fractures, n = 37) and bone microarchitectural parameters at the distal tibia after adjustment for age and body weight.

Download figure to PowerPoint

In men who had grade 1 fractures, no microarchitectural parameter differed from the control group (Table 4). Men who had grade 2 fractures had lower D.tot, D.cort, and D.trab at both skeletal sites. At the distal tibia, they also had lower Tb.N and higher Tb.Sp and Tb.SpSD. After adjustment for aBMD, D.cort and C.Th at both skeletal sites remained lower compared with controls. In 20 men who had grade 3 fractures, all the microarchitectural parameters except Tb.Th differed significantly from the controls. After adjustment for aBMD, Tb.N, Tb.Sp, and Tb.SpSD remained significantly different from the controls at both skeletal sites.

Table 4. Comparison of Bone Microarchitecture Parameters at the Distal Radius and Tibia According to the Severity of the Vertebral Fractures
 No fracture (n = 822)Grade 1 (n = 18)Grade 2 (n = 60)Grade 3 (n = 20)
  • D.tot = total volumetric bone mineral density (vBMD); D.cort = cortical vBMD; C.Th = cortical thickness; D.trab = trabecular vBMD;

  • Tb.N = trabecular number; Tb,Th = trabecular thickness; Tb.Sp = trabecular spacing; Tb.SpSD = trabecular distribution.

  • a

    p < .05;

  • b

    p < .01;

  • c

    p < .005;

  • d

    p < .001—adjusted for age and weight.

  • p < .05;

  • #

    p < .01;

  • §

    p < .005—after additional adjustment for areal bone mineral density (ultradistal radius for the parameters of the distal radius and total hip for the parameters of the distal tibia).

Age (years)70 ± 972 ± 975 ± 8d74 ± 9a
Weight (kg)79 ± 1181 ± 1179 ± 1171 ± 11c
Bone mineral density (g/cm2)
 Lumbar spine1.044 ± 0.1841.018 ± 0.1600.969 ± 0.182c0.921 ± 0.142b
 Total hip0.964 ± 0.1330.962 ± 0.1150.881 ± 0.124d0.843 ± 0.158d
 UD radius0.464 ± 0.0720.437 ± 0.0540.434 ± 0.071c0.414 ± 0.095c
Distal radius
 D.tot (mg/cm3)296.4 ± 62.1274.4 ± 57.5268.4 ± 57.1c251.2 ± 77.4c
 D.cort (mg/cm3)806.4 ± 69.9777.6 ± 69.4771.6 ± 38.7d,761.8 ± 114.2b
 C.Th (mm)0.712 ± 0.2180.621 ± 0.2170.598 ± 0.191d,0.598 ± 0.246a
 D.trab (mg/cm3)174.8 ± 38.7166.2 ± 28.3163.9 ± 38.7a140.5 ± 41.4d
 Tb.N (1/mm)1.86 ± 0.251.85 ± 0.191.80 ± 0.281.60 ± 0.34d,
 Tb.Th (µm)78.1 ± 11.974.7 ± 10.175.2 ± 10.772.7 ± 14.1
 Tb.Sp (µm)460 [412; 511]468 [433; 507]487 [419; 550]530 [473; 675]d,§
 Tb.SpSD (µm)196 [169; 226]214 [183; 221]211 [183; 264]249 [212; 312]d,
Distal tibia
 D.tot (mg/cm3)292.0 ± 58.0815.9 ± 87.7799.7 ± 96.3d797.4 ± 106.0b
 D.cort (mg/cm3)838.2 ± 58.0815.9 ± 87.7799.7 ± 96.3d,#797.4 ± 106.0b
 C.Th (mm)1.20 ± 0.291.11 ± 0.351.03 ± 0.29d,1.03 ± 0.41a
 D.trab (mg/cm3)173.5 ± 37.4165.0 ± 21.6160.2 ± 34.1a142.2 ± 45.8d
 Tb.N (1/mm)1.75 ± 0.291.70 ± 0.331.65 ± 0.27a1.49 ±0.39d,
 Tb.Th (µm)83.1 ± 13.182.8 ± 14.580.8 ± 11.078.8 ± 13.3
 Tb.Sp (µm)490 [433; 559]494 [418; 594]526 [453; 589]a578 [502; 883]d,§
 Tb.SpSD (µm)229 [194; 271]227 [190; 298]249 [211; 285]a303 [240; 412]d,§

In the polytomous logistic regression models adjusted for age and weight, all the microarchitectural parameters of the distal radius and tibia were significantly associated with the grade 2 and 3 fractures (OR = 1.25 to 1.94 per 1 SD change, p = .05 to p < .001) but not with the grade 1 fractures (OR = 1.02 to 1.68, p > .10). After adjustment for aBMD, only D.cort remained weakly significantly associated with the grade 2 and 3 fractures (radius: OR = 1.33, 95% CI 1.00–1.79, p = .05; tibia: OR = 1.34, 95% CI 1.03–1.75, p < .05).

All the calculations provided similar results when men who reported peripheral fractures but did not have vertebral fractures were excluded from the analysis.

Peripheral fractures

At the distal radius and tibia, almost all the microarchitectural parameters differed between the men who did and did not report low-trauma peripheral fractures (Table 3). After adjustment for aBMD, all the differences became nonsignificant. In the logistic regression models adjusted for age and weight, the microarchitectural parameters of the distal radius (except D.cort and C.Th) and the distal tibia (except Tb.Th) were significantly associated with the presence of peripheral fractures. All the ORs lost significance after adjustment for aBMD.

Eighty-five men who had one peripheral fracture had lower D.trab at both skeletal sites, lower C.Th at the distal tibia as well as at the distal radius, lower Tb.N, and higher Tb.Sp and Tb.SpSD (all p < .05; Fig. 2). After adjustment for aBMD, all the differences became nonsignificant. Fifteen men who had two or more peripheral fractures had lower D.tot, D.trab, and Tb.N as well as higher Tb.Sp and Tb.SpSD at both skeletal sites (p < .05 to p < .001). They also had lower C.Th and D.cort at the distal tibia (p < .005). After adjustment for aBMD, Tb.Sp and Tb.SpSD at both skeletal sites (p < .05 to p < .005) as well as Tb.N at the distal radius and D.cort at the distal tibia (both p < .05) remained significantly different in men with multiple peripheral fractures compared with controls.

thumbnail image

Figure 2. Association between the number of peripheral fractures (0 = no fracture, n = 820; 1 = one peripheral fracture, n = 85; ≥2 = two or more peripheral fractures, n = 15) and bone microarchitectural parameters at the distal tibia after adjustment for age, weight, and height.

Download figure to PowerPoint

Using polytomous logistic regression adjusted for age, weight, and height, D.trab, Tb.N, Tb.Sp, and Tb.SpSD at the distal radius and tibia were associated with multiple fractures (OR = 1.68 to 2.66 per 1 SD change, p < .001). At the distal radius, Tb.N, Tb.Sp, and Tb.SpSD remained associated weakly significantly with multiple fractures after adjustment for the ultradistal radius aBMD (OR = 1.55 to 1.93, p < .05). Associations of D.trab, Tb.N, Tb.Sp, and Tb.SpSD with single peripheral fractures were weaker (OR = 1.23 to 1.47, p < .05) and nonsignificant after adjustment for aBMD. At the distal tibia, D.cort and C.Th were associated with the presence of multiple peripheral fractures (OR = 1.87 and OR = 2.40 per 1 SD change, p < .005). After adjustment for aBMD, the associations of cortical parameters with peripheral fractures became nonsignificant.

When the analyses were limited to 49 men who had sustained the most recent low-trauma peripheral fractures 10 years or less prior to recruitment, all the OR values were similar although less significant owing to lower statistical power. All the calculations provided similar results when men who had vertebral fractures but did not report peripheral fractures were excluded from the analysis.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

In older men, poor cortical bone status was associated with the presence of vertebral fracture independent of aBMD. Multiple fractures and severe (grade 3) vertebral fractures were associated with impaired trabecular microarchitecture. The association between bone microarchitecture and prior peripheral fractures was weaker and, for most of the parameters, became nonsignificant after adjustment for aBMD. Men with vertebral fractures, but not peripheral fractures, were shorter, most probably owing to the vertebral fractures.

Moderate and severe vertebral fractures were associated with poor bone microarchitecture. We excluded high-trauma fractures and nonfracture deformities. Specificity was preferred to sensitivity. The retained fractures are probably fragility fractures. Mild vertebral deformities are often due to arthritis or Scheuerman disease. The incidence of vertebral fractures and arthritic deformities increases with age; however, only fractures are related to low aBMD.25, 26 High-trauma vertebral fractures and Scheuerman disease are more frequent in men than in women, whereas the reverse is the case for the osteoporotic vertebral fracture.27, 28

Vertebral fractures are a hallmark of osteoporosis. Many of them occur without preceding trauma. Bone microarchitecture is a determinant of the strength of vertebral bodies.13 Both men and women with vertebral fractures had poor trabecular microarchitecture.15, 29, 30 Patients with severe or multiple vertebral fractures have lower aBMD values than those with mild or one fracture.31, 32 In postmenopausal women, more severe vertebral fractures were related to poorer bone microarchitecture assessed by HR-pQCT or bone histomorphometry.15, 19 Similar association for bone microarchitecture assessed by bone histomorphometry was found in osteoporotic men.20, 21 Here we show similar trends for HR-pQCT.

Cortical bone seems to play a major role in the pathogenesis of vertebral fractures in men. D.cort entered the stepwise logistic regression models as the strongest predictor of vertebral fractures, even before aBMD. Our results are in line with the data that men with vertebral fractures had higher porosity than men without fractures.17 Men with vertebral fractures also had decreased Tb.N and poor trabecular connectivity in bone biopsy.16, 17 However, in these studies, more men had multiple vertebral fractures than in our cohort. Moreover, in both studies, the men with vertebral fractures were compared not with the general population but with men who had osteoporosis diagnosed by DXA (without fracture).

In postmenopausal women, poor trabecular microarchitecture was the major determinant of vertebral fragility, whereas cortical deterioration played a secondary role.14, 19, 29, 33 However, the morphologic basis of age-related bone loss differs between the sexes. Age-related decrease in trabecular vBMD at the spine and hip and deterioration of trabecular connectivity are greater in women than in men.34–36 The decrease in Tb.N underlying trabecular bone loss in women is more deleterious for bone strength than the trabecular thinning predominating in men.37, 38 Moreover, the age-related decrease (number of SDs below the mean in young men) is greater for D.cort and C.Th than for Tb.N and Tb.Th.39 Thus, in older men, cortical deterioration may be the weakest link determining vertebral fragility.

Men with peripheral fractures had lower D.cort and D.trab, in line with previous data.7–11 In addtion, we show that lower D.trab in these men can be due to the lower number of more heterogeneously distributed trabeculae. However, only men with multiple peripheral fractures had poor bone microarchitecture after adjustment for aBMD. Since peripheral fractures often occur after a fall, health status prior to the fracture may determine the risk of fall and fracture.40, 41 Mechanisms of peripheral fractures vary according to the skeletal site (eg, crush at the wrist, bending for the cervical fracture, torsion at the ankle). Resistance to bending and torsion depends on the bone tissue placed far from the neutral axis of the tubular bone.42 Thus, in the pathogenesis of peripheral fractures, bone size and shape may play a more important role than bone microarchitrecture inside bone.43, 44

Most of the associations became nonsignificant after adjustment for aBMD. For the distal radius, the analyses were adjusted for aBMD of the same skeletal site. Since aBMD subsumes various bone components, it may reflect bone strength better than any single microarchitectural parameter. The associations between bone microarchitecture of the distal tibia and the presence of fragility fractures also weakened after adjustment for aBMD of the total hip, which is a distant skeletal site. It indicates that bone mass at one skeletal site reflects the general skeletal status. This is also consistent with the data that aBMD at one skeletal site predicts fractures at other sites.45

The occurrence, skeletal site, and circumstances of peripheral fractures were self-reported and not ascertained further. There was no formal adjudication or inspection of radiographs or medical records. In half the men who reported peripheral fractures, the most recent fracture had occurred more than 10 years before recruitment. Thus the current status of bone microarchitecture may not correspond with that at the moment when the fracture occurred. By contrast, current microarchitectural status may predict future peripheral fracture regardless of aBMD, especially in a short-term study, as described by Scheu and colleagues.11 In addition, in the study by Scheu and colleagues, incident fractures were ascertained systematically by a physician, and the parameters that were most predictive depended in part on bone size.

The relationship between bone microarchitecture and fractures varied according to the severity and number of fractures. D.cort and C.Th are lower in grade 1 fractures (although nonsignificantly, probably owing to the low number of patients and insufficient power), but they did not differ among the three fracture groups (<0.2 SD between grade 1 and grade 3 groups). D.trab, Tb.N, Tb.Sp, and Tb.SpSD were nearly normal in the grade 1 group and deteriorated with increasing fracture severity (>0.6 SD between grade 1 and grade 3 groups). In men with multiple vertebral or peripheral fractures, microarchitectural parameters were poor compared with controls. However, in these men and in men with grade 3 fractures, only Tb.N, Tb.Sp, and Tb.SpSD were higher than in the controls after adjustment for aBMD.

Thus cortical degradation may be the first signal of lower bone strength in men. D.cort and C.Th in the grade 2 group were significant after adjustment for aBMD, which is moderately decreased and partly driven by the well-preserved trabecular bone. In men with grade 3 vertebral fractures or men with multiple vertebral or peripheral fractures, the decrease in aBMD was driven by the parallel severe cortical and trabecular bone loss. Thus only the parameters of trabecular number and distribution that are more poorly reflected by aBMD remained significant after adjustment for aBMD.

The strengths of our study are the large cohort, evaluation of vertebral fractures using validated criteria, and assessment of the bone microarchitecture at the weight-bearing and non-weight-bearing sites by HR-pQCT. We recognize limitations. The cross-sectional design limits inferences on cause and effect. Older volunteers may not be representative of men in their age range. Semiquantitative diagnosis of vertebral fractures and differentiation between mild vertebral fractures and nonosteoporotic deformities is subjective. Peripheral fractures were not formally adjudicated. In HR-pQCT, Tb.Th Tb.Sp, and C.Th are calculated, not measured. Assessment of microarchitectural parameters may be inaccurate owing to partial-volume effects. In particular, estimation of D.cort and C.Th may be erroneous mainly in the oldest men with thin cortex who had more vertebral fractures. HR-pQCT does not account for the intrinsic age-related deterioration of the cortical bone charaterized by microdamage, imperfections of bone mineral, and abnormalities in posttranslational modifications of bone proteins.46

Thus, in older men, vertebral fractures are associated with poor bone microarchitecture, even after adjustment for aBMD. The association between peripheral fractures and bone microarchitecture was weaker and lost significance after adjustment for aBMD. These cross-sectional data point to the role of bone microarchitecture as an independent determinant of bone fragility in men; however, they need to be confirmed in prospective studies.

Disclosures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

All the authors state that they have no conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

This work was supported by grants from the Roche Pharmaceutical Company, Basle, Switzerland, from Agence Nationale de la Recherche, and from Hospices Civils de Lyon, France.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References