SEARCH

SEARCH BY CITATION

Keywords:

  • RHEUMATOID ARTHRITIS;
  • BONE MICROSTRUCTURE;
  • BONE STRENGTH;
  • INFLAMMATION;
  • HIGH-RESOLUTION PQCT

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

In this cross-sectional study, we investigated volumetric bone mineral density (vBMD), bone microstructure, and biomechanical competence of the distal radius in male patients with rheumatoid arthritis (RA). The study cohort comprised 50 male RA patients of average age of 61.1 years and 50 age-matched healthy males. Areal BMD (aBMD) of the hip, lumbar spine, and distal radius was measured by dual-energy X-ray absorptiometry. High-resolution peripheral quantitative computed tomography (HR-pQCT) of the distal radius provided measures of cortical and trabecular vBMD, microstructure, and biomechanical indices. aBMD of the hip but not the lumbar spine or ultradistal radius was significantly lower in RA patients than controls after adjustment for body weight. Total, cortical, and trabecular vBMD at the distal radius were, on average, –3.9% to –23.2% significantly lower in RA patients, and these differences were not affected by adjustment for body weight, testosterone level, or aBMD at the ultradistal radius. Trabecular microstructure indices were, on average, –8.1% (trabecular number) to 28.7% (trabecular network inhomogeneity) significantly inferior, whereas cortical pore volume and cortical porosity index were, on average, 80.3% and 63.9%, respectively, significantly higher in RA patients. RA patients also had significantly lower whole-bone stiffness, modulus, and failure load, with lower and more unevenly distributed cortical and trabecular stress. Density and microstructure indices significantly correlated with disease activity, severity, and levels of pro-inflammatory cytokines (interleukin [IL] 12p70, tumor necrosis factor, IL-6 and IL-1β). Ten RA patients had focal periosteal bone apposition most prominent at the ulnovolar aspect of the distal radius. These patients had shorter disease duration and significantly higher cortical porosity. In conclusion, HR-pQCT reveals significant alterations of bone density, microstructure, and strength of the distal radius in male RA patients and provides new insight into the microstructural basis of bone fragility accompanying chronic inflammation. © 2014 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Rheumatoid arthritis (RA) is a chronic immune-mediated destructive arthritis characterized by chronic inflammatory synovitis and bone loss. Excessive production of pro-inflammatory cytokines influences bone metabolism by stimulating osteoclast activity and impeding osteoblast function, which leads to progressive bone loss in RA.[1] Several large population-based studies have documented an elevated fracture risk in RA patients compared with controls, most prominent at the hip and spine.[2-4] The increase in fracture risk is similar in male and female patients[3] and cannot be fully explained by decreased areal bone mineral density (aBMD).[5, 6] In the FRAX algorithm, developed by the World Health Organization to evaluate individual 10-year absolute risk of hip and major osteoporotic fracture, RA is the only secondary cause of osteoporosis that is considered independent of aBMD.[7] Although this can be partially attributable to impaired mobility,[8-10] a greater risk of falling,[11] or use of glucocorticoids,[2, 12] other factors, such as a deterioration in bone microstructure and material property that are not captured by aBMD measurements can significantly compromise bone strength[13] and contribute to the increased fracture risk associated with RA.

High-resolution peripheral quantitative computed tomography (HR-pQCT) is a 3D imaging technique specifically developed for in vivo quantification of volumetric BMD (vBMD) and bone microstructure. It allows separate measurement of the cortical and trabecular components of bone at the distal extremities (radius and tibia).[14, 15] The HR-pQCT image data set can be fitted into a micro-finite element (µFE) model to obtain indices of whole bone strength and to determine the relative load distribution between cortical and trabecular compartments.[16] Several HR-pQCT–based studies have investigated the pathophysiology of osteoporosis and its dependence on age, concurrent disease, and therapy.[16-23] We previously reported a HR-pQCT–based case-control study to investigate the microstructural changes at the distal radius in female RA patients.[24] This study showed that female RA patients had substantially lower vBMD as well as inferior bone microstructure and strength compared with controls and that these changes in bone quality were not detected by aBMD using dual-energy X-ray absorptiometry (DXA).

Gender has a significant influence on the clinical manifestations of RA.[25] Despite later age of disease onset, male RA patients are more likely than female patients to have specific features associated with more severe disease, such as nodules, extra-articular manifestations, and greater autoantibody production.[25-27] Previous studies on bone quality in RA patients have mainly been conducted in female patients with the proportion of male patients in these studies being relatively small.[28-30] The few studies conducted exclusively in male RA patients to date are limited by small sample size[31, 32] or by the lack of a well-matched healthy control cohort.[33-35] The aim of this study was to investigate microstructural bone loss in male RA patients. We conducted a case-control study utilizing HR-pQCT to image the distal radius in a cohort of male RA patients and age-matched healthy male controls. We measured aBMD at various anatomic sites using DXA and also measured vBMD, cross-sectional geometry, bone microstructure, and µFE-derived biomechanical indices at the distal radius using HR-pQCT. We also investigated the relationship between these measures and disease characteristics, including serum levels of androgen and pro-inflammatory cytokines.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Subjects

A consecutive cohort of 50 Chinese male patients with a diagnosis of RA was recruited from the outpatient rheumatology clinic of the Prince of Wales Hospital in Hong Kong for this cross-sectional study. All patients fulfilled the 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for RA[36] and were ambulatory at the time of the study. Exclusion criteria were as follows: 1) history of metabolic disorder that could affect bone metabolism, such as severe renal impairment defined as a creatinine clearance of less than 30 mL per minute, thyroid or parathyroid disease, and malignancy; 2) history of treatment that could affect bone metabolism, such as antiresorptive drugs, calcitonin, fluoride, thyroid or parathyroid hormone therapy, and androgen-replacement therapy. Treatment with glucocorticoids, calcium, and/or vitamin supplement was allowed for patients. Fifty age-matched male controls were recruited through word-of-mouth recommendation from the staff of the hospital. All eligible controls were: 1) ambulatory; 2) had no history of autoimmune disease, thyroid or parathyroid diseases, malignancy, or any other major medical problem that required long-term treatment, with the exception of hypertension, diabetes mellitus, and dyslipidemia (under control with or without treatment); and 3) had no history of use of medication that could affect bone metabolism including glucocorticoids. Use of calcium and/or vitamin supplements was allowed for controls. The prevalence of hypertension, diabetes mellitus, and dyslipidemia was 36%, 16%, and 20%, respectively, in controls, which was comparable to that in RA patients (46%, 16%, and 14%, respectively, all p > 0.05). The study protocol was approved by the Joint Chinese University of Hong Kong - New Territories East Cluster Clinical Research Ethics Committee (ref. no. CRE-2011.570) with all participants providing written informed consent.

Demographic, clinical, and laboratory assessment

The following demographics were recorded: age, body weight, body height, smoking and drinking habits, and fracture history of the participants and their first-degree relatives. Only low-trauma fracture, defined as a fracture arising from trauma that would not normally be expected to result in fracture, such as a fall from less than or equal to standing height, was recorded.

Disease duration of RA was defined as the duration between disease onset and date of the study. Disease activity was measured by the Disease Activity Score in 28 Joints (DAS28) using C-reactive protein levels, with higher score indicating higher disease activity. The erythrocyte sedimentation rate (ESR) was measured according to the standard Westergren method.[37] Disease severity was measured by the number of deformed joints and the presence of radiographic erosions at the hands and/or wrists. Rheumatoid factor status and the presence of extra-articular features (including nodules, rheumatic lung disease, rheumatic vasculitis, rheumatic eye disease, Felty's syndrome, and rheumatic neurological disease) were obtained from medical records. Also recorded were current use of disease-modifying antirheumatic drugs (DMARDs), including methotrexate, sulfasalazine, hydroxychloroquine, leflunomide, and azathioprine; current use of biologic agents; and current or past use of glucocorticoids.

Lateral lumbar and thoracic radiographs were obtained using a standardized radiographic technique. A semiquantitative method, developed by Genant,[38] was used to identify and classify vertebrae (T4 to L4) as normal (Grade 0), mild fracture (Grade I, 20% to 25% reduction in anterior, middle, or posterior height), moderate fracture (Grade II, 25% to 40% reduction in vertebral height), or severe fracture (Grade III, >40% reduction in vertebral height). Vertebral fracture was defined as any vertebral body graded as Grade I or above.

Serum total testosterone and sex hormone-binding globulin (SHBG) concentrations were quantified by ELISA using commercial kits (Testosterone Parameter Assay Kit, Hu SHBG Q kit, R&D Systems, Minneapolis, MN, USA). Testosterone deficiency was defined as serum total testosterone level less than 10.4 nmol/L (300 ng/dL). Serum levels of four pro-inflammatory cytokines, namely interleukin 12p70 (IL-12p70), tumor necrosis factor (TNF), IL-6, and IL-1β, were simultaneously quantified by BD FACS caliber flow cytometry (Becton Dickinson, San Jose, CA, USA) using Human Cytokine Cytometric Bead Array reagent kits (BD Biosciences, San Diego, CA, USA) according to manufacturer instructions.

DXA assessment

aBMD of the hip (total left hip and femoral neck), lumbar spine (L1 to L4), and ultradistal radius of the nondominant forearm was performed on standard DXA equipment (Hologic Delphi W, Bedford, MA, USA) by a technician certified by the International Society of Clinical Densitometry. aBMD of the hip was not obtained in one RA patient because of bilateral hip replacement. DXA scan of the ultradistal radius spanned for 15 mm, starting 10 mm proximal to the ulnar tip. Our short-term precision error of aBMD by DXA, expressed as the coefficient of variance (CV), ranged from 0.72% (lumbar spine) to 1.5% (femoral neck).[39] aBMD results were expressed in g/cm2 and T-score (femoral neck, total hip, and lumbar spine) calculated with reference to a local population norm.[39]

HR-pQCT imaging and image analyses

Cross-sectional geometry, vBMD, and microstructure were measured at the distal radius of the nondominant forearm using HR-pQCT (Scanco Medical AG, Bruttisellen, Switzerland). The participant's forearm was immobilized in a carbon fiber cast fixed within the scanner gantry. A dorsal-palmer projection image was obtained to define the tomographic scan region. The scan region was fixed 9.5 mm proximal from the mid-joint line and spanned 9.02 mm proximally in length, equivalent to a stack of 110 slices with an isotropic voxel size of 82 µm.[14]

Images were first evaluated using a standard protocol.[14] The entire volume of interest (VOI) was automatically separated into cortical and trabecular components, yielding average and trabecular vBMD in mg hydroxyapatite (HA)/cm3. Trabecular vBMD was calculated for two regions: the peripheral region adjacent to the cortex and a central medullary region. Trabecular bone volume fraction (BV/TV) was derived from trabecular vBMD, assuming fully mineralized bone to have a mineral density of 1.2 g HA/cm3. Trabecular number was defined as the mean inverse distance between 3D ridges (the centroid of trabeculae) by the distance-transformation method.[40] Trabecular thickness and separation were derived from BV/TV and trabecular number using standard histomorphometry methods. The standard deviation (SD) of trabecular separation was used to reflect trabecular network inhomogeneity. The orientation of the trabecular network was quantified using the Structure Model Index (SMI).[41] For an ideal plate and rod structure, the SMI value would be 0 and 3, respectively, the closer the SMI to 0 the greater the plate structure, whereas the closer the SMI to 3 the greater the rod structure. Studies have shown that plate structure confers a greater mechanical strength than rod structure.[42] Our short-term HR-pQCT reproducibility, expressed as CV, ranged from 0.38% to 1.03% for density measures and from 0.80% to 3.73% for microstructural measures.[24]

A fully automated cortical compartment segmentation technique was used to evaluate cortical bone density and microstructure.[43] Cortical vBMD was the mean mineralization of all voxels in the cortical VOI. The cortical to total cross-sectional area ratio yielded the cortical area fraction. The average maximal second moment of inertia (MOI) for the cortex (Ct. MOI) and the total cross-sectional (total MOI) were calculated from the segmented binary images as estimates of bending strength. A volumetric index of cortical porosity, denoted Ct. Po (%), was calculated from the cortical pore volume (Ct. PoV) divided by the mineralized cortical bone volume.[44] Other variables related to cortical microstructure included cortical pore diameter and distribution of cortical pore diameter (the SD of cortical pore diameter). Cortical thickness (ie, endosteal-periosteal distance) was determined by composite segmentation of the mineralized cortex, disregarding any intracortical pore surface in these calculations.

Fig. 1 shows representative HR-pQCT images of the distal radius for a median (by trabecular vBMD) RA patient and a median control. The automated cortical bone compartment segmentation was qualitatively acceptable in all subjects except for 10 RA patients. In these 10 patients, there was focal periosteal bone apposition, leading to focal cortical thickening beyond its normal boundary on the ulnovolar aspect of the distal radius, along with exaggerated endocortical trabecularization[45] (Fig. 2). Because of poor automated recognition of the endocortical boundary, manual adjustment to the endocortical contour was required for these 10 patients and the apposed bone included as cortical compartment. Apart from this focal periosteal bone apposition, we did not observe any other geometric abnormalities attributable to the disease at the scan region, such as a significant break in the cortex, which might interfere with the image analyses.

image

Figure 1. Representative HR-pQCT images of the distal radius of a median (by trabecular volumetric bone mineral density) control (top) and a median rheumatoid arthritis (RA) patient (bottom): distal-most slices (A, E); proximal-most slice (B, F); 3D visualization of the mineralized cortical bone (transparent gray), cortical porosity (solid red), and trabecular bone (solid green) (C, G); and 3D visualization of cortices (transparent gray) and cortical pores (solid red) (D, H).

Download figure to PowerPoint

image

Figure 2. HR-pQCT images of the distal radius from a representative rheumatoid arthritis patient with focal periosteal bone apposition leading to undue cortical thickening and a change in cortical morphology on the ulnovolar aspect of the distal radius (arrows), particularly toward the more distal end of the radius: slice no. 1 (A), slice no. 26 (B), 3D visualization of the cortex (transparent gray) and cortical pores (solid red) (C, D).

Download figure to PowerPoint

µFE analysis

All µFE analyses were performed using FE-solver included in the built-in Image Processing Language software. A special peeling algorithm, specifying a minimum cortical thickness of 6 voxels, was used to identify cortical and trabecular bone tissue. µFE analysis was performed by converting the binary image data to a mesh of isotropic brick elements.[46] Different elastic properties were specified for cortical and trabecular bone tissue. For all elements, a Poisson's ratio of 0.3 was specified. Elements representing cortical bone were assigned a Young's modulus of 20 GPa, and those representing trabecular bone tissue were assigned a Young's modulus of 17.5 GPa.[47] A uniaxial compression test with a 1000-N load was performed with an applied strain of 1% for obtaining the following indices: stiffness, apparent modulus, failure load, percentage of load carried by cortical bone at the distal and proximal surface of the VOI, average and SD of von Mises stresses for trabecular and cortical bone, and trabecular and cortical equivalent strain.

Statistical analysis

Statistical analyses were performed using the IBM Statistics Package for Social Sciences (IBM SPSS Statistics 20, SPSS Inc, Chicago, IL, USA). Demographics for RA patients and controls were compared using Student's t test or Mann-Whitney U test or chi-square test, depending on type and distribution of the data. Normality testing was performed using a combination of Kolmogorov-Smirnov test and histogram for all densitometric, geometric, microstructural, and biomechanical indices. Normally distributed indices were expressed as mean ± SD and were compared between two groups using Student's t test. Four indices—trabecular network inhomogeneity, cortical pore volume, SD of cortical stress and cortical equivalent strain—were not normally distributed, and they were expressed as median (interquartile range) and compared using Mann-Whitney U test. The relationships between densitometric, geometric, microstructural, biomechanical indices and variables related to disease activity/severity, and serum levels of pro-inflammatory cytokines were examined using Spearman correlation. Within patients and controls, we did not find a significant association between aBMD at the ultradistal radius and body weight (p > 0.05). To determine whether any differences between patients and controls were driven by body weight, testosterone level, or aBMD at the ultradistal radius, analyses of covariance (ANCOVA) were performed for statistical significance with body weight/testosterone level/aBMD at the ultradistal radius as a covariate. Interactions between body weight/testosterone level/aBMD at the ultradistal radius and each dependent variable were tested with homogeneity of regression slopes between groups being assumed. To determine whether differences between patients and controls were driven by smoking habit, the interaction between smoking habit (current smoker versus nonsmoker) and groups (patients versus controls) was tested using a two-way analysis of variance (ANOVA). Finally, densitometric, geometric, microstructural, and biomechanical indices among the 10 particular RA patients with focal periosteal bone apposition, other RA patients, and controls were compared using ANOVA with Bonferroni-corrected pairwise comparisons. For the four indices that were not normally distributed, log transformation (base 10) was conducted to assume normal distribution and ANCOVA/ANOVA tests performed on the transformed data. All hypotheses were two-tailed, and a p value <0.05 was considered statistically significant.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Characteristics of study participants

Characteristics of the study participants are shown in Table 1. RA patients and controls were comparable according to age, body height, drinking habit, personal history, and family history of fragility fracture. Body weight was significantly lower and the percentage of current smokers significantly higher for RA patients. Compared with controls, patients had a significantly lower serum level of testosterone but a similar level of SHBG. Prevalence of testosterone deficiency was also significantly higher in patients than in controls (46% versus 8%, p < 0.0001). Adjustments of body weight and testosterone level were added to univariate between-group comparisons but did not change the results significantly (Table 2). Two-way ANOVA testing the interaction between smoking habit and groups (patients versus controls) showed all interactive p > 0.05, indicating that the differences between patients and controls did not depend on their smoking habit (data not shown). The prevalence of vertebral fracture (Grade I or above) was marginally higher in patients than in controls (18% versus 6%, p = 0.065). Vertebral fracture at the lumbar spine (L1 to L4) was found in 7 patients (14%) and 2 controls (4%) (p = 0.160). Levels of all pro-inflammatory cytokines, including IL-12p70, TNF, IL-6, and IL-8, were significantly higher in RA patients.

Table 1. Characteristics of Study Participants
VariablesPatients (n = 50)Controls (n = 50)p Value
  1. SHBG = sex hormone binding globulin; IL = interleukin; TNF = tumor necrosis factor; DMARDs = Disease Modifying Anti-Rheumatic Drugs.

  2. Results are mean ± SD or median (interquartile range) unless otherwise indicated. Boldface indicates statistically significant difference.

  3. a

    At hands and/or wrists.

  4. b

    Doses are prednisolone-equivalent.

Demographics   
Age (years)61.1 ± 8.561.3 ± 8.40.933
Body weight (kg)61.1 ± 9.267.1 ± 6.9<0.0005
Body height (m)1.65 ± 0.051.67 ± 0.060.075
Current smoker (no. [%])21 (42)7 (14)0.002
Current drinker (no. [%])4 (8)4 (8)
Fracture, 1st-degree relative (no. [%])4 (8)2 (4)0.678
Fracture, at age >25 years (no. [%])3 (6)1 (2)0.617
Vertebral fracture (no. [%])9 (18)3 (6)0.065
Hormones and pro-inflammatory cytokines   
Testosterone (nmol/L)15.9 ± 12.523.4 ± 12.10.003
SHBG (nmol/L)78.8 ± 44.783.3 ± 32.10.563
IL-12p70 (pg/mL)3.3 ± 3.42.0 ± 0.50.033
TNF (pg/mL)8.2 ± 10.64.2 ± 1.70.001
IL-6 (pg/mL)6.1 ± 6.53.8 ± 2.60.020
IL-1β (pg/mL)10.4 ± 11.57.4 ± 1.20.018
Disease characteristics   
Disease duration since onset (years)12.3 (5.5, 9.0)
Disease Activity Score in 28 joints3.30 ± 1.45
No. of deformed joint(s)2 (0, 6.3)
Deformed wrist(s) (no. [%])25 (50)
Erosive diseasea (no. [%])34 (68)
Currently on DMARDs (no. [%])44 (88)
Currently on glucocorticoidsb (no. [%])18 (36)
Current dose (mg/d)5 (2.5, 10)
Cumulative dose (g)3.44 (1.09, 7.47)
Cumulative duration (months)30.1 (8.4, 44.9)
Previously on glucocorticoids (no. [%])11 (22)
Last use of glucocorticoids (years)3.6 (0.9, 16.2)
Cumulative dose (g)1.08 (0.64, 2.20)
Cumulative duration (months)5.2 (3.2, 7.2)
Exposure to glucocorticoids 6 months before the study (no. [%])19 (38)
Table 2. Density, Cross-Sectional Geometry, Microstructure, and Biomechanical Indices for Rheumatoid Arthritis Patients and Controls
VariablesPatients (n = 50)Controls (n = 50)% DifferenceaUnadjusted p valueAdjusted p valuebAdjusted p valuecAdjusted p valued
  1. BMD = bone mineral density; vBMD = volumetric bone mineral density; HA = hydroxyapatite; Tb = trabecular; pTb = trabecular bone in the peripheral region adjacent to the cortex; mTb = trabecular bone in the central medullary region; Ct = cortical; MOI = moment of inertia; BV/TV = trabecular bone volume fraction; Ct. PoV = cortical pore volume; Ct. Po = cortical porosity index; Ct. Po. Dm = cortical pore diameter; Ct. Po. Dm. SD = standard deviation of Ct. Po. Dm.

  2. Results are mean ± SD or median (interquartile range). Boldface indicates statistically significant difference.

  3. a

    Calculated based on the difference of the two means.

  4. b

    Values of p are adjusted by body weight.

  5. c

    Values of p are adjusted by serum level of testosterone.

  6. d

    Values of p are adjusted by areal BMD of ultradistal radius.

Areal BMD (g/cm2)       
Femoral neck0.66 ± 0.100.74 ± 0.09–10.6<0.00050.0260.0020.002
Total hip0.82 ± 0.160.95 ± 0.12–13.0<0.00010.009<0.0005<0.0005
Lumbar spine0.95 ± 0.160.98 ± 0.18–2.80.4230.4140.4550.854
Ultradistal radius0.43 ± 0.100.47 ± 0.07–8.70.0200.1220.096
vBMD (mgHA/cm3)       
Total vBMD310.4 ± 76.9358.4 ± 68.3–13.40.0010.0040.0100.021
Tb. vBMD124.6 ± 36.3162.2 ± 34.2–23.2<0.0001<0.0001<0.0001<0.0001
pTb. vBMD190.4 ± 36.3222.6 ± 32.3–14.5<0.0001<0.0005<0.0001<0.0001
mTb. vBMD79.1 ± 39.0120.3 ± 37.8–34.2<0.0001<0.0001<0.0001<0.0001
Ct. vBMD926.8 ± 74.3964.4 ± 44.3–3.90.0030.0050.0100.026
Cross-sectional geometry       
Cortical area fraction0.253 ± 0.0720.263 ± 0.052–3.80.4310.4460.9070.119
Total MOI (mm4)2742 ± 11642645 ± 6913.60.6170.1240.4070.040
Ct. MOI (mm4)2044 ± 9891970 ± 5033.80.6360.1450.3440.006
Trabecular microstructure       
BV/TV0.10 ± 0.030.14 ± 0.03–23.2<0.0001<0.0001<0.0001<0.0001
Tb. number (mm−1)1.53 ± 0.301.66 ± 0.24–8.10.0150.1510.0270.109
Tb. thickness (mm)0.067 ± 0.0110.082 ± 0.015–17.7<0.0001<0.0001<0.0001<0.0001
Tb. separation (mm)0.62 ± 0.150.53 ± 0.1015.40.0020.0330.0050.021
Inhomogeneity (mm)0.26 (0.22, 0.32)0.22 (0.20, 0.25)28.70.0010.0150.00420.008
Structure model index2.31 ± 0.391.77 ± 0.4130.2<0.0001<0.0001<0.0001<0.0001
Cortical microstructure       
Ct. thickness (mm)1.12 ± 0.341.14 ± 0.20–1.70.7240.8960.6270.002
Ct. PoV (mm3)22.8 (12.7, 33.7)14.2 (9.2, 21.5)80.30.0010.0020.0021<0.0005
Ct. Po (%)4.28 ± 2.492.61 ± 1.5263.9<0.00050.001<0.0005<0.0005
Ct. Po. Dm (mm)0.183 ± 0.0170.171 ± 0.0146.7<0.00050.012<0.0001<0.0005
Ct. Po. Dm. SD (mm)0.080 ± 0.0130.074 ± 0.0138.60.0160.1070.0090.010
Biomechanical indices       
Stiffness (kN/mm)80.3 ± 25.492.7 ± 17.2–13.40.0050.0660.0430.093
Modulus (Mpa)1843 ± 5462174 ± 480–15.20.0020.0050.0110.036
Failure load (N)3988 ± 12254629 ± 838–13.90.0030.0470.0260.037
% load Ct. distal end58.6 ± 11.851.0 ± 6.114.8<0.00050.001<0.0001<0.0001
% load Ct. proximal end87.2 ± 7.384.8 ± 4.42.80.0500.0800.0260.024
Tb. average stress (MPa)45.7 ± 6.955.1 ± 7.6–17.1<0.0001<0.0001<0.0001<0.0001
Tb. SD stress (MPa)31.5 ± 1.929.5 ± 1.56.9<0.0001<0.0001<0.0001<0.0001
Ct. average stress (MPa)80.5 ± 5.083.1 ± 2.8–3.20.0010.0030.0030.008
Ct. SD stress (MPa)20.6 (18.9, 23.2)19.2 (18.0, 20.4)12.10.001<0.0005<0.0001<0.0001
Tb. equivalent strain0.0084 ± 0.00110.0099 ± 0.0012–15.5<0.0001<0.0001<0.0001<0.0001
Ct. equivalent strain0.0137 (0.0133, 0.0140)0.0140 (0.0138, 0.0141)–2.7<0.0001<0.0005<0.0001<0.0005

In general, the RA patient cohort consisted of patients with mild disease activity and severity (Table 1). The majority (68%) of patients had late disease onset (>45 years old) and half had deformed wrist(s). Twenty-four (48%) patients had radiographic erosion at the wrist(s). Positive rheumatoid factor was found in 80% and extra-articular features in 32% of patients. The most common extra-articular features were nodules (14%), pulmonary fibrosis (10%), and C1/2 subluxation (8%). Most patients (88%) were currently on synthetic DMARDs, whereas only 2 patients were currently on biologic agents. Eighteen (36%) patients were currently on low-dose glucocorticoids.

aBMD

aBMD at the femoral neck, total hip, and ultradistal radius was significantly lower in RA patients than in controls (Table 2). These differences did not depend on smoking habit or testosterone level but were affected by body weight. After adjustment of body weight, differences in aBMD at the femoral neck and total hip decreased but remained significant (adjusted percentage differences –5.9% and –7.2%, respectively), whereas difference in aBMD at the ultradistal radius became insignificant (adjusted percentage difference –6%, p = 0.122). aBMD at the lumbar spine was comparable between patients and controls, even when those with lumbar vertebral fracture were excluded (data not shown). The prevalence of osteoporosis (T-score ≤–2.5) at the femoral neck, total hip, and lumbar spine in patients was 6.1%, 14.3%, and 10%, respectively, whereas the corresponding figures in controls were 0%, 0%, and 8%, with significant difference observed only in the prevalence of osteoporosis at the total hip (p = 0.006).

Geometry, vBMD, and microstructure

In contrast to aBMD at the ultradistal radius, total (–13.4%), cortical (–3.9%), and trabecular vBMD (–23.2%) values were significantly lower in RA patients than controls, and these differences were not affected by adjustment for body weight, smoking habit, testosterone level, or aBMD at the ultradistal radius (Table 2). The most marked difference in vBMD was found in trabecular vBMD in the central medullary region (–34.2%).

There were insignificant differences in indices of cross-sectional geometry, including cortical area fraction and total and cortical areal MOI, between RA patients and controls. After adjustment for aBMD at the ultradistal radius, total and areal MOI became significantly higher in patients. All indices of trabecular microstructure were significantly inferior in patients than in controls, with percentage differences ranging from –8.1% (trabecular number) to 28.7% (trabecular network inhomogeneity). SMI value was also significantly higher in patients, indicating a more rodlike trabecular network. Cortical thickness was similar between patients and controls, although cortical porosity was significantly higher in patients. Cortical pore volume and Ct. Po were, on average, 80.3% and 63.9%, respectively, higher in patients than in controls. The average cortical pore diameter was also higher with larger pores being more common in patients. Group-wise differences in microstructure did not depend on smoking habit or testosterone level, although adjustment for body weight did negate significance with regard to both trabecular number (adjusted p = 0.151) and SD of cortical pore diameter (adjusted p = 0.107), without affecting the significance of the other indices. The majority of the group-wise differences were independent of aBMD at the ultradistal radius with the exception of differences in trabecular number (adjusted p = 0.109) and cortical thickness (adjusted p = 0.002).

µFE-derived biomechanical indices

RA patients had significantly lower whole-bone stiffness, apparent modulus, and failure load than controls. Particularly at the more distal end of the scanned area, the proportion of load carried by cortical bone was significantly larger in patients than in controls. At the more proximal end, this difference was only marginally significant (p = 0.050) and became significant after adjustment for aBMD at the ultradistal radius (adjusted p = 0.024). Cortical and trabecular stress and equivalent strain were both significantly lower in patients, with stress being more unevenly distributed. After adjustment for body weight or aBMD at the ultradistal radius, group-wise differences in stiffness were no longer significant, whereas differences in other biomechanical indices remained significant.

Relationship with disease characteristics and pro-inflammatory cytokines

Table 3 and Table 4 show the relationship between indices of vBMD, geometry, microstructure, and biomechanical competence and indices of disease characteristic. In general, higher disease activity (higher ESR and DAS28 score) tended to be associated with lower areal MOI and greater impairment of trabecular microstructural indices (lower vBMD, lower BV/TV, thinner trabeculae, and lower stress and strain). Conversely, more chronic disease (longer disease duration) and more severe disease (higher number of deformed joints) tended to be associated with larger cortical area fraction and thickness but with a greater deficiency in vBMD and microstructure for both cortical and trabecular compartments, as well as lower and more unevenly distributed stresses in both compartments.

Table 3. Correlation Coefficients Between Indices of Volumetric Density, Cross-Sectional Geometry, Microstructure, and Biomechanics and Indices of Disease Activity, Severity, Use of Glucocorticoids, and Serum Levels of Pro-Inflammatory Cytokines in Rheumatoid Arthritis Patients
VariablesESR (mm/1st h)DAS 28 scoreDisease duration (years)No. of deformed jointsIL-12 (pg/mL)TNF (pg/mL)IL-6 (pg/mL)IL-1β (pg/mL)Cumulative glucocorticoid dose (g)Cumulative glucocorticoid duration (months)
  1. ESR = erythrocyte sedimentation rate; DAS 28 score = Disease Activity Score in 28 joints; IL = interleukin; TNF = tumor necrosis factor; vBMD = volumetric bone mineral density; HA = hydroxyapatite; Tb = trabecular; pTb = trabecular bone in the peripheral region adjacent to the cortex; mTb = trabecular bone in the central medullary region; Ct = cortical; MOI = moment of inertia; BV/TV = trabecular bone volume fraction; Ct. PoV = cortical pore volume; Ct. Po = cortical porosity index; Ct. Po. Dm = cortical pore diameter; Ct. Po. Dm. SD = standard deviation of Ct. Po. Dm.

  2. Boldface indicates statistically significant correlation (*p < 0.05, **p < 0.01).

vBMD (mgHA/cm3)          
Total vBMD–0.294*–0.2680.2440.077–0.334*–0.180–0.289*–0.246–0.112–0.090
Tb. vBMD–0.432**–0.238–0.178–0.348*–0.218–0.057–0.039–0.1870.0350.042
pTb. vBMD–0.502**–0.282*–0.058–0.292*–0.221–0.077–0.132–0.2150.0250.030
mTb. vBMD–0.340*–0.192–0.244–0.362**–0.211–0.0570.017–0.1680.0490.045
Ct. vBMD0.0550.0000.0770.188–0.188–0.149–0.210–0.058–0.298*–0.273
Cross-sectional geometry          
Cortical area fraction–0.214–0.2550.405**0.267–0.321*–0.233–0.313*–0.207–0.132–0.117
Total MOI (mm4)–0.470**–0.154–0.190–0.396**–0.219–0.081–0.141–0.410**0.1470.138
Ct. MOI (mm4)–0.483**–0.194–0.093–0.315*–0.314*–0.197–0.255–0.451**0.0410.045
Trabecular microstructure          
BV/TV–0.431**–0.237–0.178–0.352*–0.214–0.052–0.032–0.1850.0320.039
Tb. number (mm−1)–0.291*–0.098–0.391**–0.397**–0.132–0.0170.063–0.1310.1940.211
Tb. thickness (mm)–0.388**–0.2660.116–0.172–0.259–0.088–0.144–0.230–0.136–0.160
Tb. separation (mm)0.331*0.0910.347*0.404**0.1660.035–0.0330.159–0.160–0.176
Inhomogeneity (mm)0.2670.1140.329*0.466**0.024–0.029–0.1060.063–0.130–0.151
Structure model index0.328*0.2030.291*0.529**0.033–0.127–0.1140.054–0.039–0.046
Cortical microstructure          
Ct. thickness (mm)–0.279*–0.2760.362**0.176–0.395**–0.271–0.366**–0.308*–0.154–0.123
Ct. PoV (mm3)–0.333*–0.1990.118–0.118–0.139–0.151–0.100–0.2360.1300.118
Ct. Po (%)–0.21–0.1830.053–0.146–0.010–0.0270.063–0.1030.2420.218
Ct. Po. Dm (mm)–0.081–0.1190.295*0.200–0.048–0.258–0.072–0.051–0.010–0.043
Ct. Po. Dm. SD (mm)–0.046–0.0270.2500.331*–0.109–0.338*–0.186–0.0700.007–0.030
Biomechanical indices          
Stiffness (kN/mm)–0.468**–0.285*0.013–0.239–0.384**–0.215–0.314*–0.419**–0.073–0.046
Modulus (MPa)–0.216–0.1880.2250.066–0.268–0.113–0.211–0.104–0.242–0.199
Failure load (N)0.488**–0.281*–0.039–0.293*–0.381**–0.205–0.305*–0.443**0.0570.031
% load Ct. distal end0.1120.0430.419**0.515**–0.079–0.166–0.2760.004–0.147–0.123
% load Ct. proximal end–0.035–0.1480.309*0.236–0.268–0.310*–0.319*–0.187–0.0040.033
Tb. average stress (MPa)–0.407**–0.336*0.016–0.346*–0.1020.087–0.024–0.127–0.176–0.176
Tb. SD stress (MPa)0.1680.0030.371**0.429**–0.011–0.057–0.1550.025–0.138–0.122
Ct. average stress (MPa)0.0130.0140.0100.058–0.1010.008–0.1100.057–0.235–0.183
Ct. SD stress (MPa)–0.110.0490.1660.1570.044–0.113–0.114–0.0920.0690.018
Tb. equivalent strain–0.414**–0.356*0.001–0.373**–0.1010.080–0.025–0.128–0.141–0.143
Ct. equivalent strain0.079–0.0650.004–0.0960.0750.1620.1050.187–0.077–0.035
Table 4. Density, Cross-Sectional Geometry, Microstructure, and Biomechanical Indices for RA Patients by Wrist Deformity and by Erosive Change at Wrist Joint
VariablesDeformed wrist(s)Wrist erosion
Yes (n = 25)No (n = 25)% Differenceap ValueYes (n = 24)No (n = 26)% Differenceap Value
  1. RA = rheumatoid arthritis; vBMD = volumetric bone mineral density; HA = hydroxyapatite; Tb = trabecular; pTb = trabecular bone in the peripheral region adjacent to the cortex; mTb = trabecular bone in the central medullary region; Ct. = cortical; MOI = moment of inertia; BV/TV = trabecular bone volume fraction; Ct. PoV = cortical pore volume; Ct. Po = cortical porosity index; Ct. Po. Dm = cortical pore diameter; Ct. Po. Dm. SD = standard deviation of Ct. Po. Dm.

  2. Results are mean ± SD or median (interquartile range). Boldface indicates statistically significant difference.

  3. a

    Calculated based on the difference of the two means.

vBMD (mgHA/cm3)
Total vBMD337.0 ± 74.8283.9 ± 70.718.70.013329.7 ± 76.1292.7 ± 74.612.60.089
Tb. vBMD120.1 ± 32.4129.0 ± 39.9–6.90.394123.9 ± 30.1125.2 ± 41.7–1.00.905
pTb. vBMD187.8 ± 34.7193.0 ± 38.4–2.70.621189.6 ± 33.0191.1 ± 39.7–0.80.887
mTb. vBMD73.4 ± 34.884.9 ± 42.8–13.60.30278.5 ± 33.079.7 ± 44.5–1.50.913
Ct. vBMD940.1 ± 76.6913.5 ± 71.02.90.209934.2 ± 77.9920.0 ± 71.71.50.505
Cross-sectional geometry
Cortical area fraction0.287 ± 0.0750.219 ± 0.05231.00.0010.276 ± 0.0810.233 ± 0.05818.40.036
Total MOI (mm4)2,724 ± 1,4982,759 ± 722–1.30.9162,750 ± 1,4972,734 ± 7710.60.964
Ct. MOI (mm4)2,160 ± 1,2991,929 ± 53212.00.4172,141 ± 1,3231,954 ± 5379.60.524
Trabecular microstructure
BV/TV0.10 ± 0.030.11 ± 0.03–6.90.3940.10 ± 0.030.10 ± 0.03–0.90.914
Tb. number (mm−1)1.42 ± 0.261.63 ± 0.31–12.80.0131.51 ± 0.281.54 ± 0.33–2.30.680
Tb. thickness (mm)0.070 ± 0.0110.065 ± 0.0117.00.1480.068 ± 0.0120.066 ± 0.0112.70.576
Tb. separation (mm)0.66 ± 0.160.57 ± 0.1315.80.0360.62 ± 0.150.61 ± 0.161.30.852
Inhomogeneity (mm)0.26 (0.20, 0.27)0.27 (0.24, 0.43)28.80.0210.26 (0.22, 0.34)0.26 (0.22, 0.32)2.50.977
Structure model index2.40 ± 0.352.21 ± 0.408.90.0712.34 ± 0.332.27 ± 0.443.00.540
Cortical microstructure
Ct. thickness (mm)1.26 ± 0.390.99 ± 0.2326.90.0051.21 ± 0.411.04 ± 0.2416.90.072
Ct. PoV (mm3)22.5 (12.4, 28.8)26.7 (13.0, 43.4)72.00.30823.9 (11.2, 32.0)21.3 (13.1, 37.4)47.00.892
Ct. Po (%)4.58 ± 2.913.98 ± 2.0015.00.4014.35 ± 2.804.22 ± 2.213.10.855
Ct. Po. Dm (mm)0.192 ± 0.0170.174 ± 0.01210.0<0.00050.189 ± 0.0180.177 ± 0.0156.70.014
Ct. Po. Dm. SD (mm)0.088 ± 0.0120.073 ± 0.01020.1<0.00010.085 ± 0.0130.076 ± 0.01212.30.011
Biomechanical indices
Stiffness (kN/mm)84.0 ± 29.576.7 ± 20.49.60.31183.6 ± 28.577.3 ± 22.28.10.387
Modulus (Mpa)2,006 ± 5701,680 ± 47819.40.0341,973 ± 5631,723 ± 51114.50.107
Failure load (N)4,117 ± 1,4333,859 ± 9886.70.4614,123 ± 1,3573,863 ± 1,1036.70.458
% load Ct. distal end63.5 ± 12.653.6 ± 8.718.60.00261.2 ± 12.356.1 ± 10.89.00.135
% load Ct. proximal end89.6 ± 5.184.8 ± 8.55.70.01888.1 ± 6.386.3 ± 8.22.10.391
Tb. average stress (MPa)45.9 ± 7.245.5 ± 6.80.90.83145.9 ± 7.345.5 ± 6.80.80.853
Tb. SD stress (MPa)32.3 ± 1.830.8 ± 1.74.70.00531.8 ± 1.731.4 ± 2.01.30.457
Ct. average stress (MPa)80.6 ± 6.180.4 ± 3.70.30.87981.4 ± 4.779.6 ± 5.22.20.218
Ct. SD stress (MPa)20.0 (18.9, 21.9)22.7 (19.2, 25.6)10.30.08220.7 (18.9, 22.8)20.4 (18.8, 24.0)–0.10.954
Tb. equivalent strain0.0084 ± 0.00120.0084 ± 0.00110.30.9410.0084 ± 0.00120.0084 ± 0.00110.50.905
Ct. equivalent strain0.0138 (0.0136, 0.0140)0.0135 (0.0130, 0.0141)–2.00.1860.0137 (0.0133, 0.0141)0.0137 (0.0132, 0.0139)0.40.953

Patients with deformed wrist(s) had significantly higher total vBMD, cortical area fraction, cortical thickness, and modulus than patients without deformed wrist (Table 4). However, cortical and trabecular microstructure tended to be inferior in this group of patients, with significant differences seen in trabecular number, separation, network inhomogeneity, and cortical pore diameter. In patients with deformed wrist(s), a significantly less proportion of the load was carried by trabecular bone with trabecular stress being more unevenly distributed. Erosive disease at wrist was associated with significantly larger cortical area fraction, larger cortical pore diameter, and SD of cortical pore diameter. Higher cumulative dose of glucocorticoids was only significantly associated with lower cortical vBMD (Table 3). Otherwise, there were no significant association between cumulative dose/duration of glucocorticoids and indices of vBMD and microstructure. There were also no significant differences in vBMD and microstructural indices between patients who did or did not use glucocorticoids 6 months before the study (data not shown).

Higher levels of pro-inflammatory cytokines were associated with lower total vBMD, lower geometric indices such as lower cortical area fraction, lower areal MOI, thinner cortices, and reduced whole-bone stiffness and failure load (Table 3). Overall, the level of pro-inflammatory cytokines did not significantly affect trabecular and cortical microstructural indices.

Subgroup analyses of RA patients with focal periosteal bone apposition

The 10 RA patients identified with focal periosteal bone apposition did not differ significantly from patients without periosteal bone apposition with regard to demographics and disease characteristics except that they had significantly shorter disease duration (median 9.8 versus 20.6 years, p =0.024) and a trend toward a lower number of deformed joints (median 2 versus 4.5, p = 0.075). In this group of patients, as a result of the apposed periosteal new bone, a larger cortical area fraction and thicker cortex led to a significantly larger total and cortical areal MOI, as well as significantly higher whole-bone stiffness and failure load (Supplemental Table S1). However, because of the markedly higher cortical porosity, the load in the cortical compartment was not evenly distributed, indicated by a significantly higher SD of cortical stress and a significantly lower cortical strain.

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

In this study, we utilized HR-pQCT to investigate geometrical and microstructural changes and to estimate biomechanical competence at the distal radius in male RA patients. We found that compared with controls, male RA patients had substantially lower vBMD, compromised bone microstructure, and biomechanical competence at the distal radius. These differences were more pronounced than those revealed by aBMD at multiple skeletal sites and were independent of body weight, testosterone level, and aBMD at the ultradistal radius.

Chronic inflammation, which is the hallmark of RA, has a profound negative effect on local and systemic bone metabolism.[48, 49] As a result of the overexpression of receptor activator of NF-κB ligand on synovial fibroblasts and infiltrating lymphocytes and high levels of pro-inflammatory cytokines, bone remodeling is increased with amplified osteoclast differentiation and activation and inhibited osteoblast function, which leads to net loss of bone tissue and reduced bone strength.[1] At the distal radius, in RA patients, a lower vBMD was found but not a lower aBMD after adjustment for body weight. Microstructural changes were evident in RA patients, affecting both cortical and trabecular compartments with more porous cortices, thinned, widely and unevenly spaced trabeculae, reduced trabecular connectivity, and relatively greater resorption of platelike trabeculae as indicted by a higher SMI value. µFE analysis showed how the capacity of the bone to resist compressive load was significantly compromised in RA patients with lower whole-bone stiffness, modulus, and failure load, as well as lower and unevenly distributed stress. RA is a known risk factor for bone fragility, and its effect on fracture risk is independent of aBMD.[7, 50] This mismatch between fracture risk and aBMD cannot be fully explained by other traditional risk factors for bone fragility, such as immobility, increased risk of falling, and use of glucocorticoids.[9, 11] The inflammation-associated, aBMD-independent deterioration in both cortical and trabecular bone at the microstructural level in RA patients as shown in this study may contribute to the discrepancy found between increased fracture risk and relative preservation of aBMD in RA patients reported in previous studies.[5, 6]

Overall, microstructural differences between male RA patients and controls are similar to those found in female RA patients as reported in our previous study[24] and in a recent HR-pQCT–based study by Kocijan and colleagues,[51] and, also similar to female RA patients, they were not influenced by body weight or sex hormone level. However, compared with female patients, trabecular microstructure tends to be more compromised in male patients. For example, trabecular bone volume fraction and number were, on average, –10.8% and –5.7%, respectively, lower in female patients than in controls, whereas in male patients they were –23.2% and –8.1%, respectively, lower than in controls. At both the distal and proximal end of the scanned area, after adjustment of aBMD, a significantly greater proportion of load was carried by cortical than trabecular compartment, indicating a disproportionate loss of trabecular bone in male RA patients. Another discrepancy between male and female RA patients was related to the clinical features associated with focal periosteal bone apposition. For female patients, focal periosteal bone apposition occurred in those with longer disease duration and wrist deformity, whereas the opposite was true for male patients. The underlying cause for this discrepancy is unclear. Focal periosteal bone apposition at this site is most likely a response to ulnar head subluxation. Although the clinical expression of RA is known to be different between sexes,[25] no studies have specifically addressed whether gender difference in radioulnar subluxation exists in RA. Such a gender difference might exist with males developing ulnar head subluxation earlier than females, altering local biomechanics and stimulating bone accretion, particularly in that area of the distal radius close to the radioulnar joint, which would help explain the overall comparable prevalence of focal periosteal bone apposition between sexes.

The region of the distal radius investigated in this study is close to the radioulnar joint and radiocarpal joints, both of which are common sites for synovitis and joint deformity in RA. Indices of density, geometry, microstructure, and biomechanics correlated with indices of overall and local disease activity and severity, as well as with serum levels of several pro-inflammatory cytokines. The higher deficits in trabecular and cortical microstructure in patients with wrist deformity were likely the result of chronic inflammation, as well as the subsequent disuse and altered mechanical microenvironment. These findings indicate that chronic inflammation is an important contributor to bone loss in RA. However, microstructural indices were only weakly associated with serum levels of pro-inflammatory cytokines, suggesting that the microstructural changes observed are more likely the effect of local factors, such as local pro-inflammatory cytokines or bone hyperemia, rather than systemic factors.[52]

We did not observe significant correlation between use of glucocorticoids, quantified as cumulative dose, cumulative duration or recent exposure, and most indices of bone density and microstructure. The relationship between use of glucocorticoids and bone quality in RA is complicated and is influenced by other disease-related factors.[53] Although use of low-dose glucocorticoids may rapidly reduce markers of bone formation, increasing the risk of osteoporosis,[54] it also counteracts the effect of inflammation on the bone, lowering the rate of bone loss, particularly at skeletal sites close to synovitis.[53, 55] In a randomized placebo-controlled trial of patients with recent onset of RA, glucocorticoids decreased the degree of localized hand bone loss.[56] Findings from previous studies support a recovery of the adverse effect of glucocorticoids on the bone upon cessation of use.[57, 58] Our finding also concurs with an earlier 2-year prospective study using quantitative CT in RA patients, which concluded that low-dose glucocorticoids had only minimal effect on trabecular and cortical bone loss at the radius and tibia.[59]

Our study has several limitations. First, it had a cross-sectional design. A causal relationship between inflammation and compromised bone quality could not be established. The number of subjects with a history of fragility fracture was small, which precluded a detailed examination into the relationship between compromised bone quality and fracture risk. Second, because laboratory assessment did not include biomarkers of bone metabolism, we could not fully investigate how bone metabolism is affected by chronic inflammation in RA patients or its relationship with bone density and microstructure. We measured serum levels of several pro-inflammatory cytokines as surrogates of overall systemic inflammation. The degree of local inflammation may have a stronger relationship with compromised bone quality at distal radius. Measurement of local pro-inflammatory cytokine would have required a biopsy and, as such, was not undertaken in this study. Third, we only scanned the distal radius but not the distal tibia, which can also be assessed by HR-pQCT. The distal radius was selected because the wrist is a more common site of synovitis than the ankle in RA and because, unlike tibia, radius is non-weight-bearing. The heterogeneity of physical function among RA patients with diverse disease severity may be an important confounding factor for bone microstructure. Nonetheless, a comparison of the distal radius to the distal tibia would have been informative with regard to the relative effects of local and systemic inflammation. Fourth, the evaluation of wrist erosion was based on erosive changes at the wrist joint, rather than the erosive change at the radius. Erosive disease was not quantified using radiological scoring, and such scores may show a stronger relationship with indices of bone density and microstructure. Finally, HR-pQCT data is restricted to the peripheral skeleton and does not provide a direct measure of bone quality at axial sites such as hip and vertebrae, which are more common sites of fragility fracture. However, it has been reported that bone density, microarchitecture, and mechanical properties at distal radius could reflect mechanical competence of the axial skeleton.[60]

In conclusion, this HR-pQCT–based study investigated differences in cortical and trabecular bone density, geometry, microstructure, and biomechanical competence in male RA patients. We found that male RA patients had substantially lower vBMD, compromised cortical and trabecular bone microstructure, and biomechanical indices at the distal radius compared with controls. These differences were more pronounced than those revealed by aBMD and were independent of body weight, testosterone level, and aBMD at the ultradistal radius. Changes in density and microstructure in RA patients correlated with disease activity, disease severity, and serum levels of pro-inflammatory cytokines. This is a similar pattern of microstructural change to that recently described in female RA patients, although it appears that there is a disproportionate loss of trabecular bone in male RA patients. Also similar to female RA patients, about one-fifth of male RA patients had readily detectable focal periosteal apposition at the ulnovolar aspect of the distal radius. Disease features associated with this bone accretion, however, differed considerably from those described in female patients.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

This work was supported by a Direct Grant from The Chinese University of Hong Kong (ref. no. 2011.1.020). The authors are grateful to Ms Li Siu Wan for her generous help in the recruitment of study controls.

Authors' roles: All authors were responsible for study design; the collection, analysis, and interpretation of all data; the writing of the article; and the decision to publish.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmr2221-sm-0001-SupTab-S1.docx24KSupplementary Table S1.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.