Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health, vol. -, no. -, 1e5, 2015 Ó Copyright 2015 by The International Society for Clinical Densitometry 1094-6950/-:1e5/$36.00 http://dx.doi.org/10.1016/j.jocd.2015.02.005
Original Article
Bilateral Asymmetry of Radius and Tibia Bone Macroarchitecture and Microarchitecture: A High-Resolution Peripheral Quantitative Computed Tomography Study Erin M. Hildebrandt,1,2 Sarah L. Manske,1,2 David A. Hanley,1,2 and Steven K. Boyd*,1,2 1
Department of Radiology, Cumming School of Medicine, University of Calgary, Calgary, Canada; and 2McCaig Institute for Bone and Joint Health, University of Calgary, Calgary, Canada
Abstract Studies assessing bone health often select the dominant or nondominant limb to scan, but not both, for efficiency reasons. New scanning technology allows 3-dimensional (3D) visualization of the microarchitecture in bone, but it is not well understood whether there are differences between the dominant and nondominant limbs. Using 3D highresolution peripheral quantitative computed tomography (HR-pQCT), the aim of this study is to investigate the effect of limb dominance on bone macroarchitecture and microarchitecture. Healthy male and female participants (N 5 100; 59 female, 41 male), mean age 30.7 12.1 years, were scanned at both radii and tibiae using HR-pQCT. Hand and foot dominance were determined by the participant’s self-report. Most participants were right hand dominant (94.0%) and right foot dominant (91.0%). In the pooled cohort, the dominant radius had significantly greater cortical area (2.11%; p 5 0.002) and failure load (3.00%; p 5 0.001). At the tibia, the dominant foot had significantly lower bone mineral density ( 0.77%; p 5 0.042), cortical area ( 1.05%; p 5 0.031), and thickness ( 1.51%; p 5 0.017). For females, there were no differences at the radius, but at the tibia, the dominant side had greater crosssectional area (1.03%; p 5 0.044). Our data suggest that dominance has a small yet significant effect on macroarchitecture at both the ultradistal radius and tibia but not microarchitecture. This work emphasizes that it is important to be consistent in the selection of either dominant or nondominant limbs for HR-pQCT cohort studies; however, in the case where the opposite limb needs to be scanned, there would be small differences in macroarchitecture and no significant differences in microarchitecture anticipated. Key Words: Bilateral asymmetry; bone microarchitecture; bone mineral density; dominance; high-resolution peripheral quantitative computed tomography.
loading (1). Bones are dynamic structures that respond to mechanical stimulation at both the macroarchitectural and microarchitectural level (2), and this is most apparent in athletic populations because of increased loading patterns on dominant upper limbs. In racquet sport players, for example, the dominant forearm has significantly greater bone mineral density (BMD), bone mineral content (BMC), and crosssectional bone area (BA) when compared with the nondominant arm (3,4). Clinical studies assessing bone quality at peripheral sites, such as the distal radius, generally select either the dominant or the nondominant limb to scan. Others have shown in the general population using dual-energy X-ray absorptiometry
Introduction Asymmetry is well established in human long bones. Specifically, right side upper limb long bones have been shown to be longer and wider, regardless of handdominance, which is presumably a result of a predominance of right-handed individuals and increased mechanical Received 11/20/14; Revised 02/07/15; Accepted 02/11/15. *Address correspondence to: Steven K Boyd, PhD, McCaig Institute for Bone and Joint Health, University of Calgary, Room HRIC 3AC64, 3280 Hospital Drive NW, Calgary, Alberta, Canada T2N 4Z6. E-mail: [email protected]
1
2 (DXA) that BMC and BA are greater at the dominant distal forearm, whereas there are no differences in areal BMD (aBMD) between sides (5e7). Three-dimensional (3D) images from peripheral quantitative computed tomography (pQCT) allow assessment of volumetric BMD and differences in cortical and trabecular bone compartments. There are limited pQCT studies with mixed results assessing side-toside differences at the radius, with some studies reporting no difference and others finding significantly greater cortical area, and total and cortical BMC and BMD at the dominant radius (8e11). Lower limb bones are also significantly different between dominant and nondominant sides, yet the effects are opposite to the upper limb (6,9). Studies have found that the nondominant leg has significantly greater BMD and aBMD when compared with the dominant leg (6,9). This is postulated to be from the inverse relationship between hand skill and lower leg motor neuron excitability, in addition to biomechanically loading the nondominant leg while performing tasks such as kicking (6,12). The ability to measure in vivo human bone microarchitecture is relatively new using multislice 3D high-resolution pQCT (HR-pQCT). HR-pQCT is increasingly being used to assess bone microarchitecture in clinical research studies because it provides fine detail of both cortical and trabecular bone at the peripheral sites of the ultradistal radius and tibia. In the interest of efficiency, usually only one upper extremity and 1 lower extremity measurement are assessed per subject, typically on the nondominant limb. In cases where previous fractures have occurred, the opposite limb is selected for scanning to avoid artifacts. The importance of selecting a dominant or nondominant limb for HR-pQCT studies is not clear because little is known about differences in bone microarchitecture. Therefore, the purpose of this study is to compare bone macroarchitecture and microarchitecture between dominant and nondominant distal radius and tibia as measured using HR-pQCT.
Hildebrandt et al. High-Resolution Peripheral Quantitative Computed Tomography Participants were scanned at both ultradistal radii and tibiae by HR-pQCT (XtremeCT; Scanco Medical, Br€ uttisellen, Switzerland) using the standard human in vivo scanning protocol (60 kVp, 1000 mA, 100 ms integration time) (13). Limbs were supported in the scanner using an anatomic brace provided by the manufacturer to immobilize the joint. Using the 2D scout scan, reference lines were placed manually at the midinclination tuberosity at the radius and at the plateau of the tibial end plate at the tibia (13). The first slice of the scan was acquired 9.5-mm (radius) and 22.5-mm (tibia) proximal from the reference line, the standard patient scan locations. Each scan produced a 9.02-mm scan length with 110 slices and a nominal isotropic resolution of 82 mm. Because the scans were acquired over a number of years, more than one trained technician performed the scans; however, the same technician performed all 4 scans for each participant. Technicians also monitored the scans for motion artifact (e.g., blurring or discontinuities), and if there was a significant artifact, a second scan was performed. Scans were graded for motion artifact before analysis, with a score of 1 indicating no motion artifact and a score of 5 indicating severe motion artifact. For this analysis, any participant with a motion artifact 4 was removed (14). Images acquired were analyzed by trained technicians using the manufacturer’s standard method described in detail by others (15,16). From this standard morphologic analysis we obtained total BMD (Tt.BMD; mg HA/cm3), trabecular BMD (Tb.BMD; mg HA/cm3), and total area (Tt.Ar; mm2). Trabecular number (Tb.N; mm 1) was calculated based on the distance transformation method (17). Trabecular thickness (Tb.Th; mm) and area (Tb.Ar; mm2) and were derived as described elsewhere (13). Cortical parameters were assessed including cortical area (Ct.Ar; mm2), BMD (Ct.BMD; mg HA/cm3), thickness (Ct.Th; mm), and porosity (Ct.Po; %) (18).
Finite-Element Modeling
Materials and Methods Subjects A total of 100 healthy participants (59 female and 41 male) were recruited from Calgary, Alberta, and surrounding area over a period of 8 years. Participant mean age was 30.7 12.1 years with a minimum age of 16.6 years and maximum of 72.8 years, and reflects the age range of our ongoing study cohorts. Participant self-report determined hand and foot dominance by asking, ‘‘What hand do you write with?’’ and ‘‘What foot do you kick with?’’ Basic information about participant’s medical history was collected at scan time, including self-reported previous fracture locations and severity. Approval for all procedures was obtained from the Conjoint Health Research Ethics Board at the University of Calgary. All participants aged O18 years provided written informed consent before involvement in the study. For those participants aged !18 years, a parent provided written informed consent on behalf of their child.
Homogeneous finite-element meshes were generated from the 3D HR-pQCT image data, as has been described elsewhere (19). A uniaxial compression test using 1% axial strain was applied in the z direction on all radii and tibiae scans (19). A homogeneous tissue modulus of 6829 MPa and a Poisson ratio of 0.3 were also applied (19). Models were solved using a custom finite-element software (FAIM, version 6.0; Numerics88 Solutions, Calgary, Canada), and failure load (N) was calculated (20).
Statistical Analysis To investigate the effect of dominance on bone macroarchitecture and microarchitecture, a 2-way mixed analysis of variance with factors of sex and limb dominance was performed. A subanalysis separating the right- and leftdominant cohort participants was performed to determine if any effects of dominance were associated with a specific limb. All analyses were performed using SPSS version 20.0 (IBM Inc; Chicago, IL). Significance was defined as p !0.05.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
Bilateral Asymmetry by HR-pQCT
3
Results Descriptive Characteristics of Participants Of the 100 participants that were scanned, 95 were included in the analysis for the radius and 96 were included in the analysis for the tibia. We excluded 2 females’ radii because of excessive motion artifact. One female’s radius and 2 male tibiae were excluded because of scanning being performed at different time points. One male participant’s radius scans were excluded because of an inaccurately selected scan region. One male participant’s radius and tibia scans were excluded because of an imaging error, and 1 female participant’s tibia scan was excluded because of the history of a fracture at the distal tibia. It was determined that 94 participants (55 females, 39 males, mean age 30.9 12.4 years) were classified as right handed, and 6 were classified as left handed (4 females, 2 males, mean age 27.3 4.5 years). At the tibia, 91 participants (57 females, 34 males, mean age 31.2 12.5 years) were classified as right foot dominant and 9 were classified as left foot dominant (2 female, 7 males, mean age 25.6 4.0 years). The percent of left-handed and -footed participants is representative of the general population. Descriptive characteristics for females and males are provided in Table 1. There were no significant differences in age and hand dominance between sexes, and males had significantly less right foot dominant participants by percent ( p !0.05).
Limb Dominance In the pooled group, dominant radii had significantly greater Ct.Ar (2.11%; p 5 0.002) and failure load (3.00%; p 5 0.001; Table 2). At the tibia, the dominant foot had significantly lower Tt.BMD ( 0.77%; p 5 0.042), Ct.Ar ( 1.03%; p 5 0.049), and Ct.Th ( 1.51%; p 5 0.017; Table 3). A subanalysis investigating whether dominance was more strongly associated with the right-dominant participants vs the left-dominant participants found that at the radius there
Table 1 Descriptive Characteristics for All Subjects and Dichotomized by Sex (Mean SD) Characteristic
Whole cohort (n 5 100)
Hand dominance 94.0 (% right) Foot dominance 91.0 (% right) Age (yr) 30.7 12.1 !20 5 20 to !40 78 40 to !60 10 60 7 Abbr: SD, standard deviation.
Women (n 5 59)
Men (n 5 41)
93.2
95.1
96.6
82.9
30.9 11.7 30.3 12.7 3 2 44 34 9 1 3 4
Table 2 Distal Radius Bone Macroarchitecture and Microarchitecture Measurements for Pooled Cohort as Measured Using High-Resolution Peripheral Quantitative Computed Tomography Variable
Dominant (n 5 95)
Tt.BMD 325.0 (mg HA/cm3) 949.1 Ct.BMD (mg HA/cm3) 180.9 Tb.BMD (mg HA/cm3) 319.5 Tt.Ar (mm2) 65.7 Ct.Ar (mm2) Ct.Th (mm) 0.926 Ct.Po (%) 1.75 252.7 Tb.Ar (mm2) 2.01 Tb.N (mm 1) Tb.Th (mm) 0.0503 Failure load (N) 2533.2
Nondominant (n 5 95)
p Value
57.9
323.2 62.2
0.362
50.7
948.5 49.6
0.835
45.8
178.9 44.5
0.053
72.2 315.7 14.5 64.3 0.168 0.912 1.03 1.67 65.2 250.2 0.239 2.00 0.035 0.0507 762.3 2457.2
76.3 14.1 0.171 0.91 69.6 0.270 0.035 738.9
0.142 0.002 0.104 0.066 0.390 0.615 0.482 0.001
Note: The p values are presented for the main effects of dominance. Because sex-by-dominance interactions was not statistically significant, p values are not presented separately for females and males (mean SD). Bold p values represent statistical significance to 0.05. Abbr: Ct.Ar, cortical area; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; SD, standard deviation; Tb.Ar, trabecular area; Tb.BMD, trabecular bone mineral density; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tt.Ar, total area; Tt.BMD, total bone mineral density.
was greater Ct.Ar and failure load ( p !0.05) for righthanded participants. There was no effect for left-handed participants at the radius. At the tibia, right foot dominant participants had significantly lower Tt.BMD, Ct.Ar, Ct.Th, and Ct.Po ( p !0.05) and higher Tb.Ar ( p !0.05) at the dominant limb. Left foot dominant participants had significantly greater Tt.BMD and Tb.BMD ( p !0.05) and lower Tt.Ar and Tb.Ar ( p !0.05) at the dominant tibia.
Comparison Within Sexes There were no sex-by-dominance interactions for macroarchitecture or microarchitecture at the radius (Table 2). For females, at the tibia, the dominant side had significantly greater Tt.Ar (1.03%; p 5 0.044; Table 3).
Discussion Our findings suggest that there are few significant differences in macroarchitecture between the dominant and nondominant ultradistal radius and tibia, and those that exist are relatively small in magnitude. Because significant differences were found, it suggests for consistency that there is continued use of the nondominant limb for scanning with
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
4
Hildebrandt et al. Table 3 Distal Tibia Bone Microarchitecture and Macroarchitecture Measurements for Whole Cohort and Separated by Sex as Measured Using High-Resolution Peripheral Quantitative Computed Tomography
Variable Tt.BMD (mg HA/cm3) Ct.BMD (mg HA/cm3) Tb.BMD (mg HA/cm3) Tt.Ar (mm2) Ct.Ar (mm2) Ct.Th (mm) Ct.Po (%) Tb.Ar (mm2) Tb.N (mm 1) Tb.Th (mm) Failure load (N)
Dominant (n 5 96), mean SD 325.3 935.8 196.3 759.0 134.6 1.33 4.09 619.3 1.91 0.0862 6752.0
54.6 46.7 38.5 153.1 28.2 0.26 1.73 143.3 0.296 0.013 1553.4
Nondominant (n 5 96), mean SD
p Value
Females (n 5 58)
Males (n 5 38)
0.042 0.998 0.883 0.755 0.031 0.017 0.066 0.240 0.200 0.156 0.223
d d d 0.044 d d d d d d d
d d d 0.215 d d d d d d d
327.8 936.1 196.1 756.9 136.0 1.35 4.26 614.9 1.93 0.0853 6702.5
55.5 49.7 37.8 158.7 29.0 0.28 1.78 149.0 0.297 0.013 1593.4
Note: The p values are presented for the main effects of dominance. Where sex-by-dominance interaction was significant, p values for post hoc simple effects testing are presented for females and males. Bold p values represent statistical significance to 0.05. Abbr: Ct.Ar, cortical area; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; SD, standard deviation; Tb.Ar, trabecular area; Tb.BMD, trabecular bone mineral density; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tt.Ar, total area; Tt.BMD, total bone mineral density.
HR-pQCT. However, although these data illustrate there are significant differences, it is not particularly critical as most of the percent differences are around 1% to 2%, excluding Ct.Po, which has a relative difference of 8%. In the case where the nondominant limb cannot be scanned, such as a fracture, the dominant limb can justifiably be considered as a suitable alternative. When scanning the dominant limb, there is no need for a conversion factor between dominant and nondominant limbs to be calculated, as has been noted in other studies using DXA (21). The consistency of selecting the nondominant limb within a cohort is more important in cross-sectional studies than it is for longitudinal studies, where it is more important to scan the same limb over time. The results at the radius demonstrate that there is significantly greater Ct.Ar and failure load at the dominant side compared with the nondominant side. This is consistent with previous DXA studies that have established there is significantly greater BA and BMC at the dominant radius but no difference in aBMD (6,7). In addition, limited results from pQCT studies demonstrate greater BA, BMC, and BMD for both total and cortical bone at the radius in the general population but no differences in trabecular bone (10,11). This result is consistent with a previous study assessing sideto-side differences in strength suggesting that there is greater Ct.Ar and strength at the dominant humerus and ulna in a pooled group of elite athletes and controls (9). We found that dominance at the tibia had the opposite effect on bone compared with the radius. This opposite pattern from the upper limb to lower limb has been termed ‘‘crossed symmetry,’’ and our results align with the limited research assessing dominance at the lower extremity (9,22). Min et al., used DXA to assess the effects of dominance on aBMD at
the calcaneus, which is subjected to a similar loading environment as at the ultradistal tibia. They found that in the general population the dominant calcaneus had significantly lower aBMD than the nondominant side (6). Similarly, tibial Ct.Ar assessed by pQCT within an athletic population, namely runners, was lower at the dominant side (9). The cause of the lower BMD and area at the dominant lower limb has been postulated to be because of a combination of increased motor neuron excitability at the left limb lowerleg muscles in right-handed subjects, and conversely for left-handed subjects, and increased loading in the nondominant foot during kicking movements (6,12,23). Tan et al. (12) found that there was greater excitability of motor neurons in the left soleus muscle of right-handed participants and the converse for left-handed participants. According to Frost’s mechanostat theory, increased excitation puts greater physical stress and strain on the bone, therefore increasing bone formation (23,24). In addition, Tan et al. (12) highlighted the postural role of the soleus, which stabilizes the lower leg to facilitate movement of the dominant foot such as kicking a ball. The combination of greater motor neuron excitability and loading characteristics of the nondominant foot may lead to greater bone macroarchitecture on the nondominant side. Dominance interacted with sex for Tt.Ar at the tibia only. Of interest, Tt.Ar was significantly greater at the dominant tibia for females, which is opposite to the crossed symmetry pattern found in the pooled group. Both age and sex are important factors when considering bone health, and population-based studies using HR-pQCT have documented differing sex-related trends in age-related bone loss, particularly in cortical bone loss in women (13). This single differing
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
Bilateral Asymmetry by HR-pQCT pattern found in males and females in relation to dominance effects may be attributed to the hormonal effect on bone that cause differences in bone growth patterns for males and females, periosteal vs endocortical growth, respectively (25,26). Additionally, it could be due to a less clear assignment of leg dominance and small sample size. There are limitations to our study, including limited representation of people with left-hand and -foot dominance and a small sample size for comparisons between sexes. Furthermore, our cohort covered a wide age range and there may have been age effects within our population. Previous studies show an age effect on dominance at the radius; below the age of 40 years, aBMD was greater on the dominant side, whereas this pattern is reversed in older age groups (O60 years) (6). We did not perform a similar analysis based on age as most participants were aged !40 years. Secondly, detailed physical activity of participants was not recorded at the time of visit, which is a factor when assessing BMD. However, we are aware that there were no elite level athletes within our cohort (e.g., elite tennis players). Finally, patient hand and foot dominance was self-reported as either right or left side leaving no room for the range of dominance that may include ambidextrous people. In conclusion, there are small differences in macroarchitecture at the distal radius and tibia as measured using HR-pQCT. The differences detected were few, yet continued scanning of the nondominant limb is a prudent approach, especially in cross-sectional studies.
5
8.
9. 10.
11. 12. 13.
14.
15. 16.
Acknowledgments
17.
The authors thank the study participants who donated their time to further our research, Britta Jorgenson for her expertise in finite-element modeling, and Michelle Kan, Eva Szabo, and Shannon Boucousis for scan acquisition and analysis.
18. 19.
References 1. Roy TA, Ruff CB, Plato CC. 1994 Hand dominance and bilateral asymmetry in the structure of the second metacarpal. Am J Phys Anthropol 94:203e211. 2. Kazakia GJ, Tjong W, Nirody JA, et al. 2014 The influence of disuse on bone microstructure and mechanics assessed by HRpQCT. Bone 63:132e140. 3. Ducher G, Prouteau S, Courteix D, et al. 2004 Cortical and trabecular bone at the forearm show different adaptation patterns in response to tennis playing. J Clin Densitom 7:399e405. 4. Ducher G, Courteix D, Meme S, et al. 2005 Bone geometry in response to long-term tennis playing and its relationship with muscle volume: a quantitative magnetic resonance imaging study in tennis players. Bone 37:457e466. 5. Walters J, Koo WW, Bush A, et al. 1998 Effect of hand dominance on bone mass measurement in sedentary individuals. J Clin Densitom 1:359e367. 6. Min JY, Min KB, Paek D, et al. 2007 Side differences in the bone density of the distal radius and calcaneus in Koreans aged 4-86 years. J Clin Densitom 10:184e188. 7. Akar S, Sivrikaya H, Canikli A, et al. 2002 Lateralized mineral content and density in distal forearm bones in right-handed men
20.
21. 22. 23.
24. 25. 26.
and women: relation of structure to function. Int J Neurosci 112: 301e311. Sergi G, Perissinotto E, Zucchetto M, et al. 2009 Upper limb bone mineral density and body composition measured by peripheral quantitative computed tomography in right-handed adults: the role of the dominance effect. J Endocrinol Invest 32:298e302. Shaw CN. 2011 Is ’hand preference’ coded in the hominin skeleton? An in-vivo study of bilateral morphological variation. J Hum Evol 61:480e487. MacIntyre NJ, Adachi JD, Webber CE. 1999 In vivo detection of structural differences between dominant and nondominant radii using peripheral quantitative computed tomography. J Clin Densitom 2:413e422. Rico H, Gonzalezriola J, Revilla M, et al. 1994 Cortical versus trabecular bone massdinfluence of activity on both bone components. Calcif Tissue Int 54:470e472. Tan U. 1985 Relationships between hand skill and the excitability of motoneurons innervating the postural soleus muscle in human subjects. Int J Neurosci 26:289e300. Macdonald HM, Nishiyama KK, Kang J, et al. 2011 Age-related patterns of trabecular and cortical bone loss differ between sexes and skeletal sites: a population-based HR-pQCT study. J Bone Mineral Res 26:50e62. Pauchard Y, Liphardt AM, Macdonald HM, et al. 2012 Quality control for bone quality parameters affected by subject motion in high-resolution peripheral quantitative computed tomography. Bone 50:1304e1310. MacNeil JA, Boyd SK. 2007 Accuracy of high-resolution peripheral quantitative computed tomography for measurement of bone quality. Med Eng Phys 29:1096e1105. Boutroy S, Bouxsein ML, Munoz F, et al. 2005 In vivo assessment of trabecular bone microarchitecture by high-resolution peripheral quantitative computed tomography. J Clin Endocrinol Metab 90:6508e6515. Laib A, Ruegsegger P. 1999 Comparison of structure extraction methods for in vivo trabecular bone measurements. Comput Med Imagaing Graph 23:69e74. Burghardt AJ, Buie HR, Laib A, et al. 2010 Reproducibility of direct quantitative measures of cortical bone microarchitecture of the distal radius and tibia by HR-pQCT. Bone 47:519e528. Macneil JA, Boyd SK. 2008 Bone strength at the distal radius can be estimated from high-resolution peripheral quantitative computed tomography and the finite element method. Bone 42:1203e1213. Pistoia W, van Rietbergen B, Lochmuller EM, et al. 2002 Estimation of distal radius failure load with micro-finite element analysis models based on three-dimensional peripheral quantitative computed tomography images. Bone 30:842e848. Lekamwasam S, Rodrigo M, de Silva KI, et al. 2005 Comparison of phalangeal bone mineral content and density between the dominant and non-dominant sides. Ceylon Med J 50:149e151. Plochocki JH. 2004 Bilateral variation in limb articular surface dimensions. Am J Hum Biol 16:328e333. Meszaros S, Ferencz V, Csupor E, et al. 2006 Comparison of the femoral neck bone density, quantitative ultrasound and bone density of the heel between dominant and non-dominant side. Eur J Radiol 60:293e298. Frost HM. 1997 On our age-related bone loss: insights from a new paradigm. J Bone Miner Res 12:1539e1546. Kim BT, Mosekilde L, Duan Y, et al. 2003 The structural and hormonal basis of sex differences in peak appendicular bone strength in rats. J Bone Mineral Res 18:150e155. Seeman E. 2002 Pathogenesis of bone fragility in women and men. Lancet 359:1841e1850.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
Original Article
Bilateral Asymmetry of Radius and Tibia Bone Macroarchitecture and Microarchitecture: A High-Resolution Peripheral Quantitative Computed Tomography Study Erin M. Hildebrandt,1,2 Sarah L. Manske,1,2 David A. Hanley,1,2 and Steven K. Boyd*,1,2 1
Department of Radiology, Cumming School of Medicine, University of Calgary, Calgary, Canada; and 2McCaig Institute for Bone and Joint Health, University of Calgary, Calgary, Canada
Abstract Studies assessing bone health often select the dominant or nondominant limb to scan, but not both, for efficiency reasons. New scanning technology allows 3-dimensional (3D) visualization of the microarchitecture in bone, but it is not well understood whether there are differences between the dominant and nondominant limbs. Using 3D highresolution peripheral quantitative computed tomography (HR-pQCT), the aim of this study is to investigate the effect of limb dominance on bone macroarchitecture and microarchitecture. Healthy male and female participants (N 5 100; 59 female, 41 male), mean age 30.7 12.1 years, were scanned at both radii and tibiae using HR-pQCT. Hand and foot dominance were determined by the participant’s self-report. Most participants were right hand dominant (94.0%) and right foot dominant (91.0%). In the pooled cohort, the dominant radius had significantly greater cortical area (2.11%; p 5 0.002) and failure load (3.00%; p 5 0.001). At the tibia, the dominant foot had significantly lower bone mineral density ( 0.77%; p 5 0.042), cortical area ( 1.05%; p 5 0.031), and thickness ( 1.51%; p 5 0.017). For females, there were no differences at the radius, but at the tibia, the dominant side had greater crosssectional area (1.03%; p 5 0.044). Our data suggest that dominance has a small yet significant effect on macroarchitecture at both the ultradistal radius and tibia but not microarchitecture. This work emphasizes that it is important to be consistent in the selection of either dominant or nondominant limbs for HR-pQCT cohort studies; however, in the case where the opposite limb needs to be scanned, there would be small differences in macroarchitecture and no significant differences in microarchitecture anticipated. Key Words: Bilateral asymmetry; bone microarchitecture; bone mineral density; dominance; high-resolution peripheral quantitative computed tomography.
loading (1). Bones are dynamic structures that respond to mechanical stimulation at both the macroarchitectural and microarchitectural level (2), and this is most apparent in athletic populations because of increased loading patterns on dominant upper limbs. In racquet sport players, for example, the dominant forearm has significantly greater bone mineral density (BMD), bone mineral content (BMC), and crosssectional bone area (BA) when compared with the nondominant arm (3,4). Clinical studies assessing bone quality at peripheral sites, such as the distal radius, generally select either the dominant or the nondominant limb to scan. Others have shown in the general population using dual-energy X-ray absorptiometry
Introduction Asymmetry is well established in human long bones. Specifically, right side upper limb long bones have been shown to be longer and wider, regardless of handdominance, which is presumably a result of a predominance of right-handed individuals and increased mechanical Received 11/20/14; Revised 02/07/15; Accepted 02/11/15. *Address correspondence to: Steven K Boyd, PhD, McCaig Institute for Bone and Joint Health, University of Calgary, Room HRIC 3AC64, 3280 Hospital Drive NW, Calgary, Alberta, Canada T2N 4Z6. E-mail: [email protected]
1
2 (DXA) that BMC and BA are greater at the dominant distal forearm, whereas there are no differences in areal BMD (aBMD) between sides (5e7). Three-dimensional (3D) images from peripheral quantitative computed tomography (pQCT) allow assessment of volumetric BMD and differences in cortical and trabecular bone compartments. There are limited pQCT studies with mixed results assessing side-toside differences at the radius, with some studies reporting no difference and others finding significantly greater cortical area, and total and cortical BMC and BMD at the dominant radius (8e11). Lower limb bones are also significantly different between dominant and nondominant sides, yet the effects are opposite to the upper limb (6,9). Studies have found that the nondominant leg has significantly greater BMD and aBMD when compared with the dominant leg (6,9). This is postulated to be from the inverse relationship between hand skill and lower leg motor neuron excitability, in addition to biomechanically loading the nondominant leg while performing tasks such as kicking (6,12). The ability to measure in vivo human bone microarchitecture is relatively new using multislice 3D high-resolution pQCT (HR-pQCT). HR-pQCT is increasingly being used to assess bone microarchitecture in clinical research studies because it provides fine detail of both cortical and trabecular bone at the peripheral sites of the ultradistal radius and tibia. In the interest of efficiency, usually only one upper extremity and 1 lower extremity measurement are assessed per subject, typically on the nondominant limb. In cases where previous fractures have occurred, the opposite limb is selected for scanning to avoid artifacts. The importance of selecting a dominant or nondominant limb for HR-pQCT studies is not clear because little is known about differences in bone microarchitecture. Therefore, the purpose of this study is to compare bone macroarchitecture and microarchitecture between dominant and nondominant distal radius and tibia as measured using HR-pQCT.
Hildebrandt et al. High-Resolution Peripheral Quantitative Computed Tomography Participants were scanned at both ultradistal radii and tibiae by HR-pQCT (XtremeCT; Scanco Medical, Br€ uttisellen, Switzerland) using the standard human in vivo scanning protocol (60 kVp, 1000 mA, 100 ms integration time) (13). Limbs were supported in the scanner using an anatomic brace provided by the manufacturer to immobilize the joint. Using the 2D scout scan, reference lines were placed manually at the midinclination tuberosity at the radius and at the plateau of the tibial end plate at the tibia (13). The first slice of the scan was acquired 9.5-mm (radius) and 22.5-mm (tibia) proximal from the reference line, the standard patient scan locations. Each scan produced a 9.02-mm scan length with 110 slices and a nominal isotropic resolution of 82 mm. Because the scans were acquired over a number of years, more than one trained technician performed the scans; however, the same technician performed all 4 scans for each participant. Technicians also monitored the scans for motion artifact (e.g., blurring or discontinuities), and if there was a significant artifact, a second scan was performed. Scans were graded for motion artifact before analysis, with a score of 1 indicating no motion artifact and a score of 5 indicating severe motion artifact. For this analysis, any participant with a motion artifact 4 was removed (14). Images acquired were analyzed by trained technicians using the manufacturer’s standard method described in detail by others (15,16). From this standard morphologic analysis we obtained total BMD (Tt.BMD; mg HA/cm3), trabecular BMD (Tb.BMD; mg HA/cm3), and total area (Tt.Ar; mm2). Trabecular number (Tb.N; mm 1) was calculated based on the distance transformation method (17). Trabecular thickness (Tb.Th; mm) and area (Tb.Ar; mm2) and were derived as described elsewhere (13). Cortical parameters were assessed including cortical area (Ct.Ar; mm2), BMD (Ct.BMD; mg HA/cm3), thickness (Ct.Th; mm), and porosity (Ct.Po; %) (18).
Finite-Element Modeling
Materials and Methods Subjects A total of 100 healthy participants (59 female and 41 male) were recruited from Calgary, Alberta, and surrounding area over a period of 8 years. Participant mean age was 30.7 12.1 years with a minimum age of 16.6 years and maximum of 72.8 years, and reflects the age range of our ongoing study cohorts. Participant self-report determined hand and foot dominance by asking, ‘‘What hand do you write with?’’ and ‘‘What foot do you kick with?’’ Basic information about participant’s medical history was collected at scan time, including self-reported previous fracture locations and severity. Approval for all procedures was obtained from the Conjoint Health Research Ethics Board at the University of Calgary. All participants aged O18 years provided written informed consent before involvement in the study. For those participants aged !18 years, a parent provided written informed consent on behalf of their child.
Homogeneous finite-element meshes were generated from the 3D HR-pQCT image data, as has been described elsewhere (19). A uniaxial compression test using 1% axial strain was applied in the z direction on all radii and tibiae scans (19). A homogeneous tissue modulus of 6829 MPa and a Poisson ratio of 0.3 were also applied (19). Models were solved using a custom finite-element software (FAIM, version 6.0; Numerics88 Solutions, Calgary, Canada), and failure load (N) was calculated (20).
Statistical Analysis To investigate the effect of dominance on bone macroarchitecture and microarchitecture, a 2-way mixed analysis of variance with factors of sex and limb dominance was performed. A subanalysis separating the right- and leftdominant cohort participants was performed to determine if any effects of dominance were associated with a specific limb. All analyses were performed using SPSS version 20.0 (IBM Inc; Chicago, IL). Significance was defined as p !0.05.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
Bilateral Asymmetry by HR-pQCT
3
Results Descriptive Characteristics of Participants Of the 100 participants that were scanned, 95 were included in the analysis for the radius and 96 were included in the analysis for the tibia. We excluded 2 females’ radii because of excessive motion artifact. One female’s radius and 2 male tibiae were excluded because of scanning being performed at different time points. One male participant’s radius scans were excluded because of an inaccurately selected scan region. One male participant’s radius and tibia scans were excluded because of an imaging error, and 1 female participant’s tibia scan was excluded because of the history of a fracture at the distal tibia. It was determined that 94 participants (55 females, 39 males, mean age 30.9 12.4 years) were classified as right handed, and 6 were classified as left handed (4 females, 2 males, mean age 27.3 4.5 years). At the tibia, 91 participants (57 females, 34 males, mean age 31.2 12.5 years) were classified as right foot dominant and 9 were classified as left foot dominant (2 female, 7 males, mean age 25.6 4.0 years). The percent of left-handed and -footed participants is representative of the general population. Descriptive characteristics for females and males are provided in Table 1. There were no significant differences in age and hand dominance between sexes, and males had significantly less right foot dominant participants by percent ( p !0.05).
Limb Dominance In the pooled group, dominant radii had significantly greater Ct.Ar (2.11%; p 5 0.002) and failure load (3.00%; p 5 0.001; Table 2). At the tibia, the dominant foot had significantly lower Tt.BMD ( 0.77%; p 5 0.042), Ct.Ar ( 1.03%; p 5 0.049), and Ct.Th ( 1.51%; p 5 0.017; Table 3). A subanalysis investigating whether dominance was more strongly associated with the right-dominant participants vs the left-dominant participants found that at the radius there
Table 1 Descriptive Characteristics for All Subjects and Dichotomized by Sex (Mean SD) Characteristic
Whole cohort (n 5 100)
Hand dominance 94.0 (% right) Foot dominance 91.0 (% right) Age (yr) 30.7 12.1 !20 5 20 to !40 78 40 to !60 10 60 7 Abbr: SD, standard deviation.
Women (n 5 59)
Men (n 5 41)
93.2
95.1
96.6
82.9
30.9 11.7 30.3 12.7 3 2 44 34 9 1 3 4
Table 2 Distal Radius Bone Macroarchitecture and Microarchitecture Measurements for Pooled Cohort as Measured Using High-Resolution Peripheral Quantitative Computed Tomography Variable
Dominant (n 5 95)
Tt.BMD 325.0 (mg HA/cm3) 949.1 Ct.BMD (mg HA/cm3) 180.9 Tb.BMD (mg HA/cm3) 319.5 Tt.Ar (mm2) 65.7 Ct.Ar (mm2) Ct.Th (mm) 0.926 Ct.Po (%) 1.75 252.7 Tb.Ar (mm2) 2.01 Tb.N (mm 1) Tb.Th (mm) 0.0503 Failure load (N) 2533.2
Nondominant (n 5 95)
p Value
57.9
323.2 62.2
0.362
50.7
948.5 49.6
0.835
45.8
178.9 44.5
0.053
72.2 315.7 14.5 64.3 0.168 0.912 1.03 1.67 65.2 250.2 0.239 2.00 0.035 0.0507 762.3 2457.2
76.3 14.1 0.171 0.91 69.6 0.270 0.035 738.9
0.142 0.002 0.104 0.066 0.390 0.615 0.482 0.001
Note: The p values are presented for the main effects of dominance. Because sex-by-dominance interactions was not statistically significant, p values are not presented separately for females and males (mean SD). Bold p values represent statistical significance to 0.05. Abbr: Ct.Ar, cortical area; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; SD, standard deviation; Tb.Ar, trabecular area; Tb.BMD, trabecular bone mineral density; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tt.Ar, total area; Tt.BMD, total bone mineral density.
was greater Ct.Ar and failure load ( p !0.05) for righthanded participants. There was no effect for left-handed participants at the radius. At the tibia, right foot dominant participants had significantly lower Tt.BMD, Ct.Ar, Ct.Th, and Ct.Po ( p !0.05) and higher Tb.Ar ( p !0.05) at the dominant limb. Left foot dominant participants had significantly greater Tt.BMD and Tb.BMD ( p !0.05) and lower Tt.Ar and Tb.Ar ( p !0.05) at the dominant tibia.
Comparison Within Sexes There were no sex-by-dominance interactions for macroarchitecture or microarchitecture at the radius (Table 2). For females, at the tibia, the dominant side had significantly greater Tt.Ar (1.03%; p 5 0.044; Table 3).
Discussion Our findings suggest that there are few significant differences in macroarchitecture between the dominant and nondominant ultradistal radius and tibia, and those that exist are relatively small in magnitude. Because significant differences were found, it suggests for consistency that there is continued use of the nondominant limb for scanning with
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
4
Hildebrandt et al. Table 3 Distal Tibia Bone Microarchitecture and Macroarchitecture Measurements for Whole Cohort and Separated by Sex as Measured Using High-Resolution Peripheral Quantitative Computed Tomography
Variable Tt.BMD (mg HA/cm3) Ct.BMD (mg HA/cm3) Tb.BMD (mg HA/cm3) Tt.Ar (mm2) Ct.Ar (mm2) Ct.Th (mm) Ct.Po (%) Tb.Ar (mm2) Tb.N (mm 1) Tb.Th (mm) Failure load (N)
Dominant (n 5 96), mean SD 325.3 935.8 196.3 759.0 134.6 1.33 4.09 619.3 1.91 0.0862 6752.0
54.6 46.7 38.5 153.1 28.2 0.26 1.73 143.3 0.296 0.013 1553.4
Nondominant (n 5 96), mean SD
p Value
Females (n 5 58)
Males (n 5 38)
0.042 0.998 0.883 0.755 0.031 0.017 0.066 0.240 0.200 0.156 0.223
d d d 0.044 d d d d d d d
d d d 0.215 d d d d d d d
327.8 936.1 196.1 756.9 136.0 1.35 4.26 614.9 1.93 0.0853 6702.5
55.5 49.7 37.8 158.7 29.0 0.28 1.78 149.0 0.297 0.013 1593.4
Note: The p values are presented for the main effects of dominance. Where sex-by-dominance interaction was significant, p values for post hoc simple effects testing are presented for females and males. Bold p values represent statistical significance to 0.05. Abbr: Ct.Ar, cortical area; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; SD, standard deviation; Tb.Ar, trabecular area; Tb.BMD, trabecular bone mineral density; Tb.N, trabecular number; Tb.Th, trabecular thickness; Tt.Ar, total area; Tt.BMD, total bone mineral density.
HR-pQCT. However, although these data illustrate there are significant differences, it is not particularly critical as most of the percent differences are around 1% to 2%, excluding Ct.Po, which has a relative difference of 8%. In the case where the nondominant limb cannot be scanned, such as a fracture, the dominant limb can justifiably be considered as a suitable alternative. When scanning the dominant limb, there is no need for a conversion factor between dominant and nondominant limbs to be calculated, as has been noted in other studies using DXA (21). The consistency of selecting the nondominant limb within a cohort is more important in cross-sectional studies than it is for longitudinal studies, where it is more important to scan the same limb over time. The results at the radius demonstrate that there is significantly greater Ct.Ar and failure load at the dominant side compared with the nondominant side. This is consistent with previous DXA studies that have established there is significantly greater BA and BMC at the dominant radius but no difference in aBMD (6,7). In addition, limited results from pQCT studies demonstrate greater BA, BMC, and BMD for both total and cortical bone at the radius in the general population but no differences in trabecular bone (10,11). This result is consistent with a previous study assessing sideto-side differences in strength suggesting that there is greater Ct.Ar and strength at the dominant humerus and ulna in a pooled group of elite athletes and controls (9). We found that dominance at the tibia had the opposite effect on bone compared with the radius. This opposite pattern from the upper limb to lower limb has been termed ‘‘crossed symmetry,’’ and our results align with the limited research assessing dominance at the lower extremity (9,22). Min et al., used DXA to assess the effects of dominance on aBMD at
the calcaneus, which is subjected to a similar loading environment as at the ultradistal tibia. They found that in the general population the dominant calcaneus had significantly lower aBMD than the nondominant side (6). Similarly, tibial Ct.Ar assessed by pQCT within an athletic population, namely runners, was lower at the dominant side (9). The cause of the lower BMD and area at the dominant lower limb has been postulated to be because of a combination of increased motor neuron excitability at the left limb lowerleg muscles in right-handed subjects, and conversely for left-handed subjects, and increased loading in the nondominant foot during kicking movements (6,12,23). Tan et al. (12) found that there was greater excitability of motor neurons in the left soleus muscle of right-handed participants and the converse for left-handed participants. According to Frost’s mechanostat theory, increased excitation puts greater physical stress and strain on the bone, therefore increasing bone formation (23,24). In addition, Tan et al. (12) highlighted the postural role of the soleus, which stabilizes the lower leg to facilitate movement of the dominant foot such as kicking a ball. The combination of greater motor neuron excitability and loading characteristics of the nondominant foot may lead to greater bone macroarchitecture on the nondominant side. Dominance interacted with sex for Tt.Ar at the tibia only. Of interest, Tt.Ar was significantly greater at the dominant tibia for females, which is opposite to the crossed symmetry pattern found in the pooled group. Both age and sex are important factors when considering bone health, and population-based studies using HR-pQCT have documented differing sex-related trends in age-related bone loss, particularly in cortical bone loss in women (13). This single differing
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
Bilateral Asymmetry by HR-pQCT pattern found in males and females in relation to dominance effects may be attributed to the hormonal effect on bone that cause differences in bone growth patterns for males and females, periosteal vs endocortical growth, respectively (25,26). Additionally, it could be due to a less clear assignment of leg dominance and small sample size. There are limitations to our study, including limited representation of people with left-hand and -foot dominance and a small sample size for comparisons between sexes. Furthermore, our cohort covered a wide age range and there may have been age effects within our population. Previous studies show an age effect on dominance at the radius; below the age of 40 years, aBMD was greater on the dominant side, whereas this pattern is reversed in older age groups (O60 years) (6). We did not perform a similar analysis based on age as most participants were aged !40 years. Secondly, detailed physical activity of participants was not recorded at the time of visit, which is a factor when assessing BMD. However, we are aware that there were no elite level athletes within our cohort (e.g., elite tennis players). Finally, patient hand and foot dominance was self-reported as either right or left side leaving no room for the range of dominance that may include ambidextrous people. In conclusion, there are small differences in macroarchitecture at the distal radius and tibia as measured using HR-pQCT. The differences detected were few, yet continued scanning of the nondominant limb is a prudent approach, especially in cross-sectional studies.
5
8.
9. 10.
11. 12. 13.
14.
15. 16.
Acknowledgments
17.
The authors thank the study participants who donated their time to further our research, Britta Jorgenson for her expertise in finite-element modeling, and Michelle Kan, Eva Szabo, and Shannon Boucousis for scan acquisition and analysis.
18. 19.
References 1. Roy TA, Ruff CB, Plato CC. 1994 Hand dominance and bilateral asymmetry in the structure of the second metacarpal. Am J Phys Anthropol 94:203e211. 2. Kazakia GJ, Tjong W, Nirody JA, et al. 2014 The influence of disuse on bone microstructure and mechanics assessed by HRpQCT. Bone 63:132e140. 3. Ducher G, Prouteau S, Courteix D, et al. 2004 Cortical and trabecular bone at the forearm show different adaptation patterns in response to tennis playing. J Clin Densitom 7:399e405. 4. Ducher G, Courteix D, Meme S, et al. 2005 Bone geometry in response to long-term tennis playing and its relationship with muscle volume: a quantitative magnetic resonance imaging study in tennis players. Bone 37:457e466. 5. Walters J, Koo WW, Bush A, et al. 1998 Effect of hand dominance on bone mass measurement in sedentary individuals. J Clin Densitom 1:359e367. 6. Min JY, Min KB, Paek D, et al. 2007 Side differences in the bone density of the distal radius and calcaneus in Koreans aged 4-86 years. J Clin Densitom 10:184e188. 7. Akar S, Sivrikaya H, Canikli A, et al. 2002 Lateralized mineral content and density in distal forearm bones in right-handed men
20.
21. 22. 23.
24. 25. 26.
and women: relation of structure to function. Int J Neurosci 112: 301e311. Sergi G, Perissinotto E, Zucchetto M, et al. 2009 Upper limb bone mineral density and body composition measured by peripheral quantitative computed tomography in right-handed adults: the role of the dominance effect. J Endocrinol Invest 32:298e302. Shaw CN. 2011 Is ’hand preference’ coded in the hominin skeleton? An in-vivo study of bilateral morphological variation. J Hum Evol 61:480e487. MacIntyre NJ, Adachi JD, Webber CE. 1999 In vivo detection of structural differences between dominant and nondominant radii using peripheral quantitative computed tomography. J Clin Densitom 2:413e422. Rico H, Gonzalezriola J, Revilla M, et al. 1994 Cortical versus trabecular bone massdinfluence of activity on both bone components. Calcif Tissue Int 54:470e472. Tan U. 1985 Relationships between hand skill and the excitability of motoneurons innervating the postural soleus muscle in human subjects. Int J Neurosci 26:289e300. Macdonald HM, Nishiyama KK, Kang J, et al. 2011 Age-related patterns of trabecular and cortical bone loss differ between sexes and skeletal sites: a population-based HR-pQCT study. J Bone Mineral Res 26:50e62. Pauchard Y, Liphardt AM, Macdonald HM, et al. 2012 Quality control for bone quality parameters affected by subject motion in high-resolution peripheral quantitative computed tomography. Bone 50:1304e1310. MacNeil JA, Boyd SK. 2007 Accuracy of high-resolution peripheral quantitative computed tomography for measurement of bone quality. Med Eng Phys 29:1096e1105. Boutroy S, Bouxsein ML, Munoz F, et al. 2005 In vivo assessment of trabecular bone microarchitecture by high-resolution peripheral quantitative computed tomography. J Clin Endocrinol Metab 90:6508e6515. Laib A, Ruegsegger P. 1999 Comparison of structure extraction methods for in vivo trabecular bone measurements. Comput Med Imagaing Graph 23:69e74. Burghardt AJ, Buie HR, Laib A, et al. 2010 Reproducibility of direct quantitative measures of cortical bone microarchitecture of the distal radius and tibia by HR-pQCT. Bone 47:519e528. Macneil JA, Boyd SK. 2008 Bone strength at the distal radius can be estimated from high-resolution peripheral quantitative computed tomography and the finite element method. Bone 42:1203e1213. Pistoia W, van Rietbergen B, Lochmuller EM, et al. 2002 Estimation of distal radius failure load with micro-finite element analysis models based on three-dimensional peripheral quantitative computed tomography images. Bone 30:842e848. Lekamwasam S, Rodrigo M, de Silva KI, et al. 2005 Comparison of phalangeal bone mineral content and density between the dominant and non-dominant sides. Ceylon Med J 50:149e151. Plochocki JH. 2004 Bilateral variation in limb articular surface dimensions. Am J Hum Biol 16:328e333. Meszaros S, Ferencz V, Csupor E, et al. 2006 Comparison of the femoral neck bone density, quantitative ultrasound and bone density of the heel between dominant and non-dominant side. Eur J Radiol 60:293e298. Frost HM. 1997 On our age-related bone loss: insights from a new paradigm. J Bone Miner Res 12:1539e1546. Kim BT, Mosekilde L, Duan Y, et al. 2003 The structural and hormonal basis of sex differences in peak appendicular bone strength in rats. J Bone Mineral Res 18:150e155. Seeman E. 2002 Pathogenesis of bone fragility in women and men. Lancet 359:1841e1850.
Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health
Volume
-,
2015
