The authors have no conflict of interest.
Reproducibility and Side Differences of Mechanical Tests for Determining the Structural Strength of the Proximal Femur†
Article first published online: 22 DEC 2003
Copyright © 2004 ASBMR
Journal of Bone and Mineral Research
Volume 19, Issue 3, pages 379–385, March 2004
How to Cite
Eckstein, F., Wunderer, C., Boehm, H., Kuhn, V., Priemel, M., Link, T. M. and Lochmüller, E.-M. (2004), Reproducibility and Side Differences of Mechanical Tests for Determining the Structural Strength of the Proximal Femur. J Bone Miner Res, 19: 379–385. doi: 10.1359/JBMR.0301247
- Issue published online: 2 DEC 2009
- Article first published online: 22 DEC 2003
- Manuscript Accepted: 9 OCT 2003
- Manuscript Revised: 25 SEP 2003
- Manuscript Received: 14 JUL 2003
- bone mechanics;
- bone strength;
In this experimental study, we evaluated the reproducibility error of mechanical strength tests of the proximal femur when simulating a fall on the trochanter. Based on side differences in femoral failure loads in 55 pairs of femora, we estimated the upper limit of the precision error to be 15% for the side impact test, whereas the intersubject variability was >40%.
Introduction: Mechanical tests are commonly used as the gold standard for determining one of the main functions of bones, that is, to provide mechanical strength. However, it is unknown what magnitude of error is associated with these tests. Here we investigate the precision error and side difference of a side impact test of the proximal femur.
Materials and Methods: BMC was measured using DXA in 54 pairs of femora from donors 79.0 ± 10.6 years of age. Bones were tested to failure, simulating a fall on the greater trochanter.
Results: Failure loads were 3951 ± 1659N (CV% = 42%) on the right and 3900 ± 1652N (CV% = 42%) on the left (no significant side difference). The average random difference of femoral BMC was 7 ± 7% and that of femoral failure loads was 17 ± 12%. The correlation between BMC and failure load was 79% (r2), but the association between side differences in failure load with those in BMC was only 4%. When confining the analysis to pairs with less than 5% differences in BMC (n = 31), side differences in failure loads were 15 ± 13%. When correcting failure loads for side differences of BMC, the difference was 16 ± 15%
Conclusions: These results suggest that the upper limit of the precision error for femoral strength tests is ∼15% in a side impact configuration. Given the large intersubject variability of failure loads, this test provides an efficient tool for determining the structural strength of the proximal femur in a fall.
BIOMECHANICAL TESTS REPRESENT an established technique for determining one of the main functions of bone, which is to carry load and to provide resistance to deformation. The ability of bones to carry load without undergoing failure (fracture) is addressed as “structural” strength. This strength provides organisms with the ability to maintain posture within the gravitation field and to transfer forces across limb segments and joints during locomotion and movement. When the structural strength of bones deteriorates (for instance in osteoporosis), there is an increased risk of sustaining bone fracture. This particularly applies when high loads act on the bone, such as during a fall.
Mechanical tests have been commonly used to determine the structural strength of bones (1–3) and to estimate to what extent noninvasive “osteo-densitometric” methods(4-6) (such as DXA or CT) are able to estimate the structural strength of bones. These noninvasive methods play an important role in the context of diagnosing osteoporosis and estimating fracture risk in patients. Preclinical testing of these methods in an experimental (biomechanical) setting can thus help to determine the most efficient techniques and parameters for identifying patients at risk.(7-32) In this context, some studies have applied mechanical tests to assess whether computer simulations (finite element analysis) are able to improve the prediction of bone strength over conventional bone densitometry.(7-9)
Fractures of the proximal femur involve the most severe consequences of osteoporotic bone loss. This applies to the individual's well-being and to the economic burden on the resources of public health care systems.(33–42) Because of its high clinical relevance, there exists a large body of literature on mechanical tests in the proximal femur. Tests have been performed using the following loading configurations:
1. vertical loading of the femoral head, with the load being applied parallel to the shaft,(10–20)
2. simulating normal load transfer during the stance phase of gait (stance phase configuration), with the load being applied from superiomedial to the femoral head,(7, 8, 21–23) and
3. simulating a sideways fall on the greater trochanter (side impact configuration).(18–32)
However, it is currently unknown to what extent these mechanical tests are reproducible and what magnitude of error is associated with them. Because determination of the structural strength of bones is inevitably destructive, mechanical tests cannot be repeated in identical specimens; determination of precision errors is therefore not straight-forward.
One potential way to determine the precision error associated with mechanical tests of the proximal femur is to compare testing results in both sides of the same subject. Under these circumstances, however, precision errors must be differentiated from actual side differences in structural bone strength. Bilateral DXA measurements have suggested that BMC may differ substantially between the left and right.(43–49) Experimental work has also shown that BMC is highly correlated with failure strength and determines between 50% and 90% of the variability observed in femoral failure loads.(10-32)
This study was designed to determine the magnitude of precision error associated with a side impact test of the proximal femur. We tested the hypothesis that precision errors of this test can be estimated (1) by determining bone strength bilaterally and (2) by comparing side differences of bone strength in subjects with only small side-to-side deviations in BMC, or alternatively, by correcting side differences in failure strength for side differences in BMC.
MATERIALS AND METHODS
The study sample was initially comprised of 76 cadavers from a course of macroscopic anatomy. The criterion of inclusion in the course was the testamentary decree to the institute several years before death and can thus be assumed to be representative of the elderly population in Southern Germany. The individuals belonged to a wide range of medical and social backgrounds, but no detailed medical or social history was available. Specimens were fixed in alcohol/formalin solution.(50)
To identify specimens with bone diseases other than osteopenia, biopsy specimens were taken from the left iliac crest for histology. A 5-mm frontal section was obtained with a band saw, ∼7 cm posterior to the anterior, superior iliac spine (location of a transiliac biopsy). The samples were dehydrated and imbedded in methylmethacrylate without decalcification.(51) Five-micrometer sections were finally obtained, and these were stained with Goldner, Toluidine blue, and von Kossa. Four individuals were excluded based on histomorphometry, because they displayed signs of malignancy. In 3 subjects, neither of the two femora was available, and in 10 other subjects, only one femur was available because a total hip endoprosthesis was present on at least one side. In one subject, a fracture was detected on a standard radiograph before mechanical testing. In four subjects, DXA scans were unavailable or displayed insufficient quality to complete the analysis.
After excluding these subjects, the sample size was eventually comprised of 54 subjects 79.0 ± 10.6 years of age (age range, 52–100 years; 24 men and 30 women).
After the dissection course, the proximal femora were excised and cleaned of the surrounding soft tissue. The specimens were kept moist continuously between retrieval, radiography, DXA measurement, and mechanical testing. To exclude prior fracture or other local bone disorders, the bones were radiographed with a Faxitron X-ray system (Model 43885A; Faxitron; Hewlett Packard, McMinnville, OR, USA) at 40–85 kV, 2 mA, and time = 120 s using a 18 × 24-cm X-ray film (Agfa Structurix D7DW; Agfa, Leverkusen, Germany).
In vitro DXA scans of the femora were obtained using a standard narrow fan beam scanner with multiview image reconstruction (GE Lunar Prodigy; GE Lunar Corp., Madison, WI, USA), with the femoral specimens submerged in a water bath. The scans were automatically evaluated with the software provided by the manufacturer (Fig. 1), providing results of areal BMD (g/cm2) and BMC (g) in standard regions of interest (femoral neck, greater trochanter, intertrochanteric region [Ward], and total proximal femur).
Mechanical test of the proximal femur
Both femora were tested in a side impact configuration, simulating a sideways fall on the greater trochanter.(18–20) The femoral head and shaft faced downward and were able to move independently of one another on the support plates during loading of the trochanter (Fig. 2). One-half of a tennis ball with a lubricant was used to simulate cartilage contact with the femoral head. The load was applied to the greater trochanter through a pad, simulating a soft tissue cover. The shaft was positioned at 10° from horizontal and the neck at 15° internal rotation (Fig. 2). This loading configuration was adopted from a testing configuration developed at the Orthopedic Biomechanics Laboratory (Harvard Medical School, Boston, MA, USA)(2, 26-29) that has been developed based on experimental data on fall biomechanics.(52, 53)
Loads were applied at a rate of 6.6 mm/s; the failure load was defined as the peak of the load-deformation curve. After the test, the femora were radiographed, and the fracture pattern was classified both from the radiographs and from visual inspection. On the right side, 36 specimens displayed cervical fractures, 14 displayed intertrochanteric fractures, 2 displayed subtrochanteric fractures, and 2 displayed fractures of the shaft. On the left, 36 specimens displayed cervical fractures, 6 displayed intertrochanteric fractures, 6 displayed subtrochanteric fractures, and 6 displayed fractures of the shaft. In 30 pairs of specimens, the location of fracture (cervical, intertrochanteric, shaft) was identical on both sides, whereas in 24 pairs, fracture occurred at a different site.
In a first step, we analyzed the association between structural strength and BMC/BMD (DXA), using linear regression analysis. The highest correlation with the failure loads among different DXA parameters was found for the BMC of the total proximal femur, whereas the BMD of the total proximal femur and the BMC/BMD of other regions displayed somewhat lower associations. The correlation between total femoral BMC/BMD and femoral failure strength was r = 0.87/0.85 on the right, r = 0.85/0.78 on the left (Fig. 3), and r = 0.89/0.83 for the mean failure load at both sides versus the mean BMC/BMD from both sides. Because of the higher correlation of BMC with failure loads compared with BMD, side differences of bone properties were based on total proximal femoral BMC.
To assess precision errors and side difference in structural strength of the femora, we determined the systematic side differences of failure loads (N) between both sides as well as the systematic percentage differences (difference left versus right/measurements on the right side × 100). Systematic differences between both sides were tested for statistical significance using a paired Students t-test.
In the next step, we computed the random side differences of failure loads (N) between both sides after excluding the + and − sign of the differences and the random percentage differences (difference between left and right/mean of failure loads on both sides × 100).
To separate true side differences from precision errors of the test, we selected two approaches. In a first approach, the analysis was performed only for specimens in which the side difference of total proximal femur BMC was less than 5% (n = 31). In a second approach, the analysis was performed with failure loads on the left being corrected for side differences in total proximal femur BMC. If, for instance, the left proximal femur displayed a 7% lower BMC than the right side, the femoral failure load on the left was increased by 7% before making the comparison with the right. This approach was based on the notion that failure loads display a high linear correlation with BMC and that side differences in BMC may thus be accompanied by similar side differences of bone failure loads. Power relationships between BMC and failure loads were not considered in this context, because they have been shown to not improve the correlation in our current (Fig. 3) and previous studies on entire bones.(54)
As measures of reproducibility (precision), we computed the root mean square (RMS) SD and RMS CV% of repeated measurements (first measurement = right proximal femur, second measurement = left proximal femur).(55)
Failure loads were 3951 ± 1659N (CV% = 42%) for the right proximal femur and 3900 ± 1652N (CV% = 42%) for the left. The lowest recorded failure load in all 108 (54 × 2) mechanical tests was 660N, and the highest failure load was 8150N. No significant systematic side differences were observed in BMC and structural strength of the femora (p > 0.7). The correlation between failure loads on the left versus failure loads on the right was r = 0.89, and the SE of the estimate (y/x) was 775N (19.5% of the mean value).
The average random difference in total femoral BMC between the left and right side (after eliminating the +/− signs) was 1.9 ± 2.0 g (7 ± 7%), with a minimum of 0 g (0%) and a maximum of 9.8 g (34%). Twenty-three pairs of specimens displayed differences in total femoral BMC of >5%, 31 pairs differences of <5%, and 14 pairs differences of <3%.
The average random difference in structural bone strength (difference divided by the mean) between both sides was 627 ± 472N (17 ± 12%), with a minimum of 16N (0.3%) and a maximum of 1684N (57%). Note that the side differences plotted in Fig. 4 are up to 79%, because this graph depicts the percentage difference of the left versus right side. Average random side differences were 19 ± 13% in women and 16 ± 12% in men. When confining the analysis to the 31 pairs in which side differences of BMC did not exceed 5% (approach 1), side differences in failure loads became only slightly lower than in the total sample (532 ± 374N; 15 ± 13%). When restricting the analysis to the 14 pairs of femora with side differences in BMC of <3%, the differences in failure loads did not decrease further (480 ± 322N; 17 ± 14%)
When correcting failure loads on the left for side differences of bone mineral status (approach 2), the average random difference in failure loads became slightly lower than in the uncorrected data (553 ± 510N; 16 ± 15%). When restricting the analysis to specimens in which failure had occurred at the same anatomic region (neck, intertrochanteric, etc.) on both sides (n = 30), side differences in corrected failure loads did not become smaller in comparison with the total sample (516 ± 383N; 18 ± 14%).
The correlation between systematic side differences in failure loads and systematic side differences in bone mineral status (percent) was r = 0.20, indicating that only 4% of the side differences in bone failure loads were explained by side differences in BMC (Fig. 4). When being expressed as RMS SD and RMS CV% of repeated measures (first measurement = right proximal femur; second measurement = left proximal femur), the reproducibility of the test was 553N (15%) for the total sample, 427N (16%) for women, and 679N (14%) for men. The precision was 449N (14%) for the subsample, with a <5% difference in femoral BMC and 530N (15%) for the corrected failure loads.
Biomechanical tests represent a well-established technique for determining one of the main functions of bones, i.e., to provide structural mechanical strength to the organism. These tests are often considered a gold standard against which other methods are evaluated. However, it is unknown to what extent these tests are reproducible and what magnitude of precision error is associated with them. In contrast to determination of stiffness (Young's modulus) in cortical or trabecular bone tissue samples,(56) determination of the structural strength of the proximal femur is inevitably destructive. The tests cannot therefore be repeated on an identical specimen, and determination of precision errors is thus not straightforward.
One option to determine the reproducibility of the testing procedure is to repeatedly test artificial models of bone made from manmade material (sawbone). However, mechanical behavior of sawbones may deviate considerably from those of real bones, because it is impossible to model the natural bone architecture, including trabecular microstructure. Another option is to test real bones from both sides in the same subject. Under these circumstances, however, true side differences must be differentiated from precision errors. It is well known that there exist relevant physiological side differences in neuroanatomy and neurophysiology.(57, 58) However, most studies that have investigated side differences in musculoskeletal tissue morphology have reported a high degree of symmetry, with the left-to-right differences being considerably smaller than the intersubject variability.(59–62) Nevertheless, substantial side differences have been observed in individual cases in bilateral DXA measurements at the proximal femur, and these were found to exceed precision errors of the method.(43-49)
In this study, we observed random side differences of BMC of 7%, whereas the intersubject variability in BMC was ∼30% (CV%). As in most previous studies,(43–45, 47, 49) no systematic (left versus right) difference was found for the BMC, nor did we observe systematic differences in mechanical failure loads. However, random side differences of mechanical tests were considerably larger (17%) than those in BMC (7%). The intersubject variability of bone failure loads was 30% (CV%) for BMC and 41% for failure strength. However, regression analysis revealed that only 4% of the variability in side differences of failure loads is explained by the side differences in BMC. Side differences in mechanical failure loads decreased only slightly (15%) when confining the analysis to those specimens in which side differences of BMC were less than 5%. Also, side differences in structural strength were only slightly reduced when correcting mechanical failure loads for side differences in BMC (16%).
A limitation of this study is the lack of medical history and the fact that bones were fixed in alcohol-formalin solution. Regarding the lack of history, we made an attempt to compensate this deficit by obtaining radiographs and histology to exclude bone diseases other than osteopenia. No information was available on lower limb dominance of the subjects, but a previous study has suggested that side differences in hip BMC do not differ between right- and left-handed persons(47) and that lower limb dominance is not associated with higher muscle cross-sectional areas on the supposedly stronger side.(62) Regarding the fixation with formalin, we have previously shown that embalmment does not affect the measurement of BMC with DXA,(50) and the correlations reported here between failure load and BMC are at the higher end of those reported in previous studies on fresh specimens. We therefore assume that fixation does not affect the conclusions drawn from this study.
Our data suggest that the majority of variability in mechanical failure loads observed between left and right femora is caused by actual variability of the mechanical testing procedure and not because of side differences between both femora. We cannot exclude entirely that there exist structural differences between left and right femora that may explain part of the side difference, independent of BMC. However, among all parameters tested so far, BMC has explained a large proportion (50–90%) of variability in failure loads between specimens, and other (structural) variables have generally been found to be unable to improve the prediction versus BMC alone. At any rate, the values reported here clearly set an upper limit of precision errors to be expected in side impact testing of the proximal femur.
In conclusion, these data suggest that an upper limit of precision errors in the range of 550N (15% of the mean value) is to be expected in mechanical tests of the proximal femur in a side impact configuration. Given the large intersubject variability in bone failure loads in this and other loading configurations (here 1600N, 41%), mechanical testing seems to be adequately suited for reliably differentiating between subjects with low and high mechanical strength. Mechanical tests thus provide an efficient tool for investigating the potential of densitometric techniques to predict the structural strength of whole bones, but researchers are encouraged to investigate the quality of the test by determining side differences in relation to side differences in BMC. Given the relatively small SE (y/x) between bone mass (BMC) and bone failure loads for the femoral side impact test (750N, 19%), we hypothesize that it may be difficult to prove superiority of more sophisticated methods, such as finite element analysis, versus measurement of bone mass by DXA in relation to experimental data.
We thank Gudrun Goldmann for help with radiography. Parts of these results were obtained within the doctoral thesis of CW, which will be submitted to the Medical Faculty of the Ludwig-Maximilians-Universität München. This work was supported by Deutsche Forschungsgemeinschaft Grant DFG LO 730/2–2.
- 11991 Biomechanics of fracture risk prediction of the hip and spine by quantitative computed tomography. Radiol Clin North Am 29:1–18., ,
- 21997 Biomechanics of cortical and trabecular bone: Implications for assessment of fracture risk. In: MowVC, HayesWC (eds.) Basic Orthopaedic Biomechanics, 2nd ed. Lippincott-Raven, Philadelphia, PA, USA, pp. 69–111.,
- 32001 Understanding bone strength: Size isn't everything. Bone 29:101–104., ,
- 41996 Noninvasive assessment of bone mineral and structure: State of the art. J Bone Miner Res 11:707–730., , , , , , , , , , ,
- 52000 Advanced imaging of the macrostructure and microstructure of bone. Horm Res 54(Suppl 1):24–30., , , , , ,
- 62002 Clinical use of bone densitometry: Scientific review. JAMA 288:1889–1897., ,
- 71999 Femoral strength is better predicted by finite element models than QCT and DXA. J Biomech 32:1013–1020., , , , ,
- 81998 Prediction of femoral fracture load using automated finite element modeling. J Biomech 31:125–133., , ,
- 92002 Estimation of distal radius failure load with micro-finite element analysis models based on three-dimensional peripheral quantitative computed tomography images. Bone 30:842–848., , , , ,
- 101976 Bone mineral content and mechanical strength of the femoral neck. Acta Orthop Scand 47:503–508., ,
- 111982 The relationship between bone density, mineral content, and mechanical strength in the femoral neck. Clin Orthop 163:272–281., , , , , , ,
- 121988 Bone mineral content and mechanical strength. An ex vivo study on human femora at autopsy. Clin Orthop 227:292–297., ,
- 131989 Biomechanical properties of the proximal femur determined in vitro by single-energy quantitative computed tomography. J Bone Miner Res 4:715–722., ,
- 141990 Predicting femoral neck strength from bone mineral data. A structural approach. Invest Radiol 25:6–18., , , ,
- 151996 Prediction of fracture load at different skeletal sites by geometric properties of the cortical shell. J Bone Miner Res 11:1356–1363., ,
- 161998 Correlation of femoral and lumbar DXA and calcaneal ultrasound, measured in situ with intact soft tissues, with the in vitro failure loads of the proximal femur. Osteoporos Int 8:591–598., , , , , ,
- 172000 In situ femoral dual-energy X-ray absorptiometry related to ash weight, bone size and density, and its relationship with mechanical failure loads of the proximal femur. Osteoporos Int 11:361–367., , , , ,
- 182002 Bone strength at clinically relevant sites displays substantial heterogeneity and is best predicted from site-specific bone densitometry. J Bone Miner Res 17:162–171., , , , , ,
- 192002 Mechanical strength of the proximal femur as predicted from geometric and densitometric bone properties at the lower limb versus the distal radius. Bone 30:207–216., , ,
- 202003 Can novel clinical densitometric techniques replace or improve DXA in predicting bone strength in osteoporosis at the hip and other skeletal sites? J Bone Miner Res 18:906–912., , , ,
- 211989 Cancellous bone and mechanical strength of the femoral neck. Arch Orthop Trauma Surg 108:72–75., ,
- 221993 The relationship between femoral neck strength, bone mineral content and fracture fixation strength: An in vitro study. Osteoporos Int 3(Suppl 1):51–53., ,
- 231997 Volumetric quantitative computed tomography of the proximal femur: Precision and relation to bone strength. Bone 21:101–108., , , , , ,
- 241988 Contribution of the trabecular component to mechanical strength and bone mineral content of the femoral neck. An experimental study on cadaver bones. Scand J Clin Lab Invest 48:457–460., ,
- 251990 The use of quantitative computed tomography to estimate risk of fracture of the hip from falls. J Bone Joint Surg Am 72:689–700.,
- 261994 Effects of loading rate on strength of the proximal femur. Calcif Tissue Int 55:53–58., , ,
- 271995 Age-related reductions in the strength of the femur tested in a fall-loading configuration. J Bone Joint Surg Am 77:387–395., , ,
- 281995 Ultrasound and densitometry of the calcaneus correlate with the failure loads of cadaveric femurs. Calcif Tissue Int 56:99–103., ,
- 291996 Impact direction from a fall influences the failure load of the proximal femur as much as age-related bone loss. Calcif Tissue Int 58:231–235., , , ,
- 301997 Assessment of the strength of proximal femur in vitro: Relationship to femoral bone mineral density and femoral geometry. Bone 20:213–218., , , , , ,
- 311998 Experimental testing of a DEXA-derived curved beam model of the proximal femur. J Orthop Res 16:394–398., , , ,
- 321999 Prediction of the strength of the elderly proximal femur by bone mineral density and quantitative ultrasound measurements of the heel and tibia. Bone 25:49–54., ,
- 331992 Hip fractures in the elderly: A world-wide projection. Osteoporos Int 2:285–289., ,
- 341997 World-wide projections for hip fracture. Osteoporos Int 7:407–413., ,
- 352002 Incidence of limb fracture across Europe: Results from the European Prospective Osteoporosis Study (EPOS). Osteoporos Int 13:565–571., , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
- 361997 The socioeconomic burden of fractures: Today and in the 21st century. Am J Med 103:20S–25S.
- 371987 Secular trends in the incidence of hip fractures. Calcif Tissue Int 41:57–64., ,
- 381995 The worldwide problem of osteoporosis: Insights afforded by epidemiology. Bone 17:505S–511S.,
- 391996 A prospective study on socioeconomic aspects of fracture of the proximal femur. J Bone Miner Res 11:1935–1942., , , , ,
- 402001 Hip fracture incidence in East and West Germany: Reassessement ten years after unification. Osteoporos Int 12:136–139.,
- 411997 Outcomes and costs after hip fracture and stroke. A comparison of rehabilitation settings. JAMA 277:396–404., , , , , , ,
- 421997 Medical expenditures for the treatment of osteoporotic fractures in the United States in 1995: Report from the National Osteoporosis Foundation. J Bone Miner Res 12:24–35., , ,
- 431991 Variation between femurs as measured by dual energy X-ray absorptiometry (DEXA). Eur J Nucl Med Mol Imaging 18:38–40., ,
- 441995 Bilateral comparison of femoral bone density and hip axis length from single and fan beam DXA scans. Calcif Tissue Int 56:26–31., ,
- 452000 Bilateral measurement of femoral bone mineral density. J Clin Densitom 3:133–140., , ,
- 462000 An investigation of the diagnostic value of bilateral femoral neck bone mineral density measurements. Osteoporos Int 11:675–679., , , , ,
- 471992 Comparison and investigation of bone mineral density in opposing femora by dual-energy X-ray absorptiometry. Osteoporos Int 2:274–278., , ,
- 481997 Bone mineral density of opposing hips using dual energy X-Ray absorptiometry in single-beam and fan-beam design. Calcif Tissue Int 61:445–447., ,
- 492000 Is there a difference between right and left femoral bone density? J Clin Densitom 3:57–61., ,
- 502001 Effect of fixation, soft-tissues, and scan projection on bone mineral measurements with dual energy X-ray absorptiometry (DXA). Calcif Tissue Int 68:140–145., , ,
- 511991 Undecalcified preparation of bone tissue: Report of technical experience and development of new methods. Virchows Arch A Pathol Anat Histopathol 418:1–7., ,
- 521991 Prediction of femoral impact forces in falls on the hip. J Biomech Eng 113:366–374., ,
- 531996 Hip impact velocities and body configurations for voluntary falls from standing height. J Biomech 29:807–811., ,
- 541999 Lumbar vertebral body compressive strength evaluated by dual-energy X-ray absorptiometry, quantitative computed tomography, and ashing. Bone 25:713–724., , , ,
- 551995 Accurate assessment of precision errors: How to measure the reproducibility of bone densitometry techniques. Osteoporos Int 5:262–270., , , , ,
- 562001 Biomechanics of trabecular bone. Annu Rev Biomed Eng 3:307–333., , ,
- 571997 The genetics and evolution of handedness. Psychol Rev 104:714–727.
- 581989 Laterality and human evolution. Psychol Rev 96:492–505.
- 591993 Right-left differences in the lengths of human arm and leg bones. Kaibogaku Zasshi 68:536–543.
- 601970 Asymmetry in muscle weight and one-sided dominance in the human lower limbs. J Anat 106:553–556.,
- 611970 Functional asymmetry in the lower limbs. Acta Anat (Basel) 77:131–138.
- 622002 Side differences of knee joint cartilage volume, thickness, and surface area, and correlation with lower limb dominance—an MRI-based study. Osteoarthritis Cartilage 10:914–921., , , , ,