Clinical Biomechanics 30 (2015) 418–423
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
The effect of osteoarthritis on the regional anatomical variation of subchondral trabecular bone in the femoral head A. Tsouknidas a,⁎, K. Anagnostidis b, S. Panagiotidou c, N. Michailidis a a b c
Department of Mechanical Engineering, Aristotle University of Thessaloniki, School of Polytechnics, Building D, 54124 Thessaloniki, Greece Cardiff University Hospital of Wales, University Hospital Llandough, UK Department of Mechanical Engineering, University of Western Macedonia, Greece
a r t i c l e
i n f o
Article history: Received 24 August 2014 Accepted 16 March 2015 Keywords: Femoral head Osteopenetration Cancellous bone characteristics Osteoarthritis
a b s t r a c t Background: The subchondral trabecular bone is located deep inside the articular cartilage, with the subcapital region carrying up to 70% of the diurnal loads occurring in the hip joint. This leads to severe regional anatomical variations of subchondral trabecular bone in the femoral head and the purpose of this study was to examine whether osteoarthritis affects these topographic characteristics. Methods: 60 femoral heads were harvested during hip replacement and studied by osteopenetration at 8 predefined angles, at a penetration rate of 1 mm/s. Twenty-eight of the donors underwent surgery due to osteoarthritis, whereas the remaining were trauma patients with hip fractures. To correlate these measurements to non-invasive data, all specimens were scanned by micro Computed Tomography (μCT) prior to experimentation. A cross-sectional area, perpendicular to the needle penetration pathway, was analyzed and the deviations compared to the recorded osteopenetration energy. Findings: The experiments revealed significant topographical deviations in the trabeculae. These were more pronounced in the osteoarthritic samples which also required overall higher osteopenetration energy. A notable dependency of the directional bone strength to its cross-sectional characteristics was observed. Although the effect of “gender” on osteopenetration energy was proven to be significant, gender was not considered an independent variable in a regression model correlating osteopenetration energy to 2D trabecular bone density as this did not improve the value of the adjusted R2. Interpretation: The investigation provided refined insight into femoral head load-bearing capacity of patients suffering from osteoarthritis, as a comparison of osteoarthritic to healthy samples illustrated that subchondral trabecular bone in the femoral head region is subjected to increased remodeling and demineralization, reflected in higher osteopenetration values. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction The proximal femur is an anatomical site of major interest to clinicians as its mechanical failure is one of the most common reasons for pain and morbidity (Rockwood et al., 1990; Cooper et al., 1992). The primary bearing surface in the hip joint is the articular cartilage. The subchondral trabecular bone is located deep inside the articular cartilage and the percentage of the load carried by this tissue varies from 4% at the base of the neck to as much as 70% in the subcapital region (Lotz et al., 1995). This suggests that cancellous bone plays an important role in the mechanical strength of the proximal femur and especially of the femoral head. The biomechanical aspects of this are highly important in a number of surgical procedures, like hip resurfacing arthroplasty. Recent retrospective clinical studies have indicated loosening of the femoral ⁎ Corresponding author at: 9th floor of Building D, Department of Mechanical Engineering, Aristotle University Thessaloniki, 54124 Thessaloniki, Greece. E-mail address: [email protected] (A. Tsouknidas).
http://dx.doi.org/10.1016/j.clinbiomech.2015.03.017 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
component as one of the 4 most common failure rates in hip resurfacing procedures (Carrothers et al., 2010), while others have associated implant characteristics such as component size with the failure rate (Coulter et al., 2012). The anchorage of these implants is associated with the microstructural characteristics of the subchondral cancellous bone that drastically affects the femoral head's biomechanical response to loading. In order to adapt to its varied mechanical loads, trabecular bone is highly organized and oriented (Tsouknidas et al., 2015). This phenomenon is characterized by high complexity within joints where movement patterns result in multiaxial forces (Cappozzo, 2009) and are subject to specific loading scenarios (Cappozzo, 1991). As a result, bone strength is optimized in the orientation of its principal loading direction and a significant dependency of bone properties to the loading direction has been repeatedly documented. Investigations into the topographical variation of trabecular bone in the femoral head are usually conducted through medical imaging techniques, both ex vivo (Chiba et al., 2013) and in vivo (Chiba et al., 2011), while most heuristic studies focus only on the anisotropy of specific
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
regions (Magnussen et al., 2005). The limitations of hitherto preformed experiments originate in the applied testing procedure (e.g. compression or tension), as the extraction of the samples is a severely destructive procedure, thus not preserving the femoral head in a way able to provide multiple subchondral trabecular bone specimens. The purpose of this study was to examine how osteoarthritis affects the regional anatomical variation in the subchondral trabeculae of the femoral head, covering a wide spectrum of regions representative of various loading directions and magnitudes. This was based on the primary hypothesis that subchondral bone remodeling has been associated with the pathophysiology of osteoarthritis (Hayami et al., 2004) and therefore it stands to reason that the subcapital bone of osteoarthritic patients will exhibit differences in the femoral head region when compared to that of healthy subjects. In order to facilitate repeated measurements, a minimally destructive osteopenetration technique was applied and compared to μCT data, to determine whether potential observations could have been drawn through medical imaging techniques. Osteopenetration was originally developed to mechanically characterize trabecular bone at the knee joint during total knee arthroplasty (An and Draughn, 2000). This work was initiated in the 1970s (Sneppen et al., 1981), when mechanical loosening was still a significant problem in the semi- and non-constrained total knee designs, the long term function of which were dependent upon the mechanical quality of tibial and femoral condylar trabecular bone. The idea was to develop a tool to evaluate bone strength intraoperatively, thus leading to the introduction of the osteo-penetrometer (Hvid et al., 1984; Hvid, 1985). 2. Materials and methods This study was conducted on femoral head samples, harvested from 37 patients undergoing total hip replacement due to osteoarthritis. During the surgical procedure and after a 45° osteotomy, femoral heads were removed and stored at −60 °C until evaluation. In order to determine the samples' structural integrity, standard X-rays (anterior–posterior) of the pelvis were taken preoperatively in all cases. The control group consisted of 32 patients without osteoarthritis, who underwent hemi-arthroplasty or total hip arthroplasty for hip fracture. Written consent was obtained from all donors (osteoarthritic and control group) according to the institution's ethics rules. All donors were subjected to dual-energy X-ray absorptiometry (DXA) of the contralateral hip, to catalog their proximal femur bone mineral density and exclude osteoporotic patients from the study (e.g. candidates with T-scores lower than − 2.5). Further exclusion criteria were applied to all patients diagnosed with any type of metabolic disease, cancer, osteoporosis, large cysts or previous surgeries in the proximal femur, excluding them from the study. This resulted in a total of 28 osteoarthritic donors (11 female and 17 male) which were considered as representative for this study. The average age of these patients was 75.6 years (ranging from 63 to 88). In a similar fashion, the control group consisted of 24 patients (8 female and 16 male) with a mean age of 78.2 years (ranging from 71 to 92). All specimens were scanned with a μCT device (Werth TomoScope® HV Compact) to reconstruct their 3D shape. The measurements were conducted at a spatial resolution of 10 μm; a high image resolution was chosen as there exists a consensus throughout the literature that measurement accuracy directly affects geometric discretization and model convergence (Bevill and Keaveney, 2009). Data acquisition was in accordance with DICOM (Digital Imaging and Communications in Medicine) which allowed for the conversion of multiple 2D images into a 3D volume. Interpolation of the obtained measurements ensured higher representation accuracy, even though this process did not result in higher resolution of the sample. The smoother representation facilitated the distinct removal of the remaining soft tissue, which was neither visible nor accessible during the initial defatting process. A semiautomated segmentation technique, supported by manual correction of the threshold results was followed. During this multi threshold
419
segmentation, the mean gray-scale within the image is calculated, while sensitive edge detection filters were employed (Rathnayaka et al., 2010; Canny, 1986) to distinguish the apparent tissue types. The reconstructed volumes exhibited porosity values ranging from 68.52 to 91.38%, while all the sample characteristics were in agreement with the existing literature data (Baroud et al., 2004). To experimentally determine the regional anatomical variation of the trabecular bone in the femoral head, the method of osteopenetration An and Draughn (2000) was employed. A needle with a cross-section area of 4.9 mm2 was used, according to the literature (Hvid et al., 1984); a representative drawing and picture of this needle are illustrated in Fig. 1. The needle penetrated the surface of the femoral heads at 8 different sites, representing regions subjected to different amounts of in vivo loading (Baroud et al., 2004; Thomas and Daniel, 1983): the medial region (reg-1) located in fovea capitis femoris, the inferior region (reg-2), the medial-superior region (reg-3), the superior region (reg-4), the anterior region (reg-5), the anterior-medial region (reg-6), the posterior-medial region (reg-7) and the posterior region (reg-8). The needle was mounted on the crosshead of an electric INSTRON Testing system and engaging the surface of the bone specimens perpendicular to the aforementioned femoral head regions. The surface of the cartilage was processed prior to the experiments, by drilling away the sclerotic bone surface (drilling depth was determined by the μCT measurements) as failure to engage the relatively strong surface of the bone specimens at right angles could result in bending of the needle and invalidation of the measurements. The samples (femoral heads) were rigidly supported and the penetration speed set to 1 mm/s. Following this, the experimental results were correlated to data obtained through the μCT measurements. To facilitate this, the crosssectional area of 8 circular regions of the subchondral trabecular bone (perpendicular to aforementioned areas and 6 mm in diameter) was measured at a 7 mm depth below the surface of the femoral head (see Fig. 2). During the osteopenetration experiments, the resulting force was recorded as a function of the needle penetration depth. The initial 4 mm of the needle displacement was disregarded, and considered as the engagement profile of the needle, which was pushed into the femoral head up to a depth of 14 mm. The 4 mm was determined, by μCT measurements, as the mean value of the cortical shell and thus not considered to avoid erroneous statements due to the cortical stiffness. The area beneath the osteopenetration curve, represents the penetration energy, expressed in Joules. The resulting values are considered as representative for bone density and strength in the same area and reflected the 4–14 mm penetration course of the needle in all specimens.
Fig. 1. Osteopenetration needle and engagement directions.
420
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
Fig. 2. Cross-sectional area of subchondral bone measured 7 mm below the femoral head surface and perpendicular to the medial-superior region. Fig. 4. The mean values of penetration energy in the 8 regions for samples with healthy BMD values and osteoarthritic ones.
Differences in the osteopenetration energy as well as in the 2D trabecular bone density among the 8 different sites were statistically investigated among the 2 donor groups but also gender specifically, using repeated measurements analysis of variance (ANOVA). 3. Results 3.1. Experimental results A characteristic load-displacement diagram of the osteopenetration experiments is demonstrated in Fig. 3. The saw-like curve observed in the diagram is due to the microstructure of the trabecular bone and the gradual destruction/collapse of trabeculae inside the femoral head. As expected, significant topographical variations of trabecular bone strength in different subchondral bone regions were documented, the mean values of osteopenetration energy are documented below for the control group followed in parentheses for osteoarthritic patients. In the medial region (fovea capitis femoris) the mean osteopenetration energy (in Joules) was 0.74 (0.84). In the inferior region the mean value was 0.84 (0.91), in the medial-superior region 2.27 (2.92), in superior region 3.08 (3.88), in anterior region 1.36 (1.42), in anterior-medial region 1.63 (1.99), in posterior-medial region 1.57 (1.76) and in posterior region 1.79 (2.32) J respectively. Fig. 4 shows the mean value of penetration energy calculated at all sites, for measurements conducted on femoral head samples with healthy BMD values as well as osteoarthritic ones. The μCT measurements conducted on the same samples are summarized in Fig. 5, for case vs. control samples. Medical imaging revealed that the 2D trabecular bone density (trabecular bone crossing a circle of
Fig. 3. Typical measurement curve of penetration tests.
6 mm in diameter) was the highest in the superior and medialsuperior regions in comparison with the fovea capitis femoris region. In the inferior region the bone strength was lower (increased, however when compared with the region of fovea capitis femoris which was the weakest region). Anterior and posterior regions (reg-5, 6, 7 and 8 respectively) had intermediate values, with the posterior region exhibiting a significantly greater mean value than that of the anterior region. 3.2. Statistical analysis ANOVA accounted for both, independent (case vs. control group and male vs. female donors) as well as paired (regional variations observed within specimens of the same donor) data. For this purpose, site-specific differences were treated as “within-subject factor” (studied at 8 sites of an individual donor), while donor group and gender were treated as “between-subjects factors” (studied at an inter-patient level). ANOVA showed that all factors (“site”, “donor group” and “gender”) have a significant effect on both dependent variables (osteopenetration
Fig. 5. Mean cross-sectional areas as registered for the 8 primary subchondral regions for case vs. control in femoral head samples (Y-axis starts at 16 mm2 to intensify the apparent deviations).
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
421
energy and 2D trabecular bone density). All p-values were lower than 0.01. In detail: ➢ the case group is characterized by a higher value of osteopenetration energy compared to the control group (by 0.35 J on average) and a higher value of 2D trabecular bone density compared to the control group (by 0.22 mm2 on average) ➢ male donors incur a higher value of osteopenetration energy compared to the female ones (by 0.10 J on average) and a higher value of 2D trabecular bone density compared to the female ones (by 0.19 mm2 on average) ➢ the interaction effect of donor group and gender was not significant for either one of the response variables (“osteopenetration energy” and “2D trabecular bone density”).
As far as the site of the femoral head is concerned, post hoc tests were conducted to examine specific differences between sites and site-specific ANOVA was employed to examine the effect of case vs. control group within each site. For the osteopenetration energy case (Fig. 4) all differences between sites are significant (p-values ≪ 0.01) and the difference of case vs. control group in all sites is also significant (p-values b 0.01). For the 2D trabecular bone density case (Fig. 5) all differences between sites are significant (p-values ≪ 0.01) and the difference of case vs. control group in most sites is also significant (p-values b 0.01 in sites 1, 3, 4, 6, 7, 8, while p-value = 0.357 in site 2 and p-value = 0.301 in site 5). The relative measurements of both, osteopenetration experiments and 2D trabecular bone density correlate well within each site and donor group. The respective correlation coefficients are summarized in Table 1 and they have all proven to be statistically significant (p-values b b 0.01). According to the calculated Pearson correlation values, osteopenetration energy has a positive linear relationship with 2D trabecular bone density in all sites and for both donor groups (although the exact expression of the fitted line may be different from site to site and for case or control group). This finding is very interesting since it enables the prediction of osteopenetration energy in any site of the femoral head of both osteoarthritic and non-osteoarthritic patients using the value of 2D trabecular bone density, which can easily be obtained through μCT measurements. To this end, a regression line needs to be fitted in the osteopenetration energy–2D trabecular bone density data for every site of the femoral head and every donor group. For example, Fig. 6 presents the regression line and 95% prediction interval of the osteopenetration energy in site 4 of the femoral head of osteoarthritic patients with respect to 2D trabecular bone density. Note that, although the effect of “gender” on osteopenetration energy was proven to be significant, it was not considered as an independent variable during the prediction of osteopenetration energy, since its impact is fully reflected in the 2D trabecular bone density measurements. That is, incorporating “gender” in a regression model which relates osteopenetration energy to 2D trabecular bone density, has, in most cases, nothing to offer. For example in the case of Fig. 6 (femoral head region 4 of osteoarthritic patients), adding “gender”
Fig. 6. Regression line and 95% prediction interval of the osteopenetration energy based on 2D trabecular bone density (in site 4).
in the regression model is pointless, since it does not increase the value of the adjusted R2. Following the principle presented in Fig. 6, the correlation of osteopenetration energy (Υ) to the 2D trabecular bone density (Χ) was determined as linear in all 8 regions. The site-specific regression coefficients β0 and β1 of this dependency (Υ = β0 + β1 × Χ) are summarized in Table 1 for all sites and for both groups (osteoarthritic and control). 4. Discussion Osteopenetration is useful in describing deviations of trabecular bone strength in subchondral bone regions, it was introduced in this study to examine topographical variations in subcapital trabecular bone of the femoral head, as it is capable of providing quasi-three-dimensional information on a combination of bone density and mechanical strength of bone, not readily obtained by other methods. The eight osteopenetration sites chosen for this study, represent regions subjected to different amounts of in vivo loading. The superior region being the most heavily loaded, posterior and anterior partially loaded and medial and inferior being the least loaded. The results are in agreement with loading distribution described in other studies (Thomas and Daniel, 1983; Hodge et al., 1986). The penetration with the 2.5 mm needle resulted in minimal damage to the trabecular structure and was mainly localized around the pathway of the needle (Hvid, 1985). This was ensured through a follow-up μCT evaluation of the femoral head samples demonstrating the absence of interference between the 8 consecutive indents. This facilitated fairly closely spaced measurements in regions with minimal distance between them, providing refined insight to the topographical variation of subchondral trabecular bone in the femoral head, as this kind of data could not be obtained in the past through other destructive tests (e.g. compression tests). A limitation however of the introduced technique is associated with the depth measurements considering the actual surface of the femoral head rather than global surface, thus not compensating for the depression at the fovea.
Table 1 Configuration of the correlation coefficients (CC) and regression coefficients β0, β1 of osteopenetration energy (dependent variable) against 2D trabecular bone density (independent variable), for all sites and both patient groups. Femoral head region
Osteoarthritic group
Control group
β0= β1= CC β0= β1= CC
1
2
3
4
5
6
7
8
−9.538 0.603 0.923 −8.468 0.546 0.933
−10.300 0.636 0.923 −10.990 0.673 0.865
−4.457 0.392 0.920 −8.621 0.594 0.941
−3.243 0.340 0.898 −11.110 0.683 0.897
−4.727 0.337 0.907 −2.968 0.236 0.912
−6.653 0.457 0.884 −5.916 0.417 0.826
−5.274 0.396 0.808 −20.670 1.292 0.911
−6.600 0.464 0.928 −10.600 0.651 0.919
422
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
The results available in literature coincide with the ones obtained here. Brown and Ferguson (1980) and Martens et al. (1983) performed compression tests by loading in three directions: anterior–posterior, superior–inferior and medial–lateral. They showed that anisotropy was evident and an increase in stiffness was found in the regions traversed by the primary trabecular system. In their studies however, bone specimens were obtained from nonspecific regions of the entire femoral heads, whereas here eight different but well defined regions were examined. Sugita et al. (1999) examined the differences in anisotropy in osteoporotic bone in the primary compressive group of the femoral head. They found increased values of compressive stiffness in the parallel loading group compared to the perpendicular loading group, but the anisotropic behavior of cancellous bone was reduced, and the femoral head became isotropic as the bone density decreased (e.g. in osteoporosis). All previous findings correlate with both, osteopenetration experiments and 2D trabecular bone density measurements conducted within this study. The medial-superior region (3) and superior region (4) displayed the highest capacity to withstand the needle's penetration, whereas the fovea capitis femoris (region 1) exhibited the lowest osteopenetration energy. This correlates well to bone density in the orientation of the fovea capitis femoris, which was lower and exhibited a smaller spread of architectural variation than regions 3 and 4 (see Fig. 5). This led the needle to engage fewer trabecular struts during the penetration, thus rendering the fovea capitis femoris more susceptible to loading (mean osteopenetration energy values), whereas anterior and posterior regions reflect similar cross-sectional areas and penetration values. Grynpas et al. (1991) associated osteoarthritis with a thickening of the subchondral trabecular bone with abnormally low mineralization patterns, which was repeatedly documented in the past (Lajeunesse and Reboul, 2003). Based on these factors the overall compressive strength of the tissue was not expected to increase when compared to that of healthy subjects (Bailey et al., 2004). The sclerosis, however, of subchondral trabecular bone along with the formation osteophytes observed in other studies (Oettmeier and Abendroth, 1989) is indicative of the increased osteopenetration energy observed in this study. All sites of the osteoarthritic specimens exhibited an increased capacity to withstand the penetrating needle but the recorded augmentation of the medial-superior region (reg-3), the superior region (reg-4) and the posterior region (reg-8) was more pronounced when compared to the remaining penetration sites. This can be attributed to the increased turnover of the bone in the subcapital region leading to manifested remodeling in the high loaded regions of the femoral head (based on Wolf's principle) and has also been observed by other groups (Layton et al., 1988). The spread of architectural variation within each scanned region was noticeably higher in the osteoarthritic samples than in the controls (a behavior demonstrated in Fig. 5). A greater site variation of both bone density has also been documented by Li and Aspden (1997), who attributed this behavior to a 12% decreased mineral density of the subchondral bone when compared to that of healthy samples. Li and Aspden (1997) determined an increased stiffness of osteoarthritic specimens when compared to healthy ones which correlates well to the increased osteopenetration energy presented in this study for the case samples vs. control. 5. Conclusions In the present investigation, the effect of osteoarthritis on the regional anatomical variation of the subchondral trabecular bone in the femoral head, was examined by osteopenetration and compared to data obtained through non-invasive medical imaging techniques. The comparison of osteoarthritic to healthy samples was based on the primary hypothesis that the subchondral trabecular bone in the femoral head region is subjected to increased remodeling and demineralization.
This tendency was verified in all penetration directions and sustained non-invasively by μCT measurements conducted prior to experimentation. The 2D bone density seems to affect the required penetration force as an almost linear dependency to penetration energy was observed, indicating that micro-scale imaging of cancellous bone can be used as an indicator of both strength characteristics and topographical variation of subchondral trabecular bone in the femoral head. The results provide refined insight to femoral head load-bearing capacity that could among others improve the anchorage of patient specific hip resurfacing implants through proper component design and positioning. Acknowledgments The authors would like to acknowledge the Technical University of Serres, Greece for providing them access to their Laboratory's μCT device, and in particular Dr. D. Sagris and Prof. C. David. References An, Y., Draughn, R., 2000. Mechanical Testing of Bone and the Bone–implant Interface. CRC Press, pp. 241–246. Bailey, A.J., Mansell, J.P., Sims, T.J., Banse, X., 2004. Biochemical and mechanical properties of subchondral bone in osteoarthritis. Biorheology 41 (3–4), 349–358. Baroud, G., Falk, R., Crookshank, M., Sponagel, S., Steffen, T., 2004. Experimental and theoretical investigation of directional permeability of human vertebral cancellous bone for cement infiltration. J. Biomech. 37 (2), 189–196. Bevill, G., Keaveney, T.M., 2009. Trabecular bone strength predictions using finite element analysis of micro-scale images at limited spatial resolution. Bone 44, 579–584. Brown, T.D., Ferguson, A.B., 1980. Mechanical property distributions in the cancellous bone of the proximal femur. Acta Orthop. Scand. 51, 429–437. Canny, J., 1986. A computational approach to edge detection. IEEE Trans. Pattern Anal. 8 (6), 679–698. Cappozzo, A., 1991. The mechanics of human walking. Adv. Psychol. 78, 167–186. Cappozzo, A., 2009. The observation of human joint movement. 4th European Conference of the International Federation for Medical and Biological Engineering IFMBE Proceedings. 22, pp. 126–129. Carrothers, A.D., Gilbert, R.E., Jaiswal, A., Richardson, J.B., 2010. Birmingham hip resurfacing: the prevalence of failure. J. Bone Joint Surg. 10, 1344–1350 (British Volume). Chiba, K., Ito, M., Osaki, M., Uetani, M., Shindo, H., 2011. In vivo structural analysis of subchondral trabecular bone in osteoarthritis of the hip using multi-detector row CT. Osteoarthr. Cartil. 19, 180–185. Chiba, K., Burghardt, A.J., Osaki, M., Majumdar, S., 2013. Heterogeneity of bone microstructure in the femoral head in patients with osteoporosis: an ex vivo HR-pQCT study. Bone 56, 139–146. Cooper, C., Campion, G., Melton, L.J., 1992. Hip fractures in the elderly: a world-wide projection. Osteoporos. Int. 2, 285–289. Coulter, G., Young, D.A., Dalziel, R.E., Shimmin, A.J., 2012. Birmingham hip resurfacing at a mean of ten years: results from an independent centre. J. Bone Joint Surg. 94 (3), 315–321 (British Volume). Grynpas, M.D., Alpert, B., Katz, I., Lieberman, I., Pritzker, K.P.H., 1991. Subchondral bone in osteoarthritis. Calcif. Tissue Int. 49 (1), 20–26. Hayami, T., Pickarski, M., Wesolowski, G.A., McLane, J., Bone, A., Destefano, J., Rodan, G.A., Duong le, T., 2004. The role of subchondral bone remodeling in osteoarthritis: reduction of cartilage degeneration and prevention of osteophyte formation by alendronate in the rat anterior cruciate ligament transection model. Arthritis Rheum. 50 (4), 1193–1206. Hodge, W.H., Fijan, R.S., Carlson, K.L., Burgess, R.G., Harris, W.H., Mann, R.W., 1986. Contact pressures in the human hip joint measured in vivo. Proc. Natl. Acad. Sci. U. S. A. 83, 2879–2883. Hvid, I., 1985. Cancellous bone at the knee: a comparison of two methods of strength measurement. Arch. Orthop. Trauma Surg. 104, 211–217. Hvid, I., Andersen, K., Olesen, S., 1984. Cancellous bone strength measurements with the osteopenetrometer. Eng. Med. 13, 73–78. Lajeunesse, D., Reboul, P., 2003. Subchondral bone in osteoarthritis: a biologic link with articular cartilage leading to abnormal remodeling. Curr. Opin. Rheumatol. 15 (5), 628–633. Layton, M.W., Goldstein, S.A., Goulet, R.W., Feldkamp, L.A., Kubinski, D.J., Bole, G.G., 1988. Examination of subchondral bone architecture in experimental osteoarthritis by microscopic computed axial tomography. Arthritis Rheum. 31 (11), 1400–1405. Li, B., Aspden, R.M., 1997. Composition and mechanical properties of cancellous bone from the femoral head of patients with osteoporosis or osteoarthritis. J. Bone Miner. Res. 12 (4), 641–651. Lotz, J.C., Cheal, E.J., Hayes, W.C., 1995. Stress distributions within the proximal femur during gait and falls: implications for osteoporotic fracture. Osteoporos. Int. 5, 252–261. Magnussen, R.A., Guilak, F., Vail, T.P., 2005. Cartilage degeneration in post-collapse cases of osteonecrosis of the human femoral head: altered mechanical properties in tension, compression, and shear. J. Orthop. Res. 23, 576–583. Martens, M., Van Audekercke, R., De Meester, P., Mulier, J.C., 1983. The mechanical characteristics of cancellous bone at the upper femoral region. J. Biomech. 12, 971–983.
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423 Oettmeier, R., Abendroth, K., 1989. Osteoarthritis and bone: osteologic types of osteoarthritis of the hip. Skelet. Radiol. 8, 165–174. Rathnayaka, K., Sahama, T., Schuetz, M.A., Schmutz, B., 2010. Effects of CT image segmentation methods on the accuracy of long bone 3D reconstructions. Med. Eng. Phys. 33 (2), 226–233. Rockwood, P.R., Horne, J.G., Cryer, C., 1990. Hip fractures: a future epidemic? J. Orthop. Trauma 4, 388–396. Sneppen, O., Christensen, P., Larsen, H., Vang, P.S., 1981. Mechanical testing of trabecular bone in knee replacement. Int. Orthop. 5, 251–256.
423
Sugita, H., Oka, M., Toguchida, J., Nakamura, T., Ueo, T., Hayami, T., 1999. Anisotropy of osteoporotic cancellous bone. Bone 24, 513–516. Thomas, D.B., Daniel, T.S., 1983. In vitro contact stress distributions in the natural human hip. J. Biomech. 16, 373–384. Tsouknidas, A., Maliaris, G., Savvakis, S., Michailidis, N., 2015. Anisotropic post-yield response of cancellous bone simulated by stress–strain curves of bulk equivalent structures. Comput. Methods Biomech. Biomed. Engin. 18 (8), 839–846.
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
The effect of osteoarthritis on the regional anatomical variation of subchondral trabecular bone in the femoral head A. Tsouknidas a,⁎, K. Anagnostidis b, S. Panagiotidou c, N. Michailidis a a b c
Department of Mechanical Engineering, Aristotle University of Thessaloniki, School of Polytechnics, Building D, 54124 Thessaloniki, Greece Cardiff University Hospital of Wales, University Hospital Llandough, UK Department of Mechanical Engineering, University of Western Macedonia, Greece
a r t i c l e
i n f o
Article history: Received 24 August 2014 Accepted 16 March 2015 Keywords: Femoral head Osteopenetration Cancellous bone characteristics Osteoarthritis
a b s t r a c t Background: The subchondral trabecular bone is located deep inside the articular cartilage, with the subcapital region carrying up to 70% of the diurnal loads occurring in the hip joint. This leads to severe regional anatomical variations of subchondral trabecular bone in the femoral head and the purpose of this study was to examine whether osteoarthritis affects these topographic characteristics. Methods: 60 femoral heads were harvested during hip replacement and studied by osteopenetration at 8 predefined angles, at a penetration rate of 1 mm/s. Twenty-eight of the donors underwent surgery due to osteoarthritis, whereas the remaining were trauma patients with hip fractures. To correlate these measurements to non-invasive data, all specimens were scanned by micro Computed Tomography (μCT) prior to experimentation. A cross-sectional area, perpendicular to the needle penetration pathway, was analyzed and the deviations compared to the recorded osteopenetration energy. Findings: The experiments revealed significant topographical deviations in the trabeculae. These were more pronounced in the osteoarthritic samples which also required overall higher osteopenetration energy. A notable dependency of the directional bone strength to its cross-sectional characteristics was observed. Although the effect of “gender” on osteopenetration energy was proven to be significant, gender was not considered an independent variable in a regression model correlating osteopenetration energy to 2D trabecular bone density as this did not improve the value of the adjusted R2. Interpretation: The investigation provided refined insight into femoral head load-bearing capacity of patients suffering from osteoarthritis, as a comparison of osteoarthritic to healthy samples illustrated that subchondral trabecular bone in the femoral head region is subjected to increased remodeling and demineralization, reflected in higher osteopenetration values. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction The proximal femur is an anatomical site of major interest to clinicians as its mechanical failure is one of the most common reasons for pain and morbidity (Rockwood et al., 1990; Cooper et al., 1992). The primary bearing surface in the hip joint is the articular cartilage. The subchondral trabecular bone is located deep inside the articular cartilage and the percentage of the load carried by this tissue varies from 4% at the base of the neck to as much as 70% in the subcapital region (Lotz et al., 1995). This suggests that cancellous bone plays an important role in the mechanical strength of the proximal femur and especially of the femoral head. The biomechanical aspects of this are highly important in a number of surgical procedures, like hip resurfacing arthroplasty. Recent retrospective clinical studies have indicated loosening of the femoral ⁎ Corresponding author at: 9th floor of Building D, Department of Mechanical Engineering, Aristotle University Thessaloniki, 54124 Thessaloniki, Greece. E-mail address: [email protected] (A. Tsouknidas).
http://dx.doi.org/10.1016/j.clinbiomech.2015.03.017 0268-0033/© 2015 Elsevier Ltd. All rights reserved.
component as one of the 4 most common failure rates in hip resurfacing procedures (Carrothers et al., 2010), while others have associated implant characteristics such as component size with the failure rate (Coulter et al., 2012). The anchorage of these implants is associated with the microstructural characteristics of the subchondral cancellous bone that drastically affects the femoral head's biomechanical response to loading. In order to adapt to its varied mechanical loads, trabecular bone is highly organized and oriented (Tsouknidas et al., 2015). This phenomenon is characterized by high complexity within joints where movement patterns result in multiaxial forces (Cappozzo, 2009) and are subject to specific loading scenarios (Cappozzo, 1991). As a result, bone strength is optimized in the orientation of its principal loading direction and a significant dependency of bone properties to the loading direction has been repeatedly documented. Investigations into the topographical variation of trabecular bone in the femoral head are usually conducted through medical imaging techniques, both ex vivo (Chiba et al., 2013) and in vivo (Chiba et al., 2011), while most heuristic studies focus only on the anisotropy of specific
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
regions (Magnussen et al., 2005). The limitations of hitherto preformed experiments originate in the applied testing procedure (e.g. compression or tension), as the extraction of the samples is a severely destructive procedure, thus not preserving the femoral head in a way able to provide multiple subchondral trabecular bone specimens. The purpose of this study was to examine how osteoarthritis affects the regional anatomical variation in the subchondral trabeculae of the femoral head, covering a wide spectrum of regions representative of various loading directions and magnitudes. This was based on the primary hypothesis that subchondral bone remodeling has been associated with the pathophysiology of osteoarthritis (Hayami et al., 2004) and therefore it stands to reason that the subcapital bone of osteoarthritic patients will exhibit differences in the femoral head region when compared to that of healthy subjects. In order to facilitate repeated measurements, a minimally destructive osteopenetration technique was applied and compared to μCT data, to determine whether potential observations could have been drawn through medical imaging techniques. Osteopenetration was originally developed to mechanically characterize trabecular bone at the knee joint during total knee arthroplasty (An and Draughn, 2000). This work was initiated in the 1970s (Sneppen et al., 1981), when mechanical loosening was still a significant problem in the semi- and non-constrained total knee designs, the long term function of which were dependent upon the mechanical quality of tibial and femoral condylar trabecular bone. The idea was to develop a tool to evaluate bone strength intraoperatively, thus leading to the introduction of the osteo-penetrometer (Hvid et al., 1984; Hvid, 1985). 2. Materials and methods This study was conducted on femoral head samples, harvested from 37 patients undergoing total hip replacement due to osteoarthritis. During the surgical procedure and after a 45° osteotomy, femoral heads were removed and stored at −60 °C until evaluation. In order to determine the samples' structural integrity, standard X-rays (anterior–posterior) of the pelvis were taken preoperatively in all cases. The control group consisted of 32 patients without osteoarthritis, who underwent hemi-arthroplasty or total hip arthroplasty for hip fracture. Written consent was obtained from all donors (osteoarthritic and control group) according to the institution's ethics rules. All donors were subjected to dual-energy X-ray absorptiometry (DXA) of the contralateral hip, to catalog their proximal femur bone mineral density and exclude osteoporotic patients from the study (e.g. candidates with T-scores lower than − 2.5). Further exclusion criteria were applied to all patients diagnosed with any type of metabolic disease, cancer, osteoporosis, large cysts or previous surgeries in the proximal femur, excluding them from the study. This resulted in a total of 28 osteoarthritic donors (11 female and 17 male) which were considered as representative for this study. The average age of these patients was 75.6 years (ranging from 63 to 88). In a similar fashion, the control group consisted of 24 patients (8 female and 16 male) with a mean age of 78.2 years (ranging from 71 to 92). All specimens were scanned with a μCT device (Werth TomoScope® HV Compact) to reconstruct their 3D shape. The measurements were conducted at a spatial resolution of 10 μm; a high image resolution was chosen as there exists a consensus throughout the literature that measurement accuracy directly affects geometric discretization and model convergence (Bevill and Keaveney, 2009). Data acquisition was in accordance with DICOM (Digital Imaging and Communications in Medicine) which allowed for the conversion of multiple 2D images into a 3D volume. Interpolation of the obtained measurements ensured higher representation accuracy, even though this process did not result in higher resolution of the sample. The smoother representation facilitated the distinct removal of the remaining soft tissue, which was neither visible nor accessible during the initial defatting process. A semiautomated segmentation technique, supported by manual correction of the threshold results was followed. During this multi threshold
419
segmentation, the mean gray-scale within the image is calculated, while sensitive edge detection filters were employed (Rathnayaka et al., 2010; Canny, 1986) to distinguish the apparent tissue types. The reconstructed volumes exhibited porosity values ranging from 68.52 to 91.38%, while all the sample characteristics were in agreement with the existing literature data (Baroud et al., 2004). To experimentally determine the regional anatomical variation of the trabecular bone in the femoral head, the method of osteopenetration An and Draughn (2000) was employed. A needle with a cross-section area of 4.9 mm2 was used, according to the literature (Hvid et al., 1984); a representative drawing and picture of this needle are illustrated in Fig. 1. The needle penetrated the surface of the femoral heads at 8 different sites, representing regions subjected to different amounts of in vivo loading (Baroud et al., 2004; Thomas and Daniel, 1983): the medial region (reg-1) located in fovea capitis femoris, the inferior region (reg-2), the medial-superior region (reg-3), the superior region (reg-4), the anterior region (reg-5), the anterior-medial region (reg-6), the posterior-medial region (reg-7) and the posterior region (reg-8). The needle was mounted on the crosshead of an electric INSTRON Testing system and engaging the surface of the bone specimens perpendicular to the aforementioned femoral head regions. The surface of the cartilage was processed prior to the experiments, by drilling away the sclerotic bone surface (drilling depth was determined by the μCT measurements) as failure to engage the relatively strong surface of the bone specimens at right angles could result in bending of the needle and invalidation of the measurements. The samples (femoral heads) were rigidly supported and the penetration speed set to 1 mm/s. Following this, the experimental results were correlated to data obtained through the μCT measurements. To facilitate this, the crosssectional area of 8 circular regions of the subchondral trabecular bone (perpendicular to aforementioned areas and 6 mm in diameter) was measured at a 7 mm depth below the surface of the femoral head (see Fig. 2). During the osteopenetration experiments, the resulting force was recorded as a function of the needle penetration depth. The initial 4 mm of the needle displacement was disregarded, and considered as the engagement profile of the needle, which was pushed into the femoral head up to a depth of 14 mm. The 4 mm was determined, by μCT measurements, as the mean value of the cortical shell and thus not considered to avoid erroneous statements due to the cortical stiffness. The area beneath the osteopenetration curve, represents the penetration energy, expressed in Joules. The resulting values are considered as representative for bone density and strength in the same area and reflected the 4–14 mm penetration course of the needle in all specimens.
Fig. 1. Osteopenetration needle and engagement directions.
420
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
Fig. 2. Cross-sectional area of subchondral bone measured 7 mm below the femoral head surface and perpendicular to the medial-superior region. Fig. 4. The mean values of penetration energy in the 8 regions for samples with healthy BMD values and osteoarthritic ones.
Differences in the osteopenetration energy as well as in the 2D trabecular bone density among the 8 different sites were statistically investigated among the 2 donor groups but also gender specifically, using repeated measurements analysis of variance (ANOVA). 3. Results 3.1. Experimental results A characteristic load-displacement diagram of the osteopenetration experiments is demonstrated in Fig. 3. The saw-like curve observed in the diagram is due to the microstructure of the trabecular bone and the gradual destruction/collapse of trabeculae inside the femoral head. As expected, significant topographical variations of trabecular bone strength in different subchondral bone regions were documented, the mean values of osteopenetration energy are documented below for the control group followed in parentheses for osteoarthritic patients. In the medial region (fovea capitis femoris) the mean osteopenetration energy (in Joules) was 0.74 (0.84). In the inferior region the mean value was 0.84 (0.91), in the medial-superior region 2.27 (2.92), in superior region 3.08 (3.88), in anterior region 1.36 (1.42), in anterior-medial region 1.63 (1.99), in posterior-medial region 1.57 (1.76) and in posterior region 1.79 (2.32) J respectively. Fig. 4 shows the mean value of penetration energy calculated at all sites, for measurements conducted on femoral head samples with healthy BMD values as well as osteoarthritic ones. The μCT measurements conducted on the same samples are summarized in Fig. 5, for case vs. control samples. Medical imaging revealed that the 2D trabecular bone density (trabecular bone crossing a circle of
Fig. 3. Typical measurement curve of penetration tests.
6 mm in diameter) was the highest in the superior and medialsuperior regions in comparison with the fovea capitis femoris region. In the inferior region the bone strength was lower (increased, however when compared with the region of fovea capitis femoris which was the weakest region). Anterior and posterior regions (reg-5, 6, 7 and 8 respectively) had intermediate values, with the posterior region exhibiting a significantly greater mean value than that of the anterior region. 3.2. Statistical analysis ANOVA accounted for both, independent (case vs. control group and male vs. female donors) as well as paired (regional variations observed within specimens of the same donor) data. For this purpose, site-specific differences were treated as “within-subject factor” (studied at 8 sites of an individual donor), while donor group and gender were treated as “between-subjects factors” (studied at an inter-patient level). ANOVA showed that all factors (“site”, “donor group” and “gender”) have a significant effect on both dependent variables (osteopenetration
Fig. 5. Mean cross-sectional areas as registered for the 8 primary subchondral regions for case vs. control in femoral head samples (Y-axis starts at 16 mm2 to intensify the apparent deviations).
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
421
energy and 2D trabecular bone density). All p-values were lower than 0.01. In detail: ➢ the case group is characterized by a higher value of osteopenetration energy compared to the control group (by 0.35 J on average) and a higher value of 2D trabecular bone density compared to the control group (by 0.22 mm2 on average) ➢ male donors incur a higher value of osteopenetration energy compared to the female ones (by 0.10 J on average) and a higher value of 2D trabecular bone density compared to the female ones (by 0.19 mm2 on average) ➢ the interaction effect of donor group and gender was not significant for either one of the response variables (“osteopenetration energy” and “2D trabecular bone density”).
As far as the site of the femoral head is concerned, post hoc tests were conducted to examine specific differences between sites and site-specific ANOVA was employed to examine the effect of case vs. control group within each site. For the osteopenetration energy case (Fig. 4) all differences between sites are significant (p-values ≪ 0.01) and the difference of case vs. control group in all sites is also significant (p-values b 0.01). For the 2D trabecular bone density case (Fig. 5) all differences between sites are significant (p-values ≪ 0.01) and the difference of case vs. control group in most sites is also significant (p-values b 0.01 in sites 1, 3, 4, 6, 7, 8, while p-value = 0.357 in site 2 and p-value = 0.301 in site 5). The relative measurements of both, osteopenetration experiments and 2D trabecular bone density correlate well within each site and donor group. The respective correlation coefficients are summarized in Table 1 and they have all proven to be statistically significant (p-values b b 0.01). According to the calculated Pearson correlation values, osteopenetration energy has a positive linear relationship with 2D trabecular bone density in all sites and for both donor groups (although the exact expression of the fitted line may be different from site to site and for case or control group). This finding is very interesting since it enables the prediction of osteopenetration energy in any site of the femoral head of both osteoarthritic and non-osteoarthritic patients using the value of 2D trabecular bone density, which can easily be obtained through μCT measurements. To this end, a regression line needs to be fitted in the osteopenetration energy–2D trabecular bone density data for every site of the femoral head and every donor group. For example, Fig. 6 presents the regression line and 95% prediction interval of the osteopenetration energy in site 4 of the femoral head of osteoarthritic patients with respect to 2D trabecular bone density. Note that, although the effect of “gender” on osteopenetration energy was proven to be significant, it was not considered as an independent variable during the prediction of osteopenetration energy, since its impact is fully reflected in the 2D trabecular bone density measurements. That is, incorporating “gender” in a regression model which relates osteopenetration energy to 2D trabecular bone density, has, in most cases, nothing to offer. For example in the case of Fig. 6 (femoral head region 4 of osteoarthritic patients), adding “gender”
Fig. 6. Regression line and 95% prediction interval of the osteopenetration energy based on 2D trabecular bone density (in site 4).
in the regression model is pointless, since it does not increase the value of the adjusted R2. Following the principle presented in Fig. 6, the correlation of osteopenetration energy (Υ) to the 2D trabecular bone density (Χ) was determined as linear in all 8 regions. The site-specific regression coefficients β0 and β1 of this dependency (Υ = β0 + β1 × Χ) are summarized in Table 1 for all sites and for both groups (osteoarthritic and control). 4. Discussion Osteopenetration is useful in describing deviations of trabecular bone strength in subchondral bone regions, it was introduced in this study to examine topographical variations in subcapital trabecular bone of the femoral head, as it is capable of providing quasi-three-dimensional information on a combination of bone density and mechanical strength of bone, not readily obtained by other methods. The eight osteopenetration sites chosen for this study, represent regions subjected to different amounts of in vivo loading. The superior region being the most heavily loaded, posterior and anterior partially loaded and medial and inferior being the least loaded. The results are in agreement with loading distribution described in other studies (Thomas and Daniel, 1983; Hodge et al., 1986). The penetration with the 2.5 mm needle resulted in minimal damage to the trabecular structure and was mainly localized around the pathway of the needle (Hvid, 1985). This was ensured through a follow-up μCT evaluation of the femoral head samples demonstrating the absence of interference between the 8 consecutive indents. This facilitated fairly closely spaced measurements in regions with minimal distance between them, providing refined insight to the topographical variation of subchondral trabecular bone in the femoral head, as this kind of data could not be obtained in the past through other destructive tests (e.g. compression tests). A limitation however of the introduced technique is associated with the depth measurements considering the actual surface of the femoral head rather than global surface, thus not compensating for the depression at the fovea.
Table 1 Configuration of the correlation coefficients (CC) and regression coefficients β0, β1 of osteopenetration energy (dependent variable) against 2D trabecular bone density (independent variable), for all sites and both patient groups. Femoral head region
Osteoarthritic group
Control group
β0= β1= CC β0= β1= CC
1
2
3
4
5
6
7
8
−9.538 0.603 0.923 −8.468 0.546 0.933
−10.300 0.636 0.923 −10.990 0.673 0.865
−4.457 0.392 0.920 −8.621 0.594 0.941
−3.243 0.340 0.898 −11.110 0.683 0.897
−4.727 0.337 0.907 −2.968 0.236 0.912
−6.653 0.457 0.884 −5.916 0.417 0.826
−5.274 0.396 0.808 −20.670 1.292 0.911
−6.600 0.464 0.928 −10.600 0.651 0.919
422
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423
The results available in literature coincide with the ones obtained here. Brown and Ferguson (1980) and Martens et al. (1983) performed compression tests by loading in three directions: anterior–posterior, superior–inferior and medial–lateral. They showed that anisotropy was evident and an increase in stiffness was found in the regions traversed by the primary trabecular system. In their studies however, bone specimens were obtained from nonspecific regions of the entire femoral heads, whereas here eight different but well defined regions were examined. Sugita et al. (1999) examined the differences in anisotropy in osteoporotic bone in the primary compressive group of the femoral head. They found increased values of compressive stiffness in the parallel loading group compared to the perpendicular loading group, but the anisotropic behavior of cancellous bone was reduced, and the femoral head became isotropic as the bone density decreased (e.g. in osteoporosis). All previous findings correlate with both, osteopenetration experiments and 2D trabecular bone density measurements conducted within this study. The medial-superior region (3) and superior region (4) displayed the highest capacity to withstand the needle's penetration, whereas the fovea capitis femoris (region 1) exhibited the lowest osteopenetration energy. This correlates well to bone density in the orientation of the fovea capitis femoris, which was lower and exhibited a smaller spread of architectural variation than regions 3 and 4 (see Fig. 5). This led the needle to engage fewer trabecular struts during the penetration, thus rendering the fovea capitis femoris more susceptible to loading (mean osteopenetration energy values), whereas anterior and posterior regions reflect similar cross-sectional areas and penetration values. Grynpas et al. (1991) associated osteoarthritis with a thickening of the subchondral trabecular bone with abnormally low mineralization patterns, which was repeatedly documented in the past (Lajeunesse and Reboul, 2003). Based on these factors the overall compressive strength of the tissue was not expected to increase when compared to that of healthy subjects (Bailey et al., 2004). The sclerosis, however, of subchondral trabecular bone along with the formation osteophytes observed in other studies (Oettmeier and Abendroth, 1989) is indicative of the increased osteopenetration energy observed in this study. All sites of the osteoarthritic specimens exhibited an increased capacity to withstand the penetrating needle but the recorded augmentation of the medial-superior region (reg-3), the superior region (reg-4) and the posterior region (reg-8) was more pronounced when compared to the remaining penetration sites. This can be attributed to the increased turnover of the bone in the subcapital region leading to manifested remodeling in the high loaded regions of the femoral head (based on Wolf's principle) and has also been observed by other groups (Layton et al., 1988). The spread of architectural variation within each scanned region was noticeably higher in the osteoarthritic samples than in the controls (a behavior demonstrated in Fig. 5). A greater site variation of both bone density has also been documented by Li and Aspden (1997), who attributed this behavior to a 12% decreased mineral density of the subchondral bone when compared to that of healthy samples. Li and Aspden (1997) determined an increased stiffness of osteoarthritic specimens when compared to healthy ones which correlates well to the increased osteopenetration energy presented in this study for the case samples vs. control. 5. Conclusions In the present investigation, the effect of osteoarthritis on the regional anatomical variation of the subchondral trabecular bone in the femoral head, was examined by osteopenetration and compared to data obtained through non-invasive medical imaging techniques. The comparison of osteoarthritic to healthy samples was based on the primary hypothesis that the subchondral trabecular bone in the femoral head region is subjected to increased remodeling and demineralization.
This tendency was verified in all penetration directions and sustained non-invasively by μCT measurements conducted prior to experimentation. The 2D bone density seems to affect the required penetration force as an almost linear dependency to penetration energy was observed, indicating that micro-scale imaging of cancellous bone can be used as an indicator of both strength characteristics and topographical variation of subchondral trabecular bone in the femoral head. The results provide refined insight to femoral head load-bearing capacity that could among others improve the anchorage of patient specific hip resurfacing implants through proper component design and positioning. Acknowledgments The authors would like to acknowledge the Technical University of Serres, Greece for providing them access to their Laboratory's μCT device, and in particular Dr. D. Sagris and Prof. C. David. References An, Y., Draughn, R., 2000. Mechanical Testing of Bone and the Bone–implant Interface. CRC Press, pp. 241–246. Bailey, A.J., Mansell, J.P., Sims, T.J., Banse, X., 2004. Biochemical and mechanical properties of subchondral bone in osteoarthritis. Biorheology 41 (3–4), 349–358. Baroud, G., Falk, R., Crookshank, M., Sponagel, S., Steffen, T., 2004. Experimental and theoretical investigation of directional permeability of human vertebral cancellous bone for cement infiltration. J. Biomech. 37 (2), 189–196. Bevill, G., Keaveney, T.M., 2009. Trabecular bone strength predictions using finite element analysis of micro-scale images at limited spatial resolution. Bone 44, 579–584. Brown, T.D., Ferguson, A.B., 1980. Mechanical property distributions in the cancellous bone of the proximal femur. Acta Orthop. Scand. 51, 429–437. Canny, J., 1986. A computational approach to edge detection. IEEE Trans. Pattern Anal. 8 (6), 679–698. Cappozzo, A., 1991. The mechanics of human walking. Adv. Psychol. 78, 167–186. Cappozzo, A., 2009. The observation of human joint movement. 4th European Conference of the International Federation for Medical and Biological Engineering IFMBE Proceedings. 22, pp. 126–129. Carrothers, A.D., Gilbert, R.E., Jaiswal, A., Richardson, J.B., 2010. Birmingham hip resurfacing: the prevalence of failure. J. Bone Joint Surg. 10, 1344–1350 (British Volume). Chiba, K., Ito, M., Osaki, M., Uetani, M., Shindo, H., 2011. In vivo structural analysis of subchondral trabecular bone in osteoarthritis of the hip using multi-detector row CT. Osteoarthr. Cartil. 19, 180–185. Chiba, K., Burghardt, A.J., Osaki, M., Majumdar, S., 2013. Heterogeneity of bone microstructure in the femoral head in patients with osteoporosis: an ex vivo HR-pQCT study. Bone 56, 139–146. Cooper, C., Campion, G., Melton, L.J., 1992. Hip fractures in the elderly: a world-wide projection. Osteoporos. Int. 2, 285–289. Coulter, G., Young, D.A., Dalziel, R.E., Shimmin, A.J., 2012. Birmingham hip resurfacing at a mean of ten years: results from an independent centre. J. Bone Joint Surg. 94 (3), 315–321 (British Volume). Grynpas, M.D., Alpert, B., Katz, I., Lieberman, I., Pritzker, K.P.H., 1991. Subchondral bone in osteoarthritis. Calcif. Tissue Int. 49 (1), 20–26. Hayami, T., Pickarski, M., Wesolowski, G.A., McLane, J., Bone, A., Destefano, J., Rodan, G.A., Duong le, T., 2004. The role of subchondral bone remodeling in osteoarthritis: reduction of cartilage degeneration and prevention of osteophyte formation by alendronate in the rat anterior cruciate ligament transection model. Arthritis Rheum. 50 (4), 1193–1206. Hodge, W.H., Fijan, R.S., Carlson, K.L., Burgess, R.G., Harris, W.H., Mann, R.W., 1986. Contact pressures in the human hip joint measured in vivo. Proc. Natl. Acad. Sci. U. S. A. 83, 2879–2883. Hvid, I., 1985. Cancellous bone at the knee: a comparison of two methods of strength measurement. Arch. Orthop. Trauma Surg. 104, 211–217. Hvid, I., Andersen, K., Olesen, S., 1984. Cancellous bone strength measurements with the osteopenetrometer. Eng. Med. 13, 73–78. Lajeunesse, D., Reboul, P., 2003. Subchondral bone in osteoarthritis: a biologic link with articular cartilage leading to abnormal remodeling. Curr. Opin. Rheumatol. 15 (5), 628–633. Layton, M.W., Goldstein, S.A., Goulet, R.W., Feldkamp, L.A., Kubinski, D.J., Bole, G.G., 1988. Examination of subchondral bone architecture in experimental osteoarthritis by microscopic computed axial tomography. Arthritis Rheum. 31 (11), 1400–1405. Li, B., Aspden, R.M., 1997. Composition and mechanical properties of cancellous bone from the femoral head of patients with osteoporosis or osteoarthritis. J. Bone Miner. Res. 12 (4), 641–651. Lotz, J.C., Cheal, E.J., Hayes, W.C., 1995. Stress distributions within the proximal femur during gait and falls: implications for osteoporotic fracture. Osteoporos. Int. 5, 252–261. Magnussen, R.A., Guilak, F., Vail, T.P., 2005. Cartilage degeneration in post-collapse cases of osteonecrosis of the human femoral head: altered mechanical properties in tension, compression, and shear. J. Orthop. Res. 23, 576–583. Martens, M., Van Audekercke, R., De Meester, P., Mulier, J.C., 1983. The mechanical characteristics of cancellous bone at the upper femoral region. J. Biomech. 12, 971–983.
A. Tsouknidas et al. / Clinical Biomechanics 30 (2015) 418–423 Oettmeier, R., Abendroth, K., 1989. Osteoarthritis and bone: osteologic types of osteoarthritis of the hip. Skelet. Radiol. 8, 165–174. Rathnayaka, K., Sahama, T., Schuetz, M.A., Schmutz, B., 2010. Effects of CT image segmentation methods on the accuracy of long bone 3D reconstructions. Med. Eng. Phys. 33 (2), 226–233. Rockwood, P.R., Horne, J.G., Cryer, C., 1990. Hip fractures: a future epidemic? J. Orthop. Trauma 4, 388–396. Sneppen, O., Christensen, P., Larsen, H., Vang, P.S., 1981. Mechanical testing of trabecular bone in knee replacement. Int. Orthop. 5, 251–256.
423
Sugita, H., Oka, M., Toguchida, J., Nakamura, T., Ueo, T., Hayami, T., 1999. Anisotropy of osteoporotic cancellous bone. Bone 24, 513–516. Thomas, D.B., Daniel, T.S., 1983. In vitro contact stress distributions in the natural human hip. J. Biomech. 16, 373–384. Tsouknidas, A., Maliaris, G., Savvakis, S., Michailidis, N., 2015. Anisotropic post-yield response of cancellous bone simulated by stress–strain curves of bulk equivalent structures. Comput. Methods Biomech. Biomed. Engin. 18 (8), 839–846.
