Journal of Biomechanics 47 (2014) 65–73
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Evaluation of a subject-specific finite-element model of the equine metacarpophalangeal joint under physiological load Simon M. Harrison a,b,n, R. Chris Whitton c, Chris E. Kawcak d, Susan M. Stover e, Marcus G. Pandy a a
Department of Mechanical Engineering, University of Melbourne, Australia CSIRO Computational Informatics, Private Bag 33, Clayton South, Victoria 3169, Australia c Equine Centre, Faculty of Veterinary Science, University of Melbourne, Australia d Gail Holmes Equine Orthopaedic Research Centre, Colorado State University, USA e JD Wheat Veterinary Orthopaedic Research Lab, University of California at Davis, USA b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 8 October 2013
The equine metacarpophalangeal (MCP) joint is frequently injured, especially by racehorses in training. Most injuries result from repetitive loading of the subchondral bone and articular cartilage rather than from acute events. The likelihood of injury is multi-factorial but the magnitude of mechanical loading and the number of loading cycles are believed to play an important role. Therefore, an important step in understanding injury is to determine the distribution of load across the articular surface during normal locomotion. A subject-specific finite-element model of the MCP joint was developed (including deformable cartilage, elastic ligaments, muscle forces and rigid representations of bone), evaluated against measurements obtained from cadaver experiments, and then loaded using data from gait experiments. The sensitivity of the model to force inputs, cartilage stiffness, and cartilage geometry was studied. The FE model predicted MCP joint torque and sesamoid bone flexion angles within 5% of experimental measurements. Muscle–tendon forces, joint loads and cartilage stresses all increased as locomotion speed increased from walking to trotting and finally cantering. Perturbations to muscle–tendon forces resulted in small changes in articular cartilage stresses, whereas variations in joint torque, cartilage geometry and stiffness produced much larger effects. Non-subject-specific cartilage geometry changed the magnitude and distribution of pressure and the von Mises stress markedly. The mean and peak cartilage stresses generally increased with an increase in cartilage stiffness. Areas of peak stress correlated qualitatively with sites of common injury, suggesting that further modelling work may elucidate the types of loading that precede joint injury and may assist in the development of techniques for injury mitigation. Crown Copyright & 2013 Published by Elsevier Ltd. All rights reserved.
Keywords: Equine locomotion Fetlock injury Musculoskeletal model Cartilage stress Contact pressure
1. Introduction The metacarpophalangeal (MCP) joint is the site of a large proportion of musculoskeletal injuries in racehorses (Bailey et al., 1999; Parkin et al., 2004). Osteochondral injuries occur predominantly in the palmarodistal aspect of the third metacarpal bone (MC3) and the dorsoproximal articular margin of the proximal phalanx (P1). Subchondral bone damage to the palmar aspect of the MC3 condyle is especially common, resulting in two types of injury: parasagittal fractures of the condyles and palmar osteochondral disease (Barr et al., 2009; Parkin et al., 2006). These injuries are considered fatigue injuries, the result of a high number of stress cycles applied to cartilage
n Corresponding author at: CSIRO Computational Informatics, Private Bag 33, Clayton South, Victoria 3169, Australia. Tel.: þ61 3 9545 8450; fax: þ61 3 9545 8080. E-mail address: [email protected] (S.M. Harrison).
and bone, rather than an acute mechanical event (Stepnik et al., 2004; Norrdin and Stover, 2006). Many factors contribute to fatigue injury, but the magnitude of the stress or strain to which tissues are subjected is critical (Rapillard et al., 2006). The MCP joint experiences the largest loads of the distal limb during locomotion (Merritt et al., 2008; Harrison et al., 2010). MCP joint hyperextension during locomotion results in the storage of elastic strain energy in the long flexor tendons and the suspensory apparatus (Biewener, 1998; Bobbert et al., 2007; Butcher et al., 2009; Harrison et al., 2010; McGuigan and Wilson, 2003; Witte et al., 2004), which is subsequently utilised by the limb to increase the efficiency of locomotion (Butcher et al., 2009; Harrison et al. 2010). However, stretching of the flexor tendons imposes large forces on the MCP joint (Merritt et al., 2008; Harrison et al. 2010). While previous studies have reported on the magnitudes of the resultant forces transmitted by the MCP joint (Merritt et al., 2008; Harrison et al., 2010), the magnitudes and locations of maximum cartilage stresses are unknown.
0021-9290/$ - see front matter Crown Copyright & 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jbiomech.2013.10.001
66
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Fig. 1. Development of a subject-specific, rigid-body, musculoskeletal model and a subject-specific, deformable, finite-element joint model. (A) Bone, muscle and cartilage geometries obtained from a cadaveric forelimb were imaged using MRI and CT; (B) tendon and ligament strains, bone kinematics, and hoof forces were determined from in vitro mechanical experiments performed on a cadaveric forelimb; (C) gait analysis experiments were performed to measure joint kinematics and hoof loading during walking, trotting and cantering; (D) bone and cartilage geometries and muscle paths were determined using image segmentation; (E) outputs from the mechanical loading experiments, MR imaging, and gait analysis experiments were used to develop a subject-specific rigid-body musculoskeletal model of the distal forelimb; (F) inverse kinematics and inverse dynamics methods were used to determine the joint angles and net moments exerted about the joints in the model; (G) muscle–tendon and ligament forces were calculated based on the subject-specific rigid-body musculoskeletal model; (H) FE meshes were created from bone and cartilage geometries; muscle–tendon and ligament forces obtained from the subject-specific rigid-body musculoskeletal model (G) were input into a subject-specific deformable finite-element (FE) model of the MCP joint; and (I) subject-specific FE model of the MCP joint showing the major tendon/ligament structures and the three bone-cartilage articulations. (a) 3D imaging, (b) cadaver experiments, (c) galt experiments, (d) segmentation, (e) subject specific dynamic musculoskeletal model, (f) inverse dynamics & inverse kinematics, (g) muscle force calculations, (h) mesh creation and (i) FE modelling.
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Table 1 Muscles and ligaments incorporated in the FE model of the MCP joint. Name
Abbreviation Muscle (M) or ligament (L)
Specified force (S) or elastic spring (E)
Deep digital flexor Superficial digital flexor Common digital extensor Lateral digital extensor Interosseous ligament Medial extensor branch of the interosseous ligament Lateral extensor branch of the interosseous ligament Straight sesamoidean ligament Medial oblique sesamoidean ligament Lateral oblique sesamoidean ligament Medial short straight sesamoidean ligament Lateral short straight sesamoidean ligament Medial cruciate ligament Lateral cruciate ligament
DDF SDF CDE LDE IM EBM
M M M M L L
S S S S E E
EBL
L
E
SSL
L
S
OSLM
L
E
OSLL
L
E
SSSLM
L
E
SSSLL
L
E
MCL LCL
L L
E E
Finite-element (FE) modelling has been used to determine the distributions of contact force, cartilage pressure, and bone stress across the three-dimensional geometry of human joints (Anderson et al., 2007, 2008, 2010; Besier et al., 2008; Heino Brechter and Powers, 2002; Fernandez and Pandy, 2006; Pustoc'h and Cheze, 2009; Fernandez et al., 2011; Akbarshahi et al., accepted), as this information is difficult to obtain by direct measurement in vivo. Once validated, an FE model can be used to determine the effects of variations in loading, geometry and material properties on joint contact stresses, without the need for additional experimentation. FE models are sensitive to input parameters such as the geometry, kinematics and mechanical properties of bone and cartilage (Besier et al., 2008; Anderson et al., 2010). Subject-specific FE analysis is often considered the gold standard, but routine application of this methodology in a clinical environment is generally impractical due to the technical expertise and time required for model development. The use of simplified features such as a generic anatomical geometry would be favoured, provided there is an acceptable loss of model fidelity. A validated FE model of the equine forelimb would be a useful tool to investigate dynamic loading during normal locomotion and gain insight into the causes of joint injury. In the present study a subject-specific, three-dimensional FE model of the equine MCP joint was developed and used to determine joint contact loading and cartilage stresses for walking, trotting and cantering. The specific aims were firstly, to evaluate calculations of bone kinematics and tendon and ligament strains derived from the subject-specific FE model against measurements of the same quantities obtained from cadaver experiments performed on the same animal; secondly, to characterise the influence of gait speed on the magnitudes of joint contact forces, mean contact pressures, and locations of maximum cartilage stress; and finally, to determine the sensitivity of the FE model predictions to variations in the model inputs and parameters.
2. Methods One thoroughbred horse (mass, 500 kg; age, 5 years) free of lameness was used as a subject for this study. All protocols were approved by the Institutional Animal Use and Care Committee at Colorado State University.
67
Joint angle trajectories and ground reaction forces for the left forelimb were measured during walking (1.4 m/s), trotting (3.6 m/s), and cantering (7.5 m/s) on a high-speed treadmill. Details of these experiments are given by Harrison et al. (2012). The horse was then euthanized with an overdose of pentobarbital sodium (88 mg/kg, IV), and the left forelimb dissected free of the thorax and subsequently frozen. The whole limb was thawed and then imaged using computed tomography (CT) (Gemini TF Big Bore 16-slice scanner, Phillips Medical Systems, Nederland, BV) at 130 kV and 150 mA, with an 18 cm field of view (FOV), 512 512 matrix, and 1 mm slice thickness. Magnetic resonance imaging (MRI) of the whole limb was performed at 1.5 T using a proton density (PD) sequence with a 29 cm FOV, 256 256 matrix, 3 mm slice thickness and 0.3 mm spacing. MRI of the MCP complex (mid-MC3 to mid-P1) was performed at 3.0 T (Siemens Medical Solutions, Malvern, PA, USA) using a fat-saturated PD sequence to obtain 0.6 mm 0.6 mm 0.6 mm voxels. Bone and cartilage geometries were segmented using Amira (Mercury Computer Systems, Chelmsford, MA, USA) and smoothed using Geomagic (Research Triangle Park, NC, USA) to reduce artefacts from segmentation. A subject-specific rigid-body musculoskeletal model was developed in OpenSim (Delp et al., 2007) using the bone geometries and joint articulations created from CT and the muscle–tendon paths reconstructed from MRI (see Fig. 1). The distal limb skeleton (MC3 to hoof) was represented as a five-segment, four-degreeof-freedom, articulated linkage. Four muscles actuated the distal limb while 10 ligaments supported the MCP joint (Table 1). The mechanical properties of the tendons and ligaments (rest lengths and force– extension curves) and the relationships between tendon/ligament strains and limb pose were determined from mechanical loading experiments performed on the cadaver forelimb. The hoof was potted into a custom mould and fixed on a 6-axis load cell. The mid-humerus was potted into a mould that was pin-jointed to the crosshead of a materials testing machine (MTS Systems Corp., Minneapolis, MI, USA). The limb was loaded to produce a fetlock angle equivalent to the maximum joint angle achieved during a canter (an angle of 1201 between the dorsal aspect of P1 and MC3). The carpus remained in extension during loading. A three-camera video motion capture system (Eagle 4 system, Motion Analysis Corporation, Santa Rosa, CA) was used to measure bone kinematics. Linear variable displacement transducers (LVDTs) were systematically applied to each major tendon and ligament near to the MCP joint. A series of cutting experiments was performed to determine the force– extension properties of the soft tissues. First, the intact limb was loaded and bone positions, ground reaction force and tendon strains were measured. The most superficial tendon/ligament (e.g., superficial digital flexor, SDF) then was cut, and the loading experiment repeated. This procedure was repeated for all the major soft tissue structures supporting the limb. Joint torques were calculated for each experiment from measurements of the bone positions and the GRF. Finally, the force–extension curve for each tendon/ligament was obtained by dividing the calculated value of joint torque by the moment arm of the tendon/ligament, which was estimated using the subject-specific rigid-body musculoskeletal model. Net joint torques and tendon forces during locomotion were calculated using the subject-specific rigid-body musculoskeletal model (Fig. 1G). Joint angles were determined from marker positions using an inverse-kinematics function available in OpenSim. Joint torques were determined using an inverse-dynamics function in OpenSim, which incorporated the measured ground reaction forces and joint motions as inputs to the model (see Harrison et al. (2012) for details). The forces developed by the small extensor muscles (common digital extensor, CDE, and lateral digital extensor, LDE) were found using an EMG-driven muscle-tendon model (Lloyd and Besier, 2003) and trial-specific muscle EMG data obtained from the gait experiments (Harrison et al., 2012). Once the ligament, CDE and LDE forces were known, the forces transmitted by the SDF and deep digital flexor (DDF) tendons were found by solving a statically determinate mechanics problem (Meershoek et al., 2001). The time histories of the joint angles, joint torques and tendon forces were normalised to the duration of the stance phase. A subject-specific FE deformable model of the MCP joint (Fig. 1I) was subsequently developed using Abaqus (Simulia, Providence, RI, USA). Solid cartilage and shell bone meshes were created from the CAD geometries (Fig. 1H). The model was comprised of four bones – MC3, P1, and the proximal sesamoid bones (Ses) – with each bone represented using three-node, non-deformable (rigid) shell elements that were tied to a single reference node. Articular cartilage was modelled using solid continuum (fournode tetrahedral) elements, and the internal (subchondral) nodes were all attached to the reference node for the appropriate bone. The sesamoid bones were tied to the fibrocartilaginous intersesamoidean ligament, which was also represented by solid continuum (four-node tetrahedral) elements. Tendons were represented by wire elements and muscle forces were applied as purely contractile loads. Ligament forces were modelled using wire elements with a non-linear force–extension relationship. All tendon/ligament origins and insertions were attached to the reference nodes of the bones from which they originated and onto which they were inserted. Cartilage was modelled as an orthotropic material (Wang et al., 2003), and the elastic material parameters were based on a first-order fit between the FE model predictions of bone displacements and the results obtained from the cadaver loading experiments described above. Specifically, maximum compression of the cartilage (under full loading) was estimated using two motion capture markers, one placed at the MCP joint centre and the other at the P1 joint centre. The relative displacement of these markers, measured in 3D by the three-camera video motion capture system, was assumed to represent the total axial compression of the
68
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
articular cartilage for both the MC3 and P1 bones. The orthotropic moduli (E1–3) were modified until the FE predictions of cartilage compression were sufficiently close to the measured displacements (see Results section below). Values of E1–3 and Poisson's ratio are given in Supplementary material. The FE deformable joint model was tested for numerical convergence, then evaluated against the results of the cadaver experiments, and finally used to simulate the mechanical environment of the MCP joint during locomotion. Model outputs included the net forces acting on the medial and lateral aspects of the P1– MC3 articulation and between the Ses and MC3; the mean contact pressure in each of these regions; and the peak contact pressure and peak octahedral shear stress (represented by the von Mises (VM) stress) in the cartilage of each articulation. The sensitivity of the FE model calculations to changes in the geometric, kinetic and elastic inputs was also studied. Perturbations were made to the inputs of the nominal cantering simulation, and the results of each perturbation simulation were expressed as a percentage change from the results of the nominal cantering simulation. First, the centre of pressure (CoP) at the hoof was varied by 1 cm in the dorsal direction (CoPþ 1 cm) and muscle forces were re-calculated. Second, keeping the MCP joint torque constant, the following tendon forces were individually varied: (1) LDE and CDE forces were reduced by 50% (DE 50%); (2) forces in the medial and lateral extensor branches of the interosseus ligament (EBM and EBL) were increased by 25% (EB 125%); and (3) the stiffness of each element comprising the major components of the suspensory apparatus (SA), specifically, interosseus muscle(IM), medial oblique sesamoid ligament (OSLM), lateral oblique sesamoid ligament (OSLL) and straight sesamoid ligament (SSL), was increased by 20% (SA 120%). Third, a non-subject-
specific cartilage geometry (i.e., one acquired from a different animal and henceforth referred to as Generic Cartilage) was used in the model to study the effect of subjectspecific modelling of cartilage geometry on estimates of cartilage pressure and stress. Fourth, a uniform cartilage thickness geometry (henceforth referred to as Uniform Cartilage) was constructed by calculating the mean thickness of the original cartilage and offsetting the subchondral bone surface by this distance using geometry manipulation software (Geomagic Morrisville, NC, USA). A cartilage volume was constructed by a lofting process using the original subchondral bone surface and the offset surface. Finally, the stiffness values of cartilage were decreased by 50% (E1 50% and E2 50%) and increased by 50% (E1 150% and E2 150%) to investigate the sensitivity of the model calculations to changes in cartilage tissue properties. Based on the results of a similar study reported for humans (Segal et al., 2009), we assumed that a change of less than 20% in an output variable between the nominal and perturbed simulations represents an acceptable level of model sensitivity from an FE analysis perspective.
3. Results Mesh convergence for the FE model was achieved using tetrahedral elements with characteristic sizes of 0.4 mm, 0.5 mm and 0.8 mm to represent the volumes of articular cartilage associated with the Ses, P1 and MC3, respectively. Thus, the P1–MC3 and
Fig. 2. Comparison of the FE model predictions to data obtained from the cadaver experiments. (A) Torque exerted at the MCP joint; (B) flexion-extension angle of the sesamoid bones relative to a position 901 from the longitudinal axis of the MC3 bone; (C) strain in the medial oblique sesamoidean ligament (OSLM); (D) strain in the lateral oblique sesamoidean ligament (OSLL); and (E) strain in the tendon of the interosseous muscle (IM). Root mean square errors (RMSE) between model and experimental data are indicated at the top of each panel. (a) MCP torque, (b) Ses flexion, (c) OSLM strain, (d) OSLL strain and (e) IM strain.
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Ses–MC3 articulations were modelled using 29,740 nodes and 32,742 nodes, respectively. Cartilage material parameters of E1 ¼33 MPa in the direction of contact and E2 ¼450 MPa in the plane normal to the direction of contact produced a maximum cartilage displacement within 3% of that measured in the cadaver loading experiments. Numerical testing and model calibration results are presented in Supplementary material. Overall, there was reasonable agreement between the FE model predictions and results obtained from the cadaver experiments. The model calculations of the net torque acting about the MCP joint and the angle of rotation of Ses about the centre of rotation of the palmar aspect of the fetlock joint were within 5% of the experimental results. Peak strains calculated for the OSLL and OSLM ligaments and the IM tendon were within 13% of the values measured using LVDTs in the in vitro experiments (Fig. 2). Large differences were observed in the tendon forces, joint contact forces and cartilage stresses calculated for the walk, trot and canter (Figs. 3–5). Net joint torque, joint contact force, cartilage pressure and VM stress all increased with locomotion speed (Tables 2 and 3). While the contact forces at the P1–MC3 and Ses–MC3 joints increased moderately from trotting to cantering (i.e., by 11–14%), there were larger increases in the peak contact pressure and peak VM stress (21–32%). Medial contact areas typically experienced larger mean pressures, peak pressures and peak VM stresses, though areas of high stress were also observed on the lateral side of all articulations (Tables 2 and 3, Figs. 4 and 5, Supplementary material). Palmar metacarpal cartilage contact pressure (Ses–MC3 articulation) and VM stress were highest in, and abaxial to, the parasagittal groove, though larger areas
69
of high stress were also observed on the lateral aspect. P1–MC3 cartilage contact pressure and VM stress were highest abaxial to the parasagittal groove, especially on the medial palmar aspect. A number of focal areas of high stress, including dorsal P1 and distal palmar MC3, coincided with the location of thicker cartilage, but other areas of high stress did not; for example, the proximal palmar MC3 joint (Figs. 4 and 5, Supplementary material). Regions of high contact stress and pressure were coincident with sites commonly associated with injury in racing horses (cf. Fig. 4a and b). Perturbations in muscle and ligament forces had little effect on the contact force, mean contact pressure and local stresses incurred at the MCP joint (i.e., variations ranged from 1% to 3%). By comparison, variations in cartilage geometry and cartilage stiffness had a substantial effect on the mean contact pressure and local stresses (Tables 2 and 3, Figs. 4 and 5, Supplementary material). Use of the generic cartilage model varied mean contact pressure by up to 28% and peak pressures and stresses by at least 58%. Locations of peak contact stress and pressure were different for the generic and subject-specific cartilage models. Peak stresses in the P1 cartilage moved to the dorsal aspect and for the MC3 cartilage were more axial for the generic cartilage model. Use of a uniform cartilage model had a less significant effect on the calculations of joint contact pressure and stress – mean contact pressures were no more than 10% lower while peak local pressures and stresses were no more than 27% lower. The locations of peak pressure and stress did not appear to change. Variation in cartilage stiffness modified the mean contact pressure by at least 23%; however, peak local stresses changed significantly (by up to 88%) in some locations and less substantially (less than 2%) in others.
Fig. 3. Muscle–tendon and ligament forces calculated using the subject-specific rigid-body musculoskeletal model and incorporated in the Ses–MC3 FE model. Forces in the superficial (SDF) and deep (DDF) digital flexor tendons, the medial (EBM) and lateral (EBL) extensor branches of the interosseous muscle, and the straight sesamoidean ligament (SSL) are shown for walking, trotting and cantering.
70
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Fig. 4. (a) Common metacarpal bone damage in racehorses. The left panel is a palmarodistal view of a third metacarpal bone (MC3) with a fracture that has propagated from the lateral parasagittal groove. The right panel is a similar view showing biaxial subchondral bone discoloration and collapse with surrounding cartilage fibrillation resulting from palmar osteochondral disease; (b) the Palmar view of the MC3 cartilage (view indicated in the top right panel by an arrow) showing calculations of contact pressure and the von Mises stress during midstance, as well as measurements of unloaded cartilage thickness obtained from MRI. Results for walking, trotting and cantering using the subject-specific cartilage meshes are compared to the results for cantering obtained using uniform thickness meshes (Canter, Uniform Mesh) and meshes from a different animal (Canter, Generic Mesh). (a) Photos of injury and (b) cartilage pressure, stress and thickness. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4. Discussion The model calculations showed that (a) muscle–tendon forces, joint loads and contact stresses at the equine MCP joint all increased as locomotion speed increased from walking to trotting and finally cantering; and (b) calculations of joint contact stress in this model are most sensitive to cartilage geometry and cartilage material properties. These findings provide additional insight into
the mechanical behaviour of the MCP joint and inform future modelling investigations. There are several limitations of this study that must be acknowledged. First, because the present analysis is based on an FE model of one animal, caution should be exercised in extrapolating the results to the general equine population. Nevertheless, the input model parameters, specifically, the joint angular displacements, muscle–tendon forces, and joint contact forces are
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
71
Fig. 5. Proximal view of the P1 cartilage (view indicated in the top right panel by an arrow) showing calculations of contact pressure and the von Mises stress during midstance, as well as measurements of unloaded cartilage thickness obtained from MRI. Results for walking, trotting, and cantering using the subject-specific cartilage meshes are compared to the results for cantering obtained using uniform thickness meshes (Canter, Uniform Mesh) and meshes from a different animal (Canter, Generic Mesh). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 2 Net joint torque, joint contact force, mean contact pressures (medial and lateral), peak contact pressure and peak von Mises stress in the P1–MC3 joint calculated for walking, trotting and cantering at midstance. Perturbations to the model parameters assumed in the analysis for cantering are also presented to quantify the sensitivity of the model calculations to changes in the magnitudes of the input parameters. Percentage variations for walking and trotting relative to cantering are indicated in parentheses. Variations greater than 20% relative to cantering are indicated in bold. Load case
Net joint torque (Nm)
Walking 215 ( 72%) Trotting 658 ( 15%) Cantering 777 Cantering perturbations Uniform 777 (0%) cartilage Generic 777 (0%) cartilage COPþ1 cm 829 (þ 7%) EB 125% 777 (0%) DE 50% 777 (0%) SA 120% 777 (0%) E1, E2 50% 777 (0%) E1, E2 150% 777 (0%)
Joint contact force (kN)
Mean medial contact pressure Mean lateral contact (MPa) pressure (MPa)
Peak contact pressure (MPa)
Peak von Mises stress (MPa)
7.9 ( 70%) 23.2 ( 11%) 26.2
29 ( 39%) 48 ( þ1%) 47
20 ( 40%) 31 ( 6%) 33
110 ( 57%) 177 ( 30%) 254
159 ( 62%) 283 ( 32%) 418
26.1 (0%)
47 ( 0%)
33 ( 1%)
252 ( 1%)
418 (0%)
25.6 ( 2%)
34 ( 26%)
30 ( 8%)
106 ( 58%)
181 ( 57%)
27.5 ( þ 5%) 26.5 ( þ 1%) 26.1 (0%) 26.2 (0%) 26.0 ( 1%) 26.3 ( þ 1%)
50 48 48 47 36 58
32 33 33 33 21 43
231 ( 9%) 255 (0%) 256 (þ 1%) 247 ( 3%) 476 ( þ88%) 275 ( þ 5%)
476 ( þ 14%) 419 (0%) 423 ( þ1%) 405 ( 3%) 409 ( 2%) 443 ( þ 6%)
(þ 5%) (0%) ( þ1%) ( 1%) ( 25%) (þ 23%)
( 3%) (0%) ( 1%) ( 1%) ( 36%) ( þ31%)
72
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Table 3 Contact force, mean contact pressure (medial and lateral), peak contact pressure and peak von Mises stress in the Ses–MC3 joints for walking, trotting and cantering at midstance. Perturbations to the model parameters assumed in the analysis for cantering are also presented to quantify the sensitivity of the model calculations to changes in the magnitudes of the input parameters. Percentage variations for walking and trotting relative to cantering are indicated in parentheses. Variations greater than 20% relative to cantering are indicated in bold. Load case
Joint contact force (kN)
Walking 6.5 ( 75%) Trotting 22.6 ( 14%) Cantering 26.4 Cantering perturbations Uniform 26.9 ( þ 2%) cartilage Generic 28.0 ( þ 6%) cartilage COPþ1 cm 27.7 ( þ 5%) EB 125% 26.4 ( þ 0%) DE 50% 27.1 (þ 3%) SA 120% 27.5 ( þ 4%) E1, E2 50% 23.9 ( 9%) E1, E2 150% 27.2 ( þ 3%)
Mean medial contact pressure (MPa)
Mean lateral contact pressure (MPa)
Peak contact pressure (MPa)
Peak von Mises stress (MPa)
32 ( 48%) 54 ( 10%) 61
31 ( 29%) 41 ( 6%) 44
99 ( 68%) 243 ( 21%) 309
163 ( 74%) 458 ( 26%) 620
60 ( 1%)
40 ( 10%)
224 ( 27%)
492 ( 21%)
48 ( 21%)
32 ( 28%)
582 ( þ88%)
246 ( 60%)
63 ( þ3%) 61 (0%) 61 ( þ 1%) 62 ( þ2%) 42 ( 31%) 74 ( þ 23%)
45 44 46 45 32 57
( þ 3%) (0%) ( þ 5%) ( þ 2%) ( 28%) ( þ29%)
381 ( þ 23%) 316 ( þ2%) 312 ( þ 1%) 345 ( þ11%) 151 ( 51%) 278 ( 10%)
711 ( þ 15%) 635 ( þ2%) 628 ( þ1%) 677 (þ 9%) 584 ( 6%) 491 ( 21%)
consistent with data reported in previous studies (Harrison et al., 2010; Merritt et al., 2008). Second, the model outputs of joint pressure and stress were not directly validated, although predictions of bone kinematics and ligament forces were quantitatively compared against measurements obtained from cadaver loading experiments. Rather than disrupt and damage the joint to obtain measurements of contact pressure and contact area (Brama et al., 2001; den Hartog et al., 2009; Easton and Kawcak, 2007), we instead chose to preserve the integrity of the joint and investigate the mechanical behaviour of the intact tendons and ligaments. The FE model predictions of joint torque, sesamoid flexion angle and strains of the major ligaments agreed well with experimental measures. Third, a coarse estimation of cartilage stiffness was based on joint compression measured in the cadaver experiments. The relative displacement of the bones in the axial direction was small (0.48 mm of compression) compared to the precision of the motion capture system ( 70.10 mm). Also, cartilage may behave differently in vivo and our results show that variations in cartilage stiffness can significantly affect the calculated values of contact pressure and VM stress. Fourth, the speed of gait tested (up to 7.5 m/ s for cantering) was lower than that typically observed during racing (16–18 m/s). The magnitudes and distribution of joint contact pressure are likely to be different at higher locomotion speeds (Brama et al., 2001). Finally, subchondral bone was treated as perfectly rigid in the FE model. In reality, the subchondral bone will deform under load and may distribute the cartilage contact pressure and VM stress in a manner different to that reported here. The FE model calculations of joint contact pressure and contact area are comparable to measurements reported previously by others. In vitro measurements of joint contact pressure obtained during simulated walking indicate a mean contact pressure of approximately 15 MPa within the P1–MC3 joint (Brama et al., 2001; Easton and Kawcak, 2007), which is lower than the 20– 29 MPa calculated in the present study. It should be noted, however, that active muscle contraction, which comprises approximately one-third of the MCP joint force during walking (Harrison et al., 2010), was not accounted for in the cadaver experiments reported in the literature. Whilst similar in magnitude, the contact pressures and contact areas predicted by the FE model (Figs. 4 and 5, Supplementary material) are more localised than those observed in the joints of the cadaver specimens, where contact areas were found to extend over large areas of the P1 cartilage, even for small articular loads (Brama et al., 2001; Easton and Kawcak, 2007). More concentrated regions of contact stress may have been calculated due to the lack of deformation of the subchondral bone in our
model and the small errors resulting from geometry segmentation and mesh generation. Notably, the thickness of cartilage (Figs. 4 and 5, Supplementary Material) was small compared to the voxel size of the MR images, and finer resolution images are recommended in future studies. Contact areas were found to increase with increasing load, a trend that is also consistent with the experiment (Brama et al., 2001; Easton and Kawcak 2007). The higher mean and peak pressures observed in the Ses–MC3 joint compared to the P1–MC3 joint must have been due to the smaller surface area associated with the former joint because the forces transmitted by these two joints were similar in magnitude. The distribution of pressure across the medial and lateral surfaces of the P1–MC3 articulation is not surprising as the ground reaction force during locomotion is oriented medially (Merkens et al., 1986, 1993), applying an adduction moment about the joint and inducing higher contact pressures on the medial side. By comparison, the distribution of pressure across the medial and lateral surfaces of the Ses–MC3 joint was more uniform, which is most likely due to the ground reaction force not contributing directly to load in this articulation (Harrison et al., 2010). The model-predicted distributions of joint contact pressure are consistent with the locations of joint surface injury commonly observed in the distal equine forelimb. In thoroughbred racehorses, osteochondral injuries are more common at the Ses–MC3 joint than the P1–MC3 joint (Barr et al., 2009). Greater severity of P1 pathology is observed medially compared with laterally in both racing and wild horses (Cantley et al., 1999; Kawcak and McIlwraith, 1994). Pathology of the palmar aspect of the MC3 at the metacarpo-sesamoideal articulation may involve both the medial and lateral condyles, but is reported to be more common medially (Barr et al., 2009; Trope et al., 2011). Conversely, metacarpal condylar fractures more commonly affect the lateral condyle (Parkin et al., 2006). Further study is needed using deformable models of bone to investigate the development of stress through the depth of the subchondral bone under this type of loading. Joint contact force was found to be most sensitive to changes in net joint torque caused by a change in gait, whereas measures of stress and pressure were highly sensitive to variations in cartilage geometry as well. Whilst the use of non-subject-specific cartilage geometry would substantially simplify FE modelling of equine joints, unfortunately this approach resulted in dramatic changes in joint contact pressure and VM cartilage stress, and therefore cannot be recommended. This study also demonstrated the importance of accurate quantification of cartilage stiffness. Because the model was relatively insensitive to the magnitude of muscle–tendon force, we
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
suggest that estimates of cartilage stress may be obtained without first calculating muscle–tendon forces on the basis of a subjectspecific rigid-body musculoskeletal model. Instead, bone kinematics and net joint torques measured from a gait analysis experiment could be used in conjunction with a generic rigid-body musculoskeletal model to determine individual muscle–tendon forces. Conflict of interest statement None of the authors have a conflict of interest with respect to the work reported here. Acknowledgements The authors thank Tanya Garcia-Nolen and the staff and volunteer students of the Gail Holmes Equine Orthopaedic Research Center for their help in the collection and analysis of the gait data. Katrina Easton is gratefully acknowledged for supplying expertise, assistance and equipment for the in vitro experiments. We thank Tomas Correa for his help with the initial FE simulations. A Gait-Extract toolbox (freely available from https://simtk.org/home/c3dtoolbox) was used to extract and process the raw kinematic marker, ground reaction force and muscle EMG data obtained for each trial into a format suitable for input to the musculoskeletal model. This work was supported by the Rural Industries Research and Development Corporation of the Australian Government, the Peter Jay Sharp Foundation, the Dan Lufkin Foundation, an Australian Research Council Discovery Grant (DP0772838), a VESKI Innovation Fellowship to M.G.P., and the Grayson-Jockey Club Research Foundation. Appendix. Supplementary material Supplementary material associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jbiomech. 2013.10.001. References Akbarshahi, M., Fernandez, J., Schache, A.G., Pandy, M.G. Subject-specific evaluation of patellofemoral joint biomechanics during function alactivity. Med. Eng. Phys., accepted. Anderson, D.D., Goldsworthy, J.K., Li, W., Rudert, M.J., Tochigi, Y., Brown, T.D., 2007. Physical validation of a patient-specific contact finite element model of the ankle. J. Biomech. 40, 1662–1669. Anderson, A.E., Ellis, B.J., Maas, S.A., Peters, C.L., Weiss, J.A., 2008. Validation of finite element predictions of cartilage contact pressure in the human hip joint. J. Biomech. Eng. 130, 051008–051010. Anderson, A.E., Ellis, B.J., Maas, S.A., Weiss, J.A., 2010. Effects of idealized joint geometry on finite element predictions of cartilage contact stresses in the hip. J. Biomech. 43, 1351–1357. Bailey, C.J., Reid, S.W.J., Hodgson, D.R., Rose, R.J., 1999. Impact of injuries and disease on a cohort of two- and three-year-old thoroughbreds in training. Veterinary Record 145, 487–493. Barr, E.D., Pinchbeck, G.L., Clegg, P.D., Boyde, A., Riggs, C.M., 2009. Post mortem evaluation of palmar osteochondral disease (traumatic osteochondrosis) of the metacarpo/metatarsophalangeal joint in thoroughbred racehorses. Equine Vet. J. 4, 366–371. Besier, T.F., Gold, G.E., Delp, S.L., Fredericson, M., Beaupré, G.S., 2008. The influence of femoral internal and external rotation on cartilage stresses within the patellofemoral joint. J. Orthop. Res. 26, 1627–1635. Biewener, A.A., 1998. Muscle-tendon stresses and elastic energy storage during locomotion in the horse. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 120, 73–87. Bobbert, M.F., Alvarez, C.B.G., van Weeren, P.R., Roepstorff, L., Weishaupt, M.A., 2007. Validation of vertical ground reaction forces on individual limbs calculated from kinematics of horse locomotion. J. Exp. Biol. 210, 1885–1896. Brama, P.A.J., Karssenberg, D., Barneveld, A., van Weeren, P.R., 2001. Contact areas and pressure distribution on the proximal articular surface of the proximal phalanx under sagittal plane loading. Equine Vet. J. 33, 26–32.
73
Butcher, M.T., Hermanson, J.W., Ducharme, N.G., Mitchell, L.M., Soderholm, L.V., Bertram, J.E.A., 2009. Contractile behavior of the forelimb digital flexors during steady-state locomotion in horses (Equus caballus): an initial test of muscle architectural hypotheses about in vivo function. Comp. Biochem. Physiol. – Part B: Biochem. Mol. Biol. 152, 100–114. Cantley, C.E., Firth, E.C., Delahunt, J.W., Pfeiffer, D.U., Thompson, K.G., 1999. Naturally occurring osteoarthritis in the metacarpophalangeal joints of wild horses. Equine Vet. J. 31, 73–81. Delp, S.L., Anderson, F.C., Arnold, A.S., Loan, P., Habib, A., John, C.T., Guendelman, E., Thelen, D.G., 2007. OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans. Biomed. Eng. 54, 1940–1950. Easton, K.L., Kawcak, C.E., 2007. Evaluation of increased subchondral bone density in areas of contact in the metacarpophalangeal joint during joint loading in horses. American Journal of Veterinary Research 68, 816–821. den Hartog, S.M., Back, W., Brommer, H., Van Weeren, P.R., 2009. In vitro evaluation of metacarpophalangeal joint loading during simulated walk. Equine Veterinary Journal 41, 214–217. Fernandez, J., Pandy, M.G., 2006. Integrating modeling and experiments to assess dynamic musculoskeletal function. Exp. Physiol. 91, 371–382. Fernandez, J.W., Akbarshahi, M., Crossley, K.M., Shelburne, K.B., Pandy, M.G., 2011. Model predictions of increased knee-joint loading in regions of thinner articular cartilage after patellar tendon adhesion. J. Orthop. Res. 29, 1168–1177. Harrison, S., Whitton, R.C., King, M., Hausler, K.K., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2012. Forelimb muscle activity during equine locomotion. J. Exp. Biol. 215, 2980–2991. Harrison, S., Whitton, R.C., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2010. Relationship between muscle forces, joint loading, and utilization of elastic strain energy in equine locomotion. J. Exp. Biol. 213, 3998–4009. Heino Brechter, J., Powers, C.M., 2002. Patellofemoral stress during walking in persons with and without patellofemoral pain. Med. Sci. Sports Exercise 34 (10), 1582–1593. Kawcak, C.E., McIlwraith, C.W., 1994. Proximodorsal first phalanx osteochondral chip fragmentation in 336 horses. Equine Vet. J. 26 (5), 392–396. Lloyd, D.G., Besier, T.F., 2003. An EMG-driven musculoskeletal model to estimate muscle forces and knee joint moments in vivo. Journal of biomechanics 36, 765–776. McGuigan, M.P., Wilson, A.M., 2003. The effect of gait and digital flexor muscle activation on limb compliance in the forelimb of the horse Equus caballus. J. Exp. Biol. 206, 1325–1336. Meershoek, L.S., Bogert, A.J.V.D., Schamhardt, H.C., 2001. Model formulation and determination of in vitro parameters of a noninvasive method to calculate flexor tendon forces in the equine forelimb. Am. J. Vet. Res. 62, 1585–1593. Merkens, H.W., Schamhardt, H.C., Hartman, W., Kersjes, A.W., 1986. Ground reaction force patterns of Dutch warmblood horses at normal walk. Equine Vet. J. 18, 207–214. Merkens, H.W., Schamhardt, H.C., Van Osch, G.J.V.M., Van Den Bogert, A.J., 1993. Ground reaction force patterns of Dutch warmblood horses at normal trot. Equine Vet. J. 25, 134–137. Merritt, J.S., Davies, H.M.S., Burvill, C., Pandy, M.G., 2008. Influence of muscle– tendon wrapping on calculations of joint reaction forces in the equine distal forelimb. J. Biomed. Biotechnol. 2008. (Article ID: 165730, 9pp.). Norrdin, R.W., Stover, S.M., 2006. Subchondral bone failure in overload arthrosis: a scanning electron microscopic study in horses. J. Musculoskelet. Neuronal Interact. 6 (3), 251–257. Parkin, T.D.H., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2004. Risk of fatal distal limb fractures among thoroughbreds involved in the five types of racing in the United Kingdom. Vet. Rec. 154, 493–497. Parkin, T.D., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2006. Catastrophic fracture of the lateral condyle of the third metacarpus/metatarsus in UK racehorses – fracture descriptions and pre-existing pathology. Vet. J. 171, 157–165. Pustoc'h, A., Cheze, L., 2009. Normal and osteoarthritic hip joint mechanical behaviour: a comparison study. Med. Biol. Eng. Comput. 47 (4), 375–383. Rapillard, L., Charlebois, M., Zysset, P.K., 2006. Compressive fatigue behavior of human vertebral trabecular bone. J. Biomech. 39, 2133–2139. Segal, N.A., Anderson, D.D., Iyer, K.S., Baker, J., Torner, J.C., Lynch, J.A., Felson, D.T., Lewis, C.E., Brown, T.D., 2009. Baseline articular contact stress levels predict incident symptomatic knee osteoarthritis development in the MOST cohort. J. Orthop. Res. 27, 1562–1568. Stepnik, M.W., Radtke, C.L., Scollay, M.C., Oshel, P.E., Albrecht, R.M., Santschi, E.M., Markel, M.D., Muir, P., 2004. Scanning electron microscopic examination of third metacarpal/third metatarsal bone failure surfaces in thoroughbred racehorses with condylar fracture. Vet. Surg. 33, 2–10. Trope, G.D., Anderson, G.A., Whitton, R.C., 2011. Patterns of scintigraphic uptake in the fetlock joint of thoroughbred racehorses and the effect of increased radiopharmaceutical uptake in the distal metacarpal/tarsal condyle on performance. Equine Vet. J. 43, 509–515. Wang, C.C.B., Chahine, N.O., Hung, C.T., Ateshian, G.A., 2003. Optical determination of anisotropic material properties of bovine articular cartilage in compression. J. Biomech. 36, 339–353. Witte, T.H., Knill, K., Wilson, A.M., 2004. Determination of peak vertical ground reaction force from duty factor in the horse (Equus caballus). J. Exp. Biol. 207, 3639–3648.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Evaluation of a subject-specific finite-element model of the equine metacarpophalangeal joint under physiological load Simon M. Harrison a,b,n, R. Chris Whitton c, Chris E. Kawcak d, Susan M. Stover e, Marcus G. Pandy a a
Department of Mechanical Engineering, University of Melbourne, Australia CSIRO Computational Informatics, Private Bag 33, Clayton South, Victoria 3169, Australia c Equine Centre, Faculty of Veterinary Science, University of Melbourne, Australia d Gail Holmes Equine Orthopaedic Research Centre, Colorado State University, USA e JD Wheat Veterinary Orthopaedic Research Lab, University of California at Davis, USA b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 8 October 2013
The equine metacarpophalangeal (MCP) joint is frequently injured, especially by racehorses in training. Most injuries result from repetitive loading of the subchondral bone and articular cartilage rather than from acute events. The likelihood of injury is multi-factorial but the magnitude of mechanical loading and the number of loading cycles are believed to play an important role. Therefore, an important step in understanding injury is to determine the distribution of load across the articular surface during normal locomotion. A subject-specific finite-element model of the MCP joint was developed (including deformable cartilage, elastic ligaments, muscle forces and rigid representations of bone), evaluated against measurements obtained from cadaver experiments, and then loaded using data from gait experiments. The sensitivity of the model to force inputs, cartilage stiffness, and cartilage geometry was studied. The FE model predicted MCP joint torque and sesamoid bone flexion angles within 5% of experimental measurements. Muscle–tendon forces, joint loads and cartilage stresses all increased as locomotion speed increased from walking to trotting and finally cantering. Perturbations to muscle–tendon forces resulted in small changes in articular cartilage stresses, whereas variations in joint torque, cartilage geometry and stiffness produced much larger effects. Non-subject-specific cartilage geometry changed the magnitude and distribution of pressure and the von Mises stress markedly. The mean and peak cartilage stresses generally increased with an increase in cartilage stiffness. Areas of peak stress correlated qualitatively with sites of common injury, suggesting that further modelling work may elucidate the types of loading that precede joint injury and may assist in the development of techniques for injury mitigation. Crown Copyright & 2013 Published by Elsevier Ltd. All rights reserved.
Keywords: Equine locomotion Fetlock injury Musculoskeletal model Cartilage stress Contact pressure
1. Introduction The metacarpophalangeal (MCP) joint is the site of a large proportion of musculoskeletal injuries in racehorses (Bailey et al., 1999; Parkin et al., 2004). Osteochondral injuries occur predominantly in the palmarodistal aspect of the third metacarpal bone (MC3) and the dorsoproximal articular margin of the proximal phalanx (P1). Subchondral bone damage to the palmar aspect of the MC3 condyle is especially common, resulting in two types of injury: parasagittal fractures of the condyles and palmar osteochondral disease (Barr et al., 2009; Parkin et al., 2006). These injuries are considered fatigue injuries, the result of a high number of stress cycles applied to cartilage
n Corresponding author at: CSIRO Computational Informatics, Private Bag 33, Clayton South, Victoria 3169, Australia. Tel.: þ61 3 9545 8450; fax: þ61 3 9545 8080. E-mail address: [email protected] (S.M. Harrison).
and bone, rather than an acute mechanical event (Stepnik et al., 2004; Norrdin and Stover, 2006). Many factors contribute to fatigue injury, but the magnitude of the stress or strain to which tissues are subjected is critical (Rapillard et al., 2006). The MCP joint experiences the largest loads of the distal limb during locomotion (Merritt et al., 2008; Harrison et al., 2010). MCP joint hyperextension during locomotion results in the storage of elastic strain energy in the long flexor tendons and the suspensory apparatus (Biewener, 1998; Bobbert et al., 2007; Butcher et al., 2009; Harrison et al., 2010; McGuigan and Wilson, 2003; Witte et al., 2004), which is subsequently utilised by the limb to increase the efficiency of locomotion (Butcher et al., 2009; Harrison et al. 2010). However, stretching of the flexor tendons imposes large forces on the MCP joint (Merritt et al., 2008; Harrison et al. 2010). While previous studies have reported on the magnitudes of the resultant forces transmitted by the MCP joint (Merritt et al., 2008; Harrison et al., 2010), the magnitudes and locations of maximum cartilage stresses are unknown.
0021-9290/$ - see front matter Crown Copyright & 2013 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jbiomech.2013.10.001
66
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Fig. 1. Development of a subject-specific, rigid-body, musculoskeletal model and a subject-specific, deformable, finite-element joint model. (A) Bone, muscle and cartilage geometries obtained from a cadaveric forelimb were imaged using MRI and CT; (B) tendon and ligament strains, bone kinematics, and hoof forces were determined from in vitro mechanical experiments performed on a cadaveric forelimb; (C) gait analysis experiments were performed to measure joint kinematics and hoof loading during walking, trotting and cantering; (D) bone and cartilage geometries and muscle paths were determined using image segmentation; (E) outputs from the mechanical loading experiments, MR imaging, and gait analysis experiments were used to develop a subject-specific rigid-body musculoskeletal model of the distal forelimb; (F) inverse kinematics and inverse dynamics methods were used to determine the joint angles and net moments exerted about the joints in the model; (G) muscle–tendon and ligament forces were calculated based on the subject-specific rigid-body musculoskeletal model; (H) FE meshes were created from bone and cartilage geometries; muscle–tendon and ligament forces obtained from the subject-specific rigid-body musculoskeletal model (G) were input into a subject-specific deformable finite-element (FE) model of the MCP joint; and (I) subject-specific FE model of the MCP joint showing the major tendon/ligament structures and the three bone-cartilage articulations. (a) 3D imaging, (b) cadaver experiments, (c) galt experiments, (d) segmentation, (e) subject specific dynamic musculoskeletal model, (f) inverse dynamics & inverse kinematics, (g) muscle force calculations, (h) mesh creation and (i) FE modelling.
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Table 1 Muscles and ligaments incorporated in the FE model of the MCP joint. Name
Abbreviation Muscle (M) or ligament (L)
Specified force (S) or elastic spring (E)
Deep digital flexor Superficial digital flexor Common digital extensor Lateral digital extensor Interosseous ligament Medial extensor branch of the interosseous ligament Lateral extensor branch of the interosseous ligament Straight sesamoidean ligament Medial oblique sesamoidean ligament Lateral oblique sesamoidean ligament Medial short straight sesamoidean ligament Lateral short straight sesamoidean ligament Medial cruciate ligament Lateral cruciate ligament
DDF SDF CDE LDE IM EBM
M M M M L L
S S S S E E
EBL
L
E
SSL
L
S
OSLM
L
E
OSLL
L
E
SSSLM
L
E
SSSLL
L
E
MCL LCL
L L
E E
Finite-element (FE) modelling has been used to determine the distributions of contact force, cartilage pressure, and bone stress across the three-dimensional geometry of human joints (Anderson et al., 2007, 2008, 2010; Besier et al., 2008; Heino Brechter and Powers, 2002; Fernandez and Pandy, 2006; Pustoc'h and Cheze, 2009; Fernandez et al., 2011; Akbarshahi et al., accepted), as this information is difficult to obtain by direct measurement in vivo. Once validated, an FE model can be used to determine the effects of variations in loading, geometry and material properties on joint contact stresses, without the need for additional experimentation. FE models are sensitive to input parameters such as the geometry, kinematics and mechanical properties of bone and cartilage (Besier et al., 2008; Anderson et al., 2010). Subject-specific FE analysis is often considered the gold standard, but routine application of this methodology in a clinical environment is generally impractical due to the technical expertise and time required for model development. The use of simplified features such as a generic anatomical geometry would be favoured, provided there is an acceptable loss of model fidelity. A validated FE model of the equine forelimb would be a useful tool to investigate dynamic loading during normal locomotion and gain insight into the causes of joint injury. In the present study a subject-specific, three-dimensional FE model of the equine MCP joint was developed and used to determine joint contact loading and cartilage stresses for walking, trotting and cantering. The specific aims were firstly, to evaluate calculations of bone kinematics and tendon and ligament strains derived from the subject-specific FE model against measurements of the same quantities obtained from cadaver experiments performed on the same animal; secondly, to characterise the influence of gait speed on the magnitudes of joint contact forces, mean contact pressures, and locations of maximum cartilage stress; and finally, to determine the sensitivity of the FE model predictions to variations in the model inputs and parameters.
2. Methods One thoroughbred horse (mass, 500 kg; age, 5 years) free of lameness was used as a subject for this study. All protocols were approved by the Institutional Animal Use and Care Committee at Colorado State University.
67
Joint angle trajectories and ground reaction forces for the left forelimb were measured during walking (1.4 m/s), trotting (3.6 m/s), and cantering (7.5 m/s) on a high-speed treadmill. Details of these experiments are given by Harrison et al. (2012). The horse was then euthanized with an overdose of pentobarbital sodium (88 mg/kg, IV), and the left forelimb dissected free of the thorax and subsequently frozen. The whole limb was thawed and then imaged using computed tomography (CT) (Gemini TF Big Bore 16-slice scanner, Phillips Medical Systems, Nederland, BV) at 130 kV and 150 mA, with an 18 cm field of view (FOV), 512 512 matrix, and 1 mm slice thickness. Magnetic resonance imaging (MRI) of the whole limb was performed at 1.5 T using a proton density (PD) sequence with a 29 cm FOV, 256 256 matrix, 3 mm slice thickness and 0.3 mm spacing. MRI of the MCP complex (mid-MC3 to mid-P1) was performed at 3.0 T (Siemens Medical Solutions, Malvern, PA, USA) using a fat-saturated PD sequence to obtain 0.6 mm 0.6 mm 0.6 mm voxels. Bone and cartilage geometries were segmented using Amira (Mercury Computer Systems, Chelmsford, MA, USA) and smoothed using Geomagic (Research Triangle Park, NC, USA) to reduce artefacts from segmentation. A subject-specific rigid-body musculoskeletal model was developed in OpenSim (Delp et al., 2007) using the bone geometries and joint articulations created from CT and the muscle–tendon paths reconstructed from MRI (see Fig. 1). The distal limb skeleton (MC3 to hoof) was represented as a five-segment, four-degreeof-freedom, articulated linkage. Four muscles actuated the distal limb while 10 ligaments supported the MCP joint (Table 1). The mechanical properties of the tendons and ligaments (rest lengths and force– extension curves) and the relationships between tendon/ligament strains and limb pose were determined from mechanical loading experiments performed on the cadaver forelimb. The hoof was potted into a custom mould and fixed on a 6-axis load cell. The mid-humerus was potted into a mould that was pin-jointed to the crosshead of a materials testing machine (MTS Systems Corp., Minneapolis, MI, USA). The limb was loaded to produce a fetlock angle equivalent to the maximum joint angle achieved during a canter (an angle of 1201 between the dorsal aspect of P1 and MC3). The carpus remained in extension during loading. A three-camera video motion capture system (Eagle 4 system, Motion Analysis Corporation, Santa Rosa, CA) was used to measure bone kinematics. Linear variable displacement transducers (LVDTs) were systematically applied to each major tendon and ligament near to the MCP joint. A series of cutting experiments was performed to determine the force– extension properties of the soft tissues. First, the intact limb was loaded and bone positions, ground reaction force and tendon strains were measured. The most superficial tendon/ligament (e.g., superficial digital flexor, SDF) then was cut, and the loading experiment repeated. This procedure was repeated for all the major soft tissue structures supporting the limb. Joint torques were calculated for each experiment from measurements of the bone positions and the GRF. Finally, the force–extension curve for each tendon/ligament was obtained by dividing the calculated value of joint torque by the moment arm of the tendon/ligament, which was estimated using the subject-specific rigid-body musculoskeletal model. Net joint torques and tendon forces during locomotion were calculated using the subject-specific rigid-body musculoskeletal model (Fig. 1G). Joint angles were determined from marker positions using an inverse-kinematics function available in OpenSim. Joint torques were determined using an inverse-dynamics function in OpenSim, which incorporated the measured ground reaction forces and joint motions as inputs to the model (see Harrison et al. (2012) for details). The forces developed by the small extensor muscles (common digital extensor, CDE, and lateral digital extensor, LDE) were found using an EMG-driven muscle-tendon model (Lloyd and Besier, 2003) and trial-specific muscle EMG data obtained from the gait experiments (Harrison et al., 2012). Once the ligament, CDE and LDE forces were known, the forces transmitted by the SDF and deep digital flexor (DDF) tendons were found by solving a statically determinate mechanics problem (Meershoek et al., 2001). The time histories of the joint angles, joint torques and tendon forces were normalised to the duration of the stance phase. A subject-specific FE deformable model of the MCP joint (Fig. 1I) was subsequently developed using Abaqus (Simulia, Providence, RI, USA). Solid cartilage and shell bone meshes were created from the CAD geometries (Fig. 1H). The model was comprised of four bones – MC3, P1, and the proximal sesamoid bones (Ses) – with each bone represented using three-node, non-deformable (rigid) shell elements that were tied to a single reference node. Articular cartilage was modelled using solid continuum (fournode tetrahedral) elements, and the internal (subchondral) nodes were all attached to the reference node for the appropriate bone. The sesamoid bones were tied to the fibrocartilaginous intersesamoidean ligament, which was also represented by solid continuum (four-node tetrahedral) elements. Tendons were represented by wire elements and muscle forces were applied as purely contractile loads. Ligament forces were modelled using wire elements with a non-linear force–extension relationship. All tendon/ligament origins and insertions were attached to the reference nodes of the bones from which they originated and onto which they were inserted. Cartilage was modelled as an orthotropic material (Wang et al., 2003), and the elastic material parameters were based on a first-order fit between the FE model predictions of bone displacements and the results obtained from the cadaver loading experiments described above. Specifically, maximum compression of the cartilage (under full loading) was estimated using two motion capture markers, one placed at the MCP joint centre and the other at the P1 joint centre. The relative displacement of these markers, measured in 3D by the three-camera video motion capture system, was assumed to represent the total axial compression of the
68
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
articular cartilage for both the MC3 and P1 bones. The orthotropic moduli (E1–3) were modified until the FE predictions of cartilage compression were sufficiently close to the measured displacements (see Results section below). Values of E1–3 and Poisson's ratio are given in Supplementary material. The FE deformable joint model was tested for numerical convergence, then evaluated against the results of the cadaver experiments, and finally used to simulate the mechanical environment of the MCP joint during locomotion. Model outputs included the net forces acting on the medial and lateral aspects of the P1– MC3 articulation and between the Ses and MC3; the mean contact pressure in each of these regions; and the peak contact pressure and peak octahedral shear stress (represented by the von Mises (VM) stress) in the cartilage of each articulation. The sensitivity of the FE model calculations to changes in the geometric, kinetic and elastic inputs was also studied. Perturbations were made to the inputs of the nominal cantering simulation, and the results of each perturbation simulation were expressed as a percentage change from the results of the nominal cantering simulation. First, the centre of pressure (CoP) at the hoof was varied by 1 cm in the dorsal direction (CoPþ 1 cm) and muscle forces were re-calculated. Second, keeping the MCP joint torque constant, the following tendon forces were individually varied: (1) LDE and CDE forces were reduced by 50% (DE 50%); (2) forces in the medial and lateral extensor branches of the interosseus ligament (EBM and EBL) were increased by 25% (EB 125%); and (3) the stiffness of each element comprising the major components of the suspensory apparatus (SA), specifically, interosseus muscle(IM), medial oblique sesamoid ligament (OSLM), lateral oblique sesamoid ligament (OSLL) and straight sesamoid ligament (SSL), was increased by 20% (SA 120%). Third, a non-subject-
specific cartilage geometry (i.e., one acquired from a different animal and henceforth referred to as Generic Cartilage) was used in the model to study the effect of subjectspecific modelling of cartilage geometry on estimates of cartilage pressure and stress. Fourth, a uniform cartilage thickness geometry (henceforth referred to as Uniform Cartilage) was constructed by calculating the mean thickness of the original cartilage and offsetting the subchondral bone surface by this distance using geometry manipulation software (Geomagic Morrisville, NC, USA). A cartilage volume was constructed by a lofting process using the original subchondral bone surface and the offset surface. Finally, the stiffness values of cartilage were decreased by 50% (E1 50% and E2 50%) and increased by 50% (E1 150% and E2 150%) to investigate the sensitivity of the model calculations to changes in cartilage tissue properties. Based on the results of a similar study reported for humans (Segal et al., 2009), we assumed that a change of less than 20% in an output variable between the nominal and perturbed simulations represents an acceptable level of model sensitivity from an FE analysis perspective.
3. Results Mesh convergence for the FE model was achieved using tetrahedral elements with characteristic sizes of 0.4 mm, 0.5 mm and 0.8 mm to represent the volumes of articular cartilage associated with the Ses, P1 and MC3, respectively. Thus, the P1–MC3 and
Fig. 2. Comparison of the FE model predictions to data obtained from the cadaver experiments. (A) Torque exerted at the MCP joint; (B) flexion-extension angle of the sesamoid bones relative to a position 901 from the longitudinal axis of the MC3 bone; (C) strain in the medial oblique sesamoidean ligament (OSLM); (D) strain in the lateral oblique sesamoidean ligament (OSLL); and (E) strain in the tendon of the interosseous muscle (IM). Root mean square errors (RMSE) between model and experimental data are indicated at the top of each panel. (a) MCP torque, (b) Ses flexion, (c) OSLM strain, (d) OSLL strain and (e) IM strain.
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Ses–MC3 articulations were modelled using 29,740 nodes and 32,742 nodes, respectively. Cartilage material parameters of E1 ¼33 MPa in the direction of contact and E2 ¼450 MPa in the plane normal to the direction of contact produced a maximum cartilage displacement within 3% of that measured in the cadaver loading experiments. Numerical testing and model calibration results are presented in Supplementary material. Overall, there was reasonable agreement between the FE model predictions and results obtained from the cadaver experiments. The model calculations of the net torque acting about the MCP joint and the angle of rotation of Ses about the centre of rotation of the palmar aspect of the fetlock joint were within 5% of the experimental results. Peak strains calculated for the OSLL and OSLM ligaments and the IM tendon were within 13% of the values measured using LVDTs in the in vitro experiments (Fig. 2). Large differences were observed in the tendon forces, joint contact forces and cartilage stresses calculated for the walk, trot and canter (Figs. 3–5). Net joint torque, joint contact force, cartilage pressure and VM stress all increased with locomotion speed (Tables 2 and 3). While the contact forces at the P1–MC3 and Ses–MC3 joints increased moderately from trotting to cantering (i.e., by 11–14%), there were larger increases in the peak contact pressure and peak VM stress (21–32%). Medial contact areas typically experienced larger mean pressures, peak pressures and peak VM stresses, though areas of high stress were also observed on the lateral side of all articulations (Tables 2 and 3, Figs. 4 and 5, Supplementary material). Palmar metacarpal cartilage contact pressure (Ses–MC3 articulation) and VM stress were highest in, and abaxial to, the parasagittal groove, though larger areas
69
of high stress were also observed on the lateral aspect. P1–MC3 cartilage contact pressure and VM stress were highest abaxial to the parasagittal groove, especially on the medial palmar aspect. A number of focal areas of high stress, including dorsal P1 and distal palmar MC3, coincided with the location of thicker cartilage, but other areas of high stress did not; for example, the proximal palmar MC3 joint (Figs. 4 and 5, Supplementary material). Regions of high contact stress and pressure were coincident with sites commonly associated with injury in racing horses (cf. Fig. 4a and b). Perturbations in muscle and ligament forces had little effect on the contact force, mean contact pressure and local stresses incurred at the MCP joint (i.e., variations ranged from 1% to 3%). By comparison, variations in cartilage geometry and cartilage stiffness had a substantial effect on the mean contact pressure and local stresses (Tables 2 and 3, Figs. 4 and 5, Supplementary material). Use of the generic cartilage model varied mean contact pressure by up to 28% and peak pressures and stresses by at least 58%. Locations of peak contact stress and pressure were different for the generic and subject-specific cartilage models. Peak stresses in the P1 cartilage moved to the dorsal aspect and for the MC3 cartilage were more axial for the generic cartilage model. Use of a uniform cartilage model had a less significant effect on the calculations of joint contact pressure and stress – mean contact pressures were no more than 10% lower while peak local pressures and stresses were no more than 27% lower. The locations of peak pressure and stress did not appear to change. Variation in cartilage stiffness modified the mean contact pressure by at least 23%; however, peak local stresses changed significantly (by up to 88%) in some locations and less substantially (less than 2%) in others.
Fig. 3. Muscle–tendon and ligament forces calculated using the subject-specific rigid-body musculoskeletal model and incorporated in the Ses–MC3 FE model. Forces in the superficial (SDF) and deep (DDF) digital flexor tendons, the medial (EBM) and lateral (EBL) extensor branches of the interosseous muscle, and the straight sesamoidean ligament (SSL) are shown for walking, trotting and cantering.
70
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Fig. 4. (a) Common metacarpal bone damage in racehorses. The left panel is a palmarodistal view of a third metacarpal bone (MC3) with a fracture that has propagated from the lateral parasagittal groove. The right panel is a similar view showing biaxial subchondral bone discoloration and collapse with surrounding cartilage fibrillation resulting from palmar osteochondral disease; (b) the Palmar view of the MC3 cartilage (view indicated in the top right panel by an arrow) showing calculations of contact pressure and the von Mises stress during midstance, as well as measurements of unloaded cartilage thickness obtained from MRI. Results for walking, trotting and cantering using the subject-specific cartilage meshes are compared to the results for cantering obtained using uniform thickness meshes (Canter, Uniform Mesh) and meshes from a different animal (Canter, Generic Mesh). (a) Photos of injury and (b) cartilage pressure, stress and thickness. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4. Discussion The model calculations showed that (a) muscle–tendon forces, joint loads and contact stresses at the equine MCP joint all increased as locomotion speed increased from walking to trotting and finally cantering; and (b) calculations of joint contact stress in this model are most sensitive to cartilage geometry and cartilage material properties. These findings provide additional insight into
the mechanical behaviour of the MCP joint and inform future modelling investigations. There are several limitations of this study that must be acknowledged. First, because the present analysis is based on an FE model of one animal, caution should be exercised in extrapolating the results to the general equine population. Nevertheless, the input model parameters, specifically, the joint angular displacements, muscle–tendon forces, and joint contact forces are
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
71
Fig. 5. Proximal view of the P1 cartilage (view indicated in the top right panel by an arrow) showing calculations of contact pressure and the von Mises stress during midstance, as well as measurements of unloaded cartilage thickness obtained from MRI. Results for walking, trotting, and cantering using the subject-specific cartilage meshes are compared to the results for cantering obtained using uniform thickness meshes (Canter, Uniform Mesh) and meshes from a different animal (Canter, Generic Mesh). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 2 Net joint torque, joint contact force, mean contact pressures (medial and lateral), peak contact pressure and peak von Mises stress in the P1–MC3 joint calculated for walking, trotting and cantering at midstance. Perturbations to the model parameters assumed in the analysis for cantering are also presented to quantify the sensitivity of the model calculations to changes in the magnitudes of the input parameters. Percentage variations for walking and trotting relative to cantering are indicated in parentheses. Variations greater than 20% relative to cantering are indicated in bold. Load case
Net joint torque (Nm)
Walking 215 ( 72%) Trotting 658 ( 15%) Cantering 777 Cantering perturbations Uniform 777 (0%) cartilage Generic 777 (0%) cartilage COPþ1 cm 829 (þ 7%) EB 125% 777 (0%) DE 50% 777 (0%) SA 120% 777 (0%) E1, E2 50% 777 (0%) E1, E2 150% 777 (0%)
Joint contact force (kN)
Mean medial contact pressure Mean lateral contact (MPa) pressure (MPa)
Peak contact pressure (MPa)
Peak von Mises stress (MPa)
7.9 ( 70%) 23.2 ( 11%) 26.2
29 ( 39%) 48 ( þ1%) 47
20 ( 40%) 31 ( 6%) 33
110 ( 57%) 177 ( 30%) 254
159 ( 62%) 283 ( 32%) 418
26.1 (0%)
47 ( 0%)
33 ( 1%)
252 ( 1%)
418 (0%)
25.6 ( 2%)
34 ( 26%)
30 ( 8%)
106 ( 58%)
181 ( 57%)
27.5 ( þ 5%) 26.5 ( þ 1%) 26.1 (0%) 26.2 (0%) 26.0 ( 1%) 26.3 ( þ 1%)
50 48 48 47 36 58
32 33 33 33 21 43
231 ( 9%) 255 (0%) 256 (þ 1%) 247 ( 3%) 476 ( þ88%) 275 ( þ 5%)
476 ( þ 14%) 419 (0%) 423 ( þ1%) 405 ( 3%) 409 ( 2%) 443 ( þ 6%)
(þ 5%) (0%) ( þ1%) ( 1%) ( 25%) (þ 23%)
( 3%) (0%) ( 1%) ( 1%) ( 36%) ( þ31%)
72
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
Table 3 Contact force, mean contact pressure (medial and lateral), peak contact pressure and peak von Mises stress in the Ses–MC3 joints for walking, trotting and cantering at midstance. Perturbations to the model parameters assumed in the analysis for cantering are also presented to quantify the sensitivity of the model calculations to changes in the magnitudes of the input parameters. Percentage variations for walking and trotting relative to cantering are indicated in parentheses. Variations greater than 20% relative to cantering are indicated in bold. Load case
Joint contact force (kN)
Walking 6.5 ( 75%) Trotting 22.6 ( 14%) Cantering 26.4 Cantering perturbations Uniform 26.9 ( þ 2%) cartilage Generic 28.0 ( þ 6%) cartilage COPþ1 cm 27.7 ( þ 5%) EB 125% 26.4 ( þ 0%) DE 50% 27.1 (þ 3%) SA 120% 27.5 ( þ 4%) E1, E2 50% 23.9 ( 9%) E1, E2 150% 27.2 ( þ 3%)
Mean medial contact pressure (MPa)
Mean lateral contact pressure (MPa)
Peak contact pressure (MPa)
Peak von Mises stress (MPa)
32 ( 48%) 54 ( 10%) 61
31 ( 29%) 41 ( 6%) 44
99 ( 68%) 243 ( 21%) 309
163 ( 74%) 458 ( 26%) 620
60 ( 1%)
40 ( 10%)
224 ( 27%)
492 ( 21%)
48 ( 21%)
32 ( 28%)
582 ( þ88%)
246 ( 60%)
63 ( þ3%) 61 (0%) 61 ( þ 1%) 62 ( þ2%) 42 ( 31%) 74 ( þ 23%)
45 44 46 45 32 57
( þ 3%) (0%) ( þ 5%) ( þ 2%) ( 28%) ( þ29%)
381 ( þ 23%) 316 ( þ2%) 312 ( þ 1%) 345 ( þ11%) 151 ( 51%) 278 ( 10%)
711 ( þ 15%) 635 ( þ2%) 628 ( þ1%) 677 (þ 9%) 584 ( 6%) 491 ( 21%)
consistent with data reported in previous studies (Harrison et al., 2010; Merritt et al., 2008). Second, the model outputs of joint pressure and stress were not directly validated, although predictions of bone kinematics and ligament forces were quantitatively compared against measurements obtained from cadaver loading experiments. Rather than disrupt and damage the joint to obtain measurements of contact pressure and contact area (Brama et al., 2001; den Hartog et al., 2009; Easton and Kawcak, 2007), we instead chose to preserve the integrity of the joint and investigate the mechanical behaviour of the intact tendons and ligaments. The FE model predictions of joint torque, sesamoid flexion angle and strains of the major ligaments agreed well with experimental measures. Third, a coarse estimation of cartilage stiffness was based on joint compression measured in the cadaver experiments. The relative displacement of the bones in the axial direction was small (0.48 mm of compression) compared to the precision of the motion capture system ( 70.10 mm). Also, cartilage may behave differently in vivo and our results show that variations in cartilage stiffness can significantly affect the calculated values of contact pressure and VM stress. Fourth, the speed of gait tested (up to 7.5 m/ s for cantering) was lower than that typically observed during racing (16–18 m/s). The magnitudes and distribution of joint contact pressure are likely to be different at higher locomotion speeds (Brama et al., 2001). Finally, subchondral bone was treated as perfectly rigid in the FE model. In reality, the subchondral bone will deform under load and may distribute the cartilage contact pressure and VM stress in a manner different to that reported here. The FE model calculations of joint contact pressure and contact area are comparable to measurements reported previously by others. In vitro measurements of joint contact pressure obtained during simulated walking indicate a mean contact pressure of approximately 15 MPa within the P1–MC3 joint (Brama et al., 2001; Easton and Kawcak, 2007), which is lower than the 20– 29 MPa calculated in the present study. It should be noted, however, that active muscle contraction, which comprises approximately one-third of the MCP joint force during walking (Harrison et al., 2010), was not accounted for in the cadaver experiments reported in the literature. Whilst similar in magnitude, the contact pressures and contact areas predicted by the FE model (Figs. 4 and 5, Supplementary material) are more localised than those observed in the joints of the cadaver specimens, where contact areas were found to extend over large areas of the P1 cartilage, even for small articular loads (Brama et al., 2001; Easton and Kawcak, 2007). More concentrated regions of contact stress may have been calculated due to the lack of deformation of the subchondral bone in our
model and the small errors resulting from geometry segmentation and mesh generation. Notably, the thickness of cartilage (Figs. 4 and 5, Supplementary Material) was small compared to the voxel size of the MR images, and finer resolution images are recommended in future studies. Contact areas were found to increase with increasing load, a trend that is also consistent with the experiment (Brama et al., 2001; Easton and Kawcak 2007). The higher mean and peak pressures observed in the Ses–MC3 joint compared to the P1–MC3 joint must have been due to the smaller surface area associated with the former joint because the forces transmitted by these two joints were similar in magnitude. The distribution of pressure across the medial and lateral surfaces of the P1–MC3 articulation is not surprising as the ground reaction force during locomotion is oriented medially (Merkens et al., 1986, 1993), applying an adduction moment about the joint and inducing higher contact pressures on the medial side. By comparison, the distribution of pressure across the medial and lateral surfaces of the Ses–MC3 joint was more uniform, which is most likely due to the ground reaction force not contributing directly to load in this articulation (Harrison et al., 2010). The model-predicted distributions of joint contact pressure are consistent with the locations of joint surface injury commonly observed in the distal equine forelimb. In thoroughbred racehorses, osteochondral injuries are more common at the Ses–MC3 joint than the P1–MC3 joint (Barr et al., 2009). Greater severity of P1 pathology is observed medially compared with laterally in both racing and wild horses (Cantley et al., 1999; Kawcak and McIlwraith, 1994). Pathology of the palmar aspect of the MC3 at the metacarpo-sesamoideal articulation may involve both the medial and lateral condyles, but is reported to be more common medially (Barr et al., 2009; Trope et al., 2011). Conversely, metacarpal condylar fractures more commonly affect the lateral condyle (Parkin et al., 2006). Further study is needed using deformable models of bone to investigate the development of stress through the depth of the subchondral bone under this type of loading. Joint contact force was found to be most sensitive to changes in net joint torque caused by a change in gait, whereas measures of stress and pressure were highly sensitive to variations in cartilage geometry as well. Whilst the use of non-subject-specific cartilage geometry would substantially simplify FE modelling of equine joints, unfortunately this approach resulted in dramatic changes in joint contact pressure and VM cartilage stress, and therefore cannot be recommended. This study also demonstrated the importance of accurate quantification of cartilage stiffness. Because the model was relatively insensitive to the magnitude of muscle–tendon force, we
S.M. Harrison et al. / Journal of Biomechanics 47 (2014) 65–73
suggest that estimates of cartilage stress may be obtained without first calculating muscle–tendon forces on the basis of a subjectspecific rigid-body musculoskeletal model. Instead, bone kinematics and net joint torques measured from a gait analysis experiment could be used in conjunction with a generic rigid-body musculoskeletal model to determine individual muscle–tendon forces. Conflict of interest statement None of the authors have a conflict of interest with respect to the work reported here. Acknowledgements The authors thank Tanya Garcia-Nolen and the staff and volunteer students of the Gail Holmes Equine Orthopaedic Research Center for their help in the collection and analysis of the gait data. Katrina Easton is gratefully acknowledged for supplying expertise, assistance and equipment for the in vitro experiments. We thank Tomas Correa for his help with the initial FE simulations. A Gait-Extract toolbox (freely available from https://simtk.org/home/c3dtoolbox) was used to extract and process the raw kinematic marker, ground reaction force and muscle EMG data obtained for each trial into a format suitable for input to the musculoskeletal model. This work was supported by the Rural Industries Research and Development Corporation of the Australian Government, the Peter Jay Sharp Foundation, the Dan Lufkin Foundation, an Australian Research Council Discovery Grant (DP0772838), a VESKI Innovation Fellowship to M.G.P., and the Grayson-Jockey Club Research Foundation. Appendix. Supplementary material Supplementary material associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jbiomech. 2013.10.001. References Akbarshahi, M., Fernandez, J., Schache, A.G., Pandy, M.G. Subject-specific evaluation of patellofemoral joint biomechanics during function alactivity. Med. Eng. Phys., accepted. Anderson, D.D., Goldsworthy, J.K., Li, W., Rudert, M.J., Tochigi, Y., Brown, T.D., 2007. Physical validation of a patient-specific contact finite element model of the ankle. J. Biomech. 40, 1662–1669. Anderson, A.E., Ellis, B.J., Maas, S.A., Peters, C.L., Weiss, J.A., 2008. Validation of finite element predictions of cartilage contact pressure in the human hip joint. J. Biomech. Eng. 130, 051008–051010. Anderson, A.E., Ellis, B.J., Maas, S.A., Weiss, J.A., 2010. Effects of idealized joint geometry on finite element predictions of cartilage contact stresses in the hip. J. Biomech. 43, 1351–1357. Bailey, C.J., Reid, S.W.J., Hodgson, D.R., Rose, R.J., 1999. Impact of injuries and disease on a cohort of two- and three-year-old thoroughbreds in training. Veterinary Record 145, 487–493. Barr, E.D., Pinchbeck, G.L., Clegg, P.D., Boyde, A., Riggs, C.M., 2009. Post mortem evaluation of palmar osteochondral disease (traumatic osteochondrosis) of the metacarpo/metatarsophalangeal joint in thoroughbred racehorses. Equine Vet. J. 4, 366–371. Besier, T.F., Gold, G.E., Delp, S.L., Fredericson, M., Beaupré, G.S., 2008. The influence of femoral internal and external rotation on cartilage stresses within the patellofemoral joint. J. Orthop. Res. 26, 1627–1635. Biewener, A.A., 1998. Muscle-tendon stresses and elastic energy storage during locomotion in the horse. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 120, 73–87. Bobbert, M.F., Alvarez, C.B.G., van Weeren, P.R., Roepstorff, L., Weishaupt, M.A., 2007. Validation of vertical ground reaction forces on individual limbs calculated from kinematics of horse locomotion. J. Exp. Biol. 210, 1885–1896. Brama, P.A.J., Karssenberg, D., Barneveld, A., van Weeren, P.R., 2001. Contact areas and pressure distribution on the proximal articular surface of the proximal phalanx under sagittal plane loading. Equine Vet. J. 33, 26–32.
73
Butcher, M.T., Hermanson, J.W., Ducharme, N.G., Mitchell, L.M., Soderholm, L.V., Bertram, J.E.A., 2009. Contractile behavior of the forelimb digital flexors during steady-state locomotion in horses (Equus caballus): an initial test of muscle architectural hypotheses about in vivo function. Comp. Biochem. Physiol. – Part B: Biochem. Mol. Biol. 152, 100–114. Cantley, C.E., Firth, E.C., Delahunt, J.W., Pfeiffer, D.U., Thompson, K.G., 1999. Naturally occurring osteoarthritis in the metacarpophalangeal joints of wild horses. Equine Vet. J. 31, 73–81. Delp, S.L., Anderson, F.C., Arnold, A.S., Loan, P., Habib, A., John, C.T., Guendelman, E., Thelen, D.G., 2007. OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans. Biomed. Eng. 54, 1940–1950. Easton, K.L., Kawcak, C.E., 2007. Evaluation of increased subchondral bone density in areas of contact in the metacarpophalangeal joint during joint loading in horses. American Journal of Veterinary Research 68, 816–821. den Hartog, S.M., Back, W., Brommer, H., Van Weeren, P.R., 2009. In vitro evaluation of metacarpophalangeal joint loading during simulated walk. Equine Veterinary Journal 41, 214–217. Fernandez, J., Pandy, M.G., 2006. Integrating modeling and experiments to assess dynamic musculoskeletal function. Exp. Physiol. 91, 371–382. Fernandez, J.W., Akbarshahi, M., Crossley, K.M., Shelburne, K.B., Pandy, M.G., 2011. Model predictions of increased knee-joint loading in regions of thinner articular cartilage after patellar tendon adhesion. J. Orthop. Res. 29, 1168–1177. Harrison, S., Whitton, R.C., King, M., Hausler, K.K., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2012. Forelimb muscle activity during equine locomotion. J. Exp. Biol. 215, 2980–2991. Harrison, S., Whitton, R.C., Kawcak, C.E., Stover, S.M., Pandy, M.G., 2010. Relationship between muscle forces, joint loading, and utilization of elastic strain energy in equine locomotion. J. Exp. Biol. 213, 3998–4009. Heino Brechter, J., Powers, C.M., 2002. Patellofemoral stress during walking in persons with and without patellofemoral pain. Med. Sci. Sports Exercise 34 (10), 1582–1593. Kawcak, C.E., McIlwraith, C.W., 1994. Proximodorsal first phalanx osteochondral chip fragmentation in 336 horses. Equine Vet. J. 26 (5), 392–396. Lloyd, D.G., Besier, T.F., 2003. An EMG-driven musculoskeletal model to estimate muscle forces and knee joint moments in vivo. Journal of biomechanics 36, 765–776. McGuigan, M.P., Wilson, A.M., 2003. The effect of gait and digital flexor muscle activation on limb compliance in the forelimb of the horse Equus caballus. J. Exp. Biol. 206, 1325–1336. Meershoek, L.S., Bogert, A.J.V.D., Schamhardt, H.C., 2001. Model formulation and determination of in vitro parameters of a noninvasive method to calculate flexor tendon forces in the equine forelimb. Am. J. Vet. Res. 62, 1585–1593. Merkens, H.W., Schamhardt, H.C., Hartman, W., Kersjes, A.W., 1986. Ground reaction force patterns of Dutch warmblood horses at normal walk. Equine Vet. J. 18, 207–214. Merkens, H.W., Schamhardt, H.C., Van Osch, G.J.V.M., Van Den Bogert, A.J., 1993. Ground reaction force patterns of Dutch warmblood horses at normal trot. Equine Vet. J. 25, 134–137. Merritt, J.S., Davies, H.M.S., Burvill, C., Pandy, M.G., 2008. Influence of muscle– tendon wrapping on calculations of joint reaction forces in the equine distal forelimb. J. Biomed. Biotechnol. 2008. (Article ID: 165730, 9pp.). Norrdin, R.W., Stover, S.M., 2006. Subchondral bone failure in overload arthrosis: a scanning electron microscopic study in horses. J. Musculoskelet. Neuronal Interact. 6 (3), 251–257. Parkin, T.D.H., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2004. Risk of fatal distal limb fractures among thoroughbreds involved in the five types of racing in the United Kingdom. Vet. Rec. 154, 493–497. Parkin, T.D., Clegg, P.D., French, N.P., Proudman, C.J., Riggs, C.M., Singer, E.R., Webbon, P.M., Morgan, K.L., 2006. Catastrophic fracture of the lateral condyle of the third metacarpus/metatarsus in UK racehorses – fracture descriptions and pre-existing pathology. Vet. J. 171, 157–165. Pustoc'h, A., Cheze, L., 2009. Normal and osteoarthritic hip joint mechanical behaviour: a comparison study. Med. Biol. Eng. Comput. 47 (4), 375–383. Rapillard, L., Charlebois, M., Zysset, P.K., 2006. Compressive fatigue behavior of human vertebral trabecular bone. J. Biomech. 39, 2133–2139. Segal, N.A., Anderson, D.D., Iyer, K.S., Baker, J., Torner, J.C., Lynch, J.A., Felson, D.T., Lewis, C.E., Brown, T.D., 2009. Baseline articular contact stress levels predict incident symptomatic knee osteoarthritis development in the MOST cohort. J. Orthop. Res. 27, 1562–1568. Stepnik, M.W., Radtke, C.L., Scollay, M.C., Oshel, P.E., Albrecht, R.M., Santschi, E.M., Markel, M.D., Muir, P., 2004. Scanning electron microscopic examination of third metacarpal/third metatarsal bone failure surfaces in thoroughbred racehorses with condylar fracture. Vet. Surg. 33, 2–10. Trope, G.D., Anderson, G.A., Whitton, R.C., 2011. Patterns of scintigraphic uptake in the fetlock joint of thoroughbred racehorses and the effect of increased radiopharmaceutical uptake in the distal metacarpal/tarsal condyle on performance. Equine Vet. J. 43, 509–515. Wang, C.C.B., Chahine, N.O., Hung, C.T., Ateshian, G.A., 2003. Optical determination of anisotropic material properties of bovine articular cartilage in compression. J. Biomech. 36, 339–353. Witte, T.H., Knill, K., Wilson, A.M., 2004. Determination of peak vertical ground reaction force from duty factor in the horse (Equus caballus). J. Exp. Biol. 207, 3639–3648.
