- Email: [email protected]
Influence of gait loads on implant integration in rat tibiae: Experimental and numerical analysis
Journal of Biomechanics 47 (2014) 3255–3263
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Influence of gait loads on implant integration in rat tibiae: Experimental and numerical analysis Marco Piccinini a, Joel Cugnoni a,n, John Botsis a, Patrick Ammann b, Anselm Wiskott c a
Laboratory of Applied Mechanics and Reliability Analysis, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland Division of Bone Diseases, Department of Internal Medicine Specialities, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland c Division of Fixed Prosthodontics and Biomaterials, School of Dental Medicine, University of Geneva, Geneva, Switzerland b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 25 August 2014
Implanted rat bones play a key role in studies involving fracture healing, bone diseases or drugs delivery among other themes. In most of these studies the implants integration also depends on the animal daily activity and musculoskeletal loads, which affect the implants mechanical environment. However, the tissue adaption to the physiological loads is often filtered through control groups or not inspected. This work aims to investigate experimentally and numerically the effects of the daily activity on the integration of implants inserted in the rat tibia, and to establish a physiological loading condition to analyse the peri-implant bone stresses during gait. Two titanium implants, single and double cortex crossing, are inserted in the rat tibia. The animals are caged under standard conditions and divided in three groups undergoing progressive integration periods. The results highlight a time-dependent increase of bone samples with significant cortical bone loss. The phenomenon is analysed through specimen-specific Finite Element models involving purpose-built musculoskeletal loads. Different boundary conditions replicating the post-surgery bone–implant interaction are adopted. The effects of the gait loads on the implants integration are quantified and agree with the results of the experiments. The observed cortical bone loss can be considered as a transient state of integration due to bone disuse atrophy, initially triggered by a loss of bone–implant adhesion and subsequently by a cyclic opening of the interface. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Rat Tibia Finite element Musculoskeletal loads Implant integration Bone loss
1. Introduction To optimize biological resource allocation, bones are maintained in a state of structural balance between applied loads and mechanical resistance. Under this premise, reduced mobility (Globus et al., 1984; Maïmoun et al., 2012) or strong exercise (Fujie et al., 2004) alter a skeleton’s muscular environment and the bones adapt to the newly established stress fields. After a period of structural remodelling, a new equilibrium is established (Frost, 1990b). Similarly, endosseous implants modify the stress field within the bone bed and thus induce a tissue adaption that may also detrimentally affect the osseous casing, such as in stress shielding (Huiskes et al., 1992), where the bone is actually “underloaded” and will resorb. In spite of their reduced size, rats are applicable to studies on implant biomechanics (Stadlinger et al., 2012; Dayer et al., 2006).
n Correspondence to: Laboratory of Applied Mechanics and Reliability Analysis (LMAF), Ecole Polytechnique Fédérale de Lausanne, Station 9, CH-1015 Lausanne, Switzerland. Tel.: þ 41 21 693 59 73; fax: þ41 21 693 73 40. E-mail addresses: marco.piccinini@epfl.ch (M. Piccinini), joel.cugnoni@epfl.ch (J. Cugnoni), john.botsis@epfl.ch (J. Botsis), [email protected] (P. Ammann), [email protected] (A. Wiskott).
http://dx.doi.org/10.1016/j.jbiomech.2014.08.023 0021-9290/& 2014 Elsevier Ltd. All rights reserved.
For instance, rats have been used to assess the effects of systemic diseases or metabolic disturbances on implant osseointegration (Fiorellini et al., 1999; Glösel et al., 2010; Viera-Negrón et al., 2008; Ammann et al., 2007; Dayer et al., 2010). Rats were also employed when evaluating biocompatible coatings (Hara et al., 1999; Jaatinen et al., 2011), drug delivery systems (Peter et al., 2005; Stadelmann et al., 2009) and locking plates in studies on fracture healing (Histing et al., 2011). Problematically though, the histological process of integration always proceeds within the environment of mechanical stimuli generated during the animals’ daily activity. Indeed when the implants are placed on limbs and the animals are left unconstrained, locomotion will affect peri-implant bone healing and remodelling (Frost, 1990a). Typically, these sways are accounted for by establishing test- and control groups that account for both the animals’ genetic biovariability and their everyday musculoskeletal activation. In most instances, this approach suffices because the analyses are based on comparisons between large groups so that the effects on the outcome parameters can be extracted. In some instances though, the rats’ daily activity strongly interferes with the implants’ experimental loading and thus precludes a pertinent analysis of the research issue under scrutiny. To extract relevant information from such data sets, these
3256
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
situations require a thorough assessment of the stress systems developed during normal daily life. In humans, this theme is approached through inverse dynamics musculoskeletal analysis and Finite Element (FE) models accounting for gait loads (Phillips, 2009; Phillips et al., 2007). Concerning rats, some evidence was provided in studies on rat stride lengths, frequencies and ground reaction forces (Clarke, 1991; Clarke and Parker, 1986; Muir and Whishaw, 1999). But still, there is a considerable research deficit regarding the quantification of the forces developed by the musculature and the loads generated on the joints during the rat gait. To our knowledge, the only work addressing this issue dealt with the internal moments and forces acting on the femur of a walking rat (Wehner et al., 2010). The objective of the present study, therefore, was twofold. (i) To develop a framework of the physiological loading conditions acting on rat tibiae during gait. (ii) To assess the interplay between the proposed framework and the bone’s reactions around titanium implants inserted in rat tibiae.
2. Material and methods This work belongs to a series of investigations on the dependency of implants’ stability on the mechanical environment. According to the hypothesis of the “loaded implant” model (Wiskott et al., 2008, 2011; Zacchetti et al., 2013), two extracutaneous implants are inserted in the proximal part of the rat tibia, and their integration evolves in relation to two mechanical stimulations: the animal activity and the controlled load introduced by activating the implants’ heads. The present work is focused on the effects of the former. All animal experiments were approved by the University of Geneva animal rights committee and supervised by the local veterinary board. 2.1. Animal experiment The study was conducted on 30 Sprague-Dawley rats. The animals were 24-weeks old at the onset of the experiment. Two titanium implants were inserted into the proximal part of the right tibia with their heads protruding from the skin. As shown in Fig. 1, the distal implants were anchored bicortically while the proximal implants only penetrated the medulla by ca. 3 mm, thereby replicating single- and bi-cortical anchorage modes (Ivanoff et al., 2000). The endosseous bore was dimensioned to produce an interference of 0.02 mm with the implant cylinders, hence ensuring implant stability during the early days of healing. Both implants had been previously sandblasted and acid-etched to improve osseointegration (Cochran et al., 1996). After surgery, the animals were caged under standard conditions, with free access to food and water, and divided into three experimental groups undergoing different integration periods: 2, 4 and 6 weeks. Animals were daily inspected to monitor the implantation conditions. No infections (i.e. purulent emissions) were detected. Transient inflammations were healed by treating them with disinfectants. After the sacrifice, the implanted tibiae were dissected, cleared of their soft tissue coverage and frozen to 21 1C. 2.2. Computed tomography processing Prior to computed tomography (CT) scanning, the specimens were maintained at 4 1C for 24 h in a 0.9% solution of NaCl. Then, they were gradually brought to room temperature. The specimens were analysed using a high resolution CT imaging system (μCT-40, Scanco Medical AG, Brüttisellen, Switzerland). For each specimen, 1022 slices
were generated per 360 degrees of rotation using the following settings: isotropic voxel size: 20 μm, source potential: 70 kVp, tube current: 114 μA, integration time: 320 ms, beam hardening correction: 200 mg HA/cm3.
2.3. Morphologic analysis The CT images of all the specimens were imported into ITKsnap (Yushkevich et al., 2006) and inspected by a CT specialist who was unaware of the specimens’ experimental groups. The morphology of the peri-implant cortical bone was classified as follows. (i) ‘Flat’ when complete bone-to-implant adhesion was observed and the tibia’s normal cortical geometry was preserved. (ii) ‘Conic’ if the zone of peri-implant adhesion was contiguous to a funnel-shaped defect of cortical bone. In this instance, the depth of the cone, Δc, was recorded. (iii) ‘Open’ whenever a gap had opened between the cortical bone and the implant. Two Regions of Interest were assessed: the peri-implant bone tissue surrounding the distal (ROId) and proximal (ROIp) implants. The morphologic features, the conic depth and the inspected ROIs are highlighted in Fig. 2.
2.4. Gait based loading model The joints and muscle loads acting on the rat tibia during gait are estimated through the equilibrium of the femur and tibia rigid body models. The musculoskeletal geometry (Johnson et al., 2008) and forces (Wehner et al., 2010) are combined to establish the equilibrium equations systems. The static equilibrium is calculated at a time-step corresponding to 35% of the gait cycle (Wehner et al., 2010), i.e. when most of the muscles reach their maximum force. First, the condylar reactions acting on the femur were calculated by equilibrating a system that included (i) the femur’s internal forces and moments, (ii) the lateral and medial gastrocnemius and (iii) the loads developed in the patella taken as the resultant of pulls by the m. vastus lateralis, the m. vastus medialis and the m. rectus femoris, all acting along the bisector lines of their origin-via-insertion coordinates (Fig. 3b). The coordinates of the condylar joint reactions were compatible with the contact areas on the femur’s condyles (Dao et al., 2011). Then the tibial equilibrium system was established. It comprised the calculated condylar reactions and the forces developed by the m. vastus lateralis and medialis, the m. rectus femoris and the m. tibialis anterior. The m. biceps femoris was treated separately. Indeed although it has been demonstrated as being active during the stance phase (Gillis and Biewener, 2001), there is a lack of information regarding the force generated by this muscle during gait. Hence, the tibial equilibrium equations were solved while treating the m. biceps femoris and a single ankle reaction acting on the inter-malleolar point as unknowns (Fig. 3c). The coordinates of the muscular attachments and the calculated loads are listed in Table 1.
2.5. Finite element models and boundary conditions. A continuum FE model of a rat tibia was generated from available micro-CT data through the procedure previously developed by the authors (Piccinini et al., 2012). First, the CT images were processed with the software ITKsnap (Yushkevich et al., 2006). Bone was segmented according to grayscale thresholds and a 3D representation of the rat tibia was obtained. Second, the CT images were processed through the open-source FE model generator VoxelMesher. More specifically, the segmentation was extracted as a binary image and its boundary was discretized using a triangle meshing algorithm for 3D implicit surfaces (Boissonnat and Oudot, 2005). The surface mesh was embedded in a bounding box, where seed nodes were generated using a recursive octree decomposition of the domain. Using the bounding box, the surface mesh and the seed nodes as input, an initial constrained Delaunay tetrahedral mesh was generated using TetGen (Si, 2008). After deleting the undesired domains, a high-quality tetrahedral mesh was calculated by refining the initial one using a local element size map. Then, the linear tetrahedra were
Fig. 1. (a) 3D representation of the proximal part of a rat tibia: the insertion points of the distal and proximal implant are represented. (b) View cut showing a schematic representation of the different implantation strategies: bi-cortical (distal implant) and mono-cortical (proximal implant).
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
3257
Fig. 2. mCT scans of an implanted specimen: (a) distal, ROId, and proximal, ROIp, region of interest, (b) flat morphology, (c) conic morphology and conic depth, Δc, and (d) open morphology.
Fig. 3. (a) Simplified representation of the rat hindlimb musculoskeletal system; muscles are represented as lines connecting their origin-via-insertion (Johnson et al., 2008). (b) Equilibrium of the femur: the medial and lateral condylar reactions, Cm and Cl, are calculated equilibrating the internal forces Fi and moments Mi, the patellar load Pi and the lateral and medial gastrocnemius, Gl and Gm. (c) Equilibrium of the tibia: the ankle joint reaction Aj and the bicep femoris load Bf are estimated equilibrating the vastus medialis Vm and lateralis Vl, the rectus femoris Rf, the tibialis anterior proximal TAp and distal TAd and the lateral and medial condylar reactions Cm and Cl previously calculated. converted to quadratic elements by adding mid-side nodes. Eventually, each element was assigned a modulus of elasticity in accordance with a 3D map of elastic modulus (Taddei et al., 2007), obtained by converting the grey value of each voxel of the original CT images into a corresponding modulus of elasticity, through the relationship developed by Cory et al. (2010). Thus, the tibia mechanical behaviour was characterized by an isotropic, continuum and inhomogeneous material field that preserves the bone internal structure. Then, the implants were inserted in-silico into the tibia proximal segment whereby the distal implant’s threaded end was reduced to a cylinder with tied contact to the cortical bone. As the integration progresses during the healing period, the mechanics of the implant–bone interface evolves from a press-fit condition to a simple adhesion. To simulate this phenomenon, three boundary conditions were analysed. In a first model, the implants were “press-fitted” into their bone bed—a condition which was modelled as a homogeneous radial displacement field of 0.01 mm that was applied to the cylindrical surfaces of the implants (Natali et al., 2009). In a second FE model, bone is
tied to the implant surfaces without press fit, while in the third one the opening of the bone–implant interface is allowed by considering a hard, frictionless contact condition between the bone and the implant without interference. All four models, including that which was bare of implants, were subjected to the gait-based loading condition previously defined. The essentials of all four models are summarized in Table 2. Computations were performed through the ABAQUS-Standard solver.
3. Results 3.1. Morphologic analysis The outcomes of the morphologic analyses are shown in Table 3, all implants integrated. After two weeks of healing, conical bone loss was
3258
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Table 1 Loads acting on the tibia, calculated with respect to the reference system in Fig. 3. Structure
Rectus femorisa,b Vastus lateralisa,b Vastus medialisa,b Tibialis anterior proximala,b Tibialis anterior distala,b Biceps femorisb Lateral condyle Medial condyle Ankle jointb
Coordinates (mm)
Force (N)
x
y
z
x
y
z
39.70 39.70 39.70 39.00 4.27 33.00 40.10 39.90 0.00
0.00 0.00 0.00 2.29 0.00 0.50 2.00 2.00 0.00
2.00 2.00 2.00 1.50 0.86 0.20 1.90 1.10 0.00
5.45 2.33 1.95 2.83 1.61 0.00 12.75 0.03 4.20
0.22 2.27 0.88 0.24 0.33 3.63 0.38 0.38 1.17
3.94 2.37 1.41 0.01 2.57 5.98 7.56 5.84 2.26
Table 3 Results of the morphologic analysis: percentage of specimens characterized by flat, conic or open peri-implant morphology after two, four and six weeks of integration. Integration period
Feature
ROId (%)
ROIp (%)
2 weeks
Flat Conic Open Flat Conic Open Flat Conic Open
79.3 14.0 6.7 50.0 33.3 16.7 41.6 41.7 16.7
80.0 20.0 0.0 65.0 35.0 0.0 55.0 45.0 0.0
4 weeks
6 weeks
a Loads magnitudes from Wehner et al. (2010). The x-, y- and z-components are calculated as the projection to the insertion–origin direction. b Coordinates from Johnson et al. (2008).
Table 2 Details of the boundary conditions adopted in the FE models. The contact conditions concern ROId and ROIp. The threaded tip of the distal implant is simplified with a tie contact. Model
Implants
Bone–implant adhesion
Press Fit
Gait loads
Bare Press Fit Adhesion Opening
No Yes Yes Yes
– Yes Yes Noa
– Yes No No
Yes Yes Yes Yes
a
Frictionless hard contact.
observed in 14% of ROIds and 20% of ROIps. After six weeks, the figures were 41% and 45%, respectively. An ‘open’ condition was only observed on the distal implants (17% after 6 weeks). None of the proximal implants presented this characteristic. In all specimens, the threaded tip of the distal implant was well integrated into cortical bone. Fig. 4 lists the cone depths after 2, 4 and 6 weeks of healing. As shown, the mean cone depth increased with time and finally reached values approaching that of the thickness of the cortical bone ( 0.5 mm). The increment between 2 and 6 weeks was significant for both ROId and ROIp although for ROIp no major increase was observed between 4 and 6 weeks. 3.2. Finite element analysis When the bare tibia model was subjected to nominal gait loads, the strain pattern developed as a coupling of bending and compression typical of long bones (Fig. 5). The tensile and compressive longitudinal strains in the midshaft worked out to 1112 με and 1318 με, respectively. During the early days after surgery, the deformation state occurring in the FE model with in-silico implants blend the effects of the gait loads and the press fit, generating the peri-implant radial stress represented in Fig. 6a and b, concerning the cortical tissue. The radial compression pattern generated by the press-fit surpasses the tensions developed during gait loading, leading to normal compressive stresses in the cortical contact areas in both proximal and distal implants. In the next phase of integration, the effects of press-fitting are released to represent bone adaptation and only the gait loads act on the implanted tibia (Fig. 6c and d). The radial stresses reach peak values of tension up to 50 MPa (Fig. 6e and f) in the distal–proximal direction. Note that the radial stress-state is not uniform and evolves from large tensile stresses in the distal–proximal direction to negligible compressive stresses in the medio-lateral direction. Fig. 7 depicts the response of non-adhesive bone-to-implant interfaces in terms of mean contact pressures and openings that
Fig. 4. Conic depth measured in specimens presenting a cortical bone loss. Mean and standard deviation are calculated on groups (n ¼10) sacrificed after integration periods of 2, 4 and 6 weeks ( p o 0.05, t-test).
develop at the implant surface. Under this condition, gait loads result in compressive stresses on the implants’ circumference with exception of the superficial zones in which tension develops in the longitudinal direction and the bone detaches from the implant. This phenomenon was observed in both ROIs but was more pronounced in ROId. In ROIp no clear-cut distinction between zones of pressure and opening was possible due to irregularities in thickness of the cortical bone and its partial entanglement with trabecular tissue. Last, Fig. 8 shows the stresses occurring in the longitudinal plane in case of opening depths corresponding to 30% and 60% of the cortical thickness with a perfectly adherent implant–bone interface. The bone tissue in the proximity of the open contact is clearly unloaded, while high tensile stress concentrations characterize the area immediately below the opening zone. The stress concentrations affecting the distal implant surpass 20 MPa and are relatively independent of the opening depth (Fig. 8a and b). Contrarily, the stress concentrations are much less pronounced around the proximal implant (Fig. 8c and d) and show sign of a progressive reduction with the increase of conic depth.
4. Discussion The present experimental–numerical analysis aimed at relating the cortical bone defects often observed in implanted rat tibiae to local stress patterns and to the ensuing osseous remodelling. To this end, finite element models of tibiae were generated and subjected to a loading environment that simulated the forces generated during the animals’ daily activities and more specifically during gait. The pertinence of the numerical model was verified to match previously published values at comparable locations. For instance, the strains calculated for the bare (i.e. devoid of implants) tibia’s
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
3259
Fig. 5. Finite Element analysis of the bare rat tibia. (a) Longitudinal strains, (b) maximum and minimum longitudinal strains along the tibial midshaft. The fibula is not considered. The value in the dotted box agrees with the in-vivo measurements by Rabkin et al. (2001).
Fig. 6. Finite Element analysis of the rat tibia with in-silico implants. A full bone–implant adhesion is assumed. Cortical radial stress considering both the gait loads and the press fit in (a) ROIp and (b) ROId. Cortical radial stress considering only the gait loads in (c) ROIp and (d) ROIp. Stresses along the implant circumference considering both press fit and gait loads (blue) and only gait loads (red) in (e) ROIp and (f) ROId. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
midshaft are in close agreement with those generated in rat long bones during locomotion: 600–1200 με (Turner et al., 1994; Hillam and Skerry, 1995; Mosley et al., 1997). Similarly the longitudinal strain illustrated in Fig. 5b agrees with in-vivo measurements at the same location (Rabkin et al., 2001).
Early post-surgery, the stresses arising in the peri-implant bone bed blend the compression due to the implants’ press fit and the tensile stresses generated during the animals’ locomotion (Fig. 6a and b). As the first one is initially predominant, the compressive forces impart their stability to the implants, prevent
3260
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Fig. 7. Results of the contact analysis in case of frictionless hard contact. (a) Contact pressure and opening in ROId, (b) contact pressure and opening in ROIp. Pressure and opening mean and standard deviation are calculated on the implant circumference.
Fig. 8. Longitudinal stress in case of partial cortical loss of adhesion. Distal implant with debonding depth equal to (a) 30% of cortical thickness and (b) 60% of cortical thickness. Proximal implant with debonding depth equal to (c) 30% of cortical thickness and (d) 60% of cortical thickness.
harmful micro-motions and thus facilitate initial integration (Abdul-Kadir et al., 2008). This stabilizing effect though, comes to an end 3–4 days after implant placement due to the resilience and remodelling of the surrounding osseous tissue (Dhert et al., 1998). Now the peri-implant bone’s evolution depends on the rats’ physical activity, the adhesive strength of the bone–implant interface and amplitude of micromotions in the bone–implant cavity. As shown in Fig. 6c and d, when press-fit is released, the interface is subjected to severe radial tensile stresses at each gait cycle.
Problematically, these values (20–50 MPa) exceed the adhesive strength between the implant and the surrounding bone (1– 4 MPa) (Gross et al., 1987; Takatsuka et al., 1995). It follows that the loss of bone-to-implant contact as observed experimentally may well be initiated by a loss of adhesion due to the excess tensile stresses generated at each gait cycle. The affected bone now loses its stimulation and reacts by migrating in apical direction along the implant. A similar bone shape is observed around overloaded implants inserted in dog and
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
rabbit bones (Hoshaw et al., 1990; Duyck et al., 2001). Clinically, this phenomenon is similar to the marginal bone loss observed around dental implants (Qian et al., 2012), which has been attributed either to disease (Bodic et al., 2005) or to overload (Isidor, 1996). Interestingly, neither diseases nor overloading were overtly active in the present study; hence a different mechanism driving cortical bone loss must be postulated (Fig. 8). According to the present FE analysis, permanent gaps may develop at the implant–bone interface. If this occurs, an irreversible process is initiated as the bone tissue that is separated from the implant is in a condition of ‘stress shielding’ and is resorbed because of disuse (Engh et al., 1987). Moreover, once the process of gap formation is initiated, the stresses generated during each gait cycle are relocated further down the implant and a new zone of the bone–implant interface is subjected to tensile forces (Fig. 8). Locomotion thus causes the inward propagation of a gap along the implant surface with the ancillary effect of resorption of unloaded bone and the eventual formation of a funnel-shaped defect in the cortical bone. The stability of this progressive process is controlled at least partially by the evolution of the peak tensile stresses at the bone–implant interface and thus depends on the local bone structure and implantation strategy. Indeed, this mechanism was observed in both ROIs but the differing bony environment and implantation strategies led to different defect morphologies. The tip of the distal implant was screwed-tightened into the opposing cortex while the upper part was inserted in a thick layer of cortical bone without trabecular bone in between. Due to an extended contact area, this configuration yielded an excellent initial resistance to the loss of adhesion (lower initial stresses). Clinically, it took 4 weeks for the funnel in ROId to start to increase (Fig. 4), thus indicating a high initial resistance of the interface. However, once debonding occurred, this configuration demonstrated a fast development of a funnel shape between 4 and 6 weeks (Fig. 4). The high stress concentrations seen in the FE simulations (Fig. 8a and b) and the fact that their magnitude does not decrease when debonding propagates explains the high rate of development of conic depth and the presence of an open gap in 17% of the distal implants after 6 weeks. By contrast, the proximal implant was inserted into a thin layer of cortical bone and large amounts of trabecular bone underneath. As shown by the conic depth evolution (Fig. 4) and the relatively high percentage of conic features (Table 3) between 2 and 4 weeks, this environment seems to offer less initial resistance to periosteal tension and promotes early debonding at the outer cortical surface. However, this implantation strategy demonstrated a better
3261
adaptation to the mechanical environment after the early stages, as the rate of evolution of the conic features was much less than in ROId after 4 weeks and no open gaps were observed around the proximal implants after 6 weeks. Indeed, as shown in Fig. 8, the smooth gradient of stiffness offered by the underlying trabecular bone helped to progressively decrease the interfacial stress concentration with increasing conic depth, and thus lead to an arrest of the gap formation process. Moreover the results show that the rate of densification of trabecular bone was sufficiently high to produce a dense osseous shell with a funnel shape around the proximal implant. Thus, the absence of trabecular bone around the distal implant could be seen as a factor limiting the adaptation capacity and facilitating the formation of a fully open gap. Within the observed time period, it thus appears that large contacts zones in the cortex delay the bone–implant opening but cannot prevent the development of fully open gaps, while trabecular bone hampers the debonding and prevents the complete failure of the interface. Still, whether an animal will – or will not – demonstrate periimplant bone loss cannot be predicted. Indeed a number of individual factors are contributive: the degree of press-fit relaxation, the adhesive strength and the size of the bone–implant contact, the morphology and material properties of the surrounding bone tissue as well as the rats’ activity during the post-surgical period (different levels of pain might prevent the animal to use the implanted leg and thus reduce the number of load applications). Thus, three configurations are possible: (1) no massive debonding occurs because press fit is retained sufficiently long to allow bone in-growth and stability of the implant, (2) a local bone–implant debonding is initiated on the outer cortical surface which finally stops at a certain depth (3) an unstable evolution of debonding/ bone resorption which leads to an open gap between implant and bone. Compared to previous studies, it should be noted here that the proposed bone resorption mechanism can only occur in implantation sites in which tensile stresses are dominant during physiological activities. Other causes may lead to similar looking cortical bone loss such as overloading (Isidor, 1996; Hoshaw et al., 1990; Duyck et al., 2001) or diseases (Bodic et al., 2005). Due to the use of transcutanesous implant, the presence of debris and local inflamation or infections cannot be completely ruled out from the study. However, based on the daily inspection in the present experiment, no external signs of purulent infections were noticed and only a negligible number of animals developed local inflammations which were efficiently treated by disinfectant. According to the experience gained so far in the loaded-implant animal
Fig. 9. Simplified representation of the mechanisms leading to the cortical bone loss. (a) In the early post surgery days the implant is stable because of the compressive stress field due to the press fit. (b) When the press fit is released part of the interface is subjected to traction. (c) The cyclic loading provokes a loss of bone–implant adhesion in the periosteal area. The stress distribution changes and part of the tissue is unloaded (i.e. above the dotted line). (d) The unloaded tissue is resorbed and the opening propagates along the implant axe. (e) If no stable configurations exist the opening reaches the endosteum and the bone–implant contact is totally lost.
3262
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
model in rat tibiae (more than 160 animals, with the same implantation), the observed occurence of purulent infections (1.2%) and inflammations (16%) is much lower than the occurence of conic cortical bone loss observed in the present study (40%). Thus, the conic shaped cortical bone loss observed in this work is predominantly attributed to the effect of tensile stresses and progressive debonding–bone resorption mechanisms as described in Fig. 9.
5. Conclusions In this study, a model of a rat’s tibia subjected to musculoskeletal loads was constructed, yielding a realistic representation of the deformation fields that develop in the tibia during gait. This model enhances our knowledge of the biomechanics of rat tibiae under physiological conditions and sets the baseline according to which implantation studies on rat tibiae are to be evaluated. With respect to the ‘loaded implant’ model, it now appears that normal locomotion may tear the adhesive interface between the implant and the bone to the extent that a gap opens or at least the contact area is drastically reduced. The observed funnel-shaped bone loss is caused by bone disuse atrophy which is initiated by a loss of bone–implant adhesion and kept ongoing by the cyclic loadings on the interface due to gait cycles. Thus, in the “loaded implant” model, rat locomotion detrimentally affects implant integration. Other type of implantation, for example in dentistry, may be affected by a similar detrimental mechanism affecting the implant stability. It can thus be recommended to consider not only the external forces but also the effects of daily physiological loads when studying the implants integration.
Conflict of interest statement The authors declare no conflict of interest.
Acknowledgements Mrs Severine Clement and Mrs. Isabelle Badoud (UNIGE) are gratefully acknowledged for their work on surgery, animal care, and data acquisition. This work was partially funded by Swiss National Science Fundation (SNF) project 315230-127612. References Abdul-Kadir, M.R., Hansen, U., Klabunde, R., Lucas, D., Amis, A., 2008. Finite element modelling of primary hip stem stability: the effect of interference fit. J. Biomech. 41, 587–594. Ammann, P., Badoud, I., Barraud, S., Dayer, R., Rizzoli, R., 2007. Strontium ranelate treatment improves trabecular and cortical intrinsic bone tissue quality, a determinant of bone strength. J. Bone Miner. Res. 22, 1419–1425. Bodic, F., Hamel, L., Lerouxel, E., Baslé, M.F., Chappard, D., 2005. Bone loss and teeth. Joint Bone Spine 72, 215–221. Boissonnat, J.D., Oudot, S., 2005. Provably good sampling and meshing of surfaces. Graphical Models 67, 405–451. Clarke, K.A., 1991. Swing time changes contribute to stride time adjustment in the walking rat. Physiol. Behav. 50, 1261–1262. Clarke, K.A., Parker, A.J., 1986. A quantitative study of normal locomotion in the rat. Physiol. Behav. 38, 345–351. Cochran, D.L., Nummikoski, P.V., Higginbottom, F.L., Hermann, J.S., Makins, S.R., Buser, D., 1996. Evaluation of an endosseous titanium implant with a sandblasted and acid-etched surface in the canine mandible: radiographic results. Clin. Oral Implants Res. 7, 240–252. Cory, E., Nazarian, A., Entezari, V., Vartanians, V., Müller, R., Snyder, B.D., 2010. Compressive axial mechanical properties of rat bone as functions of bone volume fraction, apparent density and micro-ct based mineral density. J. Biomech. 43, 953–960. Dao, T.T., Pouletaut, P., Goebel, J.C., Pinzano, A., Gillet, P., Ho Ba Tho, M.C, 2011. In vivo characterization of morphological properties and contact areas of the rat cartilage derived from high-resolution MRI. IRBM 32, 204–213.
Dayer, R., Brennan, T.C., Rizzoli, R., Ammann, P., 2010. PTH improves titanium implant fixation more than pamidronate or renutrition in osteopenic rats chronically fed a low protein diet. Osteoporos. Int. 21, 957–967. Dayer, R., Rizzoli, R., Kaelin, A., Ammann, P., 2006. Low protein intake is associated with impaired titanium implant osseointegration. J. Bone Miner. Res. 21, 258–264. Dhert, W.J.A., Thomsen, P., Blomgren, A.K., Esposito, M., Ericson, L.E., Verbout, A.J., 1998. Integration of press-fit implants in cortical bone: a study on interface kinetics. J. Biomed. Mater. Res. 41, 574–583. Duyck, J., Rønold, H.J., Van Oosterwyck, H., Naert, I., Sloten, J.V., Ellingsen, J.E., 2001. The influence of static and dynamic loading on marginal bone reactions around osseointegrated implants: an animal experimental study. Clin. Oral Implants Res. 12, 207–218. Engh, C.A., Bobyn, J.D., Glassman, A.H., 1987. Porous-coated hip replacement. The factors governing bone ingrowth, stress shielding, and clinical results. J. Bone Joint Surg. Ser. B 69, 45–55. Fiorellini, J.P., Nevins, M.L., Norkin, A., Weber, H.P., Karimbux, N.Y., 1999. The effect of insulin therapy on osseointegration in a diabetic rat model. Clin. Oral Implants Res. 10, 362–368. Frost, H.M., 1990a. Skeletal structural adaptations to mechanical usage (SATMU): 1. Redefining Wolff’s law: the bone modeling problem. Anat. Rec. 226, 403–413. Frost, H.M., 1990b. Skeletal structural adaptions to mechanical usage (SATMU): 2. Redefining Wolff’s law: the remodeling problem. Anat. Rec. 226, 414–422. Fujie, H., Miyagaki, J., Terrier, A., Rakotomanana, L., Leyvraz, P.F., Hayashi, K., 2004. Detraining effects on the mechanical properties and morphology of rat tibiae. Bio-Med. Mater. Eng. 14, 219–233. Gillis, G.B., Biewener, A.A., 2001. Hindlimb muscle function in relation to speed and gait: in vivo patterns of strain and activation in a hip and knee extensor of the rat (Rattus norvegicus). J. Exp. Biol. 204, 2717–2731. Globus, R.K., Bikle, D.D., Morey-Holton, E., 1984. Effects of simulated weightlessness on bone mineral metabolism. Endocrinology 114, 2264–2270. Glösel, B., Kuchler, U., Watzek, G., Gruber, R., 2010. Review of dental implant rat research models simulating osteoporosis or diabetes. Int. J. Oral Maxillofac. Implants 25, 516–524. Gross, U., Roggendorf, W., Schmitz, H.J., Strunz, V., 1987. Biomechanical and Morphometric Testing Methods for Porous and Surface-reactive Biomaterials. ETATS-UNIS, American Society for Testing and Materials, Philadelphia, PA. Hara, T., Hayashi, K., Nakashima, Y., Kanemaru, T., Iwamoto, Y., 1999. The effect of hydroxyapatite coating on the bonding of bone to titanium implants in the femora of ovariectomised rats. J. Bone Joint Surg. Ser. B 81, 705–709. Hillam, R.A., Skerry, T.M., 1995. Inhibition of bone resorption and stimulation of formation by mechanical loading of the modeling rat ulna in vivo. J. Bone Miner. Res. 10, 683–689. Histing, T., Garcia, P., Holstein, J.H., Klein, M., Matthys, R., Nuetzi, R., Steck, R., Laschke, M.W., Wehner, T., Bindl, R., Recknagel, S., Stuermer, E.K., Vollmar, B., Wildemann, B., Lienau, J., willie, B., Peters, A., Ignatius, A., Pohlemann, T., Claes, L., Menger, M.D., 2011. Small animal bone healing models: standards, tips, and pitfalls results of a consensus meeting. Bone 49, 591–599. Hoshaw, S.J., Brunski, J.B., Cochran, G.V.B., Higuchi, K.W., 1990. Theories of Bone Modeling and Remodeling in Response to Mechanical Usage. Experimental Investigation of an in vivo Bone–Implant Interface. 391-394. Huiskes, R., Weinans, H., Van Rietbergen, B., 1992. The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res., 124–134. Isidor, F., 1996. Loss of osseointegration caused by occlusal load of oral implants: a clinical and radiographic study in monkeys. Clin. Oral Implants Res. 7, 143–152. Ivanoff, C.J., Gröndahl, K., Bergström, C., Lekholm, U., Brånemark, P.I., 2000. Influence of bicortical or monocortical anchorage on maxillary implant stability: a 15-year retrospective study of Brånemark system implants. Int. J. Oral Maxillofac. Implants 15, 103–110. Jaatinen, J.J.P., Korhonen, R.K., Pelttari, A., Helminen, H.J., Korhonen, H., Lappalainen, R., Kröger, H., 2011. Early bone growth on the surface of titanium implants in rat femur is enhanced by an amorphous diamond coating. Acta Orthop. 82, 499–503. Johnson, W.L., Jindrich, D.L., Roy, R.R., Reggie Edgerton, V., 2008. A threedimensional model of the rat hindlimb: musculoskeletal geometry and muscle moment arms. J. Biomech. (41), 610–619. Maïmoun, L., Brennan-Speranza, T.C., Rizzoli, R., Ammann, P., 2012. Effects of ovariectomy on the changes in microarchitecture and material level properties in response to hind leg disuse in female rats. Bone 51, 586–591. Mosley, J.R., March, B.M., Lynch, J., Lanyon, L.E., 1997. Strain magnitude related changes in whole bone architecture in growing rats. Bone 20, 191–198. Muir, G.D., Whishaw, I.Q., 1999. Ground reaction forces in locomoting hemiparkinsonian rats: a definitive test for impairments and compensations. Exp. Brain Res. 126, 307–314. Natali, A.N., Carniel, E.L., Pavan, P.G., 2009. Dental implants press fit phenomena: biomechanical analysis considering bone inelastic response. Dent. Mater. 25, 573–581. Peter, B., Pioletti, D.P., Laïb, S., Bujoli, B., Pilet, P., Janvier, P., Guicheux, J., Zambelli, P.Y., Bouler, J.M., Gauthier, O., 2005. Calcium phosphate drug delivery system: influence of local zoledronate release on bone implant osteointegration. Bone 36, 52–60. Phillips, A.T.M., 2009. The femur as a musculo-skeletal construct: a free boundary condition modelling approach. Med. Eng. Phys. 31, 673–680. Phillips, A.T.M., Pankaj, P., Howie, C.R., Usmani, A.S., Simpson, A.H.R.W., 2007. Finite element modelling of the pelvis: inclusion of muscular and ligamentous boundary conditions. Med. Eng. Phys. 29, 739–748.
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Piccinini, M., Cugnoni, J., Botsis, J., Zacchetti, G., Ammann, P., Wiskott, A., 2012. Factors affecting subject-specific finite element models of implant-fitted rat bone specimens: critical analysis of a technical protocol. Comput. Methods Biomech. Biomed. Eng.. Qian, J., Wennerberg, A., Albrektsson, T., 2012. Reasons for marginal bone loss around oral implants. Clin. Implant Dent. Relat. Res. 14, 792–807. Rabkin, B.A., Szivek, J.A., Schonfeld, J.E., Halloran, B.P., 2001. Long-term measurement of bone strain in vivo: the rat tibia. J. Biomed. Mater. Res. 58, 277–281. Si, H, 2008. Adaptive tetrahedral mesh generation by constrained Delaunay refinement. Int. J. Numer. Methods Eng. 75, 856–880. Stadelmann, V.A., Terrier, A., Gauthier, O., Bouler, J.M., Pioletti, D.P., 2009. Prediction of bone density around orthopedic implants delivering bisphosphonate. J. Biomech. 42, 1206–1211. Stadlinger, B., Pourmand, P., Locher, M.C., Schulz, M.C., 2012. Systematic review of animal models for the study of implant integration, assessing the influence of material, surface and design. J. Clin. Periodontol. 39, 28–36. Taddei, F., Schileo, E., Helgason, B., Cristofolini, L., Viceconti, M., 2007. The material mapping strategy influences the accuracy of CT-based finite element models of bones: an evaluation against experimental measurements. Med. Eng. Phys. 29, 973–979. Takatsuka, K., Yamamuro, T., Nakamura, T., Kokubo, T., 1995. Bone-bonding behavior of titanium alloy evaluated mechanically with detaching failure load. J. Biomed. Mater. Res. 29, 157–163.
3263
Turner, C.H., Forwood, M.R., Rho, J.Y., Yoshikawa, T., 1994. Mechanical loading thresholds for lamellar and woven bone formation. J. Bone Miner. Res. 9, 87–97. Viera-Negrón, Y.E., Ruan, W.H., Winger, J.N., Hou, X., Sharawy, M.M., Borke, J.L., 2008. Effect of ovariectomy and alendronate on implant osseointegration in rat maxillary bone. J. Oral Implantol. 34, 76–82. Wehner, T., Wolfram, U., Henzler, T., Niemeyer, F., Claes, L., Simon, U., 2010. Internal forces and moments in the femur of the rat during gait. J. Biomech. 43, 2473–2479. Wiskott, H.W.A., Bonhote, P., Cugnoni, J., Durual, S., Zacchetti, G., Botsis, J., Scherrer, S.S., Ammann, P., 2011. Implementation of the “loaded implant” model in the rat using a miniaturized setup—description of the method and first results. Clin. Oral Implants Res., 0. Wiskott, H.W.A., Cugnoni, J., Scherrer, S.S., Ammann, P., Botsis, J., Belser, U.C., 2008. Bone reactions to controlled loading of endosseous implants: a pilot study. Clin. Oral Implants Res. 19, 1093–1102. Yushkevich, P.A., Piven, J., Hazlett, H.C., Smith, R.G., Ho, S., Gee, J.C., Gerig, G., 2006. User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage 31, 1116–1128. Zacchetti, G., Wiskott, A., Cugnoni, J., Botsis, J., Ammann, P., 2013. External mechanical microstimuli modulate the osseointegration of titanium implants in rat tibiae. BioMed Res. Int., 2013.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Influence of gait loads on implant integration in rat tibiae: Experimental and numerical analysis Marco Piccinini a, Joel Cugnoni a,n, John Botsis a, Patrick Ammann b, Anselm Wiskott c a
Laboratory of Applied Mechanics and Reliability Analysis, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland Division of Bone Diseases, Department of Internal Medicine Specialities, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland c Division of Fixed Prosthodontics and Biomaterials, School of Dental Medicine, University of Geneva, Geneva, Switzerland b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 25 August 2014
Implanted rat bones play a key role in studies involving fracture healing, bone diseases or drugs delivery among other themes. In most of these studies the implants integration also depends on the animal daily activity and musculoskeletal loads, which affect the implants mechanical environment. However, the tissue adaption to the physiological loads is often filtered through control groups or not inspected. This work aims to investigate experimentally and numerically the effects of the daily activity on the integration of implants inserted in the rat tibia, and to establish a physiological loading condition to analyse the peri-implant bone stresses during gait. Two titanium implants, single and double cortex crossing, are inserted in the rat tibia. The animals are caged under standard conditions and divided in three groups undergoing progressive integration periods. The results highlight a time-dependent increase of bone samples with significant cortical bone loss. The phenomenon is analysed through specimen-specific Finite Element models involving purpose-built musculoskeletal loads. Different boundary conditions replicating the post-surgery bone–implant interaction are adopted. The effects of the gait loads on the implants integration are quantified and agree with the results of the experiments. The observed cortical bone loss can be considered as a transient state of integration due to bone disuse atrophy, initially triggered by a loss of bone–implant adhesion and subsequently by a cyclic opening of the interface. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Rat Tibia Finite element Musculoskeletal loads Implant integration Bone loss
1. Introduction To optimize biological resource allocation, bones are maintained in a state of structural balance between applied loads and mechanical resistance. Under this premise, reduced mobility (Globus et al., 1984; Maïmoun et al., 2012) or strong exercise (Fujie et al., 2004) alter a skeleton’s muscular environment and the bones adapt to the newly established stress fields. After a period of structural remodelling, a new equilibrium is established (Frost, 1990b). Similarly, endosseous implants modify the stress field within the bone bed and thus induce a tissue adaption that may also detrimentally affect the osseous casing, such as in stress shielding (Huiskes et al., 1992), where the bone is actually “underloaded” and will resorb. In spite of their reduced size, rats are applicable to studies on implant biomechanics (Stadlinger et al., 2012; Dayer et al., 2006).
n Correspondence to: Laboratory of Applied Mechanics and Reliability Analysis (LMAF), Ecole Polytechnique Fédérale de Lausanne, Station 9, CH-1015 Lausanne, Switzerland. Tel.: þ 41 21 693 59 73; fax: þ41 21 693 73 40. E-mail addresses: marco.piccinini@epfl.ch (M. Piccinini), joel.cugnoni@epfl.ch (J. Cugnoni), john.botsis@epfl.ch (J. Botsis), [email protected] (P. Ammann), [email protected] (A. Wiskott).
http://dx.doi.org/10.1016/j.jbiomech.2014.08.023 0021-9290/& 2014 Elsevier Ltd. All rights reserved.
For instance, rats have been used to assess the effects of systemic diseases or metabolic disturbances on implant osseointegration (Fiorellini et al., 1999; Glösel et al., 2010; Viera-Negrón et al., 2008; Ammann et al., 2007; Dayer et al., 2010). Rats were also employed when evaluating biocompatible coatings (Hara et al., 1999; Jaatinen et al., 2011), drug delivery systems (Peter et al., 2005; Stadelmann et al., 2009) and locking plates in studies on fracture healing (Histing et al., 2011). Problematically though, the histological process of integration always proceeds within the environment of mechanical stimuli generated during the animals’ daily activity. Indeed when the implants are placed on limbs and the animals are left unconstrained, locomotion will affect peri-implant bone healing and remodelling (Frost, 1990a). Typically, these sways are accounted for by establishing test- and control groups that account for both the animals’ genetic biovariability and their everyday musculoskeletal activation. In most instances, this approach suffices because the analyses are based on comparisons between large groups so that the effects on the outcome parameters can be extracted. In some instances though, the rats’ daily activity strongly interferes with the implants’ experimental loading and thus precludes a pertinent analysis of the research issue under scrutiny. To extract relevant information from such data sets, these
3256
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
situations require a thorough assessment of the stress systems developed during normal daily life. In humans, this theme is approached through inverse dynamics musculoskeletal analysis and Finite Element (FE) models accounting for gait loads (Phillips, 2009; Phillips et al., 2007). Concerning rats, some evidence was provided in studies on rat stride lengths, frequencies and ground reaction forces (Clarke, 1991; Clarke and Parker, 1986; Muir and Whishaw, 1999). But still, there is a considerable research deficit regarding the quantification of the forces developed by the musculature and the loads generated on the joints during the rat gait. To our knowledge, the only work addressing this issue dealt with the internal moments and forces acting on the femur of a walking rat (Wehner et al., 2010). The objective of the present study, therefore, was twofold. (i) To develop a framework of the physiological loading conditions acting on rat tibiae during gait. (ii) To assess the interplay between the proposed framework and the bone’s reactions around titanium implants inserted in rat tibiae.
2. Material and methods This work belongs to a series of investigations on the dependency of implants’ stability on the mechanical environment. According to the hypothesis of the “loaded implant” model (Wiskott et al., 2008, 2011; Zacchetti et al., 2013), two extracutaneous implants are inserted in the proximal part of the rat tibia, and their integration evolves in relation to two mechanical stimulations: the animal activity and the controlled load introduced by activating the implants’ heads. The present work is focused on the effects of the former. All animal experiments were approved by the University of Geneva animal rights committee and supervised by the local veterinary board. 2.1. Animal experiment The study was conducted on 30 Sprague-Dawley rats. The animals were 24-weeks old at the onset of the experiment. Two titanium implants were inserted into the proximal part of the right tibia with their heads protruding from the skin. As shown in Fig. 1, the distal implants were anchored bicortically while the proximal implants only penetrated the medulla by ca. 3 mm, thereby replicating single- and bi-cortical anchorage modes (Ivanoff et al., 2000). The endosseous bore was dimensioned to produce an interference of 0.02 mm with the implant cylinders, hence ensuring implant stability during the early days of healing. Both implants had been previously sandblasted and acid-etched to improve osseointegration (Cochran et al., 1996). After surgery, the animals were caged under standard conditions, with free access to food and water, and divided into three experimental groups undergoing different integration periods: 2, 4 and 6 weeks. Animals were daily inspected to monitor the implantation conditions. No infections (i.e. purulent emissions) were detected. Transient inflammations were healed by treating them with disinfectants. After the sacrifice, the implanted tibiae were dissected, cleared of their soft tissue coverage and frozen to 21 1C. 2.2. Computed tomography processing Prior to computed tomography (CT) scanning, the specimens were maintained at 4 1C for 24 h in a 0.9% solution of NaCl. Then, they were gradually brought to room temperature. The specimens were analysed using a high resolution CT imaging system (μCT-40, Scanco Medical AG, Brüttisellen, Switzerland). For each specimen, 1022 slices
were generated per 360 degrees of rotation using the following settings: isotropic voxel size: 20 μm, source potential: 70 kVp, tube current: 114 μA, integration time: 320 ms, beam hardening correction: 200 mg HA/cm3.
2.3. Morphologic analysis The CT images of all the specimens were imported into ITKsnap (Yushkevich et al., 2006) and inspected by a CT specialist who was unaware of the specimens’ experimental groups. The morphology of the peri-implant cortical bone was classified as follows. (i) ‘Flat’ when complete bone-to-implant adhesion was observed and the tibia’s normal cortical geometry was preserved. (ii) ‘Conic’ if the zone of peri-implant adhesion was contiguous to a funnel-shaped defect of cortical bone. In this instance, the depth of the cone, Δc, was recorded. (iii) ‘Open’ whenever a gap had opened between the cortical bone and the implant. Two Regions of Interest were assessed: the peri-implant bone tissue surrounding the distal (ROId) and proximal (ROIp) implants. The morphologic features, the conic depth and the inspected ROIs are highlighted in Fig. 2.
2.4. Gait based loading model The joints and muscle loads acting on the rat tibia during gait are estimated through the equilibrium of the femur and tibia rigid body models. The musculoskeletal geometry (Johnson et al., 2008) and forces (Wehner et al., 2010) are combined to establish the equilibrium equations systems. The static equilibrium is calculated at a time-step corresponding to 35% of the gait cycle (Wehner et al., 2010), i.e. when most of the muscles reach their maximum force. First, the condylar reactions acting on the femur were calculated by equilibrating a system that included (i) the femur’s internal forces and moments, (ii) the lateral and medial gastrocnemius and (iii) the loads developed in the patella taken as the resultant of pulls by the m. vastus lateralis, the m. vastus medialis and the m. rectus femoris, all acting along the bisector lines of their origin-via-insertion coordinates (Fig. 3b). The coordinates of the condylar joint reactions were compatible with the contact areas on the femur’s condyles (Dao et al., 2011). Then the tibial equilibrium system was established. It comprised the calculated condylar reactions and the forces developed by the m. vastus lateralis and medialis, the m. rectus femoris and the m. tibialis anterior. The m. biceps femoris was treated separately. Indeed although it has been demonstrated as being active during the stance phase (Gillis and Biewener, 2001), there is a lack of information regarding the force generated by this muscle during gait. Hence, the tibial equilibrium equations were solved while treating the m. biceps femoris and a single ankle reaction acting on the inter-malleolar point as unknowns (Fig. 3c). The coordinates of the muscular attachments and the calculated loads are listed in Table 1.
2.5. Finite element models and boundary conditions. A continuum FE model of a rat tibia was generated from available micro-CT data through the procedure previously developed by the authors (Piccinini et al., 2012). First, the CT images were processed with the software ITKsnap (Yushkevich et al., 2006). Bone was segmented according to grayscale thresholds and a 3D representation of the rat tibia was obtained. Second, the CT images were processed through the open-source FE model generator VoxelMesher. More specifically, the segmentation was extracted as a binary image and its boundary was discretized using a triangle meshing algorithm for 3D implicit surfaces (Boissonnat and Oudot, 2005). The surface mesh was embedded in a bounding box, where seed nodes were generated using a recursive octree decomposition of the domain. Using the bounding box, the surface mesh and the seed nodes as input, an initial constrained Delaunay tetrahedral mesh was generated using TetGen (Si, 2008). After deleting the undesired domains, a high-quality tetrahedral mesh was calculated by refining the initial one using a local element size map. Then, the linear tetrahedra were
Fig. 1. (a) 3D representation of the proximal part of a rat tibia: the insertion points of the distal and proximal implant are represented. (b) View cut showing a schematic representation of the different implantation strategies: bi-cortical (distal implant) and mono-cortical (proximal implant).
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
3257
Fig. 2. mCT scans of an implanted specimen: (a) distal, ROId, and proximal, ROIp, region of interest, (b) flat morphology, (c) conic morphology and conic depth, Δc, and (d) open morphology.
Fig. 3. (a) Simplified representation of the rat hindlimb musculoskeletal system; muscles are represented as lines connecting their origin-via-insertion (Johnson et al., 2008). (b) Equilibrium of the femur: the medial and lateral condylar reactions, Cm and Cl, are calculated equilibrating the internal forces Fi and moments Mi, the patellar load Pi and the lateral and medial gastrocnemius, Gl and Gm. (c) Equilibrium of the tibia: the ankle joint reaction Aj and the bicep femoris load Bf are estimated equilibrating the vastus medialis Vm and lateralis Vl, the rectus femoris Rf, the tibialis anterior proximal TAp and distal TAd and the lateral and medial condylar reactions Cm and Cl previously calculated. converted to quadratic elements by adding mid-side nodes. Eventually, each element was assigned a modulus of elasticity in accordance with a 3D map of elastic modulus (Taddei et al., 2007), obtained by converting the grey value of each voxel of the original CT images into a corresponding modulus of elasticity, through the relationship developed by Cory et al. (2010). Thus, the tibia mechanical behaviour was characterized by an isotropic, continuum and inhomogeneous material field that preserves the bone internal structure. Then, the implants were inserted in-silico into the tibia proximal segment whereby the distal implant’s threaded end was reduced to a cylinder with tied contact to the cortical bone. As the integration progresses during the healing period, the mechanics of the implant–bone interface evolves from a press-fit condition to a simple adhesion. To simulate this phenomenon, three boundary conditions were analysed. In a first model, the implants were “press-fitted” into their bone bed—a condition which was modelled as a homogeneous radial displacement field of 0.01 mm that was applied to the cylindrical surfaces of the implants (Natali et al., 2009). In a second FE model, bone is
tied to the implant surfaces without press fit, while in the third one the opening of the bone–implant interface is allowed by considering a hard, frictionless contact condition between the bone and the implant without interference. All four models, including that which was bare of implants, were subjected to the gait-based loading condition previously defined. The essentials of all four models are summarized in Table 2. Computations were performed through the ABAQUS-Standard solver.
3. Results 3.1. Morphologic analysis The outcomes of the morphologic analyses are shown in Table 3, all implants integrated. After two weeks of healing, conical bone loss was
3258
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Table 1 Loads acting on the tibia, calculated with respect to the reference system in Fig. 3. Structure
Rectus femorisa,b Vastus lateralisa,b Vastus medialisa,b Tibialis anterior proximala,b Tibialis anterior distala,b Biceps femorisb Lateral condyle Medial condyle Ankle jointb
Coordinates (mm)
Force (N)
x
y
z
x
y
z
39.70 39.70 39.70 39.00 4.27 33.00 40.10 39.90 0.00
0.00 0.00 0.00 2.29 0.00 0.50 2.00 2.00 0.00
2.00 2.00 2.00 1.50 0.86 0.20 1.90 1.10 0.00
5.45 2.33 1.95 2.83 1.61 0.00 12.75 0.03 4.20
0.22 2.27 0.88 0.24 0.33 3.63 0.38 0.38 1.17
3.94 2.37 1.41 0.01 2.57 5.98 7.56 5.84 2.26
Table 3 Results of the morphologic analysis: percentage of specimens characterized by flat, conic or open peri-implant morphology after two, four and six weeks of integration. Integration period
Feature
ROId (%)
ROIp (%)
2 weeks
Flat Conic Open Flat Conic Open Flat Conic Open
79.3 14.0 6.7 50.0 33.3 16.7 41.6 41.7 16.7
80.0 20.0 0.0 65.0 35.0 0.0 55.0 45.0 0.0
4 weeks
6 weeks
a Loads magnitudes from Wehner et al. (2010). The x-, y- and z-components are calculated as the projection to the insertion–origin direction. b Coordinates from Johnson et al. (2008).
Table 2 Details of the boundary conditions adopted in the FE models. The contact conditions concern ROId and ROIp. The threaded tip of the distal implant is simplified with a tie contact. Model
Implants
Bone–implant adhesion
Press Fit
Gait loads
Bare Press Fit Adhesion Opening
No Yes Yes Yes
– Yes Yes Noa
– Yes No No
Yes Yes Yes Yes
a
Frictionless hard contact.
observed in 14% of ROIds and 20% of ROIps. After six weeks, the figures were 41% and 45%, respectively. An ‘open’ condition was only observed on the distal implants (17% after 6 weeks). None of the proximal implants presented this characteristic. In all specimens, the threaded tip of the distal implant was well integrated into cortical bone. Fig. 4 lists the cone depths after 2, 4 and 6 weeks of healing. As shown, the mean cone depth increased with time and finally reached values approaching that of the thickness of the cortical bone ( 0.5 mm). The increment between 2 and 6 weeks was significant for both ROId and ROIp although for ROIp no major increase was observed between 4 and 6 weeks. 3.2. Finite element analysis When the bare tibia model was subjected to nominal gait loads, the strain pattern developed as a coupling of bending and compression typical of long bones (Fig. 5). The tensile and compressive longitudinal strains in the midshaft worked out to 1112 με and 1318 με, respectively. During the early days after surgery, the deformation state occurring in the FE model with in-silico implants blend the effects of the gait loads and the press fit, generating the peri-implant radial stress represented in Fig. 6a and b, concerning the cortical tissue. The radial compression pattern generated by the press-fit surpasses the tensions developed during gait loading, leading to normal compressive stresses in the cortical contact areas in both proximal and distal implants. In the next phase of integration, the effects of press-fitting are released to represent bone adaptation and only the gait loads act on the implanted tibia (Fig. 6c and d). The radial stresses reach peak values of tension up to 50 MPa (Fig. 6e and f) in the distal–proximal direction. Note that the radial stress-state is not uniform and evolves from large tensile stresses in the distal–proximal direction to negligible compressive stresses in the medio-lateral direction. Fig. 7 depicts the response of non-adhesive bone-to-implant interfaces in terms of mean contact pressures and openings that
Fig. 4. Conic depth measured in specimens presenting a cortical bone loss. Mean and standard deviation are calculated on groups (n ¼10) sacrificed after integration periods of 2, 4 and 6 weeks ( p o 0.05, t-test).
develop at the implant surface. Under this condition, gait loads result in compressive stresses on the implants’ circumference with exception of the superficial zones in which tension develops in the longitudinal direction and the bone detaches from the implant. This phenomenon was observed in both ROIs but was more pronounced in ROId. In ROIp no clear-cut distinction between zones of pressure and opening was possible due to irregularities in thickness of the cortical bone and its partial entanglement with trabecular tissue. Last, Fig. 8 shows the stresses occurring in the longitudinal plane in case of opening depths corresponding to 30% and 60% of the cortical thickness with a perfectly adherent implant–bone interface. The bone tissue in the proximity of the open contact is clearly unloaded, while high tensile stress concentrations characterize the area immediately below the opening zone. The stress concentrations affecting the distal implant surpass 20 MPa and are relatively independent of the opening depth (Fig. 8a and b). Contrarily, the stress concentrations are much less pronounced around the proximal implant (Fig. 8c and d) and show sign of a progressive reduction with the increase of conic depth.
4. Discussion The present experimental–numerical analysis aimed at relating the cortical bone defects often observed in implanted rat tibiae to local stress patterns and to the ensuing osseous remodelling. To this end, finite element models of tibiae were generated and subjected to a loading environment that simulated the forces generated during the animals’ daily activities and more specifically during gait. The pertinence of the numerical model was verified to match previously published values at comparable locations. For instance, the strains calculated for the bare (i.e. devoid of implants) tibia’s
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
3259
Fig. 5. Finite Element analysis of the bare rat tibia. (a) Longitudinal strains, (b) maximum and minimum longitudinal strains along the tibial midshaft. The fibula is not considered. The value in the dotted box agrees with the in-vivo measurements by Rabkin et al. (2001).
Fig. 6. Finite Element analysis of the rat tibia with in-silico implants. A full bone–implant adhesion is assumed. Cortical radial stress considering both the gait loads and the press fit in (a) ROIp and (b) ROId. Cortical radial stress considering only the gait loads in (c) ROIp and (d) ROIp. Stresses along the implant circumference considering both press fit and gait loads (blue) and only gait loads (red) in (e) ROIp and (f) ROId. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
midshaft are in close agreement with those generated in rat long bones during locomotion: 600–1200 με (Turner et al., 1994; Hillam and Skerry, 1995; Mosley et al., 1997). Similarly the longitudinal strain illustrated in Fig. 5b agrees with in-vivo measurements at the same location (Rabkin et al., 2001).
Early post-surgery, the stresses arising in the peri-implant bone bed blend the compression due to the implants’ press fit and the tensile stresses generated during the animals’ locomotion (Fig. 6a and b). As the first one is initially predominant, the compressive forces impart their stability to the implants, prevent
3260
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Fig. 7. Results of the contact analysis in case of frictionless hard contact. (a) Contact pressure and opening in ROId, (b) contact pressure and opening in ROIp. Pressure and opening mean and standard deviation are calculated on the implant circumference.
Fig. 8. Longitudinal stress in case of partial cortical loss of adhesion. Distal implant with debonding depth equal to (a) 30% of cortical thickness and (b) 60% of cortical thickness. Proximal implant with debonding depth equal to (c) 30% of cortical thickness and (d) 60% of cortical thickness.
harmful micro-motions and thus facilitate initial integration (Abdul-Kadir et al., 2008). This stabilizing effect though, comes to an end 3–4 days after implant placement due to the resilience and remodelling of the surrounding osseous tissue (Dhert et al., 1998). Now the peri-implant bone’s evolution depends on the rats’ physical activity, the adhesive strength of the bone–implant interface and amplitude of micromotions in the bone–implant cavity. As shown in Fig. 6c and d, when press-fit is released, the interface is subjected to severe radial tensile stresses at each gait cycle.
Problematically, these values (20–50 MPa) exceed the adhesive strength between the implant and the surrounding bone (1– 4 MPa) (Gross et al., 1987; Takatsuka et al., 1995). It follows that the loss of bone-to-implant contact as observed experimentally may well be initiated by a loss of adhesion due to the excess tensile stresses generated at each gait cycle. The affected bone now loses its stimulation and reacts by migrating in apical direction along the implant. A similar bone shape is observed around overloaded implants inserted in dog and
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
rabbit bones (Hoshaw et al., 1990; Duyck et al., 2001). Clinically, this phenomenon is similar to the marginal bone loss observed around dental implants (Qian et al., 2012), which has been attributed either to disease (Bodic et al., 2005) or to overload (Isidor, 1996). Interestingly, neither diseases nor overloading were overtly active in the present study; hence a different mechanism driving cortical bone loss must be postulated (Fig. 8). According to the present FE analysis, permanent gaps may develop at the implant–bone interface. If this occurs, an irreversible process is initiated as the bone tissue that is separated from the implant is in a condition of ‘stress shielding’ and is resorbed because of disuse (Engh et al., 1987). Moreover, once the process of gap formation is initiated, the stresses generated during each gait cycle are relocated further down the implant and a new zone of the bone–implant interface is subjected to tensile forces (Fig. 8). Locomotion thus causes the inward propagation of a gap along the implant surface with the ancillary effect of resorption of unloaded bone and the eventual formation of a funnel-shaped defect in the cortical bone. The stability of this progressive process is controlled at least partially by the evolution of the peak tensile stresses at the bone–implant interface and thus depends on the local bone structure and implantation strategy. Indeed, this mechanism was observed in both ROIs but the differing bony environment and implantation strategies led to different defect morphologies. The tip of the distal implant was screwed-tightened into the opposing cortex while the upper part was inserted in a thick layer of cortical bone without trabecular bone in between. Due to an extended contact area, this configuration yielded an excellent initial resistance to the loss of adhesion (lower initial stresses). Clinically, it took 4 weeks for the funnel in ROId to start to increase (Fig. 4), thus indicating a high initial resistance of the interface. However, once debonding occurred, this configuration demonstrated a fast development of a funnel shape between 4 and 6 weeks (Fig. 4). The high stress concentrations seen in the FE simulations (Fig. 8a and b) and the fact that their magnitude does not decrease when debonding propagates explains the high rate of development of conic depth and the presence of an open gap in 17% of the distal implants after 6 weeks. By contrast, the proximal implant was inserted into a thin layer of cortical bone and large amounts of trabecular bone underneath. As shown by the conic depth evolution (Fig. 4) and the relatively high percentage of conic features (Table 3) between 2 and 4 weeks, this environment seems to offer less initial resistance to periosteal tension and promotes early debonding at the outer cortical surface. However, this implantation strategy demonstrated a better
3261
adaptation to the mechanical environment after the early stages, as the rate of evolution of the conic features was much less than in ROId after 4 weeks and no open gaps were observed around the proximal implants after 6 weeks. Indeed, as shown in Fig. 8, the smooth gradient of stiffness offered by the underlying trabecular bone helped to progressively decrease the interfacial stress concentration with increasing conic depth, and thus lead to an arrest of the gap formation process. Moreover the results show that the rate of densification of trabecular bone was sufficiently high to produce a dense osseous shell with a funnel shape around the proximal implant. Thus, the absence of trabecular bone around the distal implant could be seen as a factor limiting the adaptation capacity and facilitating the formation of a fully open gap. Within the observed time period, it thus appears that large contacts zones in the cortex delay the bone–implant opening but cannot prevent the development of fully open gaps, while trabecular bone hampers the debonding and prevents the complete failure of the interface. Still, whether an animal will – or will not – demonstrate periimplant bone loss cannot be predicted. Indeed a number of individual factors are contributive: the degree of press-fit relaxation, the adhesive strength and the size of the bone–implant contact, the morphology and material properties of the surrounding bone tissue as well as the rats’ activity during the post-surgical period (different levels of pain might prevent the animal to use the implanted leg and thus reduce the number of load applications). Thus, three configurations are possible: (1) no massive debonding occurs because press fit is retained sufficiently long to allow bone in-growth and stability of the implant, (2) a local bone–implant debonding is initiated on the outer cortical surface which finally stops at a certain depth (3) an unstable evolution of debonding/ bone resorption which leads to an open gap between implant and bone. Compared to previous studies, it should be noted here that the proposed bone resorption mechanism can only occur in implantation sites in which tensile stresses are dominant during physiological activities. Other causes may lead to similar looking cortical bone loss such as overloading (Isidor, 1996; Hoshaw et al., 1990; Duyck et al., 2001) or diseases (Bodic et al., 2005). Due to the use of transcutanesous implant, the presence of debris and local inflamation or infections cannot be completely ruled out from the study. However, based on the daily inspection in the present experiment, no external signs of purulent infections were noticed and only a negligible number of animals developed local inflammations which were efficiently treated by disinfectant. According to the experience gained so far in the loaded-implant animal
Fig. 9. Simplified representation of the mechanisms leading to the cortical bone loss. (a) In the early post surgery days the implant is stable because of the compressive stress field due to the press fit. (b) When the press fit is released part of the interface is subjected to traction. (c) The cyclic loading provokes a loss of bone–implant adhesion in the periosteal area. The stress distribution changes and part of the tissue is unloaded (i.e. above the dotted line). (d) The unloaded tissue is resorbed and the opening propagates along the implant axe. (e) If no stable configurations exist the opening reaches the endosteum and the bone–implant contact is totally lost.
3262
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
model in rat tibiae (more than 160 animals, with the same implantation), the observed occurence of purulent infections (1.2%) and inflammations (16%) is much lower than the occurence of conic cortical bone loss observed in the present study (40%). Thus, the conic shaped cortical bone loss observed in this work is predominantly attributed to the effect of tensile stresses and progressive debonding–bone resorption mechanisms as described in Fig. 9.
5. Conclusions In this study, a model of a rat’s tibia subjected to musculoskeletal loads was constructed, yielding a realistic representation of the deformation fields that develop in the tibia during gait. This model enhances our knowledge of the biomechanics of rat tibiae under physiological conditions and sets the baseline according to which implantation studies on rat tibiae are to be evaluated. With respect to the ‘loaded implant’ model, it now appears that normal locomotion may tear the adhesive interface between the implant and the bone to the extent that a gap opens or at least the contact area is drastically reduced. The observed funnel-shaped bone loss is caused by bone disuse atrophy which is initiated by a loss of bone–implant adhesion and kept ongoing by the cyclic loadings on the interface due to gait cycles. Thus, in the “loaded implant” model, rat locomotion detrimentally affects implant integration. Other type of implantation, for example in dentistry, may be affected by a similar detrimental mechanism affecting the implant stability. It can thus be recommended to consider not only the external forces but also the effects of daily physiological loads when studying the implants integration.
Conflict of interest statement The authors declare no conflict of interest.
Acknowledgements Mrs Severine Clement and Mrs. Isabelle Badoud (UNIGE) are gratefully acknowledged for their work on surgery, animal care, and data acquisition. This work was partially funded by Swiss National Science Fundation (SNF) project 315230-127612. References Abdul-Kadir, M.R., Hansen, U., Klabunde, R., Lucas, D., Amis, A., 2008. Finite element modelling of primary hip stem stability: the effect of interference fit. J. Biomech. 41, 587–594. Ammann, P., Badoud, I., Barraud, S., Dayer, R., Rizzoli, R., 2007. Strontium ranelate treatment improves trabecular and cortical intrinsic bone tissue quality, a determinant of bone strength. J. Bone Miner. Res. 22, 1419–1425. Bodic, F., Hamel, L., Lerouxel, E., Baslé, M.F., Chappard, D., 2005. Bone loss and teeth. Joint Bone Spine 72, 215–221. Boissonnat, J.D., Oudot, S., 2005. Provably good sampling and meshing of surfaces. Graphical Models 67, 405–451. Clarke, K.A., 1991. Swing time changes contribute to stride time adjustment in the walking rat. Physiol. Behav. 50, 1261–1262. Clarke, K.A., Parker, A.J., 1986. A quantitative study of normal locomotion in the rat. Physiol. Behav. 38, 345–351. Cochran, D.L., Nummikoski, P.V., Higginbottom, F.L., Hermann, J.S., Makins, S.R., Buser, D., 1996. Evaluation of an endosseous titanium implant with a sandblasted and acid-etched surface in the canine mandible: radiographic results. Clin. Oral Implants Res. 7, 240–252. Cory, E., Nazarian, A., Entezari, V., Vartanians, V., Müller, R., Snyder, B.D., 2010. Compressive axial mechanical properties of rat bone as functions of bone volume fraction, apparent density and micro-ct based mineral density. J. Biomech. 43, 953–960. Dao, T.T., Pouletaut, P., Goebel, J.C., Pinzano, A., Gillet, P., Ho Ba Tho, M.C, 2011. In vivo characterization of morphological properties and contact areas of the rat cartilage derived from high-resolution MRI. IRBM 32, 204–213.
Dayer, R., Brennan, T.C., Rizzoli, R., Ammann, P., 2010. PTH improves titanium implant fixation more than pamidronate or renutrition in osteopenic rats chronically fed a low protein diet. Osteoporos. Int. 21, 957–967. Dayer, R., Rizzoli, R., Kaelin, A., Ammann, P., 2006. Low protein intake is associated with impaired titanium implant osseointegration. J. Bone Miner. Res. 21, 258–264. Dhert, W.J.A., Thomsen, P., Blomgren, A.K., Esposito, M., Ericson, L.E., Verbout, A.J., 1998. Integration of press-fit implants in cortical bone: a study on interface kinetics. J. Biomed. Mater. Res. 41, 574–583. Duyck, J., Rønold, H.J., Van Oosterwyck, H., Naert, I., Sloten, J.V., Ellingsen, J.E., 2001. The influence of static and dynamic loading on marginal bone reactions around osseointegrated implants: an animal experimental study. Clin. Oral Implants Res. 12, 207–218. Engh, C.A., Bobyn, J.D., Glassman, A.H., 1987. Porous-coated hip replacement. The factors governing bone ingrowth, stress shielding, and clinical results. J. Bone Joint Surg. Ser. B 69, 45–55. Fiorellini, J.P., Nevins, M.L., Norkin, A., Weber, H.P., Karimbux, N.Y., 1999. The effect of insulin therapy on osseointegration in a diabetic rat model. Clin. Oral Implants Res. 10, 362–368. Frost, H.M., 1990a. Skeletal structural adaptations to mechanical usage (SATMU): 1. Redefining Wolff’s law: the bone modeling problem. Anat. Rec. 226, 403–413. Frost, H.M., 1990b. Skeletal structural adaptions to mechanical usage (SATMU): 2. Redefining Wolff’s law: the remodeling problem. Anat. Rec. 226, 414–422. Fujie, H., Miyagaki, J., Terrier, A., Rakotomanana, L., Leyvraz, P.F., Hayashi, K., 2004. Detraining effects on the mechanical properties and morphology of rat tibiae. Bio-Med. Mater. Eng. 14, 219–233. Gillis, G.B., Biewener, A.A., 2001. Hindlimb muscle function in relation to speed and gait: in vivo patterns of strain and activation in a hip and knee extensor of the rat (Rattus norvegicus). J. Exp. Biol. 204, 2717–2731. Globus, R.K., Bikle, D.D., Morey-Holton, E., 1984. Effects of simulated weightlessness on bone mineral metabolism. Endocrinology 114, 2264–2270. Glösel, B., Kuchler, U., Watzek, G., Gruber, R., 2010. Review of dental implant rat research models simulating osteoporosis or diabetes. Int. J. Oral Maxillofac. Implants 25, 516–524. Gross, U., Roggendorf, W., Schmitz, H.J., Strunz, V., 1987. Biomechanical and Morphometric Testing Methods for Porous and Surface-reactive Biomaterials. ETATS-UNIS, American Society for Testing and Materials, Philadelphia, PA. Hara, T., Hayashi, K., Nakashima, Y., Kanemaru, T., Iwamoto, Y., 1999. The effect of hydroxyapatite coating on the bonding of bone to titanium implants in the femora of ovariectomised rats. J. Bone Joint Surg. Ser. B 81, 705–709. Hillam, R.A., Skerry, T.M., 1995. Inhibition of bone resorption and stimulation of formation by mechanical loading of the modeling rat ulna in vivo. J. Bone Miner. Res. 10, 683–689. Histing, T., Garcia, P., Holstein, J.H., Klein, M., Matthys, R., Nuetzi, R., Steck, R., Laschke, M.W., Wehner, T., Bindl, R., Recknagel, S., Stuermer, E.K., Vollmar, B., Wildemann, B., Lienau, J., willie, B., Peters, A., Ignatius, A., Pohlemann, T., Claes, L., Menger, M.D., 2011. Small animal bone healing models: standards, tips, and pitfalls results of a consensus meeting. Bone 49, 591–599. Hoshaw, S.J., Brunski, J.B., Cochran, G.V.B., Higuchi, K.W., 1990. Theories of Bone Modeling and Remodeling in Response to Mechanical Usage. Experimental Investigation of an in vivo Bone–Implant Interface. 391-394. Huiskes, R., Weinans, H., Van Rietbergen, B., 1992. The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res., 124–134. Isidor, F., 1996. Loss of osseointegration caused by occlusal load of oral implants: a clinical and radiographic study in monkeys. Clin. Oral Implants Res. 7, 143–152. Ivanoff, C.J., Gröndahl, K., Bergström, C., Lekholm, U., Brånemark, P.I., 2000. Influence of bicortical or monocortical anchorage on maxillary implant stability: a 15-year retrospective study of Brånemark system implants. Int. J. Oral Maxillofac. Implants 15, 103–110. Jaatinen, J.J.P., Korhonen, R.K., Pelttari, A., Helminen, H.J., Korhonen, H., Lappalainen, R., Kröger, H., 2011. Early bone growth on the surface of titanium implants in rat femur is enhanced by an amorphous diamond coating. Acta Orthop. 82, 499–503. Johnson, W.L., Jindrich, D.L., Roy, R.R., Reggie Edgerton, V., 2008. A threedimensional model of the rat hindlimb: musculoskeletal geometry and muscle moment arms. J. Biomech. (41), 610–619. Maïmoun, L., Brennan-Speranza, T.C., Rizzoli, R., Ammann, P., 2012. Effects of ovariectomy on the changes in microarchitecture and material level properties in response to hind leg disuse in female rats. Bone 51, 586–591. Mosley, J.R., March, B.M., Lynch, J., Lanyon, L.E., 1997. Strain magnitude related changes in whole bone architecture in growing rats. Bone 20, 191–198. Muir, G.D., Whishaw, I.Q., 1999. Ground reaction forces in locomoting hemiparkinsonian rats: a definitive test for impairments and compensations. Exp. Brain Res. 126, 307–314. Natali, A.N., Carniel, E.L., Pavan, P.G., 2009. Dental implants press fit phenomena: biomechanical analysis considering bone inelastic response. Dent. Mater. 25, 573–581. Peter, B., Pioletti, D.P., Laïb, S., Bujoli, B., Pilet, P., Janvier, P., Guicheux, J., Zambelli, P.Y., Bouler, J.M., Gauthier, O., 2005. Calcium phosphate drug delivery system: influence of local zoledronate release on bone implant osteointegration. Bone 36, 52–60. Phillips, A.T.M., 2009. The femur as a musculo-skeletal construct: a free boundary condition modelling approach. Med. Eng. Phys. 31, 673–680. Phillips, A.T.M., Pankaj, P., Howie, C.R., Usmani, A.S., Simpson, A.H.R.W., 2007. Finite element modelling of the pelvis: inclusion of muscular and ligamentous boundary conditions. Med. Eng. Phys. 29, 739–748.
M. Piccinini et al. / Journal of Biomechanics 47 (2014) 3255–3263
Piccinini, M., Cugnoni, J., Botsis, J., Zacchetti, G., Ammann, P., Wiskott, A., 2012. Factors affecting subject-specific finite element models of implant-fitted rat bone specimens: critical analysis of a technical protocol. Comput. Methods Biomech. Biomed. Eng.. Qian, J., Wennerberg, A., Albrektsson, T., 2012. Reasons for marginal bone loss around oral implants. Clin. Implant Dent. Relat. Res. 14, 792–807. Rabkin, B.A., Szivek, J.A., Schonfeld, J.E., Halloran, B.P., 2001. Long-term measurement of bone strain in vivo: the rat tibia. J. Biomed. Mater. Res. 58, 277–281. Si, H, 2008. Adaptive tetrahedral mesh generation by constrained Delaunay refinement. Int. J. Numer. Methods Eng. 75, 856–880. Stadelmann, V.A., Terrier, A., Gauthier, O., Bouler, J.M., Pioletti, D.P., 2009. Prediction of bone density around orthopedic implants delivering bisphosphonate. J. Biomech. 42, 1206–1211. Stadlinger, B., Pourmand, P., Locher, M.C., Schulz, M.C., 2012. Systematic review of animal models for the study of implant integration, assessing the influence of material, surface and design. J. Clin. Periodontol. 39, 28–36. Taddei, F., Schileo, E., Helgason, B., Cristofolini, L., Viceconti, M., 2007. The material mapping strategy influences the accuracy of CT-based finite element models of bones: an evaluation against experimental measurements. Med. Eng. Phys. 29, 973–979. Takatsuka, K., Yamamuro, T., Nakamura, T., Kokubo, T., 1995. Bone-bonding behavior of titanium alloy evaluated mechanically with detaching failure load. J. Biomed. Mater. Res. 29, 157–163.
3263
Turner, C.H., Forwood, M.R., Rho, J.Y., Yoshikawa, T., 1994. Mechanical loading thresholds for lamellar and woven bone formation. J. Bone Miner. Res. 9, 87–97. Viera-Negrón, Y.E., Ruan, W.H., Winger, J.N., Hou, X., Sharawy, M.M., Borke, J.L., 2008. Effect of ovariectomy and alendronate on implant osseointegration in rat maxillary bone. J. Oral Implantol. 34, 76–82. Wehner, T., Wolfram, U., Henzler, T., Niemeyer, F., Claes, L., Simon, U., 2010. Internal forces and moments in the femur of the rat during gait. J. Biomech. 43, 2473–2479. Wiskott, H.W.A., Bonhote, P., Cugnoni, J., Durual, S., Zacchetti, G., Botsis, J., Scherrer, S.S., Ammann, P., 2011. Implementation of the “loaded implant” model in the rat using a miniaturized setup—description of the method and first results. Clin. Oral Implants Res., 0. Wiskott, H.W.A., Cugnoni, J., Scherrer, S.S., Ammann, P., Botsis, J., Belser, U.C., 2008. Bone reactions to controlled loading of endosseous implants: a pilot study. Clin. Oral Implants Res. 19, 1093–1102. Yushkevich, P.A., Piven, J., Hazlett, H.C., Smith, R.G., Ho, S., Gee, J.C., Gerig, G., 2006. User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage 31, 1116–1128. Zacchetti, G., Wiskott, A., Cugnoni, J., Botsis, J., Ammann, P., 2013. External mechanical microstimuli modulate the osseointegration of titanium implants in rat tibiae. BioMed Res. Int., 2013.