3 Results

3.1 Bone-joint mechanics (humerus project)

▼ 72 

In the first part of the project, the study of intact and fractured proximal humeri under physiological loads, the following results were obtained:

3.1.1 Fixation stiffness in vitro and in the simulation

This model was validated by comparison between the stiffness obtained in vitro with the stiffness calculated in a numerical model. For comparison the obtained in vitro values are rewritten here: seven specimens fixated with the LCP-PH were mechanically tested in vitro in axial compression and torsion. From these measurements the average stiffness in the axial direction was calculated to be 957.5 ± 398.21 N/mm and in torsion to 0.48 ± 0.11 Nm/°. The specimen of poor bone quality selected for finite element analysis had an in vitro axial stiffness of 830.2 N/mm and a torsional stiffness of 0.53 Nm/° and those specimens selected to represent the reference bone had an axial stiffness of 927.2 N/mm and a torsional stiffness of 0.64 Nm/°. The corresponding finite element analyses realized in this work (humerus project) yielded a compressive stiffness of 846.4 N/mm and a torsional stiffness of 0.52 Nm/° for a poor bone quality and 1047.9 N/mm and 0.59 Nm/° for the reference bone. The composite stiffness between experiment (in vitro) and finite element analysis (this work) for the specific bones differed in compression by a maximum of 13% and in torsion by a maximum of 7.8 %.

3.1.2 Straining of intact and fractured proximal humerus under physiological loads

The intact humeri bones were loaded mainly in compression with superimposed bending and torsion (Fig. 3.1).

▼ 73 

Fig. 3.1: Finite element model of the proximal humerus under physiological-like loading conditions. Minimum principal strains of the bone surface are given for the neutral position (left), 90° forward flexion (center) and for 90° abduction (right).

In this figure both intact specimens, the reference bone and poor bone quality were compared under different physiological-like loading conditions.

For the intact situation and 90° abduction the maximal strain values were found in the cortical bone. Strains were further increased in the specimen with poor bone quality under 90° abduction.

▼ 74 

Even though the strains differed in value, the strain patterns were comparable between the neutral position, 90° abduction and 90° forward flexion under physiological-like loading conditions in the defect situation (Fig. 3.2). Comparing the strains obtained in the cortical bone (percentage of elements) for the different simulated physiological cases 90° abduction showed the largest strain values, followed by 90° forward flexion (Fig. 3.3). The lowest strain values were found for the neutral position (Fig. 3.3).

The strain distribution (percentage of elements) in the cancellous bone loaded in 90° abduction, were moderated in the specimen with reference bone quality (8% ≥ 1000 με, Fig. 3.4). The strain values were lower than the limits reported from in vivo data (Frost, 1987;Lanyon, et al., 1975;Lanyon, 1976). For the same physiological loading condition the specimen with poor bone quality showed a significant increment in the number of elements with high strain magnitudes (38% ≥ 1000 με, Fig. 3.4). A similar result was found in the surface strain distribution. For the specimen with the poor quality the strain pattern was similar but noticeably increased in magnitude compared with the specimen with reference bone quality. The modeling of the muscles as single force vectors implies a local overestimation of the strains close to the muscle attachments. This effect was considerably more pronounced in the osteoporotic specimen compared to the reference bone.

Fig. 3.2: Minimum principle strains at the anterior bone surface. Osteoporotic, poor quality bone with a bone defect stabilized by an angle stable osteosynthetic device is shown for three different arm positions: 0°, 90° forward flexion (F) and 90° abduction (A).

▼ 75 

Fig. 3.3: Distribution of cortical bone elements for different strain ranges. Strains are given as minimum principal strain values for different arm positions under physiological-like loading conditions for the specimen with poor bone quality and a 5 mm osteotomy at the level of the surgical neck of the humerus.

The specimens stabilized with a LCP-PH were not strained in a different manner: Strains of the cancellous bone were increased (> 20%) in the fractured situation compared with the intact one. However, the strain values were significantly increased in the specimen with lower bone density (Fig. 3.4). These results were reported by (Maldonado, et al., 2003).

Fig. 3.4: Distribution of cancellous bone elements for different strain ranges. Strains are given as minimum principal strain values during 90° abduction for a specimen with reference bone quality and defect, and a specimen of poor bone quality, both with and without defect. Defects were stabilized using an angle-stable osteosynthetic device.

3.2 Osteochondral healing (Galileo project)

3.2.1 Strain Analysis

▼ 76 

As explained in the material and methods section it was necessary to perform a numerical analysis of the histological sections for developing the tissue differentiation model (Fig. 3.5). Factors for growth and resorption as well as the points of the trilinear curve for each tissue type were calculated. Additionally, strains of the different tissues formed during healing were determined for the first time. The compressive field of strains in the initial defect situation was determined as well.

Fig. 3.5: Histological sections of osteochondral defect healing at 4, 6 and 12 weeks stained with Safranin-Orange von Kossa (top) and Safranin-light green (bottom). 4th week: First cartilage cells at the defect wall as well as a resorption region at the basis are visible. 6th week: the cartilage growths from the lateral borders of the defect. It is calcified forming cancellous bone as far as the subchondral plate. Fibrous tissue fills the defect with low formation of hyaline cartilage. At the center a defect region can be seen. 12th week: the cancellous bone has completely regrown; the fibrous tissue is differentiated to hyaline cartilage (Bail, et al., 2003;Duda, et al., 2005).

3.2.1.1 Histological sections

Using the finite element method histological sections of osteochondral healing were numerically analyzed. Values of minimum principal strains for each tissue type at the three points selected in the animal experimentation were obtained as follows:

▼ 77 

At the 4th week (Fig. 3.6) strain concentrations at the interface cartilage subchondral bone were found. Maximal compressive strain values of –2.0e-2 at the interface cartilage-connective tissue were registered. At the defect center these values decreased to –1.7e-2 in the connective tissue. Strain values of up to –1.2e-2 were obtained at the defect wall in the regions corresponding to the places of first cartilage differentiation. At the defect basis compressive strains values between –0.8e-2 and –1.0e-2 were achieved in the connective tissue after subchondral bone resorption.

Fig. 3.6: Left: histology at the 4th week, stained with Safranin-Orange von Kossa. Right: Straining of the formed tissues during healing in a selected histological section at the 4th week.

At the 6th week (Fig. 3.7) the first cartilage regions grew and reached the defect basis in a centripetal filling configuration. The minimum principal strains at the defect wall were then reduced by up to 15 %. The minimum principal strains at the defect center were reduced from -1.7e-2 to –1.2e-2. In the transition region between the defect center and the cartilage at the wall, the minimum principal strains varied between –1.5e-2 and –1.3e-2. Minor values of minimum principal strains were found in the area where the newly formed cartilage was differentiated to pre-cancellous bone. The connective tissue at the resorption region in the defect basis was replaced by hyaline cartilage. Minimum principal strains were then changed to –0.5e-2.

▼ 78 

Fig. 3.7: Left: histology at the 6th week, stained with Safranin-O vK. Right: Straining of the formed tissues during healing in a selected histological section at the 6th week.

At the 12th week (Fig. 3.8) the minimum principal strain values at the defect wall were reduced by up to 12% when the fibrous cartilage was replaced by hyaline cartilage. At the defect center the connective tissue was replaced by fibrous tissue and minimum principal strains between –0.5e-2 and –0.6e-2 were found. The defect basis was completely filled with subchondral bone and thereby the minimum principal strains were reduced by up to –0.2e-2.

Fig. 3.8: Left: histology at the 12th week, stained with Safranin-O vK. Right: Straining of the formed tissues during healing in a selected histological section at the 12th week.

▼ 79 

High average strain values were registered in the bone marrow areas at the subchondral bone. The compressive strain magnitudes varied between –2.0e-2 and –2.5e-2 at the 4th week. Only minor variations were registered during healing. At the 6th week these values varied between –1.7e-2 to –2e-2 and were maintained constant until the 12th week. The maximal strain values registered in the total model were –1.09e-1 at the 4th week, -5.42e-2 at the 6th week and –4.23e-2 at the 12th week. These values were found in regions localized between the interface bone marrow and cancellous bone.

The averages of the rate of change in the minimum principal strains at each control point (4, 6, and 12 weeks) were calculated. These rates defined the factors for growth and resorption for each tissue type observed during healing in vivo. These values were used to update the elastic modulus of Young as explained above (material and methods, the tissue differentiation model) and were reported in the table 2.2.

Additionally as shown in Fig 3.9, a scheme representing the initial defect was drawn on the digitized histological section in order to compare the localization and magnitudes of the straining during defect filling. That is, the remaining host tissues are not shown. The newly formed regions (resorption or connective tissue at the basis) and their strain distribution are particularly recognizable. The blue-violet zones at the interface host-new tissue and at the center indicate zones of strain concentration due to strong changes in the mechanical properties of the related tissues. Compressive loads produce high strain values at the center of the defect where the tissues will be differentiated into stiffer tissues. The identification of zones of high biological activity during osteochondral healing (resorption, differentiation, growth) related with drastic variations in the magnitudes of compressive tissue straining was for the first time confirmed and reported in this work.

▼ 80 

Fig. 3.9: Straining of the newly differentiated soft tissues at the defect region during healing after numerical analysis of histological sections at 4 (left), 6 (center), and 12 weeks (right). The quadrilateral represents the initial defect (Ø = 6mm, depth = 2mm).

3.2.1.2 Defect model

Observing the strain pattern an unloaded region at the defect base (680 με, 27% of intact), and an increased strain field at the circumference of the osteochondral defect (5000 με, 200% of intact; Fig. 3.10) were seen. Minor differences in the straining were found when the defect size was varied. An increment of 0.8 mm in the defect diameter produced an increment of 12% in the strain field and an increment of 10% was obtained when the depth was enlarged by 0.4mm.

Fig. 3.10:Straining of the initial defect situation.

3.2.2 Comparison of the simulated healing to spontaneous repair in vivo

▼ 81 

Qualitatively, the differentiation model resembles healing as observed in histology (Fig. 3.11). The healing pattern showed “centrifugal-shaped” growth from the defect borders to the center, resorption at the defect basis, and a remaining unfilled region at the center of the defect (Fig. 3.5): after the 4th week the defect was filled with connective tissue in a non structural array, showing the first tissues at the interface of the defect wall, at the 6th week the connective tissue was differentiated to fibrous tissue in a well defined structural array and a small percentage of hyaline cartilage. At the 12th week part of the fibrous tissue was differentiated to hyaline cartilage.

Quantitatively, the differentiation model showed amounts of new-formed tissues comparable with those obtained in the histomorphometrical findings.

Fig. 3.11: Simulation of osteochondral defect healing. The selected iterations correspond to the histological staining after 4 (top), 6 (center) and 12 weeks (bottom).

▼ 82 

After qualitative matching of the simulated healing with the observations of histology and similar amounts of differentiated tissues in the simulated healing with the histomorphometrical analysis, the differentiation model was considered to be validated. Only after validation can the differentiation model be used to analyze the influence of mechanical conditions on osteochondral healing.

3.3 Influence of mechanical conditions on osteochondral healing

3.3.1 Influence of the defect size

During simulated osteochondral defect healing, tissue straining led to an increase of the material properties and successively to a filling of the defect. The filling occurred from the circumference of the defect and with resorption of the defect base (Fig. 3.11). In the simulated healing the tidemark (subchondral bone plate) of the osteochondral defect area was re-established by a defect healing from the surrounding trabecular bone rather than a bone apposition from the defect base. Analyzing the defect size, which allows healing, the following results were found: when the defect depth was increased by 50% it was filled with 5% of hyaline cartilage. An increase of 33% in the defect width produced 20% of hyaline cartilage. No hyaline cartilage formation was registered when the cartilage thickness was increased by 15% (Fig. 3.12). In no case was the defect completely filled with hyaline cartilage. However, qualitatively the healing pattern did not show significant differences.

Simulated healing appears to be stable: Growth occurred without oscillations around a specific tissue type but showed smooth variations from the initial defect to the final healing stages. The defect filling occurred mainly during the first 50 iterations (approx. 70% of the initial defect area). Between the iteration 51-150, the healing rate decreased and hence only minor variations in the elastic modulus of Young were observed after the 50th iteration. Contrarily, to the observation after the 12 weeks in the in vivo situation, the model predicted full defect filling. The complete defect filling occurred later (approx. iteration 112) when the defect width was increased compared with a model of increased defect depth (approx. iteration 72). When the cartilage stiffness was increased, filling of the defect was observed approx. in the 68th iteration (Fig. 3.12).

▼ 83 

Fig. 3.12: Simulated osteochondral defect healing: With an increased cartilage thickness only fibrous tissue was generated (A). When the defect width was increased, only slight variations compared to the initial healing prediction were found (B). When the depth was increased only fibrous tissue formation was predicted (C).

For all models minor variations in the quantified newly formed tissues at the equilibrium point were found. No changes in the material properties were registered in the remaining tissues (cartilage and cancellous bone) compared to the initial values. A slight stiffness reduction or “zone of influence” at the subchondral bone surrounding the defect was observed in all models. These results were reported by (Duda, et al., 2005). After the initial resorption at the defect base, the adaptive finite element analysis predicted a restoration of the tidemark and a complete defect filling after 100 iterations. This pattern of healing appeared to be independent of the specific defect geometry or loading configuration used (Fig. 3.12). When the defect size was varied the resorption area was comparable for both models (12% vs. 15%). This area was smaller when the cartilage thickness was increased (7%). Only in the initial defect and the larger defect situation (defect width +33%) did the adaptive finite element analysis predict cartilage formation.

3.3.2 Influence of the local joint curvature

The healing pattern was in general qualitatively comparable to the findings of the examination of the histological sections. The following stages of healing were registered both in vivo and in the simulation: initial cartilage formation at the defect wall, centripetal filling starting at the lateral borders at the interface defect-cancellous bone and resorption at the basis of the cancellous bone even a slight reduction of the stiffness in the remaining bone (Fig. 3.5). This general behavior appeared to be independent of the joint curvature (Fig 3.11, Fig. 3.13).

▼ 84 

Fig. 3.13: Simulated healing process. Left: Healing process in a flat interface. Right: Healing process in a concave interface. The healing pattern appears to be independent of the local geometry. Time points compared approximately from top to bottom: 4, 6 and 12 weeks in the animal experiment.

However, the amount of the individual tissues formed was different between models during simulated healing (Fig. 3.14 – 3.17). After equilibrium was reached, only minor differences remained (< 10%).

Compared to the total number of iterations required to achieve equilibrium (150), a high percentage of the defect (90%) was filled relatively fast in the case of a flat and a convex interface and slightly slower in a concave interface. This level was reached approximately after the first 41 iterations in a flat interface, first 47 iterations in a convex interface and first 60 iterations in a concave interface. In the case of a flat interface the pre-cancellous bone becomes stiffer and was gradually differentiated to cancellous bone (approx. iterations 41-81). After equilibrium, approx. 25% of the newly formed cartilage showed a hyaline consistence (Fig. 3.14). Although the equilibrium was achieved, a remaining unfilled defect area prevailed (approx. 10%).

▼ 85 

Fig. 3.14: Tissue quantification during healing. Top: flat interface. Center: convex interface. Bottom: concave interface. The mechanical conditions in a concave interface (bottom) appear to be more favorable for healing based on the larger quantity of differentiated hyaline cartilage.

More hyaline cartilage was formed in the concave model during simulated healing. The maximum percentage of hyaline cartilage during the simulations was smaller (27%) but occurred earlier (approx. iteration 26) in the convex than in the concave model (40%, approx. iteration 40). In the histomorphometric analysis a maximum of 33.1 ± 13.2% of hyaline cartilage was registered in the 12th week (Duda, et al., 2005). After resorption, the newly formed cancellous bone at the defect base showed in general a 35% higher stiffness in the concave interface compared with the convex ones. However, these values were in both cases up to 20% smaller than the original bone stiffness. For both models an increase of 15% in the stiffness was observed in the remaining cartilage surrounding the defect compared to flat geometry.

In the case of a flat surface, approximately 25% of the total area (TA) at the defect basis (Fig. 3.14) was reabsorbed. In the case of a concave interface, this area was higher (approx. 18%) compared with the convex surface (approx. 5%). The reabsorbed bone area was restored more rapidly in the convex situation (approx. iteration 28) compared with the flat (approx. iteration 40) and the convex situation (approx. iteration 50).

▼ 86 

As shown in the figure 3.14, the stiffness of each tissue type during simulated healing was determined. The tissues were identified and classified according to the values given in the table 2.2. However, in order to analyze the evolution of the newly formed tissues, which were localized in the initial defect area, these limits were redefined. For example, for the cancellous bone stiffness initially limited between 825.1 y 2300 MPa, two limits were then labeled inside this range: (825.1 – 1500) MPa and (1500.1 – 2300) MPa. Similarly, new limits were defined for the calcified cartilage: from (12.1 – 825) MPa these limits were extended to (12.1-25, 25.1-100, 100.1-500, 500.1-825) MPa. For hyaline cartilage the initial range (8.1 – 12) MPa was extended to (8.1-9, 9.1-10, 10.1-11, 11.1-12) MPa. Fibrous cartilage was subdivided from (3.1-8.0) MPa to (3.1-4, 4.1-5, 5.1-6, 6.1-7, 7.1-8) MPa. The initial limits of the defect tissue (0.2-3) MPa remained unchanged (Fig. 3.15, 3.16, 3.17). Since the stiffness of each tissue type during healing was written in an ASCII file it was not necessary to run the algorithm again to quantify the percentage of the elements corresponding to each redefined limit. In this form, a more precise quantification of the joint curvature effect on healing can be determined.

A relation between the bone quality reached and the stiffness of the newly differentiated cartilage tissue was established. Comparing the convex model with the concave one, it can be seen clearly that although the percentage of cancellous bone with stiffness between 1500.1 and 2300 MPa (superior quality) is slowly higher in the model with a convex curvature (Fig.3.15), the quantity and quality of the newly formed hyaline cartilage is higher in the concave model (Fig. 3.16). It is important to note that although the defect filling occurs roughly at the same “time” (iteration) for the concave and the convex model, the differentiation for the fibrous, hyaline and cancellous bone tissue occurs more slowly (requires more iterations) in the concave case (Fig. 3.15 vs. Fig. 3.16). That is, the slope of differentiation from one tissue type to another was slower in the concave case compared with the convex one.

Fig. 3.15: Simulated healing for a convex interface. For each tissue type, the range of the maximal and minimal values of the stiffness was subdivided. The slope for tissue differentiation and the quantity of reached stiffnesses during healing are illustrated.

▼ 87 

Fig. 3.16: Quantification of stiffnesses reached (tissue types) and slope of differentiation during simulated healing for a concave interface.

Analyzing healing in the flat model (Fig. 3.17) the quantity of the different newly formed tissues, as expected, was similar to the convex model. Defect filling occurs approximately at the same iteration and the slope of differentiation is comparable with the convex model, too.

Fig. 3.17: Quantification of stiffnesses reached (tissue types) and slope of differentiation during simulated healing for a flat interface.

▼ 88 

Variations in the joint curvature altered the quality of the remaining cancellous bone surrounding the defect. During the first iterations a reduction in the stiffness of the cancellous bone was observed in a region localized under the defect, which was extended from the defect basis to the distal horizontal border of the model (from 1750MPa to approx. 850MPa). This “zone of influence” was registered in all models (Fig. 3.11, Fig. 3.13). Inside this zone the cancellous bone proximal to the defect basis was differentiated to connective tissue forming the characteristic bone resorption observed in vivo. The flat interface showed the lowest percentage of affected elements (10%), followed by the convex interface with 19% and for the concave interface with 32%.

The zone of influence in the subchondral bone at the defect basis was differentiated to tissues of higher stiffness (from 850MPa to 1700MPa). After resorption the bone region was differentiated from connective tissue to other tissue types with higher elastic Young’s modulus until the original bone stiffness was achieved. However, a minor percentage of the rest of the cancellous bone inside this zone of influence surrounding the horizontal border, where the mechanical boundary condition were applied, showed a lower mechanical quality compared with the native cancellous bone stiffness. In a convex interface a maximal reduction of 52.8% in the mechanical properties of the cancellous bone was registered between the iterations 1 – 43 in 7% of the elements. After the iteration 43 the cancellous bone stiffness gradually increased to achieve the original stiffness. In a concave interface in 25% of the elements conforming this region a reduction of up to 69% was observed until the iteration 85. After the iteration 85 the stiffness of cancellous bone was restored to the original values. In a flat interface 12% of the elements diminished its mechanical stiffness of up to 65% during the first 28 iterations. After the iteration 28 the initial stiffness was gradually reestablished (Fig. 3.14).

During healing, active resorption zones, at the defect basis, were observed in the cancellous bone. These zones seemed to be more active than observed in the remaining cancellous bone. The cancellous bone stiffness was gradually reduced changing its tissue type to pre-cancellous bone, cartilage, fibrous tissue and finally to connective tissue. These resorption zones achieved a maximal percentage of 25.5% of the total area for a concave curvature (iteration 53), followed by the convex one with 10.6% (iteration 48). A maximum of 7% of the total area of the cancellous bone elements in the total area (TA) (iteration 46) were differentiated to connective tissue in the model with a flat curvature (Fig. 3.14). The patterns of these remaining defect regions in the cancellous bone were analyzed (Fig. 3.11, Fig. 3.13). Size and localization of these zones varied in dependence to the joint curvature. In total three zones, with different sizes and localizations, were observed. The region (Z1) is localized at the interface defect wall - remaining cartilage tissue. The second resorption region Z2 was localized in the center of the defect limiting at the axisymmetric axis of the model (Z2). The third region (Z3) was observed in a zone localized between the regions Z1 y Z2 (Fig. 3.18).

▼ 89 

Fig. 3.18: Resorption regions observed in the cancellous bone (Z1, Z2, Z3) during simulated osteochondral healing. Specific elements at the same regions, remarked in red, were select to compare its differentiation pattern.

In the flat joint resorption regions, type Z1 and Z2 were formed. In the convex joint curvature all resorption types were observed: Z1, Z2 y Z3, while in the concave interface only resorption regions type Z1 and Z3 were registered. After a maximum of 46 iterations for a flat interface, 67 iterations for a convex interface and 115 iterations for a concave interface, these regions started to fill themselves with cancellous bone (Fig. 3.14). After equilibrium, a remaining area of 12% of unfilled cancellous bone for a flat interface, 4% for a convex interface and 0% for a concave interface was observed. Looking at the interface defect wall to cancellous bone (original) (Fig. 3.14) in a convex joint, 5% of the elements of the total defect area are still almost undifferentiated after the first 20 iterations. This zone was changed to fibrous tissue between the iterations 21 – 39. Finally, after iteration 39 it was differentiated to cancellous bone. In a concave surface, although generally the repair process appeared to be slightly slower than observed for the convex model, these events were comparable. So, at the same interface in the concave surface the connective tissue remained undifferentiated until iteration 29. It was differentiated to fibrous tissue at iteration 61 and to cancellous bone at iteration 120. At this interface (defect wall to cancellous bone) only minor changes were observed in a flat interface. The interface defect wall to host cartilage remained 37.5% partially unfilled in a convex surface (Fig. 3.11). In the concave geometry, 29.4% of the area remained unfilled during the first healing stages (iterations 16 - 25). After 26 iterations, the percentage of the unfilled elements was reduced by up to 2% and stayed constant during the rest of the healing process. At this interface minor variations were observed in a flat interface.

Fluid flow and pore pressure were studied as possible mechanical stimulus to simulate tissue differentiation (Fig. 3.19).

▼ 90 

Fig. 3.19: Pore pressure (top) and fluid flow (bottom) as mechanical stimuli to simulate differentiation. Fluid flow was unable to reproduce the characteristic resorption region at the defect basis.

These fluid related mechanical parameters were unable to reproduce some characteristics of the healing process observed in vivo. In the simulation, pore pressure used as mechanical stimulus for differentiation appears to fill the defect very fast without formation of fibrous tissue. Fluid flow, used as mechanical stimulus, fills the defect without resorption of the cancellous bone at the defect basis, which was a typical feature of the in vivo healing process.

3.3.3 Influence of the defect fillings stiffness

Comparing the quality of the formed tissues during healing, when the defect was filled with a biomaterial, the use of a plug with the same stiffness as defined for the cancellous bone (P1) allowed calcified and hyaline cartilage formation (Fig. 3.20). A reduction of 50% of the material stiffness (P2) produced a decrease in the mechanical quality of the newly differentiated cartilage (Fig. 3.21). The remaining cartilage showed a diminution of approximately 35% of its original mechanical stiffness independent of the plug used.

▼ 91 

Fig. 3.20: Tissue quantification during healing using defect fillings. Top: Graft with the same stiffness as the cancellous bone (P1). Bottom: Graft with a stiffness of 50% of the cancellous bone (P2). Although a higher percentage of resorption is observed in P1, it shows a higher percentage of hyaline cartilage compared with P2. Additionally it shows calcified cartilage formation.

In the defect region, a calcified cartilage layer was newly formed and maintained during healing only when a biomaterial with the same native bone stiffness was used (Fig. 3.20). When the stiffness was reduced by 50%, a minor percentage of calcified cartilage was formed compared with P1 (Iterations 1 – 38). A total degradation of the calcified cartilage was observed in this case, which disappeared approximately after iteration 38. Although the formed tissues varied in quantity, the strain pattern was comparable for both models. The first iterations of the healing process for P2 appeared to be lightly unstable. During the first 7 iterations (from 150 necessaries to achieve equilibrium) the material properties of the connective tissue (defect) showed oscillatory variations around of the maximal and minimal values of this tissue type, avoiding differentiation to fibrous tissue. After the 7th iteration this effect disappeared and stable growth was detected again. This effect was principally observed at the lateral surfaces of the biomaterial plug P2.

Fig. 3.21: Differentiated tissues after healing (equilibrium) using a cylindrical graft with two different stiffnesses: 100% and 50% of the cancellous bone stiffness are compared. The usage of defect fillings with the same stiffness than the cancellous bone forms tissues with a better mechanical quality during healing. However, a slight resorption at the cancellous bone is observed, as well as calcified cartilage differentiation.

▼ 92 

For both models, the first cartilage “islands” were observed not only at the lateral wall defect, as observed in the osteochondral defects, but also at the center of the biomaterial plug. In P1, the defect was approximately 80% filled after 35 iterations. This filling consisted of approximately 95% fibrous tissue. Further into the healing process, up to 20% of the fibrous tissue was differentiated into hyaline cartilage after 75 iterations. At the equilibrium state, the defect was up to 95% filled, 30% of which was fibrous tissue and 70% hyaline cartilage. In P2, the defect was approximately 90% filled consisting only of fibrous tissue after 35 iterations. Halfway through the healing process (iteration 75), up to 15% of the fibrous tissue was differentiated into hyaline cartilage. However, at the equilibrium point, 45% of the newly formed tissue consisted of hyaline cartilage. A comparison of P1 and P2 after equilibrium (at approx. iteration 150) showed that 12% more cancellous bone had formed in P2.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
XDiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML-Version erstellt am:
15.11.2007