By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of March 17 & 18. You should not experience any issues or loss of access during this time. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
Bone healing requires: cells that are capable of forming bone (osteogenicity); bioactive factors that can attract such cells and initiate bone formation (osteoinduction); a matrix that guides the bone formation (osteoconduction); adequate vascularization; and initial mechanical support to the surrounding bone, which becomes more important as the size of the defect increases.1
Autologous bone is considered the gold standard treatment for bone defects and is mostly harvested from the iliac crest. However, the harvesting procedure has a complication rate of 10–40%, including hemorrhage, nerve, and vascular lesions and post-operative pain.2 Moreover, the amount and quality of bone that can be harvested is limited, restricting its use in large defects.3 Therefore, large defects are currently treated by distraction osteogenesis, vascularized bone (fibula) grafting, or massive cortical allografts.4 All have disadvantages, such as multiple surgical procedures, high complication rates, and prolonged periods of immobility and rehabilitation.
The challenge is to develop a bone substitute material that enhances healing but also offers adequate strength. Porous titanium scaffolds are interesting, since titanium has superior mechanical properties compared to materials such as calcium phosphate ceramics and polymers.5 Although its potential has been recognized for years, development of open porous structures has been hampered by limitations in production techniques.6 With plasma spraying,7 space-holder techniques,8 powder metallurgy,9 or sintering of titanium fibers,10 it is difficult to produce a porous structure with the desired architecture that meets both osteoconductive and mechanical requirements. For osteoconduction, an open interconnected porous structure with pores in the range of 200–500 µm is required.11 From a mechanical point of view, the structure should be stiff enough to sustain physiological loads, but should not drastically exceed the stiffness of the bone being replaced to avoid stress shielding.
Better control over the structural architecture can be acquired using selective laser melting (SLM).12 SLM allows production of fine and small porous titanium structures, with struts in the range of 100–200 µm. This enables the possibility of tailoring and optimizing the structural and mechanical properties of the scaffolds while maintaining the required pore dimensions that allow for bone and vessel ingrowth. Thinner titanium struts may result in increased elastic and plastic deformation. Such deformation of the porous structure reduces stress-shielding inside the scaffold and may provide a biomechanical stimulus for the bone-forming cells, thereby resulting in more bone formation.13
In this study, we used a critically sized femoral bone defect in a rat model to test two hypotheses: porous titanium scaffolds can be a biomechanically strong osteoconductive scaffold for repair of cortical bone defects, and thinner strut sizes will result in favorable mechanical properties that will increase bone formation within the titanium scaffold thereby improving mechanical integrity of the treated bone defect.
MATERIALS AND METHODS
Porous Titanium Scaffolds
Porous titanium scaffolds were produced from Ti6Al4V using SLM (Layerwise, Leuven, Belgium). Two structural variants were designed using a dodecahedron unit cell as a template structure. One variant consisted of thin titanium struts (“titanium-120”); the second consisted of thick struts (“titanium-230”). Both variants were produced in two shapes: cylindrical scaffolds (Ø5 mm × 10 mm) for determining the compression strength and elastic modulus (Supplementary Material 1) and femur-shaped scaffolds (6 mm mid-diaphyseal segment of the femur bone, Fig. 1) for determining the ultimate compression force (UCF; Supplementary Material 1) and for in vivo implantation. All samples underwent post-production chemical and heat treatment to increase surface roughness (Supplementary Material 2).
In 27-male Wistar rats, a 6 mm segmental bone defect was created in the right femur and treated with either titanium-120 (n = 9) or titanium-230 (n = 9), or was left empty in the control group (n = 9). The local animal ethics committee approved the study. All animals were housed according to the national guidelines for care and use of laboratory animals.
A single dose of antibiotics (enrofloxacin, 5 mg/kg body weight) was administered 1 h before surgery. The operation was performed aseptically under general anesthesia (1–3.5% isoflurane). The right femur was exposed though a lateral incision of the skin and division of the underlying fascia. A 23 mm long PEEK plate (RatFix, AO Foundation, Davos, Switzerland) was fixed to the anterolateral plane of the femur. Three proximal and three distal screws fixed the plate. The periostium was removed over ∼8 mm of the mid-diaphysis before removal of the 6 mm long bone segment. The segment was removed with a tailor-made saw guide and a wire saw (RatFix, AO Foundation), and the scaffold was placed press-fit into the defect site. The fascia and skin were sutured in layers and prophylactic pain medication (buprenorphine, 0.05 mg/kg body weight) was administered twice a day for the first 3 days after surgery. Fluorescent dyes were administered at 4 (tetracyclin, 25 mg/kg body weight), 8 (calcein, 25 mg/kg body weight), and 11 weeks (xylenol orange, 90 mg/kg body weight).
Immediately after the surgery, while the rats were still anesthetized, a SkyScan 1076 scanner (Bruker micro-CT, Kontich, Belgium) was used to acquire a baseline in vivo micro-CT scan. A 36 µm-resolution protocol was used at 95 kV, 1.0 mm Al filter, and 0.6° rotation step, resulting in a 15 min scan. In vivo scans were repeated after 4, 8, and 12 weeks. For the final ex vivo scan, an 18 µm-resolution protocol was used at 95 kV, 1.0 mm Al/0.25 mm Cu filter, and 0.4° rotation step (3 h scan). CT images were reconstructed using volumetric reconstruction software NRecon version 1.5 (Bruker micro-CT).
The total bone volume (TBV) was defined as the total bone volume within the 6 mm defect segment including bone formed around the titanium scaffold (Fig. 2A) The bone volume in pores (BVp) was defined as the bone volume measured within the pore volume (PV) of the titanium scaffold (Fig. 2B), and was also expressed as a percentage of the pore volume (BVp/PV). TBV and BVp were determined using software CTAnalyser version 1.11 (Bruker micro-CT; Supplementary Material 3).
The strength of the treated femurs was measured with three-point bending tests conducted on five samples from each group. Both bending supports were chosen as close as possible to the bone-scaffold interfaces (distance <5 mm). Small distance between the interfaces and the supports ensures that the bending test measures the interface strength of bone and scaffold. The contralateral femurs served as controls. To ensure that we tested the entire spectrum, we first sorted the treated femurs according to their BVp and then included every other femur. The bending tests were carried out using a Zwick test machine (Zwick GmbH, Ulm, Germany). First, the PEEK plate was carefully removed. The femurs were then supported at the proximal and distal side using two plates that were secured with screws. A plate that exceeded the average pore size applied a downward force to the middle of the porous titanium scaffold, pushing it outside the bone defect. The tests were performed at a rate of 2 mm/min until the peak load was reached. The force–displacement curves were used to determine the maximum force.
Histology was performed on four femurs of each group to study the bone-titanium interface and bone morphology. Specimens were dehydrated in a graded ethanol series, and embedded in methylmethacrylate. Sections of ∼20 µm were obtained using a diamond saw (Leica SP1600) and stained with basic fuchsin 0.3% solution (Sigma, Zwijndrecht, The Netherlands) and methylene blue 1% solution (Sigma). Bone stains red with basic fuchsin and fibrous tissue stains blue with methylene blue. Unstained sections were examined using an epifluorescent microscope (Axiovert 200MOT/Carl Zeiss, Göttingen, Germany) with a triple filter block.
Statistical analyses were performed using SPSS Statistics 17.0 (SPSS, Chicago, IL). The data are presented as means with standard deviation. One-way ANOVA and subsequent post hoc pairwise comparisons with Bonferroni adjustment were used to analyze differences among the three groups. A repeated measures general linear model was used when examining the longitudinal in vivo micro-CT data. A Pearson's correlation coefficient was used to determine the correlation between BVp, TBV, and the maximum bending force.
Porous Titanium Scaffolds
The different strut sizes and equal pore dimensions resulted in a porosity of 88% in the titanium-120 scaffolds and 68% in the titanium-230 scaffolds (Table 1). The titanium-120 structure had fivefold lower compression strength and a fourfold lower homogenized elastic modulus than the titanium-230 structure (Table 1). There was a significant difference in the UCF (p < 0.001). The UCF of the titanium-230 scaffolds (530 ± 85 N) was higher than the corresponding bone segments (441 ± 31 N, p = 0.022), whereas the UCF of titanium-120 scaffolds (84 ± 11 N) was lower than the corresponding bone segments (p < 0.001; Fig. 3).
Table 1. Structural and Mechanical Characteristics of Porous Titanium Scaffolds29
Cortical Bone (Rat)
Pore size is presented as median and range. Compression strength and homogenized elastic modulus is presented as average + SD.
Titanium thickness (µm)
Pore size (µm)
Surface area/volume (µm2)
Compression strength (MPa)
14.3 ± 1.7
77.7 ± 12.8
140 ± 19 (29)
Homogenized elastic modulus (GPa)
0.38 ± 0.04
1.56 ± 0.21
8.80 ± 2.53 (29)
Correct positioning of the scaffolds was confirmed by micro-CT directly after surgery in all animals and no dislocation of the scaffolds was detected during follow-up. The titanium-230 structure remained completely intact in all rats, whereas breakage of some struts was seen in six of the nine rats given titanium-120. This occurred after either 4 (two cases) or 8 weeks (four cases), but did not result in loss of fixation or complete loss of structural integrity of the scaffolds. The scaffolds were well integrated with the adjacent cortical bone and a progression of bony bridging was observed over time (Supplementary Fig. S1), although in some rats small areas of the adjacent cortex underwent changes that may indicate bone resorption (Supplementary Fig. S2). In the empty control group, loss of fixation, due to breakage of the screws, occurred in six of nine rats. This happened to one rat at 4 weeks, to four rats at 8 weeks, and to one rat at 12 weeks. Those rats were taken out of the experiment at subsequent time points. In the remaining rats, no bridging of the defect had occurred, and a consistent pattern of bone resorption of the remaining cortical bone was observed (Supplementary Fig. S3).
Treatment with porous titanium scaffolds resulted in more TBV than in the empty controls at all time points (Fig. 4A). The increase of TBV was most profound between 4 and 12 weeks, whereas in the empty controls TBV reached a plateau after 8 weeks. At 12 weeks, a significant difference in TBV (p = 0.008) was found (Fig. 4B). The TBV of the titanium-120 group (18.4 ± 7.1 mm3) and the titanium-230 group (18.7 ± 8.0 mm3) were significantly higher than the TBV of the empty control group (5.8 ± 5.1 mm3, p = 0.015 and p = 0.012, respectively).
The porous structure of the scaffolds facilitated bone ingrowth given that an increase of BVp was found at all time points (Fig. 5 A). At 12 weeks, the absolute BVp was 7.4 ± 2.3 mm3 in the titanium-120 scaffolds and 6.0 ± 2.7 mm3 in the titanium-230 scaffolds (p = 0.38) (Fig. 5B). This resulted in a BVp/PV of 16 ± 5% in the titanium-120 and 20 ± 9% in the titanium-230.
The intact femurs broke at a force of 233 ± 27 N. The bending force of the titanium-120 treated femurs was 144 ± 73 N (62% of control) compared with 104 ± 38 N (45% of control) for titanium-230 treated femurs (Fig. 6A). Except for one case, all samples broke at the titanium-bone interface. BVp measured with micro-CT strongly correlated with the maximum bending force for the titanium-120 group (r2 = 0.83, p = 0.03). The two treated femurs in which >8 mm3 bone had formed within the pores had a bending force comparable with the intact femurs (Fig. 6B). For the titanium-230 group, the maximum bending force did not seem to relate to BVp (r2 = 0.02, p = 0.84).
The empty defect sites showed limited bone formation and resorption of cortical bone at the proximal and distal sites (Fig. 7A and F). Within the remaining defect area, abundant fibrous tissue was found. Histology of the titanium groups revealed formation of a major plug of new bone in the medullary canal at both ends of the bone defect. This bone most likely formed through the process of direct ossification (Fig. 7B and D). The newly formed bone extended from this plug into the porous titanium and the inner space of the scaffold. Bone was also abundant at the outer area of the scaffolds, showing signs of an attempt to bridge the defect. The area inside the porous titanium that was not filled with bone was filled with fibrous tissue. The pattern observed correlated well with the bone seen on the corresponding micro-CT images (Fig. 7G and H).
Bone formed directly on the surface of the porous titanium scaffold. At some areas, however, a thin layer of fibrous tissue between the titanium and the bone was observed (Fig. 7E). No signs of foreign body reactions or inflammation were detected. In one titanium-120 sample, a possible development of a hypertrophic non-union was seen, since a cluster of chondrocytes was found at a site suspect to breakage of titanium struts (Fig. 7C).
The injected fluorochrome labels showed the mineralized bone at 4 (red), 8 (green), and 12 weeks (yellow; Fig. 8). Bone formation was most active around the titanium-bone interface at the proximal and distal ends of the porous scaffolds (Fig. 8D). Only limited progression of the bridging of the bone defect through the medullary cannel was seen between 4 and 12 weeks (Fig. 8C), since the label injected at 4 weeks (red) was found close to the most advanced bone fronts (yellow).
This longitudinal in vivo study supports our first hypothesis that porous titanium scaffolds provide mechanical support in the early phase after implantation and facilitate bone formation (osteoconduction) over time, resulting in good mechanical strength of the treated femurs after 12 weeks. A lower titanium strut size reduced the homogenized elastic modulus of the scaffold but did not result in significantly more bone formation or higher mechanical strength of the treated femurs, meaning that these experiments did not support our second hypothesis.
The osteoconductive properties of porous titanium scaffolds were proven by the fact that more bone had formed in the bone defects treated than in the defects that were left empty. This agrees with previous reports that used a metaphyseal bone defect model in rabbits.14, 21–24 The rat defect model used here allowed for in vivo micro-CT scanning to monitor bone formation with time. Bone formation was measured using a custom-made algorithm that first removed the metal artifacts and then selected the areas of newly formed bone (Supplementary Material 3). Accurate selection of bone was verified using corresponding histological sections (Fig. 7). The in vivo bone measurements showed a gradual increase in bone formation in rats that received titanium-120 or titanium-230 scaffolds. This formation may have been ongoing, because no plateau phase was reached within the 12 weeks (Fig. 4A).
The increase in bone regeneration in the defects treated with porous titanium scaffolds may be related to the scaffold structure and its mechanical properties. The structure of osteoconductive scaffolds is well defined in terms of pore size, interconnectivity, and porosity (11), and these criteria were met for both structural variants. However, the mechanical properties of the two structural variants were different due to their different strut sizes. Reducing the strut size by ∼50% in the titanium-120 structure resulted in a large decrease of the homogenized elastic modulus (Table 1). The measured modulus for the titanium-120 is close to the lowest range reported in the literature for porous titanium8, 14–17 and within the range of human trabecular bone (0.01–2 GPa).18 Such low modulus allows for more deformation upon loading, and was therefore hypothesized to result in more bone ingrowth in the titanium-120 scaffolds. However, this did not occur (Fig. 5B); a possible explanation could be that the loads applied to the titanium-120 scaffolds after implantation were unable to deform the scaffolds.
Defining the mechanical properties that would allow deformation of the porous titanium scaffolds after implantation was complicated by a number of factors. Although the titanium-120 was significantly weaker than the femur segment that it replaced, and the titanium-230 was significantly stronger, bone is able to withstand forces that are at least twice the normal peak loading.19 Furthermore, different bones and even different areas of a bone can have different mechanical properties.18 Finally, not all the loads will be transferred through the scaffolds, since a portion of the load will be transferred to the PEEK fixation plate. Preliminary results of a finite element model of this femur bone defect indicates that the division of force is highly dependent on scaffold stiffness, contact conditions between scaffold and bone, and the loading.20 Moreover, the load distribution changes over time as more bone is generated within the scaffold. Therefore, one must take the species, the type of bone that needs to be replaced, and the applied fixation methods into account.
Implantation of the scaffolds provided sufficient support to the defect, because there was no loss of fixation, whereas in most rats with the defect empty suffered plate fixation failure. The ability to provide sufficient support likely contributed to bone formation in the defect but was only possible because the scaffold was load-bearing. The bending forces were surprisingly high, taking into account 12 weeks of implantation with only ∼20% of the pore volume occupied by newly formed bone. The broken struts seen in the titanium-120 scaffolds, which could be explained by the limited compression strength, did not negatively impact the maximum bending force. In fact, the force was somewhat higher in the titanium-120 group compared to the titanium-230 group (Fig. 6A). Interestingly, there was a strong correlation between bending force and bone volume inside the pores for the titanium-120 but not for the titanium-230 scaffolds. Possible other factors that may affect the strength of the treated femurs could be the bone-titanium bonding. Previous studies that used similar heat and surface treatments showed good bone-bonding and indicated a possible osteoinductive role of the modified surface.25 The larger surface area in the titanium-120 scaffolds (Table 1) may have resulted in a larger area of direct bone-titanium contact. This may explain why bone volume within the pores shows a better correlation with the final strength of the femurs that received a titanium-120 scaffold.
Our work shows the potential of porous titanium scaffolds to function as a load-bearing scaffold that may become relevant in clinical cases where conventional fixation methods alone may be insufficient. But before porous titanium can be used clinically, the mechanical properties should be tailored to the human situation. Another aspect of porous titanium that should be further explored is the surface. Surface modifications have been studied,26 and present an opportunity to enhance bone-titanium bonding or increase bone formation. An example would be adding a calcium phosphate coating.27 Antibiotic coatings have been developed for solid implants,28 and they may help to reduce the risk of infection, the main drawback of titanium implants. The challenge will be to combine all these techniques into one scaffold that can meet the clinical requirements.
This research forms part of the Project P2.04 BONE-IP of the research program of the BioMedical Materials institute, co-funded by the Dutch Ministry of Economic Affairs, Agriculture and Innovation. Osteosynthesis & Trauma Care foundation is acknowledged for financial support.