We used demineralised bone matrix (DBM) to augment re-attachment of tendon to a metal prosthesis in an in vivo ovine model of reconstruction of the extensor mechanism at the knee. We hypothesised that augmentation of the tendon-implant interface with DBM would enhance the functional and histological outcomes as compared with previously reported control reconstructions without DBM. Function was assessed at six and 12 weeks postoperatively, and histological examination was undertaken at 12 weeks.
A significant increase of 23.5% was observed in functional weight-bearing at six weeks in the DBM-augmented group compared with non-augmented controls (p = 0.004). By 12 weeks augmentation with DBM resulted in regeneration of a more direct-type enthesis, with regions of fibrocartilage, mineralised fibrocartilage and bone. In the controls the interface was predominantly indirect, with the tendon attached to the bone graft-hydroxyapatite base plate by perforating collagen fibres.
The clinical outcome following endoprosthetic reconstruction of large segmental bone defects can be limited by inadequate re-attachment of the soft-tissue. Bickels et al1 demonstrated significant comorbidities associated with endoprosthetic replacement of the proximal tibia, with only 78% of patients being able to walk and climb stairs without restriction. Post-operative extensor lag with ineffective active extension of the knee1,2 can occur as a result of inadequate attachment of the tendon to the implant. Successful and reliable attachment of soft tissues to the metal surfaces of implants would be a significant advance and improve functional outcomes.
Current research into tendon-metal attachment has used interpositional bone graft to create a direct-type neo-enthesis in vivo. Bone graft has been attached to both hydroxyapatite (HA) coatings3–5 and porous metal surfaces.6–9 To date, most experimental studies have used a closed clamp system for attachment of the tendon. However, poly-glycolic mesh (Vicryl, Ethicon, Edinburgh, United Kingdom) has been employed successfully to attach the patellar tendon to a metal implant augmented with autologous bone and marrow grafts.5 This mechanism provided a less enclosed environment, reducing tissue necrosis and improving histomorphometric outcomes while providing adequate mechanical fixation.5
Higuera et al9 have demonstrated that bone morphogenic protein-7 (BMP-7), also known as osteogenetic protein-1 (OP-1), used to augment tendon-implant fixation, results in similar functional and morphological outcomes as does autologous bone and marrow grafting. More recently, augmenting the tendon-bone interface with demineralised bone matrix (DBM) has resulted in functionally and morphologically superior entheses at 12 weeks.10 DBM is osseoinductive,11 providing a slow-release system for growth factors, including BMPs,12 and has the potential to improve attachment of the tendon to the implant for endoprosthetic reconstruction.
The aim of this study was to test the hypothesis that DBM augmentation of a tendon-implant interface would significantly improve the formation of a functional and morphological enthesis compared with non-augmented controls in an ovine model of reconstruction of the extensor mechanism of the knee.
Materials and Methods
The study was performed under the United Kingdom Home Office Animals (Scientific Procedures) Act 1986. We used six skeletally mature Friesland ewes weighing between 47 kg and 61 kg for the DBM Mesh group. The patellar tendon was attached to the surface of a custom-made metal implant using a four-ply sleeve of Vicryl mesh as described previously.5 The tendon-implant interface was augmented with allogenic DBM, autologous cancellous bone and a bone marrow graft harvested at the time of surgery from the ipsilateral iliac crest. Animals underwent Kistler force plate analysis (Kistler Biomechanics Ltd, Alton, United Kingdom) at six and 12 weeks after operation, and were killed at 12 weeks. The tendon-implant interfaces were harvested and examined histologically at 12 weeks. The six- and 12-week data from these animals were compared with those of a previously reported group5 (n = 6) in which the extensor mechanism reconstruction was performed in a similar manner but without DBM augmentation (Mesh group).
A custom-designed titanium alloy (Ti6-A14-V) base plate, 30 mm in length, was used to simulate the surface of a proximal tibial replacement. The base plates were plasma-sprayed with a 70 μm-thick coating of HA (Plasma Biotal Limited, Tideswell, United Kingdom) and fitted with six press-fit spikes (1.5 mm in diameter, 4 mm long) to provide mechanical fixation for the tendon-mesh construct.
DBM was manufactured according to the protocol of Urist’s described in 1965.11 The infraspinatus fossae from the scapulae of one female sheep were removed, stripped of periosteum and placed in 0.6 M HCl until radiographic analysis (Raymax, Elstree, United Kingdom) showed complete demineralisation. The demineralised bone was washed in 0.15 M NaCl and lyophilised (BOC Edwards, Crawley, United Kingdom). Strips of DBM measuring 15 mm ·30 mm were cut and then sterilised using gamma irradiation at a dose of 25 kGy (Isotron, Reading, United Kingdom). Samples were rehydrated at the time of surgery in saline for 45 minutes prior to use.
The animals were premedicated with xylazine hydrochloride 0.2 mg/kg (Bayer, Bury St. Edmunds, United Kingdom). Anaesthesia was induced with 25 mg hypnovel (midazolam; Roche, Welwyn Garden City, United Kingdom) and 2 mg/kg of ketamine hydrochloride (Ketasset, Fort Doge Animal Health Ltd, Southamptom, United Kingdom) given intravenously, and maintained with halothane 2% (Merial Animal Health Ltd, Harlow, United Kingdom) and oxygen at 4 L/min. The right patellar tendon was isolated and detached from the tibial tuberosity by sharp dissection. A tibial tubercle osteotomy was performed and the metal base plate was attached to the flat bone bed using two 2.7 mm self-tapping cortical bone screws (Synthes, Statec Medical Ltd, Welwyn Garden City, United Kingdom) with the HA coating and spikes facing away from the tibia. A layer of mesh (30 ·90 mm) was placed over the base plate, with the short axis of the mesh strip transversely aligned with the length of the base plate. Autologous cancellous bone and marrow graft (1.5 g wet weight) was harvested from the ipsilateral iliac crest at the time of surgery and packed on to the mesh-covered base plate, from the proximal margin of the plate to a distance 2 mm beyond the most distal spike to maintain consistency between individuals. In the control Mesh group5 (n = 6), the tendon was sutured into a four-ply sleeve of Vicryl mesh which extended from the distal margin of the patella to a distance of 20 mm past the end of the cut tendon. The sleeve-wrapped tendon was pressed on to the base plate spikes and two holes were made in the distal 20 mm section of the sleeve to match the positions of the screw holes in the base plate. Two 2.7 mm self-tapping cortical screws were used to anchor the distal end of the mesh sleeve through the base plate to the underlying bone. The 30 mm ·90 mm base mesh layer was then folded over the reconstruction and sutured along the medial margin of the base plate. In the Mesh DBM group (n = 6) the tendon-implant fixation was identical to that in the Mesh group; but a 2 mm to 3 mm thick piece of DBM, trimmed to the dimensions of the base plate, was placed on top of the bone graft prior to securing the tendon-mesh construct in place. The animals were given intramuscular buprenorphone hydrochloride analgesia (Vetergesic, Animal Health Ltd, Melton Mowbray, United Kingdom) for one day post-operatively and subcutaneous Ceporex antibiotic (1 mg/kg; Viovet, St Albans, United Kingdom) for five days. The animals were allowed to mobilise freely in individual pens measuring 1.6 m by 2.8 m.
Force plate analysis.
Animals underwent Kistler force plate analysis pre-operatively and at six and 12 weeks postoperatively. They were trained to walk unimpeded over the force plate, and 12 readings of the peak vertical component of the ground reaction force (GRFz) were measured from each hind leg. Each reading was normalised for weight (Fmax/weight) and data from the operated leg were expressed as a percentage of the unoperated side (%FWB).
The patella-patellar tendon-implant complexes were harvested en bloc and fixed in 10% formalin for ten days. The samples then underwent ascending graded alcohol dehydration over seven days, defatting in chloroform for five days, and embedding in LR White Hard Grade Resin (London Resin Company Ltd, Reading, United Kingdom). Sections were cut, ground and polished to 100 μm thick, and double-blind qualitative analysis was performed using image analysis software (Axiovision, Zeiss Ltd, Welwyn Garden City, United Kingdom).
Numerical data were analysed by SPSS v 12.0 for Windows (SPSS Inc., Chicago, Illinois) and analysed using non-parametric testing. Mann-Whitney U tests were used to determine differences between groups at each time point, and Wilcoxon signed-ranks tests were used to compare data between time points within each group. The results were considered to be statistically significantly different when the p-value was < 0.05. Data were expressed as median values with 95% confidence intervals (CI).
The animals recovered well immediately after operation, but one failure due to pull out from Mesh reconstruction occurred in the Mesh group after 14 days. A replacement animal was used to maintain the group size.
Force plate analysis.
At six weeks there was a significant difference in functional weight-bearing (% FWB) between the Mesh with a median of 45.27% (95% CI 34.12 to 53.79) and Mesh DBM with a median of 68.72% (95% CI 59.92 to 82.52) groups (p = 0.004) (Fig. 1⇓).
There was no significant difference in the median %FWB between the two groups at 12 weeks (p = 0.190), with the Mesh and Mesh DBM groups reaching 66.21% (95% CI 50.74 to 87.04) and 87.43% (95% CI 75.25 to 96.04), respectively.
For the Mesh and Mesh DBM groups, the differences in the median Fmax/weight between six and 12 weeks for the operated limb were 7.31 (95% CI 5.12 to 9.99) and 6.42 (95% CI 3.47 to 10.11), respectively. No significant difference between groups was observed (p = 0.345). For the unoperated limb the median values were −12.92 (95% CI −19.31 to −5.53) and 0.15 (95% CI −5.3 to 4.73). A significant difference between the groups was observed (p = 0.028), indicating that the Mesh group animals altered their gait pattern between six and 12 weeks, compensating for inadequate function in the operated limb.
For both groups a significant improvement in median %FWB was observed between six and 12 weeks (Mesh p = 0.028, Mesh DBM p = 0.028).
In the Mesh group at 12 weeks there was no evidence of residual mesh fibres or a chronic inflammatory response. The tendon mid-substance appeared to have recovered normal morphology, with crimped collagen fibres aligned parallel to the line of tensile loading (Fig. 2a⇓). The tendon-bone graft interface was predominantly an indirect type, with Sharpey-like fibres extending from the tendon into the underlying remodelling bone (Fig. 2b⇓). These regions were interspersed with fibrocartilaginous zones resembling that of a direct-type enthesis (Fig. 2c⇓). Retention of bone graft at the HA-coated base plate was not observed. Small, discrete bone graft islands were located in the inter-spike regions of the base plate remote from the HA coating (Fig. 2d⇓). Soft tissue was interposed between the bone graft islands and the HA-coated base plate. Fibrous tissue encapsulation was observed with fibroblast nuclei running parallel to the base plate (Fig. 2e⇓).
In the Mesh DBM group there was no evidence of mesh fibres or chronic inflammation. The tendon mid-substance appeared identical to that observed in the Mesh group. The tendon-bone graft interface had clearly defined regions of tendon, fibrocartilage, mineralised fibrocartilage and bone (Fig. 3a⇓). In the fibrocartilaginous region chondrocytes were observed in lacunae, in organised rows orientated along the axis of tensile loading (Fig. 3b⇓). In the mineralised fibrocartilaginous region chondrocytes were surrounded by mineralised matrix (Fig. 3b⇓). Retention of bone graft at the HA-coated base plate was not observed. However, unlike the discrete bone graft islands observed in the Mesh group, a single large bone mass was observed in the interspike region (Fig. 3c⇓). Soft tissue was interposed between the bone graft and the HA-coated base plate (Fig. 3c⇓). Fibrous encapsulation of the implant was observed.
This study demonstrated that DBM augmentation in an in vivo model of reconstruction of the extensor mechanism resulted in superior functional outcomes at six weeks and a more normal direct-type enthesis at 12 weeks, compared with autologous bone and marrow graft-augmented controls.
The Vicryl mesh was shown to have sufficient mechanical integrity to support extensor mechanism reconstruction in an ovine model. It has previously been shown to provide an environment which is more conducive to healing of the tendon-bone-implant than a compressive clamp system,5 and the results of our current study substantiate these findings. Absorbable meshes are frequently used in both orthopaedic and non-orthopaedic fields,13–15 but translation into clinical reconstruction of the extensor mechanism has yet to be generally accepted. This study indicates that Vicryl mesh has the potential to be used successfully to provide mechanical stabilisation of the healing tendon-implant interface while a biological interface capable of supporting functional weight-bearing develops.
Here, DBM has been shown to improve both functional and morphological outcomes. Studies indicate that increasing the mechanical fixation and stability of the tendon-bone interface during early healing results in an improved enthesis,16 and that biological augmentation of healing tendon-bone interfaces with a mixture of growth factors results in greater formation of new bone and fibrocartilage.17 In the normal tendon-bone interface, progressive calcification results in an increasing stiffness as the tendon approaches the bone surface, and this dictates the mechanical forces experienced by the cells throughout the enthesis. The mechanical forces and the biological environment combine to maintain the differentiated tissue types throughout the progressively mineralised interface. In our model of extensor mechanism reconstruction there are a number of cell types present immediately after the interface is reconstructed. These include tenocytes from the patellar tendon, osteoblasts from the autologous bone graft and stem cells, both in bone marrow derived from the autologous marrow graft and those recruited to the site of injury as part of the normal inflammatory response. Based on the findings of previous studies showing the chondrogenic and osteogenic nature of DBM,18–20 we postulate that augmentation of the healing tendon-bone-implant interface with DBM provides biological factors which promote osteogenic and chondrocytic differentiation of one or more of the cell types present in the healing enthesis. In the Mesh DBM group, the development of new bone and cartilage as a result of cell differentiation provides a structure more similar to a normal tendon-bone interface than with the non-augmented Mesh group controls. We hypothesise that this more natural structure, combined with the gradual slow release of growth factors from the DBM, sustains differentiation of cells at the healing interface, forming cartilage and enhancing bone formation. This cascade of events ultimately culminates in the formation of the typical four zones of tendon, fibrocartilage, mineralised fibrocartilage and bone that we observed in the DBM-augmented specimens. The precise cellular mechanisms behind the development of these tissues have yet to be established.
In 2007, Bi et al21 demonstrated the existence of tendon stem progenitor cells, but showed that BMP-2 and transforming growth-factor-®1 (TGF-®1), growth factors known to be released from DBM, inhibited these cells from expressing markers indicative of cartilaginous differentiation (including sclerix, Sox 9 and tenomodulin). In the same study, BMP-2 and transforming growth factor-®1 (TGF-®1) were shown to promote both mouse and human bone marrow-derived stem cells to differentiate along the osteoblastic lineage, expressing markers including Runx-2 and osterix. In 2003, Kasten et al22 demonstrated that human bone marrow-derived stromal cells differentiated along the osteoblastic lineage when cultured on DBM in vitro. These findings indicate that BMP-2 and TGF-®1 leached from DBM may act on the bone marrow-derived stem cells present in our in vivo model, rather than on the tenocytes or tendon-derived stem cells, and may be responsible for forming new bone at the developing entheses. As in normal tendon-bone healing, this new bone growth could then develop into fibrocartilage and mineralised fibrocartilage as the interface matures. However, Zhou et al18 showed that three-dimensional demineralised bone powder (DBP)/ collagen scaffolds induce chondrogenesis and osteogenesis in both human dermal fibroblasts (hDFs) and human marrow stromal cells (hMSCs). They showed that the latter have chondrogenic potential when treated with TGF-®1, and that DBP markedly enhances chondrogenesis in these cells. DBM activates the TGF-®/BMP signalling pathway in hDFs, normally cells with no chondrogenic potential. Also, expression of the osteoblastic phenotype was observed in both cell types when independently cultured on the scaffolds with osteogenic supplements. This study supports our supposition that the stem cell population present in healing interfaces could be responsible for the formation of bone and cartilage in the neo-entheses. However, it also shows that other cells, including tenocytes with no predilection towards chondrogenesis, could also be involved in the development of the more normal interface morphology observed in the Mesh DBM group. Further research is needed to determine the precise mechanisms involved in the development of these entheses and the role that DBM plays in enhancing interface recovery. We feel that both biological and mechanical influences are of equal importance; however, this remains to be proven.
The functional results reflect our morphological findings. At six weeks post-operatively, functional weight-bearing in the Mesh DBM group was significantly superior to that in the Mesh group. This indicates that DBM augmentation has a positive effect on early tendon-bone-implant healing. A significant difference was not observed between the groups at 12 weeks, despite the median values for functional weight-bearing in the Mesh DBM group being 20% greater, because of the large confidence intervals for the Mesh group data. In the Mesh group there was considerable variability observed in functional recovery between the animals within the group. Much less variability was observed in the Mesh DBM group, as shown by the reduced confidence intervals. For any clinical application, a consistent functional outcome is required. Although the DBM augmentation functional results were more consistent between animals at 12 weeks post-operatively, further studies are necessary to determine whether this finding persists in the long-term.
Our study was limited by retention of bone graft at the surface of the implant. This finding is similar to that of Higuera et al,9 who cited bony ingrowth as being less than 1% over their implant surface in a similar tendon-implant model. Pendegrass et al23 demonstrated that graft retention is improved by using an autologous bone block and marrow graft to augment the healing tendon-implant interface, and that this results in superior functional and morphological outcomes. We suspect that using a bone block in future studies will improve graft retention in our DBM-augmented model of proximal tibial reconstruction, and improve the long-term functional and morphological results.
The DBM used in this study provided a novel means of augmenting the tendon-implant interface, with potentially useful clinical applications and improved results over augmented controls. We hope to develop this technique into a clinically viable method of re-attaching soft tissues to metal implants.
This study was supported by Stanmore Implants Worldwide.
The author or one or more of the authors have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article.
- Received February 3, 2009.
- Accepted May 20, 2009.
- © 2009 British Editorial Society of Bone and Joint Surgery