Effects of surgical angiogenesis on segmental bone reconstruction with cryopreserved massive-structural allografts in a porcine tibia model

Abstract

Cryopreserved bone allografts (CBA) used to reconstruct segmental bone defects provide immediate structural stability, but are vulnerable to infection, non-union and late stress fracture as the majority of the allograft remains largely avascular. We sought to improve the bone vascularity and bone formation of CBAs by surgical angiogenesis with an implanted arteriovenous (AV) bundle, using a porcine tibial defect model.

Cryopreserved tibial bone allografts were transplanted in swine leukocyte antigen (SLA) mismatched Yucatan minipigs to reconstruct a 3.5cm segmental tibial defect. A cranial tibial AV-bundle was placed within its intramedullary canal to induce angiogenesis. The AV bundle was patent in 8 pigs and ligated in a control group of 8 pigs. At 20 weeks neo-angiogenesis was evaluated by micro-angiography. Bone formation was measured by quantitative histomorphometry and micro-computed tomography (micro-CT).

Seven of 8 AV-bundles in the revascularized group were patent. One had thrombosed due to allograft displacement. Total vascular volume was higher in the revascularized allografts compared to the ligated group (p=0.015). Revascularized allografts had increased levels of bone formation on the allograft endosteal surface compared to the ligated control group (p=0.05).

Surgical angiogenesis of porcine tibial CBAs by intramedullary implantation of an AV-bundle creates an enhanced autogenous neoangiogenic circulation and accelerates active bone formation on allograft endosteal surfaces.

Introduction

Treatment of large segmental bone defects created by limb-sparing tumor resection, infection and traumatic loss has significant morbidity with all currently available methods, including vascularized autografts, prosthetic replacement, bone transport, and use of structural cryopreserved bone allografts (CBA). ^{1–7; 8} Vascularized autologous bone provides superior rates of healing and remodeling than banked allogeneic bone.⁹ Unfortunately, only a few donor sites (chiefly fibula and iliac crest) are available, with generally poor size- and shape match and significant donor site morbidity. Cryopreserved allogeneic bone can be stored in tissue banks and matched preoperatively to provide an immediately stable reconstruction. Their lack of cellular and vascular viability however makes CBAs susceptible to infection, and non-union. Limited ability to revascularize over time subsequently prevents bone remodeling in response to loading, resulting in late stress fractures.

Surgical revascularization of structural allograft bone has been demonstrated by angiogenesis from implanted AV bundles or fascial flaps in small laboratory animal models resulting in improved blood flow, healing and active remodeling.^{10; 11} Using sex-mismatched transplantation, laser capture microdissection and PCR amplification, we have demonstrated the lineage of osteocytes in areas of active new bone formation to be circulation-derived from the recipient animal in a rat femoral model. ¹² Others have reported similar remodeling when a periosteal sleeve was left behind to envelop an orthotopically-placed allograft. ¹³ The angiogeneic effect has been demonstrated to be amplified with the use of growth factors when delivered by osmotic pumps, biodegradable microspheres or endothelial cell transfection (gene therapy).^{11; 14–16} Such surgical angiogenesis may prove of value in improving outcomes for many patients in whom cryopreserved bone segments are used to span large segmental gaps resulting from limb-sparing tumor surgery, trauma, congenital deformity or osteomyelitis.

Miniature swine have distinct advantages for allogeneic tissue research, particularly as a bridge to clinical application. Their size, anatomy, physiology and immunology are well known and comparable to man. Most importantly, both blood type and the swine major histocompatibility haplotypes (the swine leucocyte antigen or SLA complex) have been well defined, the latter determined by routine DNA sequencing. For bone, orthotopic reconstruction of segmental bone defects in porcine hindlimbs uses surgical techniques and implants identical to clinical practice. Their physiology, including rate of new bone formation is nearly identical to man¹⁷. A number of vascularized composite allotransplantation studies have taken place using miniature swine, including skin, muscle and bone, bone marrow, and composite knee joint^18–23 transplants. Prior to clinical implementation however, confirmation in a large animal model is desirable. The anatomic and physiologic similarities between pig and man make a porcine model ideal for such a study.

It is our hypothesis that an AV bundle implanted within a cryopreserved CBA will facilitate angiogenesis, resulting in an increased number of intramedullary vessels as well as active bone formation of the previously necrotic endosteal surface of the graft in a large animal model.

Methods

This study was approved by the Institutional Animal Care and Use Committee. A total of 20 female Yucatan minipigs were used, provided by Sinclair Bio Resources, LLC, all with defined swine leukocyte antigen (SLA) haplotypes. The animals averaged 26 kg in weight, and were 6 months of age. Sixteen pigs comprised two experimental groups of eight. Four additional pigs were used exclusively as donors for the first eight tibial allografts.

In all animals, a 3.5cm tibial segment was removed from the proximal tibial diaphysis, the defect comprises about 30% of the diaphysis with an average diaphyseal length of 11cm. This defect is in excess of a 3.0 cm critical defect used by a previous investigator¹³. In the first 8 experimental pigs, the excised tibia segment was stored at −80°C for a minimum of one month, and used as the CBA for later animals. In all animals, the tibial defect was immediately replaced with a sized-matched cryopreserved allograft, obtained from a donor with a major SLA mismatch. The donor bones for the first 8 pigs were obtained from the four additional pigs exclusively used as donors. The excised tibia segment from the first 8 pigs served as donors for the last 8 pigs. Implantation of a patent arteriovenous bundle within the tibia provided surgical angiogenesis. In the ligated group, the AV bundle was implanted but ligated proximally to prevent angiogenesis in an otherwise identical procedure. In all animals, the contralateral undisturbed tibia was used to normalize values.

To harvest the first 8 grafts, 4 Yucatan minipigs were anesthetized with Tiletamine HCL + Zolazepam HCL 5 mg/kg IM (Telazol 100 mg/ml, Zoetis inc, Kalamazoo, MI.), Xylazine 2 mg/kg IM (Xylamed 100 mg/ml, Bimeda-MTC, Cambridge ON, Canada) and euthanized with Pentobarbital Sodium 390mg/ml (Vortech Dearborn MI, 0.22 ml/kg IV). A 3.5cm tibial mid-diaphyseal segment was removed bilaterally under sterile conditions, beginning proximally at a point 1 cm distal to the tibial tubercle using a cutting jig for precision. The marrow was removed, the bone rinsed with sterile saline, and then stored at −80°C for at least one month.

All remaining surgical procedures were performed under inhalational anesthesia. Induction of anesthesia was performed with injection of Tiletamine HCL + Zolazepam HCL 5 mg/kg IM, Xylazine 2 mg/kg IM and Glycopyrrolate 0.01mg/kg IM (Novaplus/Vizient 0.2 mg/ml, Chicago IL). Venous access was gained through the ear vein and used for fluid administration (Ringer’s lactate, average 1000 ml/procedure IV) and drug administration. Preoperative antibiotic prophylaxis was given in the form of Cefalozin 1 gram IV (Hospira, Lake Forest, IL). The pig was intubated and inhalation anesthesia was maintained with isoflurane (1-3% inhalation). The leg was washed with iodine; a sterile field was created with surgical drapes and adhesive iodine incision drapes (Ioban, 3M Health Care, St Paul, MN). Through a 10cm anterolateral incision along the cranial ridge of the tibia, the cranial compartment musculature was dissected from the tibia. The cranial tibial artery and vein were identified between the cranial tibial muscle and the lateral surface of the tibia, ligated at the ankle and elevated as an AV bundle with proximal inflow. A 3.5cm tibial bone segment was removed, prepared and stored at −80°C for potential later use as described above. The pre-selected SLA-mismatched allograft was thawed in sterile saline. It was positioned into the matched tibial defect, threading the AV bundle through the medullary canal from proximal to distal. A small burr was used to create a small window at the bone junction sites to pass the AV bundle. In the ligated group, the bundle was then ligated proximally to obstruct the blood flow. (Figure 1+2) Osteosynthesis was performed with a 9-hole 2.7mm locking compression plate spanning the allograft on the medial side (Synthes, Monument, CO). The plate was applied to the allograft with two unicortical locking screws, to the proximal and distal tibial segments with 3 bicortical screws each (1 compression and 2 locking screws). The fascia and subcutaneous tissue were closed with 2-0 braided polyglactin (Vicryl, Ethicon, Somerville NJ) and the skin with subcuticular 3-0 polyglactin sutures. A compressive and occlusive dressing was applied (Dermabond Prineo Skin Closure System, Ethicon, San Lorenzo, Puerto Rico and Tegaderm, 3M, St Paul, MN). The first radiographs were taken immediately postoperatively. Full weight bearing and unrestricted movement was allowed following the surgery. Excede 5 mg/kg IM every 5days (Pfizer, NY, NY) and Baytril 10mg/kg IM daily (Bayer Healthcare, Shawnee Mission KS) were given postoperatively for 2 weeks. Buprenorphine SR 0.18mg/kg (10mg/ml ZooPharm, Windsor, CO) was given prior to surgery and repeated after 72 hours. Carprofen 4mg/kg IM (Rymadyl, Zoetis Inc, Kalamazoo, MI) was given in the first 5 days after surgery and continued when necessary.

Figure 1:

Surgical revascularization of a cryopreserved tibial bone allograft in a porcine model. First image: the anatomy of a porcine tiba bone with popliteal artery, the caudal tibial artery and the cranial tibial artery. Second image: the 3.5cm tibia bone segment is removed. Third image: the technique used in the revascularised group: transplantation of the CBA, the cranial tibial artery and vein are distally ligated and inserted into the intramedullary canal (arrows). The CBA is fixated with a plate and screws. Fourth image: the technique used in the control group: transplantation of the CBA, the cranial tibial artery/vein both distally and proximally ligated (arrow) and inserted into the intramedullary canal. The CBA is fixated with a plate and screws.

Figure 2:

Intra-operative pictures of the cryopreserved allograft with an implanted AV-bundle. (A) Porcine tibia, view from medial. (B) Arteriovenous bundle consisting of the cranial tibial artery and vein (arrow) distally ligated with a suture. (C) View from cranial: transplanted CBA with the arteriovenous bundle placed intramedullary through the burred hole proximally (arrow) end fixated with a plate medially. (D) View from medial.

Radiograph assessment: Bone Healing

Radiographs were obtained at surgery, 2, 4, 6, 10 and 20 weeks after sedation with Telazol (5 mg/kg IM) and Xylazine (2 mg/kg IM). The radiographs were taken at 91.4 cm, with exposure settings at 70kVp and 5mAs. Bone healing was assessed with a bone healing score designed and used for assessment of cortical allografts.^{24; 25} (Table 1) All radiographs were scored by two blinded observers.

Table 1.

Radiographic Evaluation of Osseous Healing and Transplant Incorporation Modified After Taira et al.²⁵

Periosteal reaction (prox. and dist. separately
None	0
Minimal (<25%)	1
Medium (25-50%)	2
Moderate (50-75%)	3
Complete (75-100%)	4
Callus Remodeling (prox. and dist. separately)
None	0
Partial	2
Complete	4
Union (prox. and dist. Separately
Total line	0
Partial line	2
Union	4
Graft appearance
Resorption	−1
No change	0
Periosteal reaction from the graft	1
Max. Total*	25

^*The maximum score for a vascularized transplant with periosteal reaction that is fully healed and remodeled both distally and proximally in 25 points.

Fluorochrome labeling: Histomorphometry

Fourteen and 4 days prior to sacrifice, calcein (Sigma, St Louis, MO 20mg/kg) and oxytetracycline hydrochloride (Vetrimycin 100mg, Boise, ID. 20mg/kg) fluorescent labels were administered intramuscularly, respectively for subsequent histomorphometric measurements of bone formation occurring during this 10-day time point. Any observed differences between the two groups can be attributed to the effect of a patent implanted AV bundle. Angiogenesis from the bundle growing into the medullary bone may allow cell migration and survival, and ultimately new bone formation. This process essentially is ‘revitalizing’ the necrotic allogeneic bone.

Sacrifice

The survival time was 20 weeks; the pigs were euthanized as described above in the first 4 donor pigs. Both lower extremities were disarticulated at the hip. A catheter was placed in the femoral artery and the vasculature of the lower extremity irrigated with 150ml of a saline/heparin solution. Subsequently, Microfil (Flow Tech, Inc. Carver, MA) was injected under physiologic pressure (100-120 mm/Hg) using a syringe pump. Microfil is a radiopaque silicone rubber that fills the capillaries with a low resistance due to the low viscosity (25 centipoise) and has been widely used in previous micro-CT studies.²⁶ Directly following sacrifice the operated and contralateral tibial bones were removed. At this time, the AV bundle was identified and noted to be patent or thrombosed based upon Microfil filling. Micro-computed tomographic scans were then made of all tibiae using a micro-CT system (Siemens Medical Solutions, Knoxville, TN, USA). Scanning parameters were 60 KeV, 500 μA, 180 projections with 188.5μm thickness slides and a low magnification. These images were used to visualize the intraosseous vasculature and analyze bone formation.

Vessel volume – Micro-CT

A 2cm segment of the central part of the graft and the contralateral tibia were used for microangiography. The segment was cut using a Diamond band saw (Exakt Technologies, Oklahoma City, OK, USA). The bone segments were fixed in 10% buffered formalin for 48 hours followed by decalcifying solution (Thermo Scientific, Baltimore, MD) for seven weeks, changed weekly. The decalcified bone segments were imaged using the micro-CT. Scanning parameters were 60 KeV, 500 μA, 360 projections with 53.9 μm thickness slides and a high magnification. The vascular volume occupied by Microfil was measured using the Bone Microarchitecture Analysis (BMA) module in the Analyze 12.0 software (Mayo Clinic, Rochester, MN). After assessing all images, an optimal signal threshold was determined to provide the best contrast between the decalcified bone and the Microfil, removing the background noise, with further refinement by including only the 250 largest connected vascular regions. Cortical and medullary vascularization were distinguished based on the difference in bone density. In this manner, a standardized vascular volume measurement in mm³ was possible. The total vessel volume and the percentage of bone volume occupied by vessel were calculated as shown in table 2.

Table 2.

Vascular Volume, contralateral side (cl), ligated group (cg), revascularization group (rg)

	Contra-lateral (cl) (n=16)	Ligated group (cg) (n=8)	Revas-cularized group (rg) (n=7)	p-value cl-cg	P-value cl-rg	p-value cg-rg
Total Vessel Volume (TVV), mm3	29 ± 13	55 ± 47	127 ± 50	0.123	0.018	0.015
TVV/Total Bone Volume (%)	0.92 ± 0.38	0.77 ± 0.46	2.0 ± 0.65	0.484	0.028	0.004
Medullary Vessel Volume (MVV), mm3	7.4 ± 6.0	2.7 ± 2.1	58 ± 22	0.208	0.018	0.001
MVV/volume medullary bone (%)	0.84 ±0.75	0.72 ± 0.68	11 ± 4.6	0.889	0.018	0.001
Cortical Vessel Volume (CVV), mm3	22 ± 10	52 ± 47	68 ± 46	0.093	0.018	0.298
CVV/volume cortical bone (%)	0.72 ± 0.27	0.75 ± 0.49	1.0 ± 0.4	0.866	0.176	0.247

Quantitative Histomorphometry

A 2.5mm bone segment of the graft was removed and embedded in methyl methacrylate. Fifteen μm-thick sections were cut using a Diamond band saw. Hematoxylin and eosin staining were performed as a standard. Unstained slides were analyzed using fluorescence microscopy at a magnification of 20x (Olympus BX51) with bone image analysis software (Osteomeasure; Osteometrics, Atlanta, GA), using the known elapsed time between differentially-colored double fluorescent calcein and oxytetracycline labels to quantify bone formation.²⁵ Histomorphometric parameters were determined for both the inner and outer cortical layer by analyzing two fields per cortical layer for each of the three sides of the triangularly-shaped bone, making a total of 12 analyzed fields per slide. The Mineralizing bone Surface to total Bone Surface (MS/BS) ratio represents the proportion of bone that is actively mineralizing, including all double-labeled surfaces and half of the single label surfaces. The Mineral Apposition Rate (MAR) represents the average rate at which new bone mineral is being deposited and was calculated from the distance between the two fluorescent labels (calcein and oxytetracycline) divided by the 10 days separating the administration of the labels. In samples where only single labels or too few double labels were present, a minimum value of 0.1 mcm/d was assigned, to decrease bias resulting from data exclusion.²⁷ The Bone Formation Rate to Bone Volume ratio (BFR/BV), a measure of bone turnover rate, was calculated by multiplying MAR by the ratio of labeled bone surface to bone volume (BFR/BV [μm³/μm³/year]).²⁸

Statistical Analysis

The results of the analyses of the operated leg were compared with the normal, non-operated contralateral leg. Histomorphometry, bone mineral density and bone volume were expressed in absolute values of the graft and normalized values as a percentage of the contralateral side. Volume was expressed in absolute values and as a percentage of the bone volume.

Statistical analysis was performed using the Mann-Whitney U test to detect differences between groups (conventional grafts versus revascularized grafts) and the paired Wilcoxon signed rank test to detect differences within each group (graft versus contralateral tibiae). Significance was set at p<0.05.

Results

The postoperative course was without significant complication in any of the 16 swine. Two animals had superficial suture-related infections that resolved with oral antibiotics. All animals ambulated without protection immediately following surgery. Two pigs avoided full weight-bearing on the operated leg for 14 days. By week 4, all pigs fully loaded the operated leg, although 8 pigs still had a slightly uneven stride. By 20 weeks all pigs had a normal stride.

CT-scan – Vessel volume

Seven AV-bundles in the revascularization group were patent and one thrombosed. This pig has been excluded from further analysis. The total vascular and intramedullary vascular volumes were significantly higher in the revascularization group compared to both the ligated group and the contralateral side (Table 2). This demonstrates the beneficial effect of an AV bundle placed within the medullary canal of a cryopreserved allograft. The absolute cortical bone vascular volume was significantly higher in the revascularization group compared to the contralateral tibia, but not significantly so in comparison to ligated controls. All patent bundles showed neovascularization extending into the cortical bone (Figure 3).

Figure 3:

Neovascularization extending into the cortical bone in the revascularized group. Upper left: Neoangiogenesis imaged without the bone by setting the specific threshold for the intravascular Microfil. Upper right: Bone and neoangiogenic vessels seen (in red) in a transverse section. Lower figures: saggital (left) and coronal (right) sections of the tibia. In all different sections is the patent arteriovenous bundle is seen, with neovascularization extending into the cortical bone.

Bone formation: Histomorphometry and Micro-CT

Histomorphometric parameters in the structural allograft are shown in Table 3 and Figure 4, Measures of endosteal bone formation revealed a significant beneficial effect of AV bundle implantation. Bone turnover as measured by the BFR/BV ratio was significantly higher in the revascularization group compared to ligated group. The proportion of actively mineralizing bone surface (the MS/BS ratio) trended higher in the revascularization group. There were no significant differences in MAR. All significant measures of bone formation were found on the endosteal cortical surface, adjacent to the AV bundle. Hematoxylin and eosin stained sections were not substantially different between the two groups (Figure 5).

Table 3.

Histomorphometry Graft of the ligated group and the revascularization group

		Outer cortical area			Inner cortical area
Parameter	Absolute/Normalized	Ligated group (n=8)	Revascularized group (n=7)	p-value	Ligated group (n=8)	Revascularized group (n=7)	p-value
MS/BS (%)	Graft values	138 ± 35	139 ± 28	0.908	66 ± 15	89 ± 29	0.083
	Normalized values (%)	49 ± 36	41 ± 22	0.949	58 ± 17	85 ± 26	0.064
	P value graft vs contralateral	0.012	0.018		0.012	0.128
MAR (μm/day)	Graft values	2.7 ± 0.4	2.4 ± 0.3	0.418	2.6 ± 0.4	2.5 ± 0.3	1.000
	Normalized values (%)	451 ± 830	450 ± 660	0.298	113 ± 21	138 ± 20	0.083
	P value graft vs contralateral	0.012	0.018		0.025	0.018
BFR/BS (μm3/μm2/year)	Graft values	1355 ± 426	1231 ± 282	0.418	631 ± 203	833 ± 316	0.247
	Normalized values (%)	171 ± 181	268 ± 521	0.563	84 ± 47	119 ± 48	0.105
	P value graft vs contralateral	0.779	0.463		0.123	0.753
BFR/BV (μm3/μm3/year)	Graft values	358 ± 235	355 ± 83	0.418	299 ± 143	381 ± 64	0.049
	Normalized values (%)	1315 ± 2624	1780 ± 3195	0.203	180 ± 78	304 ± 58	0.011
	P value graft vs contralateral	0.012	0.028		0.025	0.028

Figure 4.

Dynamic histomorphometry, example of fluorescence slides with a 20× magnification; the slides are taken from the endosteal side of the cortical bone. The green label is calcein and the clear blue label is the tetracycline. The brown staining is the microfil inside the vessels.

A: Patent AV bundle group, B: ligated control group, C: contralateral side.

In the patent group, is an increased amount of fluorescence labels and more empty lacunae seen compared to the control group and contralateral side. These images confirm a higher bone turnover in the revascularized group, with new woven bone and some via.

Figure 5.

Hematoxylin and eosin stained slides with a 20× magnification

A: vascularized group, B: allograft control group, C: contralateral side control.

The black material is Microfil contrast agent filling small vessels. There was no gross descriptive difference between groups on H&E stained slides.

Micro-CT analysis showed no difference in BV/TV or BS/BV ratios between the patent and ligated AV bundle. Both were significantly higher than contralateral tibial values (p=0.02 and p=0.01 respectively).

Bone Healing

The radiographs obtained at 20 weeks revealed 3 ligated group and one revascularization group proximal junction non-union. Failure of proximal fixation appeared to be the cause, demonstrated initially by a halo of bone resorption around proximal screws, seen as early as two weeks postoperatively. In these 4 animals all proximal screws were broken in 3 of the 4 pigs at 20 weeks. The 12 completely healed grafts had either no evidence of fixation problems (6), or changes insufficient to cause fixation failure (3 loose but not broken screws in 2 cases, 2 loose proximal screws in one case, 2 broken screws in one case and one broken screw in 2 instances) (Figure 6).

Figure 6:

Examples typical 20-week postoperative radiographs. On the left, a patent AV bundle with union both proximally and distally, with intact fixation. On the right, a ligated control tibia demonstrating complete nonunion with loss of alignment, broken or loose proximal screws and loss of anatomic reduction at 20 weeks follow up. All proximal screws are either loose or broken

Bone healing scores were obtained on all final radiographs. The average total bone healing score at 20 weeks tended to be higher in the revascularized group compared to the ligated group (21.9 ±2.2 vs 18.9 ± 4.6, P=0.2), with more proximal callus remodeling in the revascularized group versus the ligated group at 20 weeks (3.7 ± 1.0 vs 2.6 ± 1.0, p=0.035).

Complications

There was loss of alignment due to proximal screw loosening in 9 pigs, including 6 ligated animals and 3 in the revascularization group. Despite this, bone apposition exceeded 50% contact in all cases and all but 4 ultimately united.

No wound infection or deep infections occurred in the five months of postoperative survival. In the clinical setting, infections occur in 9% to 30% of structural allograft implantation, generally within the first month postoperatively.^29–31 Our use of occlusive dressings and wound sealant played a role. Risks associated with tumor excision, radiation and/or chemotherapy were of course absent in our study. One initially patent AV-bundle was found to be thrombosed due to graft displacement. This pig was excluded from further analyses. This, loss of alignment and nonunions discussed above were the major complications observed in the 16 animals.

Discussion

Reconstruction of large segmental bone defects with cryopreserved allografts preserves limb function but often at the price of significant morbidity.^{1–8; 32} CBA’s may be closely matched to each defect, are readily available in bone banks and provide good immediate stability. Although freezing reduces immunogenicity,⁹ it also reduces osteogenic and osteoinductive capacity.³³ Revitalization is limited primarily to the allograft/host junctions and periosteal surface.^34–37 CBAs resist infection poorly and have significant rates of nonunion, late stress fracture and mechanical failure.^{8; 29; 38; 39} Failure rates up to 60% at 10 years have been reported.⁷ These shortcomings might be overcome by surgically revascularizing the bone, permitting later active bone remodeling from circulation-derived cells.⁴⁰ Implantation of AV bundles has been demonstrated in small animal models to not only improve cortical bone blood flow, but also formation of new bone .^41–44 Wrapping free vascularized autogenous periosteal flaps is another potential solution, although available sources are limited in size, and are used primarily to envelop small bones such as the clavicle with poor vascularity or segmental loss⁴⁵.

Our data demonstrate that a patent AV bundle implanted within a large cryopreserved bone allograft does undergo angiogenesis, generating new intramedullary vessels that occupy a significantly greater proportion of the bone volume than controls, including ingrowth into endosteal cortical bone, seen on micro-CT angiography. Improved blood flow correlated with significantly greater endosteal allograft bone formation, as measured by quantitative histomorphometric parameters. The outer cortical bone of the allograft did not show similar bone formation rates. This is not unexpected, due to the intramedullary position of the AV bundle. Comparable studies of CBA revascularization in rat femora showed a higher capillary density and a higher mean cortical bone blood flow with AV bundle implantation.^{10; 11} In these small bones, revascularization occurred rapidly, without differences between 4 and 18 week time points¹¹. In the same model, histomorphometry demonstrated higher bone formation rate in both endosteal and periosteal bone at16 weeks.⁴⁴ It is reasonable to assume more time is required for a similar observable effect in a larger animal model.

Revascularization of allografts with periosteal sleeve preservation has also been demonstrated in a similar segmental swine tibial defect model.¹³ In this study, native periosteum was preserved when creating a segmental defect. This was used to envelop a matched allograft segment, with the addition of adipocyte-derived stem cells (ADSC) and bone morphogenic protein type 2 (BMP-2) placed in the medullary canal was tested in a second, otherwise identically-treated group. Both groups were found to heal faster than negative controls lacking periosteum or periosteum with ADSCx and BMP-2. Thus, periosteum served as source of both surgical angiogenesis and new bone formation on the cortical rather than endosteal surface.

Augmentation of surgical angiogenesis by gene therapy^{16; 46} or direct delivery of growth factors ^{15; 47} holds promise in small animal studies. Our porcine model closely approximates clinical scenarios, making it ideal for testing of such methodologies as a bridge to clinical practice. Delivery of blood flow by implantation of an AV-bundle or other vascularized tissue has proven beneficial in treating both avascular necrosis and non-union of small carpal bones clinically.^{22; 42; 48–51} Structural allograft healing is accelerated in long bone segmental defects when combined with a vascularized fibular autograft, due to similar mechanisms. In several series, allografts without fibular flaps required 14 to 23 months to heal^{37; 52–55}. The addition of a vascularized fibula shortens the time to 6-9 months due in part to a similar enhancement of allograft viability ^{8, 55–57}. At 20 weeks, we did not see a significant difference in bone healing scores. It is likely that the survival time was simply too short to demonstrate a difference. A prolonged survival period using this same model would be one logical next step for further research.

There were no major complications in the postoperative period. There were two superficial stitch abscesses but no deep infections in any animal. In clinical setting infection rates between 9% and 30% have been reported.^{29–31; 58} Dick et al reported a percentage of 13.3% deep infections of which 70% occurred within one month of initial surgery. The follow up time of five months in our study was sufficient to identify any potential deep infection. The discrepancy between conventional clinical outcomes and these study results might be due to the clinical setting, most commonly in primary bone tumors and trauma. Hernigou et al reported a higher risk of infection when adjuvant chemotherapy and/or radiation was applied.⁵⁹ The use of occlusive dressings, wound sealant and absence of wound issues related to tumor biopsy or excision, radiation and chemotherapy likely contributed to the lack of any serious deep infection in our series.

Arteriovenous bundles may be ligated distally, as in our study, or reconnected distally as a flow-through pedicle. Thrombosis of ligated AV bundles is lower than a similarly occluded artery alone, and may be further improved by inclusion of adjacent connective tissue.⁶⁰ Circulation is likely maintained by flow between small connecting vessels. Thrombosis was not a significant problem in our study. The single AV bundle that thrombosed was likely the result of damage from loss of bone reduction. The incidence of thrombosis varies widely in published studies. All rat saphenous AV bundles were patent at 16 weeks in one rat study¹⁰ and 85% patent in two others.^{11; 61} Kumta et al. had a patency rate of 50% of the femoral vascular bundle inserted in rat allografts with a follow up time of 6 and 12 weeks.⁶²

We recognize certain limitations of our study. At 6 months age, our pigs were actively growing juveniles of manageable size. Mature pigs may have differed in outcome. Although a twenty weeks’ survival time is substantial, we found cortical bone formation to be incomplete. Subsequent studies with longer survival times would serve to test whether additional bone remodeling would occur in CBAs treated with surgical angiogenesis, or conversely demonstrate rates of stress fracture and reconstructive failures seen currently with cryopreserved structural allograft use.

Implant loosening and loss of alignment were due largely to the effect of immediate unrestricted weightbearing, necessary in the swine model. A single locked plate, precise allograft matching and an intact fibula, felt likely to provide adequate stability in the planning stages of the study, proved inadequate for some animals. We have subsequently changed our methods to include dual plating.

This large animal study of segmental tibial allograft reconstruction demonstrates that implantation of locally-available AV-bundles into cryopreserved allograft segment improves allograft intramedullary blood flow and promotes new bone formation. Future studies using this model to explore the means to augment the rate and extent of cortical angiogenesis using growth factors, gene therapy, stem cells or other therapy may provide the means to further improve outcomes. The study suggests similar methods applied in reconstructing segmental long bone defects clinically may serve to improve clinical outcomes in the future.