Structural bone allografts are a viable option in reconstructing massive bone defects in patients following musculoskeletal (MSK) tumour resection and revision hip/knee replacements. To decrease infection risk, bone allografts are often sterilised with gamma-irradiation, which consequently degrades the bone collagen connectivity and makes the bone brittle. Clinically, irradiated bone allografts fracture at rates twice that of fresh non-irradiated allografts. Our lab has developed a method that protects the bone collagen connectivity through ribose pre-treatment while still undergoing gamma-irradiation. Biomechanical testing of bone pretreated with our method provided 60–70% protection of toughness and 100% protection of strength otherwise lost with conventional irradiation. This study aimed to determine if the ribose-treated bone allografts are biocompatible with host bone.

The New Zealand White rabbit (NZWr) radius segmental defect model was used, in which 15-mm critically-sized defects were created. Bone allografts were first harvested from the radial diaphysis of donor female NZWr, and treated to create 3 graft types: C=untreated controls, I=conventionally-irradiated (33 kGy), R=our ribose pretreated + irradiation method. Recipient female NZWr (n=24) were then evenly randomised into the 3 graft groups. Allografts were surgically fixed with a 0.8-mm Kirschner wire. Post-operative X-rays were taken at 2, 6, and 12 weeks, with bony healing assessed by a blinded MSK radiologist using an established radiographic scoring system. The reconstructed radii were retrieved at 12 weeks and analysed using bone histomorphometry and microCT. Kruskal-Wallis and Mann-Whitney tests were utilised to compare groups, with statistical significance when p<0.05.

Radiographic analysis revealed no differences in periosteal reaction and degree of osteotomy site union between the groups at any time point. Less cortical remodeling was observed in R and I grafts compared to untreated controls at 6 weeks (p=0.004), but was no longer evident by 12 weeks. Radiographic union was achieved in all groups by 12 weeks. Histologic and microCT analysis further confirmed union at the graft-host bone interface, with the presence of mineralising callus and osteoid. Histomorphometry also showed the bridging external callus originated from host bone periosteum and a distinct cement line between allograft and host bone was present at the union site.

Previous studies have shown that the presence of non-enzymatic glycation end products in bone can impair fracture healing. However, these studies investigated bony healing in the setting of diabetic states. Our findings showed that under normal conditions, ribose pretreated grafts healed at rates similar to controls via mechanisms also seen in retrieved human allografts clinically in use. These findings that grafts pretreated with our method are biocompatible with host bone in the rabbit help to further advance this technology for clinical trials.