Evaluation of osteoproductive biomaterials: allograft, bone inducing agent, bioactive glass, and ceramics

This thesis focuses on clinically relevant evaluations of natural and synthetic grafting materials. The intent was to develop a sound methodology to evaluate biomaterials considered for clinical use and to provide an insight on the different grafting stategies to enhance bone formation. The most traditional technique to bridge large defects involves the use of segments of frozen allogenic cortical bone. In this instance, it is crucial that cortical allografts retain biomechanical properties, allowing the repaired bone to withstand forces generated during weight bearing. In the first part of this study, the influence of long-term storage of allografts frozen in saline or dry containers was evaluated in bending and torsion. Within pairs of canine metacarpal bones, metatarsal bones, and ribs, one bone was frozen in a dry plastic container, while the other was immersed in a normal solution of sodium chloride for one year. The energy absorbed at failure and the ultimate displacement of all bones tested in bending were increased by 25%, to 30% and 18%, to 24% respectively, when the bones were frozen in isotonic saline solution. Corticocancellous grafts frozen in normal saline solution are biomechanically less fragile and brittle than grafts stored in plastic without saline solution. Where biomechanical strength is not crucial, fresh autogenous cancellous graft is still considered the "golden standard" when evaluating other biomaterials. However, the morbidity and additional surgical time associated with its collection, as well as the limited supply in each patient, have stimulated the search for substitutes. These have been classified as osteoinductive or osteoconductive, depending on their properties. The second part of the study dealt with an osteoinductive graft substitute extracted from a human osteosarcoma cell line. The osteoinductive properties of this bone-inducing agent (BIA) were tested in the Latissimus Dorsi of five dogs. Because of its clinical relevance, we also evaluated this material after orthotopic implantation, in unicortical femoral defects created in four dogs. The effects of BIA were compared with bovine collagen I (the carrier), a gelatin capsule (negative control), and fresh autogenous cancellous graft (positive control), after orthotopic and heterotopic implantation. Bone formation was evaluated via serial radiographs, dual energy Xray absorptiometry, histology, and histomorphometry. Ten milligrams of BIA did not induce bone formation six weeks after heterotopic implantation and failed to promote bone healing eight weeks after orthotopic implantation in dogs. The discrepancy between these results and those previously obtained in rodents may be related to immunogenic factors or to the dose of BIA used in this study. The third part of the study comprised osteoconductive biomaterials considered for potential use in impaction revision hip arthroplasty in man. The general purpose of this study was to evaluate the use of biomaterials to expand or replace allografts, with a specific interest in the revascularization and incorporation of impacted morselized grafts. In study III, a model was designed to simulate the environment encountered in revision hip arthroplasty and study the biological properties of several impacted aggregates of large particles in the same animal (higher vertebrate). The model was evaluated in Study III and applied to study biomaterials in Studies IV and V. A miniature impactor was designed to produce pellets of aggregates at a compactive effort similar to that generated in clinical impaction grafting. Twenty-two sheep underwent implantation of pellets into six metaphyseal defects in both rear limbs (Study IV). Another eight sheep underwent surgical implantation of four pellets in metaphyseal defects (Study V). Defects were sealed with polymethylmethacrylate in both phases. Healing of the defects was evaluated at seven weeks (n=11) and 14 weeks (n=19) with computed tomography, histology, and 2 histomorphometry. Complications in the initial phase (sheep in Study IV) included four femoral fractures and migration of the cement seal (18/102 defects). No complications occurred in the later phase (sheep in Study V). Although no difference was found between left and right limbs, osteogenesis and incorporation of biomaterials varied between implantation sites. Treatment site allocation was randomized according to a Latin square design for comparison of grafting materials. This model allows evaluation of several impacted aggregates (including large particles) in the same animal. It is particularly suitable for analyzing the biological properties of grafting materials prior to evaluation under loading conditions. The osteoproductive properties of a new silicate-free idealized morselized bioactive glass (Corglaes®), alone or combined with allograft were evaluated in Study IV. Particle size distribution of all aggregates (except the autograft) approached the ideal logarithmic grading line, and all implants were impacted at a standard compactive effort. Treatment groups consisted of (1) empty defect (negative control), (2) autograft (positive control), (3) allograft (clinical control), (4) allograft idealized with Corglaes®, (5) 50 /50 allograft / Corglaes®, and (6) Corglaes®. Healing of the defects was evaluated at seven weeks (n=6) and 14 weeks (n=16) with computed tomography, histology, and histomorphometry. Remnants of Corglaes® were found within one of the defects 7 weeks after implantation. Defects filled with mixtures containing 50% or 100% Corglaes® were less dense, and contained less bone and more fibrous tissue than defects with allograft, autograft, or allograft idealized with Corglaes®. Allograft idealized with Corglaes® may be considered for impaction grafting in revision hip arthroplasty, as well as local delivery of antibiotics. However, further studies and potential revision of the agent are required before mixtures containing concentrations of Corglaes® equal to or higher than 50% can be recommended. In Study V, the model was slightly adjusted to investigate, the properties of idealized, impacted tricalcium phosphate-hydroxyapatite (TCP-HA) aggregates, varying in chemical composition (ratio of TCP to HA) and particle size distribution (idealization with 8 versus 3 particle size ranges). Treatment groups consisted of: (1) allograft (clinical control), (2) 50/50 allograft / 80%HA/20%TCP idealized with 8 particle size ranges, (3) 50/50 allograft / 80%TCP/20%HA idealized with 8 sizes and (4) 50/50 allograft / 80%HA/20%TCP idealized with only 3 particle sizes. At 14 weeks, the pixel value measured with computed tomography in all defects containing synthetic agents was higher than in defects filled with allograft alone (p<0.01). Defects containing the idealized mixture of 80%HA/20%TCP granules (group 2) achieved lower histological scores and contained less bone than the clinical control (p<0.05), whereas groups 3 and 4 did not differ from the control. Although all synthetic agents were osteoconductive, our results suggest that increasing the ratio of TCP over HA and limiting the number of particle size ranges to 3 instead of 8 improve the performance of impacted aggregates as graft expanders. Evaluation under loading conditions of morselized allograft expanded with 80%TCP/20%HA (BoneSave®) in 3 particle size ranges is warranted.