Resorbable porous ceramics derived from chemically converted corals have been used successfully as bone graft substitutes for many years. Converted corals provide a 3D porous architecture that resembles cancellous bone with a pore diameter of 200–700 μm. The success of these corals as a bone graft substitute relies on vascular ingrowth, differentiation of osteoprogenitor cells, remodelling and graft resorption occurring together with host bone ingrowth into the porous microstructure or voids left behind during resorption. The resorption rate of the coral can be controlled by partial conversion to provide a hydroxyapatite (HA) layer via thermal modification. This study examined the resorption rates and bone formation of partially converted corals in a bilateral metaphyseal defect model.

Bilateral defects (5 mm x 15 mm) were created 3 mm below the joint line in the proximal tibia of 41 skeletally mature New Zealand white rabbits following ethical approval. Two variations of a calcium carbonate–HA coral (Pro Osteon 200 R, Interpore-Cross International, Irvine, CA) were examined with different HA thickness (200R; 14% or 200 RT; 28%). Empty defects (negative control) or defects filled with morcellised bone autograft from the defect sites (positive control) were performed. The tibiae were harvested at 6, 12, 24, 36 or 52 weeks, radiographed (standard x-rays and faxitron) in the anteroposterior and lateral planes. Tibias were processed for torsional testing and quantitative histomorphometry using back scattering scanning electron microscopy. Four additional rabbits were killed at time zero to determine the mechanical properties of the intact tibia (n=6 tibias) and 2 for tibias for time zero histomorphometry. Data were analysed using a 3-way analysis of variance.

No clinical complications were encountered in this study. Radiographic assessment revealed a progression in healing, implant resorption and bone infiltration. Cortical closure in the 200 R and 200RT treated defects was noted by 24 weeks. All specimens failed in torsional testing with a spiral fracture initiating at the distal defect site and extending into the distal diaphysis. Torsional properties reached intact control tibia levels by 24 weeks in both groups. No significant differences were noted between 200 R and 200 RT based on torsional data. SEM revealed progressive resorption of the calcium carbonate core of the 200 R and 200 RT with time, infiltration of bone and ingrowth to the HA layers. Time and measurement site (cortical versus cancellous) were significant for implant resorption, bone, and void. The thinner HA layer (200 R) resorbed more quickly compared to the thicker layer (200 RT) in the canal as well as cortical sites. Increased bone and decreased void were noted at the cortex measurement sites in the 200 R group at 24 weeks and in the 200 RT group at 12 and 24 weeks (p< 0.05). Implants were nearly completely resorbed by 52 weeks with only a few percent of implant remaining.