Lateral ankle stabilization after distal fibular resection using a novel approach: a surgical technique.

BACKGROUND: After tumor resection involving the distal fibula, the method for recreating stability of the lateral ankle remains controversial. Many reconstructive options exist, including allograft reconstruction and arthrodesis; however, each of these approaches has significant potential disadvantages. DESCRIPTION OF TECHNIQUE: The distal fibula is resected as necessary to obtain negative margins for local control of the neoplasm. Reconstruction of the lateral ankle ligamentous complex is performed using the peroneus brevis tendon to reestablish lateral and anterior stability of the tibiotalar joint. The peroneus brevis tendon is transected proximally at it myotendinous junction and then sutured to the calcaneofibular and anterior talofibular ligaments in sequence and then tenodesed to the lateral distal tibia with suture anchors and a staple. METHODS: We present three patients who underwent distal fibulectomy for tumors originating in the distal fibula. All patients who have undergone the reconstruction being described are included within this cohort study. The patients were assessed clinically and radiographically at a range of 14 months to 9.5 years (average, 4.8 years) for functional recovery, return of range of motion, stability of the ankle, and imaging evidence of arthrosis and instability. RESULTS: There were no episodes of instability or early progression to arthrosis. In addition, all patients obtained excellent ankle stability and range of motion on examination, but two had complications. One sustained a traumatic fracture to the base of the fifth metatarsal that healed with nonsurgical treatment and another who underwent further fibular shortening and bursectomy at the tip of the residual fibula with complete relief of his symptoms. CONCLUSIONS: Reconstruction of the lateral ankle after distal fibular resection is possible using the peroneus brevis tenodesed to the distal tibia and sutured to the remnants of the calcaneofibular and anterior talofibular ligaments as described in this surgical technique. In this small group, we found that patients were able to return to normal daily activities without instability or progression to tibiotalar arthrosis at short term; however, longer followup and larger series of patients are called for to confirm these findings.

Lower dose of rhBMP-2 achieves spine fusion when combined with an osteoconductive bulking agent in non-human primates.

STUDY DESIGN: A non-human primate lumbar intertransverse process arthrodesis model was used to evaluate the effectiveness of different formulations of recombinant human bone morphogenetic protein-2 (rhBMP-2) to induce consistent bone formation. OBJECTIVE: To determine if the combination of rhBMP-2/absorbable collagen sponge (ACS) wrapped around a bulking agent, consisting of a biphasic calcium phosphate/collagen composite, could achieve posterolateral spine fusion with a dose of rhBMP-2 (3.0 mg/side) that previously failed to induce posterolateral fusion in rhesus monkeys with other carriers. SUMMARY OF BACKGROUND DATA: Successful bone induction in both human and non-human primates has required higher concentrations of BMP than were required in lower order models. The Food and Drug Administration approved concentration of rhBMP-2 for interbody fusion (1.5 mg/mL) when delivered on the ACS alone without a bulking agent (doses 3-9 mg/side) has failed to induce clinically relevant amounts of bone formation in a posterolateral spine fusion model in rhesus monkeys. Previously, a higher concentration of 2.0 mg/mL of rhBMP-2 delivered on stacked sheets of a biphasic calcium phosphate ceramic/collagen compression resistant matrix (CRM) was required to achieve fusion in the rhesus monkey and was the basis for this study (doses of 6-12 mg/side). METHODS: Nine skeletally mature, rhesus macaque monkeys underwent single level posterolateral arthrodesis at L4-L5. Two different rhBMP-2 doses were evaluated in 3 delivery configurations. The first 3 monkeys received 10 mg/side (2.5 mL at 4.0 mg/mL) of rhBMP-2 loadeddirectly onto a CRM carrier (15% hydroxyapatite/85%beta-tricalcium phosphate ceramic/collagen matrix), resulting in a final concentration of 2.0 mg/mL. The second 3 monkeys received 3 mg/side (2.0 mL at 1.5 mg/mL) of rhBMP-2 loaded directly on the CRM carrier, resulting in a 0.6 mg/mL final concentration. Three additional monkeys also received the 3 mg/side (2.0 mL at 1.5 mg/mL) of rhBMP-2 delivered on an ACS, which was then wrapped around the dry CRM matrix used as a bulking agent, yielding a 1.5 mg/mL final concentration of rhBMP-2 on the sponge wrapped around the bulking agent. The monkeys were euthanized at 24 weeks after surgery. Manual palpation, plain radiographs, computerized tomography, and nondecalcified histology were used to evaluate fusion in a blinded fashion. RESULTS: The 3 monkeys with 10 mg rhBMP-2 placed directly on the CRM carrier (2.0 mg/mL final concentration) achieved solid fusion. The 3 monkeys that underwent fusion with 3 mg of rhBMP-2 placed directly on the CRM carrier (0.6 mg/mL final concentration) failed to achieve fusion. In contrast, the 3 monkeys that underwent fusion with the same 3 mg dose of rhBMP-2 dispensed only on an ACS that was wrapped around the CRM achieved solid bilateral fusion. CONCLUSIONS: This study shows the importance of carrier optimization and final implant protein concentration for the successful delivery of rhBMP-2. By combining the properties of the ACS with the CRM, the required dosage of rhBMP-2 was diminished by more than 3-fold in the non-human primate model. This finding suggests that the currently available concentration of rhBMP-2 (1.5 mg/mL) could be successful for achieving posterolateral spine fusion when combined with an osteoconductive bulking agent that can support the induced new bone formation.

Comparison of Healos/bone marrow to INFUSE(rhBMP-2/ACS) with a collagen-ceramic sponge bulking agent as graft substitutes for lumbar spine fusion.

STUDY DESIGN: A rabbit lumbar intertransverse process arthrodesis model was used to evaluate the efficacy of two different bone substitute materials: 1) collagen-hydroxyapatite sponge (Healos bone void filler) combined with heparinized bone marrow; and 2) recombinant human bone morphogenetic protein-2 delivered in a collagen sponge (INFUSE Bone Graft) wrapped around an additional collagen-ceramic sponge (Mastergraft Matrix) as a "bulking agent." OBJECTIVES: To compare the relative efficacy of two different bone graft substitutes to achieve posterolateral lumbar spine fusion in rabbits. SUMMARY OF BACKGROUND DATA: Autogenous bone graft is considered the gold standard graft material for spine fusion. Complications with its use, however, may occur in as many as 30% of patients. A variety of bone substitutes have been used for spine fusion, but there are few direct comparison experiments to determine the relative efficacy of any two alternatives. METHODS: Adult New Zealand white rabbits (n = 24) were divided into two groups and underwent bilateral posterolateral intertransverse process spine arthrodesis at L5-L6. The fusions were augmented by different bone substitute materials as follows: Group 1 (n = 12) received 3 mL of collagen-hydroxyapatite sponge (Healos bone void filler) (10 x 30 x 5 mm, two per side) with 3.0 mL of heparinized bone marrow on each side of the spine. (ratio 1:1); Group 2 (n = 12) received 1.5 mL of rhBMP-2 (0.43 mg/mL solution) on a Type 1 collagen sponge (INFUSE Bone Graft) wrapped around an additional 1.5 mL collagen-ceramic (15%HA/85%TCP) sponge (Mastergraft Matrix) as a bulking agent to provide 3 mL of graft on each side of the spine. Bone marrow was aspirated from posterior iliac crest, and 1 mL of bone marrow was sent to count number of nucleated cells. The rabbits were killed after 8 weeks; the spines were evaluated by manual palpation, radiographs (plain radiograph and CT scan), tensile mechanical testing, and nondecalcified histology. RESULTS: The bone marrow had average of total nucleated cell count 9 x 10 cells. All rabbits (100%) in Group 2 (INFUSE/Mastergraft Matrix) achieved solid spinal fusions by manual palpation and radiographs, whereas solid spinal fusion was not achieved by manual palpation and radiographs in any of the rabbits treated with Healos combined with heparinized bone marrow (Group 1). The plain radiograph and CT scans of Group 1 showed some minimal new bone formation near the transverse processes, but none of these rabbits formed a continuous fusion mass. In contrast, all of plain radiographs and CT scans in Group 2 showed continuous fusion mass and complete graft incorporation between transverse processes bilaterally. Biomechanically, the relative strength and relative stiffness values of L5-L6 (fusion segment) in Group 2 were statistically significant greater than L5-L6 in Group 1 (P < 0.001). Histologic sections confirmed the palpation and radiographic results. CONCLUSION: From the manual palpation, radiographic and biomechanical assessment of fusion, the results in this study showed that INFUSE (rhBMP-2/collagen sponge) consistently produced spine fusion when wrapped around a collagen-ceramic sponge bulking agent (Mastergraft Matrix). Meanwhile, Healos was ineffective as a bone graft material when combined with heparinized autogenous bone marrow.

The effects of doxorubicin (adriamycin) on spinal fusion: an experimental model of posterolateral lumbar spinal arthrodesis.

BACKGROUND CONTEXT: Malignant spinal lesions may require surgical excision and segmental stabilization. The decision to perform a concomitant fusion procedure is influenced in part by the need for adjunctive chemotherapy as well as the patient's anticipated survival. Although some evidence exists that suggests that chemotherapy may inhibit bony healing, no information exists regarding the effect of chemotherapy on spinal fusion healing. PURPOSE: To determine the effect of a frequently used chemotherapeutic agent, doxorubicin, on posterolateral spinal fusion rates. STUDY DESIGN/SETTING: Prospective animal model of posterolateral lumbar fusion. OUTCOME MEASURES: Determination of spinal fusion by manual palpation of excised spines. Plain radiographic evaluation of denuded spines to evaluate intertransverse bone formation. METHODS: Thirty-two New Zealand White rabbits underwent posterior intertransverse process fusion at L5-L6 with the use of iliac autograft bone. Rabbits randomly received either intravenous doxorubicin (2.5 mg/kg) by means of the central vein of the ear at the time of surgery (16 animals) or no treatment (16 animals; the control group). The animals were euthanized at 5 weeks, and the lumbar spines were excised. Spine fusion was assessed by manually palpating (by observers blinded to the treatment group) at the level of arthrodesis, and at the adjacent levels proximal and distal. This provided similar information to surgical fusion assessment by palpation in humans. Fusion was defined as the absence of palpable motion. Posteroanterior radiographs of the excised spines were graded in a blinded fashion using a five-point scoring system (0 to 4) devised to describe the amount of bone observed between the L5-L6 transverse processes. Power analysis conducted before initiation of the study indicated that an allocation of 16 animals to each group would permit detection of at least a 20% difference in fusion rates with statistical significance at p=.05. RESULTS: Eleven of the 16 spines (69%) in the control group and 6 of the 16 spines (38%) in the doxorubicin group fused. This difference was statistically significant (=.038). There was no significant correlation (p>.05) between the radiographic grade of bone formation (0 to 4) and fusion as determined by palpation. There were four wound infections in the control group and four in the doxorubicin group. However, solid fusions were palpated in three of these four spines in both the control and treatment groups. CONCLUSIONS: No significant differences in wound complications were noted with doxorubicin administration. A single dose of doxorubicin administered intravenously at the time of surgery appears to play a significant inhibitory role in the process of spinal fusion. If similar effects occur in humans, these data suggest that doxorubicin may be harmful to bone healing in a spine fusion if given during the perioperative period. Further investigation will be necessary to determine the effect of time to aid at determining whether doxorubicin administered several weeks pre- or postoperatively results in improved fusion rate, and whether bone morphogenetic proteins can overcome these inhibitory effects.

Evidence of osteoinduction by Grafton demineralized bone matrix in nonhuman primate spinal fusion.

STUDY DESIGN: A nonhuman primate posterolateral lumbar intertransverse process arthrodesis model was used to evaluate osteoinductive bone graft materials. OBJECTIVES: To test two new formulations of Grafton demineralized bone matrix (Flex and Matrix) for evidence of osteoinduction and their potential efficacy as an extender or enhancer for autogenous bone in a previously validated nonhuman primate posterolateral lumbar fusion model. SUMMARY OF BACKGROUND DATA: Whereas several demineralized bone matrix formulations have been shown to be variably osteoinductive in rodent ectopic bone assays and rabbit spine applications, few have demonstrated efficacy in higher species and in more challenging applications such as posterolateral spine fusion. The authors are not aware of any published studies describing any demineralized bone matrix that has been tested in a nonhuman primate posterolateral spine fusion model. METHODS: After approval by the institutional animal care and use committee, eight skeletally mature rhesus macaques underwent single-level posterolateral arthrodesis. In four animals, autograft (4 g/side) was implanted with a piece of human Grafton Flex demineralized bone matrix. In the other four animals, rhesus Grafton Matrix demineralized bone matrix, a new and more porous formulation of Flex, was implanted with autograft (4 g) on one side of the spine, and Matrix with half the amount of autograft (2 g) was implanted on the opposite side. Radiographs were taken at intervals until the animals were killed at 24 weeks. Spinal fusion was evaluated by manual palpation (status was fused or not fused), and computed tomography was done to visualize the amount of bone formation. RESULTS: Fusion was ascertained by palpation in two of four monkeys receiving Flex with autograft and in three of four monkeys receiving Matrix with autograft. Evidence of osteoinduction was seen in all four monkeys on the Matrix with 4 g autograft side, which had larger fusion masses than in the other treatments. Histologic examination showed that the bone formed was normal. CONCLUSIONS: The rhesus Matrix formulation performed better than the human Flex. Evidence of osteoinduction was seen in all four monkeys that received Matrix, which improved the fusion success of autograft. This alone suggested that it might play a role as a graft enhancer, not merely as a graft extender. Human studies are warranted.

Tensile properties of fresh human calcaneal (Achilles) tendons.

The purpose of this study was to measure the tensile properties of fresh human calcaneal (Achilles) tendons. Twenty fresh cadaveric (age range = 57-93 years) bone-Achilles tendon complexes were harvested within 24 hr postmortem. The calcaneus together with 15 cm of the Achilles tendon extending proximally from the insertion on the calcaneus was clamped and biomechanically tested. Each tendon was firmly fixed in clamps in an MTS Systems Corporation MTS testing machine and tension was applied at a displacement rate of 8 cm per minute until the tendon failed. The tensile force and tensile strain (as measured using an extensometer) were recorded and plotted using onboard software. The narrow age range of our donors prevented any meaningful correlation between age and tensile properties; however, the results showed that: 1) the average ultimate tensile strength (UTS) of the human Achilles tendon was 1189 N (range = 360-1,965), 2) there was a correlation between left and right legs for UTS, 3) there was a correlation between left and right legs in regard to cross sectional area, and 4) there was no correlation between UTS and cross-sectional area.

Simple carrier matrix modifications can enhance delivery of recombinant human bone morphogenetic protein-2 for posterolateral spine fusion.

STUDY DESIGN: A nonhuman primate lumbar intertransverse process arthrodesis model was used to evaluate modifications to a plain collagen sponge to deliver recombinant human bone morphogenetic protein-2 (rhBMP-2). OBJECTIVES: To evaluate the feasibility of enhancing the delivery of rhBMP-2 with the established collagen sponge carrier by adding biphasic ceramic phosphate (BCP) granules (15% hydroxyapatite, 85% tricalcium phosphate) or allograft chips to provide compression resistance for posterolateral spine arthrodesis. SUMMARY OF BACKGROUND DATA: Recombinant human bone morphogenetic protein-2 was successfully delivered with a resorbable collagen sponge in a rabbit intertransverse process fusion model. Success in nonhuman primates required a higher dose (6-9 mg) of rhBMP-2 and a more compression-resistant matrix (ceramic) than plain collagen. The limitation of the ceramic carrier was its radiopacity, which made radiographic detection of new bone formation difficult. METHODS: Nine adult rhesus monkeys underwent bilateral posterolateral intertransverse process arthrodesis at L4-L5. The animals were divided into three groups (n = 3 each) based on the graft material implanted: 1) autogenous iliac crest bone (5 cm3/side); 2) collagen sponge and 15:85 BCP granules loaded with rhBMP-2 (3 mg/side); and, 3) collagen sponge and allograft chips loaded with rhBMP-2 (3 mg/side). The monkeys were killed 24 weeks after surgery. Inspection, manual palpation, radiography, computed tomographic scans, and histology were used to assess fusion. RESULTS: All six monkeys with rhBMP-2 delivered in the collagen/15:85 BCP carrier and the collagen/allograft chips carrier achieved solid spine fusions, whereas only one of three animals fused with autogenous bone graft. Histologic analysis of the bone induced by rhBMP-2 showed normal trabecular bone and bone marrow elements. CONCLUSIONS: The addition of either 15:85 BCP granules or allograft bone chips to the existing resorbable collagen sponge matrix enhanced delivery of rhBMP-2 in the posterolateral spine. The combination matrices were more compression resistant and had improved radiographic resorption properties that permitted easy radiographic visualization of new bone. In addition, a lower dose of rhBMP-2 (3 mg/side) was successful compared with the dose previously used with the plain collagen sponge (6 mg/side).

Overcoming the immune response to permit ex vivo gene therapy for spine fusion with human type 5 adenoviral delivery of the LIM mineralization protein-1 cDNA.

STUDY DESIGN: An animal study in immune competent rabbits and athymic rats was conducted. OBJECTIVES: To develop an animal model for simulation of previous human Type 5 adenovirus (Ad5) exposure, to determine the impact of adenoviral pre-exposure on spine fusion induced with ex vivo Ad5-LMP-1, and to test strategies for overcoming any potential immune response. SUMMARY OF BACKGROUND DATA: Cells transduced with adenovirus containing the osteoinductive LMP-1 cDNA (Ad5-LMP-1) can induce spine fusion in rabbits. Because up to 80% of the human population has been exposed to adenovirus, immune responses to the vector may limit this strategy in humans. Few studies have modeled previous adenoviral exposure and tested strategies to circumvent it. METHODS: Adult New Zealand white rabbits were injected with 10 or 10 viral particles of Ad5-LacZ. At 4 or 16 weeks after Ad5 injection, autologous buffy coats were prepared from peripheral blood, and 4 million cells per side were infected ex vivo for 10 minutes with Ad5-LMP-1 (multiplicity of infection = 4). Cells were implanted on a collagen matrix instead of an autograft for posterolateral lumbar arthrodesis. Unimmunized rabbits served as control subjects. Additional immunized rabbits underwent arthrodesis at 4 weeks with increased cell number (10 million) and viral dose (multiplicity of infection = 10), or with both parameters increased. The rabbits were killed at 4 weeks, and the spines were assessed by palpation and radiograph. A parallel study was performed in athymic rats using immunized rabbits for the donor cells. RESULTS: All the unimmunized rabbits had solid spine fusions. None of the rabbits arthrodesed 4 weeks after Ad5 pre-exposure achieved fusion. At 4 weeks after Ad5 exposure, increasing the multiplicity of infection to 10 did not overcome the immune response (0/3 fused), but increasing the cell number to 10 million (2/3 fused) or increasing both cell number and multiplicity of infection (3/3 fused) did overcome the immune effects. Delaying arthrodesis until 16 weeks after Ad5 pre-exposure also overcame the immune response (3/3 fused). Similar results were seen in the athymic rat ectopic implant model, suggesting that the immune effect was mediated by humoral antibodies rather than a T-cell response. CONCLUSIONS: Two model systems were developed that simulate previous exposure to human Ad5 and could separate the cellular and humoral components of the response. There was a dose-dependent inhibition of ex vivo Ad5-LMP-1 gene transfer to cells from animals previously exposed to human Ad5. Data suggested that the inhibition of Ad5 infection was caused by humoral antibodies rather than a T-cell-based response. Minor modifications in the gene transfer protocol, such as doubling the viral dose or number of cells infected, or increasing the infection time, could overcome the immune response for an ex vivo approach.

Retention of 125I-labeled recombinant human bone morphogenetic protein-2 by biphasic calcium phosphate or a composite sponge in a rabbit posterolateral spine arthrodesis model.

The purpose of this study was to characterize the retention kinetics of recombinant human bone morphogenetic protein-2 (rhBMP-2) applied to two calcium-based delivery matrices. Biphasic calcium phosphate (BCP) and a composite containing BCP in an absorbable collagen sponge (BCP/ACS) were evaluated using a spinal fusion model in rabbits. rhBMP-2 labeled with radioactive iodine (125I) was used as a tracer to assess in vivo retention of rhBMP-2 in the presence of these materials (nine animals per material studied). Over a 36 day study period, animals were assessed for the following: percent administered dose retained at the implant site as measured by scintigraphic imaging (counting) with a gamma camera (all animals), radiography of the implant site (all animals), radioactivity in blood and plasma (all animals), and radioactivity in the urine and feces (three animals for each material). Radioactivity data were corrected for the decay of 125I and the attenuation between the implant in vivo and the gamma camera. Differences observed between the two materials for the area under the retention vs. time profile (AUC; 988%*day for BCP vs. 1070%*day for BCP/ACS, p = 0.57) and the mean residence time (MRT; 10.2 days for BCP vs. 7.6 days for BCP/ACS, p = 0.06) were not statistically significant. Initial retention/incorporation of rhBMP-2 was slightly higher for rhBMP-2/BCP/ACS than for rhBMP-2/BCP (96.8% vs. 86.0%, p < 0.05). Animals receiving rhBMP-2/BCP showed a longer terminal retention half-life (t1/2) than did those receiving rhBMP-2/BCP/ACS (7.5 vs. 4.5 days, p < 0.05). The urinary radioactivity recovery data supported the data obtained by scintigraphy. Over the 36 day collection period, essentially complete recovery of radioactivity (dose) in urine was observed for rhBMP-2/BCP and rhBMP-2/BCP/ACS and the majority of the radioactivity (approximately 95%) was soluble in trichloroacetic acid, suggesting extensive catabolism of rhBMP-2 before renal excretion. Fecal recovery of radioactivity was low, approximately 2-3%. In conclusion, rhBMP-2 was retained at the implant site when delivered with either BCP or BCP/ACS based on mean residence time and area under the retention curve vs. time profile. Use of these matrices resulted in detectable rhBMP-2 levels at the surgical site for over a week in contrast to data reported with several other matrices that lasted less time. Systemic catabolism and elimination of the rhBMP-2 was extensive and systemic presence of the protein was negligible.

Delivery of recombinant human bone morphogenetic protein-2 using a compression-resistant matrix in posterolateral spine fusion in the rabbit and in the non-human primate.

STUDY DESIGN: A rabbit and rhesus monkey model of posterolateral intertransverse process spine arthrodesis was used. OBJECTIVE: To test two new soft tissue compression resistant ceramic/collagen sponge carriers for recombinant human bone morphogenetic protein-2. SUMMARY OF BACKGROUND DATA: After determining that a plain collagen sponge was too compressible for large animals in a posterolateral fusion application, the authors demonstrated good bone induction using biphasic ceramic phosphate granules (60% hydroxyapatite/40% tricalcium phosphate) as the carrier matrix for recombinant human bone morphogenetic protein 2 in rhesus monkeys. A limitation of 60:40 biphasic ceramic phosphate was its slow resorption time caused by the high hydroxyapatite content, making radiographic detection of new bone formation very difficult. METHODS: Adult New Zealand white rabbits (n = 14) underwent posterolateral spine arthrodesis at L5-L6 using 5:95 biphasic ceramic phosphate (5% hydroxyapatite/95% tricalcium phosphate) impregnated Type I collagen sponges (17 x 35 x 2.5 mm, two per side) loaded with 0.86 mg recombinant human bone morphogenetic protein 2. Additional rabbits (n = 14) received 60:40 hydroxyapatite-tricalcium phosphate granules as the carrier for bone morphogenetic protein 2. Adult rhesus monkeys (n = 6) underwent posterolateral arthrodesis at L4-L5 with ceramic/collagen sponge carrier loaded with 9 mg recombinant human bone morphogenetic protein 2 per side. Two monkeys received ceramic/collagen sponges containing 15:85 biphasic ceramic phosphate (15% hydroxyapatite/85% tricalcium phosphate) with two pieces per side; two received sponges containing 5:95 biphasic ceramic phosphate with two pieces per side, and two received sponges containing 5:95 biphasic ceramic phosphate with four pieces per side. The rabbits were killed after 5 weeks and the monkeys after 24 weeks; the spines were evaluated by manual palpation, radiographs, tensile mechanical testing (rabbits only), and histology. RESULTS: The recombinant human bone morphogenetic protein 2 delivered in the 5:95 biphasic ceramic phosphate/collagen sponge achieved fusion in 100% of rabbits and had improved handling properties compared with the biphasic ceramic phosphate granules. Biomechanical results with 5:95 biphasic ceramic phosphate/collagen carrier were comparable to those obtained with the 60:40 biphasic ceramic phosphate granules and superior to those of autogenous bone graft (P < 0.05). The recombinant human bone morphogenetic protein 2 delivered in the 15:85 or the 5:95 biphasic ceramic phosphate/collagen sponge carrier (two pieces per side) induced fusion in nonhuman primates with normal bone histology, less residual ceramic, and more bone in the center of the carrier matrix in comparison with BCO granules alone. The 15:85 biphasic ceramic phosphate/collagen sponge resulted in fusion mass sizes closer to the original size of the matrix implanted than did the 5:95 biphasic ceramic phosphate/collagen sponge, which was considered a desirable feature. The monkeys with 9 mg recombinant human bone morphogenetic protein 2 spread over four sponges per side instead of two had half the effective recombinant human bone morphogenetic protein 2 concentration per sponge and inferior results. CONCLUSIONS: The new compression-resistant biphasic ceramic phosphate/collagen sponge matrices were biologically compatible with recombinant human bone morphogenetic protein 2 bone formation, resulted in biomechanically stiffer fusion masses than autograft, better space maintenance than plain collagen sponges, and improved handling and radiographic resorption properties over the ceramic carriers previously tested.