Bone growth into porous materials has proven to be a very effective method for attaching prosthetic implants to the osseous skeleton1-3. However, there remains a need to develop modalities that can accelerate and/or increase biologic fixation. The more rapidly that bone forms and the greater the amount of bone that forms about and/or within an implant, the faster the implant becomes mechanically secured against the disruptive forces of load bearing and the sooner patients can safely return to their activities of daily living. In situations in which the condition of the bone stock or the healing process is compromised (e.g., when the patient is elderly or has osteoporosis) or the initial stability of the implant is more tenuous or crucial (e.g., during posttraumatic, revision, or minimally invasive procedures), the construct would clearly benefit from enhanced mechanical support.
Bisphosphonates reduce bone catabolism by interfering with cell metabolism and causing osteoclast apoptosis; the resulting net gain in bone formation explains their widespread use in treating osteoporosis. There is also the possibility that some bisphosphonates have a catabolic effect by direct action on osteoblasts, although the information is variable4-6. There is growing evidence that bisphosphonate compounds can be utilized to modify the peri-implant bone response in favor of enhanced bone formation and implant fixation. This has been shown in numerous clinical and basic research studies involving the oral, systemic, and local delivery of bisphosphonates7-26. Local delivery is a sensible approach because it reduces the amount of drug used and preferentially targets the site of interest, thereby avoiding systemic exposure. Previous studies in rats and dogs have shown that direct elution of zoledronic acid from implants increased net local bone formation, but the results have been limited primarily to twelve weeks or less and the dose response is not yet fully characterized14,19,20.
We hypothesized that the extent of peri-implant net bone gain is dose dependent and that the bone gain persists over the long term. The purpose of this study was to quantify the effect of local delivery of different doses of zoledronic acid on bone growth within and about porous tantalum implants one year after surgery.
Implants that measured 9 mm in diameter and 90 mm in length (Fig. 1) were manufactured for use in a canine femoral intramedullary model. The implants were made of porous tantalum (Trabecular Metal; Zimmer, Warsaw, Indiana), a metallic biomaterial that is approximately 80% porous, with a mean pore size of about 450 µm27,28. The implants were plasma-spray-coated with a 10 to 15-µm layer of hydroxyapatite, the composition of which was 98% pure, 99% dense, and 64% crystalline and had a calcium-to-phosphate ratio of 1.67. As previously described, the hydroxyapatite served to partially immobilize the zoledronic acid through its chemical affinity for calcium phosphate14,19. After use of this delivery system, Tanzer et al.14 and Roberts et al.29 described a biphasic zoledronic acid elution profile consisting of an initial burst release within a few hours followed by a much slower, protracted, and progressive release over many weeks.
Zoledronic acid was utilized because it is a third-generation compound that is considered to be the most potent aminobisphosphonate, as much as 1000 times more potent than pamidronate30. Commercially pure zoledronic acid (Novartis Pharmaceuticals, Basel, Switzerland) was dissolved in distilled water, after which a 1-mL aliquot containing either 0.05 mg or 0.20 mg of zoledronic acid was systematically and evenly applied to implants with use of a micropipette. The implants were dried overnight in an oven at 37°C and sterilized with use of ethylene oxide. The implants were then surgically inserted through the piriformis fossa and into the femoral medullary canal with use of the same surgical technique that is used for open intramedullary nailing.
Bilateral surgery was performed on ten adult mongrel dogs. On one side, all ten dogs received a control implant; on the contralateral side, five of the dogs received an implant that was dosed with 0.05 mg of zoledronic acid and five received an implant that was dosed with 0.20 mg of zoledronic acid. After one year, the twenty femora were harvested, stripped of soft tissue, radiographed, and processed for undecalcified thin-section histologic analysis. Each bone-implant construct was sectioned transversely with a low-speed diamond cut-off apparatus (Buehler, Lake Bluff, Illinois) at 1-cm intervals to produce eight sections for subsequent analysis. The sections were contact-radiographed with use of a machine (Faxitron X-Ray, Lincolnshire, Illinois) and then digitally imaged with use of backscattered scanning electron microscopy. Semiautomated software was used to analyze the images for five parameters: (1) the percentage extent of bone ingrowth, calculated by dividing the total area of bone islands within the implant perimeter by the void space, exclusive of the area occupied by the tantalum struts (Fig. 2); (2) the percentage of bone apposition, calculated by dividing the total length of bone lying on an imaginary line traced exactly at the perimeter of the implant by the length of this line (Fig. 2); (3) the percentage filling of the medullary space exterior to the implant, calculated only for diaphyseal sections by dividing the total area of bone islands within the medullary canal by the area of the medullary canal, not including the implant (Fig. 3); (4) the cortical porosity, calculated only for diaphyseal sections by dividing the total area of intracortical pores, excluding Haversian canals, by the cortical area; and (5) the diaphyseal cortical area, calculated only for diaphyseal sections simply by measuring the bone envelope within periosteal and endosteal limits.
Statistical comparisons of these parameters were made between the control and the zoledronic-acid-dosed implants and between the 0.05-mg and the 0.20-mg zoledronic-acid-dosed implants with use of multiple two-level hierarchical models, with the level of significance set at 0.05.
Source of Funding
Financial support for the study was provided by the Canadian Institutes of Health Research.
In general, the anteroposterior and lateral radiographs of the twenty harvested femora did not demonstrate any marked differences in peri-implant bone formation between the control and zoledronic-acid-dosed implants. In total, 160 histologic sections (eighty from the ten control implants and eighty from the ten zoledronic-acid-dosed implants) were examined with use of contact radiography and backscattered scanning electron microscopy. The contact radiographs and backscattered scanning electron micrographs revealed varying degrees of peri-implant bone within the medullary canal in all sections (Figs. 4 and 5). This bone was typically more abundant in sections of zoledronic-acid-dosed implants; it had the appearance of remodeled trabecular bone and could be seen in both continuous and discontinuous clusters of varying density. The net bone formation was evident in both metaphyseal and diaphyseal sections and appeared as a localized response, largely confined to within a radius of about 3 mm of the implant edge.
Qualitative analysis of the backscattered scanning electron micrographs revealed that the gray-scale and general appearance of bone within and immediately about the implants was indistinguishable from native cortical and cancellous bone (Figs. 2, 3, and 4). Although there was more bone about the zoledronic-acid-dosed implants in seventy-seven of the eighty section pairs, the amount and difference varied among animals, as indicated in Figures 4 and 5 and in Tables I, II, and III.
Quantitative analysis with backscattered scanning electron microscopy revealed that bone ingrowth was present in all sections (Table I) and was generally more abundant nearer the implant perimeter than deeper inside the implant (Figs. 2, 4, and 5). The difference in the mean extent of bone ingrowth between control implants and paired implants dosed with 0.05 mg of zoledronic acid (1.2%, 95% confidence interval = -2.5% to 5.6%) was not significant. The difference in the mean extent of bone ingrowth between control implants and paired implants dosed with 0.20 mg of zoledronic acid (3.9%, 95% confidence interval = 1.6% to 6.2%) was significant. Comparison of the paired differences in bone ingrowth between the implants dosed with 0.05 mg of zoledronic acid and the implants dosed with 0.20 mg of zoledronic acid (2.7%, 95% confidence interval = -2.3% to 7.1%) showed that no meaningful effect of zoledronic acid dose could be identified in this small sample.
In every dog, the mean amount of bone apposition was greater in the implants with either dose of zoledronic acid than it was in the paired control implants (Table II). The difference in the mean amount of bone apposition between control implants and paired implants dosed with 0.05 mg of zoledronic acid (11.3%, 95% confidence interval = 6.3% to 16.4%) was significant (Table II). The difference in the mean amount of bone apposition between control implants and paired implants dosed with 0.20 mg of zoledronic acid (14.0%, 95% confidence interval = 6.7% to 20.3%) was also significant (Table II). The relative differences were similar, with 52% mean greater apposition for implants that had been dosed with 0.05 mg of zoledronic acid and 50% mean greater apposition for implants that had been dosed with 0.20 mg of zoledronic acid. However, when the paired data between the implants with 0.05 mg of zoledronic acid and the implants with 0.20 mg of zoledronic acid were compared, with the numbers studied, there was no apparent effect of dose on the difference in bone apposition (2.6%, 95% confidence interval = -6.1 to 10.7).
In every dog, there was greater mean filling of the medullary canal with bone about the implants that had been dosed with zoledronic acid as compared with the mean amount of filling in the paired control implants (Table III and Figs. 4 and 5). The difference in the mean amount of canal filling between control implants and paired implants that had been dosed with 0.05 mg of zoledronic acid (7.7%, 95% confidence interval = 2.1 to 13.7) was significant (Table III). The difference in the mean amount of canal filling between the control implants and the paired implants that had been dosed with 0.20 mg of zoledronic acid (10.6%, 95% confidence interval = 3.7 to 17.4) was also significant (Table III). The relative differences were somewhat similar, with 57% mean greater canal filling for implants that had been dosed with 0.05 mg of zoledronic acid and 74% mean greater canal filling for implants that had been dosed with 0.20 mg of zoledronic acid. In comparing the paired data between the implants that had been dosed with 0.05 mg of zoledronic acid and the implants that had been dosed with 0.20 mg of zoledronic acid, with the numbers studied, there was no apparent effect of dose on the difference in filling of the canal with bone (2.9%, 95% confidence interval = —6.1 to 11.7).
Mean intracortical porosity for all femora fell within a narrow range of 0.92% to 1.04%, with no significant differences within or between groups of dogs. There were also no significant differences in cortical area between the femora that received the zoledronic-acid-dosed implants and those that received the control implants.
Note: The porous tantalum implants were donated by Zimmer, Inc., Warsaw, Indiana, and the zoledronic acid was donated by Novartis Pharmaceuticals, Basel, Switzerland.