Spine fusion is pivotal in the correction and prevention of progression of spinal deformity and in the treatment of spinal instability, trauma, and reconstruction. Lumbar fusion ranks as the second most common lumbar spine procedure, with more than 250,000 spinal arthrodeses done each year in the United States.13 Posterolateral lumbar intertransverse process fusions are the most common type of fusion done. Efforts to enhance the success of lumbar spinal fusion continue to evolve.13
Biologic and biosynthetic augmentations have become routine adjuvants to enhance spinal fusion. Autogenous corticocancellous bone graft procured from the iliac crest has become the traditional standard for these procedures. It contains determined osteogenic cells, osteoinductive bone matrix proteins and osteoconductive bone mineral, and collagen. Various supplements have gained acceptance to promote fusion when the quality or quantity of autogenous graft is lacking. These supplements include allograft, demineralized bone matrix (DBM), and segmental spinal fixation systems.14 Despite the use of these alternatives, failure to achieve a solid bony union occurs in 5–35% of patients with single-level fusion and more frequently when multiple levels are attempted.4,16 The rate of reported nonunions for lumbar fusions using autograft and segmental instrumentation has improved but is in the 5–15% range. These nonunions frequently lead to unsatisfactory results and may require the need for additional surgeries. Recently, the Food and Drug Administration approved the use of recombinant human bone morphogenetic protein-2 (rhBMP-2) for use in lumbar interbody fusion; however, high costs may preclude its widespread use.
Normally, after placement of autogenous bone graft in the intertransverse process fusion bed, the following events occur: (1) appositional bone growth on the graft; (2) osteoclastic bone resorption of the incorporated graft; and (3) progressive remodeling of the graft. If there is a delay in incorporation of the autogenous bone graft (such as occurs in patients who smoke), osteoclasts will remove the corticocancellous graft before vascular invasion and formation of new bone occur, leading ultimately to nonunion. Therefore, an agent that delays removal of the bone graft from the fusion bed may help facilitate arthrodesis.
Bisphosphonates have been shown to inhibit bone resorption. Alendronate sodium is a third-generation bisphosphonate that has been shown at therapeutic levels to inhibit bone resorption at the bone-mineral interface without compromising bone formation.9,12,15,19 The interplay of bone turnover in favor of formation leads to an increase in mineral content of native bone. The decrease in bone resorption is accompanied by an increase in calcium balance and in the mineral content of bone. As a result, these compounds have found widespread acceptance in the prevention and treatment of osteoporosis in humans.2,13,17,18 In addition, preclinical studies document that alendronate has no deleterious effect on fracture healing.7,8–13,15,19
In theory, alendronate also should be successful in enhancing the biologic process of spinal fusion. This action may occur by preventing early resorption of the corticocancellous autograft at the fusion site. In addition, many elderly patients having fusion have osteoporosis and need bisphosphonate treatment. There is good correlation between the potencies of bisphosphonates on inhibiting bone resorption in rats and humans. Therefore, we tested the following hypotheses: (1) our pseudarthrosis model would show lack of radiographic union by 6 weeks after surgery; (2) a therapeutic dose of alendronate would result in stiffer spines as assessed by manual palpation and higher radiographic fusion scores at early times compared with the high dose of alendronate group and with controls; (3) high-dose alendronate would lead to no change in bone area fraction, controls would show a decrease in bone area fraction with time, and the therapeutic dose of alendronate would show less change than controls with time; and (4) the numbers of osteoclasts would decrease with alendronate in a dose-dependent manner.
MATERIALS AND METHODS
The dose of 1 μg/kg/week was based on the therapeutic equivalent for the clinical treatment of osteoporosis translated to the rat model. The 10 μg/kg/week dose was chosen as 1 log dose over therapeutic dose to ascertain the effect of a log dose in this model.
Alendronate sodium was obtained from Merck Pharmaceuticals (West Point, PA). Ten milligrams were reconstituted in phosphate-buffered saline to give a concentration of 100 μg/mL.
The Institutional Animal Care and Use Committee at our institution approved all animal experimentation. Sixty-three male Sprague-Dawley rats with a mean weight of 345 g (range, 320–415 g) and an average age of 12 weeks (range, 10–14 weeks) were housed under a 12-hour day and 12 hour night light condition with 60% humidity at 21°C. The anesthetic consisted of intraperitoneal administration of 40–80 mg/kg ketamine and 5–10 mg/kg xylazine. Maintenance of anesthesia was done using oxygen (2 L/minute) and halothane through a face mask. Monitoring was done by a pulse oximeter placed on the right forelimb. The surgical site was prepared by shaving the entire dorsum free of hair and then it was scrubbed twice with Betadine (Purdue Pharma, LP, Stamford, CT) soap and alternatingly wiped with 70% isopropyl alcohol. Two milliliters of 0.5% Marcain (AstraZeneca, Wilmington, DE) was infiltrated over the incision site. Buprenorphine 0.05–0.075 mg/kg was administered intramuscularly at the end of the procedure. The rats were placed on a water circulating heat pad and covered with soft linen until they awoke. The groups of animals receiving treatment received weekly subcutaneous injections of alendronate sodium in two different doses (1 μg/kg/week; ATherapeutic and 10 μg/kg/week AHigh-dose). Control animals received similar subcutaneous injections of normal saline. At term, animals were sacrificed by carbon dioxide asphyxiation.
The bone graft was obtained from the tailbones through caudectomy before the exposure of the spine. A dorsal incision was made over the coccyx at (Co) 3–8 after which the tail was transected between the caudal vertebrae at Co6. A V-shaped soft tissue flap was preserved for closure with 3–0 absorbable sutures. The transected tail then was dissected through a longitudinal incision. The soft tissue was dissected off using tissue forceps, rongeurs, and a scalpel. Consistently, we procured three tail bone vertebrae and morselized them longitudinally using bone-cutting rongeurs and scissors. The bone graft was weighed to quantify the harvested bone. The average weight of bone graft per rat was 0.25 g. Our laboratory previously found that seven tail bones yields a fusion rate of approximately 70% in control animals; our aim was to achieve an approximately 40% healing rate; therefore we used ½ the number of tail bones.
After procurement of the bone graft, a 4-cm longitudinal skin incision was made on the dorsal lumbar midline. After exposure of the dorsolumbar fascia, bilateral muscle splitting incisions were done to expose the transverse processes. A dental burr was used to decorticate the posterior surface of the transverse process. After irrigation, the autograft derived from the tail was placed in the defect using a 1-mL syringe cut at the tip to allow graft extrusion. After the surgical procedure, the wound was lavaged thoroughly with saline before graft placement (deep layer), the fascial layer closed with 4.0 Vicryl (Ethicon, Inc, Somerville, NJ) in continuous fashion, and then the skin was closed (superficial layer) in standard interrupted fashion with 3.0 proline suture.
The lumbar spinal columns were excised with care taken to avoid disturbing the fusion mass. The spinal column was dissected free of soft tissue to provide a clear radiographic image. The specimens then were radiographed in the frontal projection using the Faxitron xray apparatus (Model 8050-010; Faxitron xray Corporation, Wheeling, IL). For optimal observation of autograft resorption, bony continuity, and new bone formation, the specimens were exposed for 8 minutes at 50 kVp. An Al step wedge control was used as an internal control for intensity. It also helped maintain the aspect ratio of subsequent images and their distance calibration. For analysis, the radiographs were digitized using a digital camera (Kodak DCS-420c) with a resolution of 1524 × 1012 pixels.
Radiographs were quantified using a 5-point (0–4) fusion score by three blinded observers as follows: 0 = no bone, 1 = evidence of resorption with sclerosis and increased intertransverse density, 2 = evidence of autograft resorption with less than 25% intertransverse area consisting of bridging trabeculae, 3 = evidence of resorption and remodeling with as much as 50% intertransverse area consisting of bridging trabeculae, and 4 = clearly defined external bony cortex with greater than 75% intertransverse area consisting of bridging trabeculae.
Manual dissection of explanted specimens was done, leaving intact the anterior and posterior longitudinal ligaments, interspinous ligaments, and intervertebral discs. The qualitative evaluation done by three blinded observers involved the manipulation of the L-4 and L-5 level and examination of any relative motion between L-4 and L-5 on either left or right sides. Results were either scored as fused or not fused.
Spinal motion segments were fixed for 24 hours in 10% neutral buffered formalin and then decalcified in 5% nitric oxide for 4 days. After decalcification, specimens were sectioned in the midsagittal plane and dehydrated sequentially in 95% alcohol for 2 days, two changes of 100% alcohol for 2 days, and finally cleared in xylene for 1 day. Thereafter, specimens were embedded in paraffin and sectioned at a thickness of 5-μm. The slides were stained using hematoxylin and eosin and Goldner’s trichrome. Numbered consecutive sections were stained with number-matched stains. Sections were studied under a light microscope by a blinded bone pathologist. Images were recorded and then analyzed quantitatively for percentage-area of fibrous and bony tissue using the Metamorph Image Analysis and Processing System (Version 4.0, Universal Imaging Co, Downingtown, PA) computer program. A square grid field with 400 μm × 400 μm dimensions was placed over the fusion segment (Fig 1). The number of osteoclasts, the number of osteoblasts, and the bone surface were determined per field and the percent osteoblasts per bone surface were calculated as an index of osteoblastic activity between groups.
Variables were analyzed by two-factor ANOVA with interaction. The factors were dose with three levels and time with three levels (Table 1). Additional testing for differences between dose levels was done with adjusted Bonferroni comparisons. Fusion rates were analyzed by chi square tests. The alpha level was set at 0.05 for all tests. Analyses were done with SYSTAT version 8.0 statistical software (SYSTAT Software, Inc, Richmond, CA).
There were no medication-related complications in any of the animals. There were no neurologic injuries, infections, or wound complications.
The validity of the pseudarthrosis model was confirmed by radiographic analysis. More than 70% of animalshad a score of 2 or less in the control group at all three times (Table 2). At all times, the animals that received an overdose of alendronate had radiographic pseudarthroses (scores ≤ 2). The animals in the therapeutic-dose group had scores greater than 2 in more than ⅓ of animals. There were no statistically significant differences in score distribution with dose. There were no significant differences in percent fused determined by manual palpation (Table 3) or radiographic fusion scores among the dose groups.
There was a dose-related and time-related difference in the area fraction of unremodeled bone within the fusion mass among groups. The total area of residual unremodeled bone in the fusion mass was calculated as a percentage of the total fusion mass area by quantitative histologic analysis. There was no statistical difference in residual bone area within the fusion mass among dose groups at 2 weeks. At 4 weeks, the high-dose alendronate group had a greater (p < 0.001) total area fraction of residual bone compared with the control group and therapeutic groups. At 6 weeks, the high-dose alendronate group still had a greater (p < 0.001) area fraction of residual bone; however there also was a significant (p = 0.001) difference between the ATherapeutic group and control group at the same time (Fig 2).
Long-term administration of alendronate suppressed osteoclast and osteoblast function. At 2 weeks, there were fewer (p < 0.001) osteoclasts per high-powered field by manual count in the high-dose group, but the number of osteoclasts did not differ between control and therapeutic dose groups (Fig 3). However, at 4 and 6 weeks, the number of osteoclasts was suppressed (p < 0.01) in the high-dose and therapeutic groups compared with the control group. The percent osteoblasts per bone surface was lower (p < 0.05) in the high-dose group at all times compared with the other groups (Fig 4).
On qualitative histologic analysis, an early healing response was observed in all groups (Fig 5). At 6 weeks, there was a marked increase in the residual thickness of the transverse process cortex in the ATherapeutic group, representing an inhibition of bony remodeling. There still was active osteoclastic resorption occurring in the control group at 6 weeks compared with an almost complete lack of osteoclasts in the AHigh-dose group (Figs 6–8).
Failure of spinal fusion surgery because of psuedoarthrosis is not an uncommon problem. As discussed earlier, the advent of instrumentation and the use of biologic adjuvants to surgery have decreased nonunion rates; however failure to fuse remains a substantial risk particularly in the posterolateral environment. Any agent that could retain biologically active autogenous graft at the desired fusion site for a longer time potentially could lead to an enhanced fusion rate. This study investigated the hypothesis that the use of alendronate sodium could slow osteoclastic activity enough to allow retention of bone graft at the fusion bed in a pseudarthrosis model, therefore enhancing arthrodesis.
There are several limitations associated with this study. The model chosen was a challenging rat-spine pseudarthrosis model that has a healing rate of only 50%. A log lower dose from the standard therapeutic dose was not taken. Also, more rats in the early 2-week time might have shown a considerable difference in stiffness rates by manual palpation. We quantified our radiographic and manual palpation data using a scoring scheme with different sets of observers. Although this may lead to substantial variability in score interpretation, care was taken to ensure all observers were familiar with the scoring schemes and adequately blinded at the time of data collection.
Bisphosphonates are well-established therapeutic agents used clinically to retard resorption at the bone-osteoclast interface for various bone diseases including Paget’s disease, tumor-induced hypercalcemia, and metastatic bone diseases.5,6 Perhaps the best-described use has been in the prevention of the progression of postmenopausal osteoporosis. Alendronate (4-amino-1-hydroxybutylidene bisphosphonate sodium salt) is a potent amino bisphosphonate that binds to hydroxyapatite mineral on bone resorption surfaces, thereby inhibiting osteoclast-mediated bone resorption.15 Alendronate has been shown to increase the bone mineral density (BMD) of the lumbar spine and hip. Several randomized double-blind clinical trials have shown that alendronate decreases the incidence of new vertebral fractures and nonspine fractures in postmenopausal women with osteopororsis.1–3
Peter et al11 investigated the effect of alendronate on fracture healing and bone remodeling in a canine fracture healing model. Orogastrically delivered alendronate at a dose of 2 mg/kg/day was administered for 9 weeks preceding fracture, 16 weeks after fracture, or before and after fracture at 25 weeks. Their results showed at 16 weeks the calluses in dogs that received alendronate were approximately two to three times larger than those in dogs that received a placebo during the healing period. This finding was consistent with a slow rate of callus remodeling. Administering alendronate did not inhibit bone formation and mineralization. Mechanical testing showed that ultimate load at failure of the fractured and contralateral intact bone was unaffected by treatment with alendronate.
The role of alendronate in particle-induced osteolysis was examined by Millett et al.9 A specially designed polyethylene (PE) implant was implanted in the tibiae of 72 animals. Four weeks after implantation, the animals were randomized into three separate groups: (1) control group, (2) prevention group, or (3) treatment group. In the prevention group, animals received intraarticular injections of high-density PE particles at 4, 6, and 8 weeks postoperatively. Alendronate, at a dose of 0.01 mg/kg/day was administered through an implantable subcutaneous pump from the fourth week through the tenth week. Treatment group animals also were exposed to PE particles at the same times to induce bone loss, but they received alendronate subsequently, from Week 10 through Week 16. Positive (particle only) and negative (saline solution only) control groups also were assessed. Tissues were harvested at 10 weeks (prevention group) and at 16 weeks (treatment group). Histologic examination showed a rim of new bone around the implant in the control animals. Animals exposed to PE particles had bone loss and those that received alendronate had bone loss either prevented or reversed, which was confirmed by histomorphometric analysis. Millett et al concluded that alendronate effectively preserved periprosthetic bone and prevented bone loss in their osteolysis model.
A similar study investigated the effect of alendronate on bone ingrowth into anterior interbody spinal fixation devices. Zou et al20 investigated the effect in 18 pigs (nine in each group) that had anterior intervertebral lumbar spinal fusion at L2–L3, L4–L5, and L6–L7. Each level was randomly allocated to one of three implants: (1) a solid piece of porous tantalum, (2) a porous tantalum ring, or (3) a carbon fiber cage. The tantalum ring and the carbon fiber cage were packed with autograft. Alendronate was delivered orally to one of the groups at a daily dose of 10 mg from postoperative Day 2 until the end of the experiment at 3 months. Histologic examination showed that the original graft was replaced entirely by new trabecular bone in both groups. The bone volume fraction in the tantalum cage groups was larger in the pigs given alendronate than in the controls. Zou et al concluded that short-term, low-dose alendronate treatment did not impair the formation of new bone but increased bone ingrowth into the central hole of the porous tantalum ring and the pores of the porous tantalum in this porcine model.
The previously described studies examined the role of alendronate in fracture healing,11 particle-induced osteolysis,9 and interbody fusions.20 The posterolateral intertransverse spinal fusion model is a difficult healing environment compared with the cited studies. Our study was a posterolateral pseudarthrosis model, therefore, we studied the effect of alendronate in a challenging spinal healing situation.
In our study, there was no noteworthy difference in fusion based on manual palpation among the control and therapeutic groups at 4 and 6 weeks. The AHigh-dose group had pseudarthroses by radiographic, histologic, and manual palpation criteria at all times. Residual unremodeled bone area was higher at 6 weeks by histomorphometric analysis in the AHigh-dose group. This total bone area represented a volume of unremodeled autograft persisting as long as 6 weeks after the surgical procedure. Transverse process cortex comparison at 6 weeks clearly showed a residual increase in the cortex thickness in the ATherapeutic group with poor osteoblastic and osteoclastic activity in the AHigh-dose group. This can be coupled with the poorer radiographic and manual palpation fusion scores at 6 weeks in AHigh-dose compared with other groups at each time.
Similarly, osteoclast and osteoblast numbers in the AHigh-dose group declined acutely after fusion and administration of the drug. In the ATherapeutic group, there was an almost linear decrease in osteoclast numbers with time and in the control group, after the inflammatory phase subsided the osteoclast numbers also were reduced. However, this decrease in osteoclasts did not carry over to a decrease in osteoblasts; there was no difference in percent osteoblasts in the control group and ATherapeutic at all times. In this model, alendronate at ultrahigh doses inactivates osteoblasts and inhibits osteoclasts.
The effect of alendronate seems to depend on dose and animal model. Our study results confirm previous observations in models of bone healing that the administration of alendronate prevents rapid resorption and remodeling of healing bone. At therapeutic doses, alendronate does not seem to have a deleterious effect in this model of pseudarthrosis. However, at supernormal doses alendronate has an inhibitory effect on osteoclastic and osteoblastic activity.
We thank and acknowledge Fengyu Zheng, MD for his technical assistance in this project.
1. Adachi JD, Bensen WG, Brown J, et al. Intermittent etidronate therapy to prevent corticosteroid-induced osteoporosis. N Engl J Med
2. Black DM, Cummings SR, Karpf DB, et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Lancet
3. Chavassieux PM, Arlot ME, Reda C, et al. Histomorphometric assessment of the long term effects of alendronate on bone quality and remodeling in patients with osteoporosis. J Clin Invest
4. DePalma AF, Rothman RH. The nature of pseudoarthrosis. Clin Orthop
5. Fleisch H. Bisphosphonates: Pharmacology and use in the treatment of tumor-induced hypercalcemic and metastatic bone disease. Drugs
6. Hosking DJ, Eusebio RA, Chines AA. Paget’s disease of bone: Reduction of disease activity with oral risedronate. Bone
7. Lenehan TM, Balligand M, Nunamaker DM, et al. Effect of EHDP on fracture healing in dogs. J Orthop Res
8. Li J, Mori S, Kaji Y, et al. Effect of bisphosphonate on fracture healing of long bone in rats. J Bone Miner Res
9. Millett PJ, Allen MJ, Bostrom MP. Effects of alendronate on particle-induced osteolysis in a rat model. J Bone Joint Surg
10. Nyman MT, Paavolainen P, Lindholm TS. Clodronate increases the calcium content in fracture callus. Arch Orthop Trauma Surg
11. Peter CP, Cook WO, Nunamaker DM, et al. Effect of alendronate on fracture healing and bone remodeling in dogs. J Orthop Res
12. Rodan GA, Fleisch H. Bisphosphonates: Mechanisms of action. J Clin Invest
13. Russell RG, Roger MJ. Bisphosphonate: From the laboratory to the clinic and back again. Bone
14. Sandhu HS, Boden SD. Biologic enhancement of spinal fusion. Orthop Clin North Am
15. Sato M, Grasser W, Endo N, et al. Bisphosphonates in action: Alendronate localization in rat bone and effects on osteoclast ultrastructure. J Clin Invest
16. Steinmann JC, Herkowitz HN. Pseudoarthrosis of the spine. Clin Orthop
17. Storm T, Thamsborg G, Steiniche T, et al. Effect of intermittent cyclical etidronate therapy on bone mass and fracture rate in women with postmenopausal osteoporosis. N Engl J Med
18. Seeman E. Osteoporosis: Trials and tribulations. Am J Med
19. Tarvainen R, Olkkonen H, Nevalainen T, et al. Effects of clodronate on fracture healing in denervated rats. Bone
© 2004 Lippincott Williams & Wilkins, Inc.
20. Zou X, Xue Q, Li H, et al. Effect of alendronate on bone ingrowth into porous tantalum and carbon fiber interbody devices: An experimental study on spinal fusion in pigs. Acta Orthop Scand