The option to apply axial loading through dynamization is available in several devices used for intramedullary and external fixation of long bone fractures. 1,10,11,17,18 It has been suggested that dynamic loading enhances fracture healing by stimulating the osteogenic response during the healing process, but the exact biologic mechanism for such an effect is unknown. 8,11,19,23 One hypothesis has been that the stress arising from elastic deformation of bone when placed under physiologic loading enhances proliferation of periosteal callus during the early phase of bone healing and accelerates the remodeling process during the final part of the healing sequence. 11,14,24 Whether the direction of the dynamic load is of importance has not been fully investigated. However, in specific animal species and bone locations, axial loading has been shown to stimulate bone formation, whereas no such effect was observed after bending forces. 23
For externally stabilized fractures, there are at least three different types of dynamization that can be identified based on the properties of the fixator system used. Elastic dynamization occurs when the fixator frame and the pins deform under physiologic loading. 6 On unloading, the original fracture gap resumes. The second type of dynamization is described as axial dynamization when the external fixator frame allows free axial movement through a telescoping mechanism incorporated in the body of the fixator, while bending and rotation are prevented. 11 This type of dynamization, as used in the current experiment, allows the fracture gap to close and remain closed even in the absence of externally applied axial compression. The most sophisticated way to achieve dynamic loading across the fracture gap is through a powered actuator attached to the pins or the fixator frame capable of closing the fracture gap or loading the bone ends dynamically, irrespective of physiologic and functional loading. This type of mechanical stimulation is referred to as controlled dynamization. 14
Experimental studies on the effect of axial dynamization on bone fracture healing pathways and on restoration of the mechanical characteristics of fractured bone have had mixed results. 2,13 A controversy has arisen between clinical studies claiming significant improvement in bone healing when using dynamization 11,19 and experimental studies showing only a limited direct effect on the healing process. 2,3,14 The main limitation of previous experimental studies dealing with axial dynamization is that the bone tissue response has been studied at one time, usually late in the bone repair process. In the current study, a time-sequenced experimental design was used to investigate the potential effect of early axial dynamization on the healing process of experimental tibial fractures (osteotomies) in dogs, as opposed to studying only delayed axial dynamization at one time at the end of the bone fracture repair and remodeling process. 2,14,29
The hypothesis addressed was whether axial dynamization might alter the cellular and tissue changes that occur within the periosteal and endosteal callus during the early stages of fracture healing. More specifically, the question was asked whether such an effect will be reflected in the clinical, radiographic, densitometric, histologic, and biomechanical characteristics at five intervals from the initial fixation phase to final healing.
MATERIALS AND METHODS
Forty-four adult mixed-breed dogs (weight, 23–35 kg) were used for the study. All animals were used in accordance with the principles of the Guide for the Care and Use of Laboratory Animals, as compiled by the Institute of Laboratory Animal Resources, National Academy of Sciences. 16 The skeletal maturity of each dog was assessed by radiographs to ensure closure of the proximal tibial epiphysis. A standardized transverse middiaphyseal osteotomy was done on both tibias in each dog to allow paired comparisons of the results. At surgery, the osteotomies were stabilized with a 2-mm gap using a half-pin external fixator (Fig 1A). One week after surgery, the telescoping mechanism of one randomly selected fixator on each dog was unlocked to allow free axial movement, whereas the contralateral side was kept rigidly fixed with a 2-mm gap throughout the study (Fig 1B). The animals were allowed free unrestricted load bearing after surgery. Static and dynamic functional loading were monitored weekly. Anteroposterior (AP) and mediolateral radiographs were taken every week to assess periosteal callus formation. An oral daily dose of 500 mg of tetracycline was given from the time of surgery until sacrifice for continuous labeling of new bone formation. Groups of seven animals were euthanized 1 day and 1, 3, and 5 weeks after dynamization, and two groups with eight animals each were sacrificed at 8 and 11 weeks after axial dynamization. For future reference, animal groups will be identified according to weeks after surgery, unless indicated otherwise. Thus, the current study involved six animal groups: the 1-, 2-, 4-, 6-, 9-, and 12-week groups.
External Fixator Design and Rigidity
The external fixator used was a custom-made version of the Orthofix unilateral 6 half-pin external fixator (model 1410 Orthofix SRL, Verona, Italy), which included a roller-bearing telescoping sleeve to allow axial compression of the osteotomy when subjected to functional loading or soft tissue contraction. The dimensions of the fixator and the distance from the fixator body to the bone (the pin length) were constant throughout the study. The fixation pins were made of stainless steel, with a shank diameter of 6 mm and a threaded portion tapering from 4.5 mm to 3.5 mm. Three pins were used in each of the two pin brackets.
A separate in vitro study done on intact tibias from weight-matched dogs and bone-fixator constructs showed that in torsion, mediolateral and AP bending, there was no significant difference in stiffness whether the fixator was locked with a 2-mm gap (control) or unlocked (dynamized). Under bending and torsion, the stiffness with the fixator, unlocked or locked, was significantly lower than that with intact bone. Under axial compression, the stiffness with the unlocked fixator was not different when compared with the intact bone, whereas the locked fixator had significantly lower axial stiffness than did the intact bone and the unlocked dynamized fixator (Fig 2). This type of external fixation and pin configuration was rated as a rigid type in axial direction in the absence of axial dynamization when compared with other external fixators and pin designs used in previous experiments. 9,15,20,26,29
Surgical Technique and Postoperative Care
The dogs were anesthetized with sodium pentobarbital (intravenous 30 mg/kg). Both hindlimbs were prepared and draped in a sterile fashion. The fixation pins were inserted through both cortices in an anterolateral position after they were predrilled with a 3.2-mm drill using a custom-made alignment jig centered over the tibia. A longitudinal medial incision and subperiosteal elevation were done to expose the midtibia. With the guidance of an alignment jig, and under saline for irrigation and cooling, a transverse tibial osteotomy was done at the midpoint between the proximal and distal fixation pins using a Stryker saw (Model 1370, Stryker Corp, Kalamazoo, MI). The fibula was fractured manually, and a small section of the fibula shaft adjacent to the fracture site was resected to avoid interference with the bone ends in the tibial osteotomy when placed under axial loading. The pin brackets of the fixator were separated by 95 mm over the osteotomy, and the fixator body was positioned 50 mm away from the centerline of the midtibia. A 2-mm gap was achieved by using a thickness gauge as a spacer between the bone ends during adjustment, followed by locking the telescoping mechanism built in the fixator body. The soft tissue incision was closed using a standard suture technique.
Each animal was evaluated daily for signs of pin tract drainage and infection. Seven days after surgery, the telescoping mechanism of one randomly selected fixator in each dog was dynamized. Both fixators were inspected on a weekly basis to assure that the sliding mechanism on the fixator body was rigidly locked (the control side) or free to slide (the dynamization side). There were no instances with slippage of the fixators throughout the study.
Standard and uniform AP and mediolateral radiographs were taken using a special canine hindlimb positioning jig immediately after surgery, at 1 week before and after dynamization, and once every week thereafter with the dogs in the supine nonweightbearing position. During the radiographic procedure, the fixators were clamped on the custom-designed jig to ensure consistent positioning of the tibias in relation to the film to facilitate gap distance (dynamization effect) and callus area measurements. The width of the osteotomy gap and the amount and the location of periosteal callus formation were measured on each radiograph with a sonic digitizer (accuracy ± 0.2 mm) (Graf/Pen Sonic Digitizer, Science Accessory Corp, Southport, CT).
After surgery, load bearing was not restricted. Load bearing on the osteotomized hindlimbs was determined before surgery and once a week starting 1 week after dynamization by measuring static and dynamic (push-off load) load bearing on a Kistler force plate during gait. During the dynamic testing, the animals were kept on a leash with a technician walking along to ensure a consistent walking speed of approximately 1 m/second. Six valid maximum ground reaction force measurements were recorded for each hindlimb at each time. 5
Torsional tests were conducted for the 4-, 6-, 9-, and 12-week groups. After sacrifice, each tibia was transected through the metaphyses to provide a standard 11-cm segment of diaphyseal bone centered around the osteotomy site. During this procedure, one specimen from the 4-week group accidentally was bent at the osteotomy site, leading to exclusion of the whole pair from the mechanical testing. Thus, six pairs in the 4-week group were tested. The bone ends, including the pinholes, were embedded in Wood’s metal (Cerro Metal Product Co, Bellefonte, PA), leaving 5 cm of the tibial middiaphysis with the osteotomy site at its center. Specimens were tested in axial torsion by internal rotation at 15° per minute until bone failure on an Instron Electromechanical Testing Machine (Model 1125, Instron Corp, Canton, MA). 7 The test was stopped as soon as the bone failed to avoid gross displacement and facilitate reapproximation of the bone ends for histologic analysis. The low rate of loading was chosen to exclude the time-dependent effects of the bone and the healing tissues. The maximum torque before fracture was determined directly from the resulting torque versus rotation curve. Torsional stiffness was defined as the slope of the initial linear portion of the curve. The biomechanical stage of fracture union was determined according to White et al. 25
After mechanical testing, the specimens were reapproximated and embedded in methylmethacrylate for undecalcified histologic analysis. The 2-week group also was used for undecalcified histologic analysis without mechanical testing. Longitudinal sections were cut in the mediolateral plane and ground to an approximate thickness of 100 μm. The slides were surface stained with a modified Paragon stain and analyzed with a light microscope. Custom-designed, semiautomatic computer-assisted image analysis software was used to determine tissue composition in the periosteal and endosteal callus. The area investigated included all regions 5 mm proximal and distal to the ends of the osteotomy gap. Stained sections were projected with transmitted light that magnified the image on the digitizer table. The system was calibrated using standard grids overlaid on the sections. The callus was divided graphically into areas of bone, cartilage, fibrous tissue, and undifferentiated marrow tissue as assessed by the staining characteristics.
For the 1-week group, undecalcified thin sections were prepared without conducting mechanical testing. The transverse 5-μm sections were cut with a microtome (Reichert-Jung 1140, Reichert Scientific Instruments, Buffalo, NY) and stained with a modified Goldner’s trichrome stain. The sections were examined with a light microscope.
New bone formation in longitudinal histologic sections was evaluated in the periosteal and endosteal regions using ultraviolet light microscopic analysis with a Laser Confocal Microscope (L.S.M., Zeiss, Thornwood, NY) and the IBAS 20 video-based image analysis system (IBAS 20, Kontron, Munich, Germany). Samples were evaluated on the basis of the brightness of the label after thresholding for the tetracycline label and the unlabeled areas to detect new bone formation and its temporal occurrence.
A paired Student’s t test was used for evaluation of differences between the dynamized and control sides at each time. Time-sequential changes in the periosteal callus area measured in the plain radiographs and load bearing in the individual dogs in the 12-week group and biomechanical and histologic parameters at different times were analyzed using analysis of variance (ANOVA) with Tukey-Kramer’s post hoc t test. Mean and standard error of the mean are given.
There were no infections at the osteotomy site. Superficial infections were seen within 3 weeks after surgery in five of 528 pin tracts. Two of the infections occurred in the dynamized side, whereas three of the infections were found in the control side in animals from the 4-, 6-, and 9-week groups. The infections healed with local antiseptic treatment. There were no deep infections.
The gaps of the dynamized osteotomies closed significantly from 1.95 ± 0.11 mm before dynamization to bone contact immediately after unlocking of the telescoping mechanism. At 1 week, the gap on the control side was 1.94 ± 0.11 mm, which represented a slight reduction compared with 2.11 ± 0.12 mm measured immediately after surgery. No additional reduction in gap occurred after the first postoperative week. The total amount of periosteal callus on the AP and mediolateral radiographs increased with time. The periosteal callus area was higher (p < 0.05) 4 weeks after surgery and showed a strong tendency to be higher (p = 0.078) 5 weeks after surgery in the dynamized side. Nine weeks after surgery, the periosteal callus area showed a strong tendency to be higher (p = 0.079) in the control side.
The time-sequential changes in the periosteal callus area from 2 weeks to 12 weeks in the individual animals showed a significant increase at 4 weeks (p < 0.05) in both sides and leveled off between 6 and 8 weeks in the dynamized side and between 8 and 10 weeks in the control side with the maximum values, respectively (Fig 3). After the week when the callus area reached a maximum in each side, the area decreased significantly in the dynamized side at 12 weeks (p < 0.05), but the decrease was not statistically significant on the control side (Fig 3). The circumferential distribution of periosteal callus was significantly nonuniform (p < 0.05), with more callus on the lateral and posterior aspects of the tibia from Week 7 for the dynamized limbs and from Week 8 until the end of the study on the control side.
Before surgery, a mean of 21% of the total body weight was borne on each hindlimb during standing, whereas the peak vertical force during gait averaged 75% of the body weight. Load bearing during standing (static) and during gait (dynamic) was significantly greater on the dynamized limbs than on the control limbs for the first 5 and 3 weeks, respectively, after dynamization (p < 0.05). After this time, there was no significant difference between the sides (p > 0.05). Twelve weeks after surgery, static and dynamic loading still were significantly reduced when compared with preoperative loading (p < 0.04 for static and p < 0.03 for dynamic loading;Fig 4).
For dynamized osteotomies and controls, the mechanical strength of the healing osteotomy improved throughout the study (p < 0.0001). The torsional stiffness of the dynamized side was significantly higher (p < 0.05) than that of the control side 6 weeks after surgery (Fig 5A). There was a trend toward increased maximum torque (p = 0.14) in the dynamized side when compared with the control side at the same period (Fig 5B). Before and after the 6-week period, the difference in mechanical properties based on a paired comparison between dynamized and control legs was not significant. Twelve weeks after surgery, the maximum torque of the dynamized and control osteotomies was 75% and 68% of values for intact bones.
The fracture lines after torsional testing occurred through the osteotomy site for all specimens in the 4-and 6-week groups. Thus, their biomechanical fracture healing stages could be classified as Type I or Type II according to White et al. 25 At 9 weeks, four specimens (two each of dynamized and controls) were classified as Type III, whereas Type IV healing was seen in five bones in the 12-week Group (four in the dynamized group and one specimen in the control group).
Eight days after surgery (1 day after dynamization), the gap area was filled with an organized blood clot and isolated from the periosteum and endosteum on the control side (Fig 6A). The periosteum and the endosteum were repaired at the area adjacent to the osteotomy site and attached to the cortex. On the dynamized side, mesenchymal tissue in the medullary canal was forced out through the original gap to the periosteal area with a strong fiber orientation (Fig 6B). The organized blood clot filling the gap area was pushed away to the periosteal region. The periosteum was detached from the cortical bone anchor. Highly oriented longitudinal fibers were observed in the medullary canal.
There was no significant difference in the tissue composition of the periosteal or endosteal callus between dynamized osteotomies and controls at all times. Fibrous tissue was predominant in the periosteal and endosteal areas for the dynamized and control osteotomies during the first 2 weeks, after which the amount of such tissue declined. In the periosteal area, the amount of cartilage increased to a maximum at 4 weeks after surgery, with a mean of 11% on the control side and 6% on the dynamized side; the difference between the groups was not significant. At 9 weeks, the amount of cartilage in the periosteal area had decreased in both groups, although the amount of cartilage was significantly higher in the control side (6.3%) than in the dynamized side (2.3%) (p < 0.05). Corresponding figures for endosteal cartilage at 4 weeks were 18% for the control side and 10% for the dynamized side, but the difference was not statistically significant.
During early periods, fields with hypercellular activities, as determined by the surface staining uptake, were significantly greater in the endosteal areas on the dynamized side (p < 0.05). At periods later than 4 weeks, such differences were less pronounced. The proportion of bone tissue in the endosteal area and in the periosteal callus increased progressively with time for the dynamized and control sides (p < 0.05). At 9 weeks, bone occupied 83% (dynamized) and 88% (control) of the periosteal callus area, whereas at 12 weeks, the amount of bone in the same area was 92% and 97%, respectively.
The density of the endosteal bone, described as the percentage of the endosteal area covered by bone, was significantly greater (p < 0.05) on the dynamized side than on the control side at 4 weeks. However, at 12 weeks the endosteal new bone formation had a strong tendency to be higher (p = 0.09) on the control side (Fig 7A) than on the dynamized side (Fig 7B). Endosteal new bone and endosteal bone density decreased significantly (p < 0.05) between 9 weeks and 12 weeks only on the dynamized side (Table 1).
By using a time-sequenced experimental model, it was possible not only to study bone remodeling characteristics during the late stage of fracture healing, but also to investigate callus formation and maturation at regular intervals during the entire course of the healing process. In previous studies on the effect of axial dynamization on the healing of long bone fractures, only one time interval was studied. 2,3,12,13 In the current study, a bilateral paired experimental model was chosen to compensate for the potential interanimal variability in bone healing studies that use mixed-breed dogs.
The potential effect of dynamization on the fracture healing process may be attributed to the repeated cyclic loading of the bone ends because of free axial movement. However, a secondary effect of dynamization involving the closure or reduction of distance at the osteotomy gap also must be recognized.
It previously was shown that smaller osteotomy gaps return to intact bone strength faster. 2 The goal with the gap on the control side was to keep the size of it as small as possible, while keeping it large enough to ensure avoidance of direct contact between the bone ends even during full weightbearing. Based on the results of the in vitro rigidity test with the fixator, the animal’s weight, and the load during weightbearing, a gap distance of 2 mm was considered optimal. Goodship and Kenwright 14 used a 3-mm gap to study the effects of controlled dynamization on healing of osteotomies in sheep. In the current animal model, a 3-mm gap was considered too large because of the increased risk for delayed union associated with a greater gap size.
A less rigid fixation that allows more movement between bone ends is supposed to trigger more bulky periosteal callus. 6,9,11,29 Consequently, a larger periosteal callus was expected on the dynamized side. Instead, the control side appeared to develop an equal amount of external callus. In fact, the maximum amount of periosteal callus occurred later in the control group. However, at 8 weeks, when the peak callus size was seen in the dynamized group, the amount of periosteal callus in the control group was on the same level. Although dynamization induces free axial movement, once bony contact has been achieved, the relative motion at the fracture site is restricted, especially under constant axial compression. Thus, the relative motion between the osteotomy ends during load bearing in the control side might have been as large as, or larger, than that of the dynamized side.
A more uniform callus distribution has been proposed as one positive effect of dynamization because of more even load distribution and more uniform closure of the osteotomy gap on the dynamized side. In the current study, the circumferential nonuniform distribution of callus followed an asymmetric pattern similar to that previously described using a smaller gap distance. 2 Georgiadis et al 13 reported that most external callus was in the posteromedial and posterolateral parts when using an unstable tibial fracture model in the canine stabilized by intramedullary rods. Less callus being formed on the medial side of the bones in the current study might partially be attributable to the medial incision used in exposing the osteotomy site. Surgical incisions affect vascular supply and thus reduce callus formation. The difference in the total amount of callus was not correlated to the animal’s weight, a finding similar to that previously reported by Aro et al. 2 Large variations in the metabolic response of the individual dogs used, and differences in physical activities, all can contribute to the observed differences in the total amount of callus.
The greater load bearing on the dynamized side during the early weeks probably was attributable to the direct bone contact and subsequently reduced discomfort during load bearing. This result may represent a clinical advantage of axial dynamization. As healing proceeded, more axial loading gradually could be taken on the control side, thereby eliminating the difference between the sides. Higher load bearing on the dynamized side might have been responsible for the faster maturation of the periosteal callus and for the greater endosteal new bone formation.
The lower structural stiffness of the control side at 6 weeks might have been caused by a slower healing process, 21 whereas the dynamized side had normal callus maturation because of load increase, bone contact, and improved osteotomy stability. Later, during the course of healing, the structural strength of the tibia on the control side seemed to approach that of the dynamized side, as revealed by the paired comparison results on all biomechanical parameters. The lack of a significant difference in biomechanical and histologic characteristics between dynamized and control osteotomies when the bone approaches the end of the healing and remodeling period has been reported previously. 2,27 In a study comparing the healing of osteotomies fixed with a 3-mm gap with and without axial micromotion, Goodship and Kenwright 14 reported similar observations at 12 weeks. At this time, a significant difference was found in bone structural strength between the control side and the side with micromotion.
In the current study, new bone formation in the dynamized side of the animals was significantly greater when compared with those fixed with rigid internal plates using the same model at similar times in previous studies. 20,22,28 The higher density of periosteal and endosteal callus on the dynamized side at 4 weeks probably was attributable to the proximity of the fracture fragments because of dynamization, resulting in a shorter distance and smaller area to be spanned by the callus tissue. At 12 weeks, the periosteal callus and the endosteal new bone on the dynamized side seemed to be in a more advanced stage of remodeling with more bone resorption. Such results reflect the indirect effect of external fixation with axial dynamization over other forms of fixation with different fracture gap conditions. Denser callus on the dynamized side when compared with a statically locked intramedullary nailed dog tibia with an osteotomy also was reported by Georgiadis et al. 13 The same effect can be accomplished using internal or external devices.
Axial dynamization appeared to accelerate callus maturation, as reflected by higher periosteal new bone formation and greater torsional stiffness on the dynamized side compared with the control side during the early times. It also showed a faster remodeling process on the dynamized side through a more rapid reduction of the periosteal and endosteal callus. Whether these effects were caused by the bone ends being brought closer together or by the cyclic loading on the fractured bone could not be ascertained in the current study.
The gap closure by axial dynamization might alter the bone healing pathway. Mesenchymal cells in the medullary canal were forced out through the gap because of the gap closing. 21 Mesenchymal tissue replaced the hematoma in the gap and was delivered to the periosteal region. Strong fiber orientation observed from the endosteal region to the periosteal region through the gap coincides with the trabecular orientation in the gap callus perpendicular to the bone axis at the early stage of bone formation observed by Aro et al. 2 Closely packed collagen bundles have been hypothesized not only to increase callus strength, but also through a more favorable molecular spatial arrangement to enhance mineralization. 4 Time-sequential bone histomorphometric analyses based on thin undecalcified sections are needed to elucidate the possible alteration of bone healing by axial dynamization.
The current study revealed that early dynamization resulted in accelerated callus formation and maturation and induced faster remodeling of endosteal and periosteal callus tissue. The osteotomy failure strength increased more rapidly on the dynamized side during the early stages of healing, whereas biomechanical fracture healing stages were more advanced on the dynamized side toward the end of the 12-week study. Most of these findings associated with axial dynamization may be the secondary effect of early load bearing induced by a more stable fracture gap condition. However, under different fracture types and gap conditions, such biomechanical effects and their temporal relationship may be different.
1. Alho A, Ekeland A, Strømsøe K, et al: Locked intramedullary nailing for displaced tibial shaft fractures. J Bone Joint Surg 72B: 805–809, 1990.
2. Aro HT, Kelly PJ, Lewallen DG, et al: The effects of physiologic dynamic compression on bone healing under external fixation. Clin Orthop 256: 260–273, 1990.
3. Aro HT, Wahner HT, Chao EYS: Healing patterns of transverse and oblique osteotomies in the canine tibia under external fixation. J Orthop Trauma 5: 351–364, 1991.
4. Brighton CT: The biology of fracture repair. Instr Course Lect 33: 60–82, 1984.
5. Budsberg SC, Verstraete MC, Soutas-Little RW: Force plate analysis of the walking gait in healthy dogs. Am J Vet Res 48: 915–918, 1987.
6. Burny FL: Elastic External Fixation in Tibial Fractures: Study of 1421 Cases. In Brooker AF, Edwards CC (eds). External Fixation. The Current State of the Art. Baltimore, Williams & Wilkins 55–73, 1979.
7. Burstein AH: A standard test for laboratory animal bone. J Biomech 4: 155–158, 1971.
8. Chao EYS, Aro H: Biomechanics of Fracture Fixation. In Mow VC, Hayes WC (eds). Basic Orthopaedic Biomechanics. New York, Raven Press Ltd 293–335, 1991.
9. Chao EYS, Aro HT, Lewallen DG, et al: The effect of rigidity on fracture healing in external fixation. Clin Orthop 241: 24–35, 1989.
10. Court-Brown CM, Christie J, McQueen MM: Closed intramedullary tibial nailing. J Bone Joint Surg 72B: 605–611, 1990.
11. De Bastiani G, Aldegheri R, Brivio LR: The treatment of fractures with a dynamic axial fixator. J Bone Joint Surg 66B: 538–545, 1984.
12. Egger EL, Gottsauner-Wolf F, Palmer J, et al: Effects of axial dynamization on bone healing. J Trauma 34: 185–192, 1993.
13. Georgiadis GM, Minster GJ, Moed BR: Effects of dynamization after interlocking tibial nailing: An experimental study in dogs. J Orthop Trauma 4: 323–330, 1990.
14. Goodship AE, Kenwright J: The influence of induced micromovement upon the healing of experimental tibial fractures. J Bone Joint Surg 67B: 650–655, 1985.
15. Hart MB, Wu JJ, Chao EYS, et al: External skeletal fixation of canine tibial osteotomies. J Bone Joint Surg 67A: 598–605, 1985.
16. Institute of Laboratory Animal Resources, National Research Council: Guide of the Care and Use of Laboratory Animals. Washington DC, National Academy Press 1–140, 1996.
17. Kempf I, Grosse A, Beck G: Closed locked intramedullary nailing: Its application to comminuted fractures of the femur. J Bone Joint Surg 67A: 709–720, 1985.
18. Klemm KW, Borner M: Interlocking nailing of complex fractures of the femur and tibia. Clin Orthop 212: 89–100, 1986.
19. Lazo-Zbikowski J, Aguilar F, Mozo F, et al: Biocompression external fixation: Sliding external osteosynthesis. Clin Orthop 206: 169–184, 1986.
20. Lewallen DG, Chao EY, Kasman RA, et al: Comparison of the effects of compression plates and external fixation on early bone-healing. J Bone Joint Surg 66A: 1084–1091, 1984.
21. Markel MD, Wikenheiser MA, Chao EYS: Formation of bone in tibial defects in a canine model: Histomorphometric and biomechanical studies. J Bone Joint Surg 73A: 914–923, 1991.
22. Rand JA, An KN, Chao EYS, et al: A comparison of the effect of open intramedullary nailing and compression-plate fixation on fracture site blood flow and fracture union. J Bone Joint Surg 63A: 427–442, 1981.
23. Rubin CT, Lanyon LE: Osteoregulatory nature of mechanical stimuli: Function as a determinant for adaptive remodeling in bone. J Orthop Res 5: 300–310, 1987.
24. Sarmiento A, Schaeffer JHF, Beckerman L, et al: Fracture healing in rat femora as affected by functional weight bearing. J Bone Joint Surg 59A: 369–375, 1977.
25. White AA, Panjabi MM, Southwick WO: The four biomechanical stages of fracture repair. J Bone Joint Surg 59A: 188–192, 1977.
26. Williams EA, Rand JA, Chao EYS, et al: The early healing of tibial osteotomies stabilized by one-plane or two-plane external fixation. J Bone Joint Surg 68A: 355–364, 1987.
27. Wolf JW, White AA, Panjabi MM, et al: Comparison of cyclic loading versus constant compression in the treatment of long-bone fractures in rabbits. J Bone Joint Surg 63A: 805–810, 1981.
28. Woo SL-Y, Lothringer KS, Akeson WH, et al: Less rigid internal fixation plates: Historical perspectives and new concepts. J Orthop Res 1: 431–449, 1983.
© 2001 Lippincott Williams & Wilkins, Inc.
29. Wu J-J, Shyr HS, Chao EYS, et al: Comparison of osteotomy healing under external fixation devices with different stiffness characteristics. J Bone Joint Surg 66A: 1258–1264, 1984.