Transfixation pin casts are generally used for stabilization of comminuted fractures at locations where it is difficult to apply internal fixation for fracture repair, although it is not uncommon for them to be used in conjunction with internal fixation. Typically, a transfixation pin cast consists of pins that pass through the bone proximal to a fracture and then are incorporated within a cast so that most of the animal's weight is supported by the cast, thereby partially unloading the fracture. In fact, the main purpose of a transfixation pin cast is to allow placement of a load on the bone distal to the pins appropriate for fracture healing but not too excessive so as to cause fracture overloading and bone deformation (strain).1
Transfixation pin casts have been used to treat fractures of large animals for > 50 years,2 albeit not without complications including pin loosening, secondary infection, and necrosis at the bone-pin interface.3,4 The bone-pin interface is typically a site of excessive motion that ultimately results in pin loosening and signs of pain.5–7 Consequently, it is often necessary to remove or replace pins, which in turn necessitates replacement of the cast. Many alternatives to decrease complications at the bone-pin interface have been investigated, such as use of pins of varying sizes, smooth versus positive-profile pins, self-tapping versus nontapping positive-profile pins, hydroxyapatite-coated positive-profile pins, and diaphyseal versus metaphyseal pin placement.8–11 However, the optimal fixation strategy for decreasing bone-pin interface complications, minimizing motion, and unloading the fracture site remains undefined.
Traditional transfixation pin casts involving an MC3 or third metatarsal bone consist of two 6.3-mm positive-profile pins placed proximal to the fracture. The pins are placed at 30° relative to each another in the transverse plane. That configuration is preferred over parallel pin placement because it provides better resistance to torsional loading as evidenced by the fact that, in vitro, MC3s stabilized with parallel pins fractured at lower torques than MC3s stabilized with divergent pins.12 Recently, it was suggested that a transfixation pin–cast construct consisting of four 4.8-mm smooth pins spaced 2 cm apart with some degree of divergence would be a more effective construct than two 6.3-mm positive-profile pins spaced 4 to 5 cm apart.a That alternative design arose from the supposition that 6.3-mm pin designs are too rigid and inhibit normal biomechanical stimuli needed to achieve appropriate healing.13 The use of pins with a smaller diameter allows for distribution of some load across the fracture site to facilitate fracture healing while still protecting the fracture from excessive strain.14 In 1 study,15 the amount of strain at the dorsal aspect of the proximal phalanx did not differ significantly between a transfixation pin–cast construct with two 6.3-mm centrally threaded positive-profile pins and a transfixation pin–cast construct with four 4.8-mm smooth Steinman pins, which suggested that loading at the fracture site was similar between those 2 constructs. However, that supposition has yet to be confirmed. The objective of the study reported here was to compare the strain at the bone-pin and cast-pin interfaces among 3 transfixation pin-cast constructs applied to the forelimbs of horses.
Materials and Methods
Specimens
Fifteen forelimbs from 15 adult horses that were euthanized for reasons unrelated to the musculoskeletal system were used for the study. The distal portion of each limb was harvested by disarticulation of the carpometacarpal joint. Then, a reciprocating saw was used to remove the proximal 25% of the MC3 and create a flat surface for load application during biomechanical testing. All soft tissue structures were left intact in the harvested limbs. The limbs were wrapped in towels soaked with physiologic saline (0.9% NaCl) solution, sealed in plastic, and stored at −20°C until tested. The limbs were thawed at room temperature (approx 21°C) for 24 hours prior to construct application and mechanical testing.
Study design
Each limb was randomly assigned by means of a random number generator to 1 of 3 transfixation pin-cast constructs so that each construct had 5 replicates. Construct 1 consisted of two 6.3-mm centrally threaded positive-profile pins spaced 4 cm apart. Construct 2 consisted of two 6.3-mm centrally threaded positive-profile pins spaced 5 cm apart. Construct 3 consisted for four 4.8-mm centrally threaded positive-profile pins spaced 2 cm apart. Each limb and construct was instrumented with strain gauges and then axially loaded to failure. Then, FE modeling was used to further evaluate the mechanical properties (eg, bone strain and load transfer) of the constructs.
Construct application
For each limb assigned to construct 1, a stab incision was made on the lateral aspect of the metaphysis of MC3 just proximal to the lateral epicondyle. A second stab incision was made on the lateral aspect of the diaphysis of MC3 4 cm proximal to (Figure 1) and at 30° divergent from (Figure 2) the first stab incision in the transverse plane. The conventional recommendation for a 2-pin transfixation pin cast is to have a 2-cm proximodistal separation between the pins. We chose to separate the pins by 4 cm in the proximodistal direction because results of another study9 indicate that the diaphysis provides greater pin stability than the metaphysis as measured by resistance to axial extraction. Holes were drilled through the bone underlying the stab incisions by use of a sequence of 4 drill bits of increasing size (3.2, 4.5, 5.5, and 6.2 mm). Each hole was then tapped with a 6.3-mm tap. A 6.3-mm centrally threaded positive-profile pin was placed under power into each hole. Construct 2 was applied in the same manner as construct 1 except the proximal pin was placed 5 cm proximal to the distal pin.
Schematic depictions of 3 transfixation pin–cast constructs as viewed in the frontal plane that were applied to equine forelimbs and then underwent compression loading until failure to assess strain at the bone-pin and cast-pin interfaces. Centrally threaded positive-profile pins were used for all constructs, and the most distal pin was placed just proximal to the epicondyles of MC3. A—Construct 1 consisted of two 6.3-mm-diameter pins spaced 4 cm apart at 30° to each other. B—Construct 2 was the same as construct 1 except the pins were placed 5 cm apart. C—Construct 3 consisted of four 4.8-mm-diameter pins spaced 2 cm apart and at 10° to one another. An osteotomy was created in the proximal phalanx (P1) and extended at a 30° angle from the proximolateral to distomedial direction in the frontal plane. Strain gauges (yellow rectangles) were attached to the cast and on the bone just proximal to the pins and adjacent to the osteotomy. ø = Diameter. P2 = Middle phalanx. P3 = Distal phalanx.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
For each limb assigned to construct 3, a stab incision was made on the lateral aspect of the distal metaphysis of MC3 just proximal to the lateral epicondyle. Three additional stab incisions were placed proximal to the first at 2-cm intervals (Figure 1) and with 10° of divergence between incisions such that adjacent pins were not in the same plane (Figure 2). The first (most distal) and fourth (most proximal) pins were oriented at 30° to one another, mimicking the divergence between the 2 pins of constructs 1 and 2. Holes were drilled through the bone underlying each of the stab incisions by use of a sequence of 2 drill bits of increasing size (3.5 and 4.5 mm). Each hole was tapped with a 4.8-mm tap. A 4.8-mm centrally threaded positive-profile pin was placed under power into each hole.
Schematic depictions of constructs 1 and 2 (A) and construct 3 (B) of Figure 1 as viewed from the transverse plane. Gray rectangles represent pins. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
For all limbs following pin placement, a reciprocating saw was used to create an osteotomy in the proximal phalanx to simulate a fracture that would be commonly stabilized by a transfixation pin cast. The osteotomy extended at a 30° angle from the proximolateral to distomedial direction in the frontal plane (Figure 1). After the osteotomy was created, a bolt cutter was used to cut all pins so that only 4 cm protruded from the specimen in preparation for cast application.
Strain gauge instrumentation
Each limb was dissected to expose the bone proximal to each pin medially and laterally as well as the dorsal surface of the osteotomy. The subcutaneous tissues, common digital extensor tendon, and periosteum were dissected and freed from the bone by use of a No. 10 scalpel blade attached to a No. 3 scalpel handle. The exposed bone was lightly debrided with sand paper, cleaned with 70% ethyl alcohol, and allowed to dry in room air. Uniaxial strain gaugesb were secured to the bone with cyanoacrylate in accordance with the manufacturer's recommendations and best practices for application to bone in vitro.16 A strain gauge was applied to the bone so that its distal edge was located 1 to 2 mm proximal to each pin on both the medial and lateral sides of the specimen17 as well as proximal and distal to the osteotomy (Figure 1). For all constructs, the strain gauges were applied in a proximal-to-distal orientation except for those adjacent to the osteotomy, which were applied in a parallel orientation. The strain gauges and surrounding area (radius, approx 1 cm) were sealed with polyurethane coating.c The gauge leads were further secured to the limb with electrical tape to the most proximal aspect of the specimen. To protect the strain gauges during testing, the adjacent skin was undermined, pulled to cover the gauges and exposed bone, and apposed with nylon suture in a simple continuous pattern.
Cast application
A cast was applied to each limb in a standard fashion. Briefly, the cast incorporated the hoof and extended proximally to 2 cm distal to the carpometacarpal joint (or proximal end of the specimen). Two layers of stockinette were applied followed by 5 rolls of 4-inch fiberglass cast material.d The rationale for the selection of cast materials was made on the basis of results of a previous study,1 in which 2 rolls of 4-inch and 2 rolls of 5-inch fiberglass casting tape were used for in vitro transfixation pin-cast application. The pins were incorporated into the cast by making incisions through the fiberglass cast material as it was wrapped around the limb. A uniaxial strain gauge was attached to the cast so that its distal edge was located 1 to 2 mm proximal to each pin on both the medial and lateral sides of the specimen. The strain gauges were attached with cyanoacrylate in accordance with the manufacturer's recommendations for porous materials. The gauge leads were secured to the cast with electrical tape at its proximal extent. All casts were allowed to cure and dry for approximately 1.5 hours prior to biomechanical testing.
Biomechanical testing
Each construct replicate was axially loaded to failure by use of a hydraulic actuatore instrumented with a linear variable differential transformerf and load cell.g Each replicate was placed under the actuator and secured in place by use of a custom frame (Figure 3). The toe of the cast made contact with the frame, which prevented forward sliding of the specimen once axial compressive loading was applied. The load cell was fitted with an adaptor and plate, which applied load to the sectioned surface of MC3. The adaptor extended 1 cm into the marrow cavity of MC3 and was used to ensure that the loading plate was centered across the bone.
Photograph that depicts the positioning of the distal portion of a forelimb of an adult horse to which 1 of the 3 transfixation pin-cast constructs described in Figure 1 has been applied just prior to compression testing. Each construct replicate was axially loaded to failure by use of a hydraulic actuator instrumented with a linear variable differential transformer and load cell. Each replicate was placed under the actuator and secured in place by use of a custom frame. The toe of the cast made contact with the frame, which prevented forward sliding once axial compressive loading was applied. The load cell was fitted with an adaptor and plate, which applied load to the sectioned surface of MC3. The adaptor extended 1 cm into the marrow cavity of MC3 and was used to ensure that the loading plate was centered across the bone.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
Each replicate was loaded to failure in a single cycle with axial compression at a constant rate of 2 mm/s.18–20 Load, displacement, and strain data were collected at 250 Hz. Axial load was applied until catastrophic failure was observed. Failure was defined as fracture of the proximal portion of MC3, cast breakage, or bowing of the load cylinder. After biomechanical testing was completed, the specimen was returned to a freezer, where it was stored until it was dissected.
Biomechanical data analysis
Strain data at loads of 2.5 (consistent with standing21) and 7.5 (consistent with walking21) kN were used to evaluate load transfer among the 3 constructs. For each construct, descriptive data were generated for strain measures. The maximum compressive strain between medial and lateral gauges for each pin was selected for analysis. Only strain measures for the most proximal and most distal pins of each construct were compared among the 3 constructs. The null hypothesis was that the amount of stain at the bone-pin interface and cast-pin interface would not differ among the 3 constructs. The data distribution for each variable was assessed for normality by means of the Shapiro-Wilk test. For normally distributed variables, ANOVA was used for global comparisons among the 3 constructs, and when necessary, pairwise comparisons were performed with Student t tests. For variables that were not normally distributed, a Kruskal-Wall is test was used for global comparisons among the 3 constructs, and when necessary, pairwise comparisons were performed with Mann-Whitney U tests. The Bonferroni correction was used to account for multiple comparisons. All analyses were performed with commercial statistical software,h and values of P < 0.05 were considered significant.
FE modeling
Finite element modelingi was used to evaluate the mechanical behavior (eg, load transfer and bone strain) of constructs 1 and 3. Construct 2 was not evaluated with FE modeling because preliminary analyses indicated that its mechanical behavior was nearly identical to that of construct 1.
Geometry—A CT image of an equine forelimb was used to derive a simplified circular geometric representation of MC3 (Figure 4). Because the elastic modulus of trabecular bone is approximately 1/40 that of cortical bone, only cortical bone was modeled. The cortical bone was thickest (10 mm) at the midshaft region and thinnest (3.5 mm) at the distal aspect of MC3 near the metacarpophalangeal (fetlock) joint. The cast was modeled with a thickness of 12.5 mm and a gap of 5 mm between the bone and cast. Pins were transversely placed at 30° to one another for construct 1 and at 10° to one another for construct 3. The distance between the bottom pin and fetlock joint was 3.5 cm for both models. Because the tissues below the fetlock joint were the same between the 2 constructs and would have the same effect on load transfer, they were modeled as a simple cylindrical structure. The simple cylindrical structure was assumed to be completely bonded to the bone and cast. The fracture site was assumed to be located 3 cm distal to the fetlock joint, which was considered sufficiently far enough from the joint to achieve a uniform strain distribution. The 30° osteotomy was not modeled. Instead, the strain and load were monitored at the site where the osteotomy would have been created to assess how effectively each construct unloaded bone.
Series of CT and radiographic images and schematic depictions provided to facilitate understanding of the FE models created for transfixation pin-cast constructs 1 and 3 of Figure 1 to further evaluate load transfer and bone strain within each construct. A—Lateral CT image of an adult equine forelimb used to derive a simplified circular geometric representation of MC3. B—Dorsopalmar radiographic image of a limb to which construct 1 was applied prior to biomechanical testing, which was used in conjunction with the CT image of panel A to derive a simplified geometric representation of MC3. C—Schematic depiction of the simplified geometry of MC3 used for FE modeling. The cortical bone (gray-shaded area) of MC3 was thickest (10 mm) at the midshaft region and thinnest (3.5 mm) at the distal aspect of the bone near the metacarpophalangeal (fetlock) joint. D—Schematic depiction of the simplified model of MC3 and construct 3 superimposed over a dorsopalmar radiographic image of the distal portion of an equine forelimb. For both FE models, the cast (blue-shaded area) was modeled with a thickness of 12.5 mm and a gap of 5 mm between the bone and cast. Notice that an osteotomy in the proximal phalanx was not included in the FE model. Instead, the strain and load were monitored at the site where the osteotomy would have been created to assess how effectively each construct unloaded bone. E—Schematic depiction of the FE model for construct 3 complete with descriptions of material properties, which were the same for both models. Quadratic, 10-node tetrahedral elements were used to mesh the models. Isotropic, homogeneous, and linearly elastic materials were used to model cortical bone (elastic modulus [E], 20 GPa; Poisson ratio, 0.3), pins (E, 190 GPa; Poisson ratio, 0.305), and cast (E, 3.4 GPa; Poisson ratio, 0.3). A single effective E of 500 MPa (Poisson ratio, 0.3) was used to model all bony and soft tissue structures distal to the fetlock joint because those structures were the same for both constructs. That E was derived iteratively until FE-derived compressive strain results at 3 sites (proximal and distal cast sites and a proximal bone site) mimicked experimental strain measures observed during biomechanical testing of the 2- and 4-pin configurations. F—3-D schematic depiction of the FE model created for construct 3. A uniform pressure equivalent to the force during walking (7.5 kN) was applied to a rigid plate, which was placed on top of the most proximal aspect of the model to simulate loading of the construct during experimental (biomechanical compression) testing.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
Material properties—Quadratic, 10-node tetrahedral elements were used to mesh the models. Isotropic, homogeneous, and linearly elastic materials were used to model cortical bone (elastic modulus, 20 GPa; Poisson ratio, 0.3),22 pins (elastic modulus, 190 GPa; Poisson ratio, 0.305),23 and cast (elastic modulus, 3.4 GPa; Poisson ratio, 0.3).24 The modeled cast material properties were consistent with the mean longitudinal and flexural modulus properties reported by Wytch et al,24 with the assumption that flexural modulus was equal to elastic modulus. A single effective elastic modulus of 500 MPa (Poisson ratio, 0.3) was used to model all bony and soft tissue structures distal to the fetlock joint (including the fracture site). That modulus was derived iteratively until FE-derived compressive strain results at 3 sites (proximal and distal cast sites and a proximal bone site) mimicked experimental strain measures observed during biomechanical testing of the 2- and 4-pin configurations. Finite element modeling strain predictions were within ± 13% of the mean experimental values observed during biomechanical testing, which was deemed sufficiently accurate for application in this study.
Loads and boundary conditions—The most distal section of the cast was fully constrained in all directions, and the pin-bone and pin-cast contact surfaces were completely bonded to avoid any relative movement between them. We deemed that reasonable because sliding between the threaded holes and threaded pins was not expected during biomechanical testing. A uniform pressure equivalent to the force during walking (7.5 kN) was applied to a rigid plate, which was placed on top of the most proximal section of MC3 to simulate experimental test loading (Figure 4).
FE outcomes—Finite element–derived mechanical outcomes included mean bone compressive strain at the distal bone-pin interface and fracture site and the relative amount of load transferred to the cast and fracture site for both the 2-pin (construct 1) and 4-pin (construct 3) transfixation pin-cast constructs. For bone-pin interface strain measures, the mean strain values for the proximal elements around the pins were reported. For load measures, load was calculated by summing the product of stress and area in the compressive direction for each element at the fracture site. Stress values were assumed to be constant over each element volume.
Results
Biomechanical testing
The strain gauge data varied among the replicates within each construct but was particularly variable for replicates of constructs 1 and 2 (Figure 5). Specifically, strain measures obtained adjacent to the osteotomy were highly erratic and somewhat unstable. For example, load was transferred in line with the strain gauges on the dorsal side of the osteotomy for some replicates and on the palmar side of the osteotomy for others. Therefore, osteotomy strain data were not included in the statistical analysis.
Representative load-displacement curve generated during compression testing to failure for a replicate of transfixation pin–cast constructs 1 (A) and 3 (B) described in Figure 1. A load-displacement curve for a replicate of construct 2 was not provided because the load-displacement curves generated for the replicates of construct 2 did not differ significantly from those for replicates of construct 1. The load-displacement curve provided for construct 3 is for 1 of the 2 replicates (out of 5) that had plastic deformation of the pins, which led to load being transferred from the construct to the bone. Nevertheless, notice that deformation or breakdown of the pins and cast, which led to cartilage compression and transfer of the load from the pins and cast to the bone, occurred at a much lower load for construct 1 than construct 3. This suggested that construct 3 was more effective in unloading the fracture site than either construct 1 or 2. However, for both replicates depicted here, the transfer of load from the pins and cast to the bone occurred at loads just > 7.5 kN, which is the estimated load applied to a forelimb of an adult horse while walking.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
All variables evaluated except for strain at the most proximal cast pin were normally distributed. Although strain was variable, it varied significantly among the constructs; therefore, the null hypothesis was rejected. Strain did not differ between constructs 1 and 2 at any location but did differ significantly between construct 3 and constructs 1 and 2 at some locations (Figure 6). When assessed at a load consistent with standing (2.5 kN), the mean strain at the proximal pin-bone interface for construct 3 was 69% and 75% that for constructs 2 and 1, respectively, whereas the mean strain at the most proximal cast site for construct 3 was 269% and 430% that for constructs 2 and 1, respectively. When assessed at a load consistent with walking (7.5 kN), the mean strain at the distal cast site for construct 3 was 386% and 218% that for constructs 2 and 1, respectively. When the relative strain was assessed among the pins of each construct, it was apparent that the most proximal pin carried the most load for constructs 1 and 2, whereas the load was fairly evenly distributed between the 2 most proximal pins of construct 3.
Mean maximum compressive strain (microstrain) at various locations of transfixation pin–cast constructs 1 (light gray bars; n = 5), 2 (dark gray bars; 5), and 3 (black bars; 5) described in Figure 1 at loads of 2.5 (A; ie, standing) and 7.5 (B; ie, walking) kN. For pin and cast measures, values represent the maximum (negative) value between medial and lateral strain gauges at that location. For osteotomy measures, values represent the maximum (negative) value between the proximal and distal strain gauges. Error bars represent the 95% confidence interval. *Values linked by brackets differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
FE modeling
Results of FE modeling for constructs 1 and 3 were summarized (Table 1). For construct 3, the load at the fracture site was 6.4% less, the bone strain at the distal bone-pin interface was 18% less, and the compressive strain adjacent to the fracture site was 22% less than the corresponding values for construct 1. Local bone strain at the bone-pin interface was quite high, reaching approximately −15,000 microstrain for both constructs 1 and 3, although the region that underwent high local bone strain was smaller for construct 3 than construct 1 (Figure 7).
Images generated during FE modeling that depict the minimum principal strain at the bone-pin interface for transfixation pin–cast constructs 1 (A) and 3 (B) described in Figure 1. Insets represent a higher magnification of the bone-pin interface outlined by the white lines in the main image. Pin and construct deformation has been exaggerated for visualization purposes. Notice that local bone strain at the bone-pin interface was quite high, reaching approximately −15,000 microstrain for both constructs, although the region that underwent high local bone strain was smaller for construct 3 than construct 1.
Citation: American Journal of Veterinary Research 79, 12; 10.2460/ajvr.79.12.1287
Summary of FE modeling results for transfixation pin–cast constructs 1 (2-pin construct) and 3 (4-pin construct) when applied to a forelimb of an adult horse.
| Variable | Construct 1 | Construct 3 |
|---|---|---|
| Minimum principal strain adjacent to bone-pin interface (microstrain) | −3,750 | −3,090 |
| Minimum principal strain at fracture site (microstrain) | −1,875 | −1,460 |
| Load going into the fracture site (N) | 788 | 737 |
| Load going into the cast (N) | 6,712 | 6,763 |
Construct 1 consisted of two 6.3-mm-diameter pins spaced 4 cm apart at 30° to each other. Construct 3 consisted of four 4.8-mm-diameter pins spaced 2 cm apart and at 10° to one another. The distance between the most distal pin and metacarpophalangeal joint was 3.5 cm for both constructs. Minimum principal strain refers to compressive strain.
Discussion
In the present study, experimental biomechanical testing and FE modeling were used to compare the mechanical properties among 3 transfixation pin–cast constructs. Results indicated that the amount of load transferred to the cast was greatest for the 4-pin construct (construct 3) and similar between the 2-pin constructs (constructs 1 and 2). This suggested that construct 3 might provide more protection against fracture overload than the other 2 constructs. Additionally, the amount of strain at the bone-pin interface was generally greater for constructs 1 and 2 than for construct 3. Strain at the bone-pin interface can contribute to pin loosening, and that finding may explain why pin loosening is a common complication associated with 2-pin transfixation pin casts in clinical practice.
The amount of strain at various locations varied significantly between construct 3 and constructs 1 and 2 evaluated in the present study. That finding was in contrast to results of another study,15 in which the amount of strain did not differ significantly between a transfixation pin–cast construct consisting of two 6.3-mm centrally threaded positive-profile pins and a construct consisting of four 4.8-mm smooth Steinman pins. In that study,15 strain was measured at the dorsal aspect of the proximal phalanx in the absence of an osteotomy at a compressive load of 5 kN. The apparently conflicting results between the present study and that study15 might be attributed, at least partially, to mechanical differences between centrally threaded positive-profile pins and smooth Steinman pins. Results of other studies25–27 indicate that centrally threaded pins are stiffer than smooth pins. However, the magnitude of the differences between the 2-pin and 4-pin constructs of the present study were small (ie, FE modeling indicated that the difference in load transfer between constructs 1 and 3 was approx 50 N) and consistent with the findings of that other study.15 It is also important to note that, although the strain differences at the distal cast sites were fairly large between construct 3 and the 2-pin constructs, the strain measures for construct 3 reflected strain associated with loading as well as strain imposed by the proximal pins under bending. Consequently, the strain measures for construct 3 likely overestimated the extent of load that was transferred to the cast. Nevertheless, the biomechanical testing and FE modeling results of the present study suggested that a 4-pin construct might provide a modest increase in fracture site protection, compared with 2-pin constructs.
The 2-pin constructs evaluated in the present study had greater strain at the bone-pin interface than did the 4-pin construct. That was important because we believe high strain, in combination with factors associated with the 2-pin surgical technique, likely contributes to the complications commonly observed with 2-pin transfixation pin casts in clinical practice, such as pin loosening and necrosis at bone-pin interfaces. Excessive strain can impair blood flow to the affected region and lead to local bone failure and avascular necrosis.28 Local bone failure will lead to pin loosening, which can cause further bone damage owing to repetitive loading (fatigue).29 Tissue irritation at the site of a loosened pin can result in swelling and drainage at the pin track, thereby leading to further loosening of the pin. Although the external ends of the pins are typically covered, they are exposed to dirt, debris, and fecal material. If a contaminated pin comes into contact with compromised tissue (ie, an enlarged and inflamed pin track caused by excessive pin motion), osteitis can result. Regardless of loading, the surgical technique for application of a 2-pin transfixation pin cast involves the use of multiple drill bits of increasing size to create the pin holes. Inappropriate lubrication and cooling of those bits during drilling can cause thermal necrosis at the bone-pin interface.29 Also, the risk of damage to the blood supply (and thereby necrosis) of the drilled bone increases as the size of the drill bit used to create the pin holes increases.30 Collectively, those factors could presumably lead to the formation of necrotic ring sequestra, pin loosening, and signs of pain.
The mean bone strain for construct 1 (2-pin construct with 4 cm of separation between pins) did not differ significantly from that for construct 2 (2-pin construct with 5 cm of separation between pins). Although that finding may not be particularly surprising, it is clinically relevant. Radiographic evidence suggests that, in horses, the cortical bone 5 cm proximal to the lateral epicondyle of MC3 (ie, diaphysis) is thicker than that 4 cm proximal to the lateral epicondyle (ie, metaphysis). Presumably, the most proximal pin of construct 2 (located in the diaphysis) would be more rigidly constrained and thus more effective at unloading bone than the most proximal pin of construct 1 (metaphysis). Our results suggested that a 2-pin transfixation pin-cast construct with 1 pin placed in the metaphysis and 1 pin placed in the diaphysis of MC3 performed similarly to a 2-pin construct with both pins placed in the metaphysis of MC3. However, results of other studies26,31 indicate that the risk of MC3 fracture increases when pins are placed in the diaphysis, likely because although the cortical bone of the diaphysis is thicker than that of the metaphysis, its overall diameter is smaller, and placement of a pin with a fairly large diameter (eg, ≥ 10% of the bone diameter) in the diaphysis can result in areas of concentrated stress.32 Investigators of another study31 recommend that the pins of transfixation pin casts be separated by 2 or 3 cm, with the most distal pin placed near the metaphysis. Nonetheless, the pin stability offered by the thicker cortical bone of the diaphysis may offset the increased risk of stress concentrations and MC3 fracture when compared with the risk for pin loosening or failure owing to insufficient fixation when pins are placed in the thin cortical bone of the metaphysis. Further biomechanical testing including fatigue and torsional testing of 2-pin transfixation pin-cast constructs with varying pin separations (eg, 2, 3, 4, and 5 cm) is necessary to identify the most effective construct.
Historically, 2 centrally threaded positive-profile pins were recommended for application of transfixation pin casts to MC3 or the third metatarsal bone of horses. Two pins were recommended to provide the cast with sufficient strength to immobilize and transfer load from the fracture without creating a catastrophic defect in the cortical bone (as might be caused by the use of 1 pin with a large diameter).26 The biomechanical testing and FE modeling results of the present study indicated that the strain at the bone-pin interface for the 2-pin constructs was significantly greater than that for the 4-pin construct. That strain might contribute to pin loosening, and alternatives to 2-pin constructs should be considered. We believe that a transfixation pin–cast construct that uses 3 positive-profile pins with a diameter between 4.8 and 6.3 mm will have similar fracture unloading properties as the 4-pin construct evaluated in the present study with the benefit of 1 less hole being drilled into MC3; however, that supposition remains to be tested.
In the present study, bone and cast strains were measured at loads of 2.5 and 7.5 kN, which were consistent with the loads applied to the distal portion of the forelimb of adult horses while standing and walking, respectively.21 Originally, we intended to evaluate bone strain at greater loads, but strain readings at loads ≥ 10 kN were lost because the gauges loosened from the constructs or the maximum number of readings allowable was exceeded. Interestingly, the mean load at catastrophic failure was > 25 kN for all 3 constructs evaluated in this study. That finding was reassuring because the MC3s of adult horses commonly sustain loads of almost 21 kN during recovery from anesthesia.21 That finding, in conjunction with our own clinical experience, suggested that all 3 constructs were capable of withstanding loads typically applied to transfixation pin casts during anesthesia recovery provided the number of attempts to rise are limited. Therefore, we recommend the implementation of assisted anesthesia recovery methods for horses following transfixation pin–cast application.
Although all 3 constructs evaluated in the present study were able to support high loads before failure, they were not equally effective in transferring load away from the fracture site. A particularly interesting finding observed during both biomechanical testing and FE modeling pertained to the amount of stress applied to the pins of the constructs. When assessed at a load consistent with a walk (7.5 kN), the pin stress (specifically von Mises stress) for construct 3 was 280 MPa, which is close to the yield strength of stainless steel (260 MPa).23 In fact, evaluation of the load-displacement curves revealed that plastic deformation occurred under a load of approximately 7.5 kN for 2 of the 5 construct 3 replicates, and visual examination of the specimens after biomechanical testing confirmed plastic deformation in those 2 replicates. This indicated that more load was transferred through the fracture site than through the pins and cast. In fact, excessive pin deflection, pin deformation, or cast breakage was observed for replicates of all 3 constructs at loads slightly > 7.5 kN, which indicated that the bulk of the load was transferred through the cartilage, bone, and fracture site. Additional research is necessary to assess transfixation pin-cast constructs consisting of 3 or 4 pins with diameters between 4.8 and 6.3 mm to identify the optimal construct. This is important because, in clinical practice, transfixation pin casts are typically required (and expected) to last for several weeks and minimizing cast-associated complications will optimize patient comfort, fracture healing, and return to function.
The present study was not without limitations. Only 5 replicates of each construct were evaluated; therefore, the results need to be evaluated with caution and further research is necessary to validate or refute our findings. The strain gauges were placed close and parallel to the osteotomy, which proved to be problematic owing to how the load was transferred across the osteotomy gap, resulting in high variation in strain measures. In hindsight, the data generated might have been less variable had a strain gauge been applied closer to the hoof or had a load cell been applied under the hoof to estimate load transfer for each construct. It might also have been prudent to put strain gauges directly on the pins rather than at multiple bone and cast sites to estimate load transfer. We chose to place strain gauges above the pins of the constructs in a manner similar to that used in another study17 to enable comparisons between the 2 studies. However, it might have been more appropriate to place the strain gauges distal to the pins because that would have provided a more accurate assessment of the compressive strain applied to the cast owing to load transfer from the bone and pins. Moreover, the strain gauges used in this study were limited to a maximum compressive strain of ± −5% (−50,000 microstrain), which limited our analysis to loads < 10 kN. The strain gauges were also unidirectional; therefore, measurements could be obtained only in 1 direction (the compressive axis). Although we attempted to align the gauges with the long-axis of the bone, there was some degree of malalignment; thus, the strain measures reported may not accurately reflect the maximum compressive strain to which the constructs were exposed. Future studies in this area should use multidirectional strain gauge rosettes to derive maximum and minimum principal strain measures and account for strain gauge placement error. We did not precondition the specimens prior to testing to failure because we wished to assess the response of each replicate to immediate loading. Preconditioning of the specimens would have removed some of the highly compliant parts of the systems prior to biomechanical testing, and likely made the strain data less erratic. The FE models developed for this study applied simplified geometry and material properties. During FE modeling, trabecular bone was ignored and bone was modeled as an isotropic material, even though results of anisotropy indicate that bone is at least orthotropic.33 Although the FE models developed in this study could be modified to account for such factors, we believe it will be more advantageous to develop an equine-specific FE model that integrates equine geometry, orientations and contact locations, heterogeneous material properties, and trabecular orientation and anisotropy. It is important to note that a fracture was not included in the FE models to avoid confounding of results by fracture mechanics (eg, lateral movement of fractured bone, crack propagation, and cohesive elements). Instead, a single material was used to model structures distal to the fetlock joint. Because that structure was the same for both FE models, and the FE-based strain at the phantom fracture site was used only to compare unloading between the 2 construct configurations, we believed use of the simplified model was justified.
Results of the present in vitro study suggested that a 4-pin transfixation pin-cast construct was more effective in unloading a simulated fracture in the proximal phalanx of an equine forelimb than either of the 2-pin constructs evaluated. Nevertheless, the 4-pin construct may behave similarly to 2-pin constructs at loads > 7.5 kN owing to plastic deformation of the pins. Additionally, results of both biomechanical testing and FE modeling indicated that the 2-pin constructs had high strain at both the bone-pin and cast-pin interfaces, which can contribute to pin loosening, a complication commonly associated with 2-pin transfixation pin cast in clinical practice. Further research is necessary to identify the optimal transfixation pin-cast construct for immobilization of fractures distal to the fetlock joint of the forelimb of adult horses.
Acknowledgments
Supported by the Townsend Equine Health Research Fund at the Western College of Veterinary Medicine, University of Saskatchewan.
The authors declare that there were no conflicts of interest.
The authors thank Brennan Pokoyoway and Dr. David Wilson for technical assistance.
ABBREVIATIONS
| FE | Finite element |
| MC3 | Third metacarpal bone |
Footnotes
Bramlage LR, Rood and Riddle Equine Hospital, Lexington, Ky: Personal communication, 2017.
CEA-06-125UW-120, Vishay Micro-Measurements, Raleigh, NC.
M-Coat A, Vishay Micro-Measurements, Raleigh, NC.
Delta-Lite Plus, BSN Medical, Hamburg, Germany.
Model RRH-10010, Enerpac, Milwaukee, Wis.
LDI-119-200-A020A, Omega, Norwalk, Conn.
Model 1220-AF, 250 kN capacity, Interface, Scottsdale, Ariz.
SPSS, version 18.0, IBM Corp, Chicago, Ill.
ABAQUS, 3DS, Waltham, Mass.
References
1. McClure SR, Watkins JP, Bronson DG, et al. In vitro comparison of the standard short limb cast and three configurations of short limb transfixation casts in equine forelimbs. Am J Vet Res 1994;55:1331–1334.
2. Reichel EC. Treatment of fractures of the long bones in large animals. J Am Vet Med Assoc 1956;129:8–15.
3. McClure SR, Honnas CM, Walkins JP. Managing equine fractures with external skeletal fixation. Compend Contin Educ Pract Vet 1995;17:1054–1063.
4. Mahan J, Seligson D, Henry SL, et al. Factors in pin tract infections. Orthopedics 1991;14:305–308.
5. Morisset S, McClure SR, Hillberry BM, et al. In vitro comparison of the use of two large-animal, centrally threaded, positive-profile transfixation pin designs in the equine third metacarpal bone. Am J Vet Res 2000;61:1298–1303.
6. Nunamaker DM, Richardson DW, Butterweck DM, et al. A new external skeletal fixation device that allows immediate full weightbearing application in the horse. Vet Surg 1986;15:345–355.
7. Németh F, Back W. The use of the walking cast to repair fractures in horses and ponies. Equine Vet J 1991;23:32–36.
8. Zacharias JR, Lescun TB, Moore GE, et al. Comparison of insertion characteristics of two types of hydroxyapatite-coated and uncoated positive profile transfixation pins in the third metacarpal bone of horses. Am J Vet Res 2007;68:1160–1166.
9. McClure SR, Hillberry BM, Fisher KE. In vitro comparison of metaphyseal and diaphyseal placement of centrally threaded, positive-profile transfixation pins in the equine third metacarpal bone. Am J Vet Res 2000;61:1304–1308.
10. Bubeck KA, García-Lopez JM, Jenei TM, et al. In vitro comparison of two centrally threaded, positive-profile transfixation pin designs for use in third metacarpal bones in horses. Am J Vet Res 2010;71:976–981.
11. Aron DN, Toombs JP, Hollingsworth SC. Primary treatment of severe fractures by external skeletal fixation: threaded pins compared with smooth pins. J Am Anim Hosp Assoc 1986;22:659–670.
12. McClure SR, Watkins JP, Ashman RB. In vitro comparison of the effect of parallel and divergent transfixation pins on breaking strength of equine third metacarpal bones. Am J Vet Res 1994;55:1327–1330.
13. Smith GK. Biomechanics pertinent to fracture etiology, reduction, and fixation. In: Newton CD, David M, eds. Textbook of small animal orthopaedics. Philadelphia: Lippincott Williams and Wilkins, 1985;195–230.
14. Perren SM. Physical and biological aspects of fracture healing with special reference to internal fixation. Clin Orthop Relat Res 1979;175–196.
15. Williams JM, Elce YA, Litsky AS. Comparison of 2 equine transfixation pin casts and the effects of pin removal. Vet Surg 2014;43:430–436.
16. Cordey J, Gautier E. Strain gauges used in the mechanical testing of bones. Part II: “in vitro” and “in vivo” technique. Injury 1999;30(suppl 1):A14–A20.
17. Brianza S, Brighenti V, Boure L, et al. In vitro mechanical evaluation of a novel pin-sleeve system for external fixation of distal limb fractures in horses: a proof of concept study. Vet Surg 2010;39:601–608.
18. Nash RA, Nunamaker DM, Boston R. Evaluation of a tapered-sleeve transcortical pin to reduce stress at the bone-pin interface in metacarpal bones obtained from horses. Am J Vet Res 2001;62:955–960.
19. Nutt JN, Southwood LL, Elce YA, et al. In vitro comparison of a novel external fixator and traditional full-limb transfixation pin cast in horses. Vet Surg 2010;39:594–600.
20. Elce YA, Southwood LL, Nutt JN, et al. Ex vivo comparison of a novel tapered-sleeve and traditional full-limb transfixation pin cast for distal radial fracture stabilization in the horse. Vet Comp Orthop Traumatol 2006;19:93–97.
21. Turner AS, Mills EJ, Gabel AA. In vivo measurement of bone strain in the horse. Am J Vet Res 1975;36:1573–1579.
22. Rauber AA. Elastizitt und festigkeit der Knochen: anatomisch-physiologische studie. Leipzig, Germany: Verlag von Wilhelm Engelman, 1876.
23. Budynas RG, Nisbett JK. Shigley's mechanical engineering design. 9th ed. New York: McGraw-Hill, 2011;42–56.
24. Wytch R, Mitchell CB, Wardlaw D, et al. Mechanical assessment of polyurethane impregnated fibreglass bandages for splinting. Prosthet Orthot Int 1987;11:128–134.
25. Clary EM, Roe SC. Enhancing external skeletal fixation pin performance: consideration of the pin-bone interface. Vet Comp Orthop Traumatol 1995;8:1–8.
26. Lescun TB, McClure SR, Ward MP, et al. Evaluation of transfixation casting for treatment of third metacarpal, third metatarsal, and phalangeal fractures in horses: 37 cases (1994–2004). J Am Vet Med Assoc 2007;230:1340–1349.
27. Palmer RH, Hulse DA, Hyman WA, et al. Principles of bone healing and biomechanics of external skeletal fixation. Vet Clin North Am Small Anim Pract 1992;22:45–68.
28. Chamay A, Tschantz P. Mechanical influences in bone remodeling. Experimental research on Wolff's law. J Biomech 1972;5:173–180.
29. Matthews LS, Hirsch C. Temperatures measured in human cortical bone when drilling. J Bone Joint Surg Am 1972;54:297–308.
30. Perren SM. Evolution of the internal fixation of long bone fractures. The scientific basis of biological internal fixation: choosing a new balance between stability and biology. J Bone Joint Surg Br 2002;84:1093–1110.
31. Joyce J, Baxter GM, Sarrafian TL, et al. Use of transfixation pin casts to treat adult horses with comminuted phalangeal fractures: 20 cases (1993–2003). J Am Vet Med Assoc 2006;229:725–730.
32. Edgerton BC, An KN, Morrey BF. Torsional strength reduction due to cortical defects in bone. J Orthop Res 1990;8:851–855.
33. Ashman RB, Rho JY, Turner CH. Anatomical variation of orthotropic elastic moduli of the proximal human tibia. J Biomech 1989;22:895–900.
