Criteria for Selection of Cases

Medical records of all adult (> 2 years of age) horses with comminuted phalangeal fractures treated with transfixation pin casts at Colorado State University from 1993 to 2003 were reviewed. Data collected from medical records included signalment; the specific bone fractured; the size, number, and characteristics of transfixation pins used in the transfixation pin cast; whether supplemental internal fixation of the fracture was used; the duration the transfixation pin cast was maintained; and any complications associated with use of transfixation pins. Follow-up results were obtained from medical records or by contact with owners or trainers. Short-term survival was defined as being discharged from the hospital (survived), whereas longterm survival was defined as a minimum of 1 year from the time treatment was initiated.

Procedures

Before surgery, horses received penicillin G potassium (22,000 U/kg [10,000 U/lb], IV, q 6 h) or ampicillin sodium (11 mg/kg [5 mg/lb], IV, q 12 h) combined with gentamicin sulfate (6.6 mg/kg [3 mg/lb], IV, q 24 h). Horses in which internal fixation was performed received antimicrobials for a mean of 7 days (range, 5 to 17 days), and horses in which transfixation pin casting alone was performed received only perioperative antimicrobials. All horses received phenylbutazone (4.4 mg/kg [2 mg/lb], IV) before surgery, which was continued (2.2 mg/kg [1 mg/lb], PO) after surgery every 12 to 24 hours depending on whether the horse had signs of discomfort. Horses were anesthetized and placed in lateral recumbency with the affected limb uppermost. Fracture characteristics and surgeon preference determined whether internal fixation was performed in addition to placement of a transfixation pin cast. In the study reported here, fractures were considered as moderately comminuted (> 3 large fragments) or severely comminuted (multiple small fragments). No intact strut of bone extended from the proximal to the distal articular margin. Internal fixation was performed in horses with moderately comminuted fractures that contained fragments large enough to support placement of cortical bone screws alone or combined with plate fixation. Both open reduction and lag screw (4.5 or 5.5 mm) placement through stab incisions during fluoroscopic guidance were used for internal fixation. In horses in which the proximal phalanx was fractured, reduction of the proximal articular margin was performed prior to reduction of the distal articular margin in order to achieve a congruent articular surface of the metacarpo/metatarsophalangeal joint. In certain horses with comminuted middle phalangeal fractures that contained fragments large enough to support internal fixation, arthrodesis of the proximal interphalangeal joint was performed concurrently with transfixation pin casting. Internal fixation was always performed prior to placement of transfixation pins. Transfixation pin casts were used concurrently with internal fixation to protect the implants from weightbearing forces in horses in which the fixation lacked adequate strength and rigidity to permit weight bearing in the postoperative period.

Placement of transfixation pins was performed with the horse in lateral recumbency with the affected limb uppermost and the pins directed in a lateral to medial direction. Pin placement within the third metacarpal or metatarsal bone varied according to surgeon preference and ranged from the metaphysis to the proximal portion of the diaphysis. Surgeon preference and pin availability determined whether threaded or smooth transfixation pins were used as well as the number used (ie, some horses were treated prior to availability of threaded transfixation pins). Two 6.3-mm-diameter centrally threaded positive profile pins^a were placed through predrilled^b and tapped holes for horses in which threaded transfixation pins were used. If smooth Steinman pins were used (6.3 mm), predrilled holes were made and the transfixation pins were placed with a hand chuck. Transfixation pins were placed in an approximate 30° divergent angle from the frontal plane in horses in which surgery was performed during the later years of the study. After placement of transfixation pins (two or three 6.3-mm pins in all cases), a half-limb fiberglass cast was applied, incorporating the pins. Casts were constructed in a standard fashion consisting of a stockinette, 1 layer of cast foam, and a minimum of 6 rolls of 4-inch cast material. Polymethyl methacrylate was applied external to the cast over each pin and on the bottom of the cast for protection. Each horse was hand assisted during recovery from anesthesia.

Statistical analysis—Data were evaluated in a descriptive manner, and data proportions were compared when appropriate by use of Mantel-Haenzel χ² analysis.^c Values of P < 0.05 were considered significant.

Results

Medical records for 20 horses were included in the study. The mean age of affected horses was 9.1 years (median, 9 years; range, 2 to 19 years). There were 11 mares, 5 geldings, and 4 sexually intact males. Breeds represented included Quarter Horse (n = 10), Arabian (5), mixed breed (4), and warmblood (1). Twelve horses had fractures in a hind limb, and 8 horses had fractures in a forelimb. The middle and proximal phalanges were fractured in 14 and 6 horses, respectively. Six horses had fractures in the right hind limb (4 had fractures of the middle phalanx, and 2 had fractures of the proximal phalanx), 6 had fractures of the middle phalanx in the left hind limb, 6 had fractures in the left forelimb (4 had fractures of the middle phalanx, and 2 had fractures of the proximal phalanx), and 2 had fractures of the proximal phalanx in the right forelimb. Eighteen fractures were closed, and 2 fractures were open (1 of the proximal phalanx of the left forelimb and 1 of the proximal phalanx of the right hind limb). Fractures occurred during various activities including work and pasture turnout.

Eleven horses were treated with internal fixation of the fracture (screws alone or combined with plate fixation) combined with transfixation pin casts, and 9 horses were treated with transfixation pin casts alone. Of the 11 horses treated with internal fixation, 4 (3 with fractures in a hind limb and 1 with a fracture of the middle phalanx in a forelimb) underwent proximal interphalangeal arthrodesis by use of a single (n = 2) or double plate (2) combined with transfixation pin casting. The remaining 7 horses underwent internal fixation with placement of only lag screws combined with transfixation pin casting (5 with fractures of the proximal phalanx and 2 with fractures of the middle phalanx). Of those 7 horses, 3 (2 with fractures of the proximal phalanx and 1 with a fracture of the middle phalanx) underwent open reduction and 4 (3 with fractures of the proximal phalanx and 1 with a fracture of the middle phalanx) had lag screws placed through stab incisions during fluoroscopic guidance. A mean of 3 lag screws was used for reduction (range, 1 to 4 screws); 4.5-mm cortical bone screws were used to reduce 4 fractures, and 5.5-mm cortical bone screws were used to reduce 3 fractures.

Centrally threaded positive profile pins^a were used in 15 horses, and smooth Steinman pins were used in 5 horses. Two transfixation pins were placed in all horses except one, which was treated with 3 smooth Steinman pins. One-quarter-inch pins (6.3 mm in diameter) were used in all horses. Eleven transfixation pins were inserted in a parallel fashion, and 9 transfixation pins were placed at a 30° divergent angle from the frontal plane. The exact location of pin placement could not be obtained from the medical records or radiographs; however, approximate locations included the following: metaphysis (n = 4), distal aspect of the diaphysis (10), middle aspect of the diaphysis (4), and proximal aspect of the diaphysis (2). Six surgeons treated horses with fractures during this period. Three surgeons (TS, GMB, and GT) performed most of the surgeries (6, 5, and 4, respectively), whereas the other 3 surgeons collectively performed the remaining surgeries.

The mean duration that transfixation pin casts remained on affected limbs was 52 days (median, 49 days; range, 1 to 131 days). After pin removal, a halflimb cast was placed on all horses for various durations depending on the progress of fracture healing (mean, 25 days; range, 1 to 84 days). To evaluate the effect of transfixation pin cast duration on short-term survival (discharged from hospital), horses were categorized into 3 groups: horses maintained in a transfixation pin cast for < 40 days (group 1 [n = 8]), horses maintained in a transfixation pin cast from 40 to 80 days (group 2 [8]), and horses maintained in a transfixation pin cast for > 80 days (group 3 [4]). A greater proportion of horses (11/12) were discharged from the hospital if they were maintained in the transfixation pin cast for > 40 days (P = 0.02) without an increased occurrence of secondary fractures. Although the difference was not significant, secondary fractures of a third metacarpal bone (3/8) occurred more frequently than secondary fractures of a third metatarsal bone (1/12). Additionally, there was no significant difference between whether smooth (n = 5; 3 survived vs 2 euthanized) or threaded transfixation pins (15; 11 survived vs 4 euthanized) were used and short-term survival of horses. Finally, while the difference was not significant, horses with hind limb phalangeal fractures had a more positive short-term survival (10 survived vs 2 euthanized) than horses with forelimb phalangeal fractures (4 survived vs 4 euthanized).

Fourteen (70%) horses were discharged from the hospital, whereas 6 (30%) horses were euthanized during the treatment period. Reasons for euthanasia included fracture of a third metacarpal or metatarsal bone associated with transfixation pins (n = 4; 3 of the third metacarpal bone and 1 of the third metatarsal bone), avascularity of the distal aspect of the limb attributable to an open fracture (1), and fracture collapse after pin removal (1). The cast material did not fail in any of the horses that sustained a fracture of a third metacarpal or metatarsal bone. Of the 4 horses that had a secondary fracture of a metacarpal or metatarsal bone, 3 had fractures originating from the proximal transfixation bone-pin interface and 1 had a fracture at the central transfixation pin (3 pins were used in this horse). Two of those horses had transfixation pins placed in the distal aspect of the diaphysis, 1 had transfixation pins placed in the middle aspect of the diaphysis, and 1 had transfixation pins placed in the proximal aspect of the diaphysis. Two horses had transfixation pins placed in a parallel manner, and 2 horses had transfixation pins placed in a divergent fashion. Three horses had 2 centrally threaded positive profile pins placed, and 1 horse had 3 smooth Steinman pins placed. A short oblique fracture of a third metacarpal bone occurred at the proximal bonepin interface during recovery from anesthesia in 1 of the 4 horses that sustained a secondary fracture. Another sustained a metacarpal fracture originating from the proximal pin hole 7 hours after recovery from anesthesia for a cast change after being maintained in a transfixation pin cast for 33 days. The third horse sustained a metacarpal fracture through the middle transfixation pin after 36 days in the transfixation pin cast. The fourth horse sustained a fracture through the proximal pin hole of the third metatarsal bone 1 week following pin removal (after 49 days) after becoming agitated and kicking the stall wall. A double-plate fixation of the third metatarsal bone was performed, but because the fracture was originally open, the plates became infected and the horse was euthanized. Because of financial constraints, the remaining 3 horses were euthanized with no further attempts at fracture management. The horse that was euthanized because of avascularity sustained a comminuted open fracture of the proximal phalanx of the left forelimb that communicated with the metacarpophalangeal joint. After 35 days in the transfixation pin cast, debridement of the wound revealed devitalized bone and the horse was euthanized. The horse in which the fracture collapsed after transfixation pin cast removal had a severely comminuted fracture of the middle phalanx of the right hind limb that was stabilized in a transfixation pin cast alone. The horse had signs of discomfort in the transfixation pin cast, and the pins were removed 24 days after placement, and a half-limb fiberglass cast was applied. Signs of discomfort in this horse immediately disappeared, but subsequently with time, the proximal phalanx luxated through the plantar aspect of the proximal interphalangeal joint. The horse became severely lame, and the owner chose euthanasia because of the poor prognosis.

Lysis around the transfixation pins was detected in 12 of 20 (60%) horses. Results of bacterial cultures of specimens obtained at the bone-pin interface were positive in 7 of 12 (58%) horses, whereas lysis was detected in the remaining 5 horses, but results of bacterial culture were negative. Organisms isolated from pin tracts included Staphylococcus aureus and species of Pseudomonas, Enterobacter, Klebseilla, and Enterococcus. Three horses with radiographic evidence of lysis of pin tracts with (n = 1) or without (2) positive results of bacterial culture had premature pin loosening. As a result, the affected pin was removed and maintenance of the other pin was required for various periods of time. In 2 of those horses, the distal pin was removed 30 days after placement followed by removal of the proximal pin 53 days after placement, whereas in the third horse, the proximal pin was removed 14 days after placement and the distal pin was removed 28 days after placement. Staging the removal of certain transfixation pins because of premature loosening or infection did not appear to negatively affect fracture stability in these 3 horses, and none of the horses that underwent staged transfixation pin removal sustained a secondary fracture of a third metacarpal or metatarsal bone. Of the 7 horses in which pin-tract infections were detected, 5 were discharged from the hospital with no deleterious effects from the infection. Two horses were euthanized during the treatment period: 1 horse because it sustained a secondary fracture of the third metatarsal bone and the other because the fracture collapsed. The horse that sustained a fracture of the third metacarpal bone had 3 smooth Steinman pins placed in the middle aspect of the diaphysis. Lysis developed in the proximal and distal transfixation pin tracts, and the fracture originated at the middle bone-pin interface and propagated proximally and distally. In the horse in which the fracture collapsed, the transfixation pins had been removed prematurely (27 days after placement) because of infection, which was causing signs of discomfort and disuse of the limb. Although there was radiographic evidence of lysis and results of bacterial culture from the transfixation pin site were positive in that horse, pin loosening was not documented in the medical record when transfixation pins were removed. Lack of pin loosening was also detected in the remaining 9 horses with documented pin-tract lysis, in which pin removal was not staged. Lucency was detected radiographically around certain pin tracts without positive bacterial culture results in 5 other horses. Four of those horses were discharged from the hospital, whereas one was euthanized because of a secondary fracture of the third metacarpal bone sustained 7 hours after transfixation pin removal and placement of a regular cast.

Each horse that was transitioned out of the transfixation pin cast developed some degree of osteopenia. However, only 1 horse developed a complication associated with osteopenia. This horse had a comminuted fracture of the middle phalanx and was maintained in a transfixation pin cast alone for 131 days because of concurrent formation of a solar abscess on the contralateral limb. Disuse osteopenia in this horse contributed to bilateral sesamoid fragmentation and subsequent metatarsophalangeal laxity because of suspensory ligament involvement. The horse was maintained in a Kimzey splint after transfixation pin removal for 8 weeks and then placed in an extended heel shoe. The horse continued to improve and is presently not lame at pasture. No other horses in our study developed complications or laminitis in the contralateral limb.

Long-term follow-up results (minimum 1 year) were obtained for 11 of the 14 discharged horses. Owners of 3 horses could not be contacted. One horse died after discharge because of unrelated causes. Eight horses were not lame at a walk but were mildly lame at a trot, whereas 2 horses were lame at a walk. These findings were based on owners' interpretation. Internal fixation (for 3 middle and 2 proximal phalangeal fractures) had been performed in 5 of the 8 horses that were not lame at a walk. The other 2 horses that were lame at the walk had fractures of the middle phalanx of a hind limb (1 had been repaired with proximal interphalangeal arthrodesis and 1 with lag screw fixation). Owners reported that these 10 horses did not have signs of discomfort while in the pasture and that 3 of 10 horses were being used for light riding, whereas 5 of 10 horses (1 stallion and 4 mares) were being used for breeding. No owners indicated that their horses had chronic signs of pain, and all were satisfied with the outcome.

Discussion

In the study reported here, secondary fracture of the third metacarpal or metatarsal bone was the most serious complication of using transfixation pin casts in adult horses. When constructing a transfixation pin cast, maximal distance between the transfixation pins and proximal end of the cast is recommended, as results of 1 study² suggest that catastrophic failure was more common when transfixation pins were placed near the proximal end of the cast. Additionally, cortical defects from transfixation pins that are ≥ 10% of the diameter of the bone act as stress concentrators, where-as defects > 20% of the diameter of the bone linearly decrease the structural stiffness of the bone in torsional loading.^1,6 Because the diameter of the third metacarpal or metatarsal bone decreases toward the diaphysis, placement of 6.3-mm-diameter pins through the diaphysis results in a greater percentage of cortical defect than if the same size pins were placed within the metaphysis. Therefore, to minimize the stress concentration effect, it is recommended that the distal pin be placed near the metaphysis and the next pin be placed 2 to 3 cm proximal to the first. In our study, all horses that sustained secondary fractures had the proximal transfixation pin placed in the diaphysis (2 in the distal aspect of the diaphysis, 1 in the middle aspect of the diaphysis, and 1 in the proximal aspect of the diaphysis) and the second transfixation pin placed anywhere from 3 to 10 cm distal to the first. This resulted in proximal pin placement at the narrowest cross-sectional area of the third metacarpal or metatarsal bone, potentially creating a large amount of stress concentration at the bone-pin interface. Results of our study also indicated that the increased amount of weight bearing and torsional forces transmitted to transfixation pins in the forelimbs versus the hind limbs potentially contributed to an increased risk of fractures of the third metacarpal bone, as only 1 fracture of the third metatarsal bone was detected in our study, and this was attributable to the horse kicking a wall after pin removal. Furthermore, in horses, the third metacarpal bone is loaded primarily in axial compression; therefore, the dorsal quadrants are subjected to tensile stress.^8,9 Osteostixis (creating drill holes in the dorsal aspect of the third metacarpus or metatarsus as a treatment for dorsal cortex fatigue fractures) acts as a stress concentrator and was found to significantly reduce the breaking strength of the bone.⁸ Because bone is weakest under tensile stress, catastrophic fracture of the third metacarpal bone can occur after osteostixis.⁸ Similarly, horses ambulating with transfixation pin casts potentially create large torsional forces at the bone-pin interface, which may contribute to catastrophic failure. However, when loads are transmitted through transfixation pins positioned 30° divergent from the frontal plane, torsional strength of the metacarpus is significantly greater than when transfixation pins are placed parallel in the frontal plane.⁶ However, only 2 of the horses with secondary fractures in our study had transfixation pins placed parallel in the frontal plane and transfixation pins placed too proximally in all 4 horses with secondary fractures of the third metacarpal or metatarsal bone. Placing the transfixation pins too proximal in the third metacarpal or metatarsal bone was the most likely factor that contributed to the occurrence of catastrophic fractures in our study.

Results of other studies^10,11 indicate that premature loosening of transfixation pins at the bone-pin interface is the most common complication of external skeletal fixation in horses, ultimately leading to loss of stability at the fracture site. With proper insertion techniques, this complication is thought to be reduced by use of threaded transfixation pins, compared with smooth pins. More specifically, positive profile threaded transfixation pins are designed to increase the holding power of the transfixation pin without compromising the internal diameter of the shaft, therefore providing better bone-pin stability and resistance to breakage.¹² However, thermal necrosis and vascular injury during pin placement can cause osteonecrosis and ring sequestration.¹² This contributes to pin loosening and pin tract infections, which results in progressive loss of cortical bone. As the cortical defect becomes large, stress and strain at the bone-pin interface increase and the rigidity of the fixation decreases, thus increasing the potential for catastrophic failure through the bone-pin interface.¹⁰ Factors that may contribute to bone resorption and lysis around pin tracts include high-speed drilling, poor elimination of kerf, and placement of pins without predrilling. Of the 12 horses that had pin tract lysis in our study, 9 did not have any associated complications and were discharged from the hospital, whereas 3 were euthanized (2 because of secondary fractures and 1 because of fracture collapse). The horse that had 3 smooth Steinman pins placed in the middle aspect of the diaphyseal region developed pin tract lysis and infection around the proximal and distal pins. It was theorized that because of pin tract infection, the entire load was placed on the central transfixation pin, thus increasing the stress concentration at that site, which potentially contributed to the cause of the fracture in the third metacarpal bone. The other horse that sustained a fracture of the third metacarpal bone had pin tract lysis with no documented pin tract infection or transfixation pin loosening. After transfixation pin removal (after 7 weeks), a secondary fracture was detected, which was suspected to have occurred during recovery from anesthesia while the limb was in a half-limb cast. The third horse in which the fracture collapsed developed transfixation pin tract infections and subsequently had signs of discomfort while wearing the cast, which necessitated premature removal of the smooth Steinman pins (after 27 days). Whether there was an association between pin tract lysis, infection, and subsequent secondary fractures of the third metacarpal or metatarsal bone cannot be determined from results of the study reported here because of the limited number of cases and the circumstances in which these 2 horses sustained fractures of the third metacarpal bone. However, secondary fractures of the third metacarpal bone in the 2 horses that had radiographic evidence of lysis around transfixation pins occurred within the first 36 days of treatment. This suggests that progressive pin tract osteolysis as was observed in the remaining 9 horses does not increase the risk of subsequent fractures of the third metacarpal or metatarsal bones.

One of the limitations to our study included the fact that 6 surgeons treated horses with transfixation pin casts during a 10-year period. Whereas 3 of the 5 surgeons performed most (n = 15) of the surgeries, there was no association between surgeon and subsequent occurrence of secondary fractures. However, because multiple surgeons treated these fractures, there were variations among pin placement, pin design, and pin insertion techniques. Additionally, use of smooth transfixation pins has widely been replaced by threaded transfixation pins because threaded transfixation pins have increased pull-out strength with less risk of developing osteolysis and infection at the bone pin interface, compared with smooth pins.^13,14 Whereas statistically, there was no association between the use of smooth pins versus threaded pins and outcome, smooth transfixation pins were used in the early years of the study before threaded transfixation pins became available for use in adult horses. Because of the benefits of centrally threaded positive profile pins^a in providing optimal stability at the bone-pin interface,^10–13 these constructs are presently being used at our institution for transfixation pin casting.

Although certain fractures appear to be appropriate candidates for the combination of internal fixation and transfixation pin casting, use of internal fixation was not associated with short-term survival in our study. However, using some form of internal fixation was determined to be appropriate by several clinicians in our study. These horses had fractures in which the fragments were large enough to support lag screw fixation as well as certain horses in which arthrodesis of the proximal interphalangeal joint was considered beneficial. In the study reported here, more proximal phalangeal fractures were repaired with supplemental internal fixation than middle phalangeal fractures because of the necessity to reconstruct the proximal articular surface of the proximal phalanx to reduce the formation of osteoarthritis in the metacarpophalangeal or metatarsophalangeal joint. Although there was no statistical benefit to internal fixation, there were no deleterious effects associated with its use. Additionally, results of available long term follow-up in horses treated with internal fixation indicated that 5 of 7 horses in which internal fixation was performed were not lame at a walk with 2 being used for breeding purposes and 2 being used for light riding. Two of those horses had proximal phalangeal fractures in which the proximal articular margin of the proximal phalanx was reconstructed.

Presently, the duration that transfixation pin casts are maintained in horses is based on surgeon preference. Horses in our study were grouped according to the duration they remained in the transfixation pin cast. Most horses (16/20) were maintained in the transfixation pin cast for < 80 days, and this duration is consistent with current recommendations. Secondary fractures of the third metacarpal or metatarsal bone occurred within the first 36 days of treatment in 3 of the 4 affected horses, greatly decreasing the survival of horses maintained in transfixation pin casts for < 40 days. Horses maintained in a transfixation pin cast for > 40 days had a significantly increased short-term survival with no deleterious effects or increased risk of fracture occurrence. Presently, our recommendation is to maintain horses in transfixation pin casts for > 40 days and < 80 days. By maintaining horses for > 40 days, the risk of fracture collapse from early removal will be minimized, whereas maintaining them for < 80 days will reduce the occurrence of disuse osteopenia.

Transfixation pin casts have been found to effectively transfer axial weight-bearing forces from the bony column to the cast, thereby protecting the distal portion of the limb.^4,15 By doing so, disruptive forces at the fracture site are reduced, thus promoting healing and reducing pain in the horse. Only 1 horse in our study was not fully weight bearing in the postoperative period, and no horses developed laminitis in the contralateral limb. Although transfixation pin casting appears to be an acceptable treatment for horses with comminuted phalangeal fractures, methods to prevent secondary catastrophic fractures would further improve its survival. In the study reported here, horses that sustained a secondary fracture had transfixation pins placed too proximal in the diaphysis than is currently recommended. It is the authors' opinion that this was the major factor contributing to the occurrence of secondary fractures. Therefore, optimal placement of transfixation pins should have the distal transfixation pin placed in the metaphysis and the second pin placed 2 cm proximal to the first at a 30° divergent angle from the frontal plane.⁶ This will permit maximal distance between the transfixation pins and the top of the cast, thereby reducing the potential for secondary fractures to occur.⁶ Indications for use of a transfixation pin cast should not only include highly comminuted phalangeal fractures that are not amenable to internal fixation but should also be considered in certain fractures as a supplement to internal fixation. This would include mildly comminuted fractures, which do not contain an intact strut of bone to fully support internal fixation alone. Fracture and joint reconstruction combined with transfixation pin casting may provide better reduction of articular surfaces and displaced fragments, potentially contributing to less development of osteoarthritis and improved long-term survival.