SEARCH

SEARCH BY CITATION

Keywords:

  • horse;
  • Thoroughbred;
  • racehorse;
  • proximal phalanx;
  • fracture

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

Reasons for performing study

Although fractures of the proximal phalanx are one of the most common long bone fractures of Thoroughbred horses in training, limited details on variations in morphology and radiological progression have been published.

Objectives

To describe in detail the configuration of parasagittal fractures of the proximal phalanx in a group of Thoroughbred racehorses, to report fracture distribution within this group of horses and to document radiological progression of fracture healing in cases treated by internal fixation.

Study design

Restrospective case series.

Methods

Case records and radiographs of Thoroughbred racehorses with parasagittal fractures of the proximal phalanx admitted to Newmarket Equine Hospital between 2007 and 2011 were analysed.

Results

One hundred and twenty-one fractures were identified in 120 Thoroughbred racehorses. Fractures were frequently more complex than was appreciated immediately following injury; a feature that has not been reported previously. There was seasonality of fractures in 2- and 3-year-old horses, but not in older horses.

Conclusions and potential relevance

Fractures of the proximal phalanx may be more complex than recognised previously, although often their complexity cannot be identified radiographically immediately following injury. The seasonality observed in 2- and 3-year-old horses is most likely to be a consequence of the timing of the turf-racing season in the UK.

The Summary is available in Chinese – see Supporting information.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

Fractures of the proximal phalanx are common injuries in the racing Thoroughbred [1-4] and Standardbred [5, 6], but are uncommon in other disciplines [7, 8]. In Thoroughbreds, fractures occur more commonly in training than when racing [2-4, 9], and 2-year-olds are most commonly affected [3, 10]. Forelimbs are injured with much greater frequency than hindlimbs, but there is no consistent left/right predilection [1, 5, 10]. In 2 separate studies from the UK, fractures of the proximal phalanx were reported as the first [2] and second [1] most common long bone fracture of Thoroughbreds in race training, accounting for 15% of all fracture types in both studies.

The most commonly reported fracture configuration is parasagittal, extending distally from the sagittal groove of the proximal articular surface of the bone [10, 11]. Fractures may be incomplete or complete. The former have been classified previously as short (extending <30 mm from the articular surface) or long (extending >30 mm) [10, 11]. Incomplete fractures have been reported as the most common configuration in the UK [1, 10], while complete fractures are most frequent in reports from North America [5, 11]. In all previous publications, descriptions of fracture configuration have been limited to radiographs obtained at the time of case presentation. We had observed that parasagittal fractures of the proximal phalanx often appear radiologically more complex and extensive during the resorptive phase of fracture healing, but to date this observation has received only limited mention [10].

A retrospective study of parasagittal fractures of the proximal phalanx in Thoroughbred racehorses referred to a single hospital in the UK was undertaken. Our objectives were to describe in detail the configuration of fractures, to report fracture distribution within this population and to document radiological progression of fracture healing in cases treated by internal fixation.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

Part 1: description of fracture configurations

A retrospective analysis of individual case records and radiographs for all Thoroughbred racehorses admitted to a single referral hospital, Newmarket Equine Hospital, between 1 January 2007 and 31 December 2011 was undertaken. This population has been reported previously [12]. During this period, all cases were evaluated using the same radiographic equipment, data retrieval system and digital measuring program (Synapse v3.2.1)a. To be included as a case, horses had to have a radiologically confirmed parasagittal fracture of the proximal phalanx and associated clinical features, including lameness, metacarpophalangeal/metatarsophalangeal joint distension, with or without pain on palpation of the proximal phalanx (Fig S1a). Multiple injuries in the same horse at different times through the study period were included, but re-injuries at the same site were not included as separate cases.

Data collected included the age, breed and sex of the horse, date of diagnosis and limb affected.

Radiological features were documented from a minimum of 4 standard radiographic projections (including dorsopalmar/dorsoplantar, dorsal 45° lateral palmar/plantaromedial oblique, lateromedial and dorsal 45° medial palmar/plantarolateral oblique). In many cases, additional projections angled between 5 and 15° from dorsopalmar/plantar were also available for review and used to assess fractures. Features documented included the fracture length, configuration, displacement, articular comminution (more than 2 visible fracture lines extending from the articular surface on a dorsopalmar/plantar radiograph) and periosteal or endosteal new bone in the dorsoproximal quadrant of the proximal phalanx. Displacement was classified as mild or moderate. Measurements were obtained from dorsopalmar/plantar radiographs (Fig 1). The lengths of fractures were measured by determining the level of the distal-most extent of the fracture from the proximal articular surface, on a line drawn from the sagittal groove of the proximal articular surface to the sagittal groove of the distal articular surface. Fracture configurations were classified as follows.

  • Short incomplete parasagittal fractures were fractures extending in a parasagittal plane from the sagittal groove of the proximal articular surface into the epiphysis or metaphysis of the bone, <30 mm distally.
  • Long incomplete parasagittal fractures were fractures extending in a parasagittal plane from the sagittal groove of the proximal articular surface into the diaphysis of the bone, ≥30 mm distally.
  • Complete parasagittal fractures were fractures extending in a parasagittal plane from the sagittal groove of the proximal articular surface into the diaphysis of the bone and exiting into either the proximal interphalangeal joint or through any aspect of the cortex of the bone.
  • Comminuted fractures were complete multipiece (≥3 pieces) fractures of any configuration. Fractures were further subdivided and classified as moderately comminuted when an intact strut of cortical bone was present between the metacarpophalangeal/metatarsophalangeal joint and the proximal interphalangeal joint, and highly comminuted when there was no intact strut.

The above classification was used to permit comparison with previous reports. In addition, the length of the entire proximal phalanx was determined from the same sagittal line and the length of the fracture calculated as a percentage of the length of the proximal phalanx. Calculating a percentage value was performed to adjust for radiographic magnification (between cases and sequential examinations in the same case) and variations in horse size.

figure

Figure 1. Dorsopalmar radiograph of a long incomplete parasagittal fracture in a right fore proximal phalanx, demonstrating measurements obtained from preoperative radiographs. A line is first drawn between the centre of the sagittal groove of the proximal articular surface of the bone and the sagittal groove of the distal articular surface. The distance along this line to the level of the distal-most extent of the fracture was then determined to give a measurement of fracture length. Abbreviation: LAT = lateral.

Download figure to PowerPoint

Part 2: radiological progression and fracture healing

Cases from Part 1 were subjected to inclusion criteria for this part of the study (Fig S1b). Radiological progression of fracture healing was evaluated for long incomplete parasagittal and complete parasagittal fractures treated by internal fixation. In all cases, the 4 standard radiographic projections were used for evaluation of fracture healing. Short incomplete parasagittal fractures were excluded from Part 2 of the study because the majority of these cases presented as chronic injuries, without accurate details regarding date of initial fracture and without radiographs available for review at the time of fracture. Due to the large bias towards surgical repair of fractures in the present series, and differences in healing between repaired and conservatively managed cases [10], conservatively managed cases were excluded from further analysis. Post operative radiographs were grouped into different time ranges after surgery, as follows: 1–13, 14–49, 50–74, 75–99 and ≥100 days. Only cases with radiographs available for evaluation taken preoperatively, within the immediate post operative period (1–13 days after surgery) and between 14 and 49 days after surgery were selected for inclusion in Part 2 of the analysis. For each case fulfilling these criteria, additional data were collected including date of surgery, details of surgical repair and any complications related to the fracture.

Post operative radiographs were examined for evidence of additional fracture lines and/or increased length of the fracture line compared with that identified on preoperative images. Resorption along the fracture lines was recorded on radiographs obtained 14–49 days post operatively and classified subjectively as mild, moderate or marked (Fig S2). The amount and degree of organisation of periosteal and endosteal callus was documented at 14–49 days post operatively and all subsequent radiographic time points. The amount of callus was subjectively classified as mild, moderate or marked (Fig S2) and the organisation of the callus as irregular or organised. Progressive organisation and radiological healing of the fracture between time points was also documented. When there was no longer a linear radiolucency and/or radiodense material bridged the site where a fracture line had been identified previously, fractures were classified as healed. Complete healing was defined as the above, with recorticalisation of fracture planes and replacement of periosteal and endosteal callus by organised bone. Fractures were classified as delayed unions if there was no evidence of bridging bone across part or all of the fracture line on radiographs obtained ≥100 days after surgery.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

One hundred and twenty-one fractures of the proximal phalanx were identified in 120 horses. Fractures were classified as short incomplete parasagittal (n = 14), long incomplete parasagittal (n = 88), complete parasagittal (n = 14) and comminuted (n = 5; Table 1). The horse included twice suffered a complete parasagittal fracture of the right fore proximal phalanx as a 2-year-old and a long incomplete parasagittal fracture of the contralateral limb as a 3-year-old.

Table 1. Distribution of 121 fractures of the proximal phalanx
ClassificationForelimbsHindlimbs
LeftRightTotalLeftRightTotal
Short incomplete parasagittal21012112
Long incomplete parasagittal333063151025
Complete parasagittal8412112
Comminuted325000
Total464692171229

Injured horses were fillies (n = 62), colts (n = 44) and geldings (n = 15). Different fracture configurations were seen with similar frequency in fillies, colts and geldings. Horses were 2–9 years old (median 2 years old; interquartile range 2–3 years old; Fig 2). Two-year-old horses were most frequently affected, and 88% cases were either 2 or 3 years old. Overall, forelimbs were affected more frequently than hindlimbs, and this was also true for individual fracture configurations, including long incomplete parasagittal, short incomplete parasagittal, complete parasagittal and comminuted (Table 1). Left and right forelimbs were affected similarly, although short incomplete parasagittal fractures occurred more frequently in right compared with left forelimbs (Table 1).

figure

Figure 2. The distribution of 121 fractures of the proximal phalanx by horse age at the time of fracture.

Download figure to PowerPoint

One-hundred and eleven of 121 fractures (92%) occurred in horses that were unraced but in training for racing on the flat, racing in the turf flat season or in ‘all-weather’ flat races (short incomplete parasagittal n = 12, long incomplete parasagittal n = 80 and complete parasagittal n = 8). Seven fractures (6%) occurred in horses training or racing under jump-racing rules (short incomplete parasagittal n = 1 and long incomplete parasagittal n = 6). In 3 cases, it is unknown whether horses were training for flat racing on turf or all-weather surface, or under jump-racing rules. In horses aged ≥4 years old, where information is available (n = 13), 6 (46%) were in flat race training and 7 (54%) in jump race training. Only 5 fractures occurred while racing (long incomplete parasagittal racing on turf n = 2, long incomplete parasagittal racing on all-weather surface n = 1 and complete parasagittal racing on turf n = 2), whereas 116 occurred while training.

Fractures occurred during every month but exhibited seasonality, with most fractures occurring between March and October (110 of 121; 91%). However, between March and October there was a similar monthly frequency (Fig 3). The seasonality of fractures was exhibited principally by 2- and 3-year-old horses, with 66 of 69 (96%) and 35 of 38 (92%) cases, respectively, occurring during the months of March to October. No observable seasonality of injury was found in horses aged ≥4 years old (Fig 3).

figure

Figure 3. Seasonal distribution of 121 fractures of the proximal phalanx with 2- and 3-year-old horses grouped separately from horses 4 years and older.

Download figure to PowerPoint

Parasagittal fractures

In all cases, parasagittal fractures originated in the sagittal groove at the proximal articular surface of the proximal phalanx. Incomplete fractures were 7–95 mm (mean 49.53 mm) in length, and extended 7–93% (mean 51%) of the measured length of the proximal phalanx. Two complete parasagittal fractures exited through the lateral cortex of the bone in the middle third of the diaphysis (Fig S3a), and the remaining 12 exited into the proximal interphalangeal joint. In 82 (71%) cases, fractures remained in a parasagittal orientation confined to the axial third of the bone (including 2 complete parasagittal fractures). In 21 (18%) cases, fractures coursed abaxially into the lateral (n = 19) or medial (n = 2) third of the bone as they extended distally, within the proximal (n = 4), middle (n = 15) or distal (n = 2) third of the bone. In 13 (11%) cases, fractures turned from parasagittal into an oblique (dorsolateral-palmaromedial) plane in the proximal (n = 1), middle (n = 6) or distal (n = 6) third of the bone. Ten of these were complete parasagittal fractures and exited through the lateral condyle into the proximal interphalangeal joint (Fig S4). Complete fractures were displaced in 12 (86%) cases (mild n = 9 and moderate n = 3). In 7 cases, displacement was limited to a mediolateral plane, and in 5 there was both mediolateral and proximodistal displacement. Twenty-three fractures (20%) had radiological evidence of proximal articular comminution (Fig S3b), and in one case there was comminution in the mid-diaphysis of the bone. Twenty-one cases (18%) had irregular periosteal and/or endosteal new bone in the dorsoproximal quadrant of the proximal phalanx in the affected limb. Of these, 6 short incomplete parasagittal cases had been managed conservatively for periods of 6–12 weeks prior to presentation.

Comminuted fractures

Four fractures were classified as moderately comminuted and one as highly comminuted. In all moderately comminuted fractures, the intact strut of bone between the proximal and distal articular surfaces was medial.

Radiological progression and fracture healing

Forty-seven cases met the inclusion criteria for Part 2 of the analysis (long incomplete parasagittal n = 39 and complete parasagittal n = 8). Fractures were repaired by percutaneous placement of 2 (n = 9), 3 (n = 29) or 4 (n = 9) 4.5 mm AO/ASIF cortex screws (Synthes)b inserted using a lag technique [13] except for the distal (4th) screw in 2 complete parasagittal cases, where 3.5 mm screws were used. In 3 cases, the thread of a single screw was stripped, requiring replacement with a 5.5 mm screw. Screws were placed perpendicular to the radiologically defined fracture plane from preoperative images. All screws were placed in a frontal plane from lateral to medial, except for the distal screw of one long incomplete parasagittal and 5 complete parasagittal cases and the distal 2 screws of 2 complete parasagittal cases, which were inserted in an oblique (palmarolateral-dorsomedial) orientation.

Review of radiographs obtained in the immediate post operative period (1–13 days) revealed in all cases that the fracture appeared well compressed, with no evidence of progression when compared with radiographs obtained before surgery.

Review of radiographs obtained 14–49 days after surgery revealed in all cases radiological evidence of resorption of bone along the fracture line, with the fracture appearing wider than seen on radiographs obtained in the immediate post operative period. Resorption was subjectively classified as mild (n = 17), moderate (n = 25) or severe (n = 5). Three fractures previously classified as incomplete were now observed radiologically to be complete, with 2 exiting through the lateral cortex and one into the proximal interphalangeal joint. A further 8 fractures could now be observed radiologically to extend into the lateral cortex of the bone (but not exit), where previously they could not. In 30 of 39 (77%) long incomplete parasagittal fractures, the fracture line was identified further distal in the diaphysis compared with preoperative radiographs. The additional length of existing fracture lines apparent radiologically maintained the same configuration (n = 4), turned laterally in the middle (n = 9) or distal (n = 2) third of the diaphysis (Fig 4), turned medially in the middle (n = 1) or distal (n = 1) third, or spiralled into an oblique (dorsolateral-palmar/plantaromedial) plane in the middle (n = 2) or distal (n = 4) third. Additional fracture lines that became apparent at radiographs 14–49 days post operatively resulted in altered fracture configurations. In 6 cases, the fracture configuration was now an inverted Y, in either the middle (n = 3) or distal (n = 3) third of the bone. In one case, fracture configuration was now complex, with multiple fracture lines apparent in the middle third of the bone. The fracture lines were apparent further distal within the diaphysis of the proximal phalanx a mean of 15.4 mm (range 2–34 mm) or 14% (range 2–34%) of the length of the proximal phalanx. In 6 of 8 complete parasagittal fractures, there was no change in radiologically defined fracture configuration compared with earlier radiographs. In one case, there were further fracture lines apparent in the diaphysis of the bone, in frontal and oblique planes, with a further exit point in the palmar cortex proximal to the condyles (Fig S5). In the other case (in which the fracture had exited laterally on preoperative images), additional fracture lines, including an inverted Y configuration, became apparent radiologically. Radiologically evident periosteal and endosteal callus had formed in all cases and was classified as mild (n = 17), moderate (n = 25) or marked (n = 5).

figure

Figure 4. Preoperative (a) and 37 days post operative (b) dorsopalmar radiographs of Case 18 with a long incomplete parasagittal fracture of the right fore proximal phalanx. The fracture appears longer and turns laterally in the middle third of the diaphysis on the post operative radiograph. Abbreviation: LAT, lateral.

Download figure to PowerPoint

Thirty-seven cases had radiographs available for review between 50 and 99 days after surgery and, if fractures had not healed on these radiographs, further radiographs at a later time point. Radiological fracture healing occurred in 30 cases and was observed at 14–49 (n = 1), 50–74 (n = 9), 75–99 (n = 16) and ≥100 days (n = 4) post operatively. Complete fracture healing was observed in only 12 cases (11 long incomplete parasagittal and one complete parasagittal fracture), and occurred at 50–74 (n = 3), 75–99 (n = 4) and ≥100 days (n = 4) post operatively. Fracture healing commonly took longest to occur proximally, in and/or adjacent to the proximal subchondral bone plate. Seven cases were classified as delayed unions; in 6 long incomplete parasagittal fractures, healing remained incomplete proximally, with persistence of irregular dorsal periosteal new bone and in one complete parasagittal fracture with incomplete fracture healing distally. In all cases, callus organisation improved progressively from radiographs obtained 14–49 days post operatively. The greatest degree of callus formation that occurred in each case during fracture healing was classified as mild (n = 3), moderate (n = 26) or marked (n = 8).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

This study was undertaken primarily to document observations from clinical cases with fractures of the proximal phalanx. Fractures are frequently more complex than reported previously and than radiologically appreciable in the acute phase following injury. The true complexity of fractures was most apparent 14–49 days following injury, during the resorptive phase of fracture healing. During the period 14–49 days following surgery, fractures were longer and had more complex configurations than was appreciated on radiographs taken at presentation. After internal fixation, the majority of fractures healed by 100 days from injury.

Descriptive epidemiology

The case details of affected horses closely resemble those reported previously in a geographically similar pool of Thoroughbred racehorses with fractures of the proximal phalanx [10]. The seasonality of parasagittal fractures of the proximal phalanx in 2- and 3-year-old horses matches the UK flat racing season and is consistent with previous observations [1]. The apparent lack of seasonality of fractures in horses aged 4 years and over may be explained by a combination of older horses training and racing year-round on all-weather surfaces, and training and racing out of the UK flat season under jump-racing rules.

Fracture configurations

Parasagittal fracture of the proximal phalanx in this study invariably propagated distally from the sagittal groove of the proximal articular surface, which is consistent with previous reports [5, 10, 11] and observations [14]. In keeping with an earlier report from the UK [10], incomplete fractures were more common than complete parasagittal fractures. However, in the present series long incomplete parasagittal fractures were more common than short incomplete parasagittal fractures, which is a reversal of previous observations [10].

Results from this study suggest that the majority of fractures adopt a limited number of configurations as they propagate distally from the sagittal groove of the proximal articular surface of the proximal phalanx. The most common configuration observed was long incomplete parasagittal, extending into the middle third of the diaphysis. Fractures that extended further distal most commonly diverged laterally before either extending into the lateral cortex of the bone or spiralling into an oblique (dorsolateral-palmar/plantaromedial) plane approaching the distal epiphysis and proximal interphalangeal joint. These configurations were also commonly observed as part of more complex comminuted fractures. Lateral deviation as fractures extended distally was recognised in previous studies [5, 10]. Ellis et al. [10] observed very few fractures to spiral as they extended distad, whereas this configuration was relatively common in the present series, accounting for 11% of the total and 7% of complete parasagittal fractures. Other reports have not provided sufficiently detailed descriptions of fracture configurations to permit comment or comparison. The difference between the present series and that of Ellis et al. [10] may in part be explained by the much higher proportion of short incomplete parasagittal fractures in the earlier report. Fackelman [14] suggested that the direction and degree of force applied by the third metacarpal/metatarsal bone and the shape of the proximal phalanx dictated the length and configuration of fracture, with the smaller lateral articular surface of the proximal phalanx and the consequent eccentric position of the sagittal groove lateral to the mid-line dictating that fractures propagate down that side. The cause for such repeatable configurations cannot be determined from this study, although structural features of the proximal phalanx, its relationship to the third metacarpal/metatarsal bone and their movement and loading patterns are all potentially involved.

Fracture complexity

It has been observed in the present series that in the acute phase following injury, fractures appeared radiologically shorter and less complex than at subsequent time points (following surgical repair). One possible explanation for the greater length and complexity at later time points is fracture propagation following surgery, as recovery from general anaesthesia represents an enhanced risk period for fracture propagation [15]. If fracture propagation occurred during recovery from general anaesthesia, this should be identifiable on early post operative radiographs. However, in no case from the present series was propagation observed radiologically within the first 14 days following general anaesthesia and surgery; propagation during recovery from general anaesthesia is therefore considered an unlikely explanation.

Osteoclastic resorption of bone at fracture margins is a normal part of healing and occurs during the first weeks following injury [16]. Temporally normal osteoclastic resorption coincides with the time points in the present study when fractures were observed to be more extensive or complex than at the time of diagnosis. It is the authors' opinion that resorption along already existing fracture lines (and not fracture propagation) is the most plausible explanation for the increase in length and/or complexity of fractures observed between 14 and 49 days following surgical repair. Ellis et al. [10] reported that in conservatively managed cases with short incomplete parasagittal fractures, it often took days for fractures to become radiologically apparent, supporting resorption as a plausible explanation for fractures appearing more extensive at time points after initial diagnosis.

The authors have observed clinically marked disparity in the degree of lameness following surgery between horses diagnosed with apparently similar fractures of the proximal phalanx. The observation that the true fracture configuration may be more complex than recognised on radiographs at the time of diagnosis and acute phase treatment is a potential and plausible explanation.

Although fractures may be more extensive and/or complex than radiographs initially identify, there was no instance of catastrophic fracture propagation following surgery in the present series. Additionally, as more complete radiological assessment of fractures does not appear possible until ≥14 days post injury, the clinical approach to acute fracture repair is justified. However, surgeons should be cogniscant of the possibility of increased fracture complexity and the effect this may have on post operative management, fracture healing and convalescence. As a result of this study, post operative radiographs during the resorptive phase of fracture healing are recommended.

Articular comminution

Proximal articular comminution was observed with reasonable frequency (20% of parasagittal fractures) in the present series. Articular comminution has not been documented previously with incomplete parasagittal fractures of the proximal phalanx. Articular comminution associated with fractures of the lateral condyle of the third metacarpal/metatarsal bone is associated with worse outcomes following surgery. Bassage and Richardson [17] reported lower lifetime earnings post injury in Thoroughbred racehorses with complete lateral condylar fractures and concurrent articular comminution in comparison to those without articular comminution. Zekas et al. [18] reported that 11 of 21 (52%) Thoroughbred racehorses with lateral condylar fractures and articular comminution started a race following injury, in comparison with 26 of 38 (68%) cases with complete lateral condylar fractures and no articular comminution. Rick et al. [19] reported that few horses with comminution at the palmar (plantar) surface at the fracture site returned to racing soundness. The effect of articular comminution on case outcome remains undetermined for fractures of the proximal phalanx, and further work is necessary to document any significance.

Fracture healing

The pattern and time frame for fracture healing to occur in the present series appeared relatively predictable following internal fixation. Radiologically, healing was characterised by progressive bone resorption along the fracture line(s), which became apparent during the first 14–49 days post operatively, followed but overlapped by a phase of bone formation in the form of bridging periosteal and endosteal callus. Subsequently, a prolonged phase of remodelling occurred. The proximal epiphyseal portion, including the subchondral bone plate, of the fracture was usually slowest to heal radiologically, similar to observations by Rick et al. [19] on the healing of fractures of the metacarpal and metatarsal condyles. In their study, it was suggested that slower healing adjacent to the articular surface may be a result of joint fluid being forced into the fracture during weightbearing, encouraging fibrocartilage formation rather than bone healing. Long incomplete parasagittal fractures generally healed more quickly than complete parasagittal fractures, and this may relate to the degree of osseous trauma sustained at the time of injury. Delayed unions occurred in relatively few cases (19%), occurring almost invariably proximally (86%), adjacent to the articular surface. The radiological appearance in these cases also suggests that healing is usually slowest dorsally, with persistent irregular periosteal new bone in this location. The reason for slower healing dorsoproximally is unknown. Radiological observations from the cases in the present series (reported in Smith and Wright [12]) have identified prodromal changes in a proportion of cases, occurring consistently in the dorsoproximal quadrant of the bone. Prodromal changes may be contributory to the slower healing observed dorsoproximally. In cases with prodromal changes, acute fracture planes occur through already pathological bone, necessitating greater remodelling to restore normal osseous architecture. A relatively lesser degree of compression in the dorsal compared with the central region may account for the slower/poor healing dorsally. The authors have observed successful fracture healing in such cases following placement of an additional 3.5 or 4.5 mm AO/ASIF cortex screw in lag fashion from lateral to medial in the dorsoproximal quadrant of the proximal phalanx (M. R. W. Smith and I. M. Wright, unpublished data).

Limitations

To enable comparison of the present series with historical reports, absolute measurements were used to classify incomplete fractures as either long incomplete parasagittal or short incomplete parasagittal. A limitation of using absolute measurements is that it does not account for radiographic magnification or variation in horse size, potentially leading to misclassification. In addition, for the same reasons, using absolute measurements only could lead to incorrect assumptions about increases in fracture length between time points. To minimise errors from use of absolute measurements, we also calculated the ratio of fracture length to the length of the entire proximal phalanx. In this series, however, there was relatively little difference between the size of the proximal phalanx between horses, and the effect of varying degrees of radiographic magnification appeared small. The mean length of the proximal phalanx of horses with incomplete fractures was 98 mm, making the classification for long incomplete parasagittal fractures on average 31% of the length of the proximal phalanx. Using a value of 31% as a cut-off, no horse would have changed classification. Additionally, the difference in the increase in fracture length between time points when using the ratio compared with absolute measurements was small.

The precision of measurements, particularly with respect to determining increases in fracture length between time points, may have been affected by inaccuracies in measurements and slight differences in radiographic positioning, making fracture lines more or less apparent. However, the mean increase in length of fractures between time points was far greater than differences likely to be explained by the above considerations, making the overall effect of potential errors small.

Conclusions

Fractures of the proximal phalanx are relatively common injuries in young Thoroughbred horses in training for flat racing. A parasagittal configuration is most common, and in the UK long incomplete parasagittal fractures are more common than any other specific configuration. Fractures follow relatively few reproducible paths as they propagate distad in the proximal phalanx, although the true complexity of fractures often is not radiologically apparent at the time of diagnosis. Fracture healing usually occurs within 100 days when treated by internal fixation.

Authorship

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information

Matthew Smith and Ian Wright contributed equally to study conception and design. Matthew Smith collected data, executed the study and analysed data. Matthew Smith and Ian Wright both contributed to data interpretation and preparation of the manuscript.

Manufacturers' addresses
  1. aFujifilm Medical Systems (UK), Bedford, Bedfordshire, UK.

  2. bSynthes (UK), Welwyn Garden City, Hertfordshire, UK.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information
  • 1
    Ramzan, P.H.L. and Palmer, L. (2011) Musculoskeletal injuries in Thoroughbred racehorses: a study of three large training yards in Newmarket, UK (2005–2007). Vet. J. 187, 325-329.
  • 2
    Bathe, A.P. (1994) 245 Fractures in Thoroughbred racehorses: results of a 2-year prospective study in Newmarket. Proc. Am. Assoc. Equine Practnrs. 40, 175.
  • 3
    Johnson, B.J., Stover, S.M., Daft, B.M., Kinde, H., Read, D.H., Barr, B.C., Anderson, M., Moore, J., Woods, L. and Stoltz, J. (1994) Causes of death in racehorses over a 2 year period. Equine Vet. J. 26, 327-330.
  • 4
    Parkin, T., French, N., Riggs, C., Morgan, K., Clegg, P., Proudman, C., Singer, E.R. and Webbon, P.M. (2004) Risk of fatal distal limb fractures among thoroughbreds involved in the five types of racing in the United Kingdom. Vet. Rec. 154, 493-497.
  • 5
    Holcombe, S.J., Schneider, R.K., Bramlage, L.R., Gabel, A.A., Bertone, A.L. and Beard, W.L. (1995) Lag screw fixation of noncomminuted sagittal fractures of the proximal phalanx in racehorses: 59 cases (1973–1991). J. Am. Vet. Med. Assoc. 206, 1195-1199.
  • 6
    Tetens, J., Ross, M.W. and Lloyd, J.W. (1997) Comparison of racing performance before and after treatment of incomplete, midsagittal fractures of the proximal phalanx in standardbreds: 49 cases (1986–1992). J. Am. Vet. Med. Assoc. 210, 82-86.
  • 7
    Kuemmerle, J.M., Auer, J.A., Rademacher, N., Lischer, C.J., Bettschart-Wolfensberger, R. and Fürst, A.E. (2008) Short incomplete sagittal fractures of the proximal phalanx in ten horses not used for racing. Vet. Surg. 37, 193-200.
  • 8
    Dyson, S.J., Nagy, A. and Murray, R. (2011) Clinical and diagnostic imaging findings in horses with subchondral bone trauma of the sagittal groove of the proximal phalanx. Vet. Radiol. Ultrasound 52, 596-604.
  • 9
    Peloso, J.G., Mundy, G.D. and Cohen, N.D. (1994) Prevalence of, and factors associated with, musculoskeletal racing injuries of thoroughbreds. J. Am. Vet. Med. Assoc. 204, 620-626.
  • 10
    Ellis, D.R., Simpson, D.J., Greenwood, R.E. and Crowhurst, J.S. (1987) Observations and management of fractures of the proximal phalanx in young Thoroughbreds. Equine Vet. J. 19, 43-49.
  • 11
    Markel, M.D. and Richardson, D.W. (1985) Noncomminuted fractures of the proximal phalanx in 69 horses. J. Am. Vet. Med. Assoc. 186, 573-579.
  • 12
    Smith, M.R.W. and Wright, I.M. (2013) Are there radiologically identifiable prodromal changes in Thoroughbred racehorses with parasagittal fractures of the proximal phalanx? Equine Vet. J. 46, 88-91.
  • 13
    Nixon, A.J., Watkins, J.P. and Auer, J.A. (1996) Principles of fracture fixation. In: Equine Fracture Repair, Ed: A.J. Nixon , W. B. Saunders Co., Philadelphia, Pennsylvania. pp 63-86.
  • 14
    Fackelman, G. (1973) Sagittal fractures of the first phalanx (P1) in the horse: fixation by the lag screw principle. Vet. Med. Small Anim. Clin. 68, 622-636.
  • 15
    Richardson, D.W. (1984) Medial condylar fractures of the third metatarsal bone in horses. J. Am. Vet. Med. Assoc. 185, 761-765.
  • 16
    Rahn, B.A. (2002) Bone healing: histologic and physiologic concepts. In: Bone in Clinical Orthopedics, 2nd edn., Ed: G.E. Fackelman , George Thieme Verlag, Stuttgart. pp 287-326.
  • 17
    Bassage, L.H. and Richardson, D.W. (1998) Longitudinal fractures of the condyles of the third metacarpal and metatarsal bones in racehorses: 224 cases (1986–1995). J. Am. Vet. Med. Assoc. 212, 1757-1764.
  • 18
    Zekas, L.J., Bramlage, L.R., Embertson, R.M. and Hance, S.R. (1999) Results of treatment of 145 fractures of the third metacarpal/metatarsal condyles in 135 horses (1986–1994). Equine Vet. J. 31, 309-313.
  • 19
    Rick, M.C., O'Brien, T.R., Pool, R.R. and Meagher, D. (1983) Condylar fractures of the third metacarpal bone and third metatarsal bone in 75 horses: radiographic features, treatments, and outcome. J. Am. Vet. Med. Assoc. 183, 287-296.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Ethical animal research
  9. Sources of funding
  10. Acknowledgements
  11. Authorship
  12. References
  13. Supporting Information
FilenameFormatSizeDescription
evj12094-sup-0001-si.docx17K

Summary in Chinese.

evj12094-sup-0002-figs1a.tiff1521K

Fig S1: Flow diagrams demonstrating the case selection criteria for inclusion into part 1 (a) and part 2 of the study (b).

evj12094-sup-0003-figs1b.tiff1521K
evj12094-sup-0004-figs2a.tiff212K

Fig S2: Pre- and post operative radiographs demonstrating varying degrees of resorption and callus associated with the fracture between 14 and 49 days following surgery. (a–d) Pre- and 21 days post operative dorsopalmar and lateromedial projections of Case 25 with mild resorption and mild callus formation. (e–j) Pre- and 40 days post operative dorsopalmar, dorsal-45°-lateral palmaromedial oblique and lateromedial projections of Case 27 with moderate resorption and marked callus formation. (k–n) Pre- and 28 days post operative dorsopalmar and lateromedial projections of Case 37 with marked resorption and moderate callus formation.

evj12094-sup-0005-figs2b.tiff233K
evj12094-sup-0006-figs2c.tiff269K
evj12094-sup-0007-figs2d.tiff221K
evj12094-sup-0008-figs2e.tiff262K
evj12094-sup-0009-figs2f.tiff287K
evj12094-sup-0010-figs2g.tiff238K
evj12094-sup-0011-figs2h.tiff267K
evj12094-sup-0012-figs2i.tiff284K
evj12094-sup-0013-figs2j.tiff248K
evj12094-sup-0014-figs2k.tiff358K
evj12094-sup-0015-figs2l.tiff328K
evj12094-sup-0016-figs2m.tiff340K
evj12094-sup-0017-figs2n.tiff329K
evj12094-sup-0018-figs3a.tiff277K

Fig S3: (a) Dorsopalmar radiograph of Case 104 with a complete parasagittal fracture of the right fore proximal phalanx, which exits through the lateral cortex in the distal third of the diaphysis. (b) There are multiple fracture lines visible proximally, meeting the study definition of articular comminution. Abbreviation: LAT, lateral.

evj12094-sup-0019-figs3b.tiff227K
evj12094-sup-0020-figs4a.tiff263K

Fig S4: Dorsopalmar (a) and dorsolateral-palmaromedial oblique radiographs (b) of Case 108 with a complete parasagittal fracture of the left fore proximal phalanx, which spirals in the distal third of the diaphysis into an oblique plane before exiting through the lateral condyle into the proximal interphalangeal joint. Abbreviations: DLPMO, dorsal-45°-lateral palmaromedial oblique; and LAT, lateral.

evj12094-sup-0021-figs4b.tiff262K
evj12094-sup-0022-figs5a.tiff243K

Fig S5: Preoperative (a–c) and 44 days post operative (d–f) dorsopalmar, dorsal-45°-lateral palmaromedial oblique and lateromedial radiographs of Case 109 with a complete parasagittal fracture of the right fore proximal phalanx. The fracture has a more complex configuration on the post operative radiographs. Abbreviations: DLPMO, dorsal-45°-lateral palmaromedial oblique; and LAT, lateral.

evj12094-sup-0023-figs5b.tiff251K
evj12094-sup-0024-figs5c.tiff605K
evj12094-sup-0025-figs5d.tiff257K
evj12094-sup-0026-figs5e.tiff274K
evj12094-sup-0027-figs5f.tiff1046K

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.