We retrospectively identified all patients who had sustained a fracture of the talus during a sixty-seven-month period (January 1995 through July 2000) with use of a prospectively designed orthopaedic trauma database. Patients with isolated talar neck fractures without talar body involvement (seventy-seven patients), talar head fractures (nine), isolated lateral process fractures (forty-four), isolated posterior process fractures (eight), or solitary transchondral lesions (thirty-six) were excluded. The remaining fifty-six consecutive patients, who had a total of fifty-seven fractures of the body of the talus that had been treated operatively, were included in the study. The study group included thirty-nine men and seventeen women who had an average age of 34.1 years (range, fifteen to seventy-four years) (
Table I ). Thirty-three patients had an isolated talar body fracture, and twenty-three patients had a fracture of both the body and the neck of the talus. The most common mechanisms of injury were falls from a height (twenty-seven patients) and motor-vehicle accidents (twenty-one patients). The average Injury Severity Score (ISS) was 15.2 (range, 9 to 50)
17 . Twenty-seven patients had an additional injury of the ipsilateral foot or ankle, and fifteen had an injury of the contralateral foot or ankle.
Initial management consisted of assessment and resuscitation according to ATLS (Advanced Trauma Life Support) guidelines. Open wounds were inspected, and sterile dressings were applied. Intravenous antibiotics and tetanus prophylaxis were administered. When possible, urgent reductions of gross deformities and open injuries were performed. Plain radiographs of the foot (including anteroposterior, lateral, and oblique views) and of the ankle (including anteroposterior, lateral, and mortise views) were made for all patients. Fractures were classified according to Orthopaedic Trauma Association (OTA) guidelines
18 . The primary fracture line was located laterally in the sagittal plane in eighteen cases (OTA classification 72-B1), medially in the sagittal plane in nine (72-B2), and in the coronal plane in twenty (72-B3). The fracture was comminuted in ten cases (72-C2). Twenty-three patients had a concomitant talar neck fracture (72-A1). The talar neck fractures were classified according to the system of Hawkins
19 as modified by Canale and Kelly
20 . Nineteen patients had a talar neck fracture with associated subluxation or dislocation of the subtalar joint (Hawkins Group 2), three patients had a talar neck fracture with associated dislocation of the talar body at the tibiotalar joint (Hawkins Group 3), and one patient had talar neck and body fractures with dislocations of the tibiotalar, subtalar, and talonavicular joints (Hawkins Group 4). Eleven of the fifty-seven fractures were open and were classified according to the system of Gustilo and Anderson as type II (one) or type IIIA (ten)
21,22 .
All patients were treated with open reduction and internal fixation. Surgery was performed within the first twenty-four hours for nineteen fractures. Delayed presentation to our facility or associated life-threatening injuries precluded urgent operative management of the remaining thirty-eight fractures. The average time from the injury to fixation was six days (range, six hours to twenty-one days). The surgical approach varied depending on the fracture pattern, associated comminution, and the preference of the operating surgeon. Thirty-seven fractures (in thirty-six patients) were treated through two surgical approaches (medial and lateral). A single medial approach was used in thirteen cases, and a single lateral approach was used in seven cases. An adjuvant osteotomy at the ankle mortise was necessary for exposure of nineteen of the fifty-seven fractures. This consisted of an oblique medial malleolar osteotomy in sixteen patients and a fibular osteotomy in three. The talar body fractures were stabilized with small-fragment and mini-fragment stainless steel implants (
Figs. 1-A, 1-B ,
1-C, and 1-D ).
Postoperative immobilization with the ankle in neutral alignment was continued until the wounds were sealed and soft-tissue swelling had diminished. Ankle-joint-spanning external fixation was used as an adjunct to internal fixation in five patients and was removed at an average of 9.8 weeks (range, six to twelve weeks). No weight-bearing on the affected limb was recommended for the first twelve weeks postoperatively or until fracture union occurred. Progression to full weight-bearing was not delayed by radiographic evidence of osteonecrosis.
Lateral foot, lateral ankle, ankle mortise, and Canale radiographs
23 were made postoperatively, at approximately six and twelve weeks, at six months, and annually thereafter. These radiographs were used to assess the quality of the initial articular reduction and the initial alignment of the talus. Articular incongruity was measured in millimeters of step-off or diastasis. Loss of length or height of the talus or misalignment after initial fixation also were recorded. Radiographic union was defined as disappearance of the fracture lines with bridging trabeculae on plain radiographs. Osteonecrosis was defined as relatively increased density of the talar dome on plain radiographs. Arthrosis was strictly defined as any loss of joint space, osteophyte formation, or the presence of subchondral sclerosis or cysts. Magnetic resonance imaging was not routinely utilized to diagnose osteonecrosis or to follow its course.
Outcomes Evaluation
In order to evaluate the functional outcome and health status of this group of patients, we administered the Foot Function Index (FFI)
24 , a specific lower-extremity outcome-assessment index, and the Musculoskeletal Function Assessment (MFA)
25,26 , a functional outcome measure. The reliability, validity, consistency, and responsiveness of both outcomes instruments have been documented
24-26 .
The FFI is designed to measure the effect of foot and ankle pathology on overall function
24 . It is composed of three subscale scores: pain (81 points), disability (81 points), and activity limitation (45 points). Higher scores indicate greater impairment. The total foot function score is the average of the three subscale scores. The MFA has ten categories: mobility, hand and fine motor function, housework, self-care, sleep and rest, leisure and recreation, family relationships, cognition and thinking, emotional adjustment and adaptation, and employment. Each category is scored independently, and an overall score can be calculated. Values range from 0 to 100, with low scores indicating better function.
Patients were evaluated with the functional outcome questionnaires at a minimum of twelve months after the injury. Attempts were made to locate patients with use of three Internet search engines. The patients were contacted by a trained medical interviewer who was not involved in their treatment. The questionnaires were completed by means of a telephone interview or self-administration during a clinic visit.
Statistical Analysis
Statistical analysis was performed with use of the SAS statistical package (SAS Institute, Cary, North Carolina). The clinical outcome variables included arthritis of the tibiotalar or subtalar joint and osteonecrosis of the talar dome. The functional outcomes included the FFI and MFA scores. Bivariate analysis was used to test the association between possible predictive variables and these outcomes. These variables included wound type (open, closed), fracture location (body only, body and neck), fracture pattern (OTA classification 72-B1, 72-B2, 72-B3, 72-C2), comminution (yes, no), and age (less than forty years, forty years or more). Associations between clinical outcomes and functional outcomes were also examined. The Fisher exact test was used to identify the associations between the clinical outcomes, osteonecrosis, and arthritis and the predictive variables. For this analysis, fracture pattern was dichotomous (with the classification "B" indicating a partial articular fracture and the classification "C" indicating an articular fracture). The Student t test was used to identify the associations between the functional outcomes, FFI scores, and MFA scores and the possible predictive variables. The t test was also used to identify the association between the functional outcomes and the clinical outcomes. In all cases, p < 0.05 was considered significant.
Data on all fractures that are seen at our level-1 trauma center are entered into a trauma database. During the sixty-seven-month study period, a total of 9190 patients were treated operatively for fractures at our institution. The overall prevalence of talar body fractures was 0.62%. Over the same time-period, 818 patients were treated operatively for foot injuries and 230 patients were treated for talar fractures. Therefore, talar body fractures were seen in 6.8% of patients with foot injuries and in 24.3% of all patients with talar injuries during this time-period.
Thirty-eight (68%) of the fifty-six patients with talar body fractures were followed clinically for a minimum of one year (average, thirty-three months; range, twelve to seventy-six months). Of the eighteen patients who were not followed, seventeen could not be located and one refused further evaluation. Twenty-six of the thirty-eight patients had a complete set of radiographs consisting of plain lateral radiographs of the foot and ankle, mortise views, and Canale views.
Nine early complications occurred in eight (21%) of thirty-eight patients. Three patients had development of a superficial wound infection; all were managed successfully with oral antibiotics. Four patients had partial wound dehiscence, and one had skin necrosis. These five soft-tissue complications were successfully treated with dressing changes. One patient with a type-IIIA open talar body fracture and associated bone loss had development of a deep infection postoperatively. The infection resolved after two subsequent operative procedures for irrigation and débridement combined with six weeks of intravenous antibiotics.
An anatomic reduction, as determined on the basis of intraoperative assessment and plain radiographs, was achieved for twenty-one of twenty-six patients. Postoperatively, four patients had 1 to 2 mm of articular incongruity and one had 10° of angular malalignment. No nonunions occurred; however, one patient with delayed union had development of late malalignment, with the fracture uniting in that position.
Ten (38%) of the twenty-six patients with a complete set of radiographs had development of osteonecrosis of the talar body during the follow-up period. In all cases, this finding was evident within the first ten months (mean, 6.2 months; range, three to ten months). Five of these patients had progressive collapse of the talar dome. Collapse occurred within fourteen months after surgery in all cases (mean, 10.2 months; range six to fourteen months) (
Figs. 2-A, 2-B ,
2-C, 2-D , and
2-E ). The other five patients had radiographic evidence of revascularization of the talar body, without collapse, at a mean of 10.4 months after surgery.
Of the ten patients with osteonecrosis, four had a fracture of the talar body only while six had combined talar body and neck fractures. Therefore, osteonecrosis developed in 27% (four) of the fifteen patients with a fracture of the talar body only compared with 55% (six) of the eleven patients with combined fractures of both the talar body and neck (p = 0.35). Collapse of the talar dome occurred in one of the fifteen patients with a fracture of the talar body only, compared with four of the eleven patients with combined fractures of both the talar body and neck (p = 0.089). Osteonecrosis occurred in three of the four patients with open fractures, compared with seven of the twenty-two patients with closed fractures (p = 0.037). The osteonecrosis progressed to collapse in all three patients with open fractures, while revascularization of the talar body was seen in five of the seven patients with closed fractures (p = 0.17). With the numbers available, age, OTA fracture classification, and comminution of the talar body had no apparent effect on the development or progression of osteonecrosis according to bivariate analysis.
Radiographic findings consistent with posttraumatic arthritis were observed in the tibiotalar joint in seventeen (65%) of twenty-six patients and in the subtalar joint in nine (35%) of twenty-six patients. These findings included any narrowing of the joint space, subchondral sclerosis, and osteophyte formation. End-stage arthritis with complete loss of joint space was seen in the tibiotalar joint in six patients, in the subtalar joint in two patients, and in both of these joints in six patients. Fractures of both the talar body and neck were more likely to result in end-stage arthritis than were fractures of the talar body only (p = 0.04). Six of the fifteen patients with talar body comminution had development of end-stage arthritis. All patients with open fractures had advanced posttraumatic arthritis of the tibiotalar or subtalar joint, compared with ten of the twenty-two patients with closed fractures (p = 0.053). With the numbers available, age and OTA fracture classification had no apparent effect on the presence of arthritis (p = 0.35 and p = 0.14, respectively). Notably, only three (12%) of twenty-six patients had no evidence of either osteonecrosis or posttraumatic arthritis on final radiographs.
Fifteen patients had a total of eighteen secondary procedures for the relief of pain. Eleven patients underwent hardware removal. Two patients underwent below-the-knee amputation. One of these amputations was performed several days after the injury because of the severity of associated foot trauma. The other amputation was performed at five months because of intractable pain in a patient who initially had sustained a crush injury and had severe associated bone, soft-tissue, and nerve damage. One patient underwent tibiotalar arthrodesis eleven months postoperatively and subsequently returned to work. Another patient underwent tibiotalar arthrodesis sixteen months after injury, followed by subtalar arthrodesis seven months later. She subsequently underwent revision subtalar arthrodesis because of nonunion. One patient, who had fractures of both the body and neck of the talus as well as of the cuboid, underwent a triple arthrodesis twenty-five months after the injury. At the time of the most recent follow-up, she was considering total ankle arthroplasty because of end-stage tibiotalar arthritis. One patient with a comminuted type-IIIA open fracture of the talar body and neck was treated with a primary subtalar arthrodesis at the time of fracture fixation. She went on to have development of osteonecrosis with collapse six months after the injury. At the time of the most recent follow-up, she was considering additional surgery for the relief of ankle pain. Another patient had a total ankle arthroplasty twelve months after the injury. He returned to his previous occupation of raising horses.
Data on functional outcome and employment status were available for thirty patients. The mean values for the FFI subscales and the total FFI score are presented in the Appendix. The mean subscale values for patients without foot or ankle pathology have been reported as 11 for pain, 15 for disability, and 10 for activity
27 . In contrast, our patient population demonstrated greater impairment, with mean FFI subscale scores of 41 for pain, 37 for disability, and 19 for activity. In the present study, no difference was detected between patients with fractures of the talar body only and patients with combined fractures of the talar body and neck. However, higher scores were seen in association with comminuted fractures and open fractures. Osteonecrosis progressing to collapse of the talar dome also produced higher FFI scores. Additionally, posttraumatic arthritis of the ankle or subtalar joint negatively affected the FFI scores, with the greatest impact on the pain subscale. With the numbers available, the age of the patient and the OTA fracture classification had no apparent effect on the FFI scores.
Outcomes were also evaluated with use of the MFA instrument
25,26 (
Table II ). The mean reference value for patients with hindfoot injuries has been reported to be 22.1
28 . The mean standardized MFA score for all patients in our series, 29.4, was significantly higher than this reference value (p < 0.001), indicating greater disability in our population. No difference in the mean standardized MFA score was detected between patients with fractures of the talar body only and those with fractures of both the talar body and neck. However, osteonecrosis progressing to collapse adversely affected the MFA score (mean, 65.6). The presence of posttraumatic arthritis of the ankle and/or subtalar joint also produced higher MFA scores. With the numbers available, age, OTA fracture classification, comminution, and history of open fracture had no apparent effect on the MFA score.
At the time of the last follow-up, twenty of thirty patients had returned to their previous level of employment. Three of these individuals worked in occupations involving heavy labor. Two patients modified their work routine because of the talar fracture. Eight patients did not return to work. Five of these eight patients had been unemployed before the injury. One patient had retired before the injury but was satisfied with his activity level at the time of the most recent follow-up. One patient was unable to return to work because severe restrictive lung disease and the talar fracture limited his function.
Fractures of the talar body are rare and often devastating injuries. Previous reports in the literature have made little mention of these fractures and the results of their treatment. In previous studies, the majority of patients were treated nonoperatively
1,2,4-7,11 and talar body fractures were not distinguished from other fractures of the talus
8,9,29-31 . Some authors have advocated primary tibiotalar and subtalar arthrodeses in cases of comminuted talar body fractures
1,7 . Sneppen and colleagues, in the previously largest series of talar body fractures, described a classification system, based on the mechanism of injury, that was designed to guide treatment and predict results
11,32 . Of their twenty-one patients with shearing or comminuted fractures of the talar body, eighteen underwent closed treatment and three underwent open reduction and internal fixation. In that series, talar malunion was observed in 60% of patients
11 . Those authors also noted high rates of posttraumatic arthritis of the ankle and subtalar joints. Ninety-five percent of their patients had moderate or severe complaints, and the authors speculated that arthritis was dependent on both malunion and damage to the cartilage at the time of injury. They proposed that these fractures be treated with exact reduction and stable fixation to maximize the outcome. We agree that meticulous surgical technique should decrease the rates of arthrosis due to articular incongruity and talar malalignment.
While arthroscopic and percutaneous techniques have been used to treat talar body fractures to minimize soft-tissue damage and disruption of the blood supply
33,34 , all patients in the present series were treated with an open technique. Careful soft-tissue handling combined with open reduction provides visualization of and access to the fracture, thus optimizing the alignment. The fracture pattern and location will determine the choice of surgical approach. If posterior access is needed, an osteotomy of the medial malleolus
16,35,36 or the fibula should be considered. In the present series, simultaneous medial and lateral approaches, occasionally including an osteotomy, were used most commonly. Dual approaches to the talus have been described for the treatment of talar neck fractures
23,37-39 . Thordarson proposed a lateral approach to remove debris from the subtalar joint in patients with fractures of the talar body
35 . Since fractures of the talar body involve the articular surfaces of both the tibiotalar and subtalar joints, open reduction and rigid internal fixation may restore talar alignment and articular congruity. This allows for early motion of the ankle and subtalar joints, potentially limiting posttraumatic arthritis to that which is caused by chondral damage or loss due to the injury itself. The fractures in the present study were stabilized with small-fragment and mini-fragment screws, which are appropriate sizes for dealing with the small osteochondral fragments that are commonly encountered in association with talar body fractures. These implants appear to be adequate to maintain fracture alignment and to permit early motion of the adjacent joints.
Stainless steel hardware was used in the present study. Some authors have suggested the use of titanium implants to allow for the use of magnetic resonance imaging to assess osteonecrosis
2,35,37,40 . While magnetic resonance imaging can provide information regarding the location and extent of osteonecrosis
40 , false-negative images have also been reported
41 . In the present series, magnetic resonance imaging was not used even if there was evidence of vascular compromise on plain radiographs. While an attempt was made to identify osteonecrosis and/or evidence of revascularization on plain radiographs, we did not alter the weight-bearing status of our patients, even when osteonecrosis was identified radiographically. Currently, the impact of weight-bearing on the progression of osteonecrosis is unknown.
We observed osteonecrosis in ten (38%) of twenty-six patients, with collapse occurring in five of the ten. Rates of osteonecrosis of as high as 40% have been reported after talar body fractures
1,2,4-6,11 . Urgent reduction of dislocations of the talar body may help to maintain any remaining blood supply, thus decreasing the likelihood of osteonecrosis
42-44 . Despite this, the vascular supply to the talar body may be more disrupted or completely destroyed in patients with open injuries, particularly those associated with dislocations. This is consistent with our finding of increased rates of osteonecrosis and collapse in patients with open fractures.
Historically, tibiotalar joint arthritis, subtalar joint arthritis, and talar malunion have been commonly reported following the treatment of talar body fractures
1,2,4,6,7,11 . Sixty-five percent of patients in the present study had development of arthritis in the tibiotalar joint, and 35% had development of arthritis in the subtalar joint. In previous studies, the rate of posttraumatic arthritis in the subtalar joint has ranged from 48% to 90% and the rate of posttraumatic arthritis in the tibiotalar joint has ranged from 50% to 90%
1,2,4-7,11 . A more recent study of talar fractures, by Elgafy and colleagues, included eleven fractures of the talar body
2 . Nine of their eleven patients underwent initial open reduction and internal fixation to restore articular congruity, and one patient each had spanning external fixation and primary tibiocalcaneal fusion. Those authors noted posttraumatic arthritis in both the tibiotalar and subtalar joints in nine of the ten patients who were assessed for arthritis. More arthritis and worse outcomes were seen in association with talar body fractures than in association with other fractures of the talus. The authors remained optimistic, however, that the low prevalence of osteonecrosis, seen in only three of eleven patients, was attributed to early anatomic reduction and fracture stabilization.
The outcomes for polytraumatized patients with foot injuries have been shown to be significantly worse than the outcomes for polytraumatized patients without foot injuries (p = 0.008)
45 . Furthermore, worse outcomes have been reported after talar body fractures compared with talar neck fractures or lateral or posterior process fractures
2 . This was based on three different rating scores
2 . The current study is the first, to our knowledge, to assess the functional outcomes of patients with talar body fractures with use of the FFI and the MFA. FFI scores have been previously used to assess patients with arthritis of the ankle
46 and to evaluate the effect of ankle arthrodesis
27 . The mean subscale values that were reported previously for patients without foot or ankle problems (11 for pain, 15 for disability, and 10 for activity) contrast with the higher scores (indicating greater impairment) in our patient population (41 for pain, 37 for disability, and 19 for activity). The reference mean standardized MFA score for patients with hindfoot and midfoot fractures was previously reported to be 22.1 after twelve months of follow-up
28 . In comparison, the mean standardized MFA overall score for the patients in the present series was 29.4. The higher score for our patients, indicating more diminished function, suggests that fractures of the talar body are associated with poor outcomes relative to other hindfoot or midfoot injuries. Our data revealed even worse outcome scores for patients with osteonecrosis and collapse and for patients with end-stage posttraumatic arthritis.
In conclusion, fractures of the talar body are often severe injuries. Open reduction and internal fixation with small-fragment and mini-fragment implants may restore congruity of the ankle and subtalar joints. However, early complications in this series were not infrequent, and the majority of our patients had radiographic evidence of osteonecrosis and/or posttraumatic arthritis. Associated talar neck fractures and open fractures more commonly resulted in osteonecrosis or advanced arthritis. Open fractures were associated with worse functional outcomes. Although most of our patients returned to their previous level of employment, worse functional outcomes were noted for patients with advanced posttraumatic arthritis of the ankle or subtalar joint and for patients with osteonecrosis of the talar dome that had progressed to collapse. It is important to counsel patients regarding these devastating injuries and their prognosis and potential complications.
A table showing the mean Foot Function Index for all fractures, categorized according to type, is available with the electronic versions of this article, on our web site at www.jbjs.org (go to the article citation and click on "Supplementary Material") and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).