We retrospectively reviewed the records of 100 consecutive patients with
102 fractures of the talar neck (classification 72-A1 according to the system
of the Orthopaedic Trauma
Association28) who
had been managed operatively at a level-1 trauma center over a
sixty-seven-month period. Sixty male patients and forty female patients with
an average age of 32.6 years (range, thirteen to seventy-seven years) and an
average Injury Severity
Score29 of 15.8
points (range, 9 to 50 points) were identified. The mechanism of injury was a
motorvehicle accident for sixty patients, a fall from a height for
twenty-seven, a motorcycle accident for seven, a pedestrianmotor vehicle
accident for two, sports-related trauma for two, a plane crash for one, and an
industrial accident for one.
Initially, patients were assessed and resuscitated according to Advanced
Trauma Life-Support guidelines (American College of Surgeons, Chicago,
Illinois). Sterile dressings were applied to open wounds, and intravenous
antibiotics and tetanus prophylaxis were administered. Twenty-four fractures
were open; of these, one was classified as type I, one was classified as type
II, and twenty-two were classified as type IIIA according to the criteria of
Gustilo and
Anderson30,31.
Plain radiographs of the foot and ankle were made in all cases. The fractures
were classified into groups as described by
Hawkins14 and
modified by Canale and
Kelly13. Four
fractures were minimally displaced (group I), sixty-eight were associated with
subluxation or dislocation of the subtalar joint (group II), twenty-five were
associated with dislocation of the tibiotalar joint (group III), and five were
associated with dislocation of the tibiotalar, subtalar, and talonavicular
joints (group IV). Twenty-three patients had a contiguous talar body fracture
and/or fracture extension, and twenty had an associated lateral talar process
fracture. Forty-four patients had additional ipsilateral foot and ankle
injuries, and twenty-six had contralateral foot and ankle injuries.
Comminution was defined on plain radiographs as a talar neck fracture with
more than three fracture fragments.
Although closed, manipulative reduction was attempted at the time of the
initial assessment, all fractures subsequently were treated with open
reduction and internal fixation with stainless steel small-fragment and/or
mini-fragment implants. An attempt was made to perform surgery on an urgent
basis. However, associated life-threatening injuries or a delay in diagnosis
or presentation to our hospital precluded urgent treatment in some cases. Dual
anteromedial and anterolateral surgical approaches were used for ninety-one
fractures (Figs. 1-A, 1-B,
1-C, and
1-D)17.
The remaining fractures were treated through a single medial approach (eight
fractures) or a single lateral approach (three fractures). Eight patients also
had an osteotomy of the medial malleolus to enhance the surgical exposure. Six
of these osteotomies were performed in patients who had associated fracture
extensions involving the talar body.
Postoperative management consisted of immobilization of the ankle in
neutral alignment until the wounds were sealed and the swelling had
diminished. Range-of-motion exercises for the ankle and foot were then
initiated. No weight-bearing on the affected limb was recommended for a total
of twelve weeks after surgery or until fracture union had occurred.
Progressive weight-bearing, however, was not delayed by radiographic evidence
of osteonecrosis. Plain radiographs of the lateral aspects of the foot and
ankle, ankle mortise radiographs, and
Canale13
radiographs were made after fixation and at approximately six-week,
twelve-week, six-month, and twelve-month intervals postoperatively.
Computerized tomography was not used to assess fractures preoperatively or to
evaluate alignment postoperatively. Osteonecrosis was defined on plain
radiographs as any area of increased density of the talar dome relative to the
adjacent structures. Magnetic resonance imaging was not routinely used to
diagnose osteonecrosis or to follow its progression. Posttraumatic arthritis
was defined as any loss of joint space, formation of osteophytes, or
development of subchondral sclerosis or cysts.
Functional outcome measurements included the Foot Function Index (FFI) and
the Musculoskeletal Function Assessment (MFA). These measurements were
collected at the most recent clinic visit or by means of a questionnaire that
was administered over the phone or by mail. The FFI is a specific
lower-extremity outcome index consisting of scores for pain (81 points),
disability (81 points), and activity (45
points)32. The
total score is the average of these three scores, with higher scores
indicating greater impairment of function. The MFA is a general-health-status
outcome index consisting of ten categories, from which a total score can be
calculated33. MFA
values range from 0 to 100, with higher scores indicating a lower level of
overall function. Both the FFI and MFA instruments have been determined to be
valid, reliable, and
consistent32-36.
The FFI has been validated for the evaluation of individuals with systemic
arthritis and measures activity specific to the foot and ankle in individuals
with a low level of function. The MFA has been validated for the evaluation of
trauma patients with greater activity levels but includes the entire
musculoskeletal system.
Statistical analysis was performed with the SAS statistical package (SAS
Institute, Cary, North Carolina). The possible predictive variables included
the type of fracture (open or closed), fracture comminution, Hawkins'
classification, and age (less than forty years or forty years or more). The
clinical outcomes included osteonecrosis, collapse of the talar dome, and
arthritis of the tibiotalar or subtalar joint. Bivariate analysis and Fisher's
exact test were used to test the association between the possible predictive
variables and these clinical outcomes. With use of chi-square analysis, the
timing of fixation was analyzed as a continuous variable (within six hours,
eight hours, twelve hours, twenty-four hours, or more than twenty-four hours)
to test the association between surgical timing and osteonecrosis. The Student
t test was used to identify the associations between the functional outcomes
(as indicated by the FFI and MFA scores) and the possible predictive variables
as well as to identify the associations between functional and clinical
outcomes.
Forty-one patients with forty-two fractures were unavailable for follow-up.
Two patients died of unrelated causes, two patients had severe closed head
injuries and were unable to walk or to communicate, one patient was in jail,
one patient did not speak English, and thirty-five patients could not be
located. Seventeen of the thirty-five patients who could not be located had
been followed for six to twelve months, and none of them had had any clinical
problems or symptomatic radiographic abnormalities at the time of the most
recent follow-up. The remaining fifty-nine patients (sixty fractures) were
evaluated at an average of thirty-six months (range, twelve to seventy-four
months) after surgery. Forty-five of these patients had complete functional
outcome data, and thirty-nine had complete radiographic data. Six of the sixty
fractures were associated with early complications
(Table I). Two patients had
development of a superficial infection, and one had partial medial wound
dehiscence. These three patients were successfully managed with oral
antibiotics and dressing changes. Another patient with wound dehiscence had
development of a deep wound infection that required serial irrigation and
débridement and intravenous administration of antibiotics. After two
years, there had been no recurrence of infection. Two other patients had
development of a deep infection within the first three months postoperatively.
Both of these patients underwent serial irrigation and débridement
procedures, and both were managed with intravenous administration of
antibiotics for several weeks. Neither patient had had a recurrence of
infection at the time of the most recent follow-up.
Three of the sixty fractures did not demonstrate radiographic evidence of
union within the first three months. One of the patients with a deep infection
had development of a nonunion. This patient also had osteonecrosis of the
talar dome, which progressed to collapse. The talar fracture united following
a subtalar arthrodesis and revision of fixation. One other nonunion united
successfully three months after iliac-crest bone-grafting. One patient had a
delayed union and was followed with plain radiographs until fracture union
occurred at twenty-four weeks after surgery.
Radiographic evidence of osteonecrosis was identified in nineteen (49%) of
thirty-nine patients. Osteonecrosis occurred within the first ten months after
the injury (at a mean of nineteen weeks) in all cases. Osteonecrosis was seen
in association with nine (39%) of twenty-three Hawkins' II fractures (five of
which progressed to collapse of the talar dome), nine (64%) of fourteen
Hawkins' III fractures (six of which progressed to collapse), and one Hawkins'
IV fracture (which progressed to collapse). Overall, osteonecrosis with
collapse of the talar dome occurred in twelve (31%) of thirty-nine patients.
Collapse occurred at a mean of thirty-nine weeks (range, twenty-six to
sixty-five weeks) after osteonecrosis was detected on plain radiographs. Seven
(37%) of the nineteen patients with osteonecrosis demonstrated
revascularization of the talar dome on plain radiographs with no evidence of
collapse. Revascularization occurred an average of thirty-five weeks (range,
twenty-five to sixty-five weeks) postoperatively. All of these patients were
followed for at least eighteen months after the return of normal bone density.
None of them demonstrated subsequent evidence of osteonecrosis or
collapse.
The average time from injury to fixation was 3.7 days (range, four hours to
forty-eight days). With the numbers available, chi-square analysis of fixation
within six hours, eight hours, twelve hours, twenty-four hours, or more than
twenty-four hours did not demonstrate a correlation between surgical delay and
the development of osteonecrosis or collapse. The mean time from injury to
fixation was 3.4 days (range, four hours to twenty days) for patients who had
development of osteonecrosis and 5.0 days (range, four hours to forty-eight
days) for those who did not (Figs. 2-A,
2-B, and 2-C). The
mean time between injury and fixation was 12.9 hours for patients who had
development of osteonecrosis and 13.0 hours for those who did not. Time was
also analyzed as a continuous variable and, with the small numbers available,
no correlation was observed between the timing of surgery and the development
of osteonecrosis with or without collapse. This test was performed on the
entire group of patients and was repeated for just those patients who had
received fixation within twenty-four hours.
Osteonecrosis occurred in eighteen of thirty-one patients with comminution
of the talar neck (p < 0.03). Twelve of these eighteen patients had
collapse of the talar dome; a significant association was noted between
comminution of the talar neck and osteonecrosis with collapse (p = 0.02). Nine
of thirteen patients with an open fracture also had development of
osteonecrosis; osteonecrosis occurred significantly more frequently in
patients with open fractures than in patients with closed fractures (p <
0.05). Eight of nine patients who had an open fracture with development of
osteonecrosis had progression to collapse of the talar dome (p < 0.02).
With the numbers available, neither osteonecrosis nor collapse was associated
with the age of the patient, Hawkins' classification, or the presence of an
associated talar body fracture. However, our data showed a trend toward
increased rates of osteonecrosis and collapse in association with greater
initial fracture displacement according to Hawkins' classification.
Twenty-one (54%) of thirty-nine patients had radiographic findings
consistent with posttraumatic arthritis of the ankle joint and/or the subtalar
joint. End-stage arthritis with complete loss of the joint space was
identified in the subtalar joint of six patients and in the ankle joint of
seven patients. Six of twenty-three patients with Hawkins' II fractures and
six of fourteen patients with Hawkins' III fractures had end-stage arthritis
involving one or both of these joints. Trends were observed toward an
association between post-traumatic arthritis and a history of open fracture
(nine of thirteen patients, p = 0.09) and between posttraumatic arthritis and
talar neck comminution (nineteen of thirty-one patients, p < 0.07).
Seventeen patients underwent twenty-five secondary procedures. Eleven
patients had removal of symptomatic hardware. Five patients had an isolated
subtalar arthrodesis because of severe posttraumatic arthritis. One patient
had a Blair tibiotalar arthrodesis twelve months after the injury because of
symptomatic osteonecrosis and collapse. Another patient underwent a Blair
arthrodesis thirteen months after the injury and achieved partial relief of
symptoms. Several months later, she underwent a subtalar arthrodesis that was
complicated by nonunion. She recently had a revision of the subtalar
arthrodesis. Another patient had a subtalar arthrodesis for the relief of
end-stage subtalar arthritis. The ankle arthritis subsequently progressed, and
the patient underwent total ankle arthroplasty, with relief of pain. Another
patient underwent total ankle arthroplasty twenty-one months after the injury.
Subsequently, she reported no ankle or foot pain. Two other patients had a
triple arthrodesis. These procedures were performed twelve and thirteen months
after the injury. Both of these patients had continued pain and limitation
because of tibiotalar arthritis. At the time of the most recent follow-up, at
least three other patients were considering secondary procedures for pain
relief.
Of the forty-five patients who completed functional outcome questionnaires,
thirty-two (71%) had returned to work. Five were employed in jobs involving
heavy construction or industrial work. Six patients who had returned to work
had modified their work duties because of the injury. Two patients had been
unemployed before the talar fracture, and one of them was working at the time
of the most recent follow-up. The other patient was not employed outside of
the home; however, he reported that he did not feel limited by the injury. Two
women who were more than sixty-five years of age had not been employed before
the talar fracture. Both of them reported that they felt severely limited
because of the foot injury. Eleven patients never returned to any form of
employment. One of these patients was not capable of working because of a
severe closed head injury that had been sustained during the same
accident.
Functional outcomes were assessed with the Foot Function Index (FFI) and
Musculoskeletal Function Assessment (MFA) questionnaires
(Table II). The FFI
specifically addresses the foot and ankle, and the MFA is a
general-health-status index. Reference FFI values have been published
previously for patients without foot or ankle
pathology34,35.
Comparison of the mean FFI scores for our patients with these reference values
(pain, 25.3 compared with 11; disability, 34.4 compared with 15; and activity,
20.1 compared with 10) indicated increased impairment after talar neck
fracture. Worse outcomes were noted in association with comminuted fractures
(p < 0.03). With the numbers available, the age of the patient, Hawkins'
classification, and the presence of associated talar body fractures did not
have an effect on the FFI scores.
Similarly, the mean standardized MFA score in our series was 24.6.
Comparison of the mean score for our patients with published reference values
for uninjured patients (9.3), patients with hindfoot injuries (22.1), and
patients with ankle or leg injuries (19.3) suggested a greater level of
disability after talar neck fracture (p <
0.001)36,37.
Fracture comminution adversely affected the MFA score (p = 0.03). However,
with the numbers available, the age of the patient, the Hawkins
classification, a history of open fracture, and extension of the fracture into
the talar body had no discernable effect on the MFA score.
Fractures of the talar neck have been associated with high rates of early
and late
complications2,4-6,12,13,16,20,21,24-27,38.
The high-energy nature of the majority of these injuries produces not only
fracture displacement and comminution but also severe soft-tissue damage,
frequently in association with open wounds. These characteristics often
portend a poor prognosis. Early complications such as skin necrosis, wound
dehiscence, and infection have occurred in as many as 77% of
cases14,16,38.
Reports of late problems such as osteonecrosis, posttraumatic arthritis, and
osteomyelitis with associated stiffness, pain, and loss of function also have
been
common2,5,12,13,16,20,24-27.
Treatment strategies initially evolved from reduction and
immobilization6,24,26,27
to limited, often temporary,
fixation14,16,27,38.
Currently, open reduction and internal fixation is performed for most talar
neck
fractures1,3-5,7,11,12,15,17,18,20,39.
The goals of surgical treatment have been to restore overall alignment of the
talus in order to maximize function of the limb.
Surgical treatment is necessary to restore and to maintain alignment after
a displaced talar neck fracture. Most fractures are associated with some
subluxation or dislocation of the talar body. As direct exposure permits
accurate reduction in these cases, we advocate dual anteromedial and
anterolateral approaches for most talar neck
fractures1,3,7,11,17.
An adjunctive osteotomy of the medial malleolus may be performed to enhance
visualization of the talar
body40. The
anterolateral exposure, along with intraoperative radiographs, greatly aids in
establishing an anatomic reduction. The Canale
view13 is
particularly helpful for assessing deformity. Medial comminution of the talar
neck may otherwise be underestimated, and varus malalignment can
result22,23.
Once an accurate reduction has been achieved, strategic, rigid internal
fixation with plates and/or screws usually will provide adequate stability to
permit early mobilization of the adjacent
joints3, which may
reduce ankle and foot stiffness and may maximize overall
function2,5,7.
The effect of dual surgical approaches on the blood supply of the talus is
unknown. It is possible that this technique could further compromise the
vascularity of fracture fragments or that of the talar body. As 91% of the
patients in the present study had both anteromedial and anterolateral surgical
approaches, we could not assess whether this method has an advantage compared
with other surgical techniques in terms of vascular compromise.
Osteonecrosis occurs frequently after talar neck fractures. It also has
been proposed that surgical treatment may promote revascularization of the
talar
body1-7.
Hawkins developed a classification scheme that was based on the fracture
displacement that is observed on the initial plain
radiographs14. He
also provided treatment recommendations and presented prognostic information
that was based on this classification system. In his study, osteonecrosis was
identified in no Group-I patients, in 42% of Group-II patients, and in 91% of
Group-III patients. Canale and Kelly later modified this classification
scheme, describing a fourth type of talar neck fracture with associated
talonavicular
dislocation13.
Studies in the literature to date have demonstrated the occurrence of
osteonecrosis in association with as many as 13% of Hawkins' I fractures, as
many as 50% of Hawkins' II fractures, as many as 84% of Hawkins' III
fractures, and as many as 100% of Hawkins' IV
fractures2-4,18,25,38.
Our data showed a trend toward an increased rate of osteonecrosis in
association with greater initial fracture displacement as described with
Hawkins' classification. Comminution of the talar neck and open fractures were
associated with an increased risk of osteonecrosis in our study.
Some authors have advocated the use of protected weight-bearing to reduce
the possibility of collapse of the talar dome in patients with
osteonecrosis1,6,13,16,18,24,27.
This recommendation remains controversial, however, and other authors have
recommended the progression of weight-bearing regardless of the presence of
osteonecrosis14,25,38.
We agree that once fracture union has been achieved, weight-bearing as
tolerated may be started. In the present study, 37% of the patients with
radiographic evidence of osteonecrosis showed a return to normal talar dome
density on plain radiographs. No collapse of the talar dome was detected in
these patients despite weight-bearing as tolerated without bracing. Thus, we
believe that the effect of weight-bearing on the progression of osteonecrosis
remains unknown. Additional studies are needed to resolve this issue.
Talar neck fractures often are treated urgently to reduce the risk of
osteonecrosis1,4-7,20
because urgent reduction of dislocations may help to preserve any remaining
blood supply to the posterior portion of the talus. However, to our knowledge,
surgical timing has not been previously shown to impact the development of
osteonecrosis. Although the numbers in the present study were small, no
correlation was found between the timing of reduction and fixation and the
development of osteonecrosis. Because we did not have consistent information
about the duration of dislocation in our patients, we were unable to determine
the impact of prolonged dislocation on the development of osteonecrosis. Other
characteristics of the initial injury and its treatment, such as the presence
of comminution or an open fracture, may have a greater influence on whether or
not osteonecrosis will
occur7,12.
We advocate urgent fracture reduction through either closed or percutaneous
manipulation, and we recommend resorting to open reduction only when the
fracture is not reducible through closed means. Once a reduction of the talar
neck fracture has been achieved, we speculate that a delay in fixation of the
fracture will not affect the development of osteonecrosis. In some cases,
internal fixation cannot be safely undertaken on an urgent basis. This may be
because of life-threatening trauma or because of severe soft-tissue damage and
swelling of the ankle and foot. It has been suggested that a surgical delay
will allow improvement in the associated soft-tissue injury and will decrease
the rates of wound complications and infection after fractures of the tibial
plafond or
calcaneus41-46.
Previous reports have demonstrated skin necrosis, wound dehiscence, and
infection in association with as many as 77% of talar neck
fractures14,16,38.
Perhaps these complications could be minimized by achieving a closed reduction
but delaying definitive fixation in order to allow for improvement in the
soft-tissue injury before proceeding with the dual surgical approach that we
favor.
In conclusion, it has been suggested that early operative intervention
protects the already tenuous blood supply to the posterior portion of the
talus after a fracture of the talar neck. Although the numbers in this series
were small, no correlation was found between the timing of fixation and the
development of osteonecrosis. Osteonecrosis was associated with talar neck
comminution and open fractures, confirming that higher-energy injuries are
associated with more complications and a worse prognosis. This concept is
further strengthened by the poor FFI and MFA scores in patients with
comminuted fractures. We continue to recommend urgent treatment of open
injuries and reduction of dislocations. Proceeding with definitive fixation
when there is minimal soft-tissue swelling will provide rigid fracture
stability to promote fracture-healing and to achieve an earlier return to
weight-bearing function.