A traumatic dislocation of the hip associated with a femoral
head fracture was first described by Birkett1 in 1869. Almost a
century later, a classification of these injuries was proposed by Pipkin11,
whose name has since been associated with this lesion. The treatment
guidelines have evolved on the basis of a relatively limited series of
studies2,3,5,8,9,13,15. Emergency reduction of the hip is imperative,
regardless of the type or extent of the fracture13,16. Once reduction
is accomplished, further evaluation, including computed tomography,
is indicated to assess the congruity and stability of the joint.
Treatment options include nonoperative treatment, excision of fracture fragments,
open reduction and internal fixation, arthroplasty, and arthrodesis.
Complications associated with these injuries are avascular necrosis
of the femoral head, nonunion of the femoral head fragment, and
posttraumatic arthritis2,3,5-9,11-16.
We report on a patient with a Pipkin type-I fracture (fracture
of the femoral head distal to the fovea) treated with open reduction
and excision of fragments. She subsequently had a chronic recurrent posterior
dislocation of the hip. A reconstructive procedure consisting of
a posterior acetabuloplasty in conjunction with an intertrochanteric
internal-rotation osteotomy led to an excellent result.
A twenty-one-year-old, previously healthy woman was involved
in a motor-vehicle accident in which she sustained multiple injuries,
including a fracture of the distal part of the right radius, facial
lacerations, a Pipkin type-I fracture of the left hip with posterior
dislocation (Fig. 1),
and a Lisfranc fracture of the left foot. In our emergency room
at the Academic Medical Center in Amsterdam, an immediate closed
reduction of the fracture-dislocation of the left hip was performed.
Radiographs made after the reduction showed widening of the joint
space with an osteochondral fragment interposed between the femoral head
and the acetabulum (Fig. 1-B). The patient was taken to the
operating room, where a Kocher-Langenbeck approach to the left hip
was performed. This approach was chosen because the fragments appeared
to be small on the radiographs and operative excision of these fragments
was anticipated. Also, since the dislocation had been posterior
it was judged to be less traumatic to approach the hip from the
side where there was already capsular damage. The posterior aspect
of the capsule was indeed found to be torn. The acetabular labrum
was intact. After subluxating the femoral head out of the acetabulum,
two unattached osteochondral fracture fragments of the femoral head,
one centimeter in diameter each, were seen. Since these fragments
originated from the non-weight-bearing aspect of the femoral head distal
to the fovea and reduction and anatomical fixation of the fragments
was not possible because they were small, they were excised. Postoperative rehabilitation
consisted of the use of crutches to walk without bearing weight.
Four weeks after the injury, the patient was evaluated in the
emergency room of our institution after an episode, lasting only
a few seconds, in which she had pain in the left hip with clicking
and an inability to stand. Radiographic evaluation at that time
did not show any changes in the appearance of the reduced left hip
(Fig. 2-A).
She was readmitted to the hospital four weeks later with an infection
of the left foot, where she had had pin fixation of the Lisfranc
injury. During that period of hospitalization, she had pain in the left
hip after flexing it in bed, and radiographs revealed a posteriorly
dislocated left hip (Fig. 2-B). A closed reduction was done in
the ward; however, the hip soon dislocated again in a similar fashion.
Computed tomographic evaluation revealed the large traumatic defect
of the femoral head inferomedially but no loose intra-articular fragments
(Fig. 3-A).
The left hip was subsequently evaluated, with the patient under
general anesthesia, to determine its stability and with the expectation
that a reconstructive procedure would be performed at a later stage.
During that examination, we were not able to subluxate or dislocate
the hip, even with the hip in maximum flexion, adduction, and internal
rotation, and we decided to continue nonoperative treatment. Unfortunately,
in the ensuing three months the patient had five documented posterior dislocations
the left hip, which, according to her, occurred with flexion of
the hip. These dislocations were treated with closed reduction in
the emergency room. She was returned to the operating room for a
repeat examination under general anesthesia twenty-two weeks after
the initial injury. In 90 degrees of flexion and maximum internal rotation,
the left hip dislocated posteriorly. It was observed that flexing
the hip more than 90 degrees caused the inferomedial defect in the
femoral head to lock on the edge of the acetabulum and that subsequently
extending the hip caused the femoral head to lever out posteriorly.
We planned a posterior acetabuloplasty to provide a posterior
buttress to prevent posterior dislocation. After induction of general
anesthesia, the patient was placed in lateral decubitus. The previously
used Kocher-Langenbeck incision was opened. The sciatic nerve was
identified and protected, and the short external rotators of the
hip were released off the femur with protection of the quadratus
femoris muscle and the medial femoral circumflex artery to preserve
the blood supply to the femoral head. After posterior capsulotomy,
the Pipkin defect in the femoral head was easily identified. The
posterior instability was confirmed with the hip in flexion and
internal rotation. A posterior acetabuloplasty was performed. Three
iliac-crest bone grafts, one to 1.5 by three centimeters in size,
from the ipsilateral side were fixed on the posterior aspect of
the acetabulum (Fig. 3-B). The grafts were taken from the
medial side of the crista so that they were unicortical. The host
side was decorticated with a sharp osteotome. The spaces between
the three grafts were filled in with cancellous bone graft from
the iliac wing. Since the grafts were taken from the medial aspect
of the iliac crest, their concavity was comparable with the posterior
facet of the acetabulum. Additional contouring of the grafts was
done with a rongeur. Following acetabuloplasty, there was still
a tendency toward posterior subluxation of the femoral head in maximum
internal rotation and flexion of approximately 90 degrees. The femoral
head articulated with the bone grafts in maximum internal rotation.
We elected to supplement the acetabuloplasty with an intertrochanteric
rotational osteotomy. After estimating that the preferred amount
of rotational correction was 25 degrees and marking this with guiding
Kirschner wires, a 90-degree, four-hole, stepped angled blade-plate
(a fifty-millimeter blade) was introduced into the femoral neck.
After the transverse osteotomy was performed just proximal to the
level of the lesser trochanter, the distal end of the femur was
internally rotated 25 degrees and the osteotomy site was temporarily
held in place with a bone-clamp to judge the adequacy of the correction.
The hip was then taken through a range of motion, with the surgeon realizing
that the amount of internal rotation had decreased due to the osteotomy.
No dislocation or subluxation was possible with any position of
the hip. The plate was fixed to the femoral shaft (Fig. 4-A and Fig. 4-B). The capsule
was closed, after which the wound was closed in layers in a routine
fashion.
Postoperatively, the patient was placed in balanced traction
for two weeks to allow early incorporation of the cancellous graft,
after which she was permitted to walk with touch-down weight-bearing.
No additional subluxation or dislocation occurred. At six weeks
after the operation, there was early healing of the intertrochanteric
osteotomy as well as of the acetabuloplasty site. The amount of
weight-bearing was increased as tolerated. The blade-plate was removed
from the proximal part of the femur at fifteen months.
At the time of the most recent follow-up, four years and five
months after the injury and four years after the corrective procedure,
the patient had no pain in the left hip and had not had any subluxation
or dislocation. Radiographs made at that time showed a concentric
reduction with some early osteophytic changes along the lateral
aspect of the femoral head (Fig. 5-B and fig. 5-B). Although internal rotation of
the left hip was decreased as compared with that of the right hip,
this did not cause any restrictions in daily activities. The left
hip had a 100-degree arc of flexion and extension, 50 degrees of
abduction, 25 degrees of adduction, 25 degrees of external rotation,
and 10 degrees of internal rotation. The right hip had a 130-degree
arc of flexion and extension, 50 degrees of abduction, 25 degrees
of adduction, 50 degrees of external rotation, and 30 degrees of internal
rotation. During walking, the left limb was externally rotated 40
degrees as compared with 25 degrees on the right. The patient was
employed as a caregiver for disabled children and participated in
sports regularly. She had an excellent hip score (18 points) according
to the system of Merle d'Aubigné and Postel10 and a good result
according to the criteria of Epstein et al.6. This discrepancy is
due to the fact that the system of Epstein et al. consists of clinical
and radiographic components, with the poorer rating determining
the final grade. By definition, an excellent radiographic grade
is not possible with any deformity of the femoral head.
Fracture-dislocations of the femoral head, so-called dashboard
injuries, are usually sustained when the knee of a passenger who
is not wearing a seat belt strikes the dashboard with the hip flexed at
approximately 90 degrees and slightly adducted in a motor-vehicle
accident. Dislocations of the hip were classified by Thompson and
Epstein16 and later by Stewart and Milford12, with both classifications
including one type defined by the presence of a femoral head fracture.
In 1957, femoral head fractures were classified by Pipkin11 into
four types, earning him this fracture eponym. A Pipkin type-I fracture
is a femoral head fracture distal to the fossa, a type-II fracture
is proximal to the fossa, a type-III fracture is a type-I or II
fracture associated with a femoral neck fracture, and a type-IV
fracture is a type-I or II fracture associated with an acetabular-rim
fracture. A refinement of the Pipkin classification system by Brumback
et al.2, who incorporated stability of the hip after the femoral
head fracture, has not yet been widely accepted.
Traumatic dislocations of the hip with a femoral head fracture
are uncommon. The infrequency of this injury is reflected in the
literature, where roughly 200 of these fractures, with an adequate description
of the type and more than one year of follow-up, have been reported.
The Pipkin type-II fracture represents the largest group. Although there
is no consensus on the treatment of these fractures, it seems that
types I and II can be treated conservatively if alignment after
reduction of the hip dislocation is anatomical and there are no
loose fragments in the joint. Computed tomographic evaluation is
indicated to confirm the adequacy of the fracture reduction. If
open reduction and internal fixation is chosen, an anterior approach
has been reported to be preferable to the Kocher-Langenbeck approach15.
If internal fixation is not possible due to extensive comminution,
then type-I and II fractures can be treated with excision of the
fragments if they make up less than 30 percent of the femoral head6.
Type-III fractures are most often treated with a hemiarthroplasty
since they are typically associated with a high rate of avascular
necrosis. In young, active patients with a type-III injury, an attempt
at open reduction and internal fixation seems warranted. Type-IV
injuries (associated with acetabular fracture) are treated with
open reduction and internal fixation of the acetabulum and excision
or internal fixation of the femoral head fragment.
In reviewing the literature, we did not identify any case of
recurrent dislocation of the hip after a Pipkin type-I, II, or III
fracture. The best way to assess hip stability after the initial
reduction of fracture-dislocations of the hip remains controversial.
Variables such as the position of the hip at the time of examination,
the amount of force applied axially, the extent of soft-tissue damage,
and the type of anesthesia may influence the likelihood of dislocation
during physical examination. Stability of the hip after fracture-dislocation
with damage to the posterior wall can be estimated with use of the
radiographic data from the study by Calkins et al.4, who showed
that hips with less than 34 percent of the posterior wall intact
(as seen on computed tomography) were unstable and that hips with
more than 55 percent of the posterior wall intact were stable. Hips
with between 34 and 55 percent of the wall intact may or may not
be stable. Interestingly, the recurrent instability in our patient
was caused by the large traumatic defect of the femoral head in association
with a normal posterior wall of the acetabulum. To the best of our
knowledge, no data is available that addresses hip instability after
Pipkin type-I or II fractures. As noted by Brumback et al.2, the
larger the defect of the femoral head, the greater the instability
of the joint. However, in their report, no cases involving recurrent
displacement of the hip after Pipkin type-I or II fractures were
noted. In our patient, this unusual complication of a Pipkin type-I
fracture was successfully treated with a posterior acetabuloplasty
and an intertrochanteric osteotomy of the proximal aspect of the
femur. The combination of these two procedures was chosen since
the acetabuloplasty alone did not provide enough stability. On the
other hand, if only an intertrochanteric rotational osteotomy had
been chosen, more than 25 degrees of internal rotation would have
been needed.