Dealing with limb-length discrepancy is an important part of the
surgical treatment of unilateral type-2 (low) and type-3 (high)
dislocation. In comparison with the usual population of patients
who are managed with total hip replacement, these patients are younger,
tend not to use external walking aids or shoe-lifts, and are predominantly female.
In the experience of the senior author (A. G.), these patients consider
limb-length discrepancy to be one of the reasons for having the
surgery, and this aspect of the overall problem should not be minimized
by the surgeon. The limb is lengthened by lowering the acetabulum
back to or nearer to its anatomical position or by inserting a femoral component
that is longer than the length of femoral bone that is removed.
The limb can be lengthened at the time of surgery by as much as
four centimeters3. The sciatic
nerve must be carefully monitored. In all cases in which the limb
is to be lengthened by more than two centimeters, we identify the
nerve in order to check its tension after reduction of the trial
components. Excessive dissection of the nerve should be avoided because
of the danger of devascularization, but the nerve must be visualized
and palpated. The trial reduction should be carried out with the
knee flexed, and the tension in the nerve is then evaluated as the
knee is gradually extended. Experience in evaluating nerve tension
is invaluable, and a wake-up test is, at least early in one's experience,
safer and more reassuring. The senior author performs a wake-up
test only when there is some question about the degree of tension
in the nerve. If the surgeon anticipates the performance of a wake-up test,
both the anesthetist and the patient must be informed prior to the
induction of anesthesia. During the wake-up test, the patient is
instructed to dorsiflex his or her toes on command (dorsiflexion
is considered to be more sensitive than plantar flexion). If the patient
has been previously instructed to do this, he or she responds with
less lightening of the anesthetic. Another method for the evaluation
of the sciatic nerve is monitoring of somatosensory evoked potentials4. This method requires additional resources,
may be too sensitive, and should be reserved for centers at which
a high volume of complex hip operations are performed.
In our experience, the femoral nerve also may be damaged by excessive
lengthening. This nerve is not monitored intraoperatively, but it
should be evaluated postoperatively along with the sciatic nerve.
If there is a femoral nerve palsy, the patient should be positioned
in bed with the hip flexed to about 70 degrees in order to relax
the femoral nerve. If both the femoral nerve and the sciatic nerve
are damaged by stretch, then the patient should be positioned with
the hip flexed to relax the femoral nerve and the knee flexed to
relax the sciatic nerve.
If there is excessive sciatic nerve tension when a trial reduction
is done, the limb has been lengthened too much. It must be shortened
by reconstruction of the acetabulum at a higher level or by shortening
of the femur. Femoral shortening can be carried out by means of
a subtrochanteric osteotomy (Fig. 4Fig. 4)5,6 or
by resection of more bone from the proximal part of the femur.
A subtrochanteric shortening osteotomy is associated with the potential
problem of nonunion. A long stem or a modular stem may be necessary
to provide stable fixation of the osteotomy site. A step-cut, oblique,
or chevron osteotomy can be used to obtain rotational stability,
or cortical struts or a plate can be used to stabilize the osteotomy
site. Cortical struts or a plate are particularly useful when an
uncemented femoral component is used5,6.
We prefer to use cortical struts or a plate to stabilize the site
of a transverse osteotomy because this allows for easier adjustment of
the version. A step-cut, oblique, or chevron osteotomy is technically
difficult and requires adjusting the version while allowing for
the configuration of the osteotomy. Shortening by resection of more
proximal femoral bone avoids the problems associated with osteotomy
but involves the resection of metaphyseal bone, which is needed for
stabilization of the implant and osseous ingrowth.
Special components should be available during the operative treatment
of a dysplastic hip. The cup should have an inner diameter of twenty-two
or twenty-six millimeters and can be cemented or uncemented. Cups
with an outer diameter of as small as thirty-six millimeters should
be available; however, the larger the diameter of the cup, the thicker
the polyethylene, which is an advantage in these young patients.
The surgeon must achieve a balance between cup diameter and cup
coverage. The femoral component must be small and should have a
straight stem in order to accommodate the anteversion. Stems with
a distal diameter of between five and ten millimeters, or modular stems,
should be available.
Prior to surgery, an accurate clinical and radiographic measurement
of leg length is important. If the patient has a fixed pelvic obliquity,
then the apparent leg length (as measured from the umbilicus to
the medial malleolus) should be used to determine the amount of
lengthening that is necessary. The surgeon should decide on the
level of the cup and template it at that level on the radiograph.
The surgeon should note how much of the cup is uncovered and anticipate whether
or not a bone graft is going to be necessary. The final decision,
however, will depend on the intraoperative findings. Templates can
be used to determine the level of the neck cut, but the cut usually
is made at the level of the lesser trochanter because of the problem
of excessive anteversion if the cut is made more proximally.
Surgical Approach
Type-1 hips (dysplasia) can be approached through a conventional
posterior13 or lateral14 approach without disturbing the greater
trochanter, unless the trochanter is riding high due to previous
avascular necrosis of the femoral head. Under those circumstances,
the surgeon may elect to perform a trochanteric osteotomy in order
to advance the greater trochanter distally. Type-2 hips (low dislocation) and
type-3 hips (high dislocation) may require a more elaborate approach
in order to obtain extensive pelvic exposure and to advance the
greater trochanter if indicated. A trochanteric osteotomy provides
excellent pelvic exposure. This allows the surgeon to identify the
false and true acetabula and, if necessary, to reconstruct the acetabulum with
a bone graft. A trochanteric osteotomy also allows for lengthening
of the extremity. A transverse osteotomy in which the vastus lateralis
is dissected off the greater trochanter offers the best pelvic exposure
but carries the risk of nonunion and, even more importantly, of
trochanteric migration15. A trochanteric
slide osteotomy retains the attachment of the vastus lateralis,
which protects against trochanteric migration (Fig. 8Fig. 8)15,16. A trochanteric slide is our
preferred approach if a grafting procedure or lengthening of more than
three centimeters is anticipated. If the leg is lengthened, it is
necessary to keep the trochanteric fragment long in order to ensure
bone-to-bone apposition for healing.
A subtrochanteric osteotomy that is performed for femoral shortening
and derotation can also be used for exposure, allowing preservation
of the greater trochanter while still providing good pelvic exposure
for bone-grafting (Fig. 4Fig. 4)5,6.
If a trochanteric osteotomy has already been performed, special care
is needed to preserve the blood supply and muscular attachments
to the femur proximal to the subtrochanteric osteotomy.
The sciatic nerve is identified after the approach has been completed,
the hip has been dislocated, and the level of the true acetabulum
has been identified. The nerve is most easily found distal to the
true acetabulum in the fat deep to the gluteus maximus, just proximal
to the tendon. The nerve is not dissected out because of the danger
of devascularization, but it is visualized and not just palpated.
If the hip is completely dislocated, identification of the nerve
may be difficult because of the redundancy of the false capsule.
Under these circumstances, it may be easier to find the nerve when
the limb has been brought more out to length when the trial components are
in place and the hip is reduced. If this technique is used, it is
important to keep the knee flexed so that undue tension is not created
in the nerve.
Acetabulum
It is important to identify the desired level for placement of
the cup. The level of the true acetabulum can be identified with
use of the obturator foramen as a landmark. If there is any doubt
at all, a radiograph should be made before reaming is started. When reaming
is started, the fovea can often be identified. After the level for
placement of the cup has been identified, reaming should be started
with use of very small reamers, usually thirty-six millimeters in
diameter. A drill-hole is made through the medial part of the acetabulum,
and a depth-gauge is used to decide how deep to ream. The senior
author stops reaming one-half to one centimeter from the inner cortex.
A trial cup is then inserted, and, if less than 70 percent of the
cup is covered, bone-grafting is performed. The femoral head is
used for the bone graft. The cartilage is reamed off, but the subchondral
bone is kept intact. The graft is placed at the superior edge of
the acetabulum or just inside it and is fixed with two cancellous
screws, usually 4.5 millimeters in diameter. The screws are placed
in an oblique-to-vertical direction. The cancellous surface of the
graft should not be exposed to host soft tissue because this surface
is more easily resorbed than the subchondral bone is. The cancellous
surface should abut the cup. In addition, we place a flying-buttress
graft consisting of cancellous, morselized, autogenous bone between
the top of the shelf graft and the ilium (Fig. 9Fig. 9). The
ilium at the junction and just proximal to the shelf graft is roughened
or perforated with multiple small drill-holes to encourage union
and remodeling.
Femur
On the femoral side, the predetermined neck cut is usually made
at the level of the lesser trochanter. This is necessary because
cutting the neck short decreases the degree of neck anteversion,
allowing for easier insertion and fitting of the femoral component.
Small reamers and broaches should be used, and the femoral component should
have a straight and narrow stem.
Sciatic Nerve
If the limb is to be lengthened by more than two centimeters,
the sciatic nerve is identified. Reduction of the hip with the trial
components in place is carried out with the knee flexed. The knee
is then gradually extended while the tension in the nerve is monitored.
If the surgeon is still not sure about whether the nerve is being stretched
too much, then a wake-up test must be performed. If the hip is too
tight for reduction of the trial components, if the nerve is too
tense, or if a wake-up test reveals sciatic nerve weakness, then
the femur must be shortened. Proximal shortening of the femur or
subtrochanteric osteotomy may be necessary5,6.
It is our practice to shorten the femur proximally in order to avoid
the technical difficulty and potential complications (such as nonunion)
associated with a subtrochanteric osteotomy. Proximal shortening,
however, should not be performed distal to the lesser trochanter
because there will not be enough of a metaphyseal flare left to
support the femoral implant. In addition, if a trochanteric approach
has been utilized, trochanteric reattachment may be difficult. Under
these circumstances, a subtrochanteric osteotomy should be utilized.
We reviewed the results of sixty-seven total hip replacements
that were performed with use of a shelf autograft in fifty-eight patients
(forty-eight women and ten men). The average age of the patients
at the time of the operation was fifty-two years (range, thirty
to sixty-five years), and the average duration of follow-up was
ten years (range, five to seventeen years). Thirty-nine hips were
classified as type 1 (dysplasia); seventeen, as type 2 (low dislocation);
and eleven, as type 3 (high dislocation). The acetabular components
included forty-two metal cups that were inserted without cement,
twenty-two all-polyethylene cups that were inserted with cement,
two all-polyethylene cups that were inserted without cement, and
one bipolar prosthesis. On the femoral side, fifty-one stems were inserted
without cement and sixteen were inserted with cement. A transgluteal
approach was used in forty-one hips; a transtrochanteric approach
(through a transverse trochanteric osteotomy), in twenty-two; and
a Smith-Petersen approach, in four. Fifty-four hips were reconstructed
within one centimeter of the true hip center, and thirteen were
reconstructed more than one centimeter higher than the true hip
center.
Four patients had a total of five incomplete nerve palsies (two sciatic
and three femoral), all of which resolved completely. Four patients
had a hip dislocation, and two of them required revision of the
cup. Four patients had a trochanteric nonunion and escape, and one
had a fibrous union.
Eleven (16 percent) of the sixty-seven cups were revised (ten hips)
or were awaiting revision (one hip), and two others were radiographically
loose but were asymptomatic. Three of the ten acetabular revisions
were performed because of nonunion at the site of the graft and
seven, because of loosening of the cup. Three of the ten acetabula
that were revised required no further grafting, four required morselized
graft, and three required another structural graft. Four (6 percent)
of the sixty-seven femoral components were revised, and two others
were radiographically loose but were asymptomatic. Overall, thirteen
(19 percent) of the sixty-seven hips were revised (twelve hips)
or were awaiting revision (one hip). Two hips had revision of both components.
Kaplan-Meier analysis revealed that the probability of survival
at fourteen years was 78 percent for the cups and 85 percent for
the stems. The average duration of survival was 14.2 years (95 percent
confidence interval, 12.6 to 15.9 years) for the cups and 15.3 years
(95 percent confidence interval, 14.0 to 16.0 years) for the stems.
Cox regression analysis was used to assess the influence of covariates
(gender, age, type of prosthesis, degree of dysplasia, and vertical
and horizontal placement of the hip center) on survivorship. Only
the degree of dysplasia and the vertical displacement from the so-called true
anatomical center had a significant effect on survivorship (r =
0.29, p = 0.0095). Survivorship was not affected by age, gender, or
the use of cement. Seven (18 percent) of the thirty-nine type-1 hips
failed; the average duration of survival in this group was fifteen
years (95 percent confidence interval, 13.2 to 16.8 years). Three
of the seventeen type-2 hips failed; the average duration of survival
in this group was 11.5 years (95 percent confidence interval, 9.4
to 13.5 years). Three of the eleven type-3 hips failed; the average
duration of survival in this group was 10.1 years (95 percent confidence interval,
7.3 to 12.9 years).
Three hips (4 percent) had a nonunion at the site of the graft.
All three required revision of the cup. One of these hips required another
structural graft because of severe graft resorption, and the other
two healed with screw fixation and placement of morselized autograft
at the site of the nonunion. Seven hips (10 percent) had resorption
of more than one-third of the graft; six of these hips had moderate
resorption (resorption of one-third to one-half of the graft), and
one had severe resorption (resorption of more than one-half of the graft).
Of the six hips that had moderate resorption, three had revision
of the cup with no need for further structural grafting, one was
awaiting revision of a loose cup, one was revised because of a loose
femoral component, and one was asymptomatic with no loosening of
the cup. The hip with severe resorption required revision of the
cup with insertion of a new structural graft, as mentioned above.
One other hip required insertion of another structural graft following displacement
of the original graft as a result of a dislocation in the early
postoperative period.
Total hip replacement is more difficult for dislocated (type-2 and
type-3) hips than it is for subluxated (type-1) hips. Exposure of
the false and true acetabula and identification of the sciatic nerve
requires experience and a good surgical approach. In patients with
unilateral involvement, the limb-length discrepancy is more severe.
Placement of the cup at the correct anatomical level is difficult
and may require femoral shortening. This can be accomplished by
resecting more femoral bone proximally or by performing a subtrochanteric
osteotomy. Coverage of the cup must be achieved to prevent early
loosening, and there are various surgical methods that can be used
to accomplish this.
Coverage of the cup can be achieved by placing a small cup in
a superior7,8 or protruded1,9 position or by using an oblong socket12. These techniques are advantageous
in that they allow the surgeon to avoid the use of a structural
autograft, which is associated with certain disadvantages, including
technical difficulty, resorption, fragmentation, and loosening of
the cup7,8.
Advocates of the use of a high hip center have recommended placement
of a small socket in a high, but not a lateral, position7,8. The advantages of this technique
are the avoidance of a structural graft and the ability to place
a cementless cup against bleeding host bone. The disadvantages are
that the use of a small cup (and therefore thin polyethylene) may
be necessary and that there is an increased risk of femoral and
acetabular component loosening10,11.
In addition, this technique does not restore bone stock for subsequent
revision surgery. Despite these disadvantages, however, the use
of a high hip center is an attractive option if limb length can
be restored with use of a long-neck prosthesis, there is adequate
bone stock, and the hip is stable. The authors who have advocated
this technique have done so on the basis of their own poor results
with femoral head autografts7,8.
Anderson and Harris recently reported on eighteen patients (twenty
hips) who had a primary total hip replacement with use of a cementless
cup17. All of the hips had either
low (type-2) or high (type-3) dislocation. After an average duration
of follow-up of eighty-three months, none of the cups had been revised
and none were loose. Although most of the cups were placed in a
relatively high position, it was necessary to use a bulk femoral
head autograft in four hips. All four grafts united, and none resorbed.
In the twenty hips that formed the basis of the study, the center
of the hip was placed an average of twenty-eight millimeters proximal
to the interteardrop line. In nine of the hips, the center of the
hip was placed at least thirty-five millimeters proximal to the
interteardrop line; these hips were arbitrarily designated by the authors
as having a high hip center. The authors stressed that in these
nine hips, the lateral placement of the cup was normal.
Centralization of the cup by means of cotyloplasty1 or reaming to the floor of the true
acetabular fossa9 also avoids
the use of a structural autograft. We have had no personal experience
with cotyloplasty, but this technique has yielded good results and
reestablishes the hip center1.
This technique, which involves the creation of a controlled fracture
of the medial wall of the acetabulum, is complicated and should
only be undertaken by surgeons who have been specifically trained
in this method. Long-term problems of protrusio and cup-loosening
must be monitored. Hartofilakidis et al. reported on a series of
eighty-four hips (sixty-seven patients) that were treated with total
hip replacement because of a high dislocation18.
In forty-nine of the hips, a cemented cup was inserted at the correct
anatomical level with use of the cotyloplasty technique. Those authors
reported an overall success rate of 87 percent after an average
duration of follow-up of 6.4 years, with a cumulative rate of survival
of the cup of 95 percent at five years and 90 percent at ten years.
In 1988, McQueary and Johnston advocated centralization by reaming
to the inner cortex and use of cement to provide coverage at the
correct anatomical level9, but
a subsequent review of these patients revealed an unacceptable prevalence
of cup loosening19. In their 1988
study, the authors reported on a series of sixty-one hips in forty-eight
patients who had insertion of the cup at the correct anatomical
level with use of cement and without bone-grafting. After an average
duration of follow-up of 8.5 years, no revisions had been performed
for aseptic loosening but six hips (10 percent) had radiographic
evidence of loosening9. When this
same group of patients was evaluated after an average duration of
follow-up of sixteen years, survivorship analysis revealed that
the rate of survival at fifteen years was 85 percent with revision
as the end point but 68 percent with radiographic loosening as the
end point19. The authors of the
later study expressed concern about the prevalence of acetabular
loosening at fifteen years. Type-2 and type-3 hips (low and high
dislocation) had a higher prevalence of loosening than did type-1
hips (dysplasia)19.
Dorr et al. reported on twenty-four hips that were treated with
a medial protrusio technique and insertion of an uncemented cup at
the correct anatomical level20.
After an average duration of follow-up of seven years, no cup had
been revised. The authors advocated reaming medially until coverage
of 80 percent of the cup is achieved. Hips with complete dislocation
required more medialization than those with subluxation.
Numair et al. presented a series of 182 Charnley total hip replacements
in 141 patients with dysplasia21.
A cemented cup was placed in the true acetabulum without a bone
graft. After an average duration of follow-up of ten years, the
rate of acetabular revision was higher in the group of hips that
had been reconstructed because of complete dislocation (15 percent;
seven of forty-six) than it was in the group of hips that had been
reconstructed because of subluxation (9 percent; twelve of 136).
In addition, the rate of definite or possible loosening of the cup
was higher in the group of hips with dislocation (11 percent; five
of forty-six) than it was in the group of hips with subluxation
(4 percent; five of 136). On the basis of these results, the authors
concluded that contemporary grafting techniques should be considered
when total hip replacement is performed for the treatment of a completely
dislocated hip21.
Use of an oblong cup allows the cup to be placed at the anatomical
level. The oblong elongated part of the cup fills in the deficient
segment and allows further stabilization with screws. This cup has
been used mainly for revision arthroplasty of the acetabulum12. It has been advocated for primary
replacement in hips with low and high dislocation12,
but we are not aware of any reports on the long-term results of
such treatment. The primary disadvantage associated with the use
of this device for the treatment of hips with dysplasia is the failure
to restore bone stock. Moreover, reaming superior to the true acetabulum
in order to seat the elongated part of the cup destroys what valuable
bone stock there is12.
Our results with the use of structural femoral autografts to
provide coverage of the cup have been very encouraging. We advocate
the technique outlined above because it restores bone stock for
future revision surgery, restores the hip center, and restores limb
length in an anatomical way22.
A cemented or uncemented cup can be used. In our series, only three
of the ten hips that had revision of the cup required another structural
graft, confirming that this technique restores bone stock for revision
surgery. Good long-term results with such structural autografts
also have been reported by other authors23-25.
Garvin et al. reported on twenty-three hips that were followed for
an average of fourteen years23.
Six hips had required structural bone-grafting to support the cup, and
all of these reconstructions were still intact with no evidence of
loosening of the cup at the time of the latest follow-up. There
were only two acetabular revisions in the entire series: one was
performed in a type-2 hip (low dislocation), and the other was performed
in a type-3 hip (high dislocation). Martí et al. reviewed the results
of sixty-three total hip replacements after an average duration
of follow-up of ten years24. All
hips required structural grafting of the acetabulum because of congenital
dysplasia (fifty-three hips) or traumatic dysplasia (ten hips).
The authors advocated the use of two or three small structural grafts
fixed with screws and reported no extensive resorption of the grafts
at ten years. Two cups were definitely loose, and two cups were
probably loose. The authors stated that the hips with complete dislocation
had poorer clinical scores than those with dysplasia but noted that
there was no radiographic difference between these groups. Rodriguez
et al. reported on twenty-nine hips in twenty-three patients who
required structural grafting during total hip replacement for dysplasia25. After an average duration of follow-up
of eleven years, three cups (10 percent) had been revised for loosening
and eight cups (28 percent) were probably loose. Graft resorption
was not a problem.
The recent article by Bal et al. also confirms that healed bulk femoral
head grafts can provide bone stock for revision arthroplasty with
use of an uncemented cup26. In
that study, nine patients (nine hips) with congenital displacement
who had had a previous total hip replacement with use of a cemented
cup and a shelf bulk autograft (six hips) or allograft (three hips)
had a revision arthroplasty with use of an uncemented cup. After
an average duration of follow-up of seventy-six months, all nine
acetabular shells were functioning well. The authors concluded that
the previously placed bulk autograft or allograft provided bone
stock for revision with an uncemented cup at the correct anatomical level.
Replacing the dislocated hip is technically more challenging than
replacing the subluxated hip. Overall, clinical and radiograp7hic
results have not been as good for hips that are completely dislocated22,24,26. The surgical approach must allow
for identification of the false and true acetabula, identification
of the sciatic nerve, and lengthening of the leg.
In patients with a dislocated hip who are managed with total
hip replacement, coverage of the cup can be achieved by medialization,
creation of a high hip center, or use of a structural graft. Bone-grafting
allows the cup to be placed in an anatomical position, provides
bone stock for additional surgery, and restores leg length. Our
results and the results of other authors confirm that these grafts
remain intact for at least ten years and restore bone stock for
additional surgery. This is particularly important in this relatively
young population.