Fractures of the acetabulum are complex, high-energy
injuries with the potential for a poor outcome regardless of the
treatment method1-5. Recent studies
have confirmed the positive association between the accuracy of
reduction and a better long-term result2,6-8.
The value of open reduction and internal fixation in restoring acetabular
bone stock and minimizing pelvic deformity has also been described2,9-14. However, series ranging in
size from sixty to 456 patients have shown that, even when these
goals are achieved, posttraumatic arthritis occurs in up to 30% of
patients6,7,9,15,16. Contributing
factors may include an imperfect reduction, osteochondral defects
in either the acetabulum or the femoral head, chondrolysis due to articular
trauma at the time of injury, and avascular necrosis of the femoral
head or the acetabulum17. Once
symptomatic posttraumatic arthritis has developed, options for salvage
are generally limited to total hip arthroplasty and arthrodesis.
Despite inconsistent evidence suggesting a higher failure rate in
this clinical situation, total hip arthroplasty remains, for many
patients, a favorable alternative to arthrodesis18,19.
Patients undergoing total hip arthroplasty for posttraumatic
arthritis after acetabular fracture reportedly have results inferior
to those of patients undergoing the procedure for nontraumatic arthritis17,20-26. However, a review of the
literature revealed few and contradictory published reports on which
these conclusions are based. Moreover, the literature is especially
deficient with respect to reports on cementless acetabular reconstruction
in patients with posttraumatic arthritis resulting from fracture of
the acetabulum.
The purpose of this study was to compare the clinical and radiographic
results of thirty patients who had undergone primary total hip arthroplasty
for posttraumatic arthritis after acetabular fracture with those
of the same procedure in patients with nontraumatic arthritis. In
addition, the thirty patients were stratified into two groups, allowing
us to compare fifteen patients who had undergone prior open reduction
and internal fixation of the acetabular fracture (open-reduction
group) with fifteen patients who had had prior closed treatment (closed-treatment
group).
In this prospective study, thirty consecutive patients (thirty
hips) with posttraumatic osteoarthritis due to fracture of the acetabulum
were treated between 1984 and 1995 with delayed primary total hip arthroplasty.
All patients received a cementless, hemispheric, fiber-metal-mesh-coated
acetabular reconstruction. During the minimum two-year study
period, no patient died or was lost to follow-up. The average
duration of clinical and radiographic follow-up was sixty-three
months (range, twenty-four to 140 months).
The average age of the sixteen women and fourteen men was fifty-one
years (range, twenty-six to eighty-six years)
at the time of the arthroplasty. The left hip was affected in sixteen
patients and the right hip, in fourteen. All acetabular fractures
had been the result of high-energy trauma. Fifteen of the
thirty patients had had previous open reduction and internal fixation
of the acetabular fracture (open-reduction group), and fifteen had
been treated nonoperatively (closed-treatment group).
The median interval between the time of injury and the total
hip arthroplasty was thirty-seven months (range, eight to 444 months)
and followed a bimodal distribution27,
with peaks at thirteen to twenty-four months and at greater
than 120 months. In order to allow a meaningful comparison of the
open-reduction and closed-treatment groups, a statistical analysis
was performed comparing these groups with respect to various preoperative
variables (Table I).
Except for gender, with a preponderance of men in the open-reduction
group and women in the closed-treatment group, all other variables
were similar between the groups.
In the patients in whom the acetabular fracture had been surgically
treated, the approach used for the total hip arthroplasty was dictated
mainly by the exposure utilized at the time of the open reduction and
internal fixation, with the need to remove hardware or heterotopic
bone, the location of any acetabular defects, and the presence of
soft-tissue contracture taken into consideration. The approach was
posterolateral in nine patients, anterolateral (a modified Hardinge
approach28) in three patients,
and transtrochanteric in three patients in the open-reduction group.
In the closed-treatment group, the approach was anterolateral in nine
patients, posterolateral in four, and transtrochanteric in two.
Neural monitoring was not used.
All acetabular reconstructions were performed with a modular
hemispheric component consisting of a titanium shell covered with
titanium fiber-metal mesh, with multiple optional holes for supplemental
screw fixation; eight Harris-Galante-I, eighteen Harris-Galante-II,
and four Trilogy components (Zimmer, Warsaw, Indiana ) were used.
A reamer was used to prepare the acetabulum, and the component was
inserted with a 0 to 3-mm press-fit. Hardware
was removed only as needed to allow an unimpeded press-fit
of the acetabular component. Bone-grafting was performed as required
to provide two-column support for the acetabular component
and to maintain the integrity of the dome and the medial wall. In
the nine patients requiring bone-grafting of acetabular defects,
morselized cancellous graft was taken from the femoral head and from
the acetabular reamings and was impacted into the contained cavitary
defect both manually and by reverse reaming. No structural grafts
were required. The acetabular component was then initially secured
with two to six titanium (Ti-6Al-4V) screws placed
through the shell. Excellent intraoperative stability of all acetabular
components was achieved. After the shell was fixed, the ultra-high molecular
weight polyethylene insert was fastened into it. In twenty-one
patients, the standard liner was sufficient. In nine patients, a
20° elevated liner was required to provide adequate intraoperative
hip stability.
The femoral reconstruction was done with a cementless titanium-alloy
femoral component (Harris-Galante, Anatomic, MultiLock,
or VerSys [Zimmer]) with a modular head in twenty-three hips
and with a cemented cobalt-alloy component (Harris Precoat [Zimmer])
with a modular head in seven.
All thirty total hip arthroplasties were performed in an operating
room with vertical laminar airflow and with the use of body-exhaust
systems by the operating team. The patients received 1 g of intravenous cefazolin
preoperatively as prophylaxis and 1 g every eight hours postoperatively
until the closed suction drain was removed, usually at forty-eight hours
postoperatively. Alternatively, 1 g of vancomycin, with the dosage
determined according to body weight and renal function, was substituted
for patients who were allergic to penicillin or cephalosporins.
Beginning on the day of the operation, all patients received thromboembolic
prophylaxis with low-dose warfarin to maintain the prothrombin
time at 1.5 times control. Prophylaxis against heterotopic ossification
consisted of a single dose of radiation (500 to 1000 cGy) in seven
patients (four who had had open reduction and internal fixation
and three who had had closed treatment) and indomethacin (50 mg
orally, three times a day for three weeks) in one patient with a
prior open reduction and internal fixation.
Physical therapy was begun on the first or second postoperative
day. Patients with cementless femoral reconstruction were allowed
50% weight-bearing with crutches or a walker for six weeks,
and then they advanced to weight-bearing as tolerated with crutches
for six weeks, at which time a cane was used as needed. Patients
with a cemented femoral component were allowed to bear weight as
tolerated, initially with the assistance of crutches or a walker
for six weeks and then with a cane if necessary.
Each patient was evaluated clinically and radiographically at
six weeks postoperatively; at three, six, and twelve months postoperatively;
and then yearly. Preoperative and yearly follow-up Harris hip
scores29 were determined for each
patient by one of five independent evaluators other than the four
surgeons. At each follow-up visit, an anteroposterior radiograph
of the pelvis, a lateral radiograph of the acetabulum, and both
anteroposterior and lateral radiographs of the hip were made. With
the radiographs made at six weeks postoperatively used as a baseline,
the femoral and acetabular reconstructions were evaluated on subsequent
radiographs by an independent observer.
A digitizing tablet (SigmaScan; Jandel Scientific, San Rafael,
California) was used for quantitative radiographic evaluation of
the acetabulum. The acetabular angle and the vertical distance of
the center of the acetabulum from the interteardrop line were measured
as described by Martell et al.30.
The radiographs were scaled for magnification by the known diameter
of the femoral head. The center of the acetabulum was determined
with use of superimposed templates. A qualitative evaluation of the
acetabulum was also done on anteroposterior pelvic radiographs.
The implant-bone interface was evaluated for the presence
and extent of radiolucent lines according to the modification by
Martell et al. of the DeLee and Charnley technique31, which describes five periacetabular
zones (A1, A2, B1, B2, and C). Radiographs were also analyzed for
the presence of retroacetabular, marginal, or screw-associated
osteolytic lesions. In addition, the screws were evaluated for breakage,
migration, or evidence of adjacent radiolucent lines.
An acetabular component was designated as unstable if either
2 mm of migration or 2° of component rotation was found. A component
was designated as probably unstable if at least four of the five
zones had a radiolucent line and that line was 2 mm wide in at least
one zone. A component was considered to be possibly unstable if
at least four of the five zones had a radiolucent line but the radiolucency
was no greater than 2 mm wide in any zone30.
Quantitative measurements of wear were performed on serial anteroposterior
radiographs of the pelvis with use of a modification of the technique described
by Livermore et al.32. A transparent
template with concentric circles was used, and the center was digitized
with use of the SigmaScan. Wear measurements were then calculated
with use of the six-week radiograph as the baseline.
Heterotopic bone was evaluated on preoperative and postoperative
radiographs according to the method of Brooker et al.33, and acetabular defects were graded
according to the method of D’Antonio et al.34.
Survivorship analysis of the acetabular reconstructions was performed
with the Kaplan-Meier method35,
with revision or radiographic signs of loosening (an unstable or
probably unstable component) used as the end point. All thirty hips
were considered for this analysis. Nonparametric Pearson, Mann-Whitney,
and analysis of variance tests were used to compare the results
of patients who had undergone prior open reduction and internal
fixation of the acetabular fracture with those of patients who had had
nonoperative treatment of the fracture. Significance was set at
p < 0.05. The results for the thirty patients with posttraumatic
arthritis were compared with the intermediate-term results
(at an average of 104 months; range, seventy-eight to 126
months) of 204 primary cementless acetabular reconstructions used
to treat 184 patients with nontraumatic arthritis at our institution36-38.
Clinical Results
Comparison of Groups with Posttraumatic and Nontraumatic
Arthritis
The average operative time in the study group (patients treated
for posttraumatic arthritis) was 179 minutes (range, ninety to 300
minutes) compared with 122 minutes (range, forty-six to
300 minutes) (p < 0.001) for the 204 primary total hip
arthroplasties performed at our institution for nontraumatic arthritis36-38. In addition, the average intraoperative
blood loss of 898 mL (range, 250 to 2900 mL) and the perioperative
transfusion requirement of 2.2 units (range, zero to five units)
in the study group were greater than the average 413-mL
(range, 125 to 1800-mL) blood loss (p < 0.001) and 1.3-unit
(range, zero to four-unit) transfusion requirement (p < 0.001)
for the patients treated for nontraumatic arthritis36-38.
Nine patients (30%) treated for posttraumatic arthritis
required an elevated acetabular liner to achieve intraoperative
hip stability compared with only 4% of the patients treated
for nontraumatic arthritis (p < 0.001)36,38.
Similarly, nine patients (30%) treated for posttraumatic
arthritis required bone-grafting of acetabular defects compared
with 4% of those treated for nontraumatic arthritis (p < 0.001).
The average preoperative Harris hip score29 in
the study group improved from 41 points (range, 19 to 55 points)
to 88 points (range, 47 to 100 points) at the time of follow-up.
Twenty-seven hips (90%) had an excellent or good
result, one hip (3%) had a fair result, and two hips (7%)
had a poor result. One of the poor results was related to the acetabular
reconstruction, with pain secondary to migration and loosening of
the acetabular component.
These results compared favorably with those in the group treated
for nontraumatic arthritis, in which the average Harris hip score
improved by 38 points, from an average of 52 points to an average
of 90 points (p = 0.21), and 83% of the patients
had a good or excellent result (p = 0.10).
During the follow-up period, one of the thirty acetabular
reconstructions was revised because of aseptic loosening. Because
it had migrated, this component was classified as unstable. The
patient had an acetabular nonunion, which was recognized at the
time of the arthroplasty but was inadequately stabilized with the
component used as a hemispheric plate39,
without additional fixation. A cementless acetabular revision was
performed in conjunction with internal fixation of the nonunion
sixteen months following the index total hip arthroplasty, and the revised
component was stable at the time of follow-up, at seventy-seven
months (Figs. 1-A, 1-B, 1-C, and 1-D).
There was one additional reoperation related to the acetabular
reconstruction. This was an open reduction and internal fixation
of an ununited trochanteric osteotomy site, and the metal backing
of the acetabular component was retained. Three additional reoperations
(three débridements of the infected site of a trochanteric
traction pin), all unrelated to the acetabular reconstruction, were
performed in one patient during the follow-up period. No
hip required excision of heterotopic bone.
The rate of reoperations related to the acetabular reconstruction
was 3% in the group treated for nontraumatic arthritis;
no patient in that group required a reoperation for aseptic loosening
of the acetabular component.
Comparison of Open-Reduction and Closed-Treatment
Groups
The average operative time was forty-five minutes longer
in the open-reduction group than in the closed-treatment group (202
compared with 157 minutes; p = 0.01). The patients in the
open-reduction group also had approximately twice as much intraoperative
blood loss (1150 compared with 647 mL; p = 0.008), and
on the average they had transfusion of one more unit perioperatively
(2.7 compared with 1.7 units; p = 0.05) than the patients
in the closed-treatment group.
More patients in the open-reduction group than in the closed-treatment
group required an elevated acetabular liner to achieve intraoperative
hip stability (eight compared with one; p = 0.005); however, significantly
fewer patients in the open-reduction group required bone-grafting
of acetabular defects (two compared with seven; p = 0.04).
With the numbers available, there was no significant difference
in the postoperative Harris hip score, the radiographic stability
of the acetabular component, or the prevalence of postoperative
complications between the two groups with posttraumatic arthritis.
Radiographic Results
Comparison of Groups with Posttraumatic and Nontraumatic
Arthritis
Complete radiographic analysis was performed on the thirty hips
at an average of sixty-three months (range, twenty-four
to 140 months) after the arthroplasty. Ten of the thirty components
had no evidence of periacetabular radiolucency. Seventeen components
had a partial radiolucency that was <1 mm wide. A complete
radiolucency, <1 mm wide, was seen around two reconstructions.
One acetabular component was surrounded by a complete radiolucency
of >2 mm in width. This was the component that migrated
and was revised (Figs. 1-A, 1-B, 1-C, and 1-D). No other radiolucency was >1
mm in width. Thus, the only component classified as unstable according
to the criteria of Martell et al.30 was
the one component, in the open-reduction group, that migrated. Twenty-five
components met their criteria for stable fixation. The other four, including
the two components with a complete radiolucency that was <1
mm wide, were classified as possibly unstable. All four of these
components were associated with a good or excellent Harris hip score
and were therefore considered stable for the statistical analysis.
Acetabular osteolysis without loosening was seen in one patient,
who had received closed treatment of the acetabular fracture. With
the exception of this one component, no component was associated
with evidence of marginal, retroacetabular, or screw-related
osteolysis. There was no screw fracture or complication due to intrapelvic
screw placement.
The 41% prevalence of radiolucent lines in the group
treated for nontraumatic arthritis was lower than the 67% prevalence
in the group treated for posttraumatic arthritis (p < 0.001).
Comparison of Open-Reduction and Closed-Treatment
Groups
In the open-reduction group, seven of the fifteen acetabular
components had no radiolucency, seven had a partial radiolucency
that was <1 mm wide, and the one component that migrated
had a complete radiolucency that was >2 mm wide. In the closed-treatment
group, three of the fifteen acetabular components had no radiolucency,
ten had a partial radiolucency that was <1 mm wide, and
two had a complete radiolucency that was <1 mm wide. There
were no radiolucencies >1 mm wide in the closed-treatment
group.
Heterotopic Ossification
Although heterotopic ossification developed in thirteen patients
(43%) after the total hip arthroplasty, all of the lesions
were either class I (five patients) or class II (eight patients)33, and there were no adverse clinical
effects (Figs. 2-A and 2-B). This prevalence
compares favorably with the 5% rate of class-III and IV
heterotopic ossification noted in our group treated for nontraumatic
arthritis38 (p = 0.35)
and with the 8% rate noted in a previous study of cementless
acetabular revisions at our institution40 (p = 0.25).
Heterotopic bone formed in two of our eight patients who had had
prophylaxis and in eleven of our twenty-two patients who
had not (p = 0.39).
The open-reduction group had more than twice as many patients
with heterotopic ossification following total hip arthroplasty than
the closed-treatment group (60% compared with 27%;
p = 0.06). However, clinical results were not affected,
probably because there were no class-III or IV heterotopic lesions33.
Component Survival
With revision or radiographic loosening as the end point, the
Kaplan-Meier ten-year survival rate was 97% (confidence
interval, 95% to 100%). This rate compares favorably
with the 99% survival rate in the group treated for nontraumatic
arthritis. Also, component stability was comparable between the open-reduction
(93%) and closed-treatment (100%) groups.
The purpose of this prospective study was to evaluate the results
of cementless acetabular reconstruction for the treatment of posttraumatic
arthritis. Patients with prior open reduction and internal fixation
of their acetabular fracture were compared with those who had been
treated nonoperatively for their acetabular fracture.
The results of cementless acetabular reconstruction for the treatment
of posttraumatic arthritis after acetabular fracture were excellent.
The Kaplan-Meier ten-year survival rate was 97%.
In fact, the only failure occurred in a patient with an unsupported acetabular
discontinuity. We now realize that plate fixation is required in
conjunction with acetabular reconstruction in such patients. This
series of patients had postoperative Harris hip scores comparable
with those of patients who had undergone a similar reconstruction
procedure for nontraumatic arthritis36-38.
Component survival, with revision or radiographic loosening as the
end point, was also comparable between the groups treated for posttraumatic
and nontraumatic arthritis. Furthermore, there was no increase in
the prevalence of periacetabular radiolucency or deterioration of
the Harris hip score with increasing duration of follow-up.
With the numbers available, younger age at the time of the arthroplasty
had no effect on the outcome. These results support the findings
of Pritchett and Bortel41, who,
to our knowledge, reported the only other series consisting exclusively
of patients in whom cementless acetabular reconstruction had been
used to treat posttraumatic arthritis following acetabular fracture.
In contrast to the excellent results reported following cementless
acetabular reconstruction after acetabular fracture, the results
of cemented acetabular reconstruction for the treatment of posttraumatic arthritis
have been inconsistent. Boardman and Charnley27,
whose study of sixty-six patients did not include radiographic
assessment, reported that the majority of the clinical results were
good or excellent and concluded that low-friction arthroplasty
is an excellent treatment option in this patient population. However,
Romness and Lewallen24 found less
encouraging results in fifty-three patients with fifty-five
hip replacements; they reported a 79% failure rate in patients
younger than the age of sixty years at the time of the arthroplasty and
a 45% failure rate in patients older than sixty. These
acetabular failure rates were four to five times higher than those
in a comparison group consisting of patients treated for nontraumatic
arthritis25. Poor results and
high complication rates following cemented acetabular reconstruction
have been reported by several authors21-23,26.
We compared the results of thirty primary total hip arthroplasties
in patients with posttraumatic arthritis with those of 204 primary
total hip arthroplasties done with the same technique in patients
with nontraumatic arthritis38.
The prevalence of radiolucency was higher in the patients with posttraumatic
arthritis (67%) than in those with nontraumatic arthritis
(41%), with the former corresponding more closely with
that reported after revision arthroplasty (61%)42. This similarity may be associated
with the fact that acetabular defects, often requiring bone graft, are
seen more commonly in the revision setting than they are in the
primary setting43-46. As is the
case with revision operations43,
the use of bone graft was associated with radiolucency (p = 0.09),
especially in patients who had had closed treatment of the acetabular
fracture. With the exception of the radiolucency around the migrated component,
all radiolucencies were <1 mm wide, did not appear radiographically
to compromise current stability, and are of uncertain long-term
importance. Moreover, the radiolucencies did not appear to progress
over time.
Operative time, blood loss, and transfusion requirements were
substantially greater in this series than they were in the group
treated for nontraumatic arthritis, an observation that attests
to the increased complexity of arthroplasty after acetabular fracture. This
was especially true for patients who had had open reduction and
internal fixation of the acetabular fracture, possibly because of
the increase in the difficulty of the exposure caused by postoperative scarring
and the extra time and tissue dissection required for hardware removal.
Of the five complications in our series, the component migration
secondary to acetabular nonunion and the wound infections at the
site of a trochanteric traction pin are particular to the posttraumatic setting,
contributing to a complication rate (17%) that was higher
than that in the group with nontraumatic arthritis37,38 and similar to that in patients
treated with revision40,42,43.
This study cannot resolve issues pertaining to acute fracture
treatment. Although prior open reduction and internal fixation did
appear to make reconstruction more complex in our series and seemed
to render patients more susceptible to intraoperative hip instability,
clinical results were not compromised. The lower prevalence of periacetabular
radiolucency after open reduction and internal fixation (as compared
with that after closed treatment), which approached significance
(p = 0.12), and the decreased need for bone-grafting (p = 0.04)
suggest superior reconstitution of periacetabular bone, which may
positively influence long-term component survival. In summary,
with attention to detail and appropriate anticipation of these unique
technical considerations, at an average of five years postoperatively
it appears that modern techniques of cementless acetabular reconstruction
were as effective in the operative treatment of posttraumatic arthritis
after acetabular fracture as they were in the operative treatment
of nontraumatic arthritis.