Between 1990 and 2000, several surgeons performed 7121 primary total hip
arthroplasties at our institution. We retrospectively identified twenty-one
patients (twenty-one hips) who had sustained an intraoperative fracture of the
acetabulum during primary total hip arthroplasty. Patients were located with
use of the complications log of our Total Joint Registry, and institutional
review board approval and consent were obtained for this study. The study
group included eight male and thirteen female patients who had had a mean age
of sixty years (range, fifteen to seventy-nine years) at the time of surgery.
The preoperative diagnoses included osteoarthritis (sixteen patients),
osteonecrosis (two), rheumatoid arthritis (one), posttraumatic arthritis
(one), and femoral neck nonunion (one). Seventeen total hip arthroplasties
were performed with the so-called hybrid technique with use of an uncemented
acetabular component and a cemented femoral component, and the other four
arthroplasties were performed without cement.
We categorized cup designs into three general types: elliptical monoblock
(Implex; Zimmer, Warsaw, Indiana), elliptical modular (Reflection [Smith and
Nephew, Memphis, Tennessee], Peripheral Self-Locking [PSL; Stryker Howmedica
Osteonics, Mahwah, New Jersey]), and hemispherical modular (Trilogy [Zimmer]
and Harris-Galante II [HG-II; Zimmer]). Elliptical shells have a peripheral
flare (an increase in diameter) at the periphery to theoretically improve rim
press-fit (Fig. 1). These three
cup designs accounted for 4953 (92.4%) of the 5359 uncemented cups that were
implanted during the study period. The remaining 406 uncemented cups were of
various designs that were implanted infrequently over a ten-year period and
then were excluded from the analysis. No intraoperative fractures occurred in
these 406 excluded cases.
The acetabular components in the hips that sustained a fracture included
twelve Implex components, one Trilogy component, seven PSL components, and one
HG-II component.
The femoral components in these hips included nine Implex cemented
components (Zimmer), four Centralign cemented components (Zimmer), three
Implex hydroxyapatitecoated uncemented components (Zimmer), four Omnifit/ODC
cemented components (Stryker Howmedica Osteonics), and one Omnifit
hydroxyapatite-coated uncemented component (Stryker Howmedica Osteonics).
Eighteen arthroplasties were performed through an anterolateral approach,
and three were performed through a posterolateral approach. In all hips, the
difference between the diameter of the last reamer used to prepare the
acetabular bed and the true largest external diameter of the implanted shell
was =2 mm.
The location and cause of the intraoperative fracture were documented in
the operative notes, and the treatment of the fracture (e.g., a change of the
acetabular component, the addition of supplemental fixation screws, or the use
of bone-graft) was recorded.
One patient died in the postoperative period, and one was lost to
follow-up. The remaining nineteen patients were followed clinically for a
minimum of two years (mean, forty-four months; range, two to ten years).
Seventeen of the nineteen patients were followed radiographically for a
minimum of two years (mean, four years; range, two to ten years).
Clinical and radiographic data from the time of the most recent follow-up
were retrospectively reviewed to evaluate the status of cup fixation and
healing of the fracture. Clinical function was evaluated by assessing pain,
walking ability, and the need for walking aids. Radiographs were evaluated
with regard to the fixation of cemented femoral components according to the
criteria of Harris et
al.10, with regard
to the fixation of uncemented femoral components according to the criteria of
Engh et al.11, and
with regard to the fixation of cemented acetabular components according to the
criteria of Hodgkinson et
al.12. Loosening of
an uncemented acetabular component was defined as implant migration, a
complete radiolucent line at the implant-bone interface, or fixation screw
breakage. Heterotopic bone, if present, was graded according to the system of
Brooker et
al.13.
Twenty-one patients (twenty-one hips; 0.3%) sustained a fracture of the
acetabulum during primary total hip arthroplasty. Of the 7121 acetabular
components, 1762 were cemented and 5359 were uncemented. No fractures occurred
in patients undergoing acetabular component insertion with cement. The
prevalence of fracture in patients undergoing acetabular component fixation
without cement was 0.4%.
In sixteen (76%) of the twenty-one hips, the fracture was noted during
impaction of the real acetabular component; in three, it was noted during
reaming; and in two, it was noted during initial hip dislocation. The location
of the fracture was posterosuperior in twelve hips, directly posterior
(involving the posterior wall) in six, medial in two, and anterosuperior in
one. In seventeen hips, the acetabular component was judged to be stable
despite the fracture and no additional treatment was performed. Stability of
the component was typically judged qualitatively according to the surgeon's
estimation of the quality of the "press-fit" and the absence of
motion when pressing a probe on the periphery of the implanted cup. In all
instances, cancellous bone graft from reaming was placed along or into the
fracture line.
In four hips, the original acetabular component was unstable in the
presence of the fracture and was therefore changed intraoperatively to a
design that allowed for supplemental screw fixation. No acetabular fixation
required the use of a supplemental plate, lag screws, or an antiprotrusio
device. Patients were kept partial weight-bearing for six weeks, and then
weight-bearing was advanced as tolerated. No braces were used.
There were two other complications. One patient sustained an intraoperative
fracture of the proximal part of the femur and was managed with conversion to
a cemented stem. Another patient died postoperatively as a result of
pneumonia. There were cases of no clinically important heterotopic
ossification. There were no dislocations or thromboembolic complications.
At the time of the most recent follow-up, all fractures had united and all
cups had achieved radiographic evidence of osseous ingrowth without evidence
of component migration. One patient required revision of a well fixed
acetabular component two years postoperatively. This revision was performed at
the same time as revision of a debonded Centralign stem (Zimmer) because of
aseptic loosening. The cup was revised because of malpositioning (excessive
retroversion) and psoas tendon irritation. The fracture was found to be healed
at the time of revision.
Of the other eighteen patients, fifteen had no pain and three had mild
trochanteric discomfort. Only two required walking aids; in both cases,
walking aids were required because of chronic spine problems and not because
of problems related to the hip.
The difference in the fracture rate between all elliptical shells (nineteen
of 2755) and all true hemispherical shells (two of 2198) was highly
significant (p = 0.001) (Table
I). Also, elliptical monoblock designs had a significantly higher
fracture rate (twelve of 339) than did either elliptical modular shells (seven
of 2416) (p < 0.0001) or hemispherical modular shells (two of 2198) (p <
0.0001). However, there was no significant difference in the fracture rate
between elliptical modular shells (seven of 2416) and hemispherical modular
shells (two of 2198) (p = 0.18). The fracture rates associated with specific
implants are summarized in Table
II.
Intraoperative fracture of the acetabulum is a very rare complication of
total hip arthroplasty, and therefore it has been difficult to accumulate a
cohort large enough to study this problem. We noted a significant (p = 0.001)
increase in the rate of intraoperative fracture in association with designs
involving a peripheral elliptical flare that was inserted into a bed that had
been prepared with a hemispherical reamer. The hoop stresses generated during
cup impaction can be substantial; thus, we do not recommend excessive
underreaming of the
acetabulum1-9,11,14-18.
It was, and is, typical for surgeons at our institution to underream by 1 to 2
mm when using hemispherical designs and to ream line-to-line when using
elliptical designs. There is some variability in the amount of the peripheral
flare among various types of elliptical shells, and this amount may increase
with increasing shell diameter. To our knowledge, during the study period, no
manufacturer offered a built-in rim flare that was >2 mm greater than the
designated shell size.
Elliptical monoblock designs were associated with the highest fracture
rates. Monoblock designs have the theoretical (but unsubstantiated) advantage
of avoiding so-called "backside wear" caused by polyethylene,
which can migrate to the periacetabular bone through screw-holes. The
disadvantage of the monoblock design is that it does not allow the surgeon to
visually confirm complete cup seating through either a screw-hole or an apical
("polar") hole. Most fractures in the current series occurred
during the insertion of a specific type of porous tantalum elliptical
monoblock cup (Implex). Porous tantalum has a much higher coefficient of
friction than a beaded or fiber-mesh surface does, and, in comparison,
acetabular components that are made from this material can require
substantially more force to seat. We recommend that surgeons using monoblock
elliptical designs have modular components available to allow for easy
exchange and screw fixation should an intraoperative fracture occur. Since the
amount of flare typically varies according to cup size, additional reaming may
be indicated for patients with extremely sclerotic bone to avoid fracture.
Surgeons using elliptical designs should consider measuring the diameter of
the last reamer used and the real acetabular component prior to impaction. We
recommend that this difference should be =2 mm, regardless of the cup
design used. The bone around the acetabular rim is typically very dense, and
this area experiences the most stress with elliptical cup
impaction6,9.
It is likely that the 3.5% fracture rate associated with the Implex design was
due to a combination of the rim flare, the increased frictional coefficient of
the tantalum, the younger age of the patients for whom these components were
typically chosen, and, most importantly, the monoblock design, which does not
allow visual confirmation of complete component seating.
In most instances, the acetabular fracture was noticed during cup
impaction. The vast majority of acetabular components were stable despite the
fracture and were not exchanged for other components. When the acetabular
component was judged to be unstable after fracture, it was changed from a
monoblock component to a modular component and was secured with supplemental
screws. Acetabular bone from reaming was routinely inserted into and around
the fracture line.
The present study has several limitations, including the retrospective
methodology, the participation of multiple surgeons, and the likelihood that
additional fractures may have occurred but may not have been diagnosed
intraoperatively. Thus, the true prevalence of these fractures is probably
much higher. Unless the fracture line is clearly visible, many nondisplaced
fractures probably go unnoticed. However, if the obvious fractures that were
recognized in this series did not adversely affect cup longevity, it is likely
that smaller or nondisplaced fractures that go unnoticed also will not
adversely affect cup longevity. The present study indicates that if initial
cup stability is obtained, osseous ingrowth is predictable.
In summary, acetabular fracture is a rare complication of primary total hip
arthroplasty. The fracture typically occurs during cup impaction. The present
study documented significantly higher fracture rates in association with an
elliptical monoblock component fabricated from tantalum that was inserted into
a bed that had been prepared with a hemispherical reamer. Surgeons using such
designs should be cognizant of the increased risk of fracture associated with
this type of implant. It should also be noted that total hip arthroplasty was
performed through traditional exposures. As minimally invasive techniques
continue to evolve, knowledge of specific acetabular component design features
that have higher fracture rates is important because visualization is often
limited. ?