To The Editor:
As one of the principal developers of the collarless, polished,
tapered (CPT) system, and a pioneer of impaction grafting in North
America, I read "Catastrophic Failure of a Cemented, Collarless,
Polished, Tapered Cobalt-Chromium Femoral Stem Used with Impaction
Bone-Grafting. A Report of Two Cases" (81-A: 844-847, June 1999),
by Jazrawi et al., with deep interest. The authors have performed
a service to the orthopaedic community by presenting textbook cases
of incorrect use of the system.
It is not implant design but implant misuse that caused the catastrophic
failures. The authors' operative technique violated every principle stressed
by the developers2. The authors'
technique, coupled with its application to the wrong candidates,
led to the disastrous results.
In the absence of an obvious manufacturing defect (as admitted
by the authors on page 846), the four remaining factors leading
to fracture of femoral stems, as listed by the authors, contain sufficient
reason for the implants to have failed. The fact that their revisions
remained successful for two to three years, despite the flawed surgical
technique, was most probably due to the strength and durability
of the CPT forged cobalt-chromium stem.
In the radiograph of Case 1 (Fig. 1-A), 20 to 25 percent of the
proximal part of the stem was proud; it should not have been more
than six to eight millimeters above the etching on the stem. The
remainder was not stabilized with cement, leading to the varus inclination
and fatigue fracture of the distally fixed stem.
In the radiograph of Case 2 (Fig. 2-A), the distal tip of the
prosthesis was not centralized and rested on the lateral cortex,
causing a clinically important stress-riser effect. Even though
the tip was fixed distally, the stem was not fixed proximally. Moreover,
the tip was believed to be protected by a strut graft and cables,
but these should have extended at least ten centimeters distal to
the tip of the stem, preferably supplemented by an equally long
strut graft medially.
Since 1989, I have used the grafting technique in 259 cases,
initially with the Exeter system and then with the CPT system (which
became available in 1990); I have not had a single mechanical failure
thus far. The femur has to be converted to an intact tube to allow
tight impaction of the graft and proper cementing of the entire
stem with optimal axial and torsional stability1.
Circumferential soft-tissue stripping and wire meshing are detrimental
to the vascularity and stability of the reconstructed proximal part
of the femur3.
Bauer (from the Cleveland Clinic) has independently shown incorporation
of the graft with the native bone in studies involving four biopsies4 and one autopsy5.
The autopsy was performed on a patient who had died of massive myocardial
infarction three years after undergoing the fourth and only successful
revision, which was performed with the CPT system with allografting.
The cement mantle and the viable bone graft covered the entire stem.
The CPT system with allografting requires specialized training
in centers where the technique is used frequently and routinely.
It is hoped that the article by Jazrawi et al. has increased awareness
of the need to adhere to the requirements established by the designers
of the CPT system in primary and revision procedures.
W. E. Michael Mikhail, M.D.
Medical College of Ohio
Dowling Hall, Room 2440
3065 Arlington Avenue
Toledo, Ohio 43614-5807
To The Editor:
I read "Catastrophic Failure of a Cemented, Collarless, Polished,
Tapered Cobalt-Chromium Femoral Stem Used with Impaction Bone-Grafting.
A Report of Two Cases" (81-A: 844-847, June 1999), by Jazrawi et
al., with great interest. I agree with the authors that poor proximal
bone support in combination with good distal fixation is the probable
cause for these failures. My colleagues and I previously reported
on bone-remodeling and allograft incorporation in a study of biopsy
specimens taken from the proximal part of the femur after revision
with an impaction grafting technique4.
We stated in that report that vascularity is necessary for bone-remodeling
and that it is unlikely that all of the bone graft in this construct
will be replaced by viable bone. As our report was based on biopsies,
we could not evaluate the degree of remodeling and graft incorporation around
the whole implant.
My colleagues and I have also reported on a patient with chronic
renal failure, osteoporosis, osteoarthritis, and two previous hip
implants (one cemented and one uncemented)5.
He underwent revision of the uncemented side with the impaction
grafting technique, and the proximal part of the femur was reconstructed
using a wire mesh, strut grafts, and a 3.5-millimeter reconstruction
plate in the calcar region to support the graft. The clinical result
was excellent, with no measurable subsidence of the femoral component
postoperatively. Three years later, he died suddenly of a myocardial
infarction. He had donated his hip to research. The proximal part
of the femur was removed, and, after fixation, the femur was radiographed
and photographed. The implant was extracted. The femur was sectioned
longitudinally, and the cement mantle was evaluated. Transverse
cuts were then made and prepared for routine histological studies.
The cement mantle was intact, and bone-graft remodeling with graft
incorporation was found around the stem. The bone graft did not
remodel consistently. Much of the distal part of the graft appeared
to have remodeled to viable bone, in contrast with some medial and proximal
areas where nearly all bone remained necrotic.
These findings support our statement that vascularity is necessary
for bone-remodeling. Consequently, incorporation of bone graft that
has no contact with living bone seems highly unlikely, although
the bone graft may act as an adequate mechanical support for the
implant.
In the case reports presented by Jazrawi et al., I think that
the proximal femoral bone was so poor that graft incorporation and
formation of new bone was not possible, while distal graft incorporation
and formation of new bone probably occurred. This resulted in high
mechanical stress on the implant, and a fatigue fracture of the
implant occurred with time. I suspect that a fatigue fracture will
occur in any implant designed for proximal fixation under these
conditions, regardless of implant design. I believe that the most
important reason for these failures was that the proximal femoral
bone was too poor. In these cases, it would have been better to
use a concept that relied on distal fixation of the femoral component
rather than the impaction grafting technique with a short-stem CPT
prosthesis.
Lars Weidenhielm, M.D., Ph.D.
Department of Orthopaedics, Karolinska Institute
S-171 76 Stockholm, Sweden
L. M. Jazrawi, C. J. Della Valle, F. J. Kummer, E. M.
Adler, and P. E. Di Cesare reply:
We thank Dr. Mikhail and Dr. Weidenhielm for their letters. As
both are pioneers of the impaction grafting technique for proximal
femoral reconstruction, we read their response letters with great
interest.
We agree with Dr. Mikhail that operative technique, primarily
the use of the impaction grafting technique in patients with uncontained defects
of the proximal part of the femur, played a role in the catastrophic
failures of both stems. However, we also believe that implant design, particularly
the tapered distal tip, increases the susceptibility to catastrophic
failure when proximal fixation is not achieved. In both cases, the distal
stem tip was well fixed without proximal bone support, resulting
in large bending stresses generated at this junction, which has
a small cross section.
The main intent of the paper was to make the orthopaedic community
aware of a complication resulting from expanding the indications
for use of the impaction bone-grafting technique. As Dr. Mikhail
stated, in order to utilize the impaction grafting technique, the
femur has to be converted to an intact tube to allow tight impaction of
the graft. We believe that in patients with substantial uncontained
proximal bone loss in the proximal part of the femur, the use of
the impaction bone-grafting technique in conjunction with either
wire mesh or reconstruction plates may not provide an adequate solution
to the problem of bone loss. Devascularization secondary to circumferential
soft-tissue stripping and inability to adequately impact bone graft
into the wire mesh both result in poor proximal graft incorporation.
It is this proximal bone support, which was lacking in both cases,
that contributed to the stem failure. Failure to adequately contain
the graft by the wire mesh led to failure of graft incorporation
and poor calcar support, resulting in cantilever bending secondary
to good distal stem fixation.
In Case 1, the stem was slightly proud. Nonetheless, the additional
proximal bone loss that occurred at the site of the reconstruction
plate and the wire mesh further contributed to the stem failure,
as the amount of bending force experienced by the femoral stem is
proportional to the amount of femoral stem left unsupported by intact
bone.
In Case 2, the distal tip of the prosthesis was not centralized
and the strut graft ended too close to the stem tip. Since the original
procedure was not performed at our institution, we can only speculate
on the reasons why the strut allograft was placed. We do not believe
that it was used to protect the stress riser caused by the eccentrically
placed stem but rather to reconstitute proximal femoral bone loss.
The strut allografting in combination with wire mesh converted the
femur into an intact tube to allow tight impaction of the bone graft.
The soft-tissue stripping required for circumferential wire mesh
resulted in devascularization of the proximal part of the femur.
Loss of the blood supply led to poor graft incorporation. In addition,
the wire mesh construct was not as strong as an intact cortical
tube. If the graft does not incorporate, the mesh cannot support
the cement mantle and a cantilever force is created in the stem.
In reference to Dr. Weidenhielm's comments, we are happy to see
that he is in agreement with our conclusions. The histological retrieval
analysis further supports our findings. The necrotic bone contained
proximally by the wire mesh will eventually be resorbed, resulting
in poor proximal fixation and eventual catastrophic failure of the
stem.
Lastly, we agree with Dr. Mikhail that impaction bone-grafting
should be performed in referral centers where the technique is used
frequently and routinely, as this will limit the injudicious use
of the technique and allow better tracking of complications.
Laith M. Jazrawi, M.D.
Craig J. Della Valle, M.D.
Frederick J. Kummer, Ph.D.
Edward M. Adler, M.D.
Paul E. Di Cesare, M.D.
Corresponding author: Laith M. Jazrawi, M.D.
Bernard Aronson Plaza, 301 East 17th Street
New York, N.Y. 10003