Few options exist for the successful treatment of large traumatic
osteochondral lesions of the knee in young patients. Potential options include
autologous osteochondral plugs, autologous chondrocyte transplants, and fresh
osteochondral
allografts1-11.
Fresh osteochondral allograft transplantation has been utilized by the senior
author (A.E.G.) for more than thirty years and has proven to be a successful
option for young patients with large lesions of the knee, particularly those
of a posttraumatic
etiology12-19.
Despite a survival rate of 85% at ten years in a series of sixty-five fresh
tibial and seventy-two fresh femoral osteochondral
allografts12,13,
the young age of many of these patients means that total knee arthroplasty
ultimately may be required within a patient's lifetime.
The performance of a total knee arthroplasty in a knee that has had
previous treatment with a fresh osteochondral allograft potentially could be
associated with several inherent technical complexities, including the need
for bone graft or metal augments and stem extensions in order to compensate
for large osseous defects. Additionally, the frequent use of osteotomies in
and around fresh osteochondral allografts for the purpose of unloading the
graft results in additional potential for complexity due to extra-articular
deformity as well as the presence of
hardware12,13.
The purpose of the present study was to evaluate the surgical complexity
of, and the prevalence of complications related to, the performance of total
knee arthroplasty in patients who have had previous osteochondral graft
transplantation. To our knowledge, there are no reports in the literature
regarding total knee arthroplasty after fresh osteochondral allograft
transplantation.
Analysis of our prospective database revealed that, between 1974 and 2000,
thirty-three consecutive patients (thirty-five knees) underwent a total knee
replacement after a previous tibial or femoral transplantation of a fresh
osteochondral allograft. Total knee arthroplasty was performed at an average
of eight years and six months (range, two to nineteen years) after the fresh
osteochondral allograft transplantation.
A cemented posterior cruciate ligament-sacrificing prosthesis was used in
all knees, and all total knee arthroplasties were performed by a single
surgeon (A.E.G.). One knee was replaced in 1974 with use of a cemented total
condylar prosthesis (Johnson and Johnson, New Brunswick, New Jersey and
How-medica, Rutherford, New Jersey), twenty-one knees were replaced from 1978
to 1993 with use of the Insall/Burstein I posterior stabilized knee prosthesis
(Zimmer, Warsaw, Indiana), and thirteen knees were replaced from 1993 to 2000
with use of the PFC total knee prosthesis (Johnson and Johnson, Raynham,
Massachusetts). All knees were exposed through a medial parapatellar approach.
The patella was resurfaced in all cases. Two patients had had bilateral fresh
osteochondral allograft transplantation and subsequently underwent a staged
bilateral total knee arthroplasty. All patients had had at least one procedure
on the knee (mean, 2.19 procedures; range, one to five procedures) before the
total knee arthroplasty.
Perioperative documentation was evaluated with regard to etiology,
preoperative functional and subjective impairment, intraoperative technical
difficulties or complications, early and late postoperative complications, and
postoperative functional and subjective outcomes.
Two patients (three knees) died from unrelated causes less than three years
after the total knee arthroplasty and were unavailable for follow-up. Both
patients were last seen at the routine two-year postoperative follow-up visit
and were satisfied with the result of the total knee arthroplasty. The
technical data from these operations were included in the study.
The study group included nineteen women (twenty knees) and fourteen men
(fifteen knees) who had had a mean age of sixty-three years (range, twenty to
eighty-three years) at the time of the total knee replacement. Seventeen total
knee arthroplasties were on the right side, and eighteen were on the left.
At the time of allograft transplantation, twelve patients had had a
simultaneous proximal tibial osteotomy and six patients had had a distal
femoral varus osteotomy in order to decrease load on the graft. All
osteotomies were closing-wedge osteotomies. All tibial osteotomy sites were
fixed with staples, whereas all distal femoral varus osteotomy sites were
fixed with 90° blade-plates. In all knees but one, the hardware was
removed prior to the total knee arthroplasty; in the remaining knee, the
hardware was removed at the time of the total knee arthroplasty.
Preoperatively, twenty-eight knees were in varus alignment. Six of these
twenty-eight knees had had a failed proximal tibial osteotomy, and arthritis
had developed primarily in the medial grafted compartment. The other seven
knees were in valgus alignment. Three of these seven knees had had a failed
distal femoral varus osteotomy, and arthritis had developed in the lateral
grafted compartment.
Fourteen knees had received a lateral tibial plateau transplant, ten knees
had received both medial tibial plateau and medial femoral condyle
transplants, six knees had received a medial tibial plateau transplant, one
knee had received both medial tibial plateau and lateral femoral condyle
transplants, one knee had received separate medial and lateral tibial plateau
transplants, one knee had received medial femoral condyle and lateral tibial
plateau transplants, one knee had received only a medial femoral condyle
transplant, and one knee had received an intercondylar allograft transplant.
Twenty-one knees had received a meniscal transplant as part of the
osteochondral allograft procedure (Table
I).
Perioperative documentation was reviewed with regard to preoperative and
postoperative range of motion, the duration of the total knee arthroplasty
procedure, estimated blood loss, problems encountered during exposure and
patellar eversion, the need for augments or bone graft, and the use of an
extended stem in the tibia or a stem in the femur. Early and late
postoperative complications were reviewed, as was the duration until failure,
which was defined as the need for revision total knee arthroplasty. The mean
duration of follow-up after the total knee replacement was seven years and
eight months (range, three to fifteen years).
The Knee Society clinical rating system was used for clinical evaluation
beginning in
199020; therefore,
only eighteen knees were available for functional outcome analysis, whereas
the other seventeen were evaluated according to the follow-up notes and
subjective impression.
Preoperatively, nine knees had had a flexion contracture of 10° to
35° and four knees had had <90° of flexion. Pain and advanced
degenerative changes on radiographs in the affected compartment or
compartments were the indications for total knee replacement in all cases.
Technical Results
The mean duration of anesthesia during total knee arthroplasty was 2.5
hours (range, 1.5 to four hours), with a mean duration of 2.8 hours (range,
1.5 to four hours) for knees with a previous osteotomy and 2.1 hours (range,
1.5 to 2.5 hours) for knees without a previous osteotomy. All knees were
exposed through a standard medial parapatellar approach. In four of the
thirty-five knees, there were difficulties with patellar eversion; however,
three of these four knees had had a previous realignment osteotomy (two had
had a proximal tibial osteotomy and one had had a distal femoral varus
osteotomy). To achieve patellar eversion, two knees required a tibial
tuberosity osteotomy and two required a so-called rectus snip. Bone quality,
graft incorporation, and graft vascularity were judged subjectively by the
senior author (A.E.G.) during the course of the procedure, and all condylar or
tibial plateau cuts were done to bleeding bone. Intraoperative inspection
revealed that all but one of the allografts had united completely and that
bone stock was of good quality. The single graft that was found to be ununited
was in a knee that had had previous revision allograft procedures because of
aseptic nonunion of the medial tibial plateau.
In four knees, insertion of the tibial component was difficult and resulted
in intraoperative split fractures of the tibial plateau, necessitating
fixation with interfragmentary screws or intraosseous wires. Two of these
knees had had a previous osteotomy of the proximal part of the tibia, which
had caused severe deformity. In one of these knees, an uncemented tibial stem
was used to stabilize the tibial component; in the other knee, the tibial
fracture was fixed with interfragmentary screws alone. In the remaining two
knees with a split fracture, a comminuted tibial plateau fracture had been the
original etiology of the knee abnormality. Interestingly, the fresh
osteochondral medial tibial plateau allograft provided better support than did
the osteoporotic lateral tibial plateau, which was split during insertion of
the tibial component in these cases. Both knees were fixed with
interfragmentary screws and were stable enough to support the tibial component
without a stem. There was no relationship between the duration of survival of
the allograft prior to knee arthroplasty and the occurrence of a tibial
plateau fracture during total knee arthroplasty.
Only one knee, which had an ununited allograft as described previously,
required augmentation with a metal tibial wedge (a 20° hemiwedge) because
of substantial bone loss. Another two knees required morselized allograft.
However, no knee required structural allograft. All three knees that required
bone graft or a wedge had had a previous osteotomy, with two of them having
had a high tibial osteotomy and the other having had a distal femoral varus
osteotomy.
Three knees (9%) required a stemmed component on both the femoral and the
tibial side. Two of these knees had had a previous proximal tibial osteotomy,
and one had had a distal femoral varus osteotomy. The knee with the previous
distal femoral varus osteotomy had hardware removal simultaneously with the
total knee arthroplasty, and a stem was needed to bypass the screw-holes. One
of the knees that had had a proximal tibial osteotomy sustained an
intraoperative tibial plateau fracture, and a stem was needed to support the
tibial component. All stems (femoral and extended tibial) were uncemented.
The mean tourniquet time was seventy-six minutes (range, fifty-five to
ninety minutes), and the mean estimated intraoperative blood loss was 250 mL
(range, 100 to 900 mL). There were no difficulties related to wound closure
and no wound-healing complications such as drainage or dehiscence.
Functional Results
All patients had subjective improvement in terms of pain and function. The
mean Knee Society objective score for the knees that were treated after 1990
improved from 34.7 (range, 20 to 59) preoperatively to 87.9 (range, 61 to 95)
at the time of the latest follow-up. The mean Knee Society function score
improved from 45 (range, 15 to 65) preoperatively to 82 (range, 45 to 100) at
the time of the latest follow-up. With the numbers available, there was no
significant difference in the functional outcome of total knee arthroplasty
following simultaneous osteochondral allograft transplantation and realignment
osteotomy as compared with total knee arthroplasty following osteochondral
allograft transplantation alone (Table
II).
Of the eighteen knees that were evaluated with use of the Knee Society
clinical rating system, seven were classified as excellent, six were
classified as good, one was classified as fair, and one was classified as
poor. The other three knees were revised during the follow-up period. Of the
remaining seventeen knees, which were treated before 1990 and were not
evaluated with use of the Knee Society clinical rating system, ten were in
patients who reported subjective improvement in pain and daily function, two
were in patients who reported improvement in daily function but still
complained of moderate pain, two were in patients who died less than three
years following the index procedure, and three were revised during the
follow-up period. In the subgroup of twelve patients without a Knee Society
score who survived for more than three years after total knee arthroplasty and
did not have a revision, four were followed after 1990 and had a mean Knee
Society objective score of 83 (range, 75 to 88) and a mean Knee Society
function score of 75 (range, 65 to 85).
For all knees, the mean range of motion improved from 85°
preoperatively to 105° after total knee arthroplasty. Nine knees had had a
flexion contracture of >10° before total knee arthroplasty, whereas one
had a flexion contracture of >10° after total knee arthroplasty. Four
knees had <80° of flexion before total knee arthroplasty, and all knees
had at least 90° of flexion after total knee arthroplasty
(Table III). With the numbers
available, there was no relationship between the length of time from allograft
insertion to knee arthroplasty and the functional outcome of total knee
arthroplasty.
Six (17%) of the thirty-five knees underwent revision total knee
arthroplasty for the treatment of aseptic loosening within the follow-up
period. Only one of these knees had had a previous simultaneous fresh
osteochondral allograft transplantation and an osteotomy. Two knees with
aseptic loosening had had multiple operations before the fresh allograft
transplantation and were revised at twenty-four months postoperatively. Both
had undergone at least three procedures before the fresh osteochondral
allograft transplantation. The remaining revisions occurred five to thirteen
years after the index total knee arthroplasty. Three of them were due to
aseptic loosening. The fourth knee sustained a traumatic patellar fracture one
year after total knee arthroplasty and underwent a patellectomy followed by
revision total knee arthroplasty because of aseptic loosening ninety-six
months after the index total knee arthroplasty. This was the only knee that
required revision because of patellofemoral problems.
No deep infections occurred. The two knees that were revised early because
of loosening were investigated for the possibility of infection. Aspiration of
the knee joint, blood tests (including determination of the erythrocyte
sedimentation rate, C-reactive protein level, and complete blood-cell count),
and intraoperative frozen-section analysis and gram-staining ruled out
infection.
The use of fresh allograft transplants in the knee for the treatment of
large osteochondral defects has been well
established12-19,21-29.
Shasha et al. evaluated sixty-five patients who were managed with fresh tibial
osteochondral graft for the treatment of posttraumatic defects and reported
the survival rate to be 95% at five years, 80% at ten years, 65% at fifteen
years, and 46% at twenty
years12. Good to
excellent results according to The Hospital for Special Surgery scoring system
were found in 86% of patients at an average of twelve years (range, five to
twenty-four years). Despite these favorable results, many graft recipients
ultimately will require a total knee arthroplasty later in life. We are
unaware of any information in the literature regarding the surgical complexity
or the outcome of conversion from a fresh osteochondral allograft to total
knee arthroplasty.
It is known that total knee arthroplasty after femoral or tibial osteotomy
has inherent technical difficulties, higher complication rates, and inferior
results as compared with primary total knee
arthroplasty30-33.Windsor
et al. evaluated the technical considerations related to forty-five total knee
arthroplasties in forty-one patients who had had a previous proximal tibial
osteotomy and found that the extra-articular deformity was associated with
tibial stem impingement on the lateral tibial
cortex34. Other
technical difficulties were related to exposure and the loss of bone stock.
Cameron and Park, in a review of 158 patients who had undergone total knee
arthroplasty after previous proximal tibial osteotomy, reported a complication
rate of 11.2% and noted that ten revisions occurred in the first six years
after the total knee
arthroplasty35.
Their results were inferior to those of primary total knee arthroplasty for
the treatment of osteoarthritis. Parvizi et al., in a review of 118 patients
who had been managed with total knee arthroplasty following proximal tibial
osteotomy, reported one instance of intraoperative patellar tendon detachment
from the tibial tubercle and noted that thirteen knees underwent revision
arthroplasty at a mean of 5.9
years36. Several
other authors have also reported that total knee arthroplasty following
proximal tibial osteotomy involves increased technical
difficulties37-41.
The effect of distal femoral varus osteotomy on total knee arthroplasty has
been examined less frequently in the literature. Cameron and Park, in a review
of eight cases of total knee arthroplasty after distal femoral varus
osteotomy, reported that no femoral component required a stem extension and
that no additional technical difficulty was
noted42. In
contrast, Nelson et al., in a study of nine patients (eleven knees) who were
managed with total knee arthroplasty after distal femoral varus osteotomy,
concluded that total knee arthroplasty decreased pain and improved knee
function but noted that the procedure was technically demanding and was
associated with inferior results as compared with those of primary total knee
arthroplasty43.
In our experience, posttraumatic knees and knees with a previous tibial
osteotomy had a slightly higher prevalence of intraoperative complications
related to bone quality. In the cases of four knees, difficulties with
insertion of the tibial component resulted in intraoperative split fractures
of the tibial plateau. Two of these knees had had a previous realignment
osteotomy of the proximal part of the tibia, which caused substantial
deformity and resulted in difficulty inserting the tibial prosthesis. The
other two knees with split fractures had a comminuted tibial plateau fracture
as the original etiology of the knee abnormality. This technical complexity is
consistent with the findings of Saleh et al., who reported three infections
and two patellar tendon disruptions in a study of fifteen total knee
arthroplasties that had been performed after open reduction and internal
fixation of a tibial plateau
fracture44. Those
authors concluded that the procedure was technically demanding and was
associated with a high failure rate (five of fifteen).
In the present study, four of the thirty-five knees required additional
maneuvers for exposure. This rate of difficulty with exposure was higher than
that associated with standard primary total knee arthroplasty but was
comparable with that reported in other studies of total knee arthroplasty
after
osteotomy34,36,42.
Only one knee required augmentation with a metal tibial wedge to compensate
for bone loss. The need for this augment was attributed to an ununited
allograft in a multiply grafted knee. Another two previously osteotomized
knees required morselized allograft on the tibial side. Three knees that had
had an osteotomy concomitant with the allograft procedure were treated with
the insertion of stemmed components to protect against stress-risers.
The Knee Society clinical rating score revealed acceptable results in the
present series as compared with the results of previous studies of total knee
arthroplasties performed after realignment osteotomies around the
knee35,36,40,42,43.
Compared with the results of primary total knee arthroplasty for the treatment
of osteoarthritic knees, the results of the present series are slightly
inferior and the revision rate is
higher45-48.
This finding can be attributed to the multiple previous procedures and the
soft-tissue and osseous changes that we encountered.
In summary, we found that total knee arthroplasty in a patient who has had
a previous fresh osteochondral allograft transplantation and a concomitant
osteotomy can be slightly more technically complex than routine total knee
arthroplasty. In contrast, patients without a previous osteotomy had
noticeably fewer difficulties during the surgical procedure. The
intraoperative findings showed that bone stock had been restored after the
transplantation and that the need for augmentation or bone graft was minimal.
We believe that fresh osteochondral allograft transplantation of the knee does
not substantially complicate subsequent total knee arthroplasty beyond factors
related to previous osteotomy, including difficulties with exposure and the
potential need for component stem extensions. ?