Intact hyaline articular cartilage is necessary for normal knee function.
Full-thickness articular cartilage defects are common in young active
patients1. Because
articular cartilage has a limited intrinsic healing capacity, cartilage
lesions persist indefinitely, cause pain, and typically progress to global
joint
degeneration2-4.
Treatment options for these large articular cartilage defects include
intrinsic repair stimulus (abrasion arthroplasty and microfracture),
cell-based cartilage repair (autologous chondrocyte implantation), autologous
whole-tissue transplantation (mosaicplasty), allograft tissue transplantation,
osteotomy, and
arthroplasty5-16.
Allograft tissue has been successfully used for the treatment of large
chondral and osteochondral defects for many
years17-23.
There has been much interest in fresh allograft tissue as a donor source
because such grafts possess metabolically active
chondrocytes24-27.
As articular cartilage matrix is dependent on chondrocyte metabolism for
long-term maintenance, the presence of viable cartilage cells in fresh
allografts is desirable. For many years, the use of fresh osteochondral
allografts was limited to a small number of institutions in North America.
However, in 1998, commercially supplied, fresh-stored osteochondral allografts
became available in the United States for human transplantation.
It is important to consider, however, that these commercially available
grafts are typically stored for intervals that far exceed the allograft
storage periods described in previous
reports17-21.
Currently, fresh allografts are typically available approximately thirty days
after tissue harvest; this interval is used to complete processing and testing
according to the American Association of Tissue Banks
guidelines28,29.
As recent in vitro studies have demonstrated decreased chondrocyte viability
and degradation of the biomechanical properties of osteochondral allografts
stored for over fourteen days, graft storage and its effect on clinical
outcome remain an area of
interest24-27,30.
To date, no study, as far as we know, has analyzed the effectiveness of
fresh stored allograft tissue in the reconstruction of symptomatic
osteochondral defects. We wished to test the hypothesis that fresh
osteochondral allografts that have been stored for greater than fourteen days
will effectively reconstruct full-thickness chondral and osteochondral defects
of the knee and improve knee function at a minimum follow-up of two years.
Aprospective registry dedicated to the tracking of patient outcomes
following articular cartilage repair and reconstruction procedures was
implemented at The Hospital for Special Surgery in 1999. The Cartilage
Registry has been approved by the institutional review board and is subject to
annual review. All eligible patients signed an informed consent form prior to
participation. Between 1999 and 2002, twenty-two of a total of 183 registry
patients were treated with fresh osteochondral allografts for symptomatic
chondral or osteochondral defects of the knee (femoral lesions). Patients with
multiple lesions (nonfocal disease), ligamentous instability, or severe lower
extremity malalignment were excluded. All patients had baseline preoperative
clinical outcome scores recorded in the database; all patients had a minimum
follow-up of two years. All grafts were fresh stored osteochondral allografts
that had been obtained from one of two tissue processors: the Musculoskeletal
Transplant Foundation (MTF; Edison, New Jersey) or CryoLife (Kennesaw,
Georgia).
Nineteen patients met the inclusion criterion for this study. There were
thirteen men and six women. At the time of surgery, the mean age was
thirty-four years (range, nineteen to forty-nine years). The mean duration of
symptoms prior to surgery was thirty-five months (range, four to 122 months).
Prior to allograft implantation, seventeen of the nineteen patients had had a
prior operation (mean, two operations; range, zero to four operations),
including microfracture arthroplasty (seven), mosaicplasty (three), anterior
cruciate ligament reconstruction (two), meniscal repair (two), osteochondral
allograft (one), pinning of an osteochondritis dissecans lesion (one), and
partial meniscectomy (one). At the time of surgery, the mean body mass index
(and standard deviation) was 28 ± 6 (range, 20 to 38). The lesions were
characterized as a full-thickness chondral lesion (five patients),
osteochondritis dissecans (thirteen patients), and osteonecrosis (one
patient).
All nineteen grafts were placed on the femur, with fourteen on the medial
femoral condyle and five on the lateral femoral condyle. The mean lesion size
was 602 mm2 (range, 121 to 1500 mm2). The mean interval
between allograft retrieval and implantation (storage time) was thirty days
(range, seventeen to forty-two days). Nine of the nineteen patients had
concomitant procedures (see Appendix). Three patients who had a meniscal
deficiency in the compartment that was to be treated also underwent meniscal
allograft reconstruction. Two patients who had a weight-bearing line
(mechanical axis) that fell through the medial compartment underwent a
valgus-producing high tibial osteotomy.
Allograft Preparation and Surgical Procedure
Allograft tissue was size-matched to the host femoral condyles with use of
anteroposterior and lateral radiographs or coronal and sagittal magnetic
resonance imaging studies of the affected knee. Allografts were deemed an
acceptable size match if both of the measured planes fell within 2 mm of the
host hemicondylar region that was to be grafted. All grafts were procured by
local tissue banks and were processed by MTF or CryoLife. All donated tissue
was tested for bacterial contamination on initial receipt and again
immediately prior to graft delivery. All tissue was tested serologically for
hepatitis A, hepatitis B, hepatitis C, and the human immunodeficiency virus.
Whole knee specimens were stored in 750 mL of a nutritive medium at 4°C.
The medium consisted of modified Dulbecco Modified Eagle Medium with HEPES
(N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid) buffer and no phenyl
red, proteins, amino acids, antibiotics, antimycotics, or vitamin E. Once
procured, allograft tissue was placed into the storage medium. The allograft
specimens were not removed from the storage vehicle until the time of
implantation. Prior to implantation, the storage time of the allograft was
ascertained from the commercial vendor by serial number. The aforementioned
process of graft processing lasted at least fourteen days.
The surgical method of implantation began with a knee examination, with the
patient under anesthesia, and knee arthroscopy. Cartilage surfaces were graded
by the surgeon according to the modified Outerbridge classification, with
Grade 0 indicating normal articular cartilage surface; Grade 1, softening or
swelling of articular cartilage surface; Grade 2, surface fibrillation of
<50% of articular cartilage thickness and <1.5 cm in length; Grade 3,
articular cartilage thickness erosion of >50% or >1.5 cm in length; and
Grade 4, exposed subchondral
bone31,32.
In all cases, the surgeon inspected the donor allograft to ensure a gross size
match to the lesion noted at the arthroscopic examination; no allografts were
rejected. A limited knee arthrotomy (without patellar eversion) was used to
implant the allografts in the host femoral condyles. Cylindrical osteochondral
grafts were used and were placed with use of a press-fit technique that did
not require internal fixation (Figs. 1-A
and 1-B). For condylar lesions, a hemicondylar specimen (medial or
lateral) was used for graft preparation; whole knee allograft specimens were
used to prepare donor osteochondral cylinders for trochlear defects. A
commercial system was used to prepare the host knee defect and donor graft for
implantation (Arthrex, Naples, Florida); a 0.5-mm offset between the created
host defect and the donor graft facilitated press-fit fixation of the grafts.
The host defect was prepared to receive the allograft specimen by creating a
cylindrical defect with use of a power reamer over a guidewire. Next, the
corresponding area to be harvested from the donor graft was demarcated with
use of a marking pen. The donor tissue specimen was stabilized, and a
cylindrical graft was harvested with use of a core reamer. The osseous portion
of the graft was then trimmed with a saw; the typical plug height ranged from
8 to 15 mm, depending on the degree of osseous involvement at the lesion site
noted at the time of surgery. Osteochondral allograft plugs were gently placed
into the defect and tamped into position, such that the allograft articular
surface was maximally matched (flush) with the host articular cartilage. In
each case, the articular cartilage was reconstructed in such a fashion that
the surrounding remaining host cartilage was either Grade 1 or 0 according to
the classification of Outerbridge.
Following surgery, the patients were managed with a hinged knee brace and
toe-touch weight-bearing for a minimum of eight weeks. A unicompartmental
unloader brace was used for four months after the initial eight-week interval.
Continuous passive motion was used in all patients for the first six weeks
following graft placement. A supervised rehabilitation program was started two
weeks after surgery and lasted between four and eight
months33.
Discontinuation of physical therapy was predicated on a restoration of normal
gait and the development of functional quadriceps muscle control.
Functional Outcome Evaluation
Routine follow-up examinations were performed by the operating surgeon.
Questionnaires to determine the functional score were distributed to the
patients at three, six, twelve, twenty-four, thirty-six, forty-eight, and
sixty months following graft implantation. The scores were recorded in the
Cartilage Registry. The mean follow-up interval was forty-eight months (range,
twenty-one to sixty-eight months). Postoperative data collection was performed
by an independent observer. Instruments for outcome evaluation included the
Activities of Daily Living Scale of the Knee Outcome
Survey34 and the
Medical Outcome Study 36-Item Short-Form Health Survey (SF-36) physical and
mental component scales (version
1.0)35. These
outcome instruments have been previously validated at our institution and have
been previously used for prospective evaluation of other cartilage repair
techniques16,36-38.
Magnetic Resonance Imaging
Magnetic resonance imaging scans (moderate echo-time fast-spin-echo
cartilage-sensitive pulse sequencing) were used to evaluate the morphologic
characteristics of the implanted grafts. The scans were acquired for eighteen
of the nineteen patients at an average of twenty-five months (range, six to
forty-nine months) following graft implantation. Magnetic resonance imaging
was performed for study purposes only and was not done on the basis of the
development of clinical symptoms. Final magnetic resonance imaging scans were
acquired between six and twelve months postoperatively in two patients,
between twelve and twenty-four months in four patients, between twenty-four
and thirty-six months in nine patients, and greater than thirty-six months in
three patients. Fifteen patients had at least two scores.
Magnetic resonance imaging was performed on a clinical 1.5-T magnetic
resonance imaging unit (Signa Horizon LX; General Electric Medical Systems,
Milwaukee, Wisconsin) with use of a commercial send-receive extremity
phased-array knee coil (knee phased array; Medrad, Indianola, Pennsylvania).
Images included a coronal fast-spin-echo sequence utilizing a repetition time
of 4000 to 4500 msec, an echo time (TE effective) of 34 msec, a field of view
of 11 to 13 cm, a slice thickness of 3 to 3.5 mm with no interslice gap, a
matrix of 512 × 256 to 288, and two excitations at a receiver bandwidth
of 20.8 to 31.2 kHz. The sagittal fast-spin-echo sequence was performed with
use of a repetition time of 3500 to 4000 msec, an echo time of 40 msec, a
field of view of 16 cm, a slice thickness of 4 mm with no gap, a matrix of 256
× 224, two excitations with a bandwidth of 20.8 kHz, and
frequency-selective fat suppression. High-resolution sagittal fast-spin-echo
sequences were acquired with use of a repetition time of 4000 to 5000 msec, an
echo time of 34 msec, a field of view of 16 cm, a slice thickness of 3.5 mm
with no interslice gap, and a matrix of 512 × 384 at two excitations
with a bandwidth of 31.2 kHz. Axial fast-spin-echo sequences were acquired
with a repetition time of 4000 to 5000 msec, an echo time of 34 msec, a field
of view of 16 cm, a slice thickness of 3.5 mm with no interslice gap, and a
matrix of 512 × 384 at two excitations with a bandwidth of 31.2 kHz.
This multiplanar cartilage-sensitive pulse sequence has been previously
validated39.
Magnetic resonance imaging assessment was performed by a musculoskeletal
radiologist without knowledge of the patient or treating surgeon. Images were
scrutinized for the relative intensity of the cartilage signal compared with
native cartilage, with use of a region-of-interest analysis on the magnetic
resonance imaging workstation (Advantage Windows; GE Healthcare, Milwaukee,
Wisconsin). Allograft cartilage signal was defined as isointense (the same
signal properties as the host articular cartilage), hypointense (a lighter
signal intensity compared with the host articular cartilage), or hyperintense
(a darker signal intensity compared with the host articular cartilage). The
morphology or geometry of the transplant was reported as flush, depressed, or
proud. Displacement, when present, was noted. Subchondral edema was determined
to be mild (<1 cm2), moderate (1 to 2 cm2), or severe
(>2 cm2). Osseous overgrowth of the subchondral plate, if
present, was noted. The interface of the adjacent articular cartilage was
assessed as smooth, fissured in an area measuring =2 mm, or fissured in an
area measuring >2 mm. Lesion fill was determined as 0% to 33%, 34% to 66%,
and 67% to 100% on the basis of both coronal and sagittal images. Grading of
the adjacent and opposite cartilage surfaces was performed with use of a
modified magnetic resonance imaging-based Outerbridge
classification39.
Fat-pad scarring was assessed as mild, moderate, or severe. Trabecular
incorporation of the graft was assessed in three planes of imaging, with use
of the higher in-plane resolution techniques, and was subjectively determined
to be complete, partial, or poor on the basis of the appearance of crossing
trabeculae and the presence of sclerosis at the donor interface. Signal
characteristics of the bone plug were also judged to be isointense to fat,
edema, or fibrosis (a low-signal intensity on all pulse sequences). This
magnetic resonance imaging assessment
system40, with the
exception of the degree of osseous incorporation of the graft and the signal
characteristics of the plug, has previously been used in the assessment of
microfracture37.
Statistical Analysis
Intragroup comparison between parameters before and after allograft
implantation was tested with the use of a paired t test. For testing
relationships between variables, linear regression and Pearson correlation
were used. A p value of <0.05 was considered significant.
Scoring Instruments
At a mean follow-up of forty-eight months (range, twenty-one to sixty-eight
months) after allograft implantation, the mean score on the Activities of
Daily Living Scale, the mean combined SF-36 score, and the mean SF-36 physical
component score all significantly increased compared with baseline scores. The
mean score (and standard deviation) on the Activities of Daily Living Scale
increased from a baseline of 56 ± 24 (range, 20 to 100) preoperatively
to 70 ± 22 (range, 30 to 98) at the time of the final follow-up (p <
0.05). The mean combined SF-36 score increased from a mean baseline value of
51 ± 23 (range, 18 to 96) to a final score of 66 ± 24 (range, 9
to 96) (p < 0.005). Additionally, the mean SF-36 physical component score
increased from a mean preoperative value of 32 ± 10 (range, 18 to 96)
to a final follow-up score of 40 ± 12 (range, 22 to 59) (p < 0.005).
However, no significant difference was demonstrated in the mean SF-36 mental
component score, which improved from 46 ± 13 (range, 24 to 64) to 49
± 11 (range, 38 to 62) (p = 0.1).
With the numbers studied, we could not correlate patient age, body mass
index, lesion location, lesion size, lesion type (chondral compared with
osteochondral), duration of symptoms, the need for concomitant procedures
(including knee osteotomy), or length of follow-up with the baseline
functional scores. The mean baseline score on the Activities of Daily Living
Scale was higher for patients with a history of one or no prior procedure (67
± 27) compared with the patients who had had multiple procedures (45
± 14) (p < 0.05). Similarly, having one or no prior procedure
correlated with higher mean baseline scores on the Activities of Daily Living
Scale (r = 0.491, p < 0.05) compared with patients who had two or more
prior procedures.
In this study, the mean age was thirty-one years for the men compared with
thirty-nine years for the women. The mean baseline combined SF-36 and SF-36
physical component scores were significantly higher for men than for women,
whereas, with the numbers studied, the scores on the Activities of Daily
Living Scale were not significantly different between genders. The mean
baseline combined SF-36 score was 58 ± 21 for the men compared with 36
± 20 for the women (p < 0.05). The mean baseline SF-36 physical
component score was 36 ± 12 for men compared with 30 ± 4 for
women (p < 0.05). The mean combined SF-36 scores increased significantly
for the men (from 58 ± 21 at baseline to 73 ± 16 at the latest
examination) and for the women (from 36 ± 22 to 52 ± 33) (p <
0.05). Male gender also was found to correlate with better combined SF-36
scores at the time of follow-up (r = 0.452, p < 0.05). Men had a
significant change in the mean activities of daily living score (58 ±
25 to 76 ± 16) (p < 0.05), whereas the mean score for women only
improved marginally from 53 ± 23 to 58 ± 30 (p = 0.36).
With the numbers studied, we could not correlate age at the time of
surgery, age at the time of follow-up, body mass index, lesion location,
lesion size, lesion type, duration of symptoms, number of prior procedures, or
status of concomitant procedures with the follow-up functional scores.
Magnetic Resonance Imaging
Eighteen of the nineteen patients underwent magnetic resonance imaging of
the involved knee at a mean interval of twenty-five months (range, six to
forty-nine months) after graft placement. Magnetic resonance imaging
demonstrated that, in general, the thickness of the implanted allograft
articular cartilage was maintained; the signal properties of the allograft
articular cartilage were isointense (compared with normal articular cartilage)
in eight of the eighteen grafts (see Appendix). The allograft articular
cartilage signal was hyperintense in seven grafts and hypointense in three
grafts. No graft displacement was observed in the eighteen allografts. When
the graft positioning was analyzed, ten grafts were found to be flush, five
were depressed, and three were proud relative to the height of the surrounding
host articular cartilage surface. Magnetic resonance imaging demonstrated that
trabecular incorporation of the allograft bone was complete in three, partial
in eleven, and poor in four grafts. An assessment of allograft osseous signal
demonstrated an appearance that was consistent with normal fat in four grafts,
edematous in eleven, and fibrotic in three. Fissures at the graft-host
interface were typically present; four grafts demonstrated a smooth interface
with the native host cartilage (Fig.
2). In the remaining patients, four grafts demonstrated fissures
that were =2 mm, and ten grafts demonstrated fissures that were >2
mm.
On the basis of the magnetic resonance imaging parameters, the ideal
allograft appearance is a nondisplaced, well-incorporated flush graft with a
graft marrow signal consistent with fat and with articular cartilage that is
isointense to normal cartilage signal and smooth
(Figs. 3-A and 3-B). Poor
allograft appearance can be characterized by poor graft-host matching and a
lack of osseous incorporation of the graft
(Figs. 4-A and 4-B).
Magnetic resonance imaging demonstrated that no overgrowth of the
subchondral plate was apparent in any planes. The percent fill of the defect
was generally good, with fifteen grafts demonstrating 67% to 100% fill; only
two grafts demonstrated relatively poor (0% to 33%) fill, and one graft
demonstrated 34% to 66% fill. Thirteen grafts had no or mild fat-pad scarring,
while four had moderate scarring and one had severe scarring. The Outerbridge
classification for the perilesional cartilage (cartilage immediately
surrounding the grafted area) was normal (Grade 0) in two patients, Grade 1 to
2 in twelve patients, and Grade 3 in four patients. Similarly, the Outerbridge
classification for the opposing chondral surfaces was normal (Grade 0) in
three patients, Grade 1 to 2 in seven, Grade 3 in seven, and Grade 4 in one
patient. With these small numbers, the Outerbridge classification of the
surrounding or opposing surface cartilage did not correlate with functional
outcome.
Complete and partial trabecular incorporation correlated with the combined
SF-36 scores at the time of follow-up (r = 0.487, p < 0.05), but not with
the scores on the Activities of Daily Living Scale (r = 0.409, p = 0.09). The
mean combined SF-36 score at the time of follow-up was significantly higher
for patients with complete or partial integration (73 ± 20) compared
with patients with poorly integrated allografts (45 ± 31) (p <
0.05). The final scores on the Activities of Daily Living Scale were higher in
patients who had complete or partial trabecular incorporation (73 ± 21)
compared with those with poorly integrated grafts (52 ± 22); the
difference was not significant (p = 0.09). It was also observed that the
patients with more complete or partial incorporation were younger (mean age,
thirty-two years) at the time of surgery compared with patients with poorer
incorporation (mean age, forty-three years) (p < 0.05).
Prior to implantation, the average allograft storage time was thirty days
(range, seventeen to forty-two days). With the numbers studied, no correlation
was found between graft storage time and the score on the Activities of Daily
Living Scale at the time of follow-up (r = 0.126, p = 0.62), combined SF-36
score (r = —0.07, p = 0.78), or physical component SF-36 score (r =
0.119, p = 0.64). A longer graft storage time (greater than thirty days)
correlated with better graft morphology (r = 0.465, p < 0.05), less
subchondral edema (r = 0.540, p < 0.05), and either complete or partial
trabecular incorporation (r = 0.501, p < 0.05).
Complications
There were no infections and no deep venous thromboses. One patient
required manipulation of the knee while under anesthesia two months following
graft implantation. One patient underwent a second allograft procedure
thirteen days postoperatively to correct a surgical error relating to a poor
graft-host match (a 3-mm proud graft); the same graft was used and
repositioned flush to the native articular surface. Four grafts failed
clinically. Two patients underwent revision osteochondral allograft
transplantation following gross collapse and fragmentation of the first graft.
A third patient underwent total knee arthroplasty two years following
allograft placement. The fourth patient underwent autologous osteochondral
transplantation (mosaicplasty) to correct a partial graft collapse involving
the anterior 20% (surface area) of the original allograft. After retrieval of
the failed grafts, pathologic examination demonstrated articular cartilage
fragmentation and necrotic bone. At a mean follow-up interval of forty-eight
months after implantation, sixteen of the nineteen allografts still functioned
within the host knee.
Fresh osteochondral allograft transplantation in the treatment of chondral
and osteochondral defects has a long clinical history with good to excellent
results at ten to fifteen
years17,18,41-51.
In North America, this experience had been limited to a few centers that had
the capability of handling donor procurement, graft storage, and recipient
matching of fresh grafts. These grafts were typically implanted within seven
days of retrieval. The hypothetical advantage of fresh allograft osteochondral
specimens is the presence of metabolically active chondrocytes whose continued
metabolism and synthesis of ground substance could maintain the cartilage
matrix of the implanted allografts. Fresh-frozen specimens have typically been
used for massive osseous defects that require large allograft specimens; such
specimens contain no viable bone or cartilage
cells52.
Fresh-frozen tissue is inappropriate for the reconstruction of most cartilage
defects, since these grafts lack viable chondrocytes. There have been numerous
reports of attempts to maintain chondrocyte viability in stored allograft
specimens with use of
cryopreservation53-58.
Those studies have generally described a low percentage of chondrocyte
survival, along with an observed compromise of the biomechanical properties of
the matrix.
In 1998, commercially supplied allografts became available in the United
States and the demand for allografts has increased. Commercially supplied
grafts must undergo testing that typically requires at least fourteen days. We
sought to better understand the impact of prolonged cold storage on these
grafts, and to critically assess the clinical outcome and magnetic resonance
imaging appearance of stored allografts used to treat knee cartilage
defects.
The present study demonstrates that the use of fresh osteochondral
allografts that had been in cold storage (for a mean of thirty days), which
were implanted to reconstruct symptomatic chondral and osteochondral defects,
resulted in an increase in functional scores at an average follow-up interval
of almost four years. The overall clinical results were comparable with those
described in previous
reports18,44,46,50,59-61.
Those reports focused on outcomes associated with the use of fresh grafts that
were stored for a short interval (less than seven days). Agnidis et al. found
increased functional outcome with validated scoring instruments (SF-36 scores)
after allograft implantation for knee cartilage defects after five
years59. In another
study, Gross et al. reported Kaplan-Meier survivorship scores of 95% at five
years and 85% at ten years for femoral
allografts18. The
graft survivorship observed in this study (84% at four years) is lower than
that reported by Gross et al. It is likely that the relatively small sample
size of the current study adversely affected the survivorship percentage. In
addition, because of its superior soft-tissue contrast and direct multiplanar,
tomographic acquisition, magnetic resonance imaging is more sensitive than
two-dimensional radiographs in detecting focal graft collapse, peripheral
integration, and the status of the overlying cartilage.
Magnetic resonance imaging was used as a noninvasive method of evaluating
transplanted allograft tissue. One study has demonstrated a positive
correlation between visual scores at second-look arthroscopy, the histological
appearance of repair cartilage, and magnetic resonance imaging scores after
autologous chondrocyte
transplantation62.
As such, cartilage-sensitive magnetic resonance imaging is a powerful tool in
assessing graft morphology following cartilage repair procedures, and it
provides information regarding the surrounding native tissue. The magnetic
resonance imaging findings in this study demonstrated that trabecular
incorporation of the graft (osseous integration) is an important prognostic
indicator of allograft health and viability. This study positively correlated
this magnetic resonance imaging parameter with functional outcome. Thus, it is
reasonable to conclude from this observation that magnetic resonance imaging
can be used to assess allograft appearance following implantation and may be
an effective method of predicting outcome. While trabecular incorporation was
observed in many grafts, abnormal bone-signal properties were noted in the
majority of implanted allografts, suggesting that osseous integration was
incomplete at the follow-up interval. Sirlin et al. demonstrated that magnetic
resonance imaging parameters could be used to assess the morphology of
osteochondral allografts following
implantation63.
Those authors found that graft failure or collapse resulted from a weakening
of the osseous portion of the graft rather than from primary cartilage
collapse; cartilage collapse only occurred in the setting of osseous failure.
The authors also noted that antibody-positive patients demonstrated a greater
degree of bone-marrow edema pattern than antibody-negative patients, and
antibody-positive patients also demonstrated a higher proportion of surface
collapse63.
In the present study, no correlation between functional outcome scores and
graft storage time was demonstrated. Although graft storage time did not
appear to affect outcome, longer-term clinical studies are needed to further
assess this issue. Interestingly, longer graft-storage times were found to
correlate with less graft edema, better graft morphology, and partial or
complete osseous incorporation of the graft as noted on the follow-up magnetic
resonance imaging evaluations. This finding is not easily explained. It is
possible that the antigenicity of allograft tissue decreases with longer
storage times, and a decrease in graft antigenicity might explain the observed
magnetic resonance imaging findings (better osseous incorporation and less
graft edema) in grafts stored longer than thirty days. Moreover, a decrease in
the number of viable chondrocytes within the implanted graft may have also
contributed to this phenomenon. In vitro studies have demonstrated poor
chondrocyte viability and diminished material properties in stored allograft
tissue26,27.
There remains a concern that such findings may ultimately compromise the
structural integrity and clinical effectiveness of stored, fresh
grafts26,27.
The limitations of this study include the small sample size and a lack of
physical examination findings. Nonetheless, we were able to demonstrate that
fresh allograft tissue that had been stored between seventeen and forty-two
days was effective over the short term in reconstructing chondral and
osteochondral lesions of the knee. Functional outcome scores improved, and
magnetic resonance imaging assessment of these grafts demonstrated the
maintenance of the articular cartilage thickness in the majority of the
implanted grafts. This study did not address the long-term function of these
allografts, which may undergo degeneration over time.
Note: The authors thank Robert Marx, MD, Edward Jones, MD, Erica
R. Urquhart, MD, PhD, Robert L. Buly, MD, Scott Rodeo, MD, and Thomas L.
Wickiewicz, MD, for their assistance in the preparation of this
manuscript.