In 1979, a twenty-one-year-old woman presented with pain and radiolucency
in the left lateral femoral condyle related to a recurrent giant-cell tumor of
bone (Fig. 1). Fourteen months
previously, the patient had undergone curettage of the same area and the
defect had been filled with morselized allograft bone. In 1979, the lateral
femoral condyle was resected en bloc and replaced by a fresh osteochondral
allograft of the same dimensions composed of hyaline cartilage and sufficient
graft bone fashioned to replace the entire lateral femoral condyle
(Fig. 2-A). No adjuvant therapy
such as application of phenol or freezing was administered to the graft bed in
either the first or the second operation. The ex vivo time between harvesting
and transplantation was less than twelve hours. During this period, the graft
was maintained in Normosol R solution (Abbott Laboratories, Chicago, Illinois)
at 4°C as previously
described17.
In 2004, after twenty-five years of satisfactory function, total knee
replacement was necessitated by pain, instability, and joint narrowing.
Preoperative imaging demonstrated preservation of the lateral joint space
between the graft femoral condyle and the host tibial plateau
(Fig. 2-B). In contrast, the
joint space in the medial compartment was narrowed and increased radiodensity
of the tibial subchondral bone, reflecting osteoarthritis of the medial
compartment, was seen. The femoral condyle graft bone appeared to be diffusely
more radiodense than the adjacent distal femoral host bone, indicating
increased calcification typical of necrotic bone.
At the time of the operation, the articular surface of the graft femoral
condyle showed fibrillation and erosion, particularly at the central edge. The
interface between the graft and host cartilage was indistinct on the cartilage
surface. The graft appeared to be firmly affixed to the host bone. The volume
of surgically useful residual graft bone was such that a standard posterior
stabilized femoral component could be used. Accordingly, the femoral condyle
was trimmed through the graft bone, so that the technique was similar to that
used in a conventional total knee replacement.
Histopathological Findings
The resected specimen was examined with methods that have been described
previously13,17.
The articular surface of the femoral condyle allograft consisted of
cartilage with fibrillation and erosion, which was particularly evident toward
the medial (central) edge of the specimen. The graft-host interface was
macroscopically indistinct on both the articular and the resected osseous
surfaces (Fig. 3).
Histologically, the lateral side of the specimen showed host cancellous bone
subjacent to host fibrocartilage. The host bone was firmly adherent to the
graft bone by appositional bone formation on the necrotic graft bone
trabeculae and by some overgrowth of host fibrocartilage over the lateral edge
of the hyaline cartilage graft (Figs.
4 and
5). Under the central and
medial portion of the cartilage allograft, necrotic graft bone remained intact
without remodeling. This bone graft showed incremental calcification as seen
by increased radiodensity (Fig.
4). The deep portion of the graft consisted of bone trabeculae
devoid of osteocytes, indicating that the bone remained necrotic. At the
medial (central) edge, active resorption of graft hyaline cartilage by
fibrovascular pannus arising from synovium was seen
(Fig. 6, a). A
distinct border between host fibrocartilage and graft hyaline cartilage was
observed (Fig. 6, b).
A substantial portion of the articular surface consisted of graft hyaline
articular cartilage (Fig. 6,
c). Graft articular cartilage was recognized by the
presence of a preexisting tidemark with underlying structurally intact, but
necrotic, graft bone (Fig. 6,
d). On most of the graft, the articular cartilage
remained intact.
Apart from superficial fibrillation common to chronic cartilage injury,
histologic features of osteoarthritis were not observed. Multiple chondrocytes
were seen within chondrons, providing evidence of chondrocyte replication
after transplantation. Within this tissue, focal chondrons devoid of
chondrocytes, indicating previous individual chondrocyte necrosis, could be
observed. However, most of the graft cartilage showed intense safranin-O
staining within the middle zone, representing proteoglycan-rich matrix
actively maintained by chondrocytes. Chondrocytes with intact nuclei and
cytoplasm were present within chondrons in the superficial, middle, and deep
zones of the graft. The chondrocytes appeared to have more spherical shapes
(loss of polarity) than those in normal or osteoarthritic cartilage, but they
were otherwise intact (Fig. 6,
c). Electron microscopy
(Figs. 7-A and 7-B) showed
chondrons containing chondrocytes that had retained microvilli on the cell
membrane and displayed abundant rough endoplastic reticulum and mitochondria
within the cytoplasm, all features of chondrocyte viability. Within and beyond
the chondron, the extracellular matrix showed spherical electron-dense
particles typical for proteoglycan condensates. This feature is a marker of
functional chondrocyte
activity27,28.
Examination of the host articular tissues showed features of
osteoarthritis, including hyaline cartilage fibrillation, erosion, and
clustering of chondrocytes. In the medial compartment, both the femoral and
the tibial articular surface showed extensive areas of eburnated bone. The
tibial plateau also showed prominent osteophytes adjacent to the cruciate
ligament insertions.
For osteochondral lesions that are smaller than 3 cm in diameter and 1 cm
in depth, several well-documented reparative techniques are available,
including microfracture, mosaicplasty, and autologous chondrocyte
implantation29-32.
For larger defects following trauma or the treatment of a benign neoplastic
condition, the use of fresh allografts potentially allows biological
attachment sites for soft tissues important for joint function and may
postpone the necessity of joint
replacement2,4,5,33.
Furthermore, as was the case in our patient, the incorporated graft bone may
permit resection of less host bone than would otherwise be required at a
subsequent arthroplasty.
The usual osteochondral allograft procedure for repair of an osteoarticular
fracture or a similar articular lesion involves use of a graft with only a
thin shell of subchondral
bone17. Typically,
within a few years, the bone in grafts of this type is replaced by creeping
substitution with host bone, which then interfaces directly with the graft
cartilage17. In
contrast, in our patient, the graft included extensive cortical and cancellous
bone sufficient to replace the defect in the lateral femoral condyle. After
twenty-five years, there was appositional new bone on the sides of the graft,
but not in the center of the graft beneath the cartilage. The clinical
relevance of these observations is that a cartilage graft containing necrotic
bone can remain in place and function for twenty-five years or more, provided
that the graft is well fixed in place by appositional new bone formation.
Although the histologic study clearly demonstrated viable graft cartilage,
ideally DNA analysis of host and graft should be carried out to confirm that
the functioning chondrocytes seen are indeed graft chondrocytes.
Macroscopically, the borders of the graft and host cartilage were
indistinct. This was a result of the integration of host fibrocartilage with
the graft hyaline cartilage. However, it is interesting, with respect to
tissue-engineered implants, that this type of matrix integration can occur
with allograft cartilage tissue.
On the basis of our previous
observations3,17,
we believe that viable hyaline articular cartilage can be expected at
twenty-five years after transplantation if the graft remains stably in place.
From previous studies, we have determined that chondrocyte loss is an early
event related to inadequate graft
preservation17.
Focal chondrocyte loss can be expected in grafts that have survived for a long
period. Focal chondrocyte necrosis, particularly in the middle and deep zones,
can result from inadequate chondrocyte nutrition during the ex vivo period
before transplantation. Safranin-O staining in cartilage is a very sensitive
indicator of high proteoglycan density, which in turn is dependent on viable
chondrocytes. Early studies of cartilage transplantation demonstrated that
necrotic cartilage is devoid of safranin-O staining. Loss of matrix staining
by safranin O occurs within two to four days after chondrocyte
death13,17.
Furthermore, safranin-O staining can be depleted within two days by enzymatic
activity, and it can be restored within a few days provided that chondrocytes
remain viable34.
Development of joint instability at the graft site can lead to redistribution
of mechanical forces on the articular cartilage, which in turn can result in
focal chondrocyte loss in the superficial layer, similar to that caused by
induction of apoptosis in
osteoarthritis35,36.
Therefore, by itself, focal chondrocyte death in the superficial layer is not
necessarily an indicator of cartilage allograft failure.
This report illustrates that the hyaline cartilage portion of a fresh
osteochondral allograft can survive and function for twenty-five years or
more. Factors that might have been important in preserving cartilage viability
in this patient include prompt postmortem harvesting of tissue, a short
duration of ex vivo storage, and good fixation of the graft to host bone. This
report also illustrates that total replacement of graft bone by host bone is
not required to maintain graft stability. ?