Repair of a localized defect in the articular cartilage with a fresh osteochondral shell allograft is an alternative in a young adult for whom non-operative treatment has failed and prosthetic arthroplasty or arthrodesis of the affected joint is not appropriate. At our institution, the lesions most commonly treated with this technique are traumatic osteochondral defects.
Osteochondral allografts have not gained widespread acceptance, as the good results reported with their use have been based on small series with short-term follow-up from only a few centers2,6,9,12,17. This may be related to the lack of a sophisticated system for procuring donors, the lack of proved storage technology, the fear of transmitted disease, and uncertainty regarding a possible immune response to the allograft14,15. In addition, some tissue banks are reluctant to accept the liability for the dissemination of fresh tissue for a procedure that is not widely used. Yet another factor is the paucity of information on the survival and function of the allograft chondrocytes in humans. Despite the lack of substantive studies, a large number of younger patients in whom more accepted procedures have failed or who are not enthusiastic about the expected results of arthroscopic shaving or drilling of cartilaginous defects are being referred for osteochondral allografting.
The present report describes the histological characteristics, viability of the chondrocytes, and biochemical changes in an osteochondral allograft specimen that had been in situ for ten years. It should be noted that the conclusion of this case report cannot be generalized.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding sources were National Institutes of Health Grant AR 28467 and Grant SA346 from the Veterans Administration Rehabilitation, Research, and Development, San Diego, California.
†University of California, San Diego Medical Center, 200 West Arbor Drive, San Diego, California 92103. Please address requests for reprints to Dr. Convery.
A twenty-eight-year-old man, who worked as a mail carrier, sustained an injury of the left knee while playing baseball. Shortly after the injury, an arthroscopic débridement was performed because of pain and swelling, and the patient was told that he had a large defect of cartilage and bone in the left medial femoral condyle. The débridement temporarily relieved the symptoms. Eighteen months later, a second arthroscopic procedure was performed because of increasing pain, buckling, and swelling that was proportional to the degree of the patient's activity. This second procedure did not relieve the pain or buckling and, at the end of each day of delivering mail, the swelling was increased. The patient was referred for osteochondral allografting to fill a defect in July 1982 (Fig. 1).
A loose body, measuring 3.0 by 2.0 by 0.5 centimeter, was removed from the weight-bearing surface of the left medial femoral condyle and was replaced anatomically with a fresh osteochondral shell allograft that included approximately six millimeters of subchondral bone. The allograft was secured with an interference fit, without the use of internal fixation.
Approximately eight years postoperatively, the patient retired, with a disability pension, from his job as a mail carrier because of persistent pain when walking and aching in the knee after the cessation of walking. For the next two years, he worked as a meter reader. He had a normal range of motion but continued to have aching and pain after walking a long distance. Roentgenograms showed a transverse fracture and collapse of the allograft. A second allograft was inserted in September 1992 (Fig. 2).
At the time of the reoperation, the first allograft was easily removed, without the use of osteotomes, with a rim of soft, friable host articular cartilage. The sclerotic base of the defect was abraded with a high-speed burr. Autogenous bone that had been obtained locally was placed in the crater, and the defect, which measured 2.2 by 3.5 centimeters, was filled with a fresh allograft of cartilage and approximately six millimeters of subchondral bone. The allograft was secured with a press-fit supplemented with Orthosorb biodegradable polydioxanone pins (Johnson and Johnson Orthopaedics, New Brunswick, New Jersey).
The articular cartilage of the original allograft appeared to be normal, as it was smooth and glistening although depressed, whereas the host articular cartilage surrounding the allograft was soft and friable. Only a small remnant of the osseous portion of the allograft, estimated to be one to two millimeters thick, was found. The original allograft was split transversely into two pieces, and the posterior portion was sent for histological evaluation. The surrounding soft fibrous-appearing cartilage and subchondral bone were not submitted for histological assessment as they were trimmed sequentially to create a perfect press-fit. The anterior, slightly smaller portion and the surrounding rim of soft articular cartilage were collected separately for biochemical analysis.
Histological Analysis
The specimen of cartilage and bone measured two centimeters at its greatest dimension. It was trisected longitudinally and fixed in neutral buffered formalin, decalcified with EDTA, and embedded in paraffin. Six-micrometer-thick sections were stained with hematoxylin and eosin to study cellular details and with safranin O-fast green to search for the presence of glycosaminoglycans13.
The articular surface of the retrieved allograft was intact, without clefts or fissures. The chondrocytes appeared viable, and there was no cloning (Fig. 3). The remaining allograft bone was acellular, non-viable, devoid of marrow cells, and replaced by an amorphous eosinophilic material. There was no inflammatory infiltrate at the periphery of the sections. The junction between the subchondral bone and the articular cartilage was intact in two of the three sections; the remaining section had substantial separation, which may have been accentuated by the preparation of the specimen. There was intense staining with safranin O-fast green (Fig. 4), demonstrating glycosaminoglycans, which suggests normal function of the chondrocytes. There was no evidence of pannus formation or synovial inflammation.
In our practice, the usual thickness of the osseous shell is five to six millimeters. The thickness of the subchondral bone in the allograft ranged from one to three millimeters depending on the area that was examined. This indicates at least a 50 per cent resorption of the original subchondral bone. The reduced thickness of the osseous shell, which was avascular and necrotic (Fig. 3), suggests that the fracture and collapse of the allograft was the result of resorption or a lack of incorporation.
Viability of the Cells
Viability was determined with double-staining, involving the use of 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester, which is specific for living cells, and propidium iodide, which is specific for dead cells8. Thin slices of the allograft were prepared and placed in small Petri dishes with Eagle minimum essential medium, supplemented with 10 per cent fetal calf serum. The samples were incubated for twelve hours at 37 degrees Celsius in 95 per cent air and 5 per cent carbon dioxide atmosphere in a tissue culture incubator (model 3193, Water-Jacketed Incubator; Forma Scientific, Marietta, Ohio). The stain for viable cells was dissolved in dimethyl sulfoxide that contained a water solution of propidium iodide and was added to the Petri dishes that contained the medium and tissue slices to obtain a concentration of five millimolar. The samples were incubated for twenty minutes at 37 degrees Celsius in a 95 per cent air and 5 per cent carbon dioxide atmosphere and were immediately analyzed under a microscope with an objective lens of ten times magnification. A Zeiss inverted microscope (Carl Zeiss, Thornwood, New York) equipped with an ultraviolet light source and interference filters was used to select emission for both dyes (live cells and dead cells) simultaneously.
Double-staining of normal articular cartilage cells demonstrates an intense yellow fluorescence (Fig. 5-A). This is in contrast to the cells of normal cartilage that has been pretreated with concentrated acetic acid to kill the chondrocytes; such cells fail to acquire the yellow stain (Fig. 5-B). Of two double-stained sections from the allograft cartilage (Fig. 6-A and 6-B), one section had chondrocytes that appeared to be completely viable, as demonstrated by yellow-green fluorescence (Fig. 6-A). In the second section, there was a mixture of apparently viable (bright-yellow) chondrocytes but also some chondrocytes that seemed to be dying, as seen by the reddish-brown staining of the nucleus with propidium iodide (Fig. 6-B).
Biochemical Analysis
Collagen concentration was determined on the basis of hydroxyproline content. Three milligrams of dry tissue was hydrolyzed in one milliliter of 6N hydrochloric acid for three hours at 130 degrees Celsius. The acid hydrolysate was treated according to the procedure of Woessner, and the amount of hydroxyproline was determined spectrophotometrically at 557 nanometers.
Semiquantitative determination of type-II collagen was carried out by analysis of the peptides released after digestion of the tissues by cyanogen bromide1. Three milligrams of dry tissue was first suspended for twenty-four hours in four-molar guanidine hydrochloride, 0.05-molar Tris chloride (pH 7.5), to remove proteoglycans and other non-collagenous aqueous soluble proteins. After extensive rinsing in water, the residue was suspended in 70 per cent formic acid, and cyanogen bromide was added in a ratio of 3:1 (weight per weight). Digestion was carried out for six hours at 40 degrees Celsius. The liberated cyanogen bromide peptides, after separation from impurities by passage over a column (Bio-Gel P-2; BioRad, Richmond, California), were then isolated by gel-permeation high-performance liquid chromatography. A modification of a technique described by Harwood and Amiel was used to isolate and quantitate the cyanogen bromide peptides. The marker peptide, characteristic of type-II collagen, was established as a 1(II)CB10.
Total glycosaminoglycan content as hexosamine was determined with the method of Elson and Morgan, in which amino sugars are acetylated and measured colorimetrically by spectrophotometry. The results were expressed as the mean and the standard error of the mean in milligrams of hexosamine per gram of dry tissue. Normal human articular cartilage that was not used in the allografting procedure—that is, tissue from the lateral femoral condyle—was used as the control tissue.
The collagen concentration was 548 milligrams per gram of dry tissue, the glycosaminoglycan concentration was 34.8 milligrams of hexosamine per gram of dry tissue, and the percentage of type-II collagen in the articular cartilage of the allograft was more than 90 per cent. These concentrations are essentially the same as those found in normal human articular cartilage, suggesting viable chondrocytes. However, the friable articular host cartilage surrounding the allograft contained more collagen (648 milligrams per gram of dry tissue), a lower glycosaminoglycan concentration (26.7 milligrams of hexosamine per gram of dry tissue), and only 60 per cent type-II collagen, suggesting fibrocartilage.
We previously reported on fifty-nine knees that had been followed for a mean of forty-two months (range, twenty-four to 120 months) after treatment with an osteochondral allograft12. Fourteen knees (24 per cent) had been treated for a traumatic lesion; eighteen (31 per cent), for unicompartmental degenerative osteoarthrosis; twelve (20 per cent), for osteochondritis dissecans; ten (17 per cent), for chondromalacia of the patella; and five (8 per cent), for avascular necrosis. Three of the fourteen traumatic defects were in the femoral condyle, and three years postoperatively all three had been treated successfully. The remaining eleven knees had a traumatic lesion of the tibial plateau, and seven of them were available for evaluation; all seven had a good or excellent result at a mean of three years (range, two to five years). Since the initial report, we, as well as others17, have determined that osteochondral allografts are not indicated for the repair of defects secondary to localized osteoarthrosis. Subsequently, we reported on twelve knees that had been followed for an average of sixty-six months (range, sixty to seventy-two months) after the placement of an osteochondral allograft2. Eight of the nine procedures that had been performed for an isolated traumatic defect had a good or excellent result.
Of the first ninety articular cartilage allografts implanted by McDermott et al. in Toronto, forty-eight were inserted to reconstruct a traumatic defect; twenty-four were used to treat osteoarthrosis; fourteen, to treat avascular necrosis; and four, to treat osteochondritis dissecans. The mean duration of follow-up was four years (range, six months to thirteen years). Eleven of the twelve allografts implanted in the femoral condyle and twenty-five (69 per cent) of the thirty-six allografts implanted in the tibial plateau for the treatment of a traumatic defect were successful.
The preoperative diagnosis was osteochondritis dissecans in twelve of the fifty-nine knees reported on earlier by us12 but in only four of the ninety knees reported on by McDermott et al. Neither of these reports cited the specific indications for operative intervention.
The most extensive experience with the adult form of osteochondritis dissecans was reported by Garrett. He believed that "with craters larger than 2 cm (4 cm2), especially those in older individuals or in those who have failed to produce fibrous tissue following abrasion arthroplasty, pain, buckling, and swelling are common and arthritis with joint space narrowing occurs in an accelerated fashion." Garrett reported only one failure in forty patients who had osteochondritis dissecans and had been followed for one to six years.
In one study3, autoradiography with the use of sup 3Hydrogen-cytidine and sup 35S-sulphate was used to demonstrate the viability of the chondrocytes in biopsy specimens from four fresh osteochondral allografts that had been transplanted twelve, twenty-four, forty-one, and seventy-two months previously. One allograft placed in the lateral humeral condyle had united twelve months postoperatively and the surface of the joint had been maintained. The cellular viability was 99 per cent as demonstrated by the study with 3HH-cytidine and 96 per cent as shown by the study with 35S-sulphate. The osseous portion of another allograft, in the medial femoral condyle, fractured and collapsed twenty-four months postoperatively. Macroscopically, the articular cartilage appeared to be well preserved, and a second allograft procedure was done because of the fracture and collapse. Femoral and tibial osteotomies were also performed to restore alignment. The cellular viability was 69 per cent on autoradiography with 3H-cytidine and 78 per cent on that with 35S-sulphate. The third osteochondral allograft had been placed in both the medial tibial plateau and the medial femoral condyle. The biopsy was obtained forty-one months postoperatively, at the time of a corrective tibial osteotomy. The proportion of cells labeled with 3H-cytidine was 90 per cent. The allograft that was followed for seventy-two months had been used to reconstruct a traumatic defect of the lateral tibial plateau. At the time of the arthrotomy for removal of the screws, the articular surface of the allograft was found to be relatively well preserved, although degenerative changes were noted. The proportion of cells labeled with 3H-cytidine was 37 per cent.
The findings of the present study suggest that the use of a fresh osteochondral allograft to reconstruct defects of articular cartilage is a viable alternative to arthrodesis or prosthetic arthroplasty of the affected joint. The graft in our patient did not fail as a result of metabolic activity of the articular cartilage but rather as a result of resorption of the transplanted subchondral allograft bone. The operative technique used for this patient does not meet current technical standards. Specifically, at present, additional autogenous bone graft is placed beneath the allograft to fill the gap between the host bed and the allograft (Fig. 1). The supplemental use of autogenous bone graft is believed to enhance the incorporation of the osseous shell of the allograft and to decrease the amount of allograft cancellous bone needed to reconstitute a deep defect.
The gross appearance of the adjacent articular surface of the tibial plateau was normal in our patient. It did not demonstrate any evidence of chondromalacia, as is often noted in patients who have long-standing condylar defects. The absence of a reciprocal lesion suggests that the allograft had collapsed relatively recently and that it had remained functional until the collapse.
It seems self-evident that long-term survival and function of an osteochondral allograft depends on viable chondrocytes to replenish matrix components. This, however, may not be the case. Mankin et al. reported moderately successful clinical results with large frozen osteoarticular allografts implanted in 314 humans, mostly to treat tumors of the extremities10,11. In an effort to preserve the viability of the chondrocytes partially, these allografts had been treated with 10 per cent glycerol in Ringer lactate solution before freezing. Enneking and Mindell collected eight osteoarticular allografts that had been placed in humans to reconstruct osseous defects created by resection of a musculoskeletal tumor. The allografts had been in situ for six to sixty-five months (mean, sixteen months). Three of the specimens were retrieved from the distal aspect of the femur; three, from the proximal aspect of the tibia; and one each, from the distal and proximal aspects of the humerus. In seven of the eight specimens, the architecture of the matrix was well preserved but the cartilage was acellular and there was no evidence of histochemical metabolic activity. Major changes of degenerative osteoarthrosis were seen in only one specimen, whereas the other seven allografts showed only minimum erosions. The subchondral plate and adjacent trabecular bone remained necrotic and unrepaired. In the subchondral area, there was revascularization and trabecular resorption with major architectural changes. In the present study, there was no evidence of revascularization of the subchondral bone of the allograft.
It should be noted that a fresh osteochondral allograft placed in a precisely prepared bed of vascular subchondral bone in which the osseous shell is only five to six millimeters thick is not at all comparable with a massive osteoarticular allograft implanted in a limb-salvage procedure. Additional investigation is needed to determine whether an articular cartilage allograft, fresh or frozen, can function as a mechanical prosthetic replacement or if a viable, metabolically active transplant is needed for long-term preservation of the function of a joint. Although the allograft in our patient had viable chondrocytes ten years after transplantation, we have not been able to determine that living chondrocytes are necessary for a successful clinical result.