Osteochondritis dissecans (OCD) is a focal bone-cartilage lesion characterized by separation of an osteochondral fragment from the articular surface. The pathologic process involves a sequence of events, with progression from an initially intact lesion to a partially loose fragment to a totally dislocated fragment. Osteochondritis dissecans, in which an osteochondral fragment separates from a normal vascular bone bed, is quite different from osteonecrosis, in which the fragment separates from an avascular bone bed. Many etiologies have been proposed, including trauma or repetitive microtrauma and impingement of the tibial spine
1 , stress fracture with no specific injury
2 , and vascular insult
3,4 .
Over the years, osteochondritis dissecans has been treated in various ways, including surgical excision (as advocated more than 150 years ago
5 ), osteochondral autogenous grafting
6 , and allogeneic grafting
7-10 . The condition is more common in adolescents and young adults and is rare in patients less than ten years of age and in those more than fifty years of age
11,12 . OCD can be separated into a juvenile form and an adult form, with the onset occurring before and after the closure of the epiphyses, respectively. The most commonly affected joints are the knee, the elbow, and the ankle; in the knee, the lateral part of the medial femoral condyle is usually involved.
Treatment is usually nonoperative for intact lesions and operative for unstable lesions. It appears that a good clinical outcome is more likely when the femoral growth plate is open and when the lesion is small, occurs in a classic location, and is stable on magnetic resonance imaging or arthroscopic probing. When a cartilage fracture or articular defect is found on magnetic resonance imaging, the patient is likely to have a poor outcome. Arthroscopic assessment is of utmost importance in the choice of treatment.
The patient who presents with a loose fragment that is not suitable for fixation is better treated with an aggressive cartilage repair technique. Classic subchondral drilling can be performed arthroscopically, while allografting and multiple osteochondral cylinder grafting as well as perichondral and periosteal resurfacing usually need to be performed with an open technique. Since 1987, we have been using autologous chondrocyte transplantation (ACT) in combination with a periosteal cover to treat chondral and osteochondral defects of the knee
13-15 . In the current report, we review our experience with the use of autologous chondrocyte transplantation for the treatment of OCD in the fifteen years since the technique was introduced in Sweden.
Patients
Between October 1987 and April 2000, we performed autologous chondrocyte transplantation in 820 patients. Fifty-eight patients (7%) had OCD, with the first patient treated in 1990. During the follow-up period, the patients were evaluated with a questionnaire, clinical evaluation, radiography, and magnetic resonance imaging. Twenty-two patients (38%) consented to second-look arthroscopy for the evaluation of graft integrity (
Figs. 1-A ,
1-B ,
1-C ,
1-D , and
1-E ).
Chondrocyte Harvest and Culture Technique
Patients were treated in three hospitals by three different surgeons. Treatment was performed in two separate operations, as described previously
13 . Arthroscopic surgery was performed initially in a tourniquet-controlled bloodless field with the patient under general or spinal anesthesia. The defect was examined, and slivers of cartilage (300 to 500 mg) were obtained from the upper minor load-bearing area of the medial femoral condyle of the injured knee for cell culture and later transplantation. In the operating room, the cartilage samples were placed in a sterile glass tube containing 0.9% NaCl at ambient temperature; the samples subsequently were transferred to the cell-culture laboratory. Cell isolation was initiated not later than six hours after surgery. The cartilage pieces were minced and were washed twice in Ham's F-12 medium (Gibco BRL, Paisley, Scotland) supplemented with gentamicin sulfate (50 µg/mL), amphotericin B (2 µg/mL) and L-ascorbic acid (50 µg/mL). The minced cartilage was digested overnight (for sixteen to twenty hours) in a 25-cm
2 culture bottle (Costar, Cambridge, Massachusetts) containing 5 mL of Ham's F-12 medium supplemented as described above, clostridial collagenase (0.8 mg/mL, catalog no. C-9407, >1200 IU/mg; Sigma, Freehold, New Jersey) and deoxyribonuclease I (0.1 mg/mL, catalog no. D-5025; Sigma). The isolated cells were washed in Ham's F-12 medium once and were resuspended in 5 mL of culture medium containing Dulbecco's modified Eagle medium/F-12 1:1 (Gibco-BRL) with the addition of 10% of the patient's own serum and supplemented with gentamicin sulfate (50 µg/mL), amphotericin B (2 µg/mL), L-ascorbic acid (50 µg/mL) and L-glutamine (Gibco-BRL). The cells were incubated in 25-cm
2 tissue culture flasks (Falcon Primaria; Becton, Dickinson, Oxford, United Kingdom) in 7% CO2 in air at 37°C. After one week, the cells were trypsinized (trypzine-ethylenediaminetetraacetic acid 0.125%) and were transferred to a 75-cm
2 culture flask (Costar) with addition of the culture medium.
Implantation was performed fourteen to twenty-one days after the biopsy. Three days before transplantation, the cell cultures were subjected to quality-control procedures, consisting of sterility testing and photographic recording of cell morphology. Chondrocytes were trypsinized and were washed twice in Ham's F-12 medium containing 20% autologous serum (implantation medium). Cells were released for implantation if cell viability was >85% as determined with trypan-blue staining. Finally, the cells were suspended in 0.3 to 0.4 mL of implantation medium in a tuberculin syringe and were transferred to the surgery department in a sterile package. The average number of cells implanted was 5.2 × 10
6 in this patient group.
Surgical Procedure
At the time of chondrocyte transplantation, prophylactic antibiotics were given for twenty-four to forty-eight hours, beginning at the initiation of the procedure. With the patient under general or spinal anesthesia, a medial or lateral parapatellar arthrotomy was performed in a tourniquet-controlled bloodless field. The chondral lesion was débrided back to the best cartilage available. (The surrounding cartilage should have little or no fissuring, and a nerve-hook probe should be used to confirm that it is not undermined.) Care was taken not to provoke bleeding from the osseous bottom of the defect when débridement was done. A periosteal flap was harvested from the proximal-medial subcutaneous border of the tibia. The flap was fitted and sutured to the surrounding rim of the débrided cartilage with interrupted 5-0 or 6-0 Vicryl (Ethicon, Somerville, New Jersey) or Dexon (Sherwood, Davis and Geck, St. Louis, Missouri) sutures with the deep cambium layer facing toward the bone. The periosteal rim was sealed with fibrin glue (Tisseel; Immuno AG, Vienna, Austria) except in one corner, where the cultured chondrocytes were injected into the defect. Following cell injection, this corner was closed with a final suture and the application of fibrin sealant. The joint capsule, retinaculum, and skin were sutured in separate layers, and the knee was covered with an elastic bandage. Continuous passive motion was administered for forty-eight hours after surgery. Rehabilitation on crutches began with gradual weight-bearing for eight weeks, with progression to full weight-bearing by ten to twelve weeks. The emphasis during rehabilitation was on functional use of the limb and active muscle recruitment.
Bone-grafting was not done in the majority of the patients in this series (see Results and Discussion sections). In some patients, however, the defect was deeper than 8 to 10 mm and autologous bone grafting combined with ACT as a one-step procedure (i.e., the "sandwich technique") was performed. In most of these cases, the defect was excised to normal surrounding cartilage, and the sclerotic bone was removed down to bleeding cancellous bone. Drilling with a 2-mm burr into the bone-marrow cavity was used to increase bleeding. Cancellous bone from the iliac crest or from the proximal-lateral part of the tibia was harvested, depending on the volume of the defect. The osseous defect was filled with cancellous bone to the level of the subchondral bone plate. A periosteal flap the size of the defect then was harvested and was anchored, with the cambium layer facing the joint and the fibrous layer facing the bone graft, with use of 5-0 or 6-0 resorbable Vicryl sutures (Ethicon) in a horizontal or mattress fashion. Fibrin glue was applied between the periosteal flap and the bone graft, and the grafts were compressed with a sponge for sixty seconds. The tourniquet was then released, and the area was again compressed for two to three minutes to minimize bleeding from the bone-grafted area into the defect. A second periosteal flap the size of the cartilage defect was then harvested. The cambium layer was placed facing the defect, sutured, and glued, and cell implantation was carried out as described above.
Clinical Evaluation
The patient's clinical status was evaluated annually with use of five scoring systems: a modified Lysholm score
16 with a maximum score of 90, the modified Cincinnati (Noyes) knee score
17 , the overall Cincinnati knee-rating score
18 , the Wallgren-Tegner activity score
19 , and the overall Brittberg clinical grading score
13 . In addition, our Brittberg-Peterson functional assessment score
14 and a questionnaire regarding the patient's perception of the surgical outcome were used. Treatment efficacy was defined as the percentage of patients with a good or excellent result at two years who maintained this status for at least three more years. Statistical analysis was performed with use of the Stata software package (version 6.0; Stata, College Station, Texas). All reported values are two-tailed. The level of significance was set at p < 0.05.
Arthroscopic assessment of graft integrity was performed in accordance with earlier criteria
14 . The scoring system assigns a maximum of 12 points and is based on evaluation of the degree of defect repair (1 to 4 points), integration to border zones (1 to 4 points), and macroscopic appearance (1 to 4 points).
Magnetic resonance imaging was performed only in the second half of the study period in various radiology departments. The analysis was performed with the standard cartilage-imaging protocol. Fifteen patients were evaluated with magnetic resonance imaging both preoperatively and postoperatively (four patients) or postoperatively only (eleven patients).
Radiographs were made at the referring hospital both preoperatively and postoperatively. Due to the differences in radiographic techniques, varying projections were used and not all images were weight-bearing. Of the fifty-eight patients in the study, twenty-seven had both a preoperative and a two-year postoperative radiographic evaluation. The radiographs were examined by an unbiased observer who compared the preoperative and postoperative images with regard to osteophyte formation, joint-space narrowing, reconstitution of bone formation, cyst formation, sclerotic changes of the subchondral bone, flattening, and tibial spine prominence "sharpening."
We found that autologous chondrocyte transplantation is a promising technique for the treatment of OCD lesions of the femoral condyle, with >90% of the patients having a good or excellent clinical outcome after a mean duration of follow-up of 5.6 years. A good result at two years was maintained over time in fifty-two (98%) of fifty-three patients.
Since OCD lesions involve both the articular cartilage and the subchondral bone, we were hesitant to treat a large bone and cartilage defect with only articular chondrocytes. However, recent experiments have shown that several mesenchymal tissues have biological plasticity and are able to form cartilage, bone, and muscle
9,20-22 . The same phenomenon has now been demonstrated for articular cartilage cells, which, in the right environment, are able to form fat, bone, or cartilage
23 . One could thus hypothesize that the bone-cartilage defect will gradually be replaced by bone from the bottom by the transplanted cells, by osteoblasts in the defect, and by periosteal cells from the periosteal flap. Indeed, good filling of the defect by the subchondral bone was demonstrated on magnetic resonance imaging. Furthermore, in a number of patients, we observed radiographic signs of bone-remodeling over a five-year period, indicating that the newly formed cartilage is able to work as a functional unit with the subchondral bone to form a new subchondral bone plate. The fact that this phenomenon is observed over long-term follow-up is indicative of the slow turnover rate of the skeletal system.
The categorization of the disease process into juvenile and adult forms may be important at the time of diagnosis as the treatment algorithm and prognosis may differ depending on the stage of disease
24-27 . There is agreement that knees treated with removal of OCD fragments that are detached from the weight-bearing femoral condyles do poorly
28-30 . Treatment by removal of the fragments and drilling to promote fibrocartilage repair is also not recommended since the repair tissue is not durable
3,31,32 .
Linden noted that, at an average duration of clinical and radiographic follow-up of thirty-three years, thirty-eight of forty-eight patients who had had the first disease manifestation after closure of the physes (i.e., the adult form of the disease) had symptomatic and radiographic gonarthrosis
33,34 . The symptoms and radiographic signs became more frequent and approached a prevalence of 100% with time. In the present study, we found that narrowing of the joint space (noted in eight [30%] of twenty-seven patients;
Table III ) was the most important postoperative radiographic finding. These patients must be followed for long periods to assess future progression, as shown by Linden and Nilsson
33 . Seventy percent of our patients did not have narrowing of the joint space at the time of follow-up (
Table III ). However, the goals of treatment for these patients were relief of symptoms, improved functionality, and, lastly, halting the natural history of an empty defect (i.e., the development of osteoarthritis).
Patients with OCD usually are subjected to numerous surgical procedures and consume a large portion of health-care resources. We previously demonstrated that ACT provided substantial health economic benefits when used for the treatment of focal chondral lesions of the knee
35 , and we have no reason to believe that the benefits would be different when ACT is used for the treatment of OCD. Several of our patients who were treated with ACT were able to return to work and athletic activity within eighteen months after surgery.
Magnetic resonance imaging offers outstanding possibilities for examining the extent of pathological changes, especially the total extent of osseous involvement, in patients who have combined bone-cartilage lesions in association with OCD. The use of this method will facilitate the understanding of this complex problem and will lead to future improvements in surgical technique. Deep radical removal of the pathological sclerotic bone and bone-grafting of the defect along with ACT is a one-step procedure that may improve the results of treatment of these difficult pathological conditions. Magnetic resonance images, in contrast to radiographs, show not only the quality of the bone reconstruction but also that of the cartilage repair that may occur after ACT alone
36 . The overlying surface seems to consist of cartilage-like tissue on magnetic resonance imaging. In a limited number of patients, we found bone-remodeling with deep bone replacing the initially formed cartilaginous tissue. In contrast, the radiographic findings were too insensitive to demonstrate the minor changes that occur in the deep portions of the transplanted area. The cyst formations seen on radiographs were more evident on magnetic resonance imaging scans and were located in the deep zones of the repair tissue. Nine of the twelve patients with cyst formations had had previous drilling. The cysts did not communicate with the surface and were thought to represent fibrous tissue that had remained after the previous drilling.
Thus, magnetic resonance imaging is a valuable tool in the treatment of OCD both for the preoperative assessment of pathological changes as well as for the postoperative assessment of the characteristics of the repaired tissue over time. We recommend preoperative and postoperative magnetic resonance imaging for patients with OCD.
We believe that the two early failures that occurred in association with graft delamination were due to the early return to impact sports in the first two years postoperatively and that these graft delaminations occurred in deep defects. The repair of defects deeper than 8 to 10 mm may be enhanced by simultaneous bone-grafting and cartilage-grafting with ACT (the "sandwich technique").
ACT appears to be a reasonable alternative to osteochondral autografting for the treatment of osteochondral defects arising from OCD lesions, especially the adult form, that are associated with irrevocable damage or are not amenable to fixation.