The patient was a thirteen-year-old girl with juvenile rheumatoid arthritis who had a history of chronic prednisone treatment in excess of 60 mg per day, but who at presentation had been managed with Enbrel (etanercept), methotrexate, and Mobic (meloxicam). The patient had multiple concomitant polyarthropathies, predominantly involving the knees and the ankles, and had undergone two previous arthroscopic debridements for osteonecrosis of the left talus.
In October 2007, the patient presented to a pediatric hospital with intractable right knee pain, swelling, and difficulty with walking, all of which worsened with activity and improved with use of anti-inflammatory and pain medications and/or with rest. Physical examination of the right knee revealed an effusion, anterolateral joint-line tenderness, and decreased motion. Preoperative magnetic resonance imaging (MRI) demonstrated a 1-cm × 2-cm lesion within the centrolateral aspect of the lateral tibial epiphysis with subchondral collapse and desiccation of the overlying articular cartilage (Fig. 1). The patient underwent arthroscopy with partial synovectomy, biopsy, and allograft bone-grafting of the tibial lesion through a transepiphyseal approach. During this procedure, the articular surface was intact, and a subchondral cystic cavity was noted. Biopsy of the lesion demonstrated necrotic bone with reactive changes. After six weeks of non-weight-bearing, the patient was allowed to progress from non-weight-bearing to partial weight-bearing and then to walking.
Within three months postoperatively, the pain had returned. A computed tomography (CT) scan demonstrated a focal subarticular lucent lesion in the lateral tibial plateau, with focal subchondral collapse (Fig. 2). Six months after nonoperative management, a repeat CT scan demonstrated no improvement. A contralateral osteonecrotic lesion in the ankle that required operative intervention delayed treatment of the right knee.
Six months after the index procedure, the patient underwent repeat arthroscopy of the right knee, which demonstrated full-thickness cartilage loss at the site of the osteonecrotic tibial lesion communicating with the subchondral cyst. The unstable articular flaps were debrided, but repeat bone-grafting was not performed.
The patient was referred to the senior author (C.J.W.) in July 2008 for consultation and treatment. At the time of presentation, the patient complained of pain with walking and had subjective complaints of knee "instability," continuous knee pain, and an inability to comfortably perform her activities of daily living or athletics. In addition to taking her disease-modifying and anti-inflammatory arthritis medications, she was taking 20 mg of OxyContin (sustained-release oxycodone) twice daily, with 5 to 10 mg of short-acting oxycodone as necessary. The Lysholm score, at 26 of 100, was poor. On physical examination, the patient had an antalgic gait, full knee motion but with tenderness over the anterolateral joint line, normal lower-extremity alignment, and no ligamentous instability.
She had severe pain that required the continuous use of crutches for walking. Nonoperative treatments were discussed but not recommended. A "drawer-access osteotomy" to allow osteochondral allografting was devised to reconstruct the cartilage defect, decrease symptoms, and improve function.
The preoperative surgical approach and technique was initially done on a cadaver, which confirmed our opinion that this procedure was a reasonable salvage procedure for this patient (Fig. 3).
With the patient under general anesthesia, an examination confirmed that there was ligamentous stability in all planes. A diagnostic arthroscopy demonstrated severe loss of cartilage on the lateral tibial plateau and communication with the underlying cystic cavity (Fig. 4, A). On the lateral femoral condyle, there was a 10-mm area of intact but soft cartilage, which was treated with antegrade subchondral drilling.
Through a hockey-stick incision centered over the lateral joint line, the dissection was made through the iliotibial band to the tibial plateau. The lateral meniscus was tagged and directed superiorly for a submeniscal approach to the joint. A vertical corticotomy was made just lateral to the anterior cruciate and posterior cruciate ligament insertions and was connected with a horizontal corticotomy 1.5 cm distal to the plateau (Fig. 3, B). The osteotomized tibial plateau was removed en bloc for allograft reconstruction (Fig. 3, C and D;Fig. 4, D and E).
The osteonecrotic lesion measured approximately 20 mm in diameter (Fig. 4, B). With use of osteochondral autograft transfer-system (OATS; Arthrex, Naples, Florida) instrumentation, the lesion was drilled to a depth of between 10 and 15 mm, leaving a healthy bone surface with the exception of a small cyst (Fig. 4, D), which was curetted and filled with autograft bone. The donor fresh-frozen allograft was marked at the same region as the host area of osteonecrosis. A 20-mm cylindrical osteocartilaginous core was harvested with use of OATS instrumentation. The edges were measured and modified to fit accurately into the host plateau. The core was placed with excellent press fit and re-creation of the tibial plateau geometry (Fig. 4, E). The tibial plateau fragment was reduced into position and stabilized with cortical and cancellous screws placed in lag mode (Fig. 4, F). There were no intraoperative complications.
The patient was placed on a reduced dosage of disease-modifying and immunosuppressive medications that were continued for the initial two weeks after the operation. The patient was started on continuous passive motion forty-eight hours after surgery and remained non-weight-bearing for six weeks. At six weeks postoperatively, the patient was no longer taking narcotic pain medication and began graduated weight-bearing with crutches. Knee motion was 0° to 90°. Twelve weeks after the operation, the patient had full knee motion and required no narcotic pain medications, but radiographs showed incomplete incorporation of the graft. The patient was instructed to continue full weight-bearing but to use a varus unloading brace for six months to protect the graft. She had no complaints with regard to the knee, despite slow graft incorporation (Fig. 5, A). As of the twenty-month follow-up, the patient has normal knee motion, uses no brace, and walks without pain. Radiographs demonstrated complete-interval graft incorporation (Fig. 5, B) with minimal joint-space narrowing. The Lysholm score was 81 (fair) at that time, as compared with 26 (poor) preoperatively. The patient reported that her limitations in her Lysholm score were related to ankle pain from her avascular necrosis in that location. Her only knee symptom was a rare, painless mechanical clicking, and she rated her satisfaction as "excellent."
Chondral injuries in young active patients are a challenge to treat2. Although trauma and sports-related injuries are the most common causes of chondral injury in children and adolescents6, steroid-associated osteonecrosis can also be the cause of chondral and subchondral bone injury. The exact mechanism of osteonecrosis is unclear, but the femoral head and distal part of the femur are the two most common sites3. Epiphyseal lesions commonly lead to subchondral fracture and overlying chondrosis, progressing to joint collapse and secondary arthritis7.
Steroid and/or narcotic pain medications are often used to treat the severe pain and disability that can result from chondral injury. Anderton and Helm reported that prednisone given in dosages of less than 15 or 20 mg per day resulted in very low risk of the development of osteonecrosis8. Additional data support a cumulative threshold dosage of 500 mg, above which patients are placed at a higher risk of the development of osteonecrosis9,10. Our patient received dosages in excess of 60 mg per day, with an unknown total dosage.
Treatment options for osteochondral defects are somewhat limited and there is little consensus on a therapeutic algorithm due to the relatively low incidence and the lack of Level-I or II evidence to help in making clinical decisions. The mainstay of nonoperative management is activity modification and medical management3, with attempts made to avoid large dosages of steroids.
If nonoperative management is unsuccessful, surgery may be indicated. Arthroscopic debridement, tibial osteotomy, osteochondral grafting, resurfacing techniques, and partial or total knee replacement are options for surgical treatment11-14. All have been reported to be efficacious in the treatment of knee osteochondral defects. However, marrow-stimulation techniques, such as drilling and microfracture, are not indicated in the setting of subchondral bone disease and this is possibly an explanation for our patient's unsuccessful index procedures.
Most reports address the treatment of distal femoral lesions, which are more common than tibial ones5. Osteochondral allografts have been associated with a greater than 75% success rate in treating focal femoral condylar lesions that are the result of chondral injury, trauma, osteochondritis dissecans, and osteonecrotic lesions15. Görtz et al. reported on twenty-eight knees in which osteonecrotic lesions were treated with osteochondral allografts, with all knees achieving radiographic union without a change in graft position or graft fracture. This large series supports the use of fresh allografts for the treatment of osteonecrosis of the knee16.
Osteochondral allografting to the tibial plateau is more technically demanding, as the plateau is partially obscured by the overlying meniscus and femoral articular surface. Shasha et al. reported on fresh osteochondral allograft reconstruction of the entire tibial plateau in patients with large, posttraumatic osteoarticular degenerative changes. In this study, 68% of grafts were intact at an average follow-up of twelve years, with 95%, 80%, 65%, and 46% survivorship at five, ten, fifteen, and twenty years, respectively17. Ueblacker and Burkart reported on retrograde osteoarticular transplantation to the central tibial plateau in five patients (four open approaches, one arthroscopic); all of those patients were satisfied with the results at follow-up times that ranged from six to greater than thirty-five months4.
The goals of treatment in a patient like ours are to restore or recreate a durable articular surface with hyaline cartilage, to provide structural subchondral support during host growth and remodeling, and to reestablish full and painless motion of the knee during normal weight-bearing. It was our opinion that a bulk allograft (such as that described by Shasha et al.) would replace a large portion of this patient's unaffected tibial plateau with allograft bone, which would necessitate bulk healing by creeping substitution and expose our patient to the risk of catastrophic collapse of the allograft. As a result, a novel, submeniscal "drawer-exposure" osteotomy was employed to allow temporary global access to the diseased tibial plateau. With proper placement of retractors, the lateral tibial plateau was safely osteotomized, leaving all of the soft-tissue stabilizers (the anterior cruciate ligament, the posterior cruciate ligament, the lateral collateral ligament, and the posterolateral corner) intact and tension free during the procedure. Use of a size-matched fresh-frozen allograft allowed re-creation of the lateral tibial surface with near anatomic accuracy. The reconstructed plateau was immediately and anatomically reduced and secured with use of standard lag-screw fixation techniques. Immunosuppressive medications can represent a challenge with regard to healing of the osteotomy, vascular incorporation of the allograft, or possible infection, but these risks were accepted because of the patient's underlying disease. A two-week holiday from high dosages of immunosuppressive medications was employed during the early postoperative period.
Kocher et al. have demonstrated that the Lysholm knee score is both a valid and a reliable rating system for chondral injuries18. In this case report, our patient improved from a score of 26 preoperatively to 81 postoperatively, which represents a dramatic reduction in pain and improvement in function. Our goal in the publication of this case report is to present a novel technique that may be useful in the treatment of incomplete osteochondral lesions of the tibial plateau. While such cases are infrequent, it appears that this method is a viable option for further clinical study. Further refinements in the technical aspects of the allografting procedure, as well as further understanding of the biology of osteochondral allografts, should lead to improved clinical outcomes.