Arthroscopically assisted reconstruction of the anterior cruciate ligament (ACL) is a common and effective method for treatment of anterior knee instability following ACL injury. Annually, an estimated 250,000 ACL injuries are diagnosed and approximately 100,000 ACL reconstructions are performed in the United States1-3. Chondrolysis is an extremely rare but devastating complication following ACL reconstruction. It is characterized by complete destruction of the articular cartilage and a profound inflammatory response consisting of pain, swelling, and loss of joint motion. The differential diagnosis includes osteonecrosis, direct trauma related to the injury or to the surgical procedure, infection, an adverse reaction to pharmacological agents (including local anesthetic infusions), and unknown causes.
Infection after arthroscopic ACL reconstruction is rare: its incidence has been reported as 0.14% to 1.7%2,4-7. The most common causative organisms are Staphylococcus aureus, Staphylococcus epidermidis, and Staphylococcus haemolyticus2,4-6. When tissue allografts have been used for ACL reconstruction, other organisms such as Peptostreptococcus, Klebsiella, Enterobacter, and Clostridium species as well as gram-negative bacilli have been implicated6,8. Although infrequent, infection can be a devastating complication that leads to septic arthritis, chondrolysis, osteomyelitis, osteochondral destruction, and, in rare cases, death2,5,8,9.
We describe a case of medial tibial plateau osteochondral destruction secondary to an Aspergillus flavus fungal infection in an immunocompetent thirteen-year-old girl who had undergone elective ACL reconstruction. The patient and her parents were informed that data concerning the case would be submitted for publication, and they provided their consent.
To the best of our knowledge, a case of Aspergillus fungal arthritis following ACL reconstruction has not been previously reported. Furthermore, salvage options outlined in the literature have been limited to primary arthrodesis, prosthetic reconstruction, use of allograft-prosthesis composites, and use of nonarticular allograft void fillers9-11. We describe a salvage technique in which the medial tibial plateau, the associated articular cartilage, and the medial meniscus were successfully replaced with a fresh osteochondral/meniscus allograft.
A thirteen-year-old otherwise healthy female athlete sustained an acute ACL injury while playing soccer. Approximately three months after the injury, she underwent arthroscopically assisted ACL reconstruction with use of a quadruple hamstring autograft at a local outpatient surgery center. Radiographs made prior to the reconstruction revealed closure of the distal femoral and proximal tibial physes, and physeal sparing techniques were not employed for the reconstruction. The graft was fixed on the femoral side with use of the EndoButton (Smith & Nephew, Andover, Massachusetts) and distally with a large-fragment cortical screw and soft-tissue washer (Synthes, Paoli, Pennsylvania). The procedure was reported to be uneventful, and routine evaluation of the articular anatomy revealed healthy cartilage throughout the knee.
Postoperatively, the patient did not receive any type of local anesthetic-infusion device or pain pump. Although the immediate postoperative course was reported to be routine, the patient continued to have substantial medial-sided knee pain over the next two months. Rehabilitation was slow, and knee motion remained limited despite active physical therapy. On follow-up examination, the surgical incisions were seen to be healing well and there were no local or systemic signs of infection. Because of the persistent pain and poor knee motion, which was attributed to arthrofibrosis, a second-look arthroscopy was performed at the same surgery center three months after the initial reconstruction. During this procedure, substantial destruction of the medial tibial articular cartilage with associated damage to the medial meniscus was noted (Fig. 1). The remaining areas of the knee, including the lateral compartment and the patellofemoral compartment, were normal. No intraoperative culture specimens were obtained at that time. The underlying cause of the chondrolysis was undetermined, and the patient was referred to our institution for additional treatment.
In our clinic, the patient continued to complain of medial-sided joint pain. The incisions were well healed. Knee motion was 10° to 105°. All ligaments, including the ACL, were clinically intact. There was a small joint effusion but no substantial joint-line tenderness or erythema. Review of the arthroscopic and magnetic resonance images (MRI) confirmed degeneration of the medial proximal tibial articular cartilage with meniscal injury as well as bone edema in the medial tibial plateau (Fig. 2). The patient had remained afebrile without constitutional symptoms. The peripheral white blood-cell count was 6200/mm3. The differential diagnosis at that point included osteonecrosis of the tibial plateau, delayed degeneration of the articular cartilage due to the initial injury or surgery, and idiopathic chondrolysis. Infection was not considered strongly as part of the initial differential diagnosis. The initial impression was that the findings most likely represented a vascular insult to the medial tibial plateau in an aseptic environment. A fresh osteochondral hemi-tibial plateau/meniscus allograft was obtained to be used in an effort to restore the articular cartilage surface and medial meniscus.
A limited medial parapatellar approach revealed complete delamination of the medial tibial articular cartilage (Fig. 3-A) and degeneration of the medial meniscus. The autograft used in the ACL reconstruction appeared relatively intact, although it was degenerated within the tibial canal. The ACL autograft was removed to accommodate the hemi-tibial plateau allograft and was sent for culture and histological examination. The cortical screw and washer were removed from the proximal part of the tibia. Specimens were obtained from the joint and tibial tunnel for routine culture. No purulence was noted. At this point, the recipient site for the osteochondral allograft was prepared with use of unicompartmental knee-replacement cutting guides. The remaining native medial meniscus was resected. The fresh osteochondral allograft consisting of metaphyseal bone, articular cartilage, and the allograft meniscus was sutured to the knee capsule with standard suturing techniques (Fig. 3-B). The graft was secured anteriorly and anteromedially with use of four cancellous 3.0-mm cannulated screws (Synthes) (Fig. 3-C).
Postoperatively, within a four-day period, Aspergillus flavus grew in one of two cultures of specimens from the tibial canal (Fig. 4-A). To help interpret the importance of the positive culture, the direct-tissue Gram stain was restained for fungi (Calcofluor White stain), and rare fungal hyphae were identified (Fig. 4-B). On the basis of the rapid fungal growth in the culture and the findings of the Gram stain, this was considered to be a true fungal infection rather than specimen contamination. The pediatric infectious disease consultant recommended a combination therapy of intravenous (IV) voriconazole, 250 mg every twelve hours, and IV caspofungin, 100 mg every twenty-four hours, for two weeks, followed by twelve months of oral voriconazole, 200 mg twice a day. In addition, the patient returned to the operating room for irrigation and debridement of the wound as well as of the tibial canal. The tibial tunnel was thoroughly debrided, and a voriconazole-impregnated calcium-sulfate paste was introduced into the tunnel12. Two additional cultures of specimens from the second procedure were positive for Aspergillus flavus. Additionally, histological examination of the tibial tissue revealed polymorphonuclear neutrophils (PMNs) adjacent to bone fragments as well as a lake of PMNs between trabeculae. The postoperative rehabilitation protocol consisted of non-weight-bearing for twelve weeks. A hinged knee brace was placed and was locked in full extension for the first two weeks, followed by six weeks in which the knee brace was locked between 0° and 90°. Subsequently, and up to the three-month mark, the brace was unlocked with no restriction on flexion. During physical therapy, passive flexion was limited to 90° for the first six weeks.
At the twenty-eight-month follow-up visit, the patient demonstrated excellent, pain-free knee motion from full extension to 140° of flexion. She had been off all antifungal medications for over a year. She had returned to playing softball and was able to run without pain. Despite the ACL deficiency, examination showed a negative Lachman test, a negative pivot-shift test, a negative posterior-drawer test, and no joint-line opening with varus or valgus stress. Radiographs revealed osseous integration of the allograft with no evidence of osteolysis, periosteal reaction, or osteoarthritis (Fig. 5). The patient had no symptomatic knee instability but had not returned to cutting or pivoting sports. Her parents reported that she now ran slower than preoperatively. We counseled the patient and her family that she was at increased risk of developing osteoarthritis in the future. If the knee does develop instability, a revision ACL reconstruction could be considered, even with the medial tibial osteochondral allograft.