Osteochondral lesions of the talus are a well-recognized source of ankle pain and instability. They are defined as a defect in the articular hyaline cartilage, predominantly within the weight-bearing area of the talar dome, and with involvement of the underlying bone. Talar involvement was originally described in 1922 by Kappis1, who identified a strong association with prior trauma. The initial classification of talar lesions was described by Berndt and Harty in 19592. In 2001, Scranton and McDermott added a fifth stage to the Berndt and Harty classification system in order to describe the cases of patients in whom the cartilage cap is intact with the lesion involving a subchondral cyst within the talar dome3. None of the stages described in these, or other, classification systems evaluates for the size of the lesion. In 2004, Raikin proposed a sixth stage in which massive lesions with volumes of >3000 mm3 (3 cm3) were present4. Fortunately, these massive-volume deep cystic lesions are rare, but they pose a very challenging treatment dilemma for the surgeon.
Surgical intervention options for lesions that have failed to respond to nonoperative treatment include arthroscopic débridement with microfracture (marrow stimulation)5-7, osteochondral autograft transfer systems with bone harvested from the ipsilateral knee (with use of single plug or multiple small-plug mosaicplasty techniques)8-10, and laboratory-grown articular chondrocyte implantation11,12. None of these options has, however, been described to treat very large lesions.
The use of an osteochondral allograft for talar lesions has been reported, but most studies have described a plug technique for the treatment of smaller lesions13. In a preliminary investigation, the short-term follow-up results of a series of bulk osteochondral allograft replacements for large lesions with use of either fresh or fresh-frozen matched allograft were described4. That study was limited by small numbers, short-term follow-up, and a mixture of graft preservation techniques.
The current institutional review board-approved study evaluated patients with a minimum follow-up interval of two years and examined a larger cohort of patients in whom a more homogeneous fresh allograft technique was used to treat this challenging pathology.
Between 2000 and 2006, twenty-two consecutive patients with massive osteochondral lesions of the talar dome were evaluated for possible bulk allograft transplantation. All had an osteochondral lesion within the talar dome with a volume (as assessed on magnetic resonance imaging and/or computed tomographic evaluation) of =3 cm3 (=3000 mm3). Patients were not considered for transplantation if they had an age of greater than sixty years, an unfused distal tibial physis, a body mass index of >35 kg/m2 (obesity class II or above, according to the World Health Organization classification system14), or involvement of the tibial side of the ankle (ankle joint arthritis), or if they had a cyst that closely approximated or involved the subtalar joint. On the basis of those criteria, four patients were excluded and subsequently underwent tibiotalocalcaneal (two patients) or pantalar fusion (two patients). The remaining eighteen patients underwent bulk osteochondral transplantation of the talar dome. In two of the eighteen patients, a fresh-frozen allograft was used and these two patients were also excluded from the study. A twelve-year-old child with an acute talar body fracture, but open distal tibial physes, was also excluded.
The fifteen remaining patients were treated with fresh bulk osteochondral allografts and were included in this study. There were ten male and five female patients. The average age of the patients was 41.9 years (range, seventeen to fifty-six years). The left ankle was involved in six patients and the right ankle, in nine patients. The medial talar dome was involved in twelve patients, while the lateral dome was involved in three patients. Ten of the patients had a known history of trauma.
Nine patients had at least one prior ankle operation. Four patients had two prior operations, and two had three. These included open reduction and internal fixation of an ankle fracture in one patient and a talar fracture in two patients (two of these three patients subsequently had the hardware removed), eighteen ankle arthroscopies, and two attempted osteochondral autograft transfers from the ipsilateral knee.
Prospectively collected data included a preoperative visual analog scale score for pain (with use of a 10-cm graded line, with 0 indicating no pain and 10, the worst pain imaginable) and the American Orthopaedic Foot and Ankle Society (AOFAS) ankle-hindfoot score. These scores were obtained preoperatively for all patients. In addition, standard weight-bearing ankle radiographs were made for all patients, and three-plane computed tomography scans (eleven patients) (Figs. 1-A and 1-B) and/or routine protocol magnetic resonance imaging scans without contrast enhancement (eight patients) (Figs. 2-A and 2-B) were acquired. Sizing was obtained from these radiographs, with use of the computed tomography scan when available, and from the T1-weighted magnetic resonance imaging sequence (combining axial, coronal, and sagittal plane measurements) when no computed tomography scan had been performed. The average volume of the lesions was 6059 mm3 (range, 3003 to 10,120 mm3), and the average depth (caudal to cranial) was 16 mm (the shallowest depth was 11 mm). The patient demographics are summarized in Table I.
Fresh talar allografts were matched for side and size with use of radiographs with a marker of a known size to assess magnification, and they were obtained from an American Association of Tissue Banks-accredited bone bank. All allografts and donors were tested for infectious diseases as per standard transplantation protocol15,16. All allografts were implanted within sixteen days (range, eleven to sixteen days) of harvest and were preserved at 4°C to optimize cartilage cellular viability17. No cross-matching or human leukocyte antigen (HLA) tissue typing was performed for any patient, and no antirejection medication was used in any patient.
All patients were available for follow-up, which included an interview, clinical evaluation, and radiographic assessment. The average duration of follow-up was fifty-four months (range, twenty-six to eighty-eight months). All patients underwent a subjective satisfaction assessment (they were asked to rate their own outcome as either excellent, good, fair, or poor), evaluation with use of the AOFAS ankle-hindfoot scoring system (the scores are subdivided into pain, function, and motion-stability subsections) as well as a visual analog scale score for pain, and radiographic analysis (radiographs were reviewed for arthritic changes and graft stability and/or subsidence). Failure was defined as the need for conversion to arthrodesis or revision surgery. Follow-up scores for patients whose graft had failed were obtained immediately prior to the arthrodesis. Patients were additionally asked if they would undergo the same procedure again.
Statistical analysis was performed with use of a standard paired t test by an independent statistician.
Surgical Technique
Surgery was performed on an outpatient basis, with all patients placed under general anesthesia and a popliteal block used for postoperative pain control. Surgery was performed in a hospital operating-room area to allow potential admission, if needed, but all patients were able to go home the same day as the procedure and none required readmission. Routine perioperative prophylactic antibiotics were given to all patients. No other special perioperative precautions were utilized. As most patients were young and active and none had specific risk factors for deep vein thrombosis, anticoagulation prophylaxis was not routinely utilized.
Surgery was performed through a medial malleolar osteotomy in four patients, through a distal fibular osteotomy in one patient, and through an anterior approach in ten patients to allow optimal visualization of the specific lesion (Fig. 3). Ankles approached through an anterior incision underwent resection of half of the talar dome (Fig. 4), while those approached through a malleolar osteotomy underwent a more location-specific osteochondral resection in order to preserve the uninvolved cartilage. The specific approach was determined by the size and location of the lesion as directed by the preoperative computed tomography or magnetic resonance imaging scan.
The resected defect was measured, and a template that matched the resected portion of the talus was created and outlined on the allograft. A matching portion was then harvested from the donor talus (Fig. 5), taking care to protect the cartilage cap of the graft. Any cystic defects remaining in the bed of the talus following resection of the lesion were impaction grafted with cancellous bone harvested from the talar allograft. The osteochondral graft was then inserted into the defect to recreate the morphology of the talar dome (Fig. 6) and congruency of the ankle joint. Rigid fixation of the graft was achieved with headless titanium compression screws (DePuy, a Johnson and Johnson company, Warsaw, Indiana). The recreation of the talar architecture was observed clinically and under image intensification guidance. Achieving clinical congruency was thought to be more accurate than achieving radiographic congruency, as articular cartilage thickness may vary between the donor and host, leaving an apparent step-off on the radiographs at the subchondral bone line. This step-off was measured on immediate postoperative radiographs in order to provide a baseline for assessment of subsidence on follow-up studies (Figs. 7-A and 7-B). Malleolar osteotomies, if required, were fixed with titanium hardware to allow future magnetic resonance imaging examination.
Postoperatively, patients remained non-weight-bearing for ten to twelve weeks. Once the surgical wounds had healed, usually two weeks after surgery, active and passive sagittal plane range of motion was allowed. Formal physical therapy was instituted at six weeks, and progressive protected weight-bearing was begun at ten to twelve weeks, depending on radiographic evidence of healing. Use of a fracture boot was continued until cross-trabeculation between the graft and host talus was seen radiographically. This occurred at an average of 18.5 weeks (range, sixteen to twenty-six weeks) following surgery.
Source of Funding
No external source of funding was obtained or utilized for this study.
All fifteen patients returned for follow-up evaluation at an average of fifty-four months after surgery. Two patients subsequently underwent an arthrodesis of the ankle to treat graft collapse and the development of arthritis after thirty-two months and seventy-six months, respectively. For both patients, a standard ankle arthrodesis was utilized, and there was no need for bulk bone-grafting or a tibiotalocalcaneal arthrodesis. Only the scores obtained for these patients before the arthrodesis are included in the results.
The preoperative AOFAS ankle-hindfoot scores averaged 38 points (range, 24 to 59 points) of a possible 100 points. These scores (including those obtained from two patients before they had an arthrodesis) improved postoperatively to an average of 83 points (range, 54 to 97 points), with an average improvement of 45 points (p < 0.05). When component scores were examined, there was very little change in the range of motion, stability, or alignment subsections, while the greatest changes were seen in the pain, activity, and walking distance subsections of the score.
The preoperative pain score on the visual analog scale averaged 8.5 (range, 8 to 10), although three patients on their own had extended the pain line and received scores of >10. These scores improved to an average of 3.3 (range, 1 to 8), representing an average improvement of 5.2 (p < 0.05).
Radiographic analysis revealed some evidence of collapse or resorption of the graft in ten of the fifteen ankles. Additionally, nine ankles demonstrated some narrowing of the joint space overlying the graft area (Table I). With the small number of subjects, an association between the radiographic findings and the clinical results could not be identified. Of the two patients who required arthrodesis, one had extensive graft collapse and the other had narrowing of the joint develop, consistent with cartilage degeneration (Fig. 8).
Patient satisfaction with the outcome was rated as excellent by five patients, good by six, fair by two, and poor by the two patients who required an ankle arthrodesis. Despite these results, all fifteen patients (including the two patients who ultimately underwent ankle arthrodesis) were pleased that they had undergone the allograft procedure and would choose to do it again.
A recent study of the treatment of osteochondral lesions of the talar dome has demonstrated that outcomes following arthroscopic management are indirectly proportional to lesion size and lesions with a surface diameter of >1.5 cm tend to have inferior outcomes18. Osteochondral autograft transfer has been suggested as an alternative to arthroscopic microfracture in these patients, but the volume of graft that can be safely harvested from the patient's knee is limited and this procedure has been associated with persistent morbidity in an otherwise healthy knee19,20. While these problems have been specifically documented with arthrotomy and harvest from the lateral femoral condyle, and not in studies utilizing arthroscopic harvest21, similar morbidity might be expected with treatment of larger lesions of the talus, given the very large autologous harvest that would be required. Articular chondrocyte implantation may be a better alternative for lesions with a larger surface area, but this procedure requires a stable osseous bed for the chondrocytes to grow on and is not typically described for a large, deep, cystic type of lesion.
Ankle arthrodesis is another option, but this procedure requires the sacrifice of ankle motion. Furthermore, the deficiencies in bone stock within the cystic area limit the possibilities for screw fixation, as well as the surface area for fusion, making this procedure technically difficult.
Osteochondral allografts have been used for many years in reconstruction following periarticular tumor resection22-24 and for treatment of large defects of the articular surfaces of the knee25-29. The advantages of this technique are the ability to supply mature hyaline cartilage and to restore the anatomic architecture in acquired periarticular defects, without the morbidity associated with the harvest from remote anatomic sites. As with all osteochondral allografts, one concern is the viability of the articular cartilage being transplanted. Previous studies have demonstrated that freezing articular cartilage in order to decrease immunogenicity, and extend storage life, in fact results in poor chondrocyte viability30-35. Studies on fresh osteochondral allografts (stored at 2° to 4°C), however, have demonstrated that chondrocytes remain viable for up to fourteen days following harvest, after which cellular metabolism and biomechanical properties deteriorate significantly17. In an animal model, chondrocytes from similarly preserved allografts that were implanted after twenty-one days showed more severe degenerative changes than those stored for less than twenty-one days36. In a clinical series involving fresh osteochondral allograft transplantation into the human knee, good results were demonstrated with grafts stored for up to forty-two days, but the study size was too small to associate outcomes with graft storage time37.
The use of fresh osteochondral allografts in the knee has shown encouraging medium to long-term results25-28, with ten-year estimated survival rates of 85% for femoral lesions described in a study by Gross et al.38.
Currently, the process of obtaining fresh allograft is challenging. First, an appropriately matched donor must be obtained. Because of the highly congruent nature of the ankle joint, the donor talus must be matched for both side and size. This is done by sending a magnification standardized radiograph (usually a coin with a known size is taped to the ankle, allowing magnification to be calculated) to one of the few tissue banks that supply fresh osteochondral tissue. The matched talus is preferably from a young (less than forty-year-old) donor to ensure minimal cartilaginous wear and must be defect-free on examination. The graft must first be screened for a multitude of potentially infectious conditions as outlined by the American Association of Tissue Banks (which usually takes seven to fourteen days), while being maintained at a temperature of 2° to 4°C. Once the graft is cleared for transplantation and matched, the patient is contacted and implantation is recommended as soon as possible to maximize chondrocyte viability. Our protocol has been to have patients be available by cellular telephone during the waiting period (which averaged four to five months) and to be brought in for surgery on the day the allograft is due to arrive at the hospital.
The potential need for HLA genotype cross-matching of patients and donors is of unknown importance. Meehan et al.39 and Phipatanakul et al.40, in separate studies, demonstrated the presence of HLA antibodies in a majority of patients undergoing osteochondral allograft transplantation. The relevance of this immune response is currently unknown, and we did not perform HLA typing and matching or administer postoperative immunosuppressants to our patients.
Prior studies of fresh osteochondral allograft transplantation to the talus are very limited. The knee literature has a longer and more detailed history, with survival rates reported up to twenty-five years following allograft implantation41,42. Shasha et al. reported a twenty-year survival rate of 66%43. In a large retrieval study from knee bulk osteochondral allografts implanted one to twenty-five years previously, Gross et al. reported that early failure is usually due to lack of chondrocyte viability, while late failure is a result of graft fracture, resorption, or collapse42. They concluded that long-term survival depends on achieving graft stability by rigid fixation with a stable osseous base, which is consistent with our technique of rigid screw fixation of the graft to a stable talar base.
Jeng et al. reported on twenty-nine patients who had total ankle osteochondral allograft replacement and were followed for a mean duration of two years44. Only 31% (nine ankles) were classified as having a successful result, with fourteen ankles undergoing revision or conversion to either fusion or lateral ankle replacement and six others demonstrating fracture or collapse radiographically in patients who were awaiting conversion at the time of the study. In a similar study in which the Agility Total Ankle System (DePuy) cutting jig was used to perform total ankle allograft replacement in eleven ankles, Meehan et al. reported that six ankles were considered to have had a successful result at thirty-three months of follow-up39. They concluded that while this procedure offers an alternative to arthrodesis in young patients, bipolar allograft replacement of the ankle has a high rate of failure.
In a study of nine isolated osteochondral allograft implantations on the talar side for the treatment of Berndt and Harty grade-IV lesions, Gross et al. reported that six grafts survived an average of eleven years13. These ankles had lesions with a surface area diameter of at least 1 cm and a depth of >5 mm, but the actual volume and the presence or absence of a cyst were not reported. Additionally, follow-up was performed by telephone interview only, with no clinical or radiographic evaluation.
The current study specifically evaluated large-volume cystic talar lesions in which resurfacing autologous articular chondrocyte implantation or allograft replacement could not be considered. All were unipolar talar lesions in which there was no apparent compromise of the tibial articular surface. This may be why the results in the present study are better than the total ankle allograft replacements discussed above.
This study has some limitations. The number of patients is small because of the uncommon nature of this condition, and the study was performed as a retrospective review of prospectively collected data obtained at the time of the initial evaluation. An additional potential limitation of the study is that patients were assessed with the AOFAS scoring system, which, while in common use, is not a validated outcome measure.
In conclusion, large-volume cystic osteochondral defects of the talus are difficult to treat. The present study offers a surgical alternative that utilizes rigidly fixed, bulk fresh osteochondral allograft transplantation, and this technique demonstrated an 87% survival rate and a high level of patient satisfaction, with 73% of patients reporting good or excellent results after a minimum follow-up of two years (average, 4.5 years). However, the long-term survival of these allografts remains unknown. Studies of ankle arthroplasty and tibiotalar allograft replacement with a minimum follow-up of greater than five years have demonstrated worse outcomes with increasing time, and additional failures of these allografts can be expected in the future. 