➢ More than 60% of the talar surface area consists of articular cartilage, thereby limiting the possible locations for vascular infiltration and leaving the talus vulnerable to osteonecrosis.
➢ Treatment strategies for talar osteonecrosis can be grouped into four categories: nonsurgical, surgical-joint sparing, surgical-salvage, and joint-sacrificing treatments. Nonoperative and joint-sparing treatments include restricted weight-bearing, patellar tendon-bearing braces, bone-grafting, extracorporeal shock wave therapy, internal implantation of a bone stimulator, core decompression, and vascularized or non-vascularized autograft, whereas joint-sacrificing or salvage procedures include talar replacement (partial or total) and arthrodesis.
➢ In patients with a Ficat and Arlet grade-I through III osteonecrosis, evidence in favor of a specific treatment is poor, although tibiotalar or tibiotalocalcaneal arthrodesis may represent a suitable salvage operation.
Talar osteonecrosis is a complicated and often frustrating condition to treat. Although the majority of talar osteonecrosis cases are traumatic, up to 25% of cases have atraumatic etiologies, including corticosteroid use, alcoholism, hyperlipidemia, irradiation, thrombophilia, and idiopathic etiologies. Outcomes remain suboptimal for advanced stages of talar osteonecrosis despite proper surgical technique and patient compliance. The purpose of this article is to discuss the work-up and management of a patient with talar osteonecrosis and to critically analyze the literature on surgical treatments.
A thoughtful work-up and treatment plan should be developed for patients with suspected talar osteonecrosis. A comprehensive history and physical examination in patients with pain and dysfunction over several months after talar trauma are essential in guiding a clinician’s diagnostic path. The quality and location of the patient’s pain must be identified. Patient comorbidities must be identified. Operative notes should be obtained, as they are insightful into understanding fixation methods and potential vascular insults during the initial trauma or fixation.
On physical examination, one must carefully inspect the healed incision sites, as these can help one to determine the previous surgical approaches. One must evaluate the motion at the tibiotalar and subtalar joints and compare it with the contralateral extremity. Using provocative maneuvers, one might find a restricted and painful range of motion if either joint is arthritic.
A thorough vascular examination is imperative and should take into account any signs of venostasis or vasculopathy. Often, the patient’s foot is perfused by only one artery. Poor distal blood flow may serve as a predictor for future healing and wound complications. Sensation of the foot must be tested as well.
The position of the foot should be noted, as the talus is often malunited in varus (if the patient had a talar neck fracture). Other important physical examination concerns are alignment and gait. The static and dynamic assessment of varus and valgus malalignment must be performed by observing the patient anteriorly and posteriorly.
All patients should have multiple weight-bearing radiographic views of the ankle. Radiographs may show evidence of talar sclerosis, malunion, nonunion, or collapse. The tibiotalar joint is carefully assessed for step-offs or evidence of arthritis. Additional ankle or foot views may help to visualize adjacent-joint degenerative changes or malalignment, which need to be taken into consideration for preoperative planning.
Radiographs may underestimate the articular damage or the quality of bone stock. Further information can be obtained via the Canale and Kelly1 oblique talus view, which will aid in determining the extent of talar neck varus angulation and shortening. This view is obtained with the foot in maximum equinus and 15° of pronation and the tube directed cephalad at a 75° angle from the horizontal. Computed tomographic (CT) scans are routinely used to evaluate the talus in the presence of the implant to assess fracture lines, the quality of the talar bone, and posttraumatic articular changes. All of this information becomes important in planning for the surgical approaches, techniques, and procedures required during the reconstruction. A magnetic resonance image (MRI) is only useful if the patient already has had the talar hardware removed, as imaging with a retained implant typically has too much artifact to be of any use except in cases of idiopathic osteonecrosis.
More than 60% of the talus is covered by articular cartilage, thereby limiting the possible locations for vascular infiltration and leaving the talus vulnerable to osteonecrosis. The vascular supply can be divided into extraosseous and intraosseous circulation2-4.
Extraosseous circulation loops around the talar neck and along the sinus tarsi, with contributions from the anterior tibial, posterior tibial, and peroneal vasculature. The tarsal canal artery courses through the sulcus tali, forming an anastomosis with the tarsal sinus artery. A complete anastomosis connecting all portions of the talus exists in only 60% of patients2,4.
The tarsal canal artery arises from the posterior tibial artery within the deltoid complex and supplies the central and lateral two-thirds of the talar body. The deltoid branches of the posterior tibial artery provide the remaining supply to the medial one-third of the body. The superomedial half of the talar neck is supplied by the anterior tibial artery. The inferior half of the neck is supplied primarily by the tarsal sinus artery, which is formed from contributions from both the anterior tibial and peroneal arteries. The dorsalis pedis artery or branches of the anastomotic loop have also been shown to supply the anterior half of the talar neck in certain individuals. The posterior tibial artery provides circulation to the posterior tubercle. Perforating peroneal artery branches, in combination with the posterior tubercle branches, supply the medial and lateral talar tubercles2-4.
Consequently, when one performs an open reduction and internal fixation of a talar body or neck fracture, one needs to ensure that the dorsal tissues are not stripped dorsally so as to preserve some of the extraosseous blood supply.
Etiology of Osteonecrosis
Osteonecrosis can be secondary to multiple etiologies, with the predominant characteristic of the disease being insufficient blood flow to the area in question and eventual death of the bone. Patients in the early (non-collapsed) stages of osteonecrosis may not be symptomatic, rendering the disease unlikely to be diagnosed until later stages unless a prior high clinical suspicion is present. The overarching goal of treatment is to halt progression of the disease or to delay end-stage arthritis. Joint-preserving procedures are often attempted for pre-collapse stages; however, after severe subchondral collapse has occurred, arthrodesis is often necessary to relieve symptoms.
Talar osteonecrosis can be caused by both atraumatic and traumatic causes5,6. Up to 25% of cases have atraumatic etiologies, including idiopathic etiologies, corticosteroid use, alcoholism, hyperlipidemia, irradiation, and thrombophilia5. However, the majority of cases are related to talar trauma, usually involving the body or neck. The risk of developing osteonecrosis after traumatic insult to the talus can be predicted by the Hawkins classification system, which is further described below7.
Talar Neck Fracture
Up to 50% of talar trauma involves talar neck fractures8. The typical mechanism of injury is hyperdorsiflexion of the foot with an applied axial load9. Increasing amounts of dorsiflexion place the posterior capsuloligamentous structures in tension, eventually leading to disruption. This may allow talar subluxation or even frank displacement. An associated inversion force may also result in medial talar neck comminution and varus deformity9.
The Hawkins classification system7 is the most commonly used schema to describe talar neck fractures; it was initially described by Hawkins and later modified by Canale and Kelly1. The classification type is based on the amount of displacement of the fracture as well as the number of dislocated peritalar joints; it is prognostic for future risk of talar osteonecrosis (Table I). In 2006, Ahmad and Raikin correlated the risk of talar osteonecrosis with each stage within the Hawkins classification system, with the risk of talar osteonecrosis being 0% to 13% for type I, 20% to 50% for type II, 75% to 100% for type III, and 100% for type IV10.
Talar Body Fracture
Talar body fractures are intra-articular and make up 7% to 38% of all talar fractures8. One of the first classification systems was anatomically based11, dividing body fractures into transchondral or osteochondral, coronal shear, sagittal shear, lateral process, or crush fractures. A simplified scheme divides these fractures into three groups: fractures of the talar body (or cleavage fractures), lateral process fractures, or compression fractures12. Much like talar neck fractures, the degree of comminution and displacement can be correlated with the future risk of osteonecrosis.
The mechanism of injury usually results from a high-energy axial load. The resultant fracture pattern depends on the foot position at the time of impact. Shear fractures occur with a hyperdorsiflexed foot, but lateral process fractures (or snowboarder’s fractures) involve an inverted and dorsiflexed ankle13. In one study of nineteen tarsal body fractures14, seven (37%) had evidence of osteonecrosis. In another study of nine talar body fractures15, 63% went on to osteonecrosis, with all patients developing subtalar arthritis (no patients developed ankle arthritis).
Current Treatment Options
Many treatment options exist for talar osteonecrosis, ranging from conservative to salvage. Treatment strategies for talar osteonecrosis can be grouped into four categories: nonsurgical, surgical-joint sparing, surgical-salvage, and joint-sacrificing treatments. Nonsurgical treatment includes restricted weight-bearing, patellar tendon-bearing braces, and extracorporeal shock wave therapy. Surgical joint-sparing treatment includes internal implantation of a bone stimulator16, vascularized autograft, and core decompression. Joint-sacrificing procedures include talar replacement (partial or total). Salvage treatment includes arthrodesis. Decisions are often made on the basis of surgeon comfort level and experience because the literature does not provide conclusive evidence of one treatment being better than another17.
Nonoperative Treatment for Talar Osteonecrosis
Nonoperative treatment includes non-weight-bearing, patellar tendon bracing, and extracorporeal shock wave therapy. Mindell et al. reported that 46% (six) of thirteen patients had excellent or good results and 54% had fair or poor results with non-weight-bearing18. Canale and Kelly treated twenty-three patients after talar neck fractures with either prolonged non-weight-bearing (more than nine months) or a patellar tendon brace1. They reported excellent or good results in 89% of patients, with no poor outcomes, for prolonged non-weight-bearing, and 33.3% good outcomes and 66.7% fair or poor outcomes for the patellar-bracing protocol described by Hawkins7. Hawkins reported that only 29% (seven) of twenty-four patients treated with non-weight-bearing had no pain at the latest office visit. He reported only 12.5% excellent or good outcomes for these patients7.
Zhai et al.19 reported on a prospective cohort of thirty-four patients who received either extracorporeal shock wave therapy or physical therapy. The researchers reported a significant improvement of the American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot score to 92.3 points (p < 0.05) in patients who underwent extracorporeal shock wave therapy compared with those who underwent physical therapy19.
Failure of nonoperative treatment is defined by the eventual need for a surgical procedure; often, treatment is a salvage operation: arthrodesis. In Hawkins’ series of nonoperative treatment, 58.3% later needed operative treatment, including 29.2% who underwent ankle or subtalar arthrodesis, 12.5% who underwent talectomies, and another 16.7% who underwent a bone-grafting procedure7. In the series by Canale and Kelly, 39% needed an additional surgical procedure (five tibiotalocalcaneal arthrodeses, two tibiotalar arthrodeses, and two talectomies)1. In the extracorporeal shock wave therapy series by Zhai et al., only 3% (one of thirty-four) needed arthrodesis19.
Multiple authors have shown that a period of non-weight-bearing results in good or fair clinical outcomes roughly half of the time, with approximately 40% to 45% of patients requiring eventual surgical intervention1,7,18,20. Patients who had a trial of non-weight-bearing longer than six months had better functional outcomes at the time of the latest follow-up with a decreased chance of talar dome collapse1,7,20. Risks of pursuing nonoperative treatment appear to be low because, to our knowledge, no study has shown that patients who were treated nonoperatively (without talar dome collapse) for a longer period of time had a poorer outcome when a surgical procedure was finally needed.
Core decompression may be an appropriate surgical treatment in patients with atraumatic talar osteonecrosis without talar dome collapse. Core decompression in Ficat and Arlet grade-I through III stages of osteonecrosis21 (Table II) has shown large improvements between preoperative and postoperative Mazur and AOFAS hindfoot scores following drilling, with >75% of patients showing excellent or good results6,22,23. After drilling, Delanois et al.6, in their study of thirty-seven ankles in twenty-four patients, and Marulanda et al.22, in their study of forty-four ankles in thirty-one patients, reported that less than one-third of patients (30%) had talar dome collapse or signs of arthritis in the tibiotalar joint. Marulanda et al. reported a significant improvement in the AOFAS hindfoot score from 42 to 88 points (p < 0.0001) in forty-four symptomatic osteonecrotic ankles after percutaneous core decompression22, and they used a 4-mm trocar to perform the core decompression.
In these studies, 11% required an arthrodesis of either the ankle or subtalar joint6,22,23. The researchers made note that age, sex, site of osteonecrosis, and prior percutaneous surgical intervention did not affect clinical outcomes.
Non-vascularized and vascularized bone-grafting have also been used to treat osteonecrosis. Based on current literature, vascularized and non-vascularized bone-grafting have excellent results with regard to function and pain scores24,25. Yu et al. utilized both techniques, a vascularized cuneiform bone flap in conjunction with non-vascularized iliac crest bone-grafting, in a series of twenty patients (30% female patients)24. Using the Kenwright outcome measure, 90% of patients had either excellent or good results24,26. Another series of seven patients underwent vascularized bone-grafting using femoral supracondylar corticoperiosteal grafts, and six of the seven patients had no subsequent dome collapse27,28. Zhang et al. reported their outcomes in twenty-four patients who underwent local vascularized cuneiform or cuboid bone-grafting. At the time of the latest follow-up, two-thirds of ankles had normal function and 83% had an excellent or good outcome25. No patients had dome collapse.
Talar Body Prosthesis for Talar Osteonecrosis
Two studies reviewed the results of implanting custom-made talar body prostheses after talar osteonecrosis. Harnroongroj et al.29,30 inserted a customized talar prosthesis into sixteen patients. Using a custom functional and radiographic outcome measure, 94% of patients with a minimum follow-up of five years and a maximum follow-up of fifteen years postoperatively had good results. In a recent update of their patients and others who had undergone a surgical procedure since the last publication30, thirty-three patients were available for follow-up (five patients had early failures) at a range of ten to thirty-six years. In these patients, twenty-eight prostheses (85%) were still in place. The AOFAS hindfoot score did not differ significantly among patients with varying levels of follow-up after ten years. The researchers noted that patients older than sixty-five years of age had significantly lower AOFAS scores (p = 0.001) compared with their younger counterparts. Failures occurred at eight to fifty-seven months and were treated with tibiotalocalcaneal arthrodesis in three patients, talar body revision in one patient, and below-the-knee amputation in one patient. Taniguchi et al.31 described twenty-two patients, with a mean age of sixty-six years, who underwent partial talar body replacement with a prosthesis. They evaluated the outcomes of both their first generation (stemmed, eight patients) and their second generation (unstemmed, fourteen patients). Both iterations of the talar body prosthesis had significant improvements (p < 0.001) in the AOFAS scores; the first-generation scores improved from 47 to 80 points, and the second-generation scores improved from 50 to 81 points. There were no significant differences between the two groups. The implant failure rate (conversion to total talar prostheses) was 25% for the first generation and 29% for the second generation. The researchers concluded that although a partial talar body prosthesis is not recommended because of high failure rates, a total talar prosthesis may be a surgical option for patients with talar osteonecrosis31.
Arthrodesis for Talar Osteonecrosis
Arthrodesis for treatment of talar osteonecrosis may be achieved by a variety of operative techniques, each with a unique set of outcomes and complications32-39. Studies of tibiotalocalcaneal arthrodesis in the setting of talar osteonecrosis, via retrograde intramedullary nailing or screw fixation, have reported good to excellent outcomes with osseous union in approximately 82% to 100% of patients (n = 74)33,34,36,38,39. Tenenbaum et al. recently reported on outcomes of retrograde tibiotalocalcaneal nailing in fourteen patients for talar osteonecrosis, showing excellent results of this procedure even in the setting of complete talar body loss, with 100% of patients going on to fusion38. Urquhart et al. also reported excellent outcomes, but with a mean time to radiographic union of 7.3 months in their series36. Common complications include nonunion, stable pseudarthrosis, and infection, with reoperation occurring in <10% of cases33,36,38.
Studies measuring outcomes of the modified Blair hindfoot fusion for talar osteonecrosis using the Kitaoka and Patzer outcome criteria demonstrated good to excellent results in more than 70% of patients (fourteen of nineteen)32,34. The original Blair procedure was described as a removal of the talus with the addition of a free graft augment to partially restore hindfoot height35. However, with the original Blair procedure demonstrating a 28% pseudarthrosis rate, modified techniques have been developed over time with the hope of improving overall fusion rates32,34,35,37. Although differing variations exist, implants are often used to fixate the newly introduced graft, decreasing motion and potentially increasing the rate of fusion. Time to radiographic union using the modified Blair technique appears to be comparable with fusion in non-osteonecrosis settings, with Van Bergeyk et al. reporting 71% fusion (five of seven) at four months37 and Lionberger et al. reporting 80% fusion (four of five) at three months35. It should be noted that union was not confirmed with a CT scan.
There was a high complication rate in the hindfoot arthrodesis studies, ranging from pin-site infections to below-the-knee amputations32-37. Specific complications include high rates of delayed union (32% [twelve of thirty-eight]), nonunion (19% [three of sixteen]), reoperation (15% [nine of sixty-two]), and infection (16% [ten of sixty-two])32-37.
Recently, arthroscopic arthrodesis has been proposed as a treatment therapy for talar osteonecrosis. Kendal et al. recently published a study on a series of sixteen patients undergoing arthroscopic ankle arthrodesis for non-collapsed talar osteonecrosis, with results demonstrating fusion in all fifteen available patients39. Three patients who reported continued pain each underwent a subtalar arthrodesis, which had a 100% union rate. The authors concluded that arthroscopic ankle arthrodesis is an appropriate treatment for advanced talar osteonecrosis.
Mindell et al. found that patients with talar dome collapse had a shorter time of non-weight-bearing (six months) compared with patients without collapse, who had a mean time of twelve months of non-weight-bearing18. Kitaoka and Patzer34 noted that, after the surgical procedure, younger patients (younger than thirty years of age) also had more excellent or good outcomes, but patients older than forty-four years of age tended to have fair or poor outcomes; significance was not discussed. Patient weight influenced outcomes, with patients with a mean weight of 76.5 kg having better outcomes than those with a mean weight of 91.3 kg (Spearman correlation coefficient, r = 0.65), which was significant.
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group40 guidelines were used to determine the quality and strength of recommended treatment strategies for talar osteonecrosis (Table III). The best study design was the prospective randomized controlled trial examining shock wave therapy19. The studies are hampered by study design because most are retrospective case series focusing on heterogeneous patients with nonvalidated subjective outcome scores. Reporting bias is extremely high as surgeons would not likely publish negative results from certain techniques. The evidence available for all four categories of treatment is considered very low using the GRADE recommendations; therefore, any estimate of effect is uncertain.
Unfortunately, given the very low grading for the evidence in the literature, there are no treatment strategies that have consensus. In our practice, we encourage non-weight-bearing for a period of four months if the talar dome has not collapsed. If the patient continues to have pain, we try either core decompression in the early stages of talar osteonecrosis or vascularized bone-grafting (from a local source) for those in the later stages. For those with isolated talar dome osteonecrosis, we have implanted a total ankle prosthesis in certain patients. In those with dome collapse, we advocate for early tibiotalar arthrodesis. For patients with subtalar joint involvement, we perform a tibiotalocalcaneal arthrodesis with a fusion rod.
Until better evidence is presented, each of the aforementioned treatment strategies likely has a role in the treatment of osteonecrosis.
A thirty-seven-year-old man without a pertinent medical history had sustained a talar neck fracture due to a high-velocity motor vehicle collision two years prior to presentation. The patient was initially treated with open reduction and internal fixation with two cannulated anterolateral-to-posteromedial screws. The patient was made non-weight-bearing for six months. The patient eventually was allowed to weight-bear, albeit with pain. On presentation, he had an antalgic gait and painful, restricted tibiotalar motion. Radiographs (Fig. 1) and CT imaging (Fig. 2) demonstrated subtalar arthritis, talar sclerosis, and nonunion of the previous fracture. The tibiotalar joint was relatively well preserved.
Given the failure of conservative treatment and evidence of talar osteonecrosis without dome collapse, the patient elected to undergo a subtalar arthrodesis. The goals of the surgical procedure were twofold: (1) to eliminate the arthritic subtalar joint, and (2) to attempt to partially revascularize the talar body with core decompression and creeping substitution from the calcaneus.
The patient demonstrated a full subtalar fusion by six months. The patient did well until twenty-four months after the subtalar arthrodesis. Follow-up radiographs (Fig. 3) and CT scans (Fig. 4) demonstrated partial talar head collapse and fragmentation and further degeneration of the ankle joint. At that time, the patient elected to undergo a trial of non-weight-bearing for four months, but eventually had a tibiotalocalcaneal arthrodesis.
At the time of the surgical procedure, the patient was placed prone and the ankle was approached posteriorly. The talar dome was resected because the quality of bone was poor. Next, a femoral head was placed as a bone block allograft. He had a retrograde intramedullary rod placed with posterior plate fixation and a supplemental screw that was placed from a proximal and medial direction to a distal and lateral direction.
At the most recent follow-up (ten months from the time of the surgical procedure), the patient had radiographs (Fig. 5) and CT scans (Fig. 6) that demonstrated full osseous union across the femoral head allograft. The patient was able to walk without pain.
Investigation performed at the Department of Orthopaedics, Medical University of South Carolina, Charleston, South Carolina
Disclosure: No external funding was utilized for the current study. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated