Secondary Logo

Journal Logo


Ankle Fusion in Patients With Hindfoot Arthritis


Author Information
Techniques in Foot & Ankle Surgery: September 2002 - Volume 1 - Issue 1 - p 34-45
  • Free



Although the ankle and subtalar joints are two distinct articulations, it is not appropriate to think of them as independent. In the German language, the subtalar joint is called the lower ankle. They are two components of a universal joint, allowing the leg to remain vertical under the body while the foot accommodates to uneven terrain. Disease in one joint will increase stresses in the other, with the possibility of secondary degeneration. Perhaps the classic example is the patient with a severely arthritic ankle. As time passes, stiffness in the ankle increases stresses across the subtalar joint, and such a patient may present with advanced degeneration of both joints.

Patients can develop concomitant disease in the ankle and hindfoot joints for a variety of reasons. In modern times, most patients with disease of both ankle and subtalar joints suffer from post-traumatic arthritis. 1,2 Avascular necrosis of the talus, with collapse of the body, can also lead to degeneration of both joints. Historically, though, combined ankle and hindfoot disease was more commonly seen with neuromuscular disorders. 3–5 Muscle imbalance after poliomyelitis left many people with disabling deformities of both joints. Although polio is now quite rare in developed regions, the modern equivalent is Charcot-Marie-Tooth (CMT) disease. When surgery for secondary arthritis is indicated, realignment and multiple joint fusion can be performed successfully.

Combined ankle and hindfoot arthrodeses are sometimes thought of as salvage procedures, rather than reconstructions. 1,2 While we would disagree with that statement, extensive fusion does result in a foot that is extremely rigid. Some patients with the transverse tarsal joints left free retain a mild amount of flexion and extension, but in all cases the vast majority of motion is lost. 1 A stiff foot causes difficulty climbing hills or uneven terrain, and eventually takes its toll on the ipsilateral knee. Of course, many patients presenting with joint disease have already lost motion, and surgical treatment can take a foot that is painful and stiff in a bad position and make one that is less painful in a more functional position. This should be the goal of all combined hindfoot arthrodeses.

Obtaining precise limb alignment is essential, so that weight-bearing forces are distributed evenly and appropriately throughout the foot. Limb alignment includes static factors, such as the positions of bones and joints, and dynamic factors, mainly muscle balance. While appropriate alignment is essential to any foot or ankle procedure, it is never more important than in cases of combined fusions, where the lack of hindfoot motion will prevent the foot from compensating at all for any imbalance. Historically, good alignment was not always obtained in the years before rigid internal fixation was routinely used. Casts and staples do not have the holding power of lag screws, and both malunion and nonunion were not uncommon. 4,5 And though it is easy to blame poor fixation for poor results, lack of attention to alignment by the surgeon may have been an important factor historically as well.

Malunion of a combined fusion leads to inappropriate concentration of forces on relatively small areas of the foot, resulting in pain, callousing, and skin breakdown. The heel must be in neutral or slight valgus, and the metatarsal heads must be level and parallel to the floor. The toes should be straight and free of extensor contracture, so they can bear some weight as well. Most importantly, muscles must be balanced so as not to deform the remaining joints. Even when joints are fused in ideal positions, persistent muscle imbalance in a neuromuscular disease will recreate the original deformity through any remaining joints. Slight varus malpositioning at the ankle or supination of the subtalar joint will increase pressure on the lateral heel, and the base and head of the fifth metatarsal. Similarly, valgus will lead to inappropriate pressure on the first metatarsal, or in cases of medial column instability, pressure on the talar head, navicular tuberosity, or medial cuneiform. Extensor clawing of the toes will increase pressure on the metatarsal heads, both by decreasing the overall surface area for weight bearing and by pulling the plantar forefoot pad of skin distally, uncovering the metatarsal heads.

Although combined ankle and hindfoot fusions have been performed for many years, new developments in surgical technique continue to appear in the literature. For uncertain reasons, recent trends have favored retrograde intramedullary nails for combined fusions. 1,6 Theoretical benefits include a percutaneous technique and strong fixation. Unfortunately, these benefits are not realized in practice. Though the incision for nail insertion is small, neurovascular injury is possible because several major structures cross the insertion site. 7 Because the nail has no intrinsic ability to correct deformity, a percutaneous approach is not appropriate for any case with malalignment. Major deformity requires open exposures of the joints to correct, negating any benefit of a small incision for nail insertion. Furthermore, although the nail is quite strong, it relies on one or two interlock screws in the talus and calcaneus to hold alignment, which theoretically is not adequate. In vitro testing has assessed stiffness and load to failure of constructs, 8,9 but failure in vivo occurs through the bone under cyclic loading, making the in vitro studies less relevant. In osteoporotic bone, the weak link in fixation is the bone, so a rigid nail with one or two points of fixation is inferior to the multiple points of fixation achieved by several large lag screws across each joint. There are also multiple reports of stress reaction or fracture just proximal to the nail end. 10 Although there are rare cases when an intramedullary nail is useful across the tibia, talus, and calcaneus, we do no find these nails to offer any advantage for most combined fusions.


The basic indication for combined fusions of ankle and hindfoot joints is pain and limitation of activities due to deformity or degeneration of these joints. Combined fusions are most beneficial for patients with a stiff, painful foot. Successful surgery will result in a foot that is stiff, well aligned, and less painful. Because most will have significant limitations of motion preoperatively, these patients will be more willing to accept the rigidity of fusion. Thus, a well-aligned combined arthrodesis is useful for pain relief. For the rare patient with joint degeneration but good motion, the result may not be as gratifying. This patient will trade motion for pain relief, and because pain relief is rarely complete, it will be more difficult to meet expectations.

Perhaps the most common etiology of combined joint degeneration is following intra-articular or peri-articular fracture. Patients with post-traumatic arthritis of ankle and subtalar joints may have little or no malalignment, especially if previous reduction has been performed. Treatment should begin with activity modification and pain medications, but by the time they present to a foot and ankle surgeon, most already have major limitations of activity and poor response to nonnarcotic pain medications. Bracing with an ankle-foot orthosis can be attempted, but we have not been satisfied with bracing for advanced disease. Bracing may limit motion but does not decrease weight-bearing forces across the joint. When joint surfaces have degenerated, it is often the weight bearing, rather than motion across a joint, that seems to cause pain. Furthermore, braces are expensive and not covered by many insurance plans. Failing these nonoperative modalities, patients with post-traumatic arthritis are good candidates for combined fusions.

There is a distinct subset of patients with post-traumatic degeneration of the ankle and the hindfoot who present with poor alignment. These are patients who sustained tibia fractures 10 to 30 years ago. Typically, the fracture was treated with casting and healed in varus, hyperextension, and internal rotation. (Fig. 1) Especially when the fracture is in the distal shaft, patients seem to be able to compensate for the deformity through the ankle and subtalar joints. Arthritis develops in these joints presumably from uneven stresses over many years. In this case, treatment should begin with tibial osteotomy to restore alignment. Impinging osteophytes can be resected from ankle or subtalar joint as needed, but fusion should not be rushed into. With restoration of alignment, some joint symptoms will abate, and ankle or hindfoot fusion can be delayed for months to years in some patients. (Unfortunately, other patients will present with advanced disease, and realignment alone will not be sufficient.)

FIG. 1.
FIG. 1.:
Malalignment above the ankle. (A, B) This 40 year old man sustained a closed tibia fracture which healed in varus, hyperextension, and internal rotation. Although he could partially compensate for the deformity by everting through his subtalar joint, he developed pain because of excessive weight-bearing through the lateral foot. (C, D) A single cut oblique osteotomy corrected the varus and rotation well. There is still some hyperextension at the ankle, but this should pose no problem. With his alignment corrected, stresses about foot and ankle are more normal, and late arthritis should be less of a concern.

Avascular necrosis of the talar body, especially after trauma, will have pain as a chief complaint. As in other post-traumatic patients, alignment is often reasonable. If the ankle and subtalar joint surfaces are well preserved, surgical treatment should begin with reduction and fixation of a talar neck nonunion (if present), or drilling and bone grafting of the body. Unfortunately, the subtalar joint is frequently degenerative, and in this case, a subtalar fusion can be performed for pain relief and also to stimulate blood flow to the body (from the calcaneus). For the patient with collapse of the body and degeneration of both ankle and subtalar joint, combined fusion is appropriate. Usually, the talar head and neck, as well as the talonavicular joint, remain viable. A large posterior iliac crest graft is used to replace the necrotic talar body.

Patients with neuromuscular disease will have more significant joint deformity than degeneration. The typical patient with CMT will have lost anterior and lateral leg muscles, and they will develop a cavovarus deformity. Initially, the deformity is passively correctable. It is at this stage that tendon transfers and bracing should be performed to maintain alignment and function. As the disease progresses in some cases, though, a rigid deformity sets in. Bracing with an ankle-foot orthosis will help accommodate the deformity somewhat, but pain, callousing, and skin breakdown limit activity. For the fixed deformity, realignment and multiple joint fusion may be the only way to get a patient out of a wheelchair.

The deformity of CMT, or any other neuromuscular disease, arises out of muscle imbalance, and this imbalance must be addressed during any reconstruction. The chronic deformity also leads to soft tissue contractures on the concave side (medial). In some cases (particularly advanced deformity in the older patient), skin becomes thin and tight over the foot. Toes are often clawed and contracted, with poor soft tissue covering. These feet are some of the most complex to reconstruct, with major joint deformities, muscle imbalance, and soft tissue contractures.


Planning begins with outpatient assessment. Reconstruction is an elective endeavor. Nonoperative modalities, including activity modification and anti-inflammatory medication may allow surgery to be delayed for many patients. However, there are a few patients for whom surgery should not be delayed. As mentioned above, the patient with a malunion of a tibia fracture should have alignment restored sooner rather than later. By normalizing forces across the ankle and subtalar joints, the progression to osteoarthritis should be slowed. Also, the patient with a progressive neuromuscular disease should consider surgery early, especially if bracing is not controlling the deformity. Early muscle balancing may prevent the late complications of tendon rupture (from chronic imbalance) and skin breakdown (which can lead to osteomyelitis and even amputation in severe cases).

Once the decision for reconstruction has been made, a surgical plan must be devised. The goal of preoperative planning is to determine which joints are painful, where deformity lies, and which muscles are contributing to deformity. Pain can come from degenerative joints or from uneven distribution of weight bearing forces. Malalignment may arise from a bone (such as with a malunion) or from a joint. Muscle imbalance can be static (and easily visible while the patient sits on the examining table) or dynamic (more readily visible during gait). All these questions must be answered prior to beginning the surgical procedure.

Physical examination begins with an assessment of alignment. The limb is viewed from the front, back, and side during stance and gait. Any deformity above the ankle is noted. The position of the heel is assessed for varus or valgus. The pattern of callousing on the sole of the foot will provide information as well. The normal foot should have mild amounts of callous under the heel and across all the metatarsal heads. Any areas with excessive callous are noted, as is the absence of callous. A foot whose skin is soft as a “baby's bottom” throughout suggests that person is not doing much weight bearing.

Muscle strength and excursion is assessed in anterior tibialis, posterior tibialis, both peroneals, toe extensors and flexors, and of course in the gastrocnemius-soleus complex. Posterior tibialis is checked with the ankle plantarflexed, so as not to allow the anterior tibialis to assist. Overpull of the peroneus longus (evidenced by excessive plantarflexion of the first metatarsal during active ankle flexion) must be noted as well. Such overpull will contribute to forefoot-driven hindfoot varus. If the patient dorsiflexes both the toes and ankle when asked to dorsiflex only the ankle, that may be a sign of extensor recruitment, which can contribute to clawing of the toes. After assessing static muscle strength, dynamic forces are observed during gait. Although rare, an overactive posterior tibialis can pull the foot into varus. Range of motion in each joint should be examined as well. Limitation of motion can be from degenerative changes with osteophytes, or from contracted muscles or joints.

In cases where it is not clear exactly from where pain is originating, diagnostic lidocaine injections may be helpful. Because the joints of the foot and ankle are in such close proximity to each other, it is easy for patient or surgeon to be unsure exactly which joint is the source of pain. A small injection of a local anesthetic into the ankle or subtalar joint is given, and response in the office is noted. These injections are best done with a small amount of contrast, to document exactly where the injection went with an arthrogram. It is possible to attempt injection into the subtalar joint but actually be in the lateral gutter of the ankle, for example. Of course, degenerative osteophytes and hindfoot deformity will alter the normal anatomy, making it more difficult to know exactly where the needle is.

Weight bearing X-rays of ankle and foot are used for surgical planning. Mortise radiographs will reveal any tilting of the talus within the mortise. Axial radiographs can be used to visualize alignment of the heel when deformity is suspected there. With hindfoot deformity, the subtalar joint becomes less visible on plain radiographs. Occasionally, a CT scan is useful for assessing the condition of this joint. Tibial X-rays are taken whenever deformity is noted there. MRI or bone scans are not used.


The basic surgical plan is the same for any kind of combined hindfoot fusion. First, alignment must be restored. Cartilage and necrotic bone are removed from joints to be fused so that healthy bone is exposed. Then rigid internal fixation is achieved, and bone graft is added. Muscle balancing is performed, and the limb is assessed for alignment. Although the young surgeon might focus on the internal fixation as the central portion of the procedure, the experienced surgeon knows that success depends primarily on restoring alignment, preparing a good fusion bed, and attaining good soft tissue closure.

The patient can be positioned laterally or supine with a bump under the ipsilateral hip. A transfibular approach under tourniquet provides exposure to ankle and subtalar joints. A longitudinal incision is made over the distal 10 centimeters or so of the fibula, and it may curve anteriorly past the tip of the lateral malleolus. (Fig. 2) Subperiosteal dissection is performed over the anterior fibula, and the tibial-fibular and talofibular ligaments are transected. The fibula is cut with an oscillating saw a few centimeters above the ankle joint, and a second cut removes a 1 centimeter segment of fibula. The fibula is then rolled back on its posterior soft tissue sleeve to expose the joints. (Fig. 3) The same saw can be used to resect the medial cortex of the distal fibula. At the end of the procedure, the fibula will be rolled back into its anatomic position, bridging across the talus and tibia. By leaving some of the soft tissue attachments on the fibula, it may act as a vascularized bone graft for the fusion. We used to resect about 25% of the thickness of the fibula, but now we leave it as thick as possible to best preserve the normal anatomy in case of later reconstruction (see the final section on future of the technique).

FIG. 2.
FIG. 2.:
(A) A lateral incision along the distal fibula provides excellent access to the ankle and subtalar joints. (B) A small anteromedial exposure will allow preparation of the medial gutter of the ankle for fusion.
FIG. 3.
FIG. 3.:
In this intra-operative photo, the fibula is rolled back on its soft tissue sleeve, exposing the ankle and the posterior facet of the subtalar joint.

Any remaining cartilage is removed so that subchondral bone is exposed on distal tibia and talus, including the lateral talar body. Osteophyte can be removed for use as bone graft. The posterior facet of the subtalar joint can be denuded as well. A 2 mm drill perforates the bones in multiple locations to bring vascular supply through the dense subchondral bone. The medial gutter of the ankle is then prepared through a second small incision over the anteromedial ankle. The saphenous nerve and vein are protected, and any cartilage is removed. Although not always necessary, we find it useful to make this small medial approach, as it allows the surgeon to be sure the talus is translated over to the medial malleolus (and not left too lateral). Talar translation is not an issue in well-aligned ankles, but in cases with cavovarus deformity, it is necessary to position the talus properly to minimize tension on tissues (see below). Meticulous preparation of bone surfaces is essential for successful fusion. Both joints are then positioned in preparation for internal fixation.

Many cases of ankle arthritis feature anterior extrusion of the talus, with loss of bone from the anterior tibial plafond. The ankle must not be fused with any anterior translation, as this will increase the lever arm of the stiff foot on the remaining mobile joints (transverse tarsal or midfoot), and may accelerate the development of arthritis in those joints. The talus must be positioned back under the tibia, and the subtalar joint should be neutral. Temporary fixation may be placed across the joints while assessing alignment. A stout K wire can be placed from the anterolateral tibia into the lateral talus and calcaneus so that it will not be in the way of lag screws. Alternately, if an assistant is available, a 4.5 mm lag screw can be placed in the same location.

Alignment is then assessed. The talus must be back under the tibia, and the foot should be neutral (neither dorsi-nor plantarflexed). The heel must be in very slight valgus, and the metatarsal heads should be even for weight bearing. An ankle that has collapsed into valgus may have developed secondary forefoot supination to compensate. Or, the chronically flat foot will have an ankle that is secondarily tilted into valgus because of loss of medial column support. In all of these valgus ankles, the first metatarsal head will be elevated once the talus has been restored to a neutral position for fusion, so that weight bearing would only occur under the lateral forefoot. This must be corrected by flexing and fusing one or more medial column joints. Usually, fusion of the first tarsometatarsal joint in a reduced position will position the metatarsal heads appropriately. In rare cases, correction may be needed through the naviculocuneiform or even the talonavicular joints. The key principle is to restore normal anatomy, rather than to compensate for one deformity by creating another. In cases with complex deformities, intra-operative radiographs with provisional fixation will help confirm restoration of normal structure.

Rigid internal fixation is then performed. Each joint must be stabilized by at least two lag screws. A single screw will allow rotation, and cannot confer rigid fixation. If an assistant is available, cannulated screws are not necessary. First, the talus is stabilized to the distal tibia. Through a small stab incision just lateral to the Achilles tendon and several centimeters above the plafond, the first screw is introduced. Blunt dissection protects the sural nerve and peroneal tendons. This screw passes from posterolateral tibia across the ankle and into the middle of the talar neck. As the partially threaded 6.5 mm screw is tightened, the provisional fixation is loosened, so that the talus will be lagged posteriorly in the plafond. Alignment is reassessed, and then the remainder of the screws are placed. Typically, we place one screw through a stab incision in the heel, passing from calcaneal tuberosity, across the talus, and into the anterior cortex of the distal tibia. Another may be placed from plantar calcaneus up through the talus and into the tibia. A screw can pass from just above the medial malleolus through the center of the subtalar joint and into the calcaneus. At least two screws must cross each joint. (Fig. 4)

FIG. 4.
FIG. 4.:
Post-traumatic arthritis. (A) This 35 year old man developed pain and deformity in his foot many years after an injury. The ankle and subtalar joints are sclerotic with little to no joint space. (B) The ankle has tilted into valgus. (C) Realignment and fusion were performed through a trans-fibular approach. Once the hindfoot was re-aligned, the first ray was elevated (supinated), so a first tarsometatarsal alignment and fusion was performed through a dorsal foot incision. (D) All fusions healed, and the patient was fully weight-bearing at three months post-op.

Lag screw fixation can be augmented with a lateral blade plate when bone quality is poor. The blade is inserted into the calcaneus and multiple screws are placed through the plate into the calcaneus, talus, and tibia. Before inserting the blade, enough bone must be resected from the lateral foot and ankle to allow the plate to lie flush. If the plate is proud of the bone, the skin will not close.

Usually, sufficient osteophyte has been resected to use as bone graft. The graft is placed into spaces around the joints as needed. If there are large defects, bone may be harvested from proximal or distal tibia, from the distal femur, or from the iliac crest. When good bony apposition is achieved between fusion surfaces, not much bone graft is needed. The fibula is replaced anatomically and held with two 3.5 or 4.5 mm lag screws (one into the tibia and one into the talus). The tourniquet is deflated before closure for hemostasis. Deep tissues are reapproximated with absorbable suture, and skin is closed without tension.

Muscle balancing can be achieved with transfers or lengthenings as needed on a case-by-case basis. Although the patient with post-traumatic deformity may not have major hindfoot imbalance, clawing of the toes is common. In order for the toes to participate in weight bearing, they must be restored to a more normal position. In cases of extensor recruitment, we begin with extensor substitutions, in a variation of the classic Hibbs' procedure. Through a dorsolateral foot incision, the long toe extensors are transected over the metatarsal necks, and the proximal tendon is attached to the peroneus tertius to support the forefoot. The short toe extensors are transected more distally, and the distal stump of long tendon is sewn to the proximal short extensor, so that the overall effect is a lengthening and weakening. Dorsal metatarsophalangeal capsulotomies can be performed as necessary. More extensive releases are required when clawing has arisen from an untreated deep posterior compartment syndrome.


The typical case is the rigid cavovarus deformity of CMT. (Fig. 5) The same lateral transfibular approach exposes the joints, but patients with CMT typically have contracted tissues on the medial side of the foot and ankle. Reduction of the foot will place too much tension on the skin, with necrosis the result. Lateral wedges of bone, especially from distal tibia, should be removed to effectively shorten the limb as the foot is reduced. The small anteromedial incision is essential in this deformity, to release the deltoid ligament as needed for reduction, and also to ensure the talus is brought medially against the malleolus. Any lateral translation will put excess pressure on medial skin, impairing healing. Fixation proceeds with lag screws. The resected wedges of bone provide more than enough graft.

FIG. 5.
FIG. 5.:
Major deformity. (A, B) This 65 year old woman has had deformity from CMT disease for many years. Her foot is in a cavovarus position, and the ankle is tilted in the mortise. Her skin is tight over the medial foot and ankle. (C) Alignment was restored by resecting a laterally-based wedge of bone from distal tibia. Unfortunately, the wedge did not reach all the way to the medial side, so the talus ended up translated laterally. Although the wounds were easily closed without tension, medial skin necrosed after surgery. (D) The lateral view shows good healing of fusions at three months after surgery. Fully threaded screws were used to hold the positions of talus and calcaneus during healing. In general, if there is no bone loss, partially-threaded lag screws are usually preferable. (E) Once the hindfoot was corrected, the medial ray was too plantarflexed, so a fusion of the first tarsometarsal joint was performed after resection of a dorsal wedge of joint. That fusion was held with a thin dorsal 2.7mm plate.

Cavovarus deformities may require release, reduction, and stabilization of the talonavicular joint to get the forefoot out of supination and into a plantigrade position. A medial utility incision (just above and parallel to the course of the posterior tibial tendon) allows exposure of the talonavicular joint. The anterior facet of the subtalar joint can be denuded through this approach as well. In cases where the talus is exposed medially and laterally, the surgeon must be careful to protect its blood supply. The deltoid should be left intact if possible, and tissues on the dorsal neck must not be disturbed. The talonavicular capsule may need a formal release, as well as the plantar fascia through a small plantar medial incision. Two screws from the medial navicular into the talar body provide fixation for the joint.

Muscle balancing is essential in these feet. By the time advanced deformity sets in, the posterior tibialis is the only foot muscle left with any function. If it is not transferred, it will pull the naviculocuneiform joint into varus. The tendon is released from its insertion and routed behind the tibia into the peroneals, or through the interosseus membrane and into the dorsolateral midfoot. It may or may not act as a weak everting force, but more importantly it will no longer be an invertor.

The toes are often clawed, with tight skin and poor vascularity. Although extensor substitutions and intrinsicplasties are desirable in theory, the toes may not survive the dissection needed for tendon balancing. Because of poor skin vascularity, we are cautious about correcting toes in patients with CMT.


The necrotic talar body will not readily fuse, so it must be resected and replaced with healthy bone. The patient is placed laterally, and the posterior iliac crest is prepared for graft harvesting. A medial distractor is placed, with a pin in the tibial shaft and a second in the calcaneal tuberosity (posterior to the neurovascular structures). A longitudinal incision is made just lateral to the Achilles tendon, and subcutaneous dissection is performed down to the fascia over the deep posterior compartment. The sural nerve must be protected. The fascia is opened, and the flexor hallucis longus is retracted medially, exposing the posterior talus. Ankle and subtalar capsulotomies are aggressively performed, and the distractor opens the joints until normal height is restored. Appropriate alignment of hindfoot and ankle must be restored, and the distractor can hold this position. All necrotic talus is resected, usually leaving only head and neck. A block of posterior iliac crest is harvested, making sure it is large enough to fill the space between tibia and calcaneus. Cancellous bone is first packed in the gap, and then the block of bone is impacted into place. The distractor is relaxed a bit, so that the block of bone is compressed gently between calcaneus and tibia.

Fixation begins with a partially threaded 6.5 mm lag screw from posterolateral tibia into the talar neck, as usual. A fully threaded 6.5 mm screw is placed from the heel across the graft and into the anterior tibia, and then a second fully threaded screw is placed across both joints. These fully threaded position screws maintain the height of the hindfoot during healing (Fig. 6).

FIG. 6.
FIG. 6.:
Avascular necrosis of the talus. (A, B) This 44 year old woman had a painful nonunion of a subtalar fusion, which was performed for avascular necrosis. The mortise view shows collapse of the talar body. Not surprisingly, the talonavicular joint remains relatively healthy. (C, D) The talar body was resected from a posterior approach and replaced with posterior iliac crest graft. Fully-threaded screws preserve height while healing. The talonavicular joint was spared.


The limb is splinted for about 3 weeks, and then sutures are removed. A short leg cast is used for immobilization for a total of 7 weeks, during which time weight of leg weight bearing is allowed. Radiographs are checked for signs of healing, and then weight bearing is increased to full over the next few weeks. The patient should be fully weight bearing by 3 months after surgery. In cases where a block of structural bone graft has been used, weight bearing is delayed for 10 to 12 weeks to allow incorporation.

Once a patient resumes ambulation, a rocker bottom shoe is recommended to compensate for the lack of ankle motion. The front and back of the shoe may be beveled off, or a separate rocker bottom may be attached to the sole of the shoe. Our hope is that use of this rocker will delay the progression of arthritis in the remaining mobile joints.


Nonunion may occur, especially when necrotic bone is present. Several studies of combined fusions using a variety of techniques have found union rates of 85% or higher. 1,2,11–13 It is probably reasonable to say that, with modern techniques of fixation, 10 to 20% of all patients will develop a nonunion in one or more joints. The rate will be higher in smokers or in any patient with avascular bone. Because this is such a heterogeneous population of patients, union rates will vary even for one surgeon with one technique. The best treatment is prevention, with meticulous intra-operative technique to remove all nonviable bone (and any remaining cartilage), thus leaving a good fusion bed. When nonunion does occur, if the patient is ambulating comfortably with little or no symptoms, observation is appropriate. Even if screws break, there is no need to revise the fusion unless alignment is lost or pain is a problem. Revision is accomplished by adding bone graft and usually revising fixation (unless the surgeon is absolutely certain that fixation remains rigid).

Malunion is a serious problem, because the remaining mobile joints cannot accommodate any malalignment. Rates of malunion are not well reported in the literature. Considering the difficulty in obtaining perfect alignment during surgery, we suspect mild malunions are relatively common. Mild degrees of malunion may be well tolerated by the patient, especially if a custom orthotic can be used to unweight the areas of the foot that are taking excess pressure. Unfortunately, an orthotic that is raised medially to accommodate a varus foot will usually cause the foot to slide laterally, increasing the pressure on the lateral side of the foot. This dilemma limits the utility of an orthotic for varus or valgus malunion. It is best to get alignment perfect the first time, and if that is not done, then revision (with osteotomy through malunion) is necessary.

Wound healing problems will be much more likely if tissues are mishandled during surgery. Tension can be decreased on the skin closure with the use of absorbable subcutaneous stitches to bring the edges together. In the deformed foot with contracted skin, failure to shorten the foot will lead to stretching of skin on the concave side of the deformity, and may lead to necrosis. In most cases of severe preoperative deformity, it is best to shorten the limb by resecting wedges of bone from the convex side of the deformity. Skin necrosis with exposure of deeper tissues should be handled aggressively, with operative debridement and early coverage. Mild skin sloughs that have healthy tissue covering bone and tendons can be treated with dressing changes and immobilization.

Even if the procedure is performed perfectly, distal joint degeneration is a likely late complication. With the transverse tarsal or midfoot joints taking the stresses of the fused ankle and hindfoot, radiographic degenerative changes will appear quickly. Clinically significant pain is common as years pass, making combined fusions less appealing for the younger patient. A semi-rigid orthotic and rocker bottom shoe may protect those joints, but severe degeneration can be treated with fusion of the remaining joints. It is not rare for a patient at this stage to actively inquire about amputation, and in many cases a trans-tibial amputation will be the best option.

Because of our own dissatisfaction with the extensively fused foot, we have begun taking down the ankle fusion and converting it to a total ankle replacement. By freeing up the ankle, moderate symptoms in the subtalar, transverse tarsal, or midfoot joints are lessened. Sometimes, freeing up the ankle is enough to relieve distal symptoms, but other times the ankle arthroplasty is combined with distal joint fusion at the same procedure. Conversion of an ankle fusion to an arthroplasty is technically feasible when normal anatomy has been preserved at the time of fusion and we are in the process of reporting our first 25 cases.

Ankle arthroplasty using the Agility Ankle (Depuy, Warsaw, IN) absolutely requires an intact deltoid ligament and medial malleolus. Although we have performed ankle replacements in the absence of fibula and lateral ligaments, these cases seem to result in more pain and may collapse into valgus without a fibula for support. We also are concerned about varus instability when the fibula has been destroyed. For all these reasons, it is extremely important that normal anatomy be preserved during an ankle fusion. The fibula must not be resected. If a trans-fibular approach is used, as described in this paper, the fibula must be replaced in its normal position (not shortened superiorly or translated posteriorly). The fibula must not be thinned too much, so that the mortise remains wide. The talus should be replaced under the tibia when subluxation has occurred. While ankle replacement offers promise for patients with degenerative or post-traumatic arthritis, those with severe muscle imbalance (such as in advanced CMT) are not good candidates. Muscle imbalance, especially if progressive or not correctable, will lead to tilting of the ankle and early failure. Similarly, we are hesitant to replace an ankle in someone with an avascular talus. The replacement seems to require bony ingrowth for best pain relief, so the necrotic talus must be replaced with healthy bone before considering replacement.


Although there will always be a place for combined fusions, we expect the procedure to become less common in the next generation. A stiff, straight foot is better than a deformed, painful one, but the best combined ankle and hindfoot fusion is still far inferior to the normal condition. Total ankle replacement is emerging as a promising new technique, especially for the foot with multiple joint involvement. We have been pleased with our results combining total ankle arthroplasty and hindfoot fusion. Preliminary studies are showing good results at about 5 years of follow up, 14 but more experience is needed to assess current designs. Many hip replacements have shown impressive success at 5 years but then failed under longer follow up. Current standards of joint replacement require 15 to 25-year survival to be considered successful.

Some of the earliest concerns with modern ankle prostheses include the relatively large amount of bone resected; implant failure would require significant limb shortening to salvage, and bone loss would make revision impossible. We have found that revision is not only feasible, but is often straightforward. Bone loss has not yet been a major problem. Of course, as duration of follow up increases, bone loss may be more of a concern. At the same time, components for revision are developing, so that new implants will be able to deal with future problems. Ankle replacement has increasingly become a useful tool in our practice. For the patient with distal joint degeneration after ankle or combined fusions, we convert the fusion to an ankle replacement as an alternative to amputation.


1. Chou LB, Mann RA, Yaszay B, et al. Tibiotalocalcaneal arthrodesis. Foot Ankle 2000; 21:804–8.
2. Papa JA, Myerson MS. Pantalr and tibiotalocalcaneal arthrodesis for post-traumatic osteoarthritis of the ankle and hindfoot. J Bone Joint Surg 1992; 74A:1042–9.
3. Ansart MB. Pan-arthrodesis for paralytic flail foot. J Bone Joint Surg 1951; 33-B:503–7.
4. Barrett GR, Meyer LC, Bray EW, et al. Pantalar arthrodesis: a long term follow-up. Foot Ankle 1981; 1:279–283.
5. Hunt WS, Thompson HA. Pantalar arthrodesis. J Bone Joint Surg 1954; 36-A:349–360.
6. Cooper PS. Complications of ankle and tibiotalocalcaneal arthrodesis. Clin Orthop Rel Res 2001; 391:33–44.
7. Pochatko DJ, Smith JW, Phillips RA. Anatomic structures at risk:combined subtalar and ankle fusion with a retrograde intramedullary rod. Foot Ankle 1996; 16:542–7.
8. Berend ME, Glisson RR, Nunley JA. A biomechanical comparison of intramedullary nail and crossed lag screw fixation for tibiatalocalcaneal arthrodesis. Foot Ankle 1997; 18:639–43.
9. Mann MR, Parks BG, Pak SS, et al. Tibiotalocalcaneal arthrodesis: a biomechanical analysis of the rotational stability of the Biomet Ankle Arthrodesis Nail. Foot Ankle 2001; 22:731–3.
10. Thordarson DB, Chang D. Stress fractures and tibial cortical hypertrophy after tibiotalocalcaneal arthrodesis with an intramedullary nail. Foot Ankle 1999; 20:497–500.
11. Acosta R, Ushiba J, Cracchiolo A. The results of a primary and staged pantalar arthrodesis and tibiotalocalcaneal arthrodesis in adult patients. Foot Ankle 2000; 21:182–194.
12. Buratti RA, Johnson JD, Buratti D. Concurrent ankle and subtalar joint arthrodesis. J Foot Ankle Surg 1994; 33:278–82.
13. Russotti GM, Johnson KA, Cass JR. Tibiotalocalcaneal arthrodesis for arthritis and deformity of the hind part of the foot. J Bone Joint Surg 1988; 70-A:1304–7.
14. Pyevich MT, Saltzman CL, Callaghan JJ, et al. Total ankle arthroplasty: a unique design. Two to twelve year follow-up. J Bone Joint Surg 1998; 80-A:1410–20.

ankle arthritis; arthrodesis; hindfoot

© 2002 Lippincott Williams & Wilkins, Inc.