Steven L. Haddad
DEFINITION
Revision of the Agility Ankle is required for a variety of circumstances. It may be necessary either relatively early after the index procedure, or delayed due to late mechanical failure.
By definition, revision may be required around a stable arthroplasty (ie, correcting imbalance creating deformity in the prosthesis or repairing fractures around the prosthesis) or by implant removal and subsequent replacement of the prosthesis (whole or in part).
ANATOMY
The anatomy of revision total ankle arthroplasty revolves around the specific mechanism of failure of the original prosthesis. Structures of concern include:
Bone: the medial and lateral malleoli, the distal tibia, and the talus
Ligaments: the anterior talofibular and calcaneofibular ligaments, the deltoid ligament, and the syndesmotic ligament complex
Muscle and tendon: the Achilles tendon and the anterior tibial, extensor hallucis longus, extensor digitorum longus, peroneus longus and brevis, and posterior tibial tendons
PATHOGENESIS
Malleoli: Failure of the malleoli may occur from fracture of these structures. Fracture can occur early (due to a technical complication sustained during the procedure) (FIG 1A) or late (from weakened bone architecture due to osteolytic cysts, or undue stresses applied to the malleoli from deformity or altered gait mechanics) (FIG 1B). Microor macromotion may occur through the prosthesis–bone interface at the fibula from lack of a syndesmotic fusion, creating the potential for lateral malleolar fracture.
FIG 1 • Acute (A) and chronic (B) medial malleolus fractures. The former fractures occur due to an intraoperative technical error. The latter can occur from imbalance about the prosthesis with undue stresses.
Distal tibia and talus: Failure is sustained through axial load applied to this portion of bone, compounded by the physiologic effects of the prosthesis. Osteolysis may occur from shed polyethylene particles, creating a macrophage reaction and autodestruction of bone. This weakened, cystic bone allows the prosthesis to subside through the resection margin, creating deformity and failure. In addition, compromise through lack of bone ingrowth into the sintered beads may create micromotion within the prosthesis–bone interface, creating further erosions and subsidence into the tibia or talus (FIG 2A,B).
Lateral and medial ankle ligaments: Failure is sustained through ligaments that are often compromised before the surgical procedure (FIG 3). Often deformity exacerbates this problem, as chronic tension through weight bearing continues to attenuate the ligaments, creating further compromise.
Extensor tendon complex (anterior tibial tendon, extensor hallucis longus, extensor digitorum longus): Issues involving the extensor complex revolve around scar tissue and anterior wound complication (FIG 4A), which as a baseline creates altered motion in the form of decreased plantarflexion (FIG 4B). In more advanced circumstances, wound coverage is required due to tendon exposure. Compromised blood supply to the anterior skin, multiple prior incisions in posttraumatic or reconstructive situations, and direct apposition of the tendon complex against the skin may all accelerate anterior incision failure.
Infection: Early incision complication may provide a portal of entry for colonized bacteria, leading to superficial cellulitis or deep infection (FIG 5A,B). Early intervention in the form of parenteral antibiotics or operative débridement may allow salvage of the prosthesis. Deep infections involving the bone (osteomyelitis) may result from direct extension from these superficial infections or may occur from bacteria seeded at the time of surgery, lying dormant for an undefined interval before presentation. Bacteria may cling to the prosthesis, creating a situation resistant to antibiotic intervention. They may form a glycocalyx, insulating themselves from both antibiotics and operative irrigation.
FIG 2 • AP (A) and lateral (B) radiographs of talar subsidence. Note the penetration of the talus into the nascent talus. Also note the osteolytic cyst in the distal tibia.
FIG 3 • Severe valgus after failure of the deltoid ligament. This is in combination with a structural foot deformity and hindfoot valgus.
NATURAL HISTORY
The natural history of all of the above-mentioned problems follows a course dependent on the index mechanism of failure.
FIG 4 • Anterior wound complication (A) leads to excessive scar tissue, compromising the ability to plantarflex (B) the ankle.
FIG 5 • Anterior wound complication has led to secondary deep infection and exposed tendons (A). The cellulitis is obvious (B), but the infection is deep.
Malleolar fractures may occur early or late. In either case, resolution of the fracture is compromised by the limited bone available due to removal for implantation of the prosthesis. The natural history may thus progress to either successful union after repair, or nonunion. Nonunion may lead to a relative increased length of the malleoli and subsequently the ligaments, leading to deformity potential (FIG 6). Deformity, once present, prevents union.
Ligament compromise follows a similar predictable course, with the endpoint being instability and subsequent deformity about the prosthesis. Lack of medial or lateral restraint allows edge-loading of the polyethylene, leading to osteolysis and implant subsidence.
Incision compromise may begin as focal necrosis about the anterior wound. Blistering may be evident, or full-thickness necrosis. Necrosis presents as peri-incisional devascularized skin, which may be limited in extent or of greater breadth; either allows slough of the zone of injury. Full-thickness granulation tissue may develop, though it will be temporally slow. As such, scar tissue accumulates about the anterior tendon complex as motion must be restricted to provide the best healing environment. With exposed tendon, granulation tissue is less likely and plastic surgery involvement becomes a possibility.
Infection may present in conjunction with wound compromise as cellulitis in the early postoperative period. Without attention, cellulitis may allow deep bacterial infestation, creating osteomyelitis or septic arthritis of the artificial joint. Salvage of the index prosthesis becomes less likely, as a glycocalyx may form about the polyethylene, shielding the bacteria from antibiotic penetration or irrigation.
FIG 6 • Stressing the ankle in the operating room reveals significant medial bone compromise, leading to valgus deformity and medial instability.
Malleolar fracture, ligament compromise, prosthesis subsidence, incision compromise, and infection will be discussed separately.
MALLEOLAR FRACTURE
PATIENT HISTORY AND PHYSICAL FINDINGS
Evaluating the ankle for malleolar fracture should include the following:
Direct palpation to the medial or lateral malleoli, or pain with weight bearing medially or laterally. This is not the normal location of postoperative pain, so it creates clinical suspicion that fracture is present.
The examiner should look for increased swelling about the ankle joint after postsurgical resolution. The examiner should evaluate for deep vein thrombosis, but normally in combination with pain, one thinks of malleolar fracture.
IMAGING AND OTHER DIAGNOSTIC STUDIES
Plain radiographs: Malleolar fractures may be subtle or obvious. Obvious fractures are visible at the level of the prosthesis, generally at the apex or superior corners of the prosthesis. In iatrogenic cases, the fractures occur at the level of the superior saw cut line on the tibia, where the sagittal saw violates the medial or lateral malleolus. Significant distraction via the uniplanar fixator upon osteoporotic bone may create avulsion fractures at the malleoli after saw cuts, where the thinned malleoli are subject to increased force per unit area. Subtle fractures are generally delayed in appearance and may involve periosteal reactions seen at the medial malleolus proximal to the prosthesis. This type of fracture occurs from an unbalanced prosthesis (FIG 7A) placing uneven load or compression about the malleoli (FIG 7B).
Tc99 bone scans: This study is generally not helpful, as increased uptake is visible surrounding the prosthesis, making it difficult to discern a fracture from normal pooling.
CT: This test is very helpful to evaluate both the presence and healing of malleolar fractures. CT scanning is not helpful during intraoperative occurrences but has value in delayed presentation, especially in subtle cases (FIG 8). Subtraction software minimizes interference from the prosthesis.
FIG 7 • Hindfoot varus (A) places increased stress about the medial malleolus, which can create a delayed medial malleolus fracture. (B) Weight-bearing radiograph of the ankle demonstrates the most common pattern of late medial malleolus fracture: a vertical shear fracture pattern. Repetitive stress from the medial corner of the tibial implant creates the vertical fracture line that allows the prosthesis to shift into varus.
DIFFERENTIAL DIAGNOSIS
Malleolar fracture
Infection
Deep venous thrombosis
NONOPERATIVE MANAGEMENT
Conservative treatment for malleolar fractures involves cast immobilization until union is complete. Unlike malleolar fractures without ankle arthroplasty, immobilization is often extended beyond the standard 6 weeks, as the decreased surface area for healing due to the space-occupying prosthesis increases the likelihood of nonunion. If immobilization is terminated before complete union, refracture or separation of the fragments becomes likely, mandating surgical correction. A CT scan is important to quantify union before discontinuing immobilization.
FIG 8 • CT scan of subtle medial malleolus fracture in a delayed presentation due to persistent hindfoot valgus.
Use of pulsed electromagnetic fields or ultrasound to stimulate union may enhance union.
SURGICAL MANAGEMENT
Surgical repair of malleolar fractures is normally the treatment of choice after the fracture is recognized. As the rehabilitative goal of total ankle arthroplasty is early range of motion, prolonged immobilization to allow conservative union may lead to undue ankle stiffness, compromising patient satisfaction. Thus, upon visualization of a malleolar fracture (either acute or delayed), surgical repair is indicated.
Preoperative Planning
In acute or iatrogenic situations, no preoperative planning is possible.
In delayed or chronic situations, a CT scan is useful to evaluate the fracture pattern and determine screw placement. In addition, the CT scan may visualize partial union, allowing percutaneous fixation of the fracture fragments.
Positioning
Positioning is supine for this procedure, with a bump under the ipsilateral hip of a diameter to rotate the knee, ankle, and foot to a neutral position (FIG 9A). Generally, blankets are used to elevate the involved extremity above the sagittal plane of the unaffected extremity (FIG 9B). This position improves the accuracy of sagittal imaging and prevents the need to lift or manipulate the involved extremity during the more tenuous portions of the surgical procedure.
Approach
The surgical approach depends on whether the malleolar fracture is acute (noted intraoperatively) or chronic (occurs at a later date).
The surgical approach for acute malleolar fractures is performed in the index procedure (ie, anterior approach for reduction and fixation of the medial malleolus fracture and lateral approach for the lateral malleolus fracture). The anterior approach allows evaluation of the fracture reduction, while screws are placed percutaneously medially. The lateral approach provides direct access for reduction and plate fixation.
In chronic situations, the lateral approach is still used for lateral malleolar fractures. However, a medial approach is preferred for medial malleolar fractures, as direct visualization of the fracture fragments is critical, and often bone loss is present. The medial approach allows placement of either screws or plates, depending on the anatomy of the fragments.
FIG 9 • AP positioning (A) reveals a knee placed in neutral alignment. Lateral (B) positioning on operating table. Note the rigid bump under the ipsilateral hip (rolled blankets) and the leg elevation on a firm surface of blankets.
TECHNIQUES
OPEN REPAIR OF THE ACUTE MEDIAL MALLEOLUS FRACTURE
Reduce this fracture anatomically (TECH FIG 1A) during surgery and hold it with a reduction clamp.
Place guidewires from the tip of the medial malleolus into the distal tibia (TECH FIG 1B). There is adequate room for one and possibly two guidewires despite the medial bone resection for the prosthesis.
Perform screw fixation percutaneously with cannulated screws (TECH FIG 1C). Alternatively, if solid-core screws are preferred, drilling is done over the guidewire, followed by guidewire withdrawal and placement of the solid-core screws.
TECH FIG 1 • Stressing the ankle intraoperatively (A) reveals gapping at the medial malleolar fracture site. Two guidewires are placed across the fracture site (B) within the substance of the medial malleolar bone. Firm compression is achieved (C) across the fracture site.
OPEN REPAIR OF THE LATERAL MALLEOLUS FRACTURE
Currently, standard fixation of the syndesmotic fusion is done percutaneously, after placing the plate against the distal fibula through the anterior approach (TECH FIG 2A).
However, if a lateral malleolar fracture is sustained, extend the lateral approach as a standard lateral approach for this type of fracture pattern. Clean the fracture ends of debris (TECH FIG 2B).
After direct exposure, reduce the fragments (TECH FIG 2C,D). Often the fracture is at the apex of the lateral portion of the prosthesis. Thus, it is difficult to provide standard lag-screw fixation.
Prebend the plate to hook around the lateral malleolus tip (TECH FIG 2E).
Apply the plate proximal to the fracture with three screws traversing the syndesmosis.
Distal screw fixation generally allows placement of two screws, both cortical. The first, more proximal screw should be placed with lag technique. The second screw is placed intramedullary, at the tip, to provide stabilization.
TECH FIG 2 • Lateral malleolus fracture. The lateral approach is chosen (A), and the fracture fragments are exposed and curetted (B). Intraoperative fluoroscopy demonstrates the fracture location (C) and reduction with plate fixation (D). Clinical photograph reveals contouring of plate (E) and screw fixation above and below the fracture to maintain stability.
REPAIR OF A LATE MEDIAL MALLEOLUS FRACTURE
As mentioned above, a direct medial approach is used to provide better access to the fracture fragments while avoiding violation of the ankle prosthesis (TECH FIG 3A).
Use a curette to remove all fibrous tissue present (TECH FIG 3B). This is a critical step, as good-quality vascular bone must be visualized on both sides of the fracture (TECH FIG 3C).
It is critical to provide direct apposition of the fracture fragments to enhance the potential for union (TECH FIG 3D,E).
Bone graft may be required to supplement any gaps present; it is generally taken from the calcaneus (TECH FIG 3F–I).
Normally, axial fixation is preferred, as compression is achieved through standard medial malleolar fixation techniques. However, plate fixation may be used as a supplement after compression is achieved (TECH FIG 3J–O).
TECH FIG 3 • Late medial malleolus fracture. A direct medial approach is performed (A). Due to the delayed presentation and subsequent longevity of motion about the fracture site, significant bone erosion is present. The fracture site is curetted (B) and stressed (C) to check the stability of the prosthesis. The fracture is manually reduced (D) and clamped to maintain the reduction (E). The base of the fracture is packed with cancellous bone graft (F). Tricortical graft is harvested (G), templated (H), and prepared (I) for a perfect fit in the cortical defect. A plate is contoured (J) and the bone graft inserted into the defect (K,L). The plate is applied (M) under compression (note eccentric placement of proximal screws). In addition, an axial screw (N) assists with this compression. The fracture achieves successful union (O).
POSTOPERATIVE CARE
Rigid internal fixation allows early motion, which is important after an ankle arthroplasty.
After a brief period of immobilization (generally 10 days to 2 weeks), the patient is converted to a controlled-ankle-motion (CAM) boot.
Range of motion is initiated at this time without weight bearing. Weight bearing is restricted until 6 weeks postoperatively, when union is generally present. It is extremely important not to allow an increase in activity or discontinuance of the CAM boot until union is complete. Nonunion of malleolar fractures in total ankle arthroplasty can severely compromise the result of the arthroplasty by creating intra-articular deformity and edge loading.
OUTCOMES
Fracture fixation rarely results in nonunion. Without fixation, the risk of nonunion is high, compromising the arthroplasty through the potential for late deformity.
COMPLICATIONS
Nonunion is the major potential complication.
LIGAMENT INSTABILITY
PATIENT HISTORY AND PHYSICAL FINDINGS
Patients with ligament instability creating late malalignment of the prosthesis will develop increasing pain in the medial or lateral gutters around the prosthesis as the progressive tilt creates gutter impingement.
Examination of the total ankle arthroplasty for ligament instability should include the following:
Medial–lateral stress radiographs. Medial and lateral stress is applied to the ankle, specifically looking for a soft endpoint or gross ligamentous laxity. This test will assist the examiner in determining the incompetence of the ligaments directly. Without a firm endpoint to stress, the ligaments are clearly compromised.
Thumb-to-forearm test, or hyperextension for the elbow: These basic tests examine for gross ligamentous laxity, often congenitally based. Patients with gross ligament laxity are at higher risk for failure of the prosthesis due to tilt, edge loading, and subsequent osteolysis.
IMAGING AND OTHER DIAGNOSTIC STUDIES
Manual stress radiographs may assist the examiner in quantifying ligament laxity. Similar to the test mentioned above, stress is applied across the ankle joint while a mortise ankle radiograph is obtained. The examiner will specifically look for increased tilt of the talus upon the tibial tray.
Weight-bearing AP and mortise radiographs may also be used to quantify ligament laxity, as the tilt of the prosthesis can be directly measured.
Both radiographs must be interpreted with caution, for this tilt may be compounded by underlying foot deformities. Thus, these deformities must be corrected simultaneously, or any ligament reconstructive procedures will ultimately fail due to recurrent tension across the newly created ligament.
DIFFERENTIAL DIAGNOSIS
Varus or valgus foot deformity
Posterior tibial tendon insufficiency
Peroneal tendon rupture
NONOPERATIVE MANAGEMENT
Conservative care revolves around stabilizing the ankle with medial and lateral posts, often in the form of a lace-up brace, a U-shaped stirrup brace, a Ritchie-type brace, or an Arizona brace (FIG 10). All braces serve to limit varus or valgus thrust, and thus limit tilt.
FIG 10 • Protective Arizona brace (right) offers a more secure option for stability over the UCBL orthotic (left).
SURGICAL MANAGEMENT
Surgical ligament reconstruction is challenging due to the space occupied by the prosthesis. Thus, standard techniques need to be modified to accommodate the bone architecture provided by resection necessary for prosthesis implantation.
Preoperative Planning
Patients with ligament incompetence are assessed for laxity via the stress radiographs mentioned above. The patient is assessed for hindfoot and forefoot deformities that may require simultaneous correction. Gross ligament laxity must be accounted for, as balance achieved through tightening one side of the ankle may create the opposite deformity from that corrected due to a lack of contralateral restraint. Finally, a diagnostic ultrasound may be used to assess the quality of the posterior tibial tendon and peroneal tendons, as an MRI will be compromised by the prosthesis.
Positioning
Depending on the involved ligaments, the patient is positioned either laterally on the operating table for lateral ligament incompetence, or supine with a bump under the opposite hip for medial ligament incompetence.
Approach
The approach parallels the standard anterior incision performed for ankle arthroplasty, maximizing the skin bridge to minimize wound complications. Exposure is carried proximal to the ankle joint a minimum of 5 cm and distal to the ankle joint a minimum of 6 cm. This generous incision allows access to reconstruct all aspects of the failed ligaments.
TECHNIQUES
DELTOID RECONSTRUCTION
The patient is placed supine on the operating table, with a bump under the contralateral hip to externally rotate the involved extremity.
The incision is medially based, extending from 1 cm proximal to the medial malleolus past the sustentaculum tali (TECH FIG 4A).
Retract the posterior tibial tendon posteriorly, exposing the insertions of the deep and superficial components to the deltoid ligament (TECH FIG 4B,C).
Using standard EndoButton technique, place drill holes at both insertions, exiting anterior to the fibula (TECH FIG 4D–F).
TECH FIG 4 • Deltoid reconstruction. Some photos are taken from cadaveric dissection during the development of the procedure; others are intraoperative photos taken during a reconstruction involving a stemmed talar component, calcaneal osteotomy, and Cotton osteotomy. The surgical approach is medial (A). After opening the posterior tibial tendon sheath and retracting the tendon anteriorly (B), the deep deltoid is visible. This ligament may be sectioned for lateral imbrications (C). Guidewires are passed for the EndoButton in a plane that emerges anterior to the fibula (D) and transtalar (E). The guidewires are placed at the insertion points of the native deep and superficial deltoid ligaments (F). The cadaveric tendon is prepared by placing Krackow suture weaves at both ends (G) and tensioned to minimize late plastic deformation. Each end of the graft is placed into the respective deep and superficial deltoid tunnel (H,I), held in place on the lateral side of the talus by the EndoButtons (J). The ankle replacement remains in valgus (K) until the tendons are placed through a tunnel drilled from the tip of the medial malleolus directed toward the anterior distal tibia (L). The spiked ligament washer is placed proximal to the exit point of this tunnel to place the graft under maximal tension (L). The looped end of the graft is placed around the screw, and the ligament washer is tightened against it (M). The ankle is now aligned in neutral (N), and this position is confirmed under fluoroscopic imaging in the AP (O) and sagittal (P) planes. Note the position of the EndoButtons on the sagittal plane, anterior to the fibula and transtalar (P).
A cadaveric anterior tibial tendon is used for the reconstruction. Weave a Krackow suture stitch with no. 2 Ethibond through both ends of the tendon and anchor it to two separate EndoButtons with 1 cm of suture lead between the EndoButton and the tendon ends (TECH FIG 4G). The tendon is prestretched to minimize late plastic deformation of the tendon graft.
Pass the ends through the drill holes and flip the buttons, providing secure fixation at both the superficial and deep insertions (TECH FIG 4H–J).
Securely reproduce the ligament origin by placing a drill hole at the tip of the medial malleolus, directing the hole toward the anterior central tibia. This hole is placed obliquely to avoid the tibial component to the prosthesis (TECH FIG 4K,L).
Double the tendon upon itself and thread it through this drill hole, with the looped end exiting the anterior tibia.
Place a 4.5-mm drill hole proximal to the exit point of the looped tendon, 1 cm beyond the maximum stretch of the loop component of the tendon. By placing the drill hole 1 cm proximal to the extent of the loop, the ligament reconstruction will remain taut (TECH FIG 4M).
Anchor the loop with a large-fragment screw and a spiked ligament washer (TECH FIGS 4N–P).
LATERAL LIGAMENT RECONSTRUCTION
Unfortunately, a modified Brostrom procedure is not sufficient to stabilize the lateral ligaments in light of an ankle replacement. Thus, a similar cadaveric tendon transfer is performed to stabilize the deficient lateral ligaments.
Use a cadaveric anterior tibial tendon, tubularizing the tendon and weaving a Krackow suture with no. 2 Ethibond on one end. Secure an EndoButton to this suture with a 1-cm gap between the end of the tendon and the EndoButton.
Make a drill hole through the talar neck at the insertion of the anterior talofibular ligament, exiting anterior to the medial malleolus. The far cortex of the hole is smaller than the length of the EndoButton.
If sufficient fibula is present distal to the lateral portion of the tibial tray, place 7.3-mm drill holes at the origins of the anterior talofibular ligament and calcaneofibular ligament. These holes meet in the central fibula.
Place the allograft tendon through these holes, with the distal segment exiting through the inferior (calcaneofibular) hole.
If there is not sufficient fibula at the tip, carry the plate for securing the syndesmotic fusion to the tip of the fibula, and place the cadaveric tendon deep to the plate. Place a screw in the most distal hole of the plate to help stabilize the transferred tendon.
In either case, place the hindfoot into eversion and make a 4.5-mm drill hole in the calcaneus, from lateral to medial, at the insertion of the calcaneofibular ligament.
Place the cadaveric tendon under maximal tension, and use a knife to bisect the allograft proximal to the previously drilled hole.
Place a 6.5-mm large-fragment screw with a large spiked ligament washer through the cadaveric tendon at the point of the previously placed incision. This screw is placed proximal to the previously drilled hole. Again, this technique will maximize tension on the transferred tendon.
Insert the screw into the calcaneus, with the spiked washer completely engaging the tendon to provide rigid fixation.
Place the peroneal tendons deep to this transferred tendon to prevent dislocation.
If residual tendon is present, it may be doubled back over the lateral wall of the fibula and anchored to the bone to provide increased strength to the transfer.
POSTOPERATIVE CARE
If the fixation is rigid and strong, the patient is placed in a cooling boot for 2 to 3 days and admitted to the hospital. This minimizes the risk of incision complications, but the patient must remain dormant with the leg elevated to prevent stretch on the newly reconstructed ligament.
If the fixation has the potential for compromise, the patient is placed in a stirrup-type plaster splint in combination with a posterior mold splint. This construct will take tension off the ligament repair while simultaneously keeping the ankle flexed to neutral.
The patient is changed to a cast at 5 to 7 days with windows placed in the cast for direct incision observation.
Physical therapy is used at 6 weeks postoperatively to increase ankle range of motion. No inversion or eversion is attempted until 3 months postoperatively. The patient is in a CAM boot fully weight bearing at this time.
The patient is placed in a lace-up brace at 12 weeks postoperatively. This brace may be discontinued at 4 months postoperatively.
OUTCOMES
This technique is newly developed, so there are no long-term outcome studies at this time. A trial is under way.
COMPLICATIONS
Wound infection, cellulitis, wound necrosis
Nerve damage to superficial peroneal, deep peroneal, saphenous, sural, or tibial nerves
Recurrence of deformity, tendon transfer failure
Opposite deformity in cases of gross ligament laxity
PROSTHESIS SUBSIDENCE
PATIENT HISTORY AND PHYSICAL FINDINGS
Patients with subsidence of the prosthesis have collapse and eventual impingement in the medial or lateral gutters. They often complain of increasing stiffness upon a prosthesis that was previously functioning well. This is rarely an acute phenomenon (unless fracture through an osteolytic cyst created the subsidence) and is often noticed as a part of a routine office visit. Pain follows the stiffness and is normally located deep, medial, and lateral to the prosthesis.
The methods for examining the total ankle arthroplasty for subsidence include the following:
Standing flexibility (range of motion) of the ankle: The patient stands with the involved ankle anterior to the axis of the body and the opposite ankle behind the body. To determine dorsiflexion, the knee is flexed forward as far anterior as possible (runner's stretch), and the angle between the foot and the tibia is measured. Plantarflexion is assessed by having the patient lean back on the involved extremity as far as possible while keeping the entire foot flat on the ground. The angle is then measured. This should be compared at each office visit and documented. Increasing stiffness is commonly seen with subsidence of the prosthesis. If this test is done routinely as a part of each office visit, accurate and reproducible values will allow measured changes in ankle flexibility.
Direct palpation: The examiner must palpate deeply the medial and lateral gutters to elicit pain. In addition, the syndesmotic fusion is painful to palpation if a nonunion has led to tibial tray subsidence. Pain signifies increasing gutter impingement as the arthritic bone interacts due to loss of height. Pain in the syndesmotic region signifies a syndesmotic nonunion, which allows tibial tray subsidence (common in an undersized prosthesis).
IMAGING AND OTHER DIAGNOSTIC STUDIES
Standard radiographs include weight-bearing AP, lateral, and mortise views of the ankle. These radiographs reveal a tilt in the components if subsidence is not uniform. The lateral radiograph (FIG 11) is assessed for tibial tray subsidence (often at the anterior cortex of the tibia) and for talar subsidence (the talar fin may be penetrating the subtalar joint). The AP and mortise radiographs allow an assessment of measurable tilt and bone loss (FIG 12). Gutter impingement becomes obvious as contact between the talus and the mediolateral malleolus is visible.
A CT scan is critical to assess penetration of the prosthesis into the subtalar joint (FIG 13). In addition, the nonunion of the syndesmotic fusion may be visualized, contributing to tibial tray subsidence. The surgeon can assess for an undersized prosthesis by examining the coverage of the tibial tray underlying the fibula. Lack of coverage by the fibula can lead to valgus subsidence of the tibial tray. The CT scan will reveal osteolytic cysts contributing to subsidence of the tibial tray or talar component, and potential fracture through these cysts. Finally, the CT can assess for ingrowth of the prosthesis and potential loosening. Subtraction software will assist in eliminating artifact due to the prosthesis.
FIG 11 • Lateral radiograph of grossly subsided prosthesis. Note the lack of anterior coverage in the distal tibia and the subsequent anterior subsidence of the prosthesis. This preserved shelf of anterior bone can be used to re-establish the height of the tibial tray with a larger prosthesis offering better sagittal coverage. The lateral radiograph helps to assess talar subsidence, in particular looking for violation of the subtalar joint.
FIG 12 • Evaluation for tibiotalar tilt and lack of fibula coverage in the tibial tray on the AP radiograph. An assessment of deformity can guide the surgeon toward supplementary procedures done at revision surgery.
DIFFERENTIAL DIAGNOSIS
Infection (osteolysis vs. osteomyelitis)
Residual posttraumatic osteoarthritis in the medial or lateral gutters
NONOPERATIVE MANAGEMENT
Nonoperative management involves bracing to prevent pain from arthritic or collapsed bone surfaces. In particular, a nonarticulating ankle–foot orthosis may limit pain.
SURGICAL MANAGEMENT
The indications for surgical treatment of subsidence are three: pain, stiffness that is unacceptable to the patient, and prevention of substantial subsidence that will diminish bone stock to the point of preventing revision of the prosthesis in the future. Subsidence to this level may necessitate implant removal followed by arthrodesis of the ankle joint.
Preoperative Planning
Assessment of the remaining bone stock after implant removal allows the surgeon to predict whether a custom prosthesis will need to be constructed to salvage the failed joint replacement. The most accurate way to make this assessment is a careful review of the CT scan.
In addition, assessment for revision of the syndesmotic fusion may be made by CT scan. Underlying foot deformity that may have contributed to prosthesis subsidence is determined to allow planning for simultaneous flatfoot–cavovarus foot correction. The revised prosthesis must be placed upon a plantigrade foot to prevent uneven stresses leading to secondary failure.
FIG 13 • CT scan of the ankle in the axial plane (A) reveals cystic changes to the distal tibia consistent with osteolysis. The coronal plane cuts (B,C) demonstrate the magnitude of subsidence and the syndesmotic nonunion. The sagittal cut (D) reveals the lack of penetration of the talar fin into the subtalar joint.
Infection is assessed via markers such as the white blood cell (WBC) count, erythrocyte sedimentation rate, and C-reactive protein. We have found limited value in a tagged-WBC scan, as the prosthesis itself may create an inflammatory foreign body reaction that mimics osteomyelitis on nuclear medicine scans.
Bone stock and osteoporosis may be assessed via CT scan, plain radiographs, and a DEXA scan. If osteoporosis is present, a concerted effort to increase bone mass will assist in providing structural support to the prosthesis and should be performed before revision.
Positioning
The patient is placed supine on the operating table.
A bump is placed under the hip to rotate the extremity to neutral with respect to the knee.
The lower leg is placed on multiple blankets to provide a firm working surface while simultaneously allowing shootthrough lateral radiographs to enhance the assessment of prosthesis coverage (anterior and posterior pillar).
Approach
The surgical approach is anterior, following the initial incision done for the primary arthroplasty. The interval is the same as that mentioned above.
TECHNIQUES
TALAR SUBSIDENCE
Expose the talus by removing all associated scar tissue and necrosis. The fixator is applied to provide distraction and easy component access while providing stability (TECH FIG 5A–D).
Remove the talar component. It is normally not ingrown and comes out simply after replacing the insertion rod and joysticking the component (TECH FIG 5E,F).
Make saw cuts to restore the axis of the prosthesis to be perpendicular to the axis of the tibia (TECH FIG 5G–I). These cuts may use a standard (though larger) cutting block.
There is often a rectangular void in the talus after component removal (TECH FIG 5J,K). Curette any fibrous tissue within that void, achieving healthy bleeding bone. There is normally a rim of excellent cortical bone in the nascent remaining talus, as the original Agility talar component does not conform to the entire talus after the initial cut.
This rim of bone is advantageous, as the new Agility LP talar component does conform to the entire talus. As such, the talar component will sit on this rim, establishing the height of the nascent talus made with the original saw cut at the index procedure.
The fin on the Agility LP will have no stability, due to the rectangular defect present in the central talus (TECH FIG 5L). One could consider bone grafting this defect, but I do not believe the bone graft will provide adequate stability in the short term, allowing the talar component to rotate or shift from the desired position after closure and rehabilitation. In addition, in the long term, it would be unusual for the bone graft to incite ingrowth into the Agility LP talar component.
Thus, I use cement fixation for the talar fin. I fill the void present with polymethylmethacrylate while it is still softer and malleable. I then onlay the Agility LP talar component onto the talus, allowing it to sit on the rim of nascent bone while the fin conforms to the cement in the desired position (TECH FIG 5M–O). To do this, the tibial component must be in place so that appropriate rotation and mediolateral positioning can be determined. The remaining native talus provides some element of bone ingrowth into the prosthesis, while the cement interdigitates with the residual sintered beads on the talar component.
I allow the cement to harden while I manually reduce the ankle joint, maintaining the correct position until the cement cures. The talus is now stable and ready to articulate (TECH FIG 5P–T).
TECH FIG 5 • Operative management of severe subsidence of both the tibial and talar components via conversion of the primary replacement to an Agility LP component, with the addition of polymethylmethacrylate for stability. The anterior approach is reproduced, and no visible components are present due to the severe subsidence (A). The fixator is applied (B), allowing distraction and stabilization (C, reverse image). After distraction, the components become visible (D) and the magnitude of subsidence becomes obvious. After component removal (E), the bone defects are apparent (F) and the residual bone is truly appreciated. The saw cuts are made (G) after application of a large-sized cutting block (H), revealing a thinned but present medial malleolus (I). Clinical inspection demonstrates a preserved cortical rim in the tibia (J) and talus (K) with the central rectangular defect expected. The trial prosthesis confirms adequate rim support for the Agility LP (L), with restoration of height. The central defect is apparent and is filled with polymethylmethacrylate. The final components (M) are two sizes larger than the original. The syndesmosis has also been grafted, with fixation revised. The rim coverage is apparent on intraoperative fluoroscopy (N,O). At 1 year postoperatively (P–R) the prosthesis is stable and balanced, without laxity or tilt. The range of motion has been improved substantially by providing an appropriately articulating ankle replacement (S,T).
TIBIAL COMPONENT SUBSIDENCE
Expose the tibia using the above-mentioned technique. Remove all associated scar tissue, and define the defects present.
Remove the tibial component by replacing the insertion rod and joysticking the component.
Generally, when the tibial component subsides, it creates significant bone loss in the distal tibia at the resection site. Often there is lack of appropriate fibula coverage on the tibial component, particularly those components that subside into valgus. This allows the surgeon to plan appropriate tibial tray coverage to ensure that the revised tibial tray will cover the fibula. In these instances, simply inserting a larger tibial tray will complete the revision.
However, if bone loss is present, the height of the tibial component must be re-established (TECH FIG 6A). In this instance, a custom tibial component must be created (TECH FIG 6B). It is important to note the size of the original component when designing the revision component. The component itself may be based on standard mediolateral dimensions, simply adding height to the tibial tray. In this instance, it is important to use knowledge based on the size of the initial component in combination with calibrated radiographs (TECH FIG 6C) to determine the size (width) necessary to provide appropriate fibula coverage. This will lessen the risk of future tibial component subsidence.
Unlike the talar component, which does not require a new cut in the talus due to the isolated central core subsidence, tibial component subsidence does erode the supportive bone, and as such will require a new cut perpendicular to the plane of the tibia. One may use the standard cutting blocks to make this cut, although custom cutting blocks can be manufactured.
Once cuts are made, the revision component fits securely into its new space. Unlike revision of the talar component, polymethylmethacrylate is not necessary, as the new tibial component has the ability for ingrowth into the newly cut surfaces.
TECH FIG 6 • This patient had tibial tray subsidence leading to ligament laxity and a varus deformity (A). By revising the tibial component to a custom tray with increased height, ligamentotaxis is established and the varus deformity is eliminated (B). The use of calibrated radiographs (C) allows accurate size estimation for the revision components (note metal spheres).
POSTOPERATIVE CARE
If the fixation is rigid and strong, the patient is placed in a cooling boot for 2 to 3 days and admitted to the hospital. This minimizes the risk of incision complications. We normally do not cast in the operating room under these circumstances, for preventing tension on the anterior surgical incision is the best method to avoid incision complications.
The patient is changed to a cast at 5 to 7 days with windows placed in the cast for direct incision observation.
Physical therapy is used at 2 weeks postoperatively to increase ankle range of motion, assuming the incisions have healed. Full weight bearing may be instituted before the standard 6-week interval if the patient had a previous successful fusion of the syndesmosis. If that is not the case, then weight bearing is restricted until the syndesmosis is fused.
The patient may discontinue use of the CAM boot before 12 weeks if strength is appropriate and the syndesmosis is fused. Stability is enhanced if polymethylmethacrylate is used, and thus weight bearing without assistive devices is accelerated.
OUTCOMES
These techniques are newly described, so there is no literature to support their use at this time. Anecdotal experience supports the techniques, however, with short-term outcomes (1 year) demonstrating substantial improvement at this time.
COMPLICATIONS
Wound infection, cellulitis, wound necrosis
Nerve damage to superficial peroneal, deep peroneal, saphenous, sural, or tibial nerves
Subsidence of the newly placed components into poor-quality bone. This is a particular problem in patients with rheumatoid arthritis or other systemic conditions.
INFECTION
PATIENT HISTORY AND PHYSICAL FINDINGS
The presentation of infection depends on the organisms involved. However, the common pathway is cellulitis, with or without a wound complication. Patients with joint infections often have fever and may have chills. Pain is often present about a previously painless prosthesis.
IMAGING AND OTHER DIAGNOSTIC STUDIES
Plain radiographs may reveal a lucent line around the prosthesis, documenting lack of ingrowth. Cystic changes may be visualized in the bone surrounding the prosthesis.
A CT scan is more specific for lucency about the prosthesis and poor ingrowth. In addition, the CT scan is more specific for bone cysts. Air or gas in the soft tissues is visible on CT scan. Contrast-enhanced CT scans can reveal a soft tissue abscess or a sinus tract communicating with an anterior wound complication.
A nuclear medicine Tc scan combined with a tagged WBC scan may assist in differentiating infection from aseptic loosening of the components (FIG 14). The results should be interpreted with caution, however, for even in cases of aseptic loosening, inflammation around the prosthesis and foreign body reaction can create a false-positive WBC scan. Thus, this scan must be interpreted in combination with clinical and hematologic findings.
FIG 14 • A Tc99 scan is tagged with an indium (white blood cell) scan to increase the accuracy of the findings. The four images on the left are the Tc99 study done at 300 seconds and 10 minutes, and the equivalent temporal study using indium is on the right. This study suggests osteomyelitis, which was confirmed at débridement.
Blood work must include a complete blood count with differential, an erythrocyte sedimentation rate, and C-reactive protein. Again, these results must be interpreted in combination with clinical and radiographic findings.
DIFFERENTIAL DIAGNOSIS
Aseptic component loosening and subsidence
NONOPERATIVE MANAGEMENT
Antibiotic management is the staple in treating infection but must be done in a thoughtful manner.
Empiric therapy is appropriate only in cases of pure cellulitis without deep infection.
If deep infection is suspected, débridement and deep cultures should be obtained before starting antibiotics. The exception to this rule would be circumstances where the patient's life is in danger (ie, the patient is septic and hemodynamically unstable).
SURGICAL MANAGEMENT
The indications for surgery are the suspicion of a deep infection. If any suspicion is present, this is an indication for surgery.
Preoperative Planning
The above-mentioned studies are all performed and interpreted.
The incision is assessed for closure and the need for plastic surgery. If a wound complication is present, a preoperative assessment by a plastic surgeon is appropriate but not mandatory. If wound closure does not seem likely, the surgeon should plan on having a Wound VAC readily available for intraoperative application.
Positioning
The patient is placed supine on the operating table.
A bump is placed under the hip to rotate the extremity to neutral with respect to the knee.
The lower leg is placed on multiple blankets to provide a firm working surface while simultaneously allowing shootthrough lateral radiographs to enhance the assessment of prosthesis coverage (anterior and posterior pillar).
Approach
The surgical approach is anterior, following the initial incision done for the primary arthroplasty. The interval is the same as that mentioned above.
TECHNIQUES
IMPLANT REMOVAL
Localize the implant through deep dissection. Follow any sinus tracts present to confirm direct communication with an infected incision.
Remove the implant (all components) by applying the insertion rods and joysticking the components.
Remove screws and plates about the syndesmotic fusion, as they are in direct communication with deep infection.
Curette the bone surfaces and remove all necrotic tissues. Perform bone débridement. All residual tissue should be viable and demonstrate good vascularity. If a tourniquet is used, it should be released at this time.
Perform irrigation with at least 6 liters of antibioticimpregnated normal saline.
Prepare the polymethylmethacrylate mixed with a heatstable antibiotic (vancomycin, gentamicin). Often two bags are needed to manufacture enough cement to fill the void left by removing the prosthesis.
Insert the cement into the newly created space while placing some distraction between the tibia and talus (TECH FIG 7A). This will maintain height by allowing the cement to fill the entire void. Ensure that the cement contacts the cut surfaces of the tibia and talus to maximize the local effect of the antibiotic and to provide enough stability to allow some weight bearing (TECH FIG 7B,C).
Close the wound over a large suction drain, or, if this is not possible, apply a Wound VAC.
TECH FIG 7 • Implantation of an antibiotic-impregnated cement spacer (A) allows structural support and maintenance of height by providing complete fill in the AP (B) and sagittal (C) planes.
IMPLANT REINSERTION
After 6 weeks of intravenous antibiotics and 6 weeks off of antibiotics (3 months total), hematologic tests are done to determine the potential for residual infection.
The identical anterior approach is used.
If bone loss was present, custom components are created based on both CT data and measured plain radiographs.
Remove the spacer (TECH FIG 8A) and curette the bone of any defects or necrotic tissue.
Make new bone cuts with a cutting block (TECH FIG 8B) to ensure that there is viable tibial and talar bone for ingrowth (TECH FIG 8C).
Insert the prosthesis. If stability is in question, place polymethylmethacrylate on the fins of the tibia and talus before insertion (TECH FIG 8D,E). This limited application of cement allows bone ingrowth to the remaining prosthesis (the majority) while providing enough stability to allow early range of motion and weight bearing (TECH FIG 8F–H).
TECH FIG 8 • Revision after cement spacer. The spacer is first removed (A), followed by revision of the saw cuts (B). This allows a larger prosthesis seated against quality vascular bone (C). The prosthesis is balanced and stable (D,E). At 1 year postoperatively, there is no visible subsidence or cystic changes in the bone to suggest recurrent infection (F,G), and the incision demonstrates no compromise (H).
POSTOPERATIVE CARE
After the index débridement, further débridement may be required with gross infections. In addition, if necessary, plastic surgery is performed once Wound VAC application has resulted in a stagnant incision. It is important to obtain excellent soft tissue coverage (often necessitating a free flap) to improve resolution of the infection and allow implantation of the revision prosthesis. In this scenario, by the time revision surgery is performed, the flap has healed sufficiently to allow the anterior approach.
After revision surgery, the protocol is no different from that for subsidence.
OUTCOMES
These techniques are newly described, so there is no literature to support their use at this time. Anecdotal experience supports the techniques, however, with short-term outcomes (1 year) demonstrating substantial improvement.
COMPLICATIONS
Recurrent infection, osteomyelitis
Recurrent wound breakdown
Nerve damage to superficial peroneal, deep peroneal, saphenous, sural, or tibial nerves
Subsidence of the newly placed components into poor-quality bone. This is a particular problem in those with rheumatoid arthritis or other systemic conditions.
REFERENCES
1. Fevang BT, Lie SA, Havelin LI, et al. 257 ankle arthroplasties performed in Norway between 1994 and 2005. Acta Orthop 2007;78: 575–583.
2. Haddad SL, Coetzee JC, Estok R, et al. Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a systematic review of the literature. J Bone Joint Surg Am 2007;89A:1899–1905.
3. Hurowitz EJ, Gould JS, Fleisig GS, et al. Outcome analysis of agility total ankle replacement with prior adjunctive procedures: two to six year follow-up. Foot Ankle Int 2007;28:308–312.
4. Knecht SI, Estin M, Callaghan JJ, et al. The Agility total ankle arthroplasty: seven to sixteen-year follow-up. J Bone Joint Surg Am 2004; 86A:1161–1171.
5. Kotnis R, Pasapula C, Anwar F, et al. The management of failed ankle replacement. J Bone Joint Surg Br 2006;88B:1039–1047.
6. Kurup HV, Taylor GR. Medial impingement after ankle replacement. Int Orthop 2008;32:243–246.
7. SooHoo NF, Zingmond DS, Ko CY. Comparison of reoperation rates following ankle arthrodesis and total ankle arthroplasty. J Bone Joint Surg Am 2007;89A:2143–2149.
8. Spirt AA, Assal M, Hansen ST Jr. Complications and failure after total ankle arthroplasty. J Bone Joint Surg Am 2004;86A:1172–1178.
9. Young J, May M, Haddad SL. Infected total ankle arthroplasty following a routine dental procedure. Foot Ankle Int 2009;30:252–256.