Revision of the aseptic and septic total ankle replacement

Revision of the Aseptic and Septic Tot a l A n k l e R e p l a c e m e n t Norman Espinosa, MD*, Stephan Hermann Wirth, MD KEYWORDS  Revision  Total  Ankle  Replacement KEY POINTS  Total ankle replacement has become popular in the treatment of ankle osteoarthritis.  Longevity of total ankle replacement is still limited.  Revision total ankle replacement represents an appealing solution to maintain function and to protect the adjacent joint.  In nonsalvageable total ankle replacement, arthrodesis is a good solution. INTRODUCTION In 1970, Lord and Marotte1 were the first to implant an unconstrained cemented artificial ankle joint, which sparked interest in ankle replacements and was followed by other surgeons using a multitude of implants. As a result of an overly constrained design in combination with cemented fixation, high shear stresses along the bonecement-implant surfaces were induced. The high shear stress led to impaired osseous integration and premature failure of total ankle replacement. As a consequence of the very high failure rate of total ankle replacement in the 1970s and 1980s, there was a period in which total ankle replacement was almost completely abandoned in clinical practice.2,3 A more profound understanding of ankle biomechanics and an improved design of total ankle replacements has led to the evolution of better second-generation and third-generation prostheses and resurgence of interest in this procedure. Increasing experience and availability of modern total ankle replacement has led to the stretching of indications (ie, total ankle replacement implanted in younger patients or severe deformity), which will result in a higher rate of failure than can be expected from the reported literature. Longevity of total ankle replacement remains a problem. There are few total ankle replacement designs that offer the possibility to exchange the Department of Orthopaedic Surgery, University of Zurich, Balgrist Hospital, Forchstrasse 340, Zurich 8008, Switzerland * Corresponding author. E-mail address: [email protected] Clin Podiatr Med Surg 30 (2013) 171–185 http://dx.doi.org/10.1016/j.cpm.2012.10.004 podiatric.theclinics.com 0891-8422/13/$ – see front matter Ó 2013 Elsevier Inc. All rights reserved. 172 Espinosa & Wirth prosthesis, and few surgeons have experience with revision total ankle replacement. Therefore, in the case of a failed total ankle replacement, conversion to an ankle arthrodesis has remained the gold standard. However, arthrodesis of the ankle joint leads to abnormal biomechanical transmissions of forces and gait alterations. As a result, the adjacent joints compensate for the loss of motion and become overloaded and arthritic, and therefore the use of ankle fusion should be weighted critically.4 In order to preserve motion at the ankle, and thus to protect the adjacent subtalar and midtarsal joints, the ability to exchange the implant successfully is desirable. However, there is sparse information in the literature regarding revision total ankle replacement.5,6 BIOMECHANICAL ASPECTS OF TOTAL ANKLE REPLACEMENT BEHAVIOR AND FAILURE During normal gait the ankle joint is loaded with a force approximately 6 times body weight.7 This force is reduced to 3 times body weight in a degenerated ankle joint.8 However, it is assumed that for total ankle replacement, the strength of bone should be at least 3 times greater than in normal conditions. Proper fixation techniques are needed to compensate for those forces exerted under high-performance activities to avoid any subsidence of the components. In addition, under optimal conditions, the ultrahigh-molecular-weight polyethylene (UHMWPE) insert should be as thick as possible to avoid premature wear. UHMWPE wear depends on geometry, strength (ultrastructure), and alignment of the components.9 At present, there is no appropriate recommendation about the UHMWPE thickness that should be used in total ankle replacement. From a logical standpoint, an optimal UHMWPE insert should be thin and strong without the risk of impairing bony strength at the bone-implant interface. In addition, a perfect prosthesis should replicate the ankle joint in an anatomic way and mimic kinetics and kinematics of a normal joint.10–12 Thus, a total ankle replacement should maximize conformity and optimize constraint. The high conformity of bearing surfaces avoids peak pressures and wear. In contrast, an artificial ankle joint needs sufficient constraint to provide stability but without increasing shear stresses at the bone-implant interface, which could lead to premature failure of the implant. Contemporary 3-component total ankle replacement designs are more anatomic, present improved biomechanical performance, and use biologic integration of the components.13,14 The surfaces are covered with calcium-hydroxyapatite variably combined with the porous coating of the component. The advantages of an anatomic design and biologic cementless fixation include less extensive resections of the tibia and talus, smaller sizes of implants, reduction of body wear, and avoidance of heat destruction of the soft tissues and bones. These advantages make revision total ankle replacement and easier conversion into an ankle arthrodesis after failed primary total ankle replacement a possibility. Although the results of cemented and uncemented first-generation total ankle replacement designs were limited and disappointing, contemporary third-generation 3-component designs with meniscal bearing fulfill anatomic and biomechanically sound criteria. A more anatomically designed total ankle replacement better replicates the normal ankle joint range of motion with better tolerance of congruent mobile bearing designs with regard to malalignment and even pressure distribution within the ankle joint compared with a 2-component fixed bearing design.10–12,15 When a total ankle replacement becomes loose, the tibial and talar components behave abnormally with increased motion in the frontal, transverse, and sagittal planes. This abnormal movement of components results in increased stress transmission across the supporting bone. According to Wolff’s law, remodeling processes take place resulting in either strengthening or weakening of the osseous ultrastructure. Revision of the Total Ankle Replacement In the case of tibial component loosening, the ring-shaped cortex at the metaphysis of the tibia becomes sclerotic and, in the center, a reduction of cancellous bone mass or formation of cysts takes place. Some total ankle replacement designs include fixation stems (eg, STAR Ankle, Small Bone Innovations, Inc, Morrisville, PA; INBONE, Wright Medical Technology, Arlington, TN; Mobility, DePuy Orthopedics, Warsaw, IN; Salto Tolaris, Tornier, Edina, MN). It is important to anticipate greater bone defects in these types of total ankle replacement during removal of the implant.16 Therefore, a flat revision component of the tibia that holds well against the cortical ring of the tibial metaphysis and does not rely on the weakened or absent cancellous bone is preferred. This cortical rim support provides enough support to act against possible tibial subsidence. In the case of talar component loosening, the component starts to swing in an anterior-posterior and proximal-distal fashion. This effect results in condensed or sclerotic bone mass in the anterior and posterior parts of the talus. These locations are also where cyst formations can be found. REVISION RATES OF CONTEMPORARY TOTAL ANKLE REPLACEMENT DESIGNS The design of a total ankle replacement plays an important role in its biomechanical behavior and therefore influences overall results.3 Recently published short-term and midterm data for the HINTEGRA (NEWDEAL International, Lyon, France), a highly anatomic design, showed revision rates of approximately 7% to 14%.13,17 These rates are less than the 23% revision rate identified by SooHoo and colleagues.18 More sophisticated techniques and improved instrumentation combined with the experience of a surgeon can achieve low infection rates (3% superficial, 1% deep). Hintermann and colleagues17 reported loosening of the talar component in 5.5% of cases. However, almost no loosening of the tibial component was observed (0.7%). Henricson and colleagues19,20 reported revision rates for the HINTEGRA between 14% and 22%.19,20 In contrast, the short-term results for the first 100 Mobility total ankle replacements (DePuy Orthopedics, Inc. Warsaw, IN) were reported by Wood and colleagues21 and the revision rate was even lower (5%) than that reported by Hintermann and colleagues.17 However, 10-year survival rates of the HINTEGRA or Mobility total ankle replacements are currently not available. Long-term studies are needed to prove whether those designs show any superior behavior and outcome compared with the 2-component and 3-component total ankle replacement designs that are presently available. The direct comparison between HINTEGRA and Mobility total ankle replacements did not reveal any difference regarding complication rates.14,22 Both designs preserve enough bone stock and the HINTEGRA design offers the possibility of revision total ankle replacement.22 More recently, Schenk and colleagues23 presented the results of 401 Salto Tolaris total ankle replacements after a mean follow-up of 29 months. The calculated 5-year survivorship was 88% and the revision rate 11%. Regarding the STAR Ankle, there are numerous reports in the literature. With specific focus on revision rate, values ranging between 14% and 48% for singlecoating designs and 7% and 24% for double-coated designs have been derived from national registries in Scandinavia.24,25 Nunley and colleagues26 recently reported a low revision rate for the STAR Ankle, averaging 6%. GENERAL THOUGHTS ON REVISION TOTAL ANKLE REPLACEMENT Reasons for total ankle replacement failure include aseptic and septic loosening.27 Aseptic loosening may occur secondary to poor osseous integration, inaccurate sizing, malalignment, and UHMWPE insert wear.15 The surgeon performing revision surgery faces serious problems. Loss of bone stock occurs as a result of resection 173 174 Espinosa & Wirth for prosthetic implantation and/or secondary wear with periprosthetic osteolysis. In addition, the soft tissue is vulnerable, especially in rheumatoid patients, making salvage surgery at the ankle more difficult than revision of failed total ankle replacements in other joints. In addition, variable degrees of fixed hindfoot deformities and contractures, which may be caused by concomitant subtalar osteoarthritis and tibial or talar component subsidence, can complicate revision surgery.22 The presence of poor bone quality impairs fixation and therefore specific fixation strategies must be selected. Any imbalance at the ankle must be detected and addressed to prevent malalignment of the total ankle replacement, which has detrimental effects on longevity of the implant if not corrected.15,22,28 This process includes assessment of possible incompetence of the lateral or medial ligaments. Osteotomies or arthrodesis are occasionally required to balance and stabilize the hindfoot to restore and maintain neutral alignment. Therefore the most important question is whether the surgeon can implant a new total ankle replacement under stable conditions. There is a paucity of reports in the literature regarding the treatment of failed total ankle replacement, with no clear indication of how to proceed in such difficult cases.5,6,29 There are certain thoughts that should be considered before embarking on revision total ankle replacement. Table 1 lists some principles, as discussed by Hintermann and colleagues.22 As long as those thoughts are considered, revision total ankle replacement is likely to be successful. Septic loosening of total ankle replacement is a serious condition that affects not only the osseous but also the soft tissue parts. Despite this, it can lead to serious systemic and life-threatening medical conditions if not treated well. As such, the indication to revise such an ankle is more aggressive than that for an aseptic loosening. PREOPERATIVE ASSESSMENT The patient should be examined barefoot during walking and in a standing position, followed by evaluation of leg and hindfoot alignment. Equinus contracture involving either the gastrocnemius or the Achilles tendon should be determined because they may play an important role in correcting the hindfoot and must be addressed surgically Table 1 Considerations for total ankle replacement revision surgery Osseous support The components should be placed on well-perfused and viable bone stock A 3-point support is optimal for stability Any osseous defect should be filled with either autogenous graft or allograft Ligamentous tensioning The goal is to restore the joint line The larger a bone defect, the greater the components or custom-made components Medial malleolar or fibular osteotomies might be considered to improve ligament tension Ligament reconstructions can help to improve stability at the ankle Hindfoot alignment A correctly aligned hindfoot supports a balanced total ankle replacement Corrective osteotomies should be liberally used Fusions of the hindfoot and midfoot can be necessary to create a stable and well-aligned socket for total ankle replacement Revision of the Total Ankle Replacement if present. Potential deformities (eg, varus or valgus malalignment and midfoot pronation or supination) and their flexibility are assessed. Rotatory alignment of the hindfoot is assessed using both malleoli to mark out the axis and comparing it with the patella. In addition, the condition of the soft tissues and neurovascular status must be evaluated. Performance of standardized weight-bearing anterior-posterior and lateral radiographs of the ankle follows the thorough clinical examination. The hindfoot alignment views as described by Saltzman and colleagues or, preferably, a long leg axial view are used to assess any valgus or varus deformity and to evaluate prosthetic migration and bone loss.30–32 In order to rule out adjacent joint arthritis and to assess possible peritalar instability, anterior-posterior and lateral views of the foot are mandatory. The anterior-posterior and lateral views of the ankle allow proper assessment of the tibial component in the frontal and sagittal plane. However, the bone stock underneath the talar component cannot be accurately determined with plain radiographs. In those cases, computed tomography (CT) is helpful to determine the extent of bony destruction and to anticipate possible need for grafts or custom-made total ankle replacement components (when there is insufficient remaining talus). The use of single-photon emission CT and fluorodeoxyglucose positron emission CT might be helpful to identify pathologic processes around the total ankle replacement components.33–35 SURGICAL MANAGEMENT If the joint can be preserved/salvaged, the authors also use the HINTEGRA total ankle replacement system (Newdeal, Lyon, France) for primary and revision total ankle replacement. When considering revision total ankle replacement, the authors refer to an algorithm proposed by Hintermann and colleagues,22 which is based on the size of bony defect at either the tibial or talar site. The standard tibial component of the HINTEGRA has a thickness of 4 mm. There are revision tibial components available with 8-mm and 12-mm thickness but they are not frequently used because most revision cases can be addressed by sole implantation of a standard tibial component. The talar revision component has a flat undersurface and long pegs to provide strong fixation within the talar bone. The shape of the talar component is conical with different medial and lateral radii, and therefore is as anatomic as possible. The technique of revision total ankle replacement using the HINTEGRA system as performed by the senior author is described later.36 The same anterior approach as used in the primary total ankle replacement is used. The skin conditions in the anterior part of the ankle joint are critical. Careful handling of the soft tissues is mandatory to prevent wound healing problems such as skin necrosis. Therefore, no sharp forceps or retractors are used during surgery. Any thickened scar tissue in front of the total ankle replacement needs to be removed and, at times, there is also the necessity to remove osseous debris to access the failed total ankle replacement. To remove the total ankle replacement from the underlying bone it is recommended to use osteotomes or chisels. In our daily practice, we always obtain 3 different samples of tissue, which are sent to pathology to rule out or to confirm an infectious process. In cases of septic loosening and/or purulent infection, a 2-stage procedure is performed starting with removal of the total ankle replacement and insertion of an antibiotic-loaded polymethyl methacrylate cement spacer. If it is not possible to salvage the infected total ankle replacement, a 2-staged conversion to an arthrodesis should be considered. Patients who qualify for ankle arthrodesis are those with aseptic loosening of their total ankle replacement and/or eradicated prosthetic joint infection associated with a massive bone loss and the unfeasibility 175 176 Espinosa & Wirth of revision total ankle replacement. In general, patients suffering from early prosthetic joint infection (ie, present for less than 3 weeks), with high susceptibility of the microorganisms against antibiotics, good soft tissue conditions, and adequate implant stability could be treated by retention of the implant. Otherwise prosthetic joint infection results in implant removal. If the implant is removed, the prosthetic components are removed and the joint debrided, followed by implantation of an antibiotic-loaded cement spacer. Adequate parenteral antibiotic treatment is administered for 4 to 8 weeks. After this period, antibiotic treatment is halted and the infectious parameters checked to make sure that infection no longer exists. In the absence of infection, ankle arthrodesis can be considered. Fig. 1 shows an algorithm of how to approach the infected total ankle replacement.27 If the prosthetic implant can be retained, the total ankle replacement components are carefully removed while avoiding further damage to the adjacent osseous structures. In total ankle replacement with tibial stems, it is necessary to create anterior cortical windows. It is crucial to limit this window to a minimum because any resection of the anterior distal tibia cortex weakens the bone and therefore impairs fixation of subsequent total ankle replacement. This step is followed by a thorough inspection of the tibial and talar remnants. Cysts are debrided until the subchondral bone plate is visible. Then the cysts are filled with either allograft or autograft bone impacted into place. For this purpose, a surgeon might have to harvest bone from the proximal tibia or iliac crest. The medial and lateral gutters of the ankle are then cleaned out. The posterior capsule is resected while avoiding the neurovascular structures at the posteromedial aspect of the ankle. By means of the alignment jig of the revision total ankle replacement, the tibial cut is made from anterior to posterior. The tibial resection should be limited to a minimum. The goal is to obtain a flat cut while preserving the Fig. 1. An algorithm of how to proceed in a case of septic total ankle replacement. (Data from Refs.16,43–46) Revision of the Total Ankle Replacement cortical ring of the tibial metaphysis. The talar cut is made flat and parallel to the tibial plane. Using a distractor on the medial part of the ankle joint, neutral alignment with balanced ligamentous tension is achieved. Sometimes release of the collateral ligaments is needed to achieve proper balance. The distance between the tibial and talar surfaces is measured. According to Hintermann and colleagues,22 in the case of a distance up to 18 mm, standard tibial and talar components can be used (Fig. 2). When the gap is greater than 18 mm but is less than 25 mm, a standard tibial and revision component can be used. Gaps with distances more than 25 mm and almost no talar body left require custom-made revision total ankle replacement components. If the tibial cut is too large, a revision tibial component should be considered, which is fixed with 2 screws placed from anterior to posterior. The trial components are inserted and the stability of the ankle joint is checked. Once a stable condition is achieved, the final components are inserted. Sometimes it is necessary to fill the medial and lateral gutters with autologous or allogenic bone graft to enhance stability (Fig. 3). ADDITIONAL SURGERY Malalignment and other potential reasons for instability should be addressed at the time when revision total ankle replacement is performed. Adjusting the tibial cuts can compensate a varus or valgus malalignment of up to 10 . Greater deformities should be corrected either by supramalleolar (closing or open wedge) or calcaneal osteotomies (medial or lateral sliding or Z-shaped). Discrepancies in fibular length are addressed by distraction together with bone block insertion (if too short) or shortening (if too long) (Fig. 4). In the case of lateral ligamentous instability, a repair of the anterior talofibular ligament, the calcaneofibular ligament, or both should be Fig. 2. Decision making in a case of loss of bone stock. (From Hintermann B, Barg A, Knupp M. [Revision arthroplasty of the ankle joint]. Orthopäde 2011;40(11):1000–7. [in German]; with permission.) 177 178 Espinosa & Wirth Fig. 3. Anterior-posterior (A) and lateral (B) radiographs of a female patient suffering from symptomatic posttraumatic ankle osteoarthritis following plate and screw removal. Anterior-posterior (C) and lateral (D) radiographs after implantation of an Agility total ankle replacement. Anterior-posterior (E) and lateral (F) radiographs 5 years after surgery with painful aseptic osteolysis and pain. Anterior-posterior (G) and lateral (H) radiographs after revision total ankle replacement. The Agility total ankle replacement has been replaced by a revision HINTEGRA prosthesis. Autologous iliac crest graft has been inserted to fill the defects in the medial and lateral gutters. performed. When there is no viable ligament tissue left, reconstruction of the lateral ligaments by transfer of an allogenic or autologous free hamstring tendon graft (gracilis or semitendinosus) should be considered. In the case of marked anterolateral ankle instability, a peroneus longus to peroneus brevis tendon transfer can be effective. Arthritic changes in the adjacent joints of the ankle that are associated with hindfoot, midfoot, and forefoot deformity should be addressed by arthrodesis to create a stable and well-aligned socket for revision total ankle replacement. After surgery, the patient is allowed to ambulate in a walking cast or removable boot. In patients who have had additional surgery on their feet, partial or non–weightbearing regimens are recommended. RESULTS AFTER REVISION TOTAL ANKLE REPLACEMENT In contrast with hip and knee revision surgery, there is almost no information available regarding revision total ankle replacement; information is limited to case reports only.5,6,29,37 Hintermann and colleagues22 published the largest series in the German literature. In their study, 83 revision surgeries in 79 patients were performed. Fiftythree percent of cases revealed an aseptic loosening, 41% suffered from painful dysfunction, and 6% from a septic loosening of the total ankle replacement. Fiveyears after surgery, 83% of patients were satisfied with the result, 14% judged the Revision of the Total Ankle Replacement Fig. 4. Anterior-posterior (A) and lateral (B) radiographs following HINTEGRA total ankle replacement in a male patient suffering from a pes cavovarus foot deformity and unrecognized peritalar instability as shown by parallelism between the talus and the calcaneus. Note the remarkable varus malalignment. Anterior-posterior (C) and lateral (D) intraoperative image intensification views following derotational subtalar arthrodesis and revision total ankle replacement (tibial component). Note that the talus and its relationship to the tibia and calcaneus have been improved. In addition, to balance the hindfoot a fibular lengthening osteotomy with interposition of allogenic bone graft has been performed and secured with 2 screws. A shortening fibular osteotomy has been added to correct hindfoot axis. Note the correct alignment of the hindfoot. result as fair, and 2% as poor. Of all patients, 59% were pain free at time of follow-up with an acceptable range of motion at the ankle joint (34 ). In addition to exchange of the metallic components, 36 concomitant procedures (ie, arthrodesis, osteotomies, ligament repairs, and peroneus longus to peroneus brevis transfers) were performed to balance the hindfoot. WHEN ARTHRODESIS IS REQUIRED If a total ankle replacement cannot be retained, arthrodesis represents a viable solution. We prefer to use an anterior compression plating system.16,27 The septic total ankle replacement is approached through an approximately 10-cm longitudinal incision, which is made over the anterior aspect of the ankle joint. Care is taken not to injure 179 180 Espinosa & Wirth the superficial peroneal nerve and its terminal branches. The sheath of the anterior tibial tendon is opened laterally together with the proximal extensor retinaculum. The incision is deepened between the tendons of the anterior tibial and extensor hallucis longus muscles. The neurovascular bundle is retracted laterally together with the extensor hallucis longus tendon. The joint capsule and periosteum are incised. A laminar spreader is inserted medially into the tibia and talus and the total ankle replacement explored. In the case of component loosening, hardware removal can be achieved with a chisel. Areas of remaining fibrous, sclerotic, and necrotic tissue must be debrided by means of an oscillating saw (to create even surfaces) and/or curette and chisel until vital subchondral bone is available. The surfaces of the tibia, talus, and fibular are drilled with a 2mm drill. Depending on the magnitude of bone loss, tricortical iliac crest autograft only or a combination of iliac crest autograft and femoral head allograft is used. The defect zone is bridged by means of the bone graft(s). The surfaces are compressed while carefully checking the hindfoot position. The desired alignment of an arthrodesis is neutral dorsiflexion, 5 of hindfoot valgus, and an equal or slightly more external rotation than the opposite foot to allow adequate propulsion over the medial ray at final stance. In a completely stiff foot, a slight dorsiflexion position of the ankle joint could be considered. Fixation is achieved by means of anterior double plating. Double plating can either be done by using two 3.5-mm, 5-hole to 6-hole, limited contact dynamic compression titanium plates or by applying more anatomic designs (Tibaxys System, Newdeal, Lyon, France). Both techniques use the plates, with 1 being placed anterolateral and the other more anteromedial. To obtain adequate compression, we recommend using an Arbeitsgemeinschaft fur Osteosynthesefragen (AO) tensioning device. Together with the Achilles tendon, the anterior double-plating system acts as a tension band system that puts the ankle joint under maximum but evenly distributed compression. The fibular is fixed to the tibia and talus with 2 to 3 titanium screws (3.5 mm). Cancellous bone graft is harvested from the proximal tibia or iliac crest. After thorough irrigation, the cancellous bone is circumferentially inserted, filling all defects, building a bridge of autologous bone between the distal tibia and the talar remnant. Before wound closure, the tourniquet is deflated and hemostasis obtained. A closed suction drain is inserted to prevent hematoma formation. The joint capsule and the tendon sheaths as well as the subcutaneous tissues are closed with single monofilament absorbable sutures and the skin is closed with a single nonabsorbable suture. A split short leg plaster cast is applied over a sterile dressing. The patient is kept in bed for 48 hours with the operated leg elevated. After this time, a well-padded removable short leg cast is applied. The patient is mobilized on 2 crutches and is non–weight bearing until complete wound healing is verified. When verified, the patient can augment weight bearing up to 15 to 25 kg for an additional 6 weeks. Sutures are removed after 2 weeks. At 8 weeks after surgery, the patient is checked clinically and radiographically. If radiographic assessment shows sufficient signs of osseous consolidation at this point, weight bearing is progressively increased. Full healing might take 3 to 6 months. If there is any doubt of consolidation and bone graft integration, we recommend the performance of CT. Deep venous thrombosis prophylaxis should be continued until cast-free full weight bearing is achieved. The patient is provided with shoe modifications to improve the gait pattern and to save the adjacent joints from early overload and progressive arthrosis. RESULTS AFTER FUSION Little information is available regarding the management of failed total ankle replacement by means of ankle arthrodesis. Zwipp and Grass38 reported on 4 patients Revision of the Total Ankle Replacement undergoing ankle arthrodesis after failed total ankle replacement. Two of them were done by screw fixation alone, whereas the remaining 2 failures were treated by anterior plating using two 3.5-mm limited contact dynamic compression titanium plates. Of the latter, 1 patient needed revision surgery because of nonunion. Groth and Fitch39 described ankle arthrodesis without bone grafting. However, such a procedure leads to significant shortening. Hopgood and colleagues40 published their report on 23 ankles that were converted to arthrodesis. Among those there were only 8 cases that used compression screw fixation, but all of them achieved complete union. In patients with rheumatoid arthritis, tibiotalocalcaneal arthrodesis performed better than ankle fusion alone. The investigators of the same study also stated that total ankle replacement design plays an important role in determining whether larger bone grafts should be used to bridge the gap. The more resurfacing the total ankle replacement Fig. 5. Anterior-posterior (A) and lateral (B) radiographs from a hemophilic male patient who received a right Salto Tolaris total ankle replacement 6 years ago. At the time of presentation the patient suffered from pain and the radiographs reveal aseptic loosening of the components with large osteolytic reactions and malalignment of the tibial component. Anterior-posterior (C) and lateral (D) radiographs after revision total ankle replacement using the HINTEGRA system with revision components. Note the tibial revision component with a greater thickness than the standard one and the longer pegs for the talar component. 181 182 Espinosa & Wirth Fig. 6. Anterior-posterior (A) and lateral (B) radiographs from the same hemophilic male patient depicted in Fig. 5 who received a left Salto Tolaris total ankle replacement that had also developed aseptic loosening of the components with less severe osteolytic reactions and malalignment of the tibial component. Anterior-posterior (C) and lateral (D) radiographs after revision total ankle replacement using the HINTEGRA system with standard components. has, the less bone loss and the easier the reconstruction. Culpan and colleagues41 presented a more homogenous series of patients who had had conversion of failed total ankle replacement in ankle arthrodesis. All patients were treated using compression screw fusion with use of tricortical iliac crest grafts. All patients but 1 achieved solid union and no complications were reported. More recently, Berkowitz and colleagues42 reported on salvage arthrodesis after failed total ankle replacement. They compared 12 patients who had an ankle fusion with 12 patients who had tibiotalocalcaneal arthrodesis. In the group with tibiotalocalcaneal arthrodesis, nonunions occurred and were identified as a risk factor for worse outcome. SUMMARY Contemporary third-generation designs of total ankle replacement provide improved anatomic and biomechanical behavior. However, their longevity is limited and even the best total ankle replacement can fail at some point and need revision surgery. Revision of the Total Ankle Replacement Conversion into arthrodesis has remained the mainstay in treating failed total ankle replacement but at the cost of function of the hindfoot. In order to preserve hindfoot motion and function while protecting the adjacent joints, the idea of total ankle replacement exchange is appealing. Current total ankle replacement systems offer the possibility to revise and exchange parts or the complete implant (Figs. 5 and 6). It is important to perform a thorough preoperative evaluation before embarking on revision total ankle replacement. Associated disorders, such as, extra-articular malalignment, instability, and potential causes of impingement, should be identified and corrected at the same time. Recent reports including larger patient populations are encouraging. The future will show whether revision total ankle replacement will yield similar results to those seen in hip and knee surgery. When there is no chance to preserve the total ankle replacement, salvage arthrodesis remains a viable option with reasonable results. REFERENCES 1. Lord G, Marotte JH. Total ankle prosthesis. Technique and 1st results. Apropos of 12 cases. Rev Chir Orthop Reparatrice Appar Mot 1973;59(2):139–51 [in French]. 2. Jensen NC, Kroner K. Total ankle joint replacement: a clinical follow up. Orthopedics 1992;15(2):236–9. 3. Cracchiolo A 3rd, Deorio JK. Design features of current total ankle replacements: implants and instrumentation. J Am Acad Orthop Surg 2008;16(9):530–40. 4. Coester LM, Saltzman CL, Leupold J, et al. Long-term results following ankle arthrodesis for post-traumatic arthritis. J Bone Joint Surg Am 2001;83(2):219–28. 5. Kharwadkar N, Harris NJ. Revision of STAR total ankle replacement to hybrid AES-STAR total ankle replacement-a report of two cases. Foot Ankle Surg 2009;15(2):101–5. 6. Assal M, Greisberg J, Hansen ST Jr. Revision total ankle arthroplasty: conversion of New Jersey low contact stress to agility: surgical technique and case report. Foot Ankle Int 2004;25(12):922–5. 7. Perry J, Schoneberger B. Gait analysis: normal and pathological function. Edited. Thorofare (NJ): Slack Inc; 1992. 8. Valderrabano V, Nigg BM, von Tscharner V, et al. Gait analysis in ankle osteoarthritis and total ankle replacement. Clin Biomech 2007;22(8):894–904. 9. Gill LH. Principles of joint arthroplasty as applied to the ankle. Instr Course Lect 2002;51:117–28. 10. Valderrabano V, Hintermann B, Nigg BM, et al. Kinematic changes after fusion and total replacement of the ankle: part 3: talar movement. Foot Ankle Int 2003;24(12):897–900. 11. Valderrabano V, Hintermann B, Nigg BM, et al. Kinematic changes after fusion and total replacement of the ankle: part 2: movement transfer. Foot Ankle Int 2003;24(12):888–96. 12. Valderrabano V, Hintermann B, Nigg BM, et al. Kinematic changes after fusion and total replacement of the ankle: part 1: range of motion. Foot Ankle Int 2003;24(12):881–7. 13. Hintermann B, Valderrabano V, Dereymaeker G, et al. The HINTEGRA ankle: rationale and short-term results of 122 consecutive ankles. Clin Orthop Relat Res 2004;424:57–68. 14. Lee KT, Lee YK, Young KW, et al. Perioperative complications of the MOBILITY total ankle system: comparison with the HINTEGRA total ankle system. J Orthop Sci 2010;15(3):317–22. 183 184 Espinosa & Wirth 15. Espinosa N, Walti M, Favre P, et al. Misalignment of total ankle components can induce high joint contact pressures. J Bone Joint Surg Am 2010;92(5): 1179–87. 16. Espinosa N, Wirth S, Jankauskas L. Ankle fusion after failed total ankle replacement. Tech Foot Ankle Surg 2010;9(4):199–204. 17. Hintermann B, Valderrabano V, Knupp M, et al. The HINTEGRA ankle: short- and mid-term results. Orthopade 2006;35(5):533–45 [in German]. 18. SooHoo NF, Zingmond DS, Ko CY. Comparison of reoperation rates following ankle arthrodesis and total ankle arthroplasty. J Bone Joint Surg Am 2007; 89(10):2143–9. 19. Henricson A, Carlsson A, Rydholm U. What is a revision of total ankle replacement? Foot Ankle Surg 2011;17(3):99–102. 20. Henricson A, Skoog A, Carlsson A. The Swedish ankle arthroplasty register: an analysis of 531 arthroplasties between 1993 and 2005. Acta Orthop 2007; 78(5):569–74. 21. Wood PL, Karski MT, Watmough P. Total ankle replacement: the results of 100 mobility total ankle replacements. J Bone Joint Surg Br 2010;92(7):958–62. 22. Hintermann B, Barg A, Knupp M. Revision arthroplasty of the ankle joint. Orthopade 2011;40(11):1000–7 [in German]. 23. Schenk K, Lieske S, John M, et al. Prospective study of a cementless, mobilebearing, third generation total ankle prosthesis. Foot Ankle Int 2011;32(8):755–63. 24. Fevang BT, Lie SA, Havelin LI, et al. 257 ankle arthroplasties performed in Norway between 1994 and 2005. Acta Orthop 2007;78(5):575–83. 25. Skytta ET, Koivu H, Eskelinen A, et al. Total ankle replacement: a populationbased study of 515 cases from the Finnish Arthroplasty Register. Acta Orthop 2010;81(1):114–8. 26. Nunley JA, Caputo AM, Easley ME, et al. Intermediate to long-term outcomes of the STAR total ankle replacement: the patient perspective. J Bone Joint Surg Am 2012;94(1):43–8. 27. Espinosa N, Wirth SH. Ankle arthrodesis after failed total ankle replacement. Orthopade 2011;40(11):1008–17 [in German]. 28. Klammer G, Benninger E, Espinosa N. The varus ankle and instability. Foot Ankle Clin 2012;17(1):57–82. 29. Myerson MS, Won HY. Primary and revision total ankle replacement using custom-designed prostheses. Foot Ankle Clin 2008;13(3):521–38. 30. Buck FM, Hoffmann A, Mamisch-Saupe N, et al. Hindfoot alignment measurements: rotation-stability of measurement techniques on hindfoot alignment view and long axial view radiographs. AJR Am J Roentgenol 2011;197(3):578–82. 31. Reilingh ML, Beimers L, Tuijthof GJ, et al. Measuring hindfoot alignment radiographically: the long axial view is more reliable than the hindfoot alignment view. Skeletal Radiol 2010;39(11):1103–8. 32. Saltzman CL, el-Khoury GY. The hindfoot alignment view. Foot Ankle Int 1995; 16(9):572–6. 33. Knupp M, Pagenstert GI, Barg A, et al. SPECT-CT compared with conventional imaging modalities for the assessment of the varus and valgus malaligned hindfoot. J Orthop Res 2009;27(11):1461–6. 34. Pagenstert GI, Barg A, Leumann AG, et al. SPECT-CT imaging in degenerative joint disease of the foot and ankle. J Bone Joint Surg Br 2009;91(9):1191–6. 35. Fischer DR, Maquieira GJ, Espinosa N, et al. Therapeutic impact of [(18)F]fluoride positron-emission tomography/computed tomography on patients with unclear foot pain. Skeletal Radiol 2010;39(10):987–97. Revision of the Total Ankle Replacement 36. Hintermann B. Total ankle arthroplasty: historical overview, current concepts and future perspectives. Edited. Vienna (NY): Springer-Verlag; 2005. 37. Gould JS. Revision total ankle arthroplasty. Am J Orthop 2005;34(8):361. 38. Zwipp H, Grass R. Ankle arthrodesis after failed joint replacement. Oper Orthop Traumatol 2005;17(4–5):518–33 [in English, German]. 39. Groth HE, Fitch HF. Salvage procedures for complications of total ankle arthroplasty. Clin Orthop Relat Res 1987;224:244–50. 40. Hopgood P, Kumar R, Wood PL. Ankle arthrodesis for failed total ankle replacement. J Bone Joint Surg Br 2006;88(8):1032–8. 41. Culpan P, Le Strat V, Piriou P, et al. Arthrodesis after failed total ankle replacement. J Bone Joint Surg Br 2007;89(9):1178–83. 42. Berkowitz MJ, Clare MP, Walling AK, et al. Salvage of failed total ankle arthroplasty with fusion using structural allograft and internal fixation. Foot Ankle Int 2011;32(5):493–502. 43. Trampuz A, Zimmerli W. Prosthetic joint infections: update in diagnosis and treatment. Swiss Med Wkly 2005;135(17–18):243–51. 44. Trampuz A, Zimmerli W. New strategies for the treatment of infections associated with prosthetic joints. Curr Opin Investig Drugs 2005;6(2):185–90. 45. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med 2004;351(16):1645–54. 46. Zimmerli W. Prosthetic joint infection: diagnosis and treatment. Curr Infect Dis Rep 2000;2(5):377–9. 185