Vol 10, No 3, May/June 2002 157 Introduced in the 1970s, total ankle arthroplasty (TAA) had unaccept- ably high complication rates com- pared with ankle arthrodesis, the standard for surgical management of advanced ankle arthrosis. In- terest has been renewed based on two factors. The first is experience gained with refined surgical tech- niques, and the second is improved elements of implant design derived from hip and knee arthroplasty applied to the ankle. More recent, less constrained designs require less bone resection and have en- hanced fixation. These factors, combined with a greater awareness of soft-tissue management, compo- nent alignment, and ligament bal- ancing, have resulted in promising intermediate results. The long-term outcome of ankle arthrodesis is cur- rently being examined, especially regarding the development of dis- abling adjacent hindfoot arthrosis. However, the renewed interest in second-generation TAA remains tempered by the poor history of earlier prostheses, the difficulty of perfecting the surgical technique, troublesome complications, and the difficulty of salvage and revision. First-Generation Total Ankle Arthroplasty TAA was developed as an alterna- tive to ankle arthrodesis. Subse- quent reports of severe osteolysis, component loosening, impinge- ment, infection, and soft-tissue breakdown 1 led to its abandonment in most centers. The failures are believed to be a result of poor pros- thesis design, inadequate fixation, poor soft-tissue management, mal- alignment, and lack of soft-tissue balancing. 2,3 Ankle arthrodesis, tibiocalcaneal arthrodesis, or below- knee amputation were often the results of failed TAA. 4 The highly constrained first-gen- eration designs excessively trans- ferred large shear, compressive, and rotatory forces associated with physiologic weight bearing to the relatively small surface area of the prosthesis-bone interface. 3,5 To allow for cementation of the compo- nent, considerable bone resection moved the prosthesis-bone interface into mechanically poor bone, thus decreasing component fixation and stability. 2 As a result of the gener- ous bony resections, subsidence was a common mode of failure, 1 and subsequent revision or arthrodesis was technically challenging. Con- versely, highly unconstrained first- generation implants, while causing less component loosening, were often Dr. Easley is Assistant Professor, Division of Orthopaedic Surgery, Duke University Medical Center, Durham, NC. Dr. Vertullo is Fellow, Division of Orthopaedic Surgery, Duke University Medical Center. Dr. Urban is Resident, Division of Orthopaedic Surgery, Duke University Medical Center. Dr. Nunley is Professor, Division of Orthopaedic Surgery, Duke University Medical Center. Reprint requests: Dr. Easley, Box 2950, Duke University Medical Center, Durham, NC 27710. Copyright 2002 by the American Academy of Orthopaedic Surgeons. Abstract First-generation total ankle arthroplaty designs had unacceptably high compli- cation and failure rates compared with ankle arthrodesis. More recent prosthe- ses have had encouraging intermediate results because of refined surgical tech- niques and improved designs. Mobile-bearing designs theoretically offer less wear and loosening through full conformity and minimal constraint. The less complex fixed-bearing designs avoid bearing dislocation and the potential for added wear from a second articulation. Four second-generation designs have demonstrated reasonable functional outcomes: the Scandinavian Total Ankle Replacement, the Agility Ankle, the Buechel-Pappas Total Ankle Replacement, and the TNK ankle. Intermediate results are promising but should be interpret- ed with care. J Am Acad Orthop Surg 2002;10:157-167 Total Ankle Arthroplasty Mark E. Easley, MD, Christopher J. Vertullo, MBBS, FRACS, W. Christopher Urban, MD, and James A. Nunley, MD Perspectives on Modern Orthopaedics unstable and suffered from impinge- ment on malleoli or soft tissue. Problems With Total Ankle Arthroplasty Achieving a successful arthroplasty of the ankle is complex. While the load-bearing area of the ankle is less than that of either the hip or knee, weight bearing in the ankle is esti- mated to be in the range of between 1.5 to 7 times body weight, 6 depend- ing on activity. At a 500-N load, ankle joint contact area averages 250 mm 2 , compared with 1,120 mm 2 for the knee and 1,100 mm 2 for the hip. 7,8 For similar loads, the forces across the ankle are much greater than at the knee or hip. 9 The ankle is not a simple hinge joint but has complex mechanics involving rotation and translation, 10 with a changing instant center of rotation. The joint has three articulations: tibia and superior talus, fibula and lateral talus, and medial malleolus and medial talus. The mechanical support supplied by the distal tibia and talus declines as a function of distance from the joint. 11 This is especially evident at 1.5 cm from the articular surface. 12 The mechanical properties of the distal tibia bone to resist compression are on average 40% less than those of the talus, making tibial subsidence a concern. However, attempts to reduce bony resection are compro- mised by the need to maximize com- ponent polyethylene thickness and obtain fixation. The soft-tissue envelope sur- rounding the ankle is thinner and less vascular than that around other weight-bearing joints, making it more susceptible to local swelling. Wound edge devascularization may result in wound dehiscence and infection despite meticulous surgi- cal technique. 2 The arthritic process often pro- duces mechanical malalignment, soft-tissue contracture, and/or instability. Unlike total knee arthro- plasty, the surgical principles of achieving proper alignment and soft-tissue balancing with TAA are not yet well established. What is understood is that failure to obtain a plantigrade, neutrally aligned, sta- ble, weight-bearing hindfoot gener- ally results in poor outcomes. 13,14 Cemented and Cementless Implants The majority of early failures with TAA occurred in cemented im- plants. In general, uncemented TAA has better results than did cemented components. 15 However, it is difficult to distinguish poor out- comes resulting from inadequate design from those resulting from method of fixation. The initial earlier implant designs were cemented, whereas more recent, improved de- signs tend to be cementless. How- ever, well-designed cemented im- plants of other weight-bearing joints have demonstrated outcomes simi- lar to or better than those of cement- less implants, rendering the concept of “cement disease” obsolete. Hence, fixation failures that occurred in first-generation cemented TAAs cannot be discussed without regard for the implant’s overall design problems. No randomized study has com- pared different methods of bone fix- ation using the same TAA prosthe- sis. The Scandinavian Total Ankle Replacement (STAR) prosthesis (Waldemar-Link, Hamburg, Ger- many) has been inserted in both cemented and uncemented versions in a nonrandomized consecutive series, with 9-year survivorship analysis done using Rothman’s cor- rection. 16 The cemented version showed an 81.1% survival rate and the uncemented version, 95.4%; however, the consecutive nature of the study introduced a selection bias for the uncemented implant. Ankle Arthrodesis Compared With Total Ankle Arthroplasty Although ankle arthrodesis remains the gold standard for treating advanced ankle arthrosis, it is not without problems. These include nonunion, pseudarthrosis, mal- alignment, wound breakdown, decreased gait speed, poor mobility over uneven surfaces, increased oxygen consumption, need for shoe modification, and infection. 17-19 Although techniques of arthrodesis using compression have decreased the rate of nonunion, relatively recent studies report nonunion rates of 12%. 17 Long-term follow-up studies of successful ankle arthrode- ses reveal that the majority of patients have substantial and accel- erated arthritic changes in the ipsi- lateral hindfoot, ranging from 44% to 100%, but not in the ipsilateral knee or contralateral hindfoot. 20 No randomized prospective study has compared ankle arthro- desis with TAA. Existing compara- tive studies of arthrodesis and TAA have a number of methodologic flaws that make it difficult to com- pare outcomes directly. 21 Although arthrodesis remains the current treatment of choice, inherent prob- lems secondary to eliminating ankle motion remain, including altered gait and a high incidence of arthro- sis in adjacent joints in the foot. Indications for Total Ankle Arthroplasty Indications for TAA are still being defined, particularly because the mean periods of follow-up for all current TAA prostheses are relative- ly short. Appropriate patient selec- tion is critical to successful outcome in any joint arthroplasty, including TAA. Previous recommendations that TAA is contraindicated in any situation 1 have been modified; the Total Ankle Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 158 optimal patient is older, with low demands and multiple joint ar- throsis. 15 Kofoed and Lundberg- Jensen 22 reported similar outcomes in patients less than 50 years old (mean, 46 years) and those more than 50 years old (mean, 64 years). Others 13 have commented that TAA may have a role in less active, non- obese patients less than 50 years old with posttraumatic arthrosis. How- ever, in general, older patients are less likely to place excessive de- mands on the arthroplasty than are young, active individuals. Patients with degenerative change in other joints, such as the subtalar, mid- tarsal, or contralateral ankle, may benefit more from TAA than from arthrodesis because ankle arthrode- sis shifts abnormal loads onto the neighboring joints and thus acceler- ates degenerative changes, espe- cially in the tarsal articulations. 20 Patients with a previous triple ar- throdesis may benefit more from TAA than from a pantalar arthrodesis. Patients with posttraumatic ankle arthrosis, especially younger patients, have worse reported out- comes and more likelihood of revi- sion with TAA than from other causes of arthrosis. 23 TAAs done for rheumatoid arthritis and osteo- arthritis have similar outcomes in comparative studies, 24 although some authors 14 suggest that higher subsidence rates occur in patients with rheumatoid arthritis. Absolute contraindications to TAA include active infection, pe- ripheral vascular disease, inadequate soft-tissue envelope, and Charcot neuroarthropathy. Relative con- traindications include young, active patients, previous infection, severe lower extremity malalignment, marked ankle instability, marked osteoporosis, and osteonecrosis of the talus. Osteonecrosis is a relative contraindication for two reasons: talar components that require a minimal amount of bone resection are prone to subsidence in the face Mark E. Easley, MD, et al Vol 10, No 3, May/June 2002 159 of underlying bone necrosis, 14 and cementless components cannot ob- tain stable osseointegration by ingrowth in the presence of osteo- necrosis. On the other hand, de- signs that remove more talar bone can tolerate focal or superficial bone necrosis. Second-Generation Total Ankle Arthroplasty New Prosthesis Designs Current TAA designs have been developed with several major modi- fications. Four second-generation TAA implants have demonstrated reasonable functional outcomes. Two designs for cementless use are currently FDA class III devices: the STAR and the Buechel-Pappas Total Ankle Replacement (Endotec, South Orange, NJ). The TNK ankle (Nara, Japan) is also designed for cement- less use. The Agility Total Ankle System (DePuy, Warsaw, IN) is an FDA class II device for cemented use. Two design philosophies are ap- parent in these second-generation implants: mobile bearings and fixed bearings. Mobile-bearing ankles are characterized by a moving polyeth- ylene bearing separating the convex talar component from the flat tibial component, resulting in two separate articular surfaces. Fixed-bearing ankles have only one articulation be- tween a tibial and talar component. To understand the two design philosophies, some terms must be defined. Constraint is the resistance of an implant to a particular degree of freedom, such as anteroposterior translation or axial rotation. Ex- cessive constraint leads to early component loosening. For example, the axial constraint in a hinged total ankle is infinite because all axial torque forces are transferred directly to the bone-prosthesis interface. Reducing constraint minimizes transmission of shear forces at the prosthesis-bone interface. Confor- mity is a geometric measure of closeness of fit of the articulation; fully conforming prostheses have articular surfaces with the same sagittal radii of curvature, resulting in full articular contact. Fully con- forming articulations typically have low wear rates because the polyeth- ylene contact stress remains below its fatigue threshold for delamina- tion and pitting. The term “partially conforming” includes a wide range of articulations, from round-on-flat designs to articulations with radii of curvature that vary by only a few millimeters. With fixed-bearing designs, fully conforming articulations create high axial constraint and result in excessive axial loosening torque. Mobile-bearing implants attempt to overcome this constraint-conformity conflict by offering two separate, fully conforming articulations that function together to reduce axial and shear constraint. The mobile bearing increases design complex- ity, adds the risk of bearing disloca- tion, and may generate greater wear particles from the second articula- tion. To reduce constraint, fixed- bearing ankles must be only partially conforming, but as a result they have theoretically higher wear rates because polyethylene contact stresses are increased. With less conformity, wear is typically greater. This is particularly true when the prosthe- sis is not perfectly balanced, in which case edge loading may occur, leading to high contact stress on thin polyethylene. The goal in de- signing a fixed-bearing ankle is to maximize conformity and minimize constraint; however, the optimum configuration is unknown. Fixed- bearing designs have the single articulation that can produce wear particles and also have a much lower risk of dislocation. The STAR and Buechel-Pappas implants have mobile bearings with full conformity and minimal con- straint. They are designed to de- crease polyethylene contact stress and reduce load transfer to the pros- thesis-bone interface while maintain- ing ankle kinematics. Theoretically, wear in these fully conforming artic- ulations occurs primarily by abra- sion. Both of these designs require only minimal bone resection and simply resurface the talus. The Agility and TNK implants have fixed bearings. They decrease the prosthesis-bone interface stress by being less conforming than first- generation designs. The Agility im- plant includes a syndesmotic fusion to prevent tibial component subsi- dence as well as resurfacing of the medial and lateral ankle recesses to enhance fixation and alignment. The TNK design attempts to solve fixation difficulties by means of hydroxyapatite-coated ceramic and a tibial component screw. Theo- retically, wear in these partially con- forming articulations is primarily by delamination and pitting, with some secondary abrasion. STAR Implant Background and Design The initial design of the STAR prosthesis was a cemented, fixed- bearing, two-component arthroplas- ty that was then revised to a cement- ed, three-component, mobile-bear- ing design. The current STAR implant is a cementless, minimally axially constrained, fully conform- ing mobile-bearing implant (Fig. 1). The mobile, ultra-high-molecular- weight polyethylene (UHMWPE) inlay, or meniscus, articulates supe- riorly with a flat cobalt-chrome (Co- Cr) tibial glide plate and inferiorly with a longitudinally ridged convex Co-Cr talar component. This design feature allows for unconstrained motion of the polyethylene inlay on the tibial component in sagittal and coronal translation and axial rota- tion, restricted only by the malleoli. The longitudinal ridge on the talar component restricts the polyethyl- ene motion on the talus to anterior and posterior translation, preventing rotation. The anatomically shaped talar component has wings that re- place degenerate medial and lateral talar facets (a potential source of pain), allow additional load trans- fer, 25 and essentially resurface the talar dome. The tibial component has a porous surface superiorly and two cylindrical bars dorsally for insertion into parallel drilled holes into the tibial subchondral bone. Between 9 and 14 mm of bone is resected to insert the tibial and talar components, both of which are only 2 mm thick. The UHMWPE bearing varies in thickness from 6 to 10 mm; the talus is available in four sizes and the tibia in three. The European version of the STAR ankle lacks a porous surface; osseointegration is achieved by a hydroxyapatite coat- ing. The USA FDA class III version has porous ingrowth Co-Cr surfaces rather than hydroxyapatite. Results Garde and Kofoed 26 found the stability of the STAR prosthesis in a study of eight patients to be similar to that of the normal contralateral ankle. Kofoed has reported interme- diate results of the current cement- less STAR design in 74 ankles. 4 The mean follow-up was 4.4 years (range, 1 to 10 years). Two cases were revised, one for malalignment and the other for subsidence. The remaining 72 TAAs showed no radi- ographic loosening or subsidence. Schernberg 27 reported intermediate results from a European multicenter study of 131 cementless STAR ankles. The 7-year survival rate was 87.3%, with all failures occurring in the first 2 years, suggesting that if Total Ankle Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 160 A B C Figure 1 A, The mobile-bearing STAR implant. Left, The talar component. Center, The UHMWPE mobile bearing or meniscus. Right, The tibial component. Postoperative anteroposterior (B) and lateral (C) radiographs. the STAR is inserted incorrectly, it will fail early. Long-term results of the earlier cemented STAR versions are avail- able. 24 However, results of both the two-component and the later three- component design were grouped together, thus failing to provide insight into the differences in out- come between them. Fourteen-year survivorship analysis in the osteo- arthritis group was 72.7% and in the rheumatoid group, 75.5%. 24 Ten centers in the United States are currently participating in the pivotal FDA trial of the STAR pros- thesis. One center has reported pre- liminary data. Mann et al 28 reviewed 50 STAR ankle replacements in 49 patients (average age, 60 years) with a 28-month average follow-up. The average American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot ankle score improved from 44 to 82 points, and patient satisfaction of 92% was reported. Average range of motion improved from 32° to 39°. Lysis was observed on only 2% of radiographs. The implant survival rate was 96%. The single conversion to arthrodesis was for talar component subsi- dence. Three deep infections were effectively managed acutely with irrigation, débridement, polyethyl- ene exchange, and intravenous antibiotics. Concerns The mobile bearing introduces an extra-articular surface that increases the potential for wear, especially because the bearing has multidirec- tional motion on the tibia that may produce more abrasive wear than occurs with unidirectional motion. 29 Dislocation of the bearing has not been reported 9 with the STAR pros- thesis but could potentially occur in the absence of proper soft-tissue balance. Only the talar aspect of the medial and lateral ankle recesses is resurfaced, and this may be a later potential cause of pain. 9 Results reported by Kofoed and Lundberg- Jensen 22 are of the STAR with hydroxyapatite-coated, nonporous ingrowth Co-Cr surfaces. Hydroxy- apatite coatings have been shown to resorb with time, 30 leaving non- porous Co-Cr to osseointegrate; Co- Cr may be an inferior surface for ingrowth compared with titanium. 31 The ingrowth of Co-Cr is typically incomplete, characterized by “spot welding” with an intervening fibrous layer. As noted, the FDA class III STAR utilizes a porous ingrowth Co-Cr surface rather than a hydroxyapatite-coated surface. Inversion and eversion can occur only when the inlay lifts off one of the other two components, resulting in edge loading. The Buechel- Pappas ankle has a more accommo- dating talar articulation that allows for some inversion and eversion. The STAR ankle’s lack of inversion and eversion may transfer excessive load to the prosthesis-bone interface and cause excessive contact stress on the polyethylene. Theoretically, this effect would be exacerbated by a stiff hindfoot, although to date this has not been a clinical problem. 4 Agility Total Ankle System Background and Design The Agility Ankle is a porous- coated, fixed-bearing implant (Fig. 2) with a partially conforming articula- tion. The modular tibial component consists of a concave UHMWPE insert, available in a variety of thick- nesses, and a titanium tibial tray. The convex Co-Cr talus is able to articulate with the top and sides of the insert, depending on its position in the mortise, thus increasing the load transfer area 5 and avoiding malleolar impingement. This is made possible by a syndesmotic fusion that enables malleolar flanges to resurface the medial and lateral recesses. In addition, the syndesmotic fusion increases the surface area of the tibial compo- nent’s prosthesis-bone interface to resist subsidence while also allow- ing the fibula to share some of the load. To simulate normal transmal- leolar ankle alignment, the talar component articulates with the tib- ial component along an axis that is in 23° of external rotation relative to the longitudinal axis of the tibia. 15 This feature requires that the talar component be tapered posteriorly to avoid impingement against the pos- terior tibial tendon and medial malleolus. The obliquely rectangu- lar talus also affords greater stability and conformity during the stance phase of gait during push-off. Slight translation in the coronal plane and axial rotation is made possible by the partially conforming design, decreasing load transfer to the prosthesis-bone interface. Six different implant sizes are available. Two thicknesses of poly- ethylene inserts (standard and stan- dard plus 2 mm) allow the surgeon to recreate appropriate ligamentous tension. The surgical technique includes the application of an exter- nal fixator that permits controlled distraction with correction of mal- alignment. The design requires more bone removal from the talus than do other designs, which may make this implant more suitable for ankles with significant distortion. 14 The Agility Ankle has undergone four different development phases. The initial design was implanted in 22 ankles. Phase II, with a thickened tibial component, was implanted in 207 ankles. Phase III, with a poste- rior augmentation of the tibial com- ponent, was utilized in 104 patients. Phase IV introduced six different component sizes and represents the current design. Results Intermediate results (range, 2.8 to 12.3 years; mean, 4.8 years) of the first 100 Agility Ankles were independently assessed. 5 Five had undergone revision, three for com- Mark E. Easley, MD, et al Vol 10, No 3, May/June 2002 161 ponent design factors that have since been addressed. Talar loosen- ing requiring revision occurred in two titanium components. As a result, the talar component was reformatted to Co-Cr. Two tibial components fractured (one of which was revised), and the component has been redesigned to be thicker. Of the 82 patients still alive, 79% were satisfied and 13% were ex- tremely satisfied. No deep infec- tions occurred. Of greatest concern was the high association between syndesmotic nonunion or delayed union and migration (19% of ankles), ballooning osteolysis (37%), and cir- cumferential radiolucency (26%). More recently, Saltzman and Alvine 32 reviewed the outcomes of 294 ankles implanted in 280 patients available for a minimum follow-up of 1 year. Ninety-four percent (262 patients [275 TAAs]) stated that they had an improved quality of life; 92% (258 [271]) would have undergone the procedure again; and 95% (265 [276]) would recom- mend the procedure to a friend. On a scale of 0 to 10, with 0 being no pain and 10 the worst imaginable pain, 72% of patients (202) rated their pain as 0 to 3, 16% (46) as 4 to 6, and 9% (26) as 7 to 10. Kaplan- Meier survivorship analysis sug- gested a 14-year survival rate of 61% for phase I ankles, with a revi- sion incidence of 27%, and a 9-year survival rate of 76% for phase II ankles, with a revision incidence of 7%. Survivorship analysis for phase III and IV ankles has not been pub- lished yet. Follow-up data from other current studies are not yet available. Concerns Although it has the advantage of increasing the weight-bearing sur- face and area for bone ingrowth of the tibial component, the syn- desmotic arthrodesis is generally performed through a separate anterolateral incision. This requires considerable soft-tissue dissection and places the anterior skin at risk. To address this concern, many sur- geons perform the syndesmotic arthrodesis through the same ante- rior incision, with a limited lateral approach for screw placement. In addition, delayed union or non- union of the arthrodesis has been reported to occur in more than one third of patients, raising concerns about prosthesis stability in these individuals. 5 In an effort to im- prove syndesmotic fusion rates, Myerson has applied a lateral plate that further compresses the fibula against the arthrodesis site and lat- eral component. This technique shows promise in decreasing not only rates of nonunion but also the incidence of ballooning osteolysis. 33 Some partially conforming joint arthroplasty designs have been reported to exceed polyethylene’s threshold for fatigue failure, pro- ducing accelerated delamination wear at areas of point or edge con- tact. 34 This wear is exacerbated when the polyethylene is <6 mm thick. 35 The standard polyethylene thickness of the Agility Ankle varies from 3.73 mm to 4.7 mm, 5 depend- ing on component size, although thicker “plus 2 mm” polyethylene inserts are also available that in- crease the polyethylene thickness to 5.73 to 6.7 mm, depending on com- ponent size. Using thin polyethyl- ene may predispose the component to failure; however, to use thicker polyethylene requires more bone resection. Total Ankle Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 162 A B C Figure 2 A, The fixed-bearing Agility Ankle. The tibial component’s porous surface is superior and the talar component is inferior, with the polyethylene visible between. Postoperative anteroposterior (B) and lateral (C) radiographs. By resurfacing the medial and lat- eral recesses, a comparatively large amount of bone resection in the medial-lateral plane is necessary. This may render later revision or con- version to arthrodesis more difficult and may increase the possibility of malleolar fracture. Another detri- mental effect of the recess resurfacing is narrowing of the talar component, thus increasing the contact load, especially in smaller ankles. 9 Com- pared with talar resurfacing designs, more bone is resected in the vertical plane (approximately 2.5 cm 9 ), fur- ther reducing bone stock. Occasion- ally, greater talar bone resection leads to talar component subsidence, par- ticularly posteriorly, where the com- ponent is relatively narrow. This problem usually is addressed ade- quately by positioning the posterior aspect of the talar component on the residual posterior cortical rim of the talus. More recently, a modified talar component has been available that is wider posteriorly and includes a flare at the base that affords greater sur- face area and support. A final concern with the Agility Ankle prosthesis is the use of the external fixator for distraction. Although deformity typically can be easily corrected with external fixator adjustments, this method of realign- ment should not be a substitute for proper ligament balancing and con- gruent bony resection. Without proper balance, the malalignment will persist postoperatively when the external fixator is removed after implantation. Buechel-Pappas Total Ankle Replacement Background and Design The Buechel-Pappas ankle re- placement is a mobile-bearing, fully conforming, titanium, porous-coated, cementless design (Fig. 3) devel- oped from the New Jersey LCS (low contact stress) Ankle Replacement (Endotec, South Orange, NJ). Simi- lar to that of the STAR, the design philosophy for the Buechel-Pappas ankle is one that combines mobility and full conformity in an effort to achieve low wear and low con- straint forces. The single-radius, concave talar component necessi- tates the removal of a minimal amount of bone and is stabilized by two fins. The initial design (LCS) used a single fin and shallow sulcus, but the modification (Buechel- Pappas) to two fins and deeper sul- cus appears to reduce the rate of talar component subsidence. Con- trol of medial-lateral translation and prevention of dislocation are pro- vided by a longitudinal sulcus on the talar component that articulates with a matching ridge on the UHMWPE bearing. The sulcus allows some inversion and eversion without producing edge loading, a potential advantage over the STAR design. The superior surface of the UHMWPE bearing articulates with the flat tibial component, allowing unconstrained motion. The tibial component is supported by a cen- tral stem implanted in the tibial metaphysis. The design does not resurface the lateral gutters. To re- duce UHMWPE wear, the Buechel- Mark E. Easley, MD, et al Vol 10, No 3, May/June 2002 163 A B C Figure 3 A, The mobile-bearing Buechel-Pappas Total Ankle Replacement. The tibial component with the stem is superior, the talus with its central sulcus is inferior, and the UHMWPE bearing is between them. (Reproduced from Buechel FF Sr, Buechel FF Jr, Pappas MJ: Eighteen-year evaluation of cementless meniscal bearing total ankle replacements. Instr Course Lect 2002;51:143-151.) Postoperative anteroposterior (B) and lateral (C) radiographs. Pappas ankle uses a nitride ceramic film on the titanium bearing surfaces that has shown improved wear char- acteristics with UHMWPE over Co- Cr in vitro. 36 The Buechel-Pappas ankle is available in six sizes. Results Drzala et al 37 reported the inde- pendently assessed intermediate results of 38 Buechel-Pappas ankles. Mean follow-up was 4.5 years (range, 2 to 8 years). Three of the ankles (7.9%) had undergone revi- sion. Radiographs showed two ankles with eccentric wear and lat- eral bearing subluxation. Two of the three patients with a preoperative diagnosis of osteonecrosis sustained lateral talar collapse. Patients with osteonecrosis had poor results on both the Foot Function Index and Medical Outcomes Study 36-Item Short Form. San Giovanni et al 38 reported the intermediate results of 21 Buechel- Pappas ankles, with a mean follow- up of 5.5 years (range, 3.3 to 9 years). Eighty-six percent of patients rated their pain as mild/occasional. Three ankles (14%) subsided; however, no component suffered circumferential radiolucency or bearing subluxation. Three ankles were considered fail- ures, one for deep infection, one for talar subsidence, and one for tibial subsidence. Buechel et al 39 have presented data with a follow-up of 2 to 18 years (average, 10 years) for the New Jersey LCS ankle and of 2 to 10 years (average, 6.5 years) for the Buechel-Pappas ankle. The rate of talar subsidence was 15% (40 ankles) in the LCS group and 2% (50 ankles) in the Buechel-Pappas group. Ten- year survivorship analysis for the LCS implant was 86.3% and for the Buechel-Pappas, 93.5%. Eighteen- year survivorship for the LCS im- plant was 76.7%. Patient satisfaction (good to excellent) was 75% in the LCS group and 94% in the Buechel- Pappas group. Concerns Drawbacks with the New Jersey LCS ankle are bearing dislocation, fracture of the tibial component, and subsidence of the talar compo- nent. 40 The design changes in the Buechel-Pappas ankle included deepening the sulcus to avoid dislo- cation; however, with any mobile- bearing design, dislocation remains a concern. Other design changes included adding a talar fin to avoid subsidence and thickening the tibial component. However, these design modifications may not prevent simi- lar problems occurring in the long term. Titanium has intrinsically poor wear characteristics compared with Co-Cr. 41 Even though the titanium nitride ceramic film shows improved wear in vitro against UHMWPE, recent studies 42,43 of similarly treated retrieved femoral heads raise con- cern about this technology’s long- term in vivo performance. Increased wear-particle generation from a sec- ond articulation is a common con- cern intrinsic to any mobile-bearing design. The Buechel-Pappas ankle does not resurface either the medial or lat- eral ankle recesses. This may be a source of continued pain. A final concern is that the tibial component needs to be implanted through an anterior cortical window, which compromises the cortical integrity proximal to the implant (the cortical fragment is replaced as part of the procedure). Despite this, the risk of tibial component loosening or subsi- dence appears to be offset by the support afforded by the stem. TNK Ankle Background The current TNK ankle (Nara, Japan) is a fixed-bearing, cementless, ceramic-on-UHMWPE prosthesis with a partially conforming articula- tion (Fig. 4). Bony fixation methods include hydroxyapatite-coated beads and a tibial screw. The design was developed from a variety of prostheses that suffered from high rates of subsidence and failure. One hundred percent of an initial metal- UHMWPE design and 71% of a sub- sequent ceramic-UHMWPE design showed subsidence and loosening at 5 years. 44 The ceramic concave talus articulates with a flat UHMWPE tray that includes a flange to re- surface the medial recess. The UHMWPE tray is secured to the ceramic tibial tray. Results The outcome at intermediate fol- low-up (mean, 4.8 years; range, 2 to 8 years) of 40 current-generation TNK ankles revealed that subsidence con- tinues to be a problem with this implant, especially in patients with rheumatoid arthritis. 44 Twenty-one percent of components showed some degree of subsidence, and 15% of screws had broken. Six ankles (15%) required revision, five for subsidence or impingement because of subsi- dence. Despite this, 77.5% of the implants were described as good or excellent. Concerns The current TNK design has the same articulation as its predecessor, which was marked by loosening and subsidence; they differ only in fixation method. This articulation may transfer excessive shear and torque to the prosthesis-bone inter- face. The prosthesis requires re- moval of more bone stock than do the mobile-bearing designs, result- ing in fixation into mechanically inferior bone. The long-term out- come of a partially conforming ceramic-UHMWPE implant is un- known. Highly conforming ceramic- UHMWPE hip articulations have been reported to undergo rapid fail- ure or fracture. 45,46 In addition, the performance of hydroxyapatite- coated ceramic in osseointegration is unknown. Total Ankle Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 164 Surgical Technique Although several approaches have been described for TAA, most sur- geons prefer an anterior approach, typically utilizing the interval be- tween the tibialis anterior and ex- tensor hallucis longus muscles. 5,47 The importance of a long incision to decrease the tension on the skin, of thick skin flaps to maintain vascu- larity, and of avoiding skin retrac- tors has been emphasized in the lit- erature. Avoidance of the tibialis anterior sheath prevents tendon bowstringing and its resultant wound complications. As noted, the Agility Ankle uses an external fixator to distract the joint and correct alignment; how- ever, lamina spreaders also can be used for this purpose. Most sys- tems have alignment jig systems for undertaking accurate bone cuts; however, instrumentation is still being improved. 9,47 Postoperatively, it is imperative to achieve a stable, neutrally aligned, plantigrade, weight-bearing position of the ankle and hindfoot. 2 Ligament reconstruction, tendon transfers, osteotomies, heel cord lengthening, and arthrodesis may be necessary during the arthroplasty to achieve this goal. Excessive ligamentous laxity can lead to instability, espe- cially in mobile-bearing prostheses. Obtaining equal and adequate ten- sion in the deltoid and lateral liga- ment complexes is essential. 14 With neutral axis bone cuts and ligament balancing, the tensioned joint space should be rectangular when the hindfoot is placed in neutral align- ment. Modified Brostrom repairs, lateral tendon reconstructions, and/or medial deltoid releases are commonly required with preopera- tive varus malalignment. Various sizes of UHMWPE allow incremen- tal tensioning of the ligament com- plexes to achieve a stable joint replacement. Techniques for achiev- ing proper alignment and ligament balance have not been standardized for TAA as they have been for total hip and total knee arthroplasty. However, some consistency is devel- oping in the management of mal- alignment in TAA. As with most surgical procedures, a learning curve exists and even for experi- enced surgeons, TAA is challeng- ing. Future Directions Recently, several new TAA prosthe- ses have been developed, but inter- mediate to long-term results are not available. Most of these implants are mobile-bearing, three-compo- nent designs with features not in- cluded in the implants already dis- cussed. Such features include ana- tomically shaped talar components with dual radii of curvature and separate fibular replacement com- ponents. Preliminary results of these implants are promising. Summary The second generation of TAA implants has shown encouraging intermediate results. Mobile-bear- ing designs theoretically offer less wear and loosening because of full conformity and minimal constraint. Fixed-bearing designs avoid bearing dislocation and the potential for added wear from a second articula- tion. The optimal articulation con- figuration is currently unknown. The STAR, Agility Ankle, Buechel- Pappas ankle, and TNK implant all have some untested features, includ- ing their bearing surfaces and bony fixation methods. Problems inherent in some designs include peripros- thetic radiolucency and subsidence, excessive bone stock removal, and syndesmotic nonunion or delayed union. All four TAAs discussed have features that could be improved. Evidence-based, stepwise introduc- tion of new orthopaedic devices 48 allows safe and controlled imple- mentation of new technologies while exposing as few patients as possible to the potential risk of fail- ure. Multicenter trials of TAA in institutions with adequate facilities for reporting outcomes would pro- vide a sufficient number of patients for statistical analysis free of perfor- mance bias. Current investigations Mark E. Easley, MD, et al Vol 10, No 3, May/June 2002 165 A B C Figure 4 A, The ceramic-on-UHMWPE, fixed-bearing TNK ankle. The tibial and talar components are ceramic, with polyethylene between. Postoperative anteroposterior (B) and lateral (C) radiographs. include finite-element and gait analysis of TAA, pressure measure- ments of articular surface contact stresses, assessment of component micromotion, and evaluation of oxy- gen tension on the anterior skin dur- ing and after the surgical approach. In addition, as occurred with the development of total hip and total knee arthroplasty, objective and con- trolled research is now replacing anecdotal experience with TAA. Total Ankle Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 166 References 1. Bolton-Maggs BG, Sudlow RA, Free- man MA: Total ankle arthroplasty: A long-term review of the London Hospi- tal experience. J Bone Joint Surg Br 1985;67:785-790. 2. Alvine FG: Total ankle arthroplasty, in Myerson M (ed): Foot and Ankle Dis- orders. Philadelphia, PA: WB Saunders, 2000, pp 1085-1102. 3. Neufeld SK, Lee TH: Total ankle ar- throplasty: Indications, results, and biomechanical rationale. Am J Orthop 2000;29:593-602. 4. Kofoed H (ed): Current Status of Ankle Arthroplasty. Berlin, Germany: Springer- Verlag, 1998. 5. Pyevich MT, Saltzman CL, Callaghan JJ, Alvine FG: Total ankle arthroplasty: A unique design. Two- to twelve-year follow-up. J Bone Joint Surg Am 1998; 80:1410-1420. 6. Stauffer RN, Chao EY, Brewster RC: Force and motion analysis of the nor- mal, diseased, and prosthetic ankle joint. Clin Orthop 1977;127:189-196. 7. Newton SE III: Total ankle arthroplas- ty: Clinical study of fifty cases. J Bone Joint Surg Am 1982;64:104-111. 8. Brown TD, Shaw DT: In vitro contact stress distributions in the natural human hip. J Biomech 1983;16:373-384. 9. Saltzman CL, McIff TE, Buckwalter JA, Brown TD: Total ankle replacement revisited. J Orthop Sports Phys Ther 2000;30:56-67. 10. Komistek RD, Stiehl JB, Buechel FF, Northcut EJ, Hajner ME: A determina- tion of ankle kinematics using fluo- roscopy. Foot Ankle Int 2000;21:343-350. 11. Hvid I, Rasmussen O, Jensen NC, Nielsen S: Trabecular bone strength profiles at the ankle joint. Clin Orthop 1985;199:306-312. 12. Kofoed H: Cylindrical cemented ankle arthroplasty: A prospective series with long-term follow-up. Foot Ankle Int 1995;16:474-479. 13. Gould JS, Alvine FG, Mann RA, Sanders RW, Walling AK: Total ankle replacement: A surgical discussion. Part I: Replacement systems, indications, and contraindications. Am J Orthop 2000;29:604-609. 14. Gould JS, Alvine FG, Mann RA, Sanders RW, Walling AK: Total ankle replacement: A surgical discussion. Part II: The clinical and surgical expe- rience. Am J Orthop 2000;29:675-682. 15. Saltzman CL: Total ankle arthroplas- ty: State of the art. Instr Course Lect 1999;48:263-268. 16. Kofoed H: Comparison of cemented and cementless ankle arthroplasty, in Kofoed H (ed): Current Status of Ankle Arthroplasty. Berlin, Germany: Springer- Verlag, 1998, pp 47-49. 17. Mann RA, Rongstad KM: Arthrodesis of the ankle: A critical analysis. Foot Ankle Int 1998;19:3-9. 18. Waters RL, Barnes G, Husserl T, Silver L, Liss R: Comparable energy expen- diture after arthrodesis of the hip and ankle. J Bone Joint Surg Am 1988;70: 1032-1037. 19. Lynch AF, Bourne RB, Rorabeck CH: The long-term results of ankle ar- throdesis. J Bone Joint Surg Br 1988;70: 113-116. 20. Coester LM, Saltzman CL, Leupold J, Pontarelli W: Long-term results fol- lowing ankle arthrodesis for post-trau- matic arthritis. J Bone Joint Surg Am 2001;83:219-228. 21. Kofoed H: Comparison of ankle arthro- plasty and arthrodesis: A prospective series with long term follow-up. Foot 1994;4:6-9. 22. Kofoed H, Lundberg-Jensen A: Ankle arthroplasty in patients younger and older than 50 years: A prospective series with long-term follow-up. Foot Ankle Int 1999;20:501-506. 23. Hintermann B: Short- and mid-term results with the STAR total ankle pros- thesis [German]. Orthopade 1999;28: 792-803. 24. Kofoed H, Sorensen TS: Ankle arthro- plasty for rheumatoid arthritis and osteoarthritis: Prospective long-term study of cemented replacements. J Bone Joint Surg Br 1998;80:328-332. 25. Lambert KL: The weight-bearing func- tion of the fibula: A strain gauge study. J Bone Joint Surg Am 1971;53:507-513. 26. Garde L, Kofoed H: Meniscal-bearing ankle arthroplasty is stable: In vivo analysis using stabilometry. Foot Ankle Surg 1996;2:137-143. 27. Schernberg F: Current results of ankle arthroplasty: European multi-center study of cementless ankle arthroplas- ty, in Kofoed H (ed): Current Status of Ankle Arthroplasty. Berlin, Germany: Springer-Verlag, 1998, pp 41-46. 28. Mann RA, Mann JA, Jaakkola J, Kennedy MP: Abstract: Short-term results with 50 Scandinavian Total Ankle Replacements. AOFAS 17th Annual Summer Meeting Final Program, San Diego, California. Seattle, WA, American Orthopaedic Foot and Ankle Society, 1999, p 46. 29. Jones VC, Barton DC, Fitzpatrick DP, Auger DD, Stone MH, Fisher J: An experimental model of tibial counter- face polyethylene wear in mobile bear- ing knees: The influence of design and kinematics. Biomed Mater Eng 1999;9: 189-196. 30. Soballe K, Overgaard S: The current status of hydroxyapatite coating of prostheses. J Bone Joint Surg Br 1996; 78:689-691. 31. Head WC, Bauk DJ, Emerson RH Jr: Titanium as the material of choice for cementless femoral components in total hip arthroplasty. Clin Orthop 1995; 311:85-90. 32. Saltzman CL, Alvine FG: The Agility Total Ankle Replacement. Instr Course Lect 2002;51:129-133. 33. Myerson MS: Salvage after complica- tions of total ankle arthroplasty. Foot Ankle Clin, in press. 34. Collier JP, Mayor MB, McNamara JL, Surprenant VA, Jensen RE: Analysis of the failure of 122 polyethylene inserts from uncemented tibial knee compo- nents. Clin Orthop 1991;273:232-242. 35. Engh GA, Dwyer KA, Hanes CK: Polyethylene wear of metal-backed tibial components in total and unicom- partmental knee prostheses. J Bone Joint Surg Br 1992;74:9-17. 36. Pappas MJ, Makris G, Buechel FF: Titanium nitride ceramic film against polyethylene: A 48 million cycle wear test. Clin Orthop 1995;317:64-70. 37. Drzala M, Lin SS, Eng KO: Abstract: Independent evaluation of Buechel- Pappas second-generation cementless total ankle arthroplasty intermediate