Journal of Orthopaedic & Sports Physical Therapy 2OOO;3O (2):56-67

Total Ankle Replacement Revisited

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Charles L. Saltzman, MD Terence E. MclK PhD2 Joseph A. Buckwalter, MD3 Thomas D. Brown, PhD4

The surgical treatment of painful, end-stage ankle arthritis includes ankle arthrodesis and total ankle replacement. In the past decade, total ankle replacement has become a viable alternative to ankle arthrodesis. Modern implant designs either involve a syndesmosis fusion and resurfacing of the medial and lateral recesses of the ankle joint or the use of a 3component, mobile bearing implant. In limited clinical series, the early results of both these prosthetic design approaches are encouraging. In selected patients, ankle arthroplasty is an effective approach to relieving pain and improving function. The purposes of this paper are to review the clinical results from total ankle replacement and ankle arthrodesis; discuss indications, contraindications, design features, postoperative rehabilitation, and initial results for the major current total ankle designs; and present concepts for future total ankle development. In particular, this article explores the advantages and concerns with 2 prevalent but different design approaches. It also discusses future directions for total ankle replacement. ) Orthop Sports Phys Ther 2000;30:56-67.

work revolutionized the surgical treatment of patients with hip arthritis. Approximately 600,000 patients undergo total hip arthroplasty annually for incapacitating arthritic hip pain. The results of these surgeries, at an average of 20 years postoperative, are truly remarkable.= In honor of his critical contributions to the develop ment of joint arthroplasty, John Charnley was knighted by Queen Elizabeth. Sir John Charnley's early discovKey Words: arthritis, arthroplasty, total ankle replacement eries and successes with metal and polyethylene low-friction arthroplasty led to great excitement and optimism in the orthopaedic comhe brilliant and prodigious work of John Charnleyn" ushmunity. His principles were a p ered in the modern era of joint replacement. In the early plied throughout the body to treat 1960s, working mostly on the hip, Charnley developed the painful, arthritic joints. This led to "low-friction arthroplasty" principle whereby a relatively a proliferation of designs and trismall metal femoral head was made to rotate against a als during the early 1970s. Joints polyethylene acetabular cup. He identified the acrylic cement polymethwith a high prevalence of endylmethacrylate as an ideal grouting material to accommodate uniform stage arthritis and relatively easy load transfer between the smooth-surfaced prosthesis and the irregular access for surgical exposure untexture of bone. His first clinical trials involved the use of a Teflon-like derwent rapid and successful dematerial articulating with the bearing surface of the metal femoral velopment of prosthetic replacehead. This was met with immediate and disastrous wear-related failure. ment designs. The knee is now as Undeterred, John Charnley continued his quest, identifying a high-mocommonly replaced as the hip, lecular-weight form of polyethylene (similar to that used in plumbing with equal or better success, and pipes) as an appropriate material for the bearing surface. His seminal shoulder arthroplasty is now a common and predictable procedure. In contrast, improvements I Associate proiessor, Orthopaedic Surgery, Biomedical Engineering, University of lowa, lowa City, in prosthetic designs for joints lowa. with a low prevalence of degeneraResearch fellow, Orthopaedic Biomechanics Laboratory, University o i lowa, lowa City, lowa. tive disease, poor anatomical acProiessor, Orthopaedic Surgery, University of lowa, lowa City, lowa. cess, and less forgiving soft tissue Proiessor, Biomedical Engineering, and Professor, Orthopaedic Surgery, University of lowa, lowa coverage has been slow. In particCity, lowa. Send correspondence to Charles L. Saltzman, Department of Orthopaedic Surgery-01017/Pe Uni- ular, ankle joint replacement deversity o i lowa, 200 Hawkins Drive, lowa City, IA 52242- 1088. E-mail: [email protected] signs are still evolving.

TABLE 1. Satisfaction rates after total ankle replacements--older designs. Device Smith ICLH TPR Bath and Wessex TPR LCS Smith Mavo

Study

Number of patients

Average follow-up, mo

Satisfaction rate, OO/

Dini and Bassett" Coldie and HerbertsI1 jensen and Kronertb Carlsson et al: Kumar" Buechel et alb KirkupI9 Kitaoka and PdtzerI3

21 18 23 52 37 40 18 160

27 36 59 60 60 72 84 108

46 60 69 81 52 85 61 19

ICLH indicates Imperial College, London Hospital; TPR, Thompson-Parkridge-Richards;and LCS, low-contact stress.

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

ANKLE ARTHRITIS-SCOPE OF THE PROBLEM Compared with other major lower extremity joints, the ankle joint possesses unique anatomic, biochemical, and biologic characteristics. Unlike the hip and knee, which are prone to the development of primary osteoarthritis, ankle arthritis is usually the end result of a traumatic event. Ankle articular cartilage has characteristic differences from hip or knee cartilage, which may protect the ankle from developing primary osteoarthritis. Ankle articular cartilage preserves its tensile stiffness and fracture stress better than hip articular cartilage." Biochemical differences between knee and ankle articular cartilage may help to further explain the relative rarity of primary ankle ~steoarthritis.~" Recently, physicians have noted a progressive increase in the incidence of disabling ankle arthritis in developed nations. This increase may result in part from the combined effects of the widespread use of life-protecting thoraceabdominal-level air-bag restraints that d o not protect the ankle and the general aging of the population. The increased incidence of painful posttraumatic ankle osteoarthritis has spurred interest in finding therapeutic solutions to this often functionally limiting problem. Bony anatomy and ligaments of the ankle joint determine the planes and ranges of the ankle's joint motion for a high degree of stability and congruence when the joint is loaded. The ankle joint is a %bone joint involving the tibia, the fibula, and the talus, supporting 3 sets of opposing articular surfaces. The tibia, medial malleolus, and medial facet of the talus form the medial articular surfaces. The fibular lateral malleolus and the talar lateral articular surface form the lateral articular surfaces, and the distal tibia and superior dome of the talus form the superior articular surface. These articular surfaces and the periankle ligaments guide and restrain movement between the talus and the mortise so that the talus has a continuously changing axis of rotation as it moves from maximum dorsiflexion to maximum plantar flexion. The talus and mortise widen slightly from posterior to anterior. Thus, when the talus is plantar flexed, the narrowest portion of the talus remains in the relaJ Orthop Sports Phys Ther-Volume 30. Nriniber 2- Febnian 2 0 M

tively wide mortise, allowing talar rotation. When the talus is maximally dorsiflexed, the tibiofibular syndesmosis spreads, and the wider portion of the talar articular surface locks into the ankle mortise, allowing little o r no talar rotation. When loaded, the human ankle joint has a smaller area of contact between the opposing articular surfaces than does the hip. At 500 Newtons of load, the contact area of the ankle joint averages 250 mm2,2.1x compared with 1120 mm2 for the kneelg and 1100 mm2 for the hip.g Although in vivo contact stress has not been measured in the ankle, the smaller contact area suggests that with equal loading, the spatial average contact stresses in the ankle are higher than in the knee or the hip.

HISTORY OF TOTAL ANKLE ARTHROPLASTY Early results with total ankle arthroplasty were generally disappointing. In the search for a workable ankle design, a number of different approaches have been tried. Our current ability to critically analyze design strategies is limited by the paucity of available data on the results of total ankle arthroplasties. Most clinical series include 20 to 40 patients, followed for an average of 5 years or less (Table 1). Patient satisfaction with firstgeneration, cemented ankle implants varied from 19 to 81~.&i.ll.lJ.I6.lY:10:13 The length of follow-up was a major factor with patient satisfaction because patients with longer follow-up generally had declining degrees of satisfaction. Rates of radiographic loosening with early implant designs were extremely high, ranging from 22 to 75% (Table 2) .~i.l3.l4.16.I!~~LO~13 A review of the data suggests that the major factors implicated with loosening were highly constrained designs and cement fixation. We d o not know whether the use of cement alone or whether the combination of the use of cement and the need to create adequate space for cement fixation was the chief contributing factor leading to these high rates of loosening. Total ankle arthroplasty has also been plagued with an unusually high incidence of wound p r o h lems. Soft tissues around the ankle, especially in patients who are elderly or have rheumatoid arthritis,

,.

w .

--:

TABLE 2. Radiographic loosening after total ankle replacement--older designs. Study

Number of patients

Average follow-up, mo

Loosening, %

Goldie and HerbertsI1 Helm and Stevens" Jensenand KronerIb Carlsson et al: Kumar" Bolton-Maggset all KirkupIq Kitaoka and Patzer1]

18 14 23 52 37 41 18 160

36 54 59 60 60 66 84 108

22 57 52 67 26 32 39 75

Device ICLH ICLH TPR Bath and Wessex TPR ICLH Smith Mayo

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

ICLH indicates Imperial College, London Hospital; TPR, Thompson-Pdrkridge-Richards.

provide a relatively thin envelope for arthroplasty containment. Problems with superficial and deep infections, resection arthroplasty, attempted reimplantations or arthrodeses, and the need for the occasional below-knee amputation to control infection dampened the enthusiasm of orthopaedic surgeons performing total ankle replacement. Following the experience with early designs, many centers decided there was little role for total ankle arthroplasty in the treatment of patients with endstage ankle arthritis. After reviewing the Mayo Clinic experience with the "Mayo" ankle implant, Kitaoka and Patzer stated, "We no longer recommend ankle arthroplasty and the constrained Mayo implant for rheumatoid arthritis or osteoarthrosis of the ankle.""'(~'"") Similarly, after reviewing London Hospital's experience with their ankle implant, BoltonMaggs and associates stated, "In view of the high complication rate, and the generally poor long-term clinical results, we recommend arthrodesis as the treatment of choice for the painful, stiff arthritic ankle regardless of the underlying pathological proc e ~ s . " ~ ( pAfter ' ~ ~ ) publication of such reports and with disappointing personal clinical experiences, most orthopaedic surgeons had abandoned total ankle arthroplasty by the early 1980s and recommended ankle arthrodesis as the surgical treatment of choice for end-stage ankle arthritis.

ANKLE ARTHRODESlS RESULTS Virtually all major medical orthopaedic textbooks currently recommend ankle arthrodesis as the surgical treatment of choice for end-stage ankle arthritis. Indeed, most ankle arthrodeses d o relieve pain, at least in the short-term. Nonetheless, the operation is not without complications, and the long-term results are not completely understood (Table 3).12m Although there are few encouraging reports of the results of ankle fusions at an average of 7 to 10 years after surgery, there are many reports that describe both short-term and long-term problems with ankle fusion^.^' Problems commonly described include eventual development of subtalar and midtarsal degenerative joint disease, pain when walking or standing, decreased functional ability, need for ambulatory aids, and need for permanent shoe modifications. I . J ~ , ~ . . Recently, Y I techniques for performing ankle arthrodesis have improved markedly with modern techniques of limited periosteal stripping, rigid internal fixation, and meticulous attention to alignment and position. The current recommendation for alignment is neutral position in the sagital plane, 5" of valgus in the coronal plane, and equal external rotation to the opposite side (typically 10"). The longterm effects of these improved approaches and vigilance with ultimate ankle positioning remain unknown, however. From a theoretical point of view, a

TABLE 3. Published long-term results after ankle arthrodesis.

Study

Number of patients

Average follow-up, y

36 12 41 41 58 101 62 34

8 8 8 12 9 10 7 8

Said et all" Mazur et all6 Morrey and WiedemanZB Ahlberg and Henricksen' BoobbyeP Morgan et alx Lynch et atz5 Click et all2

Majort complications, O/O

24

*

48 32 21 10 34 6

Continued pain, %

* 25 76 68

* * *

Hindfoot degenerative joint disease, %

>SO 100 50 44 *

* *

*

Data not reported. t Deep infection, nonunion, or amputation. Orthop Sports Phys Ther .\'oltime 30. Nutnher 2. F e h n ~ a n2000

tibiotalar fusion will naturally cause increased stress in contiguous joints that are forced to perform motions the fused ankle is unable to perform. Because increased mean contact stress is a primary causal factor for the development of joint degeneration, the widespread advocacy of ankle arthrodesis as the final treatment for ankle arthritis seems imprudent. These concerns have led several investigators to continue to search for a workable ankle joint replacement that will allow the normal functional range of motion of the ankle joint, reduce the pain and disability resulting from arthritis, and preferably preserve the articular integrity of adjacent articulations.

CURRENT APPROACHES TO ANKLE ARTHROPLASTY

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Indications and Contraindications

didate? What if she is 45, 35, or 25 years old? What if she is a heavy laborer or hopes to continue jogging for exercise? Surgeons caring for patients with knee and hip arthritis previously have addressed the same concerns. A hip or knee fusion is clearly a more functionally limiting surgery than is an ankle fusion. Nevertheless, there is a parallel set of concerns and evolution of clinical thinking. Over the past 3 decades, with improved confidence in prosthetic designs, superior surgical techniques, and the availability of revision components, the minimum age and maximal functional level for hip or knee replacement has gradually declined. Today hip and knee replacements are performed in younger patients with increasing confidence. In fact, in the United States, lower extremity malignancies of young adults are commonly treated with hip or knee joint replacements. Although the clinical issues of the hip and knee are clearly different than that of the ankle, the history of surgeons recognizing and reporting satisfactory results after joint replacement in active patients will stimulate ankle replacement surgeons to test and find the limits of age and activity restrictions with ankle joint replacement. Surgeons should explain the inherent risks, alternatives, and benefits of ankle arthroplasty. Patients must understand that total ankle implants are mechanical devices, and, as with all mechanical devices, will they eventually fail. The longevity of implant depends on several factors, including the loading history of the implant. Vigorous, repetitive impact loading of an implant will lead to earlier mechanical failure. The salvage options are limited. Theoretically, surgeons can replace a mobile bearing, or they may be able to use a larger tibia1 or talar component if there is local bony loosening or destruction. In many instances, revising the initial components may not be possible, and an arthrodesis may be required, possibly using structural autograft from the iliac crest. All these surgeries carry risks of surgical failure and morbidity (including intractable infection requiring amputation); at minimum, they require months of postoperative convalescence. Although the benefits for many are great, for those who develop serious complications, the morbidity can be considerable. The decision to proceed with either an ankle arthrodesis or arthroplasty therefore must be made carefully after considering all these factors.

No consensus exists regarding the indications for ankle arthroplasty. Many surgeons treat all patients with debilitating ankle arthritis that is unresponsive to nonoperative approaches with an arthrodesis procedure. As surgeons gain more experience using modern ankle arthroplasty techniques, surgical confidence in the outcomes of ankle joint replacement has risen, and they are performing ankle replacement more commonly. The surgeons who perform ankle replacement are each defining for themselves the population of patients who may benefit from ankle joint replacement. The "ideal" patient for ankle joint replacement is an elderly person with low physical demands who has good bone stock, normal vascular status, no immunosuppression, and excellent hindfoot-ankle alignment. The patient with bilateral ankle arthritis or ipsilateral hindfoot arthritis who has or will require a triple arthrodesis (fusion of the talonavicular, subtalar, and calcaneocuboid joints) is particularly appropriate to consider as a candidate for ankle arthroplasty because bilateral ankle fusions and pantalar (ankle and triple) arthrodeses function poorly. Contraindications include neuroarthropathic degenerative joint disease (Charcot ankle), active or recent infection, avascular necrosis of the talus (especially for noncemented talar fixation), severe benign joint hypermobility syndrome, nonreconstructable malalignment, severe soft tissue problems around the ankle, and sensory o r motor dysfunction of the foot or leg. The contraindications for ankle arthroplasty are relatively clear-cut. Similarly, for surgeons with expeFixation rience, knowledge, and confidence in their ability to All current ankle arthroplasty designs rely on bony perform ankle arthroplasty, the indications outlined ingrowth for implant stability. There are several maabove for this procedure are fairly straightforward. jor advantages to bony ingrowth. First, compared The real controversy resides in the large gray area with cement fixation, less bone resection typically is that lies between these clinical extremes. Is the 5 5 year-old secretary with unilateral ankle arthritis a can- required because no space between bone and imJ Orthop Sports Phys Ther.\'olume 30.Nu111her 2eFehn1an 2000

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

plexity of the design must increase compared with plant is necessary for fixation. Second, inadvertent designs that replace the superior joint alone. One cement displacement or spillage is avoided. Third, design (TNK ankle)33.:" extends the polyethylene rebecause acrylic cement curing is an exothermic prosurfacing of the tibia along the medial joint surface cess, damage to local soft tissues from the relatively to permit some coverage of the medial malleolar and high levels of heat is avoided by the use of bone intalar articulation. To resurface the lateral side as well growth approaches. Ingrowth prostheses incorporate one of the follow- as the medial, a more complex design is required. ing: beaded surfaces along the interface with bone, a The 2 major design choices are to develop a single tibial prosthesis that resurfaces lateral, medial, and hydroxyapatite layer covering, o r a combination of superior surfaces or to develop a separate lateral-surbeads and hydroxyapatite coating. Beaded surfaces allow bone ingrowth. Micromotion between the pros face replacement component. Currently, there are no designs incorporating an independent lateral thesis and bone must be minimized during the period of ingrowth (usually 6 to 12 weeks). Unsuccessful component because of concerns that loosening bony ingrowth results in fibrous ingrowth, which may would cause failure of any separate component anchored into the distal fibula. The one design that increase the propensity of the component to midoes resurface the lateral joint surface (Agility Angrate. Hydroxyapatite coating encourages molecular bonding between the crystals of the calcium hydroxy- kle) includes a relatively robust metal tibial component with polyethylene resurfacing of the superior, apatite and the bony bed. The time course of this bonding process appears to be shorter (3 to 6 weeks) lateral, and medial articulations. This requires synthan that required for ingrowth into beads. Nonethe- desmosis fusion for stabilization of the tibial component and conversion of the natural %bone ankle less, with time the hydroxyapatite coating ultimately joint into a 2-bone joint. will be totally resorbed, leaving concerns regarding In addition to the issue of what joints should be long-term component fixation. Perhaps the most atreplaced, a second major issue arises regarding motractive current approach is the combined use of a tion. Normal human ankle motion involves rotation beaded surface with hydroxyapatite. Theoretically, and sliding. There is no fixed axis of rotation, but this strategy allows the early fixation of the implant through hydroxyapatite bonding and the later secure rather a continuously moving axis. Prosthetic designs that permit these normal movements will minimize ingrowth of bone within the beaded interstices. interfacial shear between bone and prosthesis and reduce ligament strain. Prosthetic designs that force Component Design abnormal motion may lead to increased bone and implant shear, ligament failure, component loosenThe differences among current ankle prostheses are striking. Many questions still remain regarding ing, migration, and accelerated wear. In the natural ankle, the talus not only dorsiflexes ankle replacement design. The first question is, "Exand plantar flexes, it also internally rotates and exactly what joints need resurfacing?" Of the current ternally rotates approximately 5-10" during normal designs, the articulations that are resurfaced include functioning. Because highly constrained designs have the following: (1) the superior tibiotalar joint alone been associated with failure, perhaps because of in(Buechel-Pappas Ultra Total Ankle, Endotec, Inc, South Orange, NJ)24; (2) the superior and medial ar- creased loading at the interface between the bone and implant, more long-lasting approaches must perticulations (TNK ankle, Kyocera, Co, Nan, Ja(3) the superior joint with a metal-poly con- mit some degree of internal and external rotation. The natural ankle also has inversion and eversion stastruct and a hemiarthroplasty on the medial and latbility conferred from the biconvex shape of the supeeral joints (Scandinavian Total Ankle Replacement rior talar surface. This permits some inversion and [STAR] ankle, Link, Inc, Hamburg, germ an^)^'^; eversion motion when the ankle is plantar flexed, (4) completely resurfacing the medial, lateral, and superior joints (Agility Ankle, DePuy, Inc, Wausau, but there is less intrinsic stability. Optimally, some inInd).x' In a general sense, increasing the number of version and eversion movement should be permitted replaced articulations leads to more involved surgerby the ankle replacement during plantar flexion. ies. For example, a surgery to simply replace the inIn the past 2 decades, 2 entirely different generic approaches to total ankle replacement design have ferior aspect of the tibia and the superior dome of emerged to permit semiconstrained motion. The first the talus involves relatively less dissection and resecapproach is a 2component design with a wider tibial tion and fewer technical challenges than does a proinferior articular surface than talar component, alcedure involving resurfacing of all 3 joint surfaces. lowing some sliding and rotational motion because Currently, we do not know which patients can be of this congruency mismatch. The prototype of this treated effectively by replacing the superior articuladesign is the Agility Ankle (Figure 1). The other gention alone and which patients require replacement of all 3 joints. eral design has been a k o m p o n e n t design. This a p proach incorporates a flat metal tibial plate, a metal To replace the medial and lateral joints, the comJ Orthop Sports Phys Tl~er.\'olnme 90- Number 2 February 2000

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 1. (A) Anteroposterior and (B) lateral view of the Agility Ankle 4 years after implantation in a 78-year-old man with osteoarthrosis. This implant design incorporates a syndesmosis fusion and cementless bead fixation. In this case, the medial malleolar screws were used to fixate an intraoperative fracture.

talar resurfacing component, and a mobile polyethylene bearing between the 2 metal components. From a theoretical point of view, the k o m p o n e n t designs permit flexion and extension at the talo-polyethylene prosthetic interface, and internal and external rotation at the tibio-polyethylene prosthetic interface. Currently, there are two k o m p o n e n t designs: the STAR (Figure 2) and the Buechel-Pappas Ultra Total Ankle (Figure 3). The main differences between these mobile bearing designs are in the capture mechanism of the polyethylene, the amount of intrinsic inversion and eversion stability, and their bone interface anchoring methods, requiring different amounts of bone resection.

THE 2-COMPONENT DESIGN WITH SYNDESMOSIS FUSION Advantages Completejoint r q b l a c a a t The concept with the 2component design incorporating syndesmosis fusion is to permit resurfacing of not only the inferior and superior surface of the tibia and talus, but also resurJ Orthop Sports Phys Ther.Volume SO-Number 2.Febnla1-y 2000

facing of the medial and lateral recesses of the ankle joint. Eliminatingfibular motion The use of this implant involves converting a %bone joint into a 2-bone joint, which from a theoretical point of view would simplify the mechanical problems involving the ankle. In normal gait, the fibula is constantly in motion with small amounts of rotation, as well as superior and inferior and medial and lateral translation. With a satisfactory syndesmosis fusion, this motion and a potential source of continuous shear stress to the ankle joint is eliminated. Greater bony suf@rt f m the tibial component The syndesmosis fusion provides an increased surface area for tibial component fixation. This may be especially important for patients with osteopenic bone. This advantage must be weighed against the cost of increased bone resection.

Concerns Mme surgery Converting a %bone ankle joint to a 2-bone joint relies on successful syndesmosis fusion. This additional operation requires more dissection

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 2. (A) Anteroposterior and (B) lateral view of the Scandinavian Total Ankle Replacement ankle 7 years after implantation in a patient with ankle arthritis secondary to hemochromatosis.This 3-component design involves a floating polyethylene meniscus (marked by the thin wire) and cementless hydroxyapatite fixation. (Photo courtesy of Dr Hakon Kofoed.)

and increases the likelihood of soft tissue compromise. Inadequate fusion of the syndesmosis may also require a revision surgery to obtain component stability. Increased bone resection The need for a noncemented approach in total ankle replacement compels the use of metal-backed components. A metalbacked polyethylene tibial component permitting ingrowth into the metal involves resurfacing of the medial and lateral ankle joint recesses and requires robust design at the medial and lateral superior corners to avoid metal fatigue failure. The size of the component, therefore, must be larger in a 2component design with resurfacing of the medial and lateral ankle joint recesses than in any design that does not resurface the medial and lateral ankle joint recesses. Resurfacing of the medial and lateral ankle joint recesses requires wider cuts into the relatively thin bone of the medial and lateral malleoli. This raises the potential for transecting the medial or lateral malleoli or for tendon injury, especially postero-

medially, with inadvertent laceration of the posterior tibial or flexor digitorum longus tendons. Talar component subsidence With a metal-backed, polyethylene tibial component, size considerations also require that the talar component be smaller than the natural width of the talus. This occurs because the design must incorporate sufficient metal and polyethylene on each side of the talus to minimize metal fatigue or polyethylene wear. As the talar component size narrows, so does the contact area of the talar component on the bony talar bed, thus increasing the contact stresses between the component and the talar bone. Accelerated polyethylene wear To allow sliding of the components, a 2component design must intentionally incorporate noncongruence between the polyethylene inner sleeve of the tibial component and the metal talar component. Noncongruence implies that at some instant of loading, there is a potential for line or even point loading between the polyethylene and the metal. In particular, any varus J Orthop Sports Phys Ther .Volume SO. Number 2. February 2000

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 3. (A) Anteroposterior and (B) lateral view of the Buechel-hppas Ultra Total Ankle 2 years after implantation in a patient with posttraumatic ankle arthritis. This 3-component design involves a floating polyethylene meniscus, a trochlear shaped talar component, and cementless fixation. (Photo courtesy of Dr Frederick Buechel.)

or valgus deformity can lead to line loading (all the force is transmitted to the fixed polyethylene bearing along an edge of the talar component), which will accelerate the rate of polyethylene wear and, theoretically, increase the need for early revision surgery. Salvage fm failure The 2component designs incorporating syndesmosis fusion require removal of approximately 1 in of bone, measured in the superior and inferior direction. Additionally, they require considerable resection of bone mediolaterally. This makes salvage by either revision arthroplasty or secondary fusion difficult. Unlike the knee or hip, which have multiple specifically designed revision systems, there are no versatile revision systems developed for ankle arthroplasty with major bone loss. Likewise, arthrodesis in the face of a minimum 1-in bone stock loss requires either a tibiotalar bone compression with a planned 1-inch minimum leg-length discrepancy or the use of structural autograft or allograft associated with increased risks of nonunion, hardware failure, and the need for reoperation.

Clinical Results The prototypic replacement with 2component design involving syndesmosis fusion is the Agility Ankle J Orthop Sport5 Phys Ther-Volume SO Number 2. February 2000

(Figure 1). The first 100 patients who underwent total ankle arthroplasty using an Agility Ankle were reviewed, independent of the operating urgeo on.^ The mean age at surgery was 63 years old (range, 28-81 years); the average length of follow-up was 4.8 years (range, 2.8-12.3 years). Radiographic follow-up with a minimum of 2 years was reported on 98 ankles, questionnaire follow-up was reported on 86 ankles, and independent clinical evaluation follow-up was reported on 56 ankles. In this serieP' there were 5 major postoperative complications involving 3 talar component revisions, 1 tibial component revision, and 1 total ankle resection with arthrodesis. There were 2 superficial wound infections and no deep infections. At followup, 54% of patients had no pain, 29% had mild pain, 16% had moderate pain, and no patients had severe pain. Most patients were satisfied with their results, with 79% rating their satisfaction level as extremely satisfied, 13% rating their satisfaction as satisfied, and 8% rating their satisfaction as indifferent, disappointed, or unhappy. The radiographic results suggested that early syndesmosis fusion was critical to tibial component stability,"' and 62% had early union, defined as radio-

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

graphic union by 6 months. Of the remaining patients, 29% had delayed union, defined as a union occurring after 6 months postoperatively; 9% had a nonunion. There were 19 cases of component migration with 12 tibial components and 7 talar components demonstrating radiographic migration. Of the ankles with tibial component migration, 8 out of 12 (67%) were associated with delayed or nonunion of the syndesmosis. Talar component migration was independent of syndesmosis union. The results of this study suggest that the use of the Agility Ankle, the prototypic 2component design incorporating syndesmosis fusion, resulted in good, reliable function and radiographic stability in most patients at an average of 5 years follow-up.2YRadiographic data emphasized the importance of syndesmosis fusion and may become an important factor in terms of component survivorship with further followup of this patient cohort.

THE 3-COMPONENT MOBILE BEARING DESIGN Advantages Congruency The k o m p o n e n t designs that incorporate a mobile bearing have intrinsic congruency between the metal and polyethylene. This leads to low mean peak polyethylene stresses. The experience with mobile bearing designs in the knee show that these designs can be associated with very low wear rates of the polyethylene. Low wear rates of polyethylene are highly desirable because submicroscopic polyethylene wear debris can cause periprosthetic osteolysis and component loosening and because they lead to less frequent revision surgeries for polyethylene exchange. Multiplane motion The k o m p o n e n t mobile bearing designs permit multiplane motion while maintaining congruency. With the current models, flexion and extension occurs at the polyethylene-to-talar component interface, and internal and external rotation occurs at the tibialcomponent-to-polyethylene interface. Inversion and eversion can be incorporated into the design and currently are somewhat permitted by the Buechel-Pappas Ultra Total Ankle but not the STAR. Minimize bone resection The k o m p o n e n t designs have the potential to minimize bone resection. The tibial component is a flat surface and because there is no need for medial and lateral capture, this component can be relatively thin (approximately 3 mm thick). Polyethylene mobile bearing sizes are as small as 6 mm thick. The amount of resection of the superior surface of the talus can be relatively minimal (23 mm). Therefore, the entire construct can remove approximately 1.2-1.4 cm of superior and inferior bone without removing substantial bone medially and laterally. 64

Easier salvage With less bone resection, at least theoretically, salvage of nonfunctional replacements should be easier. If salvage is performed for component breakage or wear, the remaining bone stock should be sufficient to permit a relatively easy transition to either a larger arthroplasty, such as the 2component design incorporating syndesmosis fusion, or a k o m p o n e n t design with a larger mobile bearing polyethylene insert. Additionally, conversion of the replacement to a fusion, if necessary, should be somewhat easier than for a 2component design involving syndesmosis fusion because the amount of original bone resection is less.

Concerns Mobib bearing dislocation One of the potential disadvantages of any mobile bearing design is the risk that the polyethylene-bearing insert will dislocate. The capture mechanism must be able to withstand loads and wear that can cause the mobile bearing to slide out of its position. With current designs, the greatest concern for dislocation is from an inversion stress such as might occur from an ankle sprain. Although these concerns are real, in the limited series reported to date, there have been no dislocations or spinouts (dislocations of the nonfixed, mobile bearing) with the 2 modern k o m p o n e n t mobile bearing designs in common u ~ e . " : ~ ~ ~ Abnormal ligament strain To avoid the obvious potential complication of mobile bearing spinout, the components must be placed in the ankle in a tightly packed position. After placement of the polyethylene inserts, the bearing should not be able to displace from the capture mechanism. This requires that the periankle ligaments be stretched at the time the bearing is inserted. A major concern with this a p proach is that the periankle ligaments can become so taut that they decrease mobility of the "mobile" bearing construct, leading to elevated ankle implant stresses and decreased range of motion. Rotational axis malalignment The current designs for the talar component are cylindrical, restricting flexion to occur about a single axis of rotation relative to the talus. If the mobile bearings are inserted in a tight-fit fashion, the axis of rotation obligatorily becomes the axis of rotation of the talar component. Variability in placement of the components relative to the normal axis of the ankle means that the replaced ankle will often be forced to rotate about an axis that is substantially different from that of the intact ankle. The axis of rotation of the talar component is, almost invariably, more horizontal and fixed than that of the normal ankle. Thus, the attempt to use a mobile bearing approach to have a relatively less constrained construct paradoxically results in an unnaturally constrained replacement. Because the ligaments are arranged to accommodate normal rotaJ Orthop Sports P l y Ther.\'olr~me 30. h'umbr 2 Febnrary 2000

tion of the ankle, any change in orientation or location of the rotation axis will result in abnormal tension, slack, or both in the ligaments throughout the joints' full range of motion. Insufficient treatment of medial and lateral arthritic problems The k o m p o n e n t designs incorporating a mobile bearing do not completely resurface the medial and lateral ankle joint recesses. One design, the STAR ankle, has a talar component that covers the medial and lateral surfaces of the talus. Contact between the component in the medial and lateral malleolar articular surfaces works as a hemiarthroplasty of these articulations. The other design, the BuechelPappas Ultra Total Ankle, does not resurface the medial and lateral ankle joint recesses, and relies on pain relief solely from decreasing contact between the inferior tibial surface and the superior talar surface.

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Clinical Results The results for the STAR (Figure 2) are encouraging. The intermediate-term results of the cementless STAR prosthesis used for treatment of ankle osteoarthrosis have been reported for 31 ankles, followed for an average of 3.5 years (range, 1-8 years)." In this series, there was 1 revision performed for malalignment. All patients were reported to be clinically improved as a result of the operation, and none had evidence of component loosening or subsidence. The Buechel-Pappas Ultra Total Ankle Replacement is an improved approach to an earlier mobile bearing prosthesis (New Jersey LCS Total Ankle, Endotec; Figure 3). The LCS design was found to occasionally result in mechanical bearing subluxation, fracture of the tibial plate, fracture of the meniscal bearing, and an irreducible dislocation. The original design was modified in the current implant (Buechel-Pappas Ultra Total Ankle Replacement) to include a deeper talar sulcus for containing the bearing component and an additional talar fixation fin and porous coating covered by titanium nitride thinfilm ceramic. Intermediate-term results of the Buechel-Pappas cementless total ankle arthroplasty have been reported.24In this series, 30 of 38 ankles were followed for an average of 4.5 years (range, 2-8 years). Mean age of patients at surgery was 49 years (range, 25-81 years). There were 12 complications including 6 wound healing problems, 2 cases of s u b talar degenerative joint disease secondary to osteonecrosis and component collapse, 2 revisions (1 talar component, 1 tibial component), 1 arthrodesis for pain, and 1 malleolar nonunion. At follow-up, 5 patients had no pain, 11 patients reported slight pain, and 8 patients reported moderate pain. Over all, the results were rated as excellent in 13 patients, good in 6, fair in 6, and poor in 5. All 3 patients with osteonecrosis had a poor result. J Orthop Sports Phys Ther.Volume SOoN~unber29Febnrary 2000

REHABILITATIONAFTER TOTAL ANKLE ARTHROPLASTY Optimal postoperative care requires that the incision heals without dehicence, that the implant becomes solidly fixed to the bony bed, and that the patient is able to regain adequate motion of the ankle. In the early postoperative period, the initial goal is to heal the incision. Because the soft tissues surrounding the ankle joint are thin and often tenuous, great care is exercised perioperatively to avoid the formation of a wound hematoma. The wounds are sometimes drained by sterile, closed-suction techniques, and intravenous antibiotics are used to cover the surgical site until after the drain is removed. A well-padded short leg splint is used until the skin sutures are removed around 10-14 days after surgery. Below knee non-weight-bearing immobilization is maintained until satisfactory bony ingrowth occurs. With hydroxyapatite coated metal component., ingrowth is usually satisfactory 3 weeks after surgery, and with metal bead coated components ingrowth is generally satisfactory at 6 weeks. Some physicians allow the patients to initiate non-weight-bearing exercises immediately after surgery, although the authors typically wait until satisfactory bony ingrowth is a p parent radiographically. Whether early postoperative motion has beneficial (greater final range of motion) o r detrimental (impairs implant fixation) effects remains a matter of controversy and will require further controlled trials. Weight bearing is initiated after ingrowth of bone into the prosthetic components has been confirmed radiographically. We instruct patients to use a simple ankle ligament stabilizing brace for the first 6 weeks of weight bearing. In patients who are elderly, we will sometimes suggest the temporary use of a cane when walking to help regain balance control and confidence. After 3 months, the patients should have a p proximately 80% of the range of motion noted by the surgeon in the operating room, and by 6 months, all of the motion should be regained. The amount of motion permissible depends on many host and intraoperative factors, including the type of implant used, component positioning, tensioning of the periankle ligaments, and whether other surgeries were simultaneously performed. In general, the goal is to obtain 10" of dorsiflexion and 30" of plantarflexion. More motion is rarely achieved with current implant designs. Many patients present with almost no motion in their ankle resulting from joint space collapse and osteophytic abutment. For these individuals, any motion is an improvement, and even 20" of painless motion in a functional range is considered to be a clinically significant improvement. If patients fail to regain their intraoperatively determined range of motion in a satisfactory manner or have difficulties with balance or strength,

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

physical therapy is initiated. Currently, in our practice, approximately 20% of patients require physical therapy to make satisfactory postoperative progress. On rare occasions, a manipulation under anaesthesia or percutaneous Achilles tendon lengthening is necessary to restore functional motion.

talar component should be as wide as possible at the bone interface to reduce contact stress. Additionally, medial and lateral superior edges of the talar component should be smooth to reduce the potential for line-loading of the polyethylene and resultant accelerated wear.

FUTURE DIRECTIONS

The 3-Component Mobile Bearing Designs

In the past 2 decades, considerable progress has been made in the design of total ankle replacements. Currently, we can offer patients with painful ankle osteoarthritis a total ankle replacement with reasonable assurance that they should have good function during the first 5 years after implantation. We d o not have long-term results on the current designs, and ankle arthroplasty is still plagued with relatively high perioperative complication rates. There are several areas where further work clearly needs to be done.

The challenge for k o m p o n e n t mobile bearing designs is to find a way to permit a relatively normal range of motion without mobile bearing spin-out. Current designs are based on a cylindrical concept and tend to highly constrain the axis of rotation. Future designs should be directed toward reducing ligament strain, restoring the normal axis of rotation, and maintaining mobile bearing stability.

Improved Techniques for Implantation The quality of the instrumentation available for insertion of the components still can be improved. In particular, the alignment systems need to be improved so that the cuts made for implantation are able to be performed in a consistent and reliable fashion and without inadvertent injury to local tissue. Use of intraoperative fluoroscopy to image the entire leg, align the cutting jigs, and guide surgery, probably should be incorporated more often. External fixation of the ankle at the time of surgery may also help with accomplishing this task in a controlled manner.

Prevention of Wound Problems Meticulous handling of soft tissues cannot be overemphasized in total ankle replacement. Many of these patients are immunocompromised from medications used to reduce inflammation. Use of self-retaining retractors should be avoided. Closure of all tissues must be performed with great care. A suction drain should always be used to avoid development of an expanding hematoma under the thin soft tissue envelope. Wound problems are often disastrous in this area, and a high degree of caution must be emphasized.

Redesign of the 2-Component Prosthesis Incorporating a Syndesmosis Fusion The current 2component design, which incorporates a syndesmosis fusion, requires relatively large resection of bone. Finite element modeling needs to be done to optimize the size of the implant and reduce its bulk as much as possible. Furthermore, the

CONCLUSION In the past 2 decades, many advances have been made in total ankle arthroplasty. We can now offer selected patients a total ankle replacement as an alternative option to arthrodesis in the treatment of end-stage ankle arthritis. Further development of prosthetic designs, which minimize bony resection and restore normal ankle motion, is needed.

REFERENCES Ahlberg A, Henricson AS. Late results of ankle fusion. Acta Orthop Scand. 1981;52:103-105. Beaudoin AJ, Fiore SM, Krause WR, Adelaar RS. Effect of isolated talocalcaneal fusion on contact in the ankle and talonavicular joints. Foot Ankle. 1991;12:19-25. Bolton-Maggs BG, Sudlow RA, Freeman MAR. Total ankle arthroplasty. A long-term review of the London Hospital experience. J Bone Joint Surg. 1985;67B:785-790. Boobbyer GN. The long-term results of ankle arthrodesis. Acta Orthop Scand. 1981;52:107-110. Brown TD, Shaw DT. In vitro contact stress distribution in the natural human hip. J Biomech. 1983;16:373-384. Buechel FF, Pappas M, lorio LJ. NJlow contact stress total ankle replacement: biomechanical rationale and review of 23 cementless cases. Foot Ankle. 1988;8:270-290. Carlsson AS, Henricson A, Linder L, Nilsson JA, RedlundJohnell I. A survival analysis of 52 Bath-Wessex ankle replacements. Foot. 1 994;4:34-40. Charnley J. Acrylic Cement in Orthopaedic Surgery. Edinburgh: Livingstone; 1970. Charnley J. Low-Friction Arthroplasty of the Hip. New York: Springer-Verlag Inc; 1978. Chubinskaya S, Huch K, Mikecz K, et at. Chondrocyte matrix metalloproteinase 8: upregulation of neutrophil collagenase by interleukin-l beta in human cartilage from knee and ankle joints. Lab Invest. 1996;74:232-240. Dini AA, Bassett FH. Evaluation of the early result of Smith total ankle replacement. Clin Orthop. 1980;146: 228-230. Click JM, Morgan DD, Myerson MS, Sampson TG, Mann JA.Ankle arthrodesis using an arthroscopic method: longterm follow-up of 34 cases. Arthroscopy. 1996;12:428434. J Orthop Sports Phys Ther .Volwne SO. Number 2 Febn~arv2000

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on January 21, 2017. For personal use only. No other uses without permission. Copyright © 2000 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

13. Goldie IF, Herberts P. Prosthetic replacement of the ankle joint. Reconstr Surg Trauma. 1981;18:205-2 10. 14. Helm R, Stevens J. Long-term results of total ankle replacement. ] Arthroplasty. 1986;1:271-277. 15. Ihn JC, Kim SJ, Park IH. In vitro study of contact area and pressure distribution in the human knee after partial and total meniscectomy. Int Orthop. 1993;17:214-218. 16. Jensen NC, Kroner K. Total ankle joint replacement: a clinical follow-up. Orthopedics. 1992;15:236-240. 17. Kempson GE. Age-related changes in tensile properties of human articular cartilage: a comparative study between the femoral head of the hip joint and the talus of the ankle joint. Biochem Biophys Acta. 1991;1075:223-230. 18. Kimizuka M, Kurowawa H, Fukubayashi T. Load-bearing pattern of the ankle joint: C-contact area and pressure distribution. Arch Orthop Trauma Surg. 1980;96:45-49. 19. Kirkup J. Richard Smith ankle arthroplasty. ] Roy Soc Med. 1985;78:301-304. 20. Kitaoka HB, Patzer GL. Clinical results of the Mayo total ankle arthroplasty. 1 Bone loint Surg. 1996;78A:16581664. 21. Kofoed H, Danborg L. Biological fixation of ankle arthroplasty. Foot. 1995;5:27-31. 22. Kofoed H, Toben S. Ankle arthroplasty for rheumatoid arthritis and osteoarthritis: prospective long-term study of cemented replacements. IBone loint Surg. 1998;80B: 328-332. 23. Kumar DJR.Total ankle arthroplasty. A review of 37 cases. 1Tenn M e d Assoc. 1988;81:682-685. 24. Lin S, Drzala M. Independent evaluation of Buechel-Pappas 2nd generation cementless total ankle arthroplasty

J Orthop Sports Phys Ther.Volume SO. Number 2. Febniary 2000

25. 26. 27. 28. 29. 30. 31.

32. 33.

34.

(TAA): intermediate term results [abstract]. Available at: http:Nwww.aofas.org/98winter/9803220953.html. Accessed October 24, 1999. Lynch AF, Bourne RB, Rorabeck CH. The long-term results of ankle arthrodesis. ] Bone loint Surg. 1988;70B:113116. Mazur JM, Schwartz E, Simon SR. Ankle arthrodesis; longterm follow-up with gait analysis. / Boneloint Surg. 1979; 61A:964-975. Morgan CD, Henke JA, Bailey RW, Kaufer H. Long-term results of tibiotalar arthrodesis. IBone loint Surg. 1985; 7A:54&550. Morrey BF, Wiedeman GP. Complications in long-term results of ankle arthrodeses following trauma. ] Boneloint S~rg.1980;62A:777-784. Pyevich MT, Saltzman CL, Callaghan JJ, Alvine FG. Total ankle arthroplasty: two to twelve year follow-up of a unique design. ] Bone loint Surg. l998;8OA:l4lO-l42O. Said E, Hunka L, Siller TN. Where ankle fusion stands today. ] Bone loint Surg. 1978;60B:2 11-21 4. Shulte KR, Callaghan JJ, Johnston RC, Kelley SS. The outcome of Charnley total hip arthroplasty with cement after minimum 20-year follow-up. ] Boneloint Surg. 1993;75A: 961-975. Takakura Y, Tanaka Y, Akiyama S, Tamai S. Results of total ankle arthroplasty. ] Foot Surg. 1996;11:9-16. Takakura Y, Tanaka Y, Sugimoto K, Tamai S, Masuhara K. Ankle arthroplasty. A comparative study of cemented metal and uncemented ceramic prostheses. Clin Orthop. 1990;252:209-216. Waugh TR, Evanski PM, McMaster WC. lrvine ankle arthroplasty. Clin Orthop. 1975;114:1880-1884.