Secondary Logo

Journal Logo

SECTION I SYMPOSIUM: Complications in Foot and Ankle Surgeries

Complications of Total Ankle Replacement

Conti, Stephen F. MD; Wong, Yue Shuen MD

Author Information
Clinical Orthopaedics and Related Research: October 2001 - Volume 391 - Issue - p 105-114
  • Free


Total ankle arthroplasty is emerging as a viable treatment for patients with symptomatic tibiotalar arthritis who have not responded to nonoperative treatment. Although enthusiasm for first generation ankle replacement prostheses was high in the 1970s and the early part of the 1980s, they subsequently were considered to have significant complications, leading orthopaedic surgeons to abandon their use. 1,3,4 The most common complications reported with these prostheses were wound healing problems, deep infection, and aseptic loosening. Salvage of these complications was difficult, with extensive fusions and amputation occurring frequently.

Ankle Arthritis

To understand the particular problems with ankle replacement arthroplasty it is necessary to first understand the basic problem, namely ankle arthritis. It is human nature to correlate new information with existing familiar patterns based on study, experience, and training. Ankle arthritis, however, is different from other types of degenerative joint disease that the orthopaedic surgeon routinely encounters in clinical practice. For example, hip and knee osteoarthritis is predominantly of degenerative etiology and is seen in older patients. Each joint essentially is a two-bone system with significant ligament balancing issues occurring in a small percentage of patients. The soft tissue surrounding the joints generally is not traumatized and the tissue envelope is more substantial. Regaining range of motion (ROM) within the arc needed to walk rarely is a problem.

Contrast hip and knee arthritis to the typical case of ankle arthritis. It is estimated that 80% of ankle arthritis is posttraumatic 1 in origin, and therefore occurs in patients younger than those with hip or knee osteoarthritis. Although there essentially is one bone above the ankle, there are 26 bones and as many joints below the ankle that can affect the alignment and functioning of the prosthesis. The normal soft tissue envelope around the ankle is thin, and because of the antecedent trauma and initial surgical repairs, the soft tissue envelope often is scarred and inelastic. These same reasons, combined with the initial period of immobilization, lack of adequate physical therapy, chronic pain, and progressive periarticular osteophyte formation, often lead to significant loss of ROM of the ankle, which may not improve with replacement of the tibiotalar joint.

Complications of First Generation Prostheses

Understanding the problems associated with first generation prostheses also is essential to understanding the rationale behind the current generation of implants and the problems being encountered with them. The complications associated with first generation prostheses may be broadly grouped into preoperative, prosthetic design, intraoperative and postoperative complications.

Preoperative complications predominantly were related to patient selection. The development of ankle prostheses paralleled hip and knee systems that were used successfully in patients with rheumatoid arthritis and osteoarthritis. However, ankle replacements in patients with rheumatoid arthritis have many potential problems, including the risk of wound healing, subsidence of the prosthesis because of poor bone quality, and late aseptic loosening. This last complication was associated with the fact that many patients with rheumatoid arthritis have pronated feet attributable to chronic subtalar synovitis and ligament incompetence. Many patients either had their hindfoot deformities uncorrected at the time of ankle replacement or had previous in situ hindfoot fusions leaving them with a calcaneovalgus deformity. This may have led to deleterious effects on the loading of the prosthesis with subsequent higher rates of failure.

Prosthetic design features certainly were an issue. The first problem was that most first generation prostheses were cemented designs, which has been associated with poorer outcome. 5 Cemented implants may have required more bone resection, leading to increased loading of the weaker metaphyseal bone. In addition, with respect to the differences between the ankle and other joints, the limited confines of the tibiotalar joint make doing proper cementing difficult, even with an understanding of modern cementing technique. Finally, less was understood about cement pressurization in the 1970s, and those cementing techniques now are considered inadequate by today’s standards.

The second problem was in understanding the biomechanics of the ankle. Constrained designs, although more stable, were associated with high rates of aseptic loosening, possibly because of increased forces being transmitted across the bone-prosthesis interface. In contrast, nonconstrained designs were associated with instability and malleolar impingement. 6

Intraoperative complications included fractures and malpositioning of components. Instrumentation for ankle arthroplasty was fairly crude, requiring a fair amount of free-hand technique. Proper alignment of the prosthesis must take into account bony anatomy and ligament balancing issues. Bone alignment must take into consideration not only the bow of the tibia in various planes, but also the position of the hindfoot. In addition, the degree of bone loss caused by the initial trauma or progressive erosion must be taken into account. Ligament balancing issues are critically important in the ankle and are very complex. Even today, one of the least understood aspects of ankle replacement surgery is ligamentous balancing.

Postoperative complications included wound healing, deep infection, subsidence, and aseptic loosening. 3 Wound healing clearly was related to patient selection. The injured soft tissue envelope after trauma, or the thin skin associated with rheumatoid arthritis made wound healing problems more likely than in the hip and knee. Subsidence and aseptic loosening already have been discussed.

Finally, perhaps the most significant problem with the first generation prostheses were the reconstructive options. Because many prostheses required significant bone resections and were cemented in place, it became difficult to revise the components when they failed.

Complications of Second Generation Prostheses

With the beginning of the current article providing some background, it now is possible to discuss the complications of second generation total ankle arthroplasties more extensively. Two basic types of designs currently exist: two-component designs and threecomponent designs. Although there are definite design feature differences between these two systems, they share certain characteristics that lead to common complications. In other cases, each type of design has unique qualities that lead to unique problems.

There are three ankle replacement systems currently in use in the United States. The Agility Ankle System by DePuy (Warsaw, IN) is a two-component system consisting of a Ti tibial component with an attached polyethylene articular surface and a CoCr talar component. The STAR ankle replacement by Link and the Buechel-Pappas LCS ankle replacement are three-component systems, which use a mobile meniscal-bearing component sandwiched between the tibial and talar components. The Agility Ankle System was developed by Frank Alvine, MD and first implanted in 1984. To date, there have been more than 2000 prostheses implanted and more than 200 surgeons trained to use the prosthesis. The STAR ankle system is undergoing clinical trials and is not yet approved by the Food and Drug Administration, and the Buechel-Pappas ankle system has limited distribution. At the author’s institution, more than 100 Agility Ankle Systems have been implanted. For these reasons most of the remainder of the current study will be devoted to the extensive experience with the Agility Ankle System at the authors’ institution.

As mentioned previously, the Agility Ankle System is a two-component system, consisting of a semiconstrained hinge joint angled in 20° of external rotation. It is unique among the current generation of ankle prostheses in that the distal tibiofibular syndesmosis is fused routinely, providing a larger surface for implant fixation and force transmission. The implant also is inserted under distraction, which achieves two goals. One goal is to remove a minimal amount of bone. The other goal is to tension and balance the periarticular soft tissues.

A review of the first 100 patients who underwent total ankle arthroplasty were reviewed independently of the surgeon. 8 Of these patients, 45 had posttraumatic arthritis, 26 had primary osteoarthritis, 26 had rheumatoid arthritis, two had septic arthritis, and one had psoriatic arthritis. The mean age of the patients was 63 years (range, 28–81 years) and average followup was 4.8 years (range 2.8–12 years). Fourteen patients died, leaving 86 patients in the study group. Three patients required talar revision and one patient required tibial revision. One patient had a revision to arthrodesis. There were two superficial wound infections and no deep infections. Fifty percent of the patients had a plantar flexion contracture requiring a heel lift. Twenty-nine percent had a delayed union of the syndesmosis, and 9% had a nonunion. Twelve tibial components and seven talar components showed radiographic evidence of migration. Of the tibial component migrations, eight were associated with delayed or nonunion of the synostosis.

The availability of published results of second generation ankle replacement systems are limited by the relatively short history of their use and the cautious introduction to large scale use. Although a fair amount of experience has been gained in the use and complications of these devices, the understanding of them is necessarily anecdotal. Nevertheless, the information given below is a true reflection of problems related to the current generation of implants and possible solutions.

The complications of current second generation total ankle arthroplasty will be divided in a similar fashion: preoperative or patient selection problems, complications related to prosthetic design, intraoperative complications, and postoperative complications. Solutions, or the controversies surrounding those complications that have no obvious solution, will be discussed when appropriate.

As with first generation arthroplasty, patient selection is critical to limiting complications and obtaining a satisfactory result. Older patients with nontraumatic primary osteoarthritis, minimal angular deformity at the ankle and excellent ROM before surgery would seem to be the best candidates for ankle arthroplasty. Age is a somewhat contentious issue in ankle replacement surgery. Clearly there is no correct answer as to whether younger patients should be offered ankle replacement as a treatment of their arthritis. Older patients, because of their advanced years and lower impact activity level, are intuitively less likely to require revision surgery because there is more likelihood that the implant will outlive them. However, most symptomatic ankle arthritis occurs in younger patients. The two possible surgical options in this situation are replacement and arthrodesis. In the case of ankle replacement, patients can hope to retain much of their preoperative tibiopedal motion, thereby reducing stress at the other foot joints and slowing down the progression of arthritis at these joints. If arthritis eventually does occur at these joints, subsequent fusion is possible, leaving the patient with a fused midfoot or hindfoot and a mobile ankle. The disadvantage with ankle replacement is that eventual failure of the prosthesis inevitably will result, and there are few data on the effectiveness of revision surgery at this time. One possible scenario would be removal of the prosthetic components, inability to do an isolated tibiotalar arthrodesis, and the necessity of doing a tibiotalcalcaneal fusion with significant shortening of the involved extremity. Although ankle arthrodesis has been the gold standard for symptomatic arthritis in this age group, this procedure is not without its own problems. There are several studies outlining the long-term problems associated with ankle arthrodesis, not the least of which is the development of arthritis in the hindfoot and midfoot joints. 7,9–12 If a young patient undergoes ankle arthrodesis, there is a significant likelihood that he or she will have hindfoot arthritis develop during the next 20 years, necessitating additional fusion surgery. An isolated ankle fusion may progress into a pantalar fusion with its increased limitations and morbidity. There also have been instances where the extent of the arthritis is such that below knee amputation has been recommended as a surgical option (oral communication, S. Hansen, MD, 2000).

The condition of the soft tissue envelope is another important preoperative consideration that may influence complications. The more damage there is to the soft tissues around the ankle, the more potential there is for wound healing difficulties, chronic swelling, and loss of ROM. Multiple scars from previous surgeries make wound healing difficulty a distinct possibility. In addition, the normal tissue planes are damaged, making it harder to identify underlying neurovascular structures with the increased likelihood that these structures will be damaged. Chronic soft tissue damage also is associated with chronic soft tissue discomfort. This discomfort may not diminish with ankle replacement surgery. Patients with significant soft tissue envelope disease preoperatively should be warned that they still might experience some chronic discomfort and swelling postoperatively.

Preoperative alignment also influences the complication rate. Large tibial bows or tibial malunions, fibular malunions, and hindfoot pronation or supination are major contributors to technical difficulty during surgery and less than successful outcomes. Clinical assessment is paramount to understanding the mechanical axis of the lower extremity. In some cases, full-length tibial and ankle radiographs can be useful. If the prosthesis is not placed parallel to the ground and loaded symmetrically, there is a greater possibility of polyethylene wear, aseptic loosening, and subsidence. The initial examination should include a careful assessment for knee pain and determination of knee alignment. Patients who have concomitant degenerative joint disease of the knee with asymmetric medial or lateral compartment narrowing should be considered for total knee replacement before proceeding with total ankle replacement. If the need for knee replacement is greater than 3 years away, then proceeding with ankle replacement surgery seems reasonable. Next in the evaluation process should be an assessment of the mortise line. If it is not parallel to the ground, there may be supramalleolar malalignment or tibial bone loss from erosion. Supramalleolar malalignment should be corrected through a tibial osteotomy before ankle replacement surgery if greater than 10° in any plane. Distal tibial bone loss usually can be accommodated through proper placement of the cutting jig. Because the Agility prosthesis relies on cortical rim support, metaphyseal bone erosions medially or laterally have little significance. This may not be the case with the STAR or Buechel-Pappas prostheses.

Unrecognized or uncorrected foot deformity below the ankle is a major reason for poor results after surgery. A plantigrade foot with weightbearing stability provided by the tripod position of the first and fifth metatarsal heads and the heel is an essential element to support the ankle prosthesis. In cases of symmetric ankle collapse, standard examination techniques will determine whether a plantigrade foot is present. If not, the surgeon must make the decision to correct the deformity before or during prosthetic implantation. 13 Operations that leave the foot supple are preferable to those that cause stiffness to obtain the plantigrade position. An example of the former is a first metatarsal medial cuneiform fusion combined with a medial slide calcaneal osteotomy for a pronated foot. The latter would be best exemplified by a triple arthrodesis. The difficulty comes in patients with asymmetric ankle collapse that is fixed. In these cases, the surgeon cannot determine the proper position that constitutes a plantigrade foot because the ankle deformity is fixed and it is unknown how much of the deformity will be corrected at the time of surgery. Often the surgical correction involves some combination of bone cut compromise to account for distal medial tibial metaphyseal bone erosion and deltoid ligament contracture. In these cases, it is preferable to do the ankle replacement surgery first followed by surgical attainment of a plantigrade foot through additional surgery at the time of implantation or several months later.

Finally, failure to adequately assess gastrocnemius-soleus tightness preoperatively can lead to significant complications. This is best assessed using the Silverskjold test. The patient is seated in front of the examiner and, with the knee extended, maximum ankle dorsiflexion is obtained by a combination of active and passive motion. While the patient continues to dorsiflex the ankle, the knee is bent and the position of the lateral border of the foot to the anterior tibial crest is assessed again. If the patient cannot dorsiflex to greater than 5° with the knee straight then a heel cord lengthening is done at the time of ankle replacement surgery. When knee flexion substantially increases dorsiflexion a selective gastrocnemius lengthening is done at the musculotendinous junction of the gastrocnemius-soleus muscle. If knee flexion does not alter ankle position, then a distal Achilles tendon lengthening would be necessary. Selective gastrocnemius lengthening is done before inflation of the tourniquet during the replacement surgery. Distal Achilles tendon lengthenings are done percutaneously just after implantation. Care is taken not to overlengthen the tendon when distal procedures are done, because significant plantar flexion weakness will result. Problems arise when significant periarticular osteophytes are present that make preoperative assessment unreliable. In these cases, the choice of Achilles tendon procedure is determined after prosthetic implantation.

Intraoperative complications seem to diminish with experience and the ankle surgeon who is beginning to do replacements may notice a steep learning curve associated with this procedure. Incisional complications include releasing the tibialis anterior tendon from its sheath and scarring or transection of the superficial or deep peroneal nerves. The anterior ankle incision is made just lateral to the tibialis anterior tendon. The superficial peroneal nerve always is beneath the incision and should be swept laterally. Most descriptions of an anterior approach usually recommend the deeper dissection in the interval between the tibialis anterior and extensor hallucis longus tendons. When the authors have done this, the tibialis anterior tendon often bowstrings by coming out of its sheath. Therefore, it is preferable to place the incision more laterally and purposely take the extensor hallucis longus tendon out of its sheath, trying to preserve the restraints of the tibialis anterior tendon. Immediately below this lies the deep peroneal nerve, which must be identified and retracted during incision of the capsule. It also must be identified again at the end of the case to avoid suturing it while closing the capsule.

Malpositioning of the prosthetic components probably is the most common intraoperative complication. This can occur in multiple planes and each would need to be discussed separately. The following comments will be confined to the Agility Ankle System except where indicated. 14

Ideally the ankle joint axis, approximated by the tips of the medial and lateral malleoli, would be reestablished after implantation. Proximal displacement of the cutting jig will displace the joint line proximally resulting in several problems. The first problem is shortening of the gastrocnemius-soleus complex with resulting plantar flexion weakness. More commonly, proximal displacement leads to medial gutter impingement. This is because the position of the medial malleolus remains unchanged while the talus migrates proximally, and the two impinge with weightbearing. Proximal displacement also may cause a significant problem with the STAR prosthesis where proximal implantation results in seating of the prosthesis on softer metaphyseal bone with greater risk of subsidence. Distal displacement of the cutting jig results in excessive bone resection of the talus with potential for subsidence or subtalar joint compromise. The joint line is moved distally, placing greater tension on the tendoAchilles resulting in loss of dorsiflexion motion and altered gait. Because inability to fully dorsiflex is a common postoperative result and causes an abnormal gait, the authors consider this a significant problem.

Another malalignment problem is varus or valgus malpositioning of the prosthesis. Extramedullary alignment rods are available but are only rough guides to proper final component positioning. Varus of the prosthesis is tolerated poorly just as varus of the foot is not well tolerated. The two most common clinical problems associated with this are medial gutter pain from impingement and lateral foot pain from excessive weightbearing on the fifth metatarsal base. Valgus positioning of the prosthesis is tolerated better than varus positioning of the prosthesis. Valgus malpositioning can cause subfibular impingement and pronation deformity of the foot with resulting arch discomfort. Either varus or valgus can result in abnormal stress on the polyethylene with the theoretical long-term disadvantage of polyethylene wear. Varus or valgus positioning of the talar component beneath the tibial component will result in many of the same problems but in addition, cause edge loading of the prosthesis on the polyethylene resulting in early wear.

One of the most complex intraoperative complications is medial or lateral malpositioning of the components. It is complicated because of the interplay between bone and soft tissue tensioning. If the tibial component is positioned too medially, it will result in a thin medial shoulder that may lead to malleolar fracture. Too lateral displacement of the tibial component often results in lateralization of the talus because the talar component only can go as medial as the shoulder of the tibial component will allow. This often puts excessive tension on the deltoid ligament resulting in varus of the talus. This may occur whether the preoperative joint space narrowing was symmetric or asymmetric. Attempts at subsequent deltoid ligament release then can cause incompetence of the ligament and severe valgus deformity.

Improper sizing of the components intraoperatively can result in complications similar to malpositioning attributable to displacement. In addition, proper sizing is important to prevent impingement, syndesmosis nonunion, and late migration of the tibial component. A smaller prosthesis seems to result in a higher rate of malleolar impingement symptoms but may make it easier to obtain a syndesmotic fusion. A larger prosthesis is just the opposite. Also, as the tibial baseplate gets bigger more of the prosthesis overhangs the lateral cortex of the tibia, because its medialization is limited by the amount of medial malleolus that can be resected. This results in a significant portion of the prosthesis underlying the syndesmosis. If the syndesmosis unites as it should this probably is not an issue, but when the syndesmosis fails to fuse the lateral edge of the prosthesis can subside into the nonunion site. This is more likely to occur when a greater proportion of the prosthesis lies beyond the lateral cortex of the distal tibia.

Distraction of the ankle is another area of potential complications. Underdistraction results in excessive bone resection. Overdistraction will allow for minimum bone resection but with resultant overstuffing of the joint causing pain and loss of motion.

Many of these intraoperative positioning problems can be lessened with the use of fluoroscopy at the time the cutting jig is placed. Anteroposterior and lateral radiographs can be helpful in obtaining a more accurate positioning of the components. The authors routinely use fluoroscopy on all patients undergoing total ankle arthroplasty to assess proper distraction, correct position of the cutting jig, and to ensure that the prosthesis is seated on the posterior cortex of the tibia and talus. Because fluoroscopy often is used to check the external fixator pin lengths and the positioning of the syndesmotic screws, the image intensifier is readily available during the procedure and its use adds little to the total operating time.

Other complications include malleolar fractures and tendon injuries. Although most malleolar fractures occur from excessive bone resection, it also is possible to cause avulsion fractures from overdistraction with the external fixator, or during implantation of the components. Excessive bone resection was discussed previously. If the components are implanted with the external fixator in place then as they are driven in, the tibia and talus are distracted vertically with tension on the deltoid and medial malleolus. To prevent this, the authors often remove the external fixator just before implantation. Tendon injuries can occur to the posterior tibial, peroneal or flexor hallucis longus tendons. Care must be taken when working behind either malleoli or in the posterior aspect of the ankle to prevent these injuries.

Postoperative complications can occur early or late. Early complications include wound healing problems, syndesmotic nonunions, swelling, infection, and deep venous thrombosis. Wound healing problems seem to occur for two main reasons. The first reason is too small of an incision is made with excessive retraction subsequently leading to wound dehiscence. The second reason is a pressure necrosis phenomenon over the anterior ankle skin. Making a longer incision is mandatory to gain surgical exposure and prevent excessive skin traction. Pressure necrosis occurs when the surgical dressing is put on at the end of surgery with the ankle in plantar flexion. When the ankle is dorsiflexed before the splint is applied, it compresses the skin on the anterior ankle and can lead to pressure necrosis. Wound dehiscence is recognized by failure of the tissue edges to heal. In contrast, pressure necrosis is confined to a large area just lateral to the skin incision.

Surgeons should take preventative actions if they anticipate that wound-healing complications may occur. In addition, modifications in technique may lessen the impact of such complications. Specific consideration should be given to capsular closure at the end of the procedure. The capsule often is thickened and scarred from trauma and chronic degenerative joint disease. Excising the capsule during the surgical approach has the advantages of allowing greater exposure and decreased pain postoperatively. However, should wound dehiscence occur, direct exposure of the implant is likely, necessitating immediate plastic surgery techniques to secure coverage, or removal of the implants, coverage and then subsequent reimplantation. Therefore, retaining the capsule and closing it at the end of the procedure seems prudent.

Syndesmotic nonunions are another early complication. Arthrodesis of the distal tibiofibular joint is dependent on several factors including width of the syndesmosis, complete decortication of the opposing surfaces, and rigid fixation. Prosthetic size considerations were discussed previously. Complete decortication is difficult because complete visualization of the joint is difficult, whether through the anterior or lateral incision. The authors have used a technique of rotating cortical flaps on adjacent sides of the syndesmosis posteriorly to thoroughly decorticate the joint and allow vascularized segments of the tibia and fibula to contact one another. The resected piece of distal tibia removed for prosthesis implantation then is used to fashion one block of cancellous bone, which is positioned in the syndesmosis and fixation is applied. Two, 3.5-mm cortical lag screws are placed across the syndesmosis with the most distal screw placed at the superior edge of the distal tibiofibular joint. Proximal placement of the screw will result in toggling of the distal fibula away from the prosthetic and increase the risk of fibular fracture.

Swelling always has been a persistent problem with multiple surgeries about the ankle, especially after trauma. With total ankle arthroplasty the problem is exaggerated for numerous reasons. First, there often is significant soft tissue damage accompanying the initial ankle fracture in patients with posttraumatic arthritis. There may be multiple subsequent surgeries obliterating tissue planes and additionally injuring the lymphatic system in the area. Long-standing arthritis then ensues with associated chronic edema and the alterations that occur in the periarticular tissues. Finally, during ankle replacement surgery there is additional soft tissue damage, creation of dead space, and increased movement about a preoperatively stiff joint. The result is swelling that may persist for as many as 1.5 years. Preoperative counseling for patients with chronic soft tissue disease will lessen their anxiety. The authors insist on 2 to 4 weeks of strict elevation above the level of the heart postoperatively. At 6 weeks after surgery, any patient with significant swelling is assessed for lymphedema treatment, consisting of pneumatic pumps, wraps and/or compression garments depending on the degree of involvement.

Infection and deep venous thrombosis are uncommon complications after total ankle arthroplasty. Adherence to good surgical technique, appropriate perioperative antibiotic use, and standard deep vein thrombosis prophylaxis will lessen these complications.

Late postoperative complications include malunion, syndesmotic nonunion, loss of dorsiflexion motion, deep infection, subsidence, and polyethylene wear. The first two complications already have been discussed. Loss of dorsiflexion is a serious functional complication. One study showed that 20° to 25° total tibiopedal motion is necessary to walk without a limp on flat, level surfaces. 15 Implicit in this statement is the assumption that this ROM is within physiologic limits, 5° dorsiflexion to 15° to 20° plantar flexion. When patients cannot dorsiflex to 5°, they will have an abnormal gait when walking barefoot. The authors’ experience with preoperative and postoperative ROM suggest that patients obtain approximately the same motion after surgery as they had before surgery. The difference has been that the ROM has been moved into a more physiologic range.

Postoperative plantar flexion contracture can occur for several reasons. The first possibility is an existing preoperative contracture. Because of soft tissue injury and prolonged use of a cast after the original injury, many patients have lost ROM. This is best released by stripping the posterior capsule off the distal tibia just before implantation of the components. Achilles tendon lengthening also is an integral part of total ankle arthroplasty. Patients with a positive Silverskjold test will have a selective gastrocnemius lengthening done at the musculotendinous junction through a posterior midline incision. Other patients will have a percutaneous distal Achilles tendon lengthening using a three-incision technique. Great care must be exercised when considering whether to lengthen the Achilles. Leaving a patient with a plantar flexion contracture after total ankle arthroplasty causes abnormal gait and possibly can lead to increased midfoot stress and arthrosis with time. Distal Achilles lengthening will increase dorsiflexion but at a risk of causing push-off weakness during gait. Without proper push-off strength, the terminal plantar flexion that usually occurs will be lost resulting in a gait pattern that resembles the gait of a patient who had ankle fusion. The theoretical advantage of selective gastrocnemius lengthening is that it can increase ROM while not causing significant plantar flexion strength deficits.

Deep infection is a potential complication of any joint implant. Ankle arthroplasty is no exception. The American Academy of Orthopaedic Surgeons has a position statement on the use of prophylactic antibiotics after joint arthroplasty. 2 The authors follow the same recommendations. All patients undergoing dental, genitourinary, or gastrointestinal manipulations who are within 2 years of their arthroplasty surgery are given 2 g amoxicillin 2 hours before the procedure, or 600 mg clindamycin 1 hour before, if allergic to penicillin. After 2 years, only patients at high-risk (patients with diabetes, patients with hemophilia, or patients who are immunosuppressed) are given prophylaxis antibiotics. The authors also consider patients with rheumatoid arthritis who are taking medication for their disease to be in this high-risk category.

Aseptic loosening and failure of porous ingrowth are other long-term complications. To try to increase porous ingrowth and proper seating of the prosthesis, the authors allow patients to bear weight at the bedside, starting on Day 1 at 30 to 40 lb, wearing the plaster splint. Periprosthetic radiolucent lines are defined poorly around ankle prostheses. In the authors’ experience, cases of significant balloon osteolysis around the fibula can occur without significantly affecting the medium-term outcome. The authors have seen cases of periprosthetic lucencies resolve with time.

Subsidence is another late complication whose etiology remains elusive. The Agility Ankle System is a cortical rim support device. Subsidence of the tibial component has not been a problem except in cases of syndesmotic nonunion where a high percentage of the tibial tray lies beyond the lateral margin of the distal tibia. Talar component subsidence is an issue and now is considered the primary mechanism of failure of the total ankle system. The authors think that there may be several considerations under the surgeon’s control. The primary consideration may be patient weight. Although no absolute cut-off exists, the basic principle of patient weight compared with implant size is more significant than weight alone. The author think that patients who weigh more than 230 lb who have a Size 4 talar component or smaller have a significant risk of talar subsidence. The second factor under the surgeon’s control is positioning of the components intraoperatively. The tibial and talar components should be positioned posteriorly enough so that they rest on the posterior cortices of the tibia and talus, respectively. In addition, in all three total ankle replacement systems discussed previously, minimizing the bone resection results in remaining bone stock that is structurally more supportive, and therefore is less likely to result in subsidence. Finally, implant designs that increase the available surface area of contact, theoretically will lessen the risk of subsidence.

Total ankle arthroplasty with this second generation of prostheses is gaining increasing popularity. Complications do and will occur with this group of replacements just as they do with all other arthroplasties whether it is in the knee, hip, elbow, shoulder, or wrist. The current study is a review of the many complications possible with one total ankle arthroplasty. What is clear is that even if the complications are recognized and the preventative measures or solutions are well thought out, there often is no clear correct answer to the problem. Each potential solution will have an effect on another portion of the biologic system, possibly creating another complication. Therefore, the surgeon contemplating total ankle arthroplasty should have an understanding of anatomy and lower extremity biomechanics, combined with a thorough knowledge of the total ankle system they have decided to use. Although there are many obstacles to a successful outcome, most of the complications reported here are preventable with good patient selection and technique.


1. Agility Total Ankle System: Surgical Technique Guide. Warsaw, IN, DePuy Orthopaedics 1999.
2. American Academy of Orthopaedic Surgeons Web site. Available at: Accessed September, 2000.
3. Beuchel FF, Pappas MJ, Iorio LJ: New Jersey low contact stress total ankle replacement: Biomechanical rationale and review of 23 cementless cases. Foot Ankle 8: 279–290, 1988.
4. Bolton-Maggs BG, Sudlow RA, Freeman MAR: Total ankle arthroplasty: A long term view of the London Hospital experience. J Bone Joint Surg 67B: 785–790, 1985.
5. Boobyer GN: The long term results of ankle arthrodesis. Acta Orthop Scand 52: 107–110, 1981.
6. Frey C, Halikus NM, Vu-Rose T, et al: A review of ankle arthrodesis: Predisposing factors to non-union. Foot Ankle Int 15: 581–584, 1994.
7. Fuller AE, Rostrup O, Huckell Jr: Ankle arthrodesis: A clinical review. J Bone Joint Surg 56B: 587, 1974.
8. Gould JS: Total ankle replacement: A surgical discussion. Part I. Replacement systems, indications, and contraindications. Am J Orthop 29: 604–609, 2000.
9. Gould JS: Total ankle replacement: A surgical discussion. Part II. The clinical and surgical experience. Am J Orthop 29: 675–682, 2000.
10. Kitaoka HB, Patzer GL: Clinical results of the Mayo Total Ankle Arthroplasty. J Bone Joint Surg 78A: 1658–1664, 1996.
11. McGuire MR, Kyle RE, Gustilo RB, et al: Comparative analysis of ankle arthroplasty versus ankle arthrodesis. Clin Orthop 226: 174–181, 1988.
12. Neufeld SK, Lee TH: Total ankle arthroplasty: Indications, results, and biomechanical rationale. Am J Orthop 29: 593–602, 2000.
13. Newton E: Total Ankle Arthroplasty. In Jahss M (ed). Disorders of the Foot and Ankle: Medical and Surgical Management. Philadelphia, WB Saunders 816–825, 1982.
14. Pyevich MT, Saltzman CL, Callaghan JJ, et al: Total ankle arthroplasty: A unique design. J Bone Joint Surg 80A: 1410–1420, 1998.
15. Wagner F: Ankle fusion for degenerative arthritis secondary to collagen diseases. Foot Ankle 3: 24–31, 1982.

Section Description

Osaretin B. Idusuyi, MD; and G. James Sammarco, MD, Guest Editors

© 2001 Lippincott Williams & Wilkins, Inc.