SALVAGE AFTER COMPLICATIONS OF TOTAL ANKLE ARTHROPLASTY
January 1st, 2002
Mark S. Myerson, M.D. and Stuart D. Miller, M.D.
SYNOPSIS
Recent advances in total ankle arthroplasty (TAA) have generated a tremendous interest in this procedure and an increase in the number of prostheses inserted. As with many new procedures, there is an inevitable perioperative failure rate. Although the experiences with TAA will indeed improve with time, we and others have identified complications and initial failures. In this article we describe some of the complications we have experienced with TAA, and we describe the surgical management of these problems.
INTRODUCTION
The Agility Ankle Replacement prosthesis (DePuy, Warsaw, IN) is the total ankle prosthesis with which we have had extensive experience and which is currently available and approved by the FDA for use in the United States. In an earlier review of the first 84 total ankle arthroplasties (TAAs) with the Agility ankle prosthesis,9 only 6% of these procedures had been revised. However, a number of tibial and talar components had settled during the follow-up period.
Complications of TAA can be divided into two major categories: those in which the joint replacement can be salvaged and those in which the ankle must be fused. The vast majority of complications are fortunately minor and can be resolved either intraoperatively, (i.e., fracture of the malleoli) or later (i.e., malalignment or syndesmosis nonunion). More severe collapse with avascular necrosis, gross loosening, or bone loss may require an arthrodesis of the ankle for salvage in order to stabilize the hindfoot and ankle. Fortunately this has been a rare occurrence, and in the senior author’s (MSM’s) experience of more than 175 TAAs using the Agility prosthesis, two patients were treated with an arthrodesis. Despite the relative rarity of complete failure, it is important to view each TAA as an alternative to arthrodesis and to be sure patients understand the potential need for a fusion should the TAA fail. The presumed longevity and the potential for eventual failure due to prosthetic wear is an issue with which we do not as yet have any experience, but this and the possibility of catastrophic failure are constant concerns. One report has described a patient who underwent a below-the-knee amputation following deep unresolved infection and failure of arthrodesis.10
While technical errors remain a persistent problem, the incidence of these seems to be decreasing. This is clearly based upon a very difficult initial learning experience with total ankle arthroplasty, the so-called learning curve. This has been addressed by the increasing availability of basic and advanced training and instruction given to users of the Agility (DePuy) TAA. Nonetheless, technical errors continue to be a factor. In a review of the senior author’s first consecutive 50 TAAs performed (Myerson M.S., Mroczek K., submitted for publication), patients were divided into Group A (the first 25) and Group B (the second 25) procedures performed. These authors noted six perioperative fractures in Group A and two in Group B; four lacerations in Group A (posterior tibial tendon, flexor hallucis longus tendon, superficial peroneal nerve, and deep peroneal nerve) and none in Group B; six minor wound complications in Group A and 2 in Group B (there were no deep infections or major wound complications noted in either group); and ideal radiographic position noted in 18 of 25 in Group A and 22 of 25 in Group B patients. With the increased availability of training and education, it is unlikely that these complications will be experienced at this rate by all surgeons. However, complications may occur even with experienced surgeons, although the incidence of complications decreases with increasing experience with the procedure. This paper addresses the issues involved with salvage of the TAA, including patient selection, hindfoot malalignment, fractures, component loosening, infection, and some of the decisions regarding fusion after TAA.
PATIENT SELECTION
Patient selection is critical with respect to their activity level, lifestyle, limb alignment, quality of the periarticular bone, and the presence of adjacent joint arthritis. Factors such as smoking, athletic interests, medications, and systemic disease may also have an impact on surgical planning. Perhaps more important is to exclude those patients who present too high a risk, including those with peripheral vascular disease, diabetes, severe peripheral neuropathy, and skin conditions that preclude the appropriate surgical approach. This is particularly relevant after trauma where the skin may not permit exposure.
Smokers have a much higher incidence of difficulty with ankle fusion,4 and similar care should be taken with smokers in ankle arthroplasty due to anticipated problems with wound healing and associated increased incidence of peripheral vascular disease. Avascular necrosis (AVN) of the talus does not preclude TAA, particularly since this is usually partial or patchy. However, extensive or complete AVN is a contraindication to the surgical procedure. If AVN is partial and there is a concern for the possibility of talar component subsidence, check the condition of the subtalar joint. If there is any associated arthritis or involvement of the inferior portion of the body of the talus with AVN, consider performing a subtalar arthrodesis as a staged procedure prior to the TAA.
Advanced age does is not a contraindication, especially in patients with bilateral ankle arthrosis for whom bilateral ankle fusions might significantly impact the ergonomics of gait. Rheumatoid disease is an indication for TAA regardless of age, especially in those patients who have undergone multiple other joint replacements. Weight remains an important consideration when selecting a patient for TAA. While the prostheses have been tested to approximately 800 lbs. of load, some obese patients with small bones may have difficulty distributing the forces on a smaller sized prosthesis. The larger heavier patient does not present a concern, provided that a proportionately larger prosthesis can be inserted and provided that periarticular osteopenia is not present in these patients.
Figure XXX shows a lateral radiograph of a middle-aged woman who walked on the limb immediately after surgery while inebriated. The talar component was driven through the talus, and an arthrodesis was performed as a salvage procedure. This failed when she subsequently fell down a flight of stairs again while intoxicated and sustained midshaft tibia and fibula fractures, which were treated with an external fixator as a further salvage procedure. The general medical status and health of the individual are obviously factors which also come into decision making with TAA.
HINDFOOT ABNORMALITIES
It is imperative that hindfoot deformity be recognized and corrected before the TAA is performed. Depending on the magnitude of the deformity, this can be performed either before or in conjunction with the arthroplasty procedure. A triple arthrodesis should be staged prior to the arthroplasty procedure. In the presence of minimal deformity, an isolated subtalar arthrodesis can be performed at the same time as the TAA, as can an osteotomy of the calcaneus. However, it may be that too much stripping takes place with the addition of a subtalar arthrodesis, limiting the available blood supply to the talus. This is particularly the case when AVN of the talus is present. Currently we prefer to stage most hindfoot fusion procedures, waiting approximately 4 months before embarking upon TAA. This is based upon anecdotal experience, however, and others may have success performing the TAA in conjunction with hindfoot correction simultaneously.
In addition to focusing on structural deformity, we emphasize the need for soft-tissue balancing both medially and laterally. Instability of the ankle often accompanies hindfoot or tibiotalar varus deformity, and a lateral ankle ligament reconstruction must be performed, usually simultaneously.1 Typically, for a varus hindfoot with an unstable ankle, the deltoid is also carefully released (Fig. XXX).
INTRAOPERATIVE COMPLICATIONS
Misalignment
The most difficult technical aspect of TAA with the Agility prosthesis seems to be alignment of the cutting jigs and creation of the appropriate bone resection. Visualization of the cutting block with intraoperative fluoroscopy can be time-consuming, but care in making the bone cuts is essential and is probably the most important step of this procedure. Correct sizing and alignment of the tibial component will decrease the incidence of fracture of the malleoli and component malalignment. There is no need to adhere rigidly to the preoperative sizing, and if the tibial component does not seem to fit, it is preferable to size up or down accordingly. If one is uncertain about the size of the prosthesis based upon the preoperative template, then the tibial cut can be made without the cut for the central fin. This allows one to re-size with a larger or smaller component as required. Alternatively, if the cuts have already been made, it is still possible to recut the tibia or talus and later fill the defect with bone graft.
The talar cut is also difficult, and imprecise positioning, particularly in the sagittal plane, will often lead to an anterior-posterior tilt that might be difficult to correct. The surgeon must make sure that the foot is completely dorsiflexed to neutral before making the talar cut. If the foot is plantarflexed at the time of the talar cut, then the posterior subtalar joint is at risk for penetration. A good way to prevent this is to make a 5mm cut in the anterior talus, insert a free saw blade or Kirschner wire into the talus to check the position on the imaging unit.(Fig. XXX) (Fig. XXX) shows a posterior tilt to the component that has not affected early functional results but will inevitably affect ankle dorsiflexion. If progressive, this condition will lead to impaction or collapse into the subtalar joint.
When hindfoot varus or valgus deformity is present and is not corrected with the external fixator prior to the cut on the talus, this will affect the final hindfoot position. Theoretically, the fixator should be inserted on the ankle perpendicular to the axis of the foot such that the talus alignment is neutralized prior to the bone cut. The later problem, however, may be that after a neutral cut on the talus, soft tissue contracture of the deltoid, for example, may again tilt the ankle into varus. Clearly the deltoid needs careful release, although an alternative would be to cut the talus (for example, with hindfoot varus) offset, with more bone removed medially than laterally. This angular cut on the talus is not ideal, and it is always preferable to obtain ligament balance with a neutral cut on the talus. (Fig. XXX) demonstrates this tilt leading to varus deformity postoperatively, which required correction of the talar cut into a more planar position.
Prior deformity of the distal tibia can present challenges. The typical deformity is of the tibial plafond, caused by erosion of the tibia joint surface by a varus or valgus talus. The surgeon must not be overzealous in resection of the distal tibial bone, as the metaphyseal bone significantly weakens with distance from the subchondral bone. We have found it better to leave a lateral void and fill it with bone graft rather than make a cut proximal enough to get a smooth surface for the tibial component. However, this depends on the size of the defect created or the amount of bone that needs to be resected. The malalignment of the distal tibial plafond can thus easily be accommodated provided that the foot is brought into a neutral position for the tibial bone cuts. Malalignment of the distal third of the tibia, secondary to prior fracture, can be severe and may require supramalleolar osteotomy. If the angulation of the tibia is greater than ten degrees, then it is preferable to perform a distal tibia osteotomy as a planned and staged procedure.
Medial Malleolus Fracture
The medial malleolus is at risk several times during the insertion of the ankle prosthesis. One should try to use the largest sized prosthesis available that will fit adequately, but care should be taken not to encroach too boldly on the medial malleolus. The component is press-fit and can be very snug. The cut for the tibial fin may be slightly too medial in some cases, pushing the component against the medial malleolus and causing fracture. When sizing the prosthesis after making the bone cuts, be prepared to change the size as needed. Avoid cutting more bone from the medial malleolus when possible. Instead take some bone laterally from the fin cut, provided that the syndesmosis is not distracted to any great extent. Perhaps the most common error causing fracture of the medial malleolus is due to an oblique (not vertical) cut of the medial malleolus, such that when the tibial component is inserted, the posterior margin of the component abuts against the obliquely cut medial malleolus. If fractured, the medial malleolus must be fixed using small cannulated screws or a tension band during the same procedure. Fig XXX shows a healed fracture of the medial malleolus with internal fixation. The difficulty with these fractures is the potential need to immobilize the ankle postoperatively, causing arthrofibrosis and limited motion. If the fixation is stable and if no motion of the fracture is present, then the rehabilitation phase should not change and early motion is instituted.
Fibula Fracture
The operative procedure for the Agility ankle prosthesis emphasizes the enhanced stability of the tibial component by syndesmotic fusion. Once the tibiofibular syndesmosis is fused, the incidence of subsidence of the tibial prosthesis, ballooning lysis, and overall failure diminishes. Fracture of the fibula may occur as a result of an inappropriate fibula cut, leaving insufficient bone laterally. It is of course important to protect the fibula during the tibial cut as well, since the fibula lies far posteriorly relative to the tibia. Here again, the proper cutting jig alignment and care with prosthesis placement may help minimize this complication.
Another source of fracture occurs as a result of the syndesmotic screw fixation with overcompression of the fibula. Since we have been routinely using a plate on the fibula to enhance syndesmosis fusion, the complication of fibula fracture from screw insertion is no longer relevant. In the patient with post-traumatic arthritis after ankle fracture, if the fibula plate is still present, then the syndesmosis screws are inserted through the plate (Fig. XXX). If a fracture of the fibula occurs during insertion of the tibial component or prosthesis, it must be fixed with a 1/3 tubular plate. We do not recommend observation of a fibula fracture occurring either intraoperatively or postoperatively. Valgus instability with lateral shift and tilt of the talus may occur.
The displacement of such a fracture is usually minimal because the distractor is in place during component placement, but one should resist the temptation to leave the fracture without stable internal fixation.
Talus Fracture
Of all the intraoperative complications, few are as troubling as fracture of the talus. With advanced cases of joint deterioration, the talus may be smaller and weaker than necessary to support the talar component. Careful surgical technique, such as checking the talar cut intraoperatively by fluoroscopy as described earlier, can help prevent loss of talar support. If AVN of the talus is present, catastrophic failure may occur. If in doubt, an MRI or CAT scan should be obtained, determined by the presence of hardware. If fracture occurs, a subtalar arthrodesis or bone grafting of the talus with limited internal fixation can be considered. If this occurs postoperatively, the same options hold although a tibiotalocalcaneal fusion should also be considered.
POSTOPERATIVE COMPLICATIONS
Syndesmosis Nonunion
A successful syndesmosis fusion seems to be tantamount to stability of the tibial component and success of the TAA. When the syndesmosis fails to fuse, a revision procedure with bone graft and more rigid fixation is advisable. Pyevich et al.9 reported an incidence of delayed union of 29% and a nonunion rate of 9% in 86 patients. A total of 67% of tibial component migration was associated with either delayed union or nonunion of the syndesmosis. The incidence of ballooning lysis was also considerably higher in patients with delayed union or nonunion of the syndesmosis, and lysis was present by 2 years in all patients with a syndesmosis nonunion. However, there did not seem to be a correlation with syndesmosis failure and clinical outcome in their series. This has been a source of constant concern, and the senior author (MSM) found a 34% rate of combined syndesmosis delayed union and nonunion in the first 90 TAA performed, similar to that reported by Pyevich et al. However, most recent 78 TAA have been performed by the senior author using a 5-hole semitubular plate on the fibula combined with stabilization of the syndesmosis with two or three fully threaded cancellous screws gaining purchase on 4 cortices. (Fig XXX) Since this was instituted, we have not encountered a single case of nonunion of the syndesmosis. When applying the fixation through the plate, a visible compression of the plate into the tibial component occurs, even though fully threaded screws are used. The significant improvement in our results since the application of the plate must have to do with distribution of the stress over the distal fibula and with compressing the fibula into the lateral margin of the tibia and tibial component.
Bone graft of the syndesmosis is most important distally. We utilize cancellous chips harvested from the bone resections for the TAA components. We do not advocate using a bone block in the syndesmosis because this can inadvertently push the fibula away from the tibia, again causing a nonunion. Care should be taken to remove as much cartilage and cortical bone from the graft, but plenty of bone is usually available. One element important in preventing nonunion is to avoid use of the lamina spreader to separate the syndesmosis when making the bone cuts. The spreader is helpful when exposing the fibula for preparation but can easily overdistract the syndesmosis, hindering correct fit of the tibial component. The tibial component should not push the fibula away. If this is recognized, either more bone must be cut from the fibula or a smaller sized tibial component must be inserted.
The occurrence of nonunion remains a worrisome situation. If this occurs, we advocate revision with cancellous bone graft. The procedure is relatively simple, involving exposure of the syndesmosis with roughening of the cortical bone followed by insertion of cancellous allograft or autograft. The fibula should be stabilized with a 5- or 6-hole semitubular plate as described for use in the primary TAA. (Fig XXX) Following revision fusion, non-weightbearing is important, but too much immobilization in a cast may limit eventual ankle motion. Full range of motion with therapy is permitted until fusion is noted.
Component Loosening
Pain occurring postoperatively, particularly “start-up” pain, may suggest loosening of one or both components. If loosening is not evident on radiographs, a CAT scan will diagnose loosening. Normally, following stable bone healing to the porous coated prosthesis, striations are present radiating out from the prosthesis proximally. These striations are not present if loosening is present. Larger components are available for revision procedures, and a 2+ polyethylene liner is sometimes helpful to maintain soft-tissue balance. Care must be taken to avoid removing too much bone to insert the next larger prosthesis because the tibial component must rest under the stronger subchondral bone. Resecting too deeply could cause the tibial plate to rest on softer bone and may lead to subsidence of the prosthesis.
Fusion
When all else fails, the ultimate salvage of TAA is an ankle fusion. The difficulty of this procedure has been addressed by several studies in treating failed ankle arthroplasty in the past,3,5,7 and one surgeon has reported using external fixation to fuse nonunions resulting from such attempts.6 The main factor in decision making with respect to arthrodesis is the manner in which the procedure will be performed. The method has implications with respect to the type of fixation used, the method of bone grafting, and the question of whether the subtalar joint is included in the arthrodesis. Wherever possible, the fusion should be limited to the tibiotalar joint alone, preserving whatever remaining motion is present in the subtalar joint. The problem seems to be with the remaining body of the talus, since fixation, either internal or external, will fail if the talus is avascular, collapsed, or largely absent. The bone loss resulting from the attempted TAA may therefore necessitate a tibiotalocalcaneal fusion. Under normal circumstances, 12 mm of bone is removed during the Agility TAA. If this is complicated by failure necessitating arthrodesis, the bone loss is usually more substantial, and a decision must be made, sometimes intraoperatively, about the viability and mechanical capacity of the talus to be fused to the tibia and to support the body weight. The tourniquet can be released intraoperatively to assess the bleeding bone of the talus after component removal and resection. With the availability of intramedullary fixation devices which span the ankle as well as the subtalar joints (tibiotalocalcaneal arthrodesis), there seems to be a preference to fuse both joints without attempting to save the subtalar joint.
The size of the defect after removal of the components will determine the type of bone graft. For more substantial defects, if iliac crest structural autograft is used, then two separate grafts can be positioned side by side. If infection is present, then the arthrodesis must be staged, with an antibiotic bone cement spacer, and intravenous antibiotics. (Fig XXX) When no local or systemic findings of infection are present, bone graft is inserted with internal fixation.
Allograft has been shown to be a very reliable substitute for autograft in lower extremity fusion procedures. In a recent review of 75 procedures performed in 73 patients, Myerson et al. demonstrated that structural femoral head allograft was successful in obtaining bone healing with either osteotomy or arthrodesis in 92% of cases.8 When iliac crest allograft is used, it should be oriented to take advantage of the structural strength of a tricortical graft. Care should be maintained not to obliterate the graft with too large a cannulated screw. Augmentation of the allograft with cancellous autograft may stimulate the rate of arthrodesis by adding growth factors; spraying the allograft with buffy coat proteins derived from intraoperative plasma harvest (Symphony System, DePuy, Warsaw, IN) also holds promise for increasing the likelihood of fusion. Some surgeons prefer to use implantable bone stimulators (EBI, Parsipanny, NJ) to help increase the likelihood of fusion in difficult situations. A method of arthrodesis with which we do not have experience, but which may have promise, is the use of a cage or cages filled with cancellous graft. The cage(s) are orientated vertically and provide adequate structural support, obviating the need for structural graft of any sort.
Infection
Treatment of an infection should follow the guidelines developed for total knee arthroplasty. Unfortunately, minor wound problems are common, and minor wound dehiscence anteriorly is a problem in about 8% of our patients. These problems heal quickly with topical wound management. More advanced infections should be addressed with removal of the components, insertion of an antibiotic spacer, and administration of intravenous antibiotics. While antibiotic beads can be made to fill the gap, a solid cement spacer acts to keep the joint space open and more stable.
Groth and Fitch 5 advised direct tibiotalar fusion without grafting for septic arthroplasty. Such an approach may lead to unacceptable shortening of the limb but remains a possible treatment option. Bishop and co-workers2 studied the use of a free vascularized bone graft to the septic ankle with either contralateral fibula or iliac osteocutaneous graft, but reported reactivation of infection in half of the patients studied. Thordarson and co-workers11 report better results with more aggressive debridement and delay in graft placement. Aggressive debridement is the key factor in eliminating osteomyelitis. Later revision to a TAA or, more appropriately, an ankle fusion, can be accomplished once the bone is free of infection.
CONCLUSION
The problems that arise during surgery and after failure of TAA may be formidable to even the most experienced surgeon. As with any operative procedure, the consideration of this procedure should be tempered with the difficulty in salvage. This article is an early summary of some of the initial problems with the Agility (DePuy) total joint ankle arthroplasty.
REFERENCES
1. Acevedo JI, Myerson MS: Technique tip: modification of the Chrisman-Snook technique. Foot Ankle Int 21:154, 2000
2. Bishop AT, Wood MB, Sheetz KK: Arthrodesis of the ankle with a free vascularized autogenous bone graft. Reconstruction of segmental loss of bone secondary to osteomyelitis, tumor, or trauma. J Bone Joint Surg Am 77:1867, 1995
3. Carlsson AS, Montgomery F, Besjakov J: Arthrodesis of the ankle secondary to replacement. Foot Ankle Int 19:240, 1998
4. Frey C, Halikus NM, Vu-Rose T, et al: A review of ankle arthrodesis: predisposing factors to nonunion. Foot Ankle Int 15:581, 1994
5. Groth HE, Fitch HF: Salvage procedures for complications of total ankle arthroplasty. Clin Orthop 224:244, 1987
6. Kitaoka HB: Salvage of nonunion following ankle arthrodesis for failed total ankle arthroplasty. Clin Orthop 268:37, 1991
7. Kitaoka HB, Romness DW: Arthrodesis for failed ankle arthroplasty. J Arthroplasty 7:277, 1992
8. Neufeld SK, Uribe J, Myerson MS: The use of structural allograft to compensate for bone loss in arthrodesis of the foot and ankle. Foot Ankle Clin ?:?, 2002
9. Pyevich MT, Saltzman CL, Callaghan JJ, et al: Total ankle arthroplasty: a unique design. Two- to twelve-year follow-up. J Bone Joint Surg 80A:1410, 1998
10. Saltzman CL: Total ankle arthroplasty: state of the art. Instr Course Lect 48:263, 1999
11. Thordarson DB, Ahlmann E, Shepherd LE, et al: Sepsis and osteomyelitis about the ankle joint. Foot Ankle Clin 5:913, 2000
|
 |
