Ankle arthrodesis was first described by Albert in 1879. 1–10 Since that time, more than 30 techniques have been described for fusion. These include the use of external and internal fixation. Recently, with the advent of cannulated screws, more clinicians have used internal fixation devices to gain rigid compression across the fusion site. Rigid internal fixation has the advantage of leading to earlier fusion and the ability to institute early range of motion of the adjacent joints. In a recent review by Coester and Saltzman et al., 1 the loss of adjacent range of motion was shown to be a main risk factors for decreasing long term satisfaction.
Unlike osteoarthritis of the hip and knee, ankle arthritis is most commonly associated with trauma. The initial traumatic even may have an impact on the adjacent joints and may contribute to long-term problems of arthritis in those joints. With the advent of new prosthetic designs for total ankle replacement, the possibility of sparing the adjacent joints by maintaining motion at the ankle is intriguing. In conjunction with the FDA, Mann and Coughlin et al. are currently conducting a multi-center study to compare the results of ankle arthroplasty and ankle fusion. This may provide insight into the prognosis of the adjacent joints.
The potential complications and difficulties associated with the failure of the ankle prosthesis, the need for revision, and the difficulties and complications associated with conversion to fusion must be weighed against the incidence and effect of arthrosis in the adjacent joints. At the present time, ankle arthrodesis still remains the gold standard for the treatment of ankle arthritis. As long-term studies are conducted to compare the results of ankle arthroplasty to ankle fusion more information will be forthcoming.
The main indication for ankle fusion is the development of arthritis of the ankle joint. The majority of cases will result from trauma. Other causes include rheumatoid arthritis, septic arthritis, hemarthrosis, and congenital deformity.
Careful history of the patient's symptoms and concomitant medical problems should be obtained. The degree of pain is a significant factor in the decision making process. The presence of arthritis on x-rays does not always correlate with the patient's degree of symptoms. Conservative management with the use of a molded ankle foot polypropylene ankle-foot orthosis (AFO) or double upright AFO can be successful in the management of the patient's pain. The use of nonsteroidal antiinflammatories may also be considered. These modalities are helpful if the patient's generalized medical condition would preclude the patient undergoing surgery.
The condition of the adjacent joints should be assessed. The presence of subtalar and transverse tarsal joint arthritis would decrease the probability of a successful pain-free result. These joints may also require fusion, which would lead to further stiffening of the hindfoot. The presence of full unrestricted motion of the subtalar and transverse tarsal joints is a good prognostic factor. Alignment of the hip and knee should also be assessed, as well as alignment of the midfoot and forefoot. Malalignment caused by involvement of the more proximal joints may need to be addressed before proper alignment of the ankle fusion can be obtained. Similarly, misalignment of the forefoot may impact on the ability to maintain a plantigrade position of the foot following fusion.
In patients with a history of trauma, injury to the adjacent joints and tendons may have a negative impact on the prognosis. In rheumatoid arthritis, involvement of the adjacent joints may also influence the ultimate outcome of the procedure. A history of previous joint infection would be a relative contraindication to the use of internal fixation. Studies have also shown that a history of smoking decreases the chances of a successful outcome. The author attempts to persuade all patients undergoing this procedure to quit smoking before surgery.
The presence of an open draining wound would be a contraindication to proceeding with surgery. A staged approach with initial debridement and culture, followed by an appropriate course of antibiotics, may allow for conversion to fusion. In patients with rheumatoid arthritis consideration should be given to stopping Embril and Methotrexate during the perioperative period.
After completion of a thorough history of the patient's initial injury and any associated medical problems, a complete physical examination should be performed. Observation of the patient's gait can be helpful in advising the patient of expected amount of improvement of their gait following surgery. Measurement and documentation of the degree of preoperative deformity should be performed. The degree of internal and external rotation of the foot compared with the contralateral normal extremity should be noted.
Complete neurologic evaluation and vascular evaluation of the extremity should be performed. The presence of any decreased sensation or Tinel's over any of the sensory nerves should be documented. The condition of the skin with the presence of scars or soft tissue defects should be evaluated, especially when a previous open injury caused the patient's arthrosis. The posterior tibial and dorsalis pedis pulses should be palpated and documented. If these pulses cannot be palpated then arterial brachial index studies should be requested.
It is critical to obtain standing, full weight bearing radiographs of both feet and ankles. AP, lateral and mortise views of the ankle and AP, lateral and oblique views of the foot should be reviewed. The presence of any arthritic changes or misalignment of any of the other joints in the foot should be documented. The alignment of the talus within the ankle mortise should be noted. The presence of an uneven wear pattern can influence the need for bone grafting to fill in these defects.
The talus and distal tibia should be assessed for any areas of avascular necrosis or residual infection. If this possibility exists based on the patient's history, but the radiographs are not conclusive, then an MRI scan with gadolinium should be performed. This allows planning of bone grafting for any areas of avascular change. If infection is suspected the MRI with gadolinium may delineate the areas of residual infection. A labeled white scan can be considered to help differentiate between inflammatory changes of arthritis and the presence of residual infection. White blood cell count, sedimentation rate, and C-reactive protein levels should also be assessed and may be helpful in evaluating the presence of infection and the possible need for intraoperative bone cultures and postoperative antibiotic management.
Transfibular approach is performed through an incision over the lateral aspect of the leg beginning 6–8 cm proximal to the level of the ankle joint directly over the fibula (Fig. 1). The incision is carried to the tip of the fibula and then curved toward the base of the fourth metatarsal. Dissection is carried out through the subcutaneous tissues to the level of the distal fibula. The distal fibula is exposed subperiosteally to avoid any injury to the sural nerve posteriorly and superficial peroneal nerve anteriorly (Fig. 2).
As the dissection is carried from the distal fibula towards the base of the fourth metatarsal, the fat of the sinus tarsi and origin of the extensor digitorum brevis may be elevated and reflected medially. It is important that that the muscle be reflected in the medial direction towards the origin of its vascular and neural supply. This approach allows visualization of the posterior, medial, and anterior facets of the subtalar joints without violation of those joint surfaces.
An osteotomy of the distal fibula is performed 1.5 cm proximal to the tibial plafond (Fig. 3). It is directed from proximal lateral to distal medial. The distal medial end of the cut should end just proximal to the most lateral aspect of the tibial plafond (Fig. 4). The osteotomy may be made more proximal if distal fibula is needed for bone graft. The anterior talofibular ligament, calcaneofibular ligament, posterior talofibular ligament, and syndesmotic ligaments should then be resected, and the distal fibula should then be removed from the wound (Fig. 5). The distal fibula can be split and the cancellous bone can be harvested for lateral use as bone graft.
Subperiosteal dissection can then be carried out over the distal anterior portion of the tibia. By releasing the joint capsule anteriorly and elevating the periosteum, visualization of the anterior spurs on the distal tibia and neck of the talus can be accomplished. These spurs should be debrided with the use of a rongeur. In some instances, where there is excessively large spur formation, a small sagittal saw may be helpful in restoring the normal contour of the distal tibia. Dissection should be minimized over the neck of the talus to avoid injuries to the arteries supplying the talus. A synovectomy may be performed at the juncture if required.
Visualization of the ankle joint is aided by placing a cob elevator into the ankle joint and using it to help free up any adhesions, both anteriorly and posteriorly. The cob can be moved into the posterior aspect of the ankle to help release the capsular tissue. Care should be taken to keep the cob in the ankle joint itself. If posterior spur formation is present, then this can be debrided. Once the dissection has been carried out visualization of the joint can be aided by placing a laminar spreader into the ankle joint (Fig. 6). This allows distraction of the ankle for inspection of the dome of the talus, tibial plafond, and under surface of the media malleolus. In situations where minimal bone loss is present, the joint surfaces can be prepared by utilizing a sagittal saw to remove the remaining subchondral bone on the distal tibia plafond and dome of the talus. This then provides two broad cancellous surfaces. However, if there has been uneven wear of the tibial plafond or the dome of the talus, or if there areas of avascular necrosis, then the author's preference is to try to maintain the initial shape of the ankle joint and denude all of the surfaces of any residual cartilage and subchondral bone using rongeurs and curettes. In either instance, the goal is to obtain a broad surface of cancellous bone and to remove the residual subchondral bone. Once the joint surfaces have been prepared in this fashion, they are then feathered with a 1/4-inch osteotomy to increase the surface area.
The talus should then be reduced into the ankle joint. At this time, the need for structural bone graft can be assessed. The ultimate alignment of the joint should place the foot in neutral, dorsiflexion and plantar flexion. Rotational alignment should be assessed by comparing the contralateral side. In most instances, the midpoint of the patella should align with the midpoint of the ankle and the second or third web space. An attempt should be made to align the foot in the same position as the contralateral leg. The heel should be placed in approximately 5 degrees of valgus. Again, the exact amount of valgus can be assessed by comparing the contralateral ankle on physical exam when the patient is standing during the preoperative evaluation.
In order to achieve this alignment, it may be necessary to add structural bone graft or to perform a correction through the cuts that are made on the dome of the talus or tibia plafond. Care should be taken not to remove an excessive amount of bone, which can then lead to shortening. This may then cause impingement on the medial side of the ankle from the medial malleolus. The removal of the distal portion of the medial malleolus runs the risk of disrupting the deltoid ligament and the vascular supply that comes in through this ligament to the talus.
If additional bone graft is needed, other than that harvested from the distal fibula, then the use of a cancellous graft from the calcaneus or iliac crest can be considered. The use of allograft can be considered, although it is the author's preference to use autogenous bone.
Once the alignment of the joint has been established and the appropriate bone graft (if needed) has been placed, initial fixation is obtained using the guide wires from the large cannulated screw system. The author prefers the use of the 7.3 system. Several techniques of screw placement have been reported. The authors preference is to use parallel screws from infra lateral directed superomedial 11,12 (Fig. 7). Cross screws have also been described. In addition, in instances where the structural integrity of the fusion site is questionable because of previous trauma, a third screw is often used from the posterolateral aspect of the distal tibial directed down into the body and neck of the talus 13 (Fig. 8) The guide wires should be placed and radiographs should be obtained to assess the appropriate placement. In addition, the clinical alignment of the foot and ankle should be reassessed prior to placing the final fixation devices. If the radiographs are satisfactory then the pins are over drilled and fixed with cancellous lag screws (Fig. 9). After removal of the guide wires, the joint is thoroughly irrigated. The subcutaneous tissues are closed in layers using Vicryl sutures over a drain. The skin is then closed with staples.
The placement of a postoperative ankle or popliteal nerve block is helpful in reducing postoperative pain. The patient is then placed in a bulky compressive dressing using plaster splints to maintain stability.
Results to date from the procedure have been satisfactory. Our initial report on 18 patients demonstrated 17 of our 18 ankles fused on an average of 14.7 weeks. 12 Subsequent reports utilizing this technique demonstrated 88% arthrodesis at an average of 13.8 weeks of the 12 patients that had initial nonunion 9 healed after revision of surgery at an average of 22.9 weeks. 14 Overall, 90% of patients believe that the surgery helped them, 70% were satisfied without reservation, 18% were satisfied with reservation and 12% were dissatisfied. Reasons for dissatisfaction included pain and nonunion.
Other than perioperative complications from wound healing or infection that occur with any surgical procedure, the primary early complication is delayed or nonunion. It has been the author's experience that the rate of nonunion remains low. When it occurs, revision using the same technique has proved to be satisfactory. 14 In most instances, bone graft is not used in revision. The hardware is removed and the joint surfaces are prepared by dissection down to cancellous surfaces, which are then feathered again as described. Fixation is then reinserted. If there are structural defects then the author does not hesitate to add cancellous bone graft. Other studies have reported on the use of bone graft. 15,16 These have shown a 92% rate of successful fusion.
Malalignment of the fusion can cause significant problems. Excessive dorsiflexion will place stress on the heel pad and can cause pain when walking. In a patient with a neuropathic foot this can be devastating, as it may lead to the development of a neuropathic ulcer or infection.
Excessive plantar flexion will cause the patient to ambulate with a back knee thrust and vaulting gait. This puts excessive stress on the knee. Patients may compensate by externally rotating the leg. This causes increased stress across the posterior tibial tendon, spring ligament, and the medial collateral ligaments of the knee. Excessive valgus deformity similarly adds excessive stress across the medial aspect of the knee and hindfoot. This can lead to disruption of the spring ligament and posterior tibial tendon dysfunction, as well as knee valgus.
Varus deformity increases the stresses on the lateral border of the foot and can cause increased instability in the subtalar joint. Patients will often complain of pain under the fifth metatarsal base as they ambulate on the lateral aspect of their foot. These deformities can often be managed with shoe modification. In a patient with excessive dorsiflexion, a rocker bottom shoe may be helpful. With excessive plantar flexion, the use of a solid ankle cushion heel shoe (SACH) may help compensate. Varus and valgus deformities may be ameliorated with appropriate orthotic device to compensate for the foot alignment.
Rotational deformities are more difficult to correct. The patient generally better tolerates excessive external rotation. Excessive internal rotation will often cause the patient to have greater difficulty with ambulation, present problems with clearing the foot in the swing phase of gait, and add excessive stresses to the hip and knee.
The development of osteoarthritis in the subtalar or transverse tarsal joint may present the most worrisome long-term complications. Reports with a long-term duration of follow-up of 8–12 years demonstrate that some patients experience limitation in activities and degeneration of adjacent joints in the foot and ankle. A recent review from Coester and Saltzman et al. demonstrated symptoms in the subtalar joint in 2/3 of the patients, with a mean follow up of 22 years. 1 Eleven percent demonstrated severe arthritis, 33% moderate arthritis, and 52% demonstrated none. The authors noted that prolonged immobilization might be a contributing factor in the etiology of the degeneration seen in these patients. In addition, they were unable to evaluate the degree of initial trauma to those adjacent joints in the patient's initial injury. They concluded that patients should be advised of the potential for development of osteoarthritic changes in the adjacent joints. They noted that the majority of subjects in their study were satisfied with the overall outcome of the ankle arthrodesis, would choose this same treatment again, and would recommended it to another patient in a similar situation.
Patients are maintained in bulky compressive dressings and plaster splints for the first 10–14 days. They are instructed in nonweight-bearing crutch ambulation. The drain is generally removed the following day when there has been less than 30 mL of drainage in an 8-hour shift.
The patients are seen in follow-up 10–12 days following surgery. If satisfactory wound healing has occurred, staples are removed from the wound and the patient is placed in a short leg cast that is maintained nonweight bearing. At 6 weeks, the patient is assessed radiographically. If there is evidence of early bony union, they are allowed to commence partial weight bearing ambulation. At this junction, if there is satisfactory healing, they are allowed to remove the boot and begin active subtalar and transverse tarsal motion exercises for 5 minutes three times a day. Otherwise, they are instructed to stay full-time in the cast walker.
At 10 weeks, if progression to solid bony union is observed on radiographs, the patient is allowed to begin full weight-bearing ambulation. The frequency of the range of motion exercises is increased as well as the duration. The patients do 10 minutes of active motion exercises four times a day. At 12–14 weeks, the patient is started on home strengthening exercises with theraband. Once solid union is achieved radiographically, the patient is then weaned out of the cast walker and back into the appropriate shoe wear. In general, we recommend that patients begin ambulation in a shoe 4 hours a day for the first 4 days and then return to the boot. We then increase the amount of time that the patient uses the shoe by 1 hour every 4 days until they are completely weaned into shoe wear.
Once the patient has reestablished shoe wear it is then necessary to assess whether they need any accommodative devices. Some patients, especially those with excellent subtalar and transverse tarsal joint motion, generally do not need any assistive devices. If there is stiffness in the subtalar or transverse tarsal joints, then consideration should be given to adding a SACH heel and rocker bottom to help relieve the stress on the adjacent joints.
POSSIBLE CONCERNS, FUTURE OF THE TECHNIQUE
With the advent of the newer designs for total ankle joint replacement, the long-term concerns over arthritic changes in the adjacent joints of the foot and ankle are being re-evaluated. It has been documented that these changes will occur in a high percentage of patients. However, long-term satisfaction with ankle fusion remains high. Whether the long-term results for ankle replacement will be able to reproduce or improve on these results has not yet been determined.
The complications associated with revision of a failed ankle arthrodesis to a fusion have been well documented. 17–20 Revision of prosthetic devices are not yet widely available. Long-term studies will be needed to provide reliable information on whether the risks and benefits associated with ankle replacement are favorable or unfavorable when compared with arthrodesis. In a younger patient, the attractiveness of an ankle replacement to help preserve the other joints in the foot and ankle is readily apparent. However, the long-term complications, including the need of subsequent revisions, are not yet known. Further studies to evaluate this will hopefully provide those answers.
The transfibular approach to ankle fusions has been shown to provide a reliable, reproducible method for the treatment of end stage degenerative arthritis of the ankle. It provides the advantage of having a low rate of skin problems and a high rate of successful fusion. The ability to obtain rigid fixation and therefore the ability to begin earlier range of motion of the adjacent joints may hopefully diminish the development of degenerative changes over time. This technique remains a widely accepted and reliable procedure for the orthopaedic surgeon.
1. Coester LM, Saltzman CL, Leupold J, et al. Long term results following ankle
arthrodesis for post traumatic arthritis. J Bone Joint Surg 2001; 83A( 2):219–28.
2. Ahlberg A, Henricson AS. Late results of ankle fusion
. Acta Orthop Scand 1981; 52:103–5.
3. 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 1995; 77:1867–75.
4. Boobbyer GN. The long term results of ankle
arthrodesis. Acta Orthop Scand 1981; 52:107–110.
5. Buck P, Morrey BF, Chao EYS. The optimum position of arthrodesis of the ankle
. A gait study of the knee and ankle
. J Bone Joint Surg 1987; 69:1052–62.
6. Lynch AF, Bourne RB, Rorabeck CH. The long-term results of ankle
arthrodesis. J Bone Joint Surg 1988; 70:113–6.
7. Morgan CD, Henke JA, Bailey RW, et al. Long term results of tibiotalar arthrodesis. J Bone Joint Surg 1985; 67:546–9.
8. Said E, Hunka L, Siller TN. Where ankle fusion
stands today. J Bone Joint Surg 1978; 60:211–4.
9. Gruen GS, Mears DC. Arthrodesis of the ankle
and subtalar joints. Clin Orthop 1991; 268:15–20.
10. Albert E. Zur resektion des kniegelenkes. Wien Med Press 1879: 20:705–8.
11. Wapner KL. Ankle
fusions. Seminars in Arthroplasty 1992; 3( 1):39–42.
12. Mann RA, Van Manen JW, Wapner KL, et al. Ankle fusion
Clin Orthop 1991; 268:49–55
13. Holt HS, Hansen ST, Mayo KA, et al. Ankle
arthrodesis using internal screw fixation. Clin Orthop 1991; 268: 21–29.
14. Mann, RA. Arthrodesis of the Foot and Ankle
. In: Coughlin, MJ, Mann, RA eds. Surgery of the Foot and Ankle.
St. Louis, Mosby, 1999:651–699.
15. Kitaoka HB, Anderson PJ, Morrey BF. Revision of ankle
arthrodesis with external fixation for non-union. J Bone Joint Surg 1992; 74:1191–1200.
16. Kirkpatrick JS, Goldner JL, Goldner RD. Revision arthrodesis for tibiotalar psuedarthrosis with fibular onlay-inlay graft and internal screw fixation. Clin Orthop 1991; 268:29–36.
17. Kitaoka HB. Salvage of nonunion following ankle
arthrodesis for failed total ankle
arthroplasty. Clin Orthop 1991; 268:37–43.
18. Kitaoka HB, Romness DW. Arthrodesis for failed ankle
arthroplasty. J Arthroplasty 1992; 7( 3):277–284.
19. Kitaoka HB. Fusion
Techniques for failed total ankle
arthroplasty. Sem Arthroplasty 1992; 3( 1):51–57.
20. Wapner KL. Salvaged of failed and infected total ankle
replacements with fusion
. In: Eds - Instructional Course Lecture. Chicago, AAOS 2002.