We inserted a prosthesis designed to replace the body of the talus in sixteen patients between 1974 and 1990 (Table I). Twelve patients had avascular necrosis: seven of them had had a displaced fracture of the neck of the talus with subluxation of the subtalar joint (a Hawkins type-II fracture), one had had a fracture of the neck of the talus and dislocation of both the subtalar and the tibiotalar joint (a Hawkins type-III fracture), and one had had a displaced coronal shear fracture through the body of the talus. Three of the twelve patients had been managed by a bone-setter, and the original radiographs were not available. Of these three patients, two were believed to have had a Hawkins type-II fracture and one was thought to have had a Hawkins type-III fracture. The remaining four of the sixteen patients had a severe crush injury.
Fourteen patients had been involved in a traffic accident. The remaining two patients had fallen from a height; one of them had had a concomitant compression fracture of the twelfth thoracic vertebra without a neurological deficit.
The ages of the patients at the time of the talar replacement ranged from twenty-one to forty-eight years (mean, thirty-two years). The body weight ranged from forty-eight to seventy-five kilograms (mean, fifty-eight kilograms) at the time of the operation and from fifty to seventy kilograms (mean, sixty kilograms) at the time of the latest follow-up evaluation. Thirteen of the patients were men, and three were women. The right talus was involved in twelve patents, and the left was involved in four.
The twelve patients who had avascular necrosis had had persistent symptoms for six months to two years (mean, fifteen months) (Table I). Radiographs showed that the talar body was sclerotic in all twelve patients and that it was fragmented in five of them. Total collapse occurred in three of these patients; partial collapse, in seven; and no collapse, in two. There was surrounding bone atrophy in twelve patients: nine who had avascular necrosis and three who had a crush fracture.
In order to be a candidate for the procedure, the patient had to have had more than one of the following findings for at least six months: pain in the region of the ankle and the foot while standing, walking, or at rest; swelling in the region of the ankle; a limp; more than 15 degrees of limitation of the active range of motion of the ankle joint; limitation of daily activities; and inability to perform a job. The replacement procedure also was indicated for patients who had a severe crush fracture of the talar body that could not be treated with open reduction and internal fixation. Involvement of the articular surface of the posterior facet of the calcaneus or of the distal aspect of the tibia was considered a contraindication to the procedure.
The radiographic indications for the talar body replacement in the patients who had avascular necrosis were evidence of definite avascular necrosis of the talar body with absence of the Hawkins sign, sclerosis of the body with surrounding bone atrophy, and fragmentation or collapse of the body.
All patients gave informed consent after having been told about the predicted results of the prosthetic replacement compared with those of arthrodesis or talectomy.
Determination of the Size of the Prosthesis
The size of the talar body prosthesis is determined by measuring the normal, contralateral talus with use of a slit-scanogram technique15 to avoid magnification. Thus, the exact size of the talus can be determined.
In the current series, the patient was placed supine on the operating table and the position of the foot and the ankle was adjusted in a foam block by one of us (T. H.). If there is any space between the foot and the foam block, plasticine (Faber-Castell AW, Smith Field, Australia), a substance that is similar to clay but that does not harden, is put into the space to ensure that the ankle and the foot are fixed and will not move while radiographs are being made.
An anteroposterior radiograph of the ankle is made, with the long axis of the foam block and the patient positioned at a right angle to the direction of the moving x-ray tube. The ankle and the foot are placed in the foam block, with the long axis of the foot in the vertical position. The position of the ankle is checked with portable fluoroscopy.
Next, an anteroposterior radiograph is made with the slit-scanogram technique by moving the x-ray tube across the ankle, and the width, the transverse concavity of the superior surface, and the inclination of the medial and lateral surfaces of the talar body are measured (Fig. 1-A).
A lateral radiograph of the ankle joint is made in the same manner, and the anteroposterior curve of the superior surface and the length of the talar body are measured (Fig. 1-B). The height of the talar body is measured by moving the x-ray tube along the leg.
Forty-degree lateral and medial oblique axial radiographs, as described by Isherwood, then are made with the slit-scanogram technique. The curve and width of the posterior facet of the talus are then measured (Figs. 2-A and 2-B).
Preparation of the Prosthesis
The unfinished talar body prosthesis is made of medical-grade 316L stainless steel (Kobe Steel, Kobe, Japan) (Fig. 3). A different prosthesis is used for the right and left sides. The prosthesis has superior, medial, lateral, and posterior surfaces, all of which are beveled to prepare the talar dome. The inferior surface of the unfinished prosthesis has a crude concave contour at the posterior aspect of the prosthesis. The plane of this contour meets the plane of the long axis of the prosthesis at a 45-degree angle. The contour is constructed as an inferior concave curved surface to serve as the posterior facet for articulating with the posterior facet of the calcaneus. There is a flat area at the anterior aspect of the inferior surface of the unfinished prosthesis, which is constructed as a convex curved surface for articulating with the middle facet of the calcaneus. The anterior aspect of the unfinished prosthesis is transverse and has a stem for insertion into the neck of the talus. The base of the stem is ten millimeters from the medial surface and eight millimeters from the superior surface. The stem is twelve millimeters in length, five millimeters in diameter at the tip, and nine millimeters in diameter at the base. The long axis of the stem is parallel to the long axis of the prosthesis.
The final shape of the prosthesis is constructed from the unfinished prosthesis with use of templates for sizing. The final shape of the templates was cut by one of us (T. H.) from the slit-scanogram radiograph. A transparent plastic plate, 5.5 centimeters wide, eleven centimeters long, and two millimeters thick, is placed on the anteroposterior radiograph of the ankle (Fig. 1-A). The width, the transverse concavity of the superior surface, and the medial and lateral inclinations of the talar body are copied onto the plastic plate with use of a pointed drawing pencil (Fig. 4, A). The plastic then is cut along the line of the copy with a fret saw and is used as the first template. A second template is prepared for the anteroposterior curve of the superior surface of the talar body (Fig. 4, B). A third template, for the thickness of the body (Fig. 4, C), and a fourth template, for the length of the body (Fig. 4, D), are copied from the lateral radiograph of the ankle (Fig. 1-B). Finally, a fifth and a sixth template, for the curve and width of the posterior facet of the talus, respectively (Fig. 4, E and F), are copied from the two Isherwood radiographs of the foot (Figs. 2-A and 2-B).
An unfinished prosthesis that is the same size as the talus that is to be replaced is selected. The final shape of the prosthesis is determined by reversing the templates on the surface of the unfinished prosthesis as the templates were obtained from the contralateral side. Lines then are drawn as landmarks on the surface of the unfinished prosthesis to serve as a guide for removing the excess metal.
The first template is placed on the anterior aspect of the unfinished prosthesis and is adjusted so that the base of the prosthetic stem is four millimeters from the superior aspect and five millimeters from the medial aspect of the template (Figs. 5-A and 5-B).
The second template is placed on the medial surface of the unfinished prosthesis and is adjusted so that the long axis of the template is parallel to the anterior cut surface of the prosthesis. The anterior end of the curve of the template then is adjusted so that it is at the same level as the highest point of the superomedial curve of the first line drawn on the prosthesis. This leaves the superior portion of the curve of the second template at the top of the talar dome. The anterior cut surface of the prosthesis corresponds with the osteotomy surface of the neck of the talus for insertion of the prosthesis (Fig. 5-B).
For the inferior surface of the prosthesis, the fifth template is placed on the anterior aspect of the crude inferior concave contour of the unfinished prosthesis and is adjusted so that the mid-point of the curve of the template is at the middle of the contour (Fig. 5-B).
The third, fourth, and sixth templates are used to confirm the thickness and length of the talar body and the width of the posterior facet of the talus when the final shaping of the prosthesis has been completed.
The prosthesis was shaved along the drawn lines by one of us (T. H.) with use of grinding machines. The anteroposterior curve of the superior surface is shaved first, followed by the width and the transverse concave curve of the superior surface of the talar body. The proper length of the prosthesis is achieved by shaving off its posterior aspect. Accuracy is controlled with use of the fourth template.
The medial and lateral surfaces of the prosthesis then are shaved by narrowing inward from the anterior to the posterior aspect of the prosthesis at a 5-degree angle. Next, the crude inferior concave contour of the inferior aspect of the unfinished prosthesis is shaved to form a perfect concave curved surface that will serve as the posterior facet. During this process, the diameter of the curve and the proper thickness of the talar dome can be achieved with accuracy with use of the fifth and third templates. The proper width of the inferior concave surface can be achieved by shaving off the anterior aspect of the curve with use of the sixth template to maintain the accuracy of the curved surface.
Our experience has shown that the posterior facet of the prosthesis should be slightly flatter and wider than the contralateral talus in order to ensure a satisfactory articulation with the posterior facet of the calcaneus. The flat area of the inferior aspect of the unfinished prosthesis distal to the inferior concave curve is beveled so that it forms a convex curve five millimeters deep just proximal to the base of the prosthetic stem in order to approximate the contour of the middle facet of the calcaneus; however, it does not directly match the contour of the articular surface of this facet so that direct contact of the prosthesis with the middle facet can be avoided. This is done because the distal portion of the middle facet remains intact as the osteotomy through the neck of the talus passes through the mid-portion of the middle facet.
The accuracy of the preparation of each surface of the final prosthesis is confirmed with use of the corresponding template (Fig. 5-C). Because the bone that is to be replaced may not be exactly the same size as the normal talus, because the appearance of the foot may vary with small changes in the radiographic position of the foot and the ankle, and because the thickness of the articular cartilage may affect the size of the final prosthesis, four more sizes of prostheses are prepared to ensure a proper fit. Two of these additional sizes are 0.5 and one millimeter smaller, and the other two are 0.5 and one millimeter larger. New templates are constructed for these additional prostheses by drawing a line 0.5 or 1.0 millimeter outside the line of the original templates for the two larger prostheses and inside the line for the two smaller ones. Finally, the prostheses are polished to obtain smooth surfaces (Fig. 6).
Operative Technique
The talus is exposed with use of a transmedial malleolar approach. The skin incision is started seven centimeters proximal to the tip of the medial malleolus. The line of the incision runs along the posterior border of the distal part of the tibia, continues downward to the posterior edge of the medial malleolus, curves one centimeter distal to the tip of the medial malleolus, and extends anteriorly to the insertion of the anterior tibial tendon. The skin flap is undermined anteriorly. The anterior border of the medial malleolus and the distal part of the tibia are identified, and a capsulotomy of the anteromedial aspect of the ankle joint is performed.
The posteromedial aspect of the ankle joint is exposed by splitting the sheath of the posterior tibial tendon, and the tendon is retracted posteriorly so that the posterior edge of the medial malleolus and the posterior aspect of the distal part of the tibia can be visualized. Care is taken to avoid injury to the posterior tibial nerve and vessels. A capsulotomy of the posteromedial aspect of the ankle joint, as well as a 45-degree oblique osteotomy at the base of the medial malleolus, then are performed. The medial malleolus and the deltoid ligament are freed up and pulled distally. A portion of the deltoid ligament is incised along the fibers to free the medial malleolus downward for good exposure of the distal portion of the body of the talus.
The talar neck and body, sustentaculum tali, posterior part of the talar body, and posterior facet of the subtalar joint are identified. The entire talar body is removed with an osteotomy at the junction of the body and the neck. The osteotomy is started at the medial aspect of the talar body; the direction is in line with the anterior aspect of the superior articular surface of the talus. The surface of the osteotomy must be perpendicular to the long axis of the talar neck.
After removal of the talar body, it is possible to visualize the posterior facet of the calcaneus, the posterior half of the middle facet of the calcaneus, and the articular surface of the distal part of the tibia and of both malleoli. The curvature and the width of the posterior facet of the calcaneus and the width of the distal articular surface of the tibia are estimated with use of bendable wire. The posterior half of the middle facet of the calcaneus is ignored because the convex curved surface at the anterior aspect of the inferior surface of the prosthesis does not come into direct contact with the articular surface of this facet.
A hole, one centimeter in diameter and 1.5 centimeters in depth, is made in the talar neck with use of a drill and curets. The entry point of the hole is four millimeters distal to the superior aspect and five millimeters lateral to the medial aspect of the talar neck. The best size of prosthesis is determined by sequentially trying the five prepared implants, starting with the smallest. The prosthesis that is selected is inserted in the ankle mortise, and the talar neck is reduced onto the prosthetic stem. The articulation of the prosthesis with the posterior facet of the calcaneus and the articular surface of the distal part of the tibia is examined to ensure that the match is good. It is not necessary to try the other prostheses after the proper size has been selected.
It is very important that the posterior facet of the talus sit accurately on the posterior facet of the calcaneus; if the remaining talar neck is too long, the position of the prosthesis will be more posterior and there will be only a partial articulation of the posterior facet of the prosthesis with the posterior facet of the calcaneus. If this is the case, the talar neck must be shortened in order to achieve good articulation of the prosthesis in the ankle mortise and with the posterior facet of the calcaneus. A maximum of four to five millimeters of shortening can be performed without creating a problem with insertion of the stem.
Next, the proper thickness of the prosthesis is confirmed by observing whether the long axis of the talar neck and that of the prosthesis are in the same straight line. If they are not, the plane of the osteotomy of the talar neck should be evaluated to determine whether it is at a right angle to the long axis of the talar neck. If it is not, the osteotomy must be revised. If the osteotomy is perfect, the entry hole for the prosthetic stem should be enlarged to accept the stem in a line parallel to the long axis of the neck.
The prosthesis is removed, and the hole in the talar neck is half filled with bone cement. It is difficult to obtain good seating of the stem in the neck if more cement is used because additional advancement of the stem is blocked by the excess bone cement inside the hole. The prosthesis is reinserted in the ankle mortise, and the stem is reduced into the hole in the talar neck. The pusher device is applied at the posterior aspect of the prosthesis to move the prosthesis forward and obtain good seating of the stem in the talar neck. Simultaneously, the heel is kept in the neutral position and the ankle is kept in neutral. Axial force is applied at the heel in a neutral direction along the tibia in order to eliminate medial or lateral tilt of the prosthesis. The posterior tibial and flexor hallucis longus tendons are checked to ensure that there is no entrapment in the region of the posterior facet. The ankle is maintained at 10 degrees of dorsiflexion. The medial malleolus then is reduced and is fixed with two malleolar screws. The position is checked with portable fluoroscopy before the skin is closed. Suction drainage is put in place, and a soft splint is applied around the ankle.
Postoperative Management and Follow-up
An active range of motion of the ankle and the subtalar joint is started on the fifth postoperative day, and progressive walking is allowed after three months. The malleolar screws are removed after the site of the osteotomy of the medial malleolus has healed, usually between eight and twelve months postoperatively.
At three weeks, the congruency of the articulation of the prosthesis in the ankle mortise and with the posterior facet of the calcaneus is evaluated radiographically with use of the slit-scanogram technique. Anteroposterior and lateral radiographs of the ankle are made to evaluate the articulation of the prosthesis in the ankle mortise, and two Isherwood radiographs of the foot are made to evaluate the posterior subtalar facet.
Slit scanograms of the ankle and the posterior facet of the subtalar joint were made at six-month intervals during the first two years and at two-year intervals during the next eight years. After ten years, radiographs of the ankle and the subtalar joint were made at two-year intervals or when the patient had pain in the ankle or the foot. The clinical results were evaluated by one of us (T. H.) at five years, at six to ten years, or at eleven to fifteen years after the operation and were graded as satisfactory or unsatisfactory with use of scoring systems for the talocrural and the subtalar joint (Tables II and III). The over-all result was graded as satisfactory when both joints had a satisfactory result, and it was graded as unsatisfactory when one or both joints had an unsatisfactory result.
The congruency of the prosthesis with the tibia and with the posterior facet of the calcaneus was good in all but one patient (Case 3), in whom the diameter of the curvature of the posterior facet of the prosthesis was smaller than that of the posterior facet of the calcaneus. This patient had persistent pain and swelling of the ankle, and radiographs showed erosion of the posterior facet of the calcaneus (Fig. 7).
Four patients had mild swelling of the ankle and temporary mild pain in the posterior aspect of the ankle during walking. The pain resolved in four to five months, and the swelling abated within three months.
No patient had a wound infection, wound necrosis, a neurovascular injury, or a disturbance of the function of the tibialis posterior or flexor hallucis longus tendon.
The radiographs showed no sclerosis of subchondral bone or irregularity of the joint spaces of the ankle or the mid-tarsal joint or of the posterior facet of the calcaneus except in one patient (Case 1, as will be discussed).
Three patients who were evaluated five years postoperatively had a satisfactory result. One patient (Case 3) had an unsatisfactory result at eight months (Table I and Fig. 7), at which time the prosthesis was removed and an arthrodesis of the tibiotalar neck was performed. Three patients who were followed for six, seven, and ten years had a satisfactory result (Table I). Of nine patients who were followed for eleven years or more, all but one had a satisfactory result (Table I). The ninth patient (Case 1) had a satisfactory result until thirteen years postoperatively (Fig. 8-A). The prosthesis then failed, and the result was judged to be unsatisfactory. This patient had increased density of the osseous trabeculae around the posterior facet of the calcaneus and the distal part of the tibia. The stem of the prosthesis had sunk forward and downward into the talar neck (Fig. 8-B). A revision was performed with use of the same operative technique as previously, but instead of an osteotomy at the junction of the talar body and neck the prosthesis was removed with a spreader and the bone cement was removed with a 4.5-millimeter drill-bit and curets. Intraoperatively, the articular surface of the posterior facet of the calcaneus and of the ankle mortise appeared normal. There was no corrosion of the prosthetic surface. A prosthesis of the same size as was used in the primary procedure was inserted, and the stem was fixed into the talar neck with bone cement (Fig. 8-C). At the time of the most recent follow-up, the patient had good function of the ankle and the foot.
Because arthrodesis of the ankle or talectomy for the treatment of avascular necrosis of the body of the talus produces disability of the ankle and the foot5,8,10,11, one of us (T. H.) developed a prosthesis to replace the talar body. The body of the talus is covered mostly with articular cartilage and has no muscle origins or tendon insertions5,10,13; therefore, stability of the talus depends on the bones that constitute the ankle mortise and on the anatomical shape of the body itself. The anterior aspect of both the talus and the ankle mortise is wider than the posterior aspect, and this provides stability against a posterior shift of the talus in the ankle mortise5,10,13. The major ligaments of the ankle, including the deltoid and lateral collateral ligaments, attach to the talar neck and the calcaneus but not to the talar body. These ligaments provide anterior, lateral, and medial stability.
Avascular necrosis of the talus involves the body but not the head or the neck2,4,10,11,14. Thus, there is bone remaining for insertion of the prosthetic stem. The undersurface of the prosthesis is supported by the spring ligament, and the prosthesis is seated into a locked position in the medial longitudinal arch, which is an effective mechanical structure for absorbing the force transmitted from the talus5,13. The inferior surface of the talar body is curved in the region of the posterior facet; this curve is easy to construct. Because the posterior facet of the subtalar joint has a narrow range of motion in eversion and inversion13, the prosthesis can be used for a long period. In addition, the inclination of the posterior facet of the calcaneus provides a stable platform for the prosthesis as well as posterior locking of the prosthesis in the medial longitudinal arch of the foot5,13.
The operative technique does not affect the subsequent stability of the prosthesis in the ankle mortise, as the major portions of the deltoid and lateral collateral ligaments are preserved.
The follow-up radiographs showed no subchondral osseous sclerosis of the joints around the prosthesis. This may be related to the fact that the foot consists of a number of joints that surround the talus and form the arches of the foot. This configuration is highly effective in transmitting force from the talus5,13, in contrast to the mechanics of other joints such as the hip. Thus, there is not a high concentration of force around the prosthesis.
The preparation and the construction of the prosthesis used in the current study were based on an old technique for the shaving and polishing of metal and were performed manually. Therefore, not all of the prostheses used in our sixteen patients were perfect. Measurements of the size of the prosthesis were determined on the basis of the contralateral talus with use of a slit-scanogram technique, an old method that is not readily available at the present time. Magnetic resonance imaging can be used for these measurements, but the construction of the templates may present a problem. Further development with use of good biomaterial and the best available technology is needed to achieve the best prosthesis possible. Despite these disadvantages, we believe that the talar body prosthesis may be useful in the treatment of avascular necrosis or severe crush injuries of the talus.