G. Mendolia* and the TALUS Group**
*Centre MCO Côte d'Opale, Boulogne S/Mer**

J. Berger (Strasbourg), C. Cermolacce (Marseilles), J.Y. Coillard (Lyon),
G. Copin (Illkirch), P. Delponte (Chateauroux), P. Diebold (Nancy),
J. Hummer (Nancy), G. Laroche (Boulogne S/Mer), G. Mendolia (Boulogne S/Mer),
G. Peltre (Paris), M. Perrin (Dijon), H. Rocher (Bordeaux), G. Rougereau (Paris)

ANKLE ARTHROPLASTY
The Ramses Ankle Replacement

Arthrodesis of the ankle joint gives satisfactory short- and medium-term results; however, in the longer term, it frequently leads to subtalar and mid-tarsal osteoarthritis (OA), which is difficult to treat.

For this reason, we have tried to perform ankle replacement, in some of our patients, as an alternative to fusion.

Implant designs have been improved; surgical technique has become simpler; and the causes of implant failure are better understood. All these developments have made it possible to obtain better results with ankle replacement.

In some conditions (discussed below), ankle prostheses are becoming a valid alternative to arthrodesis of the ankle joint.

This article deals with the results of our first 38 cases managed with the Ramses ankle replacement.

We shall be looking at

(1) the rationale for the Ramses device;
(2) the surgical technique;
(3) a description of our first 38 cases;
(4) the causes implant failure encountered
- submalleolar impingement
- short Achilles tendon syndrome; and
(5) fusion after ankle replacement.

DESIGN RATIONALE OF THE RAMSES ANKLE REPLACEMENT

There are many ankle replacement patterns; however, results overall have not been very encouraging. Rather than use the design concepts of the past, we felt that it would be necessary to obtain a better understanding of the biomechanical causes of ankle replacement failure, and to draw up a specification of what was expected of a satisfactory ankle implant.

A number of surgeons involved, in their daily practice, in foot and ankle surgery got together in the Talus group, to find the necessary answers.

We compared our experience gathered with previous designs and surveyed the recent literature. This exercise allowed us to draw up a list of requirements that any new ankle replacement would have to meet.

In this article, we shall confine ourselves to the four chief requirements identified:-
(1) The mobile bearing;
(2) The preservation of the anterior tibial cortex;
(3) The sides of the tibial component;
(4) the curvature in the coronal plane of the talar component.

(1) The mobile bearing
Congruent ankle designs consisting of only two pieces (a tibial and a talar component) should no longer be employed. Whilst this design is simple and apparently anatomically correct, it will almost invariably result in failure.

The reason for this failure is easy to see: With every step, heel-strike imparts a powerful postero-anterior translational force which acts on the anterior aspect of the tibial component. This force will weaken the anterior tibial cortex and make the tibial component tilt backwards and downwards. This will happen regardless of the fixation technique used, and has been seen even with large central stems.

As a rule, this complication will occur within one to two years from implantation.

The use of a mobile bearing is beneficial, since it converts shear and torque into the translational movement of a meniscal component.

It also makes the anatomical centres of rotation of the tibia and the talus match as the joint goes through its range of flexion and extension. A congruent ankle design without a mobile bearing cannot produce this point: Even though the two curvatures are the same, the anterior or posterior positioning of one component with regard to the other will lead to a loss of correspondence of the anatomical centres of rotation. This, in turn, will lead to eccentric wear, which will be

• anterior, if the tibial component is too far towards the back; and
• posterior, if the tibial component is too far towards the front.

(2) Preservation of the anterior tibial cortex
Some ankle replacements require a window in the anterior tibial cortex for their insertion. This window will weaken the bone at this level.

This region is increasingly stressed, with weight-bearing, as the foot goes into dorsiflexion.

The stress pattern has been elegantly shown in the strain gauge studies performed by Blaimont et al, who were looking into the biomechanical behaviour of the ankle joint under weight-bearing conditions (see diagram, from the paper presented by those authors).

The study of the biomechanical behaviour of the distal tibia showed the following pattern:-

(a) When the ankle joint is at neutral or plantar flexed, there will be uniform compression, with a weight-bearing axis along the midline of the section, and with full congruence of the joint.

(b) When the foot is dorsiflexed, the weight-bearing axis will move forwards, the joint will be eccentrically loaded, and the anterior part will be compressed.

It will be seen that the anterior cortex of the distal end of the tibia is a vital support, especially under conditions of single-limb weight-bearing during heel strike, when the foot is dorsiflexed.

Strain gauge measurements. Tibial strain diagram. From: Blaimont - Cahiers d'Enseignement de la SOFCOT

Cutting a window in the anterior tibial cortex will weaken the bone, and will lead to anterior cortical impaction, and to the backward tilting and eventual loosening of the tibial component.

Together with the removal of bone stock to create a socket for the a large central stem, the cortical window will produce a "flare", which will cause technical problems in cases that require secondary fusion.

(3) The sides of the tibial component (Figs. 1 and 2)
The ankle replacement must ensure a maximum of joint containment; however, the lateral and medial joint spaces must not be involved in the stabilization of the joint.

Transverse stability must be provided by the implant itself, whose sides must, therefore, be suitably designed.

Figure 1

Figure 2

If this requirement is not taken into account, transverse stability will have to be provided by the medial and lateral malleoli, which will be stressed by direct abutment on the medial and lateral aspects of the calcaneus.

Not only does an ankle replacement with an excessively wide mortise cause the talus to wobble, impingement on the malleoli will also cause pain, which may be severe.

(4) Curvature in the coronal plane of the talar component (Figs 3-4 and 5)
It is completely erroneous to believe that the ankle joint has only one axis of flexion/extension. There is ample evidence from all the biomechanical studies conducted (Barnett & Napier, 1952; Hicks, 1953; Inman, 1976; de Vogel & van Langelaan, 1970) to show that there are at least two distinct axes, one for plantar flexion and one for dorsiflexion; or, rather, that motion in the ankle joint occurs around a series of instantaneous centres of rotation.

The main movement of the ankle joint is flexion/extension; however, this movement is associated with abduction/adduction, and pronation/supination in the joint.

Whilst the greater part of abduction/adduction and pronation/supination is provided by the subtalar joint, a certain, albeit small, amount also occurs in the ankle joint.

Figure 3

Figure 4

Figure 5

The ankle and the subtalar joint work together to allow the foot to move and to compensate for ground irregularities.

The curvature in the coronal plane of the talar component allows varus/valgus movements to occur without loss of joint congruence. Some implant designs do not have a curved talar dome but a trapezoidal pattern; in these devices, varus/valgus movements produce loss of congruence, with gaping of the artificial joint either laterally or medially.

This process will off-load the gaping side of the bearing, and put increased stress on the opposite side. This overloading will lead to the crushing of the polyethylene (PE) at this site. The risk is particularly great in implants whose bearings do not have the required thickness of PE. These very thin PE components are very susceptible to damage by overloading.

The domed talar component is also a means of restoring, in the ankle joint, the overall movement pattern of what Blaimont and Bonnel have termed the "hind-foot joint complex."

SURGICAL TECHNIQUE

The technique required for the implantation of an ankle prosthesis must be straightforward enough to be mastered by any orthopaedic surgeon. On no account should it involve lengthy and tedious bone sculpting to match complex implant curves.

It is the implant that must match the host bone and the surgeon's requirements, not the other way round.

For the Ramses device, it was decided to have plane horizontal cuts, both at the tibial and at the talar end. Except for the spacing to be provided, these cuts are identical to the well-known resections practised for arthrodesis.

(1) Patient positioning
The patient is positioned supine, with a pad under the ipsilateral buttock, to prevent external rotation of the foot. The opposite limb is draped. A tourniquet is applied and inflated.

(2) Approach (Fig. 6)
Commonly, a longitudinal anterocentral ca. 10 cm long incision is employed.

Figure 6

The incision may be routed more laterally or more medially, to use a previous anterolateral or anteromedial approach.

This incision allows the surgeon to stay away from any strictly lateral or strictly medial approaches used in the past. Such incisions should not be reused, since they are ill-suited to the requirements of arthroplasty.

After the fascia has been incised in a longitudinal fashion, the incision is deepened between the tendon of the tibialis anterior medially, and the extensor digitorum longus tendons laterally. The neurovascular bundle is reflected (usually laterally) and protected.

The anterior aspect of the tibia must be exposed with great care, from the medial to the lateral malleolus; while the talus is exposed as far as the upper part of its head; also, a complete anterior capsulectomy needs to be performed.

Anterior osteophytes will need to be chiselled off, to provide a complete view of the joint line.

(3) Tibial cut (Fig. 7)
The tibial cutting guide is placed against the anterior tibial cortex.

Since the tibial joint surface is curved in the sagittal plane, a thickness of 1 cm of the cut in the anterior part corresponds to a thickness of 0.5 cm half-way along the a.p. distance.

Figure 7

The width of the tibial cut must be such as to extend from the base of the medial malleolus to the anterior tubercle laterally.

Care should be taken to ensure that a sufficient amount of bone is left standing at the base of the medial malleolus, so as not to weaken this structure.

The cutting guide comes in three widths (small, medium, or large). The correct placement of the guide is ensured by a pin aligning the guide on the axis of the tibia, and by the use of an image intensifier.

The cutting guide is tacked to the tibia by two pins inserted horizontally (in an antero-posterior direction). The cut, made with an oscillating saw, must be horizontal, and divide the bone completely, in an antero-posterior direction.

Resection is completed by two vertical, sagittal cuts, flush against the medial malleolus on the medial side, and the tibiofibular joint laterally.

These two sagittal cuts may be difficult to effect while the talus is still intact. In such cases, it would be expedient to resect the talus prior to the performance of the sagittal tibial cuts.

(4) Talar cut (Fig. 8)
The talar cutting guide is placed over the tibial guide, making sure that the sizes agree (small tibial cutting guide and small talar cutting guide, and so on through the size range).

Figure 8

The foot is placed at right angles to the axis of the lower limb, in such a way as to have the plant of the foot horizontal with regard to the body axis. The talar cutting guide is tacked to the bone with two horizontal (antero-posterior) pins.

The talar cut is made with an oscillating saw; it must be horizontal, and completely divide the bone in the sagittal direction. Ideally, it should be taken along the upper part of the neck of the talus, and remove 15 mm of the talar dome in its thickest portion.

Once this resection has been performed, the two sagittal tibial cuts can readily be made, and the tibial bone fragment removed.

The total "tibial + talar" resection thickness will vary with size; roughly, it amounts to 20 mm.

(5) Anterior, lateral, medial, and midline debridement
Once the tibial and talar resections have been performed, and the bone removed, complete joint debridement should be performed: This involves capsulectomy, the removal of all osteophytes, and the complete release of the lateral and medial talomalleolar joint spaces.

This stage of the procedure is very important (see below: Causes of failure - submalleolar impingement). A periosteal elevator must fit comfortably into the talomalleolar joint space, and it must be possible to make the joint lines gape with axial traction.

(6) Drilling of fixation holes in the tibia and the talus (Fig. 9)
The drill guide must be of the same size (small, medium, or large) as the previously chosen tibial cutting guide. It must also be of the same width as the tibial cut surface. The upper stop is placed against the anterior tibial cortex.

Figure 9

The foot rests on its heel on the operating table; this way, it will spontaneously go into anterior subluxation.

This subluxation should be reduced by a posterior drawer movement, to align the lower part of the drill guide with the anterior edge of the cut talar surface.

In this way, the anterior part of the drill guide will be aligned, on the tibial side, with the anterior cortex by means of its stop; and on the talar side, with the anterior part of the talar dome.

The holes for the two tibial pegs and for the talar peg are drilled, using drills with drill stops.

(7) Component trials (Fig. 10)
The drill holes made will allow the implant components to align automatically in the sagittal plane.

On the tibial side, a trial corresponding to the cutting guide used (small, medium, or large) is inserted.

Figure 10

On the talar side, the choice of implant size should be governed by the size of the cut surface to be covered. If need be, the talar size may be different from the size chosen for the tibia, since all talar sizes will work with all tibial sizes.

The trial mobile bearing is inserted. This bearing must be of the same width as the tibial component. The bearing is supplied in five sizes, making it possible for the surgeon to restore the ankle joint to its ideal height.

The ideal height is the one at which the ligament tension on the medial and the lateral side of the joint is such as to give good stability in the coronal plane.

Varus/valgus tests are performed. The talus must be seen to slide in the coronal plane, under the mobile bearing. During these manoeuvres, the joint surfaces must remain in contact. Any gaping and tilting of the bearing are a sign of instability, and must be remedied forthwith by the insertion of a thicker bearing.

The height of the implant must be such as to establish a clearance of at least 1 mm in the talomalleolar joints. Also, with the trials in place, the ankle joint must have a good range of movement (ROM). As a rule, plantar flexion will be good (40° to 60°). Above all, dorsiflexion must be at least 10° to 20°.

If, after all the steps described have been taken, dorsiflexion is found to be insufficient, or if there is an equinus deformity, the following action will need to be taken:

• Do not insert a thinner mobile bearing, since doing so would cause instability and pain.

• Lengthen the Achilles tendon, since that is the only procedure that will preserve a correct ligament tension and give a satisfactory ROM.

This action will prevent the occurrence of a short Achilles tendon syndrome (discussed below, under Causes of Failure).

(8) Insertion of final implant components (Fig. 11)
The final components may be cemented, if the quality of the patient's bone stock is less than optimal; or simply press-fitted. Whether or not cement should be used will be decided on a case-by-case basis.

Figure 11

Once the final tibial and talar components have been inserted, the final size of mobile bearing is chosen. If need be, a size different from that of the trial bearing may be used. The mobile bearing is simply impacted into place. A slight anterior drawer of 5 mm is normal, as are antero-posterior movements of the intermediate bearing during flexion/extension movements of the ankle joint.

(9) Wound closure and postoperative management
The incision is closed over a suction drain. The retinacula are closed over the tendons, and the skin is sutured.

A removable splint holding the foot at right angles is applied. This permits immediate weight-bearing using elbow crutches, which are usually discarded at about three weeks.

The splint is removed daily for hygiene purposes. Gentle passive and later active flexion/extension is practised in bed, without weight-bearing.

Active mobilization with weight-bearing is commenced at one month. The immobilizing splint is replaced by a flexible ankle brace, which supports the ankle and prevents oedema. The brace is discarded at three months, once proprioception has been recovered.

THE RAMSES ANKLE REPLACEMENT
OUTLINE OF REHABILITATION

1st month
Gradual return to full weight-bearing (2 walking aids, followed by one walking aid), wearing an immobilizing splint.

Without weight-bearing: gentle active and passive flexion/extension exercises.

2nd month
The immobilizing splint is discarded and replaced by flexible ankle brace.

Full weight-bearing without walking aids.

Active and passive flexion/extension exercises with weight-bearing.

Gait training.

Proprioceptive exercises.

3rd month
Flexible ankle brace continues to be worn.

Intensification of active and passive weight-bearing flexion/extension exercises.

Proprioceptive exercises and varus/valgus movements.

Gait training and motor coordination work.

End of 3rd month:
Brace is discarded - free use of ankle joint.

OUR FIRST 38 CASES

This was a somewhat heterogeneous series, done by a number of surgeons of whom many were at the bottom of the learning curve (several "first-timers").

The patients were followed up very regularly, and could not be included in the series unless they had been followed up for a minimum of two years.

(1) Indications
Clinically, all the patients had severe pain in the ankle joint and a pronounced limp.

The case mix was as follows: Seventeen cases of post-traumatic OA, after bimalleolar or tibial pilon fractures; six cases of avascular necrosis of the talus; seven cases of rheumatoid arthritis. Eight patients had primary OA of the ankle joint; when a more detailed history was taken, it appeared that these cases were, in actual fact, secondary to chronic instability of the ankle joint after neglected ankle sprains.

Arthroplasty was always offered as a possible means of postponing arthrodesis, rather than as an outright alternative to fusion. We think that good results can be achieved only if the patient has been fully informed, and is fully cooperative.

(2) Intraoperative complications
The following intraoperative complications were encountered:-
- One fracture of a weakened medial malleolus, which occurred during an attempt to insert an excessively wide mobile bearing;
- three instances of varus/valgus malpositioning;
- two cases of ankle instability, where an excessively thin bearing had been used. The bearing did not, however, dislocate.
- eight cases of insufficient dorsiflexion. These cases occurred at the start of our experience with this device. As discussed below, we would now perform Achilles tendon lengthening more readily than we did earlier on.
- three cases of painless stiffness of the ankle joint;
- seven painful ankles. It is interesting to see the detailed break-down of this group:

• In two cases, the pain was comparatively mild, and the patients happy with their result.
  
• In five cases, the pain was severe, and led to the removal of the implant. Two of these cases had a loosened implant, while in three cases the implant was soundly fixed. Pain was found to be secondary to subfibular impingement. This condition is discussed below.

Fusion was performed five times, using two iliac crest bone grafts wedged into the gap. Fusion was achieved in normal time (three months) in four of these cases.

One patient had a tight non-union, which was slightly painful.

The result of these five fusions was pain relief in three cases (those who had had implant loosening); and persistent lateral malleolar pain in the two cases where the implants had not been loose.

Pain relief was obtained when the lateral malleolus was partially resected. As discussed below, it appears, with hindsight, that this resection was all that was required, and that the implant (which was not responsible for the pain) could have been left in situ.

(3) Longer-term results
The review of our first 38 cases at a minimum of two years' follow-up showed the following pattern:-
- 28 satisfactory results, with a stable and pain-free ankle that did not interfere with walking or with activities of daily living;
- three cases of pain relief, but major stiffness, with a combined range of dorsi- and plantar flexion of only 10°;
- seven patients had severe pain. Two implants were left in situ, while five devices were removed - three for loosening, and two that were not loose.

The five secondary fusions performed produced
- three fair results, and two cases of persistent pain, of which one has a non-union line.

(4) Functional results
A review form was created, to allow comparison of the postoperative status with the pattern seen before surgery.

Pain
Of the patients reviewed, 20 were pain-free; eight had starting-up pain, often in the forefoot; three had weather aches; seven had severe pain. These seven patients are discussed in greater detail above.

Standing
Thirty-one patients had a stable ankle in single-leg stance. Seven were unstable; these were the patients with ankle pain.

Walking
Twelve patients were able to walk unlimited distances; 20 could walk > 1 km; while six were restricted to < 1 km.

Limp
No limp was found 13 times; a slight limp when walking unaided was seen 17 times; while an occasional slight limp with walking aids was observed eight times.

Stairs
Thirty patients could negotiate stairs normally; six had some disability; one had to use a banister; one was unable to go up or down stairs.

ROM
There was no fixed equinus deformity. Mobility was expressed as combined motion, i.e. dorsiflexion plus plantar flexion.
The ROM was found to be 40° to 50°, in 28 cases; 40° to 30°, in three cases; and 30° to 20°, in two cases. Five ankles were stiff in neutral position; of these, three were painless, while two were painful and resulted in secondary fusion.

Stability
Stability was good in 28 cases; seven patients had chronic instability, though only when walking on uneven terrain; three patients had chronic instability even when walking on even ground.

CAUSES OF FAILURE RELATED TO SURGICAL TECHNIQUE

(1) Submalleolar impingement
This condition was described by Isbister following calcaneal fractures. Isbister used the term calcaneo-fibular abutment. The somewhat wider term submalleolar impingement is used by us to cover any impingement of the malleoli (lateral as well as medial) on the sides of the talus or the calcaneus.

This impingement has a variety of causes (calcaneal fracture; malleolar OA; etc.).

Often, the condition is iatrogenic in origin; in particular, it may occur following ankle replacement, but has also been observed after fusion.

The mechanism is a reduction in the height of the joint space by overzealous resection or because of insufficient thickness of the intermediate bearing used as part of the joint replacement.

This loss of height makes the talus and the calcaneus ride up in the ankle mortise. Loads are transferred, not through the implant, but exclusively via the contact surfaces of the lateral and medial malleoli and the matching surfaces on the talus and the calcaneus.

As submalleolar clearance is lost, tendons are squeezed between the unyielding fascia and the bony malleolus.

This syndrome accounts for the oedema formation and the severe ankle pain noted after some joint replacement operations, without any loosening of the implant. It also explains why there should be persistent pain after compression screw fusion of an ankle joint where more than 5 mm of bone has been resected.

(a) Impingement after fusion If the joint has been soundly fused, the pain is often attributed to subtalar OA, and the patient is subjected to double (ankle and subtalar) fusion.

Since subtalar fusion involves a wide opening of the submalleolar space and the release of the peroneal tendons, it often results in pain relief - which could have been obtained in the first instance, by simply effecting this release (Figs. 12, 13, 14 - the Disaster Scenario).

(b) Impingement after joint replacement The surgeon must have several thicknesses of intermediate bearing available, so as to be able correctly to restore the normal joint line level. If there is pain after joint replacement, it should not be taken for granted that this pain is being caused by implant loosening, neither should the implant be automatically removed. The simple exchange of the initial bearing for a thicker one should distract the submalleolar region sufficiently to relieve the impingement and, hence, the pain.

Because of this need for maintaining the joint line at its correct level, arthrodesis should never be performed by simply bringing together the bone surfaces left after the removal of an ankle replacement. It makes more sense to wedge iliac bone grafts into the space between the tibia and the talus, to ensure sound fusion and, at the same time, to distract the joint to a height that will prevent submalleolar problems.

It is futile to expect graftless fusion after implant removal to work. Because of the original resections required to accommodate the metal and the PE components, graftless fusion will be either impossible to achieve, or will be fraught with the risk of pain and disability secondary to submalleolar impingement.

(2) The short Achilles tendon syndrome
If, at arthroplasty, a dorsiflexion deficit or (worse) an equinus deformity are allowed to persist, walking and weight-bearing will cause major compressive forces to act on the anterior part of the implant.

Since the defect is caused by the contraction of a powerful tendon, any attempt to overcome the deficit or the deformity will drive the anterior tibia firmly onto the implant. This will result in anterior osteolysis, progressive backward tilting of the implant, and, eventually loosening.

Clinically, anterior pain is a very early feature. Pain is brought on by dorsiflexion, which produces a compression syndrome.

Patients with this syndrome will need their Achilles tendon lengthening. Preferably, this lengthening should be performed as a preventative measure, whenever 20° of dorsiflexion cannot be obtained at arthroplasty.

DISASTER SCENARIO
Chapter One:
One useless procedure

The implant does not establish clearance in the malleolar joints, and is therefore useless.

Loads are transferred through the malleoli, and pain persists (Fig. 12).

Figure 12

DISASTER SCENARIO
Chapter Two:
Two useless procedures

Since the patient is still in pain, the implant is removed. An attempt at graftless fusion makes the talus and calcaneus ride up in the mortise, making the compression of the peroneal tendons and the submalleolar syndrome worse than before. The patient is still in pain (Fig. 13).

Figure 13

DISASTER SCENARIO
Chapter Three:
Three useless procedures

Since the patient still has submalleolar pain, and since a CT scan shows some subtalar lesions, subtalar OA is diagnosed. Subtalar fusion is performed (Fig. 14).

Figure 14

DISASTER SCENARIO
Continuation and (almost) end of story

"Oddly enough", this last-mentioned procedure leaves the patient very much improved. In fact, subtalar fusion requires a generous release of the peroneal tendons, and often involves the partial resection of the lateral malleolus. If these procedures had been employed in the first instance, the patient probably would not have had a pain problem.
Fifteen years later, the patient has worsening and refractory mid-tarsal pain, and is being considered for mid-tarsal fusion, as the only possible way of managing this condition.

FUSION AFTER FAILED JOINT REPLACEMENT

This should be a straightforward operation. Since fusion by direct apposition of the tibial and talar bone surfaces is dangerous, and very often not feasible, tricortical iliac crest grafts should be wedged between the two plane surfaces. As a rule, two such grafts should be inserted.

(1) Patient positioning and the surgical approach are as described for arthroplasty.

(2) Removal of tibial component
With the Ramses prosthesis, this is straightforward, since there is no bulky peg. If the device to be removed has a large central fixation peg, removal may be more difficult.

On no account should a cortical window be made in the anterior tibia, since the removal of the large peg will have left the bone with a large internal "flare", which would be difficult enough to fuse as it is, and would be even more difficult to arthrodese if the front portion were missing.

(3) Removal of the talar component
This component presents less difficulty. Following removal, the talar dome should be trimmed horizontally, to establish a smooth, level, and well vascularized surface.

(4) Clearance of the medial and lateral malleolar joint spaces
This is done as for arthroplasty, so as to create a gap between the malleoli and the talus.

(5) Insertion of grafts
Various patterns may be encountered:-

(a) The tibial and talar surfaces are smooth and level In that case, two tricortical iliac crest grafts should be inserted.
Note: These grafts must be sufficiently long to create the necessary clearance between the malleoli and the talus.

Fixation may be by means of crossed screws, using the Méary technique. The screws can be readily routed through the grafts (Fig. 15). Equally, an external fixator may be used (Fig. 16). As a third possibility, a Crawford Adams osteotomy may be made, followed by screw fixation through the fibula and the tibia, and through the fibula and the talus (Fig. 17).

Figure 15

Figure 16

Figure 17

Providing that the joint line level has been correctly restored, this technique should protect fully against the occurrence of a submalleolar syndrome. Results have been very satisfactory.

(b) The talar surface is satisfactory, but the inside of the distal end of the tibia is "flared" As a first step, an external fixator should be applied. The fixator will be set to distract, so as to create a gap between the malleoli and the talus. A long corticocancellous graft is harvested from the outside of the iliac bone. Its base is placed on the talus, while its upper end is wedged into the flare.

The remainder of the defect is filled with autologous cancellous bone mixed with very small chips of highly porous bone substitute.

In this particular pattern, the Crawford Adams technique is less often indicated, since it tends to weaken or disrupt the flare when the proximal fixation is applied.

(c) The talar surface is satisfactory, but the anterior portion of the internal tibial flare is broken or non-united As a first step, an external fixator is applied, and set to distract. The only way in which the large defect can be filled, with any prospect of graft integration, is by the use of large amounts of cancellous bone from a bone bank.

In the presence of infection, the operation may be performed along the lines of the Papineau technique, with the wound packed open between the different stages of the procedure.

If the attempted arthrodesis fails to unite, a transplantar locking nail may be used; alternatively, talectomy may be performed and distraction applied as described by Judet, to salvage particularly difficult cases.

Female, age 48 years
Post-traumatic OA of the right ankle

(1) Preoperative a.p. view	(2) Intraoperative lateral view	(3) Following insertion of the Ramses ankle replacement
(4) Postop. a.p. view	(5) postop. lat. view	(6) At 1 year

(7) At 2 years - To date (at 6 years): satisfactory result.

CONCLUSION

To our way of thinking, ankle replacement constitutes a means of postponing fusion, in suitably cooperative patients. The causes of postoperative problems following this type of arthroplasty are now much better understood: We now know that pain after ankle joint replacement is not necessarily indicative of loosening, and should not automatically lead to the implant being removed.

Surgeons need to be aware of the short Achilles tendon syndrome and the syndrome of subfibular impingement. These conditions are often responsible for postoperative problems, and can be remedied by specific procedures, to produce a good result of arthroplasty.

REFERENCES ASCENCIO J.C. La prothèse de cheville aujourd’hui : réalité ou utopie. Maitrise orthopédique 1993 - N° 20 BARNETT C.H. NAPIER J.R. The axis of rotation at the ankle joint in man. Its influence upon the form of the talus and the mobility of the fibula. J. Anat. 1952, 86 - 1 BESNIER J.P. La cheville. Ed. Frison-Roche 1992 BIGAN N., RICHTER. Résultat à long terme du traitement des fractures de la pince malléolaire (table ronde. Ann. Orthop Ouest 1994, 95-151 BIGAN N., WEISANG C. L’arthrose après fractures bi-malléolaires. Ann. Orthop. Ouest 1970, 12, 110 - 146 BONNEL F., BLOTMAN F. Bases anatomiques et Principes Biomécaniques de la cheville, Cheville et Médecine de Rééducation. Masson éditions PARIS 1982, 1-6 BROQUIN J., EMANI A., MAURER P., TOMENO B. Arthrodèse Tibio-tarsienne ; Etude des complications et de la tolérance. Rev. Chir. Orthop. 1979, 65, 393-401 BUECHEL F., PAPPAS MJ., IORIO J. Neuw Jersey. Low contact stress total ankle replacement biomechanical rationale and review of 23 cementless cases. Foot Ankle, 8, 279-190, 1988. BUECHEL F. Total Ankle : replacement State of the Art. Disorders of the Foot and Ankle in Jahss MH . Ed. 1991 - Secon Edition Vol III, 2671-2687. CARRET J.P., SCHNEPR J. Etude Cinématique de l’articulation tibio-tarsienne, Cheville et Médecine de Rééducation. Masson éd. PARIS 1982. 7-9. CLOSE J.R. - INMAN V.R. The action of the ankle joint. Prosthetic Devices Research Project. Institute of Engineering Research. University of Berkeley. Série 11, issue 22. The Project, Berkeley, 1952. CLOSE J.R. Some applications of the functionnal anatomy of the ankle joint. J. Bone Joint Surg. 1956. 38 A 761. DANAN J.P., TOMENO B. Panarthrodèse de l’arrière-pied : technique, tolérance et résultats lointains. Rev. Chir. Prthop. 19979, 65, 433-433. DEMOTTAZ J.D. - MAZUR J.M. - THOMAS W.H. - SLEDGE C.B. - SIMON S.R. Clinical study of total ankle replacement with gait analysis. A preliminary report. J.Bone Joint Surg. 1979, 61 A, 964. DE VOGEL P.L. Enige fuctionneel-anatomische aspecten van het bovenste spronggewricht Theisis. Leiden. 1970. DINI A., BASSET F. Evaluation of the Early résult of SMITH Total Ankle replacement. Clin. Orthop. 146, 228, 1980. DUPARC J. - ALNOT J.Y. Fractures malléolaires. Classification et indications thérapeutiques. Ann Chir. 1969, 23, 853-868 DUQUENNOY A., MESTDAGH H., TILLIE B., STAHL P. Résultats Fonctionnels de l’arthrodèse tibio-tarsienne. Rev. Chir. Orthop. 1985, 71, 151, 161 FICK R. Handbuch der Anatommie und Mechanik der Gelenke unter Berücksichtigung der bewegenden. Muskeln (Part1), Gustav Fischer Verlag, Jena, 1904. HENKE J.W. Handbuch der Anatomie und Mechanik der Gelenke. C.F. Winter’sche, Leipzig, Verlagshandlung, 1863. HICKS J.H. The mechanics of the foot : the joints, J.Anat., 1953, 87-345. HUSON A. Perpectives in human-joint kinematics, biomechanics : principles and application, sélected proceedings of the 3 rd General Meeting of the European Sociéty of Biomechanics. Nijmegen, the Netherlands, 31-46, 1982. INMAN VT.T. The joints of ankle, Baltimore, Williams and Wilkins Co, 1976. ISBISTER J.F. Calcanéo-fibular abutment following crush fracture of the calcanéum. JBJS (Br), 1974, 56, 274- 278. KAPANDJI I.A. Physiologie articulaire (Fasc. 2), Paris, Maloine, 1977. KLEIN P. Implications biomécaniques de l’analyse des surfaces de contact de l’articulation tibio-tarsienne. Mémoire pour le garde de licencié en Kinésithérapie. Bruxelles, ULB, 1978. LE COEUR P. La pince bimalléolaire. Physiologie normale et pathologie du péroné. Thèse du Doctorat. Paris 1938. LEDERMANN M. - CORDEY J. Messmethodik und Standard-werte zur Untersuchung des Dynamik der Syndesmose beim Menschen. In : Funktionnelle Anatomie und Pathomechanik desSprunggelenks. Stuttgart, Georg Thieme Verlag. 1984. LIBOTTE M. - KLEIN P. - COLPAERT H. - ALAMEH M. - BLAIMONT P. - HALLEUX P. Contribution à l’étude biomécanique de la pince malléolaire. Rev. chir. orthop, 1982,. 68, 299-305. LORD G. - MAROTTE J.H. L’arthroplastie totale de cheville - Expérience sur 10 ans, à propos de 25 observations personnelles. Revue chir. orthop. 1980,66, 527. LORD G., MAROTTE J.H. Prothèse totale de cheville à propos de 12 cas. Rev. Chir. Orthop. 1973, 59, 139-151 Mac GUIRE M., KYLER R., GUSTILLO R., PREMER R. Comparative Analysis of Ankle Arthroplasty Versus Ankle Arthrodesis. Clin. Orthop. 226, 174-181, 1988. MANN R.A. Bio Mechanics of the foot. In the American Académy of orthopedic surgeons. Atlas of orthotics St Louis CV Mosby Compagny. MENDOLIA G. LAROCHE G. DONNEAUD B. Biomécanique et cahier des charges de la prothèse de cheville. Revue de Médecine et de chirurgie du Pied - 1994. MENDOLIA G. IVème journées Nimoises de pathologie locomotrice. Pr G. ASCENCIO - arthroplastie de cheville. Octobre 94. MENDOLIA G. L’arthrose tibio-tarsienne et son traitement. Journée d’enseignement du Centre Hospitalier d’ALBI le 15 mars 1995. Journée d’enseignement du Centre Hospitalier de DUNKERQUE le 31 Mai 1996. NANCREDE C.B. Some new observations upon the anatomy and fuctions of the ankle and tibio-fibular joints, Med Times, 27 mars 1880, P.316. PIDORHZ. Cahier d’enseignement de la SOFCOT. Les arthroplasties de cheville. RABISCHONG P. Anatomie Comprehensive du Pied. Le pied : Actualités en Médecine, Chirurgie et rééducation. Masson ed. PARIS, 1981, 9-1 STAUFFER R., SEGAL N. Total Ankle Arthroplasty, Four Year’s Expériance. Clin. Orthop. 160, 117-220, 1981 STAUFFER R.N. - CHAO E.Y.S. - BREWSTER R.C. Force and motion analysis of the normal, diseased ans prosthetic ankle joint. Clin. orthop. 1977, 127-189. TOMENO B., CORMIN M. Que faut-il penser de l’arthroplastie de la cheville ? Rev. Chir. Orthop. 1979, 67, 141-146. VAN LANGELAAN E.J. A kinematical analysis of the tarsal joints. An X-ray photogrammetric study. Acta orthp. scand, 1983, 54, suppl 204 VICHARD P. - PINON P. - PELTRE G. Déformation «en cupule et dôme» de l’articulation tibio-tarsienne rélélatrice d’une synostose congénitale du tarse. Rev. Chir. orthop. 1980, 66, 387. VILADOT A. Patologia del Antepie, Barcelona. Toray, 1974. VILADOT A. A propos des Arthroplasties de la cheville. Cheville et Médecine de Rééducation. Masson éd. PARIS 1982, 237-243 WAGNER J. - BOURGEOIS R. - HERMANNE A. Biomécanique du cadre tibio-péronier : rôle mécanique et physiologique du péroné. Cahiers d’enseignement de la Sofcot, N0 19, Paris, Expansion Scientifique Française, 1983.