Ankle arthrodesis (fusion)- Arthroscopic assisted Ankle Fusion

One key part of the local clinical assessment pre-operatively is whether any deformity is present and as important whether it is flexible or fixed.
In the coronal plane varus and valgus are easily identified.
As important is the sagittal plane, and especially any equinus which may need correcting.
Here the plantar surface of the calcaneus is appropriately positioned to the long axis of the tibia (at right angles to it). What is not functionally positioned is the relationship of the calcaneus to the midfoot, where there is equinus present. If this is a fixed midfoot deformity and if it is ignored, then once the ankle is fused correctly the patient will not be able to heel-strike. The midfoot therefore needs correction also.

This is a typical AP Xray of an “in-line” arthritic ankle.

The lateral view is also very typical with subtle anterior subluxation of the talus within the ankle mortis, very minimal anterior joint space and a fair sized anterior cheilus.
This cheilus will probably have to be burred away as an initial stage of the arthroscopic preparation to improve access for preparing the joint and also easier assessment of the level that the tibial side of the joint is being prepared back to (congruent preparation being a key part of the technique).

This varus ankle may also potentially treated with an arthroscopic ankle fusion.
If fully passively correctible during examination then this will be straightforward.
What is evident on closer inspection of the ankle is that there is a fair amount of hypertrophy of the medial malleolus, secondary to the arthritic process. If the deformity is not correctible then the medial joint will need to be to an extent “cored out” in the medial gutter to create a space for the medial talus to reduce into.

In any ankle fusion it is absolutely key that the compensatory joints(subtalar and immediate midfoot joints) are very largely (if not entirely) clear of arthritic disease.
A plain X-ray assessment of the subtalar joint is inevitably limited and cross-sectional imaging should often be performed.
This can also provide useful insight into the exact location of any deformity present.
If the location of pain may either be ankle or subtalar then an arthrogram with a radio-opaque dye and lignocaine into whichever joint is under consideration for intervention is used. In the majority the injectate will remain in the joint injected (as here in the subtalar joint in this CT arthrogram) and inference can then be drawn about the location of pain.

For the antero-medial portal a green needle and 20ml syringe is used to inflate the joint with saline, angulating laterally and posteriorly towards the midline approximately of the ankle.It should not encounter resistance if it is on the correct line. If difficulty is encountered move superiorly or inferiorly a few mm each time.
Move the ankle again to confirm the level of the joint. If still no joy increase the traction. If this still fails it is possible that there is a medial cheilus blocking entry to the joint in which case try using an anterolateral portal initially.
Some surgeons use the anterolateral portal as their initial point of entry and this is really just a matter of preference. It sits in the line of the 3/4 th toes and at the same level horizontally.

As the needle is being withdrawn from medial to the tibialis anterior tendon a 0.5-1cm or so longitudinal incision is made just through the skin in exactly the same location and with the same orientation.

The trocar and sheath for the scope are carefully passed in the same direction of an artery clip that has been used to make a tract following the blade.

The camera is placed in the anteromedial portal, giving a lateral view of the joint.

With the camera angled slightly anteriorly, and placed from the medial portal, the anterior joint space can be assessed in lateral profile. At this early stage a clearance of the anterior joint space is made of fibrosis and synovitis and including an anterior cheilectomy if required.The cheilus will usually either be medial or lateral, though can rarely span the whole distal tibial lip.
In this case there is a fair amount of synovitis (1). If a cheilus were present it would be evident from the anterior aspect of the tibial plafond (2)

With the camera angled more posteriorly the main joint space can be seen, again in lateral profile.
Here a large area of synovitis(3) is seen deeply at the junction of fibula (1) and tibia (2).

The antero-lateral portal is sited in the line of the 3/4th toe and at the level of the joint line, indicated by the location of antero-medial portal.Before the entry point is chosen it is worthwhile routinely flexing the 4th toe which will often reveal the line of the Superficial Peroneal nerve at the ankle and distally. This is best avoided and the portal placed to avoid the nerve .

The following is simply a description of one routine that can be used in preparation of the joint. In reality once you are used to what can be achieved with the different portals you will probably move freely between the two throughout the operation.
The initial approach for arthroscopic ankle fusion is with the shaver through the antero-lateral portal and scope in the antero-medial portal, giving a lateral appearance of the ankle joint, position as shown previously.
In this position with an aggressive soft tissue shaver remove any soft tissue which is making identification of the joint margins (anterior, lateral, medial and posterior) difficult.
Ensure adequate removal of the anterior tibial cheilus if the ankle pre-op is in persistent equinus or entry into the joint is difficult.
The cheilectomy is also best performed with the scope through the antero-medial portal and a small (3.5 -4mm) spherical bony burr the antero-lateral portal. This gives a lateral view of the joint.
Use initially the soft tissue blade to define the margins of the cheilus, followed by the spherical burr to remove it through the antero-lateral portal.
A key part of the technique throughout is to move the camera easily in & out to give both a close-up bone surface view as well as a more complete view of the joint to ensure equal (or appropriate) amounts of bone are being removed from all areas being prepared. On this point the resolution of the joint surface should be clear enough to see once one is through the subchondral bone plate and into healthy subchondral bone.
The surgeon will eventually be moving frequently between visualisation through medial portal (with its lateral view of the joint) to the lateral portal (with its anterior view). This helps build up an appreciation of the 3D state of the joint during preparation.

Once the initial soft tissue and cheilus clearance has been performed the camera is swapped to the anterolateral portal and the shaver placed medially.The joint surface will now be prepared methodically using a burr, starting with the talus and then progressing to the tibia once the talus is finished. The talus is easiest prepared first with a burr to run from the posterior talus to anterior talus, commencing in the medial aspect of the joint, with the burr at 90 degrees to the long axis of the tibia as can be seen here.

The joint is assessed again and the soft tissue resector specifically used to ensure the joints outer margins are easily defined and seen before bony preparation of the joint surfaces commences.With the camera is inserted into the antero-lateral portal a more “A-P” view of the joint is seen, as shown here.
It is key to ensure that no raised ridges remain at the outer limits of the joint surfaces following the bony preparation, which may hamper joint compression in the final stages of the operation.
Often there is a mixture of bare bone and chondral cover, whereas this joint is almost entirely back to bare bone.
The choice of shaver blade for joint surface preparation will be effected by the joint surface. An Acromioniser blade, cylindrical and suited to both bone and soft tissue, is useful if there is much remaining chondral cover.

Once preparation of the joint surfaces starts attention is turned in the first instance to the medial gutter and the medial malleolus. The scope is in the antero-lateral portal and the shaver the medial portal, a small spherical burr the easiest instrument to commence this part of the preparation.
The neighbouring surfaces of the medial malleolus and talus need to be taken back to healthy subchondral bone and some space may need to be created between the medial malleolus and medial and dorsal surfaces of the talus to allow easy access for the burr.

With the scope is in the antero-lateral portal and the shaver the medial portal the neighbouring surfaces of the medial malleolus and talus are being taken back to healthy subchondral bone, often starting with a spherical burr(though here this has now progressed to a cylindrical burr).Bleeding is more readily observed from the distal tibia usually rather than the talus and if one is unsure about whether an adequate depth bone has been prepared to then turning the inflow fluid off and waiting a few moments will help encourage bleeding, if the depth of preparation is correct.
On occasion little bleeding will be observed in healthy, vascular bone , despite turning fluid in-flow off so turning on suction may be required.
The appearance of the bone surface is also of equal importance in determining whether an adequate depth has been reached, not simply the presence of bleeding. This is a matter of experience, though these pictures do give a good idea of what subchondral bone looks like.
During preparation specifically enough bone should be removed at this medial entry point to allow the cylindrical bone burr a flat trajectory into the joint(parallel to the dorsal surface of the talus) so that the talus can be dealt with easily in the next step of preparation.

A general principle in preparation during arthroscopic ankle fusion is that the margins of the articular surfaces should not be left proud, such that they could compromise joint compression. Here the tip of the medial malleolus is being prepared taken back along those lines.
The same applies to the margins of the posterior talus, lateral malleolus and lateral gutter.

Once joint surface preparation has started, to judge whether enough bone has been removed, turn the fluid inflow off which allows vascular subchondral bone to bleed and confirm enough bone has been removed.This prepared surface of the medial malleolus gives a good view of the type of picture which should be seen once the subchondral bone is exposed.
It is useful also to make a habit of viewing the joint surface debridement close up to ensure an appropriate amount of bone is removed. Next move back and get a larger picture of the entire surface just prepared to ensure the overall alignment is good. In other words confirm that all parts of the surface in question are having the same amount removed.

Preparation of the talar dome commences postero-medially and continues in a controlled fashion under direct vision antero-medially, using a cylindrical burr placed through the antero-medial portal.Note the angle of entry of the cylindrical burr which is clearly, and correctly, parallel to the joint line. This line of approach is possible having burred out some medial bone from both talus and tibia, as described.
The scope is antero-lateral and the shaver antero-medial.
Note also the difference in colour of the talar and tibial bone. The tibial bone also usually bleeds better than the talus. This may just be the effect of gravity.

Getting to the back of the postero-medial talus with the cylindrical burr.
Generally the talus is prepared first and then the tibia. If the ankle is very tight and difficult to access then it is occasionally necessary to prepare both surfaces concurrently, progressively moving from medial to lateral across the joint
The majority of talar preparation can be visualised best with the scope laterally and the burr medially, giving an A-P view.
The tibial preparation is though easiest to visualise with the scope medially.
Once experience is gained in the technique these positions are less relevant.

The most posterior part of the talus is often not accessible with the burr, as it drops away posteriorly. This can be dealt with (later) by using a large sharp hoofed curette to remove cartilage and subchondral bone from this aspect of the talus.
It should also be understood that it is not usually necessary to prepare the most posterior aspect of the talus.
Whilst the ankle is being distracted the talus is plantar-flexed, which brings its most posterior aspect beneath the tibial articular surface. Once the traction is released, and the joint positioned into neutral, the most posterior part of the talus sits posterior to the back of the tibia and won’t become part of the fusion mass. It can therefore be ignored.

The 3.5 or 5.5 mm spherical burr is used to prepare the tibia. It is important to identify the posterior and inferior rim of the tibia to make sure it is taken back appropriately and not left prominent.

For the tibial preparation the camera position has been changed and it is now in the antero-medial portal, producing more of a lateral view of the joint. The tibia is prepared systematically from posteromedial to anteromedial and then stepwise across the width of the tibia.It is useful to make a habit of debriding the joint surface close up, to ensure an appropriate amount of bone is removed. Then move back and get a larger picture to ensure the overall alignment is good and not too much is being removed.
Here the yellow line demarcates the junction between prepared and non-prepared bone.

It is much more evident that this is a lateral view of the articular surface.
The fibula (F) can be left unprepared (as can the lateral gutter), unless a case of avascular necrosis of the talus in which case both are also debrided.
Argument though can also be made for deriding the lateral gutter to aid congruent compression of the joint (and this in fact is my normal practice).
The tibia is marked TI and talus TA.

Any of the posterior talus, not accessible with the burr, is taken back to non-prominent bone using a large hoofed curette so that this part of the joint will not compromise compression during fixation.This is not a huge point however. It should be appreciated that during preparation of the ankle arthroscopically it is plantar-flexed with the distraction. This posterior aspect of the talus will therefore automatically assume a more posterior position once the ankle is placed into a neutral position prior to fixation. This will place the most posterior aspect of the talus posterior to the tibial articular surface and therefore not relevant with respect to the surfaces that are to unite.
The anterior extent of talar resection is just as far as the start of the talar neck.
Medially and laterally into the gutters the talar walls are normally prepared but need not necessarily be if reduction/compression is not going to be compromised by leaving them.
If peri-articular cysts are present they need to be opened, debrided , their bases drilled and filled with graft if particularly large.
In particular the base of cysts should not be used as a guide the depth of more general joint surface preparation as this will result in excessive bone resection .

The final routine stage of joint surface preparation is to drill deep into any areas which do not display good bleeding subchondral bone, using whatever large threaded guidewires will be used in the next step for screw placement. The guidewire will be introduced through each of the arthroscopic portals as required to access the joint surface in need of more preparation.
Stop fluid inflow at this stage to demonstrate that vascular bone has been exposed.

Before the arthroscopic set up is removed (but with the traction released) one should ensure that the ankle is placeable into a neutral position for the arthroscopic fusion.That is to say the foot should sit squarely beneath the tibia in the AP plane and perpendicular to the long axis of the tibia in the sagittal plane.
Consider removing more anterior cheilus or a triple cut of the achilles if equinus persists.
A padded sterile bowl (1) is placed beneath the lower calf and a sandbag is placed beneath the ipsilateral buttock to correct the natural tendency of most limbs to roll into external rotation.
A neutral alignment makes correct radiological alignment of screws easier.

When the guide wires are placed it is important that the position of the ankle is strictly controlled. The foot is held in neutral in both sagittal and coronal planes by the surgeon.The dimensions of the medial face of the tibia need to be appreciated. The anterior crest and the medially adjoining 1-2cm of bone are too thick to attempt screw placement and should be avoided.The posterior surface should also be identified and avoided.

The midpoint of the bone of the subcutaneous border of the tibia is the ideal point for insertion of the screws for arthroscopic ankle fusion. A large fragment titanium cannulated screw system is used and the thread of screw chosen should be the shortest available.Some guide wires supplied with sets as standard are rather thick (up to 3mm) and leave sizable stress risers in the bone if they need to be re-positioned. Guide wires can be downsized to a smaller diameters as long as the length is the same, remembering that an undersized guide wire will lead to significant swarf wedged in the cannulated drill.
Position the guide wires under image intensification. The aim of the first guide wire for arthroscopic ankle fusion is that it should run from medial tibia to lateral talus on the AP view and sit in the mid talus on the lateral view.
Initially drill the guide wire perpendicular to the long axis of the tibial shaft to gain purchase, then angulate the wire into the desired direction. Advance initially under X-Ray control.

The aim of the first guide wire for arthroscopic ankle fusion is that it should run from medial tibia to lateral talus on the AP view and sit in the mid talus on the lateral view.A lateral projection of the ankle during arthroscopic ankle fusion showing the relationship of the Fibula to the talus. Generally it sits slightly posterior to the midline of the talus . A guide wire aimed from medial tibial cortex just anterior to the fibula should sit approximately in the mid-talus on saggital projection.
The tip of the Fibula is an easily palpated landmark and can therefore be usefully used as a point to triangulate the tip of the guide wire towards, without an immediate lateral X-ray image.

It is very important not to use overlong screws which impinge upon the subtalar joints articular surfaces.
A direct lateral X-ray (as seen here) may not give an accurate enough idea of the screw lengths with respect to the subtalar joint. Because of the joint geometry of the inferior aspect of the talus the proximity of the wires to the joint is not adequately shown with this view.

The truest view of the guide wire lengths is given by an oblique lateral image intensifier image as shown here and this should be done in every case before deciding on the appropriate size of implant, as well as a true lateral image.Perhaps surprisingly, this is the oblique lateral view of the previous direct lateral image.
The subtalar joint should be seen to be free of guide wire penetration.
In deciding upon the screw length to use one should always select shorter than the measurement, to allow for the effect of compression of the joint space. If this is not allowed for then overlong screws are likely to be selected which will have a greater chance of breaching the subtalar joint (and predisposing to iatrogenic joint damage, though this also depends upon the location of the breach and will be less likely if it is within the area of the sinus tarsi).

The placement of the leg on a padded bowl during arthroscopic ankle fusion means that it is easier to use the pistol grip wire driver which otherwise would be abutting on the table and compromising guide wire placement.

The Ideal initial AP screw placement for arthroscopic ankle fusion is into the lateral third of the talus. It should remain fully contained within the talar bone. If any question about this exists then the AP view can be rotated to a true mortice. The starting point on the medial tibial cortex should be approximately a hands breadth above the tip of the medial malleolus. This will be the longer of the two screws.

The second guide wire for arthroscopic ankle fusion is placed parallel to the first but is for a shorter screw.
Its entry point on the medial aspect of the Tibia will be lower and one is aiming for it to in the region of the angle between medial malleolus and tibial articular surface. The wires should be placed far enough apart that their exists enough space for both the threads and screw heads of each screw.

To increase the area of purchase in the talus the second wire should be angulated to diverge from the first in the sagittal plane, as shown here. This second guide wire runs to the medial aspect of the talus in the coronal plane.The second guide wire (2) is placed parallel to the first (1), which is superior and runs further lateral in the A-P plane.
The guide wires are inserted under Image Intensification. The Intensifier needs to be sterile draped and good images of the AP and lateral appearance of the ankle need to be obtained before proceeding with guide with guide wire insertion.

The second screw need not be precisely parallel in the lateral projection, as this allows more purchase upon the talus for arthroscopic ankle fusion.

It is a good idea to under-drill the depth of each screw hole. Using II this is straightforward to judge. This gives still some unprepared talar bone most inferiorly which is likely to give better purchase to a screw than if the whole screw track is over-drilled bone.In tough adult bone it may be necessary to cut the initial few threads of the tibial screw channel with a large cortical tap.
If the bone is soft it may be that the 6.5mm screw will not provide sound enough fixation. Further drilling is generally not required, simply up-size to an 8.0mm diameter screw.

The screw thread chosen for arthroscopic ankle fusion should normally be the shortest available. Care should be taken to avoid a thread length which crosses the joint line, limiting compression.In a normally aligned ankle place the longer & lateral screw first, the medial screw second.It should be checked on both AP and Lateral and oblique lateral projections in determining an appropriate length has been chosen and in particular that the subtalar joint has not been breached.
If fixation with one screw is poor then consider using a longer thread , a wider diameter screw or a washer if the cortex is poor . An alternative is to add additional screws from a lateral , antero-lateral or antero-medial approach . Either arthroscopic portal can be expanded for these latter two.

The appearance on table at the time of an arthroscopic fusion can initially be concerning with fluroscopy suggesting gaps of a magnitude that could not be tolerated with standard open surgery.
As important is that the grip of the screws and compression is excellent.
This is however by no means the most common on-table appearance, which would be one showing good bone to bone contact.
If there is concern that the joint may not be compressing one of the portals can be extended and the joint directly inspected.

A typical appearance of the portals and approach immediately post-operatively.
Closure is with non-absorbable sutures.

The immediate post operative X-rays after a typical arthroscopic ankle fusion will often show small radiographic gaps (as here). This appearance is not concerning (as it would be after an open fusion). The gaps rapidly fill in with bone as union occurs.

The X-Rays at 6 weeks after the arthroscopic ankle fusion show that bone has formed almost completely across the arthrodesis site.
This is rapid even in the context of an arthroscopic fusion.
The post-operative protocol is not deviated from never the less and weight-bearing continues to be restricted and graduated.

The appearance at 3 months post-surgery of the soft tissue envelope.

From 10-15 years post ankle fusion there should be an expectation of developing arthritis in the subtalar and/or Choparts joints, as shown here. It is not however probable that these joints will require subsequent surgical management.

Avascular necrosis of the talus is one relative indication for fusion of the ankle. Given the lack of further de-vascularisation when performed arthroscopically this is certainly worth considering as the technique of choice for such cases if an isolated ankle fusion is required.
Given the position of the talus spanning both ankle and subtalar joints either or both joints may be effected and cross sectional imaging is worth performing in all cases. An MRI in particular is also of use to identify the extent of avascular change, which is often incomplete with remaining patches of vascular bone.
What is also notable in both these images is how well preserved the joint space is despite significant avascular changes.
An initial arthroscopic view of the joint may well suggest healthy condral surfaces, with no obvious defects. This will easily peel away once joint surface preparation is commenced.

The initial appearance of a joint with an avascular talus may be very normal upon arthroscopic inspection to the extent that chondral surfaces may appear uneffected.
Once preparation begins however the cartilage will peel off with little effort. Avascular bone itself is both visibly and tangibly different from normal cancellous bone.
Arthroscopic preparation of the talus for fusion involves performing a generous resection of the avascular bone. The pre-operative scans should be studied and understood especially to guide the surgeon to areas of remaining blood supply, which should be well opened-up during joint preparation. In most cases the tibia will not be effected similarly.
There may be large islands of loose, unattached bone within the talar dome which will need to be removed. The surfaces are then thoroughly and multiply drilled with the heavy guide wires. Iliac crest bone graft is added is most cases and in particular to where bony defects exist.
Post operatively weight bearing is delayed until 6 weeks post operatively at least and progressed with more slowly than normal.

A multi-cystic varient of osteoarthritis of the ankle exists that requires a few nuances in technique if being dealt with by arthroscopic fusion.
The extent and depth of mutli-cystic change is hinted at by the plain X-ray appearance but a CT is useful pre-operatively to show the true extent and location of the bony defects.
When deep cysts are present it is not practical to resect the bone back to the level of the bases of the cysts. In doing so the congruencies of the opposing joint surfaces will be compromised, in particular because of malleolar impingement preventing compression of the relatively over-resected joint surfaces.
I recommend that the cysts are simply de-roofed and contents curetted out. The sclerotic cyst bases are dealt with by drilling into them under direct vision with the guide wires on the screw set and/or broken-up by hammering into them using a fine Lambotts osteotome. Large cysts are packed with iliac crest bone graft. This is done semi-openly by extending one of the anterior arthroscopic portals for limited open access to the joint.
Fixation is then in a standard fashion.