Tibial fracture non-union correction using Taylor Spatial Frame (Smith and Nephew)

This gentleman had an anterolateral plating of a distal tibia fracture with delayed post-operative wound healing and persistent pain with limited mobility. He had a history of significant venous disease, smoking and obesity. Given this background, plating with a small fragment plate, was perhaps optimistic and he may have been better served with a nail at the index operation- a load sharing device inserted with a closed technique.
The lateral image confirms a non-union with broken plate and an apex anterior deformity of the tibia and fibula. There is callus formation and therefore this is likely to be a hypertrophic non-union.

The AP image confirms almost 100% translation of the distal fragment with varus and shortening.
Clinically the soft tissue envelope was intact but with changes of chronic venous disease. There was clear mobility at the fracture site confirming non-union.

Nuclear imaging was not suggestive of infection, however the history of wound problems was a cause for concern and therefore I planned a staged reconstruction.
In terms of our 4S assessment of this non-union we have to address the following areas;
Straighten (align) the tibia,
Stabilise and possibly Sterilise (ie eradicate any infection).
The first stage would involve debridement of the non-union and sampling. Given the magnitude of the deformity and poor soft tissue envelope I planned to perform a gradual deformity correction in a ring fixator. The TSF is ideal for this type of complex, multi-plane deformity.
The second stage, assuming microbiology cultures were negative and after the deformity was corrected, would involve exchange of the frame for an intramedullary nail. This allows a significant reduction in the time the patient has to spend in the frame.

The patient is prepped to above the knee to allow rotation to be assessed intra-operatively.The deformity and chronic venous changes to the limb can be seen.

The old incision is opened to access the metalwork and fracture site.As far as possible dissection remains in the scar tissue from the previous surgery.

The bone is relatively superficial beneath the incision but care is taken to lift full thickness skin flaps and avoid undermining the skin and compromising wound healing.Once the tibia is exposed the fibrous tissue over the plate is mobilised with a combination of sharp dissection and gentle persuasion with an osteotome. At this point dissection and elevation of the scar tissue remains on the bone or plate, this ensures we are safely away from any neurovascular structures.

The plate (1.) is partially covered in bone which is elevated with an osteotome.
The osteotome is worked along the superior edge of the plate, lifting the overlying bone off the plate.

The whole of the proximal part of the plate has now been exposed and the screws are removed.
Screws are only removed once the whole screw head is visible and all the scar tissue cleared from inside the head, this ensures the screwdriver will sit properly in the head and avoids ’rounding off’ the screw head. In this case the screws were intact and removed easily but it is sensible to anticipate problems and have a broken screw set, such as the Synthes Operace system, available- the broken screw set includes corers to remove broken screw fragments and reverse cutting taps to extricate any screws where the head has rounded off and they cannot be removed with the screwdriver.

A pair of forceps (1.) is useful to help remove screws which lack hold in the bone and won’t back out.
An assistant applies gentle pressure to the shaft of the screw as it is unwound, this gives the screw thread something to grip onto and allows the screw to be removed.

The non-union site and distal end of the plate are now exposed.In case the non-union is easy to identify due to the significant translation and mobility.
Often it is not as easy to locate the non-union, particularly hypertrophic ones where there may be abundant callus, and it is important to have a clear mental image based on studying the pre-operative imaging of the patho-anatomy.
I often use intra-operative imaging and k-wires to mark and confirm the site of the non-union

As dissection proceeds distally care must be taken for the neurovascular structures crossing from the anterior and lateral compartments toward the midline.Given that this is revision surgery assumptions as to where important structures may lie are not valid. There is a risk of encountering both the superficial peroneal nerve in the superficial part of the wound and the anterior tibial vessels and deep peroneal nerve in the deeper layers.
Constant surgical vigilance is the only way to avoid accidental injury.

A periosteal elevator is used to expose the most distal part of the plate which sits just above the ankle joint.
As before, the elevator stays on the plate/bone to avoid damage to the structures anterior to the distal tibia (anterior neurovascular bundle, tendons of long extensors) and gently lifts the scar tissue from the plate.

Once all the plate is exposed the screws are removed.Once all the screws have been removed the plate fragments can be lifted from the bone, this is often best achieved with an osteotome inserted at the edge of the plate and the bone to free the fibrous soft tissue attachments.
If the plate is not freeing up as anticipated it is worth checking with an X-ray that all the screws have been removed.

Once the bed of the plate is exposed and samples are taken.It is important that no dirty (ie previously used) instruments, swabs or suction are placed in the field being sampled.
Samples of bone and fibrous tissue from the bed of the plate and the non-union are taken using rongeurs and scalpels.
Samples are taken according to protocol, 5 seperate samples, each taken with fresh, unused instruments and placed in a separate container are sent for microbiology and two samples for histology.

Samples are placed directly into specimen pots which are then passed out to the theatre team to be labelled, they should be sent to the lab as soon as possible.
Once sampling is completed the anaesthetist gives antibiotics, our protocol is to use vancomycin and meropenem for suspected bone infection cases, this combination covers over 98% of organisms identified in Fracture Related Infections in our unit.

The proximal end of the tibia is now mobilised prior to preparation.Fibrous tissue in the non-union site is carefully removed with a nibbler, this is relatively safe as the posterior neurovascular structures lie behind the tibia separated by tibialis posterior. However tissue should only be removed under direct vision. The fibrous tissue is removed until the bone end is completely exposed.

The bone ends will be ‘squared off’ to facilitate deformity correction and compression of the non-union site.
This can be done with a drill and osteotomes but here we are using a sagittal saw. Whichever technique is used the key is to avoid heat from the instrument and thermal necrosis of the bone ends. The saw blade is continuously cooled with saline (1.).

The Image intensifier is used to confirm that the tibial debridement cut will be perpendicular to the anatomical axis of the tibia, in both planes.Note that 2 screws remain in the distal segment- it was not possible to access these before mobilising the non-union site and they were therefore removed after the non-union had been derided.

The saw is used only with the blade under direct vision and with regular pauses to ensure no heat is generated.

The tibial cuts are completed with an osteotome to avoid the saw blade plunging into the soft tissues on the far side of the bone.The posterior neurovascular structures lie behind the tibia, deep to tibialis posterior. The osteotome is gently tapped with a mallet to complete the saw cut.

The loose fragment can then be removed with the nibblers.
There may be soft tissue attachments on the far side of the tibia, these are gently elevated with a small periosteal elevator.

The tibial bone ends are inspected to ensure it is viable.Viability is assessed by the capacity to bleed- punctate bleeding can be seen on the cut cortical surfaces, the ‘paprika sign’.
Areas with doubtful viability are carefully debrided to bleeding bone with a bone nibbler by gently nibbling back the bone edge until bleeding is seen.

The process is repeated at the distal segment.
This is exactly the same methodical process with a saw cut perpendicular to the long axis of the tibia, completed with an osteotome and then the bone ends checked for viability.
If there is no bleeding and the bone does not appear viable then more should be excised until only viable bone ends are left- this is vital for the success of the operation.

The bone ends have now been mobilised and prepared and some correction of the deformity is possible.
There is scarring across the medullary canal, which is common in this situation, however the cortices are seen to be bleeding and the bone ends are therefore viable with the capacity to heal the fracture.
There will inevitably be some limb shortening. Several centimetres of acute shortening can be tolerated without compromising the limb vascularity, as a rule of thumb 15% of the segment length or up to 5cm, however acute limb shortening must always be performed with careful vascular monitoring of the limb- for larger shortenings I perform on table doppler studies to confirm good distal perfusion. The shortening was discussed pre-operatively with the patient with an agreement that lengthening of the limb would only be considered if shortening was excessive. This would generally be if the limb was shortened by more than 2 cm.
In this case bifocal treatment in a ring fixator would be the preferred option with compression of the non-union site distally and restoration of leg length by distraction osteogenesis through a proximal corticotomy.

The next stage is to perform a fibula osteotomy.This will allow completion of the deformity correction and some compression at the non-union site. otherwise the intact fibula acts to splint the two ends of the non-union apart.
The fibula osteotomy is placed close to the tibial non-union to facilitate correction.
When selecting the site attention should be paid to neurovascular structures particularly the superficial peroneal nerve which may cross the fibula and also to avoid being too close to the syndesmosis which would destabilise the ankle.
A longtitudinal incision is made over the fibula, the fascia is then incised in line with the skin incision and the peroneii mobilised with dissecting scissors. The perisosteum over the site of the osteotomy is elevated.

The exposed fibula is then cut with a small sagittal saw.The soft tissues are retracted with Homan retractors placed around the fibula and the saw blade cooled with saline irrigation.

In this instance a segment of fibula will be removed owing to the shortening in the tibia.
Two oblique cuts are made in the fibula and then completed with an osteotome. The peroneal artery lies medial to the fibula and is at risk if the saw or ostetome plunge beyond the fibula. In some cases it is possible, and safer, to make the cuts from anterior to posterior rather than from lateral to medial.
Oblique cuts are used to allow the 2 ends of the fibula to slide past each other during deformity correction/compression.

The fibula segment is removed carefully after gently removing soft tissues with an elevator.
The excised fragment (1.) and gap in the fibula (2.) can be seen.

The final step is to ‘petal’ the ends of the tibia.A small, sharp, osteotome is used to make a series of cuts in the tibial ends to encourage bleeding.

The wound is then thoroughly irrigated with saline.Six litres of saline via a large bore giving set is used to irrigate the wound.

Wounds are closed in a layered fashion, I use an interrupted subdermal suture with vicryl to oppose the skin edges and prefer interrupted sutures to the skin using either mattress or the Allgower-Donati technique to preserve the blood supply to the skin edges. Following wound closure the leg is reprepped and draped and new gowns/gloves donned.

The frame is now planned with image intensifier.I will be using a Taylor Spatial frame to allow hexapod correction of the deformity – the hexapod is an engineering concept, the use of 6 adjustable struts allows deformity correction in any plane.
A 3 ring construct will be used to allow the whole tibia to be spanned ,given the size of the limb and force through the frame during correction.
“Plan A” is to use the frame to correct the residual deformity gradually over a few weeks and then convert to an intramedullary nail to reduce the patients time in frame- thus the frame is planned to allow nailing of the tibia with the frame on the leg.

The distal ring position is marked on the skin first.This will be parallel with the ankle joint (essentially perpendicular to the anatomical axis) and sufficiently proximal to avoid wires within the joint itself (the capsule extends approximately 10mm proximal to the joint line anteriorly and up to 19mm in the tibio-fibular recess). Intra-articular wires risk seeding infection into the joint.

The position of the middle ring is then marked, which will be on the other side of the non-union approximately 150mm from the distal ringThis position will allow medium size TSF struts to be used for the correction (and thus plenty of scope for changes in strut length during the correction).
The third and most proximal ring will be at the level of the proximal tibia giving a 2 ring construct over the long proximal segment to give better control of the forces in the frame during correction.

The most proximal ring, just below the knee joint, will be a 2/3 ring, open posteriorly to allow a good range of motion at the knee.

The proximal ring is connected to a full ring of the same diameter (this will be the middle ring of the final construct).The 2 rings are connected with threaded rods. Traditionally 4 rods are used although here I am using 3 which have equivalent stability, these are evenly spaced around the ring and positioned so that they will on the more distal full ring they will not interfere with the struts that will be used on the distal segment (these will be mounted on the other side of this ring). Once built the proximal 2 ring block is slid over the foot onto the leg.

Place a 1.8mm wire in the proximal tibia from lateral to medial.This is the proximal reference wire, accurate placement is important to ensure the frame is correctly aligned with the limb.
An assistant holds the leg ensuring the patella (1.) is facing upwards.
The wire is inserted under II control, parallel to the floor.
My right hand is holding the wire in a chlorhexidine soaked swab (which keeps the wire cool).
The wire is held in the driver and pushed through the skin and onto bone. It is then passed across the proximal tibia with intermittent pulses with the wire driver which avoids overheating the wire which would result in thermal necrosis of the bone surrounding the wire and thereby loosening of the wire.
Once the wire is through the far side of the bone it is tapped with a mallet through the skin, this prevents any nearby neurovascular structures being caught in a spinning wire.

The image intensifier is used to ensure that the wire is place perpendicular to the anatomical axis of the tibia.Note that this is not parallel with the knee joint which normally has a slight varus alignment.
This wire thus passes across the middle of the tibia, is orthogonal to the anatomic (and mechanical) axis and is parallel to the floor with the leg on the table and the patella facing the roof. The wire is proximal and anterior to the common peroneal nerve which should not be at risk provided the wire is placed from lateral to medial.

The rings are now connected to the wire.The ring is placed so that the anterior tab (1.) is aligned with the patella (2.) which ensures good rotational alignment and that the open part of the ring is sitting posteriorly to allow knee flexion.
It is important to ensure good soft tissue clearance around the ring, normally this means at least 1 fingers breadth of clearance anteriorly. In this case we have deliberately sited the ring a little closer as we are planning to perform an intramedullary nailing once the deformity its corrected (approximately 2 weeks post op) and siting the ring closer to the skin allows room for the tibial nail insertion jig to pass in front of the frame without impinging.
Once the ring is correctly aligned it is mounted on the wire with fixation bolts and one end of the wire is tightened in the fixation bolt with a nut using 10mm spanners (3.).

The other end of the wire is now tightened with the wire tensioner (1.).First the wire is positioned with an untightened fixation bolt. Next the wire tensioner is placed over the wire and onto the ring (2.). An assistant stabilises the rings.
My right hand is steadying the tensioner while my left hand turns the tensioning device clockwise- this allows the tensioner to grasp and tighten the wire.
Here I have tightened to 130Nm (adult patient with healthy bone)although less tension may be used in other anatomical areas such as the foot.

This image is from a subsequent step of the operation but demonstrates very well how the wire should sit within the tensioner and also how the tensioner needs to engage with the frame, before tension is applied.

Once the wire is tensioned the fixation bolt is tightened to secure the wire at the chosen tension..The tensioner is then removed by turning the handle anticlockwise until the wire is released and it can be removed.

X-rays are taken in the coronal plane to ensure correctly aligned the rings.The threaded rod is parallel to to the anatomical axis of the tibia confirming correct alignment.
Completely accurate alignment is not essential when using a hexapod frame as the software can account a discrepancy in the alignment of the frame to the limb, however well aligned frames tend to be more stable.
If the frame alignment in the coronal plane does need to be altered this can be done by placing washers between the reference wire and the ring on one side to make small adjustments to varus/valgus, larger adjustments may need the reference wire to be repositioned.

A wire is now inserted into the distal ring of the proximal block in the transverse plane.This is again placed from lateral to medial and pushed through the lateral soft tissues onto bone to avoid catching any neurovascular structures (tibilias anterior and deep peroneal nerve within the anterior compartment in this case).
This allows fine tuning of the sagittal alignment of the rings by adjusting the position that the wire is attached to the ring.
The position is checked on a lateral xray, the threaded rods should be aligned with the posterior cortex of the tibia.

Once satisfied with the coronal and sagittal alignment of the rings further fixation is added to the proximal segment.This follows Ilizarov principles- we want at least 2 fixation points on each ring and the wires to cross as close to 90 degrees as possible for stability.
On the proximal ring I have added 2 further wires, one passes through the fibula head (ensuring it is in the anterior half to avoid injury to the peroneal nerve as it winds around the fibula neck- the entry point can be screened with x-ray if not readily palpable) from posterolateral to antero-medial.
A further wire is passed from the medial side in a postero-medial to anterolateral direction.
All wires are planned to pass within safe corridors to avoid neurovascular injury. The cross sections describing the safe corridors are outlined in detail in Catagni’s Atlas for the insertion of transosseous wires and half pins (see below).

On the distal ring in this segment I have added one further wire, the anatomical constraints mean it is not possible to achieve a 90 degree crossing angle of the wires without risking injury to a named nerve or vessel.
Here the wire is placed from antero-lateral to poster-medial beneath the medial face of the tibia trying to avoid capturing muscle in the anterior or posterior compartments.
Catching muscle in the wire will cause fretting on the wire every time the ankle moves and will be a persistent source of pin inflammation and pain.

First we place the distal reference wire for the distal segment ring.This segment is too short to be able to position 2 rings so I have used a single ring above the ankle (as above).
An assistant holds the foot with the 2nd toe pointing directly upwards and aligned with the patella to ensure correct rotation, and the wire is placed from lateral to distal under X-ray control.
The wire must be at least 10mm above the ankle joint to stay outside of the capsule. It is placed parallel with the floor and parallel with the ankle joint (and thus perpendicular to the anatomical axis), the entry point is just anterior to the fibula so only the tibia is captured by the wire.
As before it is pushed onto bone and then intermittently drilled through the bone while held with a chlorhexidine soaked gauze to reduce heat generation.

The wire is then tapped through the far side with a mallet to avoid catching any neurovascular structures, this is made easier by holding the end of the wire as shown, keeping the wire as straight as possible so that when it is hit with the mallet it advances through the bone.

The ring is mounted and attached/tensioned as before.An AP xray shows it is correctly aligned with the ankle (ie parallel with the ankle joint).

The sagittal alignment can now be adjusted on the lateral xray by rotating the ring around the wire. I am trying to align it to be perpendicular to the anatomical axis rather than the ankle joint (which faces anteriorly).
Once the alignment is corrected the ring can be temporarily held with a cut 1.8mm Ilizarov wire placed anteriorly in the tibia- capturing 2 cortices, but not passing through the posterior soft tissues, and fixed to the ring with a fixation bolt.

With the ring alignment controlled with our temporary wire (1.) further fixation is now added to the ring. As we are only using one ring we aim for a minimum of three wires.
In addition to the transverse reference wire I have passed one from posterolateral to anteromedial, through the fibula and tibia, and one from posteromedial to antero-lateral, again respecting the safe corridors for wire placement to avoid neurovascular injury.
I have also added a ‘drop wire’. This is a wire deliberately placed away from the ring, proximally, in a transverse plane and mounted to the ring with posts which adds further stability in this short segment.
If we could not gain sufficient stability in this segment a further option would be to extend the frame onto the foot- this is a subjective judgement based on the size of the segment (poorer control in very short distal tibial segments), the bone quality and patient size etc. One advantage of extending onto the foot, for instance in pilon fractures, is that it allows immediate weight bearing. The main disadvantage is that wires in the foot are often poorly tolerated.

The proximal and distal segments are now connected with TSF struts.Six struts are used at designated positions on the rings to create the ‘hexapod’. This configuration then allows multi-plane deformity correction around a ‘virtual hinge’ as determined by the computer software.

The struts have several features to allow them to be adjusted easily.
They are attached to the ring at fixed points with ‘shoulder bolts’ (1.).
A lock can be used to prevent inadvertent changes in strut length by locking the adjustment mechanism (2.).
The sleeve can be unlocked to allow rapid changes in strut length intra-operatively (3.).
Gradual adjustments are made using the winding mechanism (4.), the length of the strut is read from the printed scale (5.).

Once the struts are in place the deformity is partially corrected by adjusting the position of the rings.An assistant stabilises the proximal block (1.) and the distal ring (2.) can then be manipulated to perform an ‘acute’ reduction.
The struts are then locked in position by closing the rapid adjustment sleeves (3.). At this point the strut lengths are determined entirely by the position we have put them in on table. Post-operatively a TSF software programme will be used to further adjust the struts and complete the correction.

Final lateral X-ray shows some residual translation of the distal segment.
This will be gradually adjusted using the TSF software post-operatively.

Non-adhesive dressings are used and the leg is dressed with gauze/wool/crepe.The pin sites themselves are dressed with blue gauze soaked in chlorhexidine held in place with clips (1.)

Xray post-operatively showing correction of sagittal alignment.

AP xray with restoration of alignment.