Total Hip replacement(2 stage revision)- Combined intra-pelvic and extended trochanteric osteotomy approach

Not a great starting position. Even after re-starting the suppressive antibiotics, the infection had not been controlled.

As can be seen on the image. The acetabular component and augment have moved medially and proximally. This movement was not acute and the implant had migrated over the course of 6-7 months. Prior to the cessation of antibiotics, the implant had been stable.
The radiographs reveal a cemented dual mobility acetabular component (metal cemented shell with polythylene onlay), an uncemented augment and an uncemented stem. The operation notes of the primary hospital were sought.
There are many classification systems for acetabular bone defects. I use the Paprosky classification system as classification is based around plain radiographs and staging allows some prediction of the levels of reconstruction required. I am not going to teach the classification here but I would strongly recommend the reader to read either the seminal paper by Paprosky. The summary of the classification system is excellently reviewed by Telleria. I will obviously be discussing this in the CDP.
With increasing grade there is increasing loss of supporting structures (columns, medial wall) and progressive migration.
Using the Praposky classification the defect is a 3B. Further migration is inhibited by the trochanters abutting the outer pelvis.
Acetabular defect classification and surgical reconstruction in revision arthroplasty. Paprosky et al. J. Arthroplasty 1994
Classifications in Brief: Paprosky Classification of Acetabular Bone Loss. Telleria & Lee. CORR 2013

CT scans are very useful for planning and can identify areas of bone loss, discontinuity and component positioning. However scatter often remains a problem and subtle problems can be missed.
The images above are from a manufacturer who plan and construct 3D custom implants. I hope to show more of these when I reconstruct this hip.
It was noted pre-op that dislocating the hip would be exceptionally difficult therefore before starting the surgery the length of ETO required (measured from the tip of trochanter) was calculated from templated radiographs. The distal extent finished 1-2 cm proximal to the tip of the stem (or cement tail for cemented implants).

The composite image shows the path of the external iliac artery. As can be seen in the uppermost image, the artery has been pushed medially by the inflammatory mass of the infection.
The artery continues distally coming to lie pretty close to the implant as can be noted in the report below;
“The superior gluteal artery lies a least 2 cm from the left acetabulum . The main branches of the left internal iliac artery which lie relatively close to the medial aspect of the acetabular component are the superior and inferior vesical arteries.
The left external iliac artery passes across the antero-medial aspect of the acetabular component lying within 3-4 mm of the metalwork”
The ureter was miles away at 15mm.

How to pitch this to you is really difficult. This was a long operation as you will soon see, and therefore I have decided to start a quarter of the way through the first part. There are numerous reasons for this.
Firstly, I am not vascular or general surgeon. I do not know the anatomy very well.
Secondly I do not know how to deal with any complication that could arise. For that reason a general and vascular surgeon started.
Thirdly, if I put in great detail the approach for you I am worried that someone, somewhere would ‘have a go’. Not good.
However, this is an extra-peritoneal approach.
The bowel in it’s peritoneum is at the top of the image, the external iliac vessels have been gently mobilised and a sling place around them.

The common iliac artery runs inferolaterally with psoas and divides at the level of the SI joints. The internal iliac descends into the pelvis and the external continues anteriorly and inferiorly with psoas, passing below the lateral 1/3 of the inguinal ligament to form the femoral artery.
In normal anatomy it is branches of the internal iliac that would be more closely associated with the acetabular roof. With the vessels out of the way, an inflammatory mass was seen (A). Clinically, this was over the site of the acetabular component, Palpation through the mass revealed a mobile hard structure.

An incision was made distally around the mass and it was ‘peeled’ away from the underlying structures. Again the vascular surgeon completed this part of the operation.

Once freed, the mass was sent for histology.
This revealed;
“a reactive lymph node with follicular hyperplasia and a sinus histiocytosis (I looked that one up too!). No granuloma or abscess formation. The cells within the sinuses contain metallic fragments”

A Cobb elevator was used to free the tissue off the hard mass below.
The Cobb is between what is left of the acetabular roof (essentially just hardened thickened infected fibrous tissue) and the acetabular implant.

As you can see in this image, there is evidence of green bone cement, and perhaps some metalwork.
Following lavage and removal of some of the cement with osteotomes is became apparent it was the acetabular component.

This looks like a massive step, from blood and fibrinous tissue to a shiny implant with polyethylene but it did only involve 2 sharp hits with an osteotome to remove the overlying cement and lavage to reveal this.
However I was not convinced the implant was free of pelvic attachments. Therefore tissues were released further; here anteriorly, working carefully between the pelvic inflammatory scar tissue and the implant.

And here posteriorly. There was a posibility that the metal shell could have been delivered through the abdominal approach but it soon became apparent that the shell was not completely loose. It was attached to something laterally.
At this stage we were not sure what it was. Most likely was that the metal shell was attached to the metal augment (seen on initial radiographs) which was in turn still attached to the pelvis. Therefore to preserve what bone remained we decided not to remove the implant. However we were confident that none of the important vessels were attached to metal shell and we were confident that extraction from the ‘hip side’ would not result in any traction injury.
The initial incision over the cup was closed (to be seen later!) and the wounds closed and dressed.

The patient was completely re-prepped and draped.
The scar, including the sinus, was completely excised.

The wound was deepened, completely removing the sinus tract until the fascia was seen. This plane was then developed.
Despite its size, this wound is simply a posterior approach with distal extension.

The fascia was opened, freed from underlying scar to reveal the underlying structures.
Although very scarred it is possible to see vastus lateralis (A) and the intact abductors (B).
It is also apparent that there is no obvious normal posterior structures to orientate for a ‘standard’ posterior approach.

Therefore the leg was internally rotated as much as possible (not a lot). The incision was made with diathermy along the posterior border of the greater trochanter. I use the posterior border of vastus as my starting point before moving proximally. This landmark is more often than not preserved.
I was able to palpate a soft spot most proximally which represented the apparent joint space. The incision was continued until I was in an area in keeping with the original position of piriformis before moving the incision through 90 degrees to the posterior ‘border’ of the acetabulum.

As can be seen in this image, The space was ‘tight.’ We did release a lot of scar tissue (see slide 29 for mishap!) and released G.max. However (as expected from pre-op planning) the decision was made to perform an extended trochanteric osteotomy.

To prepare for an osteotomy the posterior border of vastus has to be released and retracted anteriorly. It was difficult to get these images in this case. The slide shown here of a right hip reveals how far this posterior vastus lateralis release can be extended , though is taken from an alternate revision case.
Often a lot of vastus is stripped by the trauma of a peri-prosthetic fracture, releasing only what is necessary is prudent.
Slide 46 of this technique shows the posterior vastus lateralis incision and its forward elevation.

Gluteus maximus is released. Care is taken to find the perforators which are ligated (I use liga-clips). Vastus remains attached at its origin.
Once the femur is exposed, the bony cuts can be made. The distal extent has been calculated from the template radiographs.
The challenge with this case was getting the correct angle for the ETO. As a rule the blade of the saw should be parallel to the tibia with the knee flexed
The position on the femur is at the posterior 1/3 – 2/3 junction (the saw blade should just miss the posterior aspect of the femoral stem / body)

There are many ways to complete the far cortex osteotomy. Some have described using multiple drill holes, others use a high speed pencil tip burr. I personally use the saw but in in a very ‘haptic’ fashion. The saw is used to score the distal cortex but it doesn’t go completely through.
Having completed the AP cut the lateral cortex is cut. This can be angled or bevelled. I then recheck all the cuts to make sure they are satisfactory.
Two broad osteotomes are then placed under the lateral fragment and elevated, completing the osteotomy. Feedback is important here and if there is no ‘give’ with gentle elevation the cuts all need to be re-checked. If there are areas of posterior or lateral bony bridging, excessive pressure with elevation can shatter the trochanteric fragment.

The trochanteric fragment was elevated anteriorly. It can be seen in this slide that the lateral border of the stem can be seen (A), it should also be noted that most proximally (B) the bone has been replaced with an infected / inflammatory mass. This was removed with currettage.

Again 30 seconds of currettage and copious lavage reveals the implants in great detail. In this image you can see the stem (A), the acetabular shell (B), the augment (C) and trochanteric fragment (D).

Chronic movement had deformed the augment and screw. I was able to simply slide the augment off the deformed screw.

Medial to the augment it can be seen that there is chronic infection (the fibrinous purple jelly-like material). This all has to be excised.

The screw was removed with bone nibblers and the screw tract over-drilled.

The next step was somewhat ‘left field’. As can be seen the trochanteric fragment was large and was elevated anteriorly (A) however the stem was still encased in thickened bone, a neocortex (B). This bone was bleeding (i.e. alive) and as excessive bone removal is to be avoided the decision was made to keep this cortex intact. I am still not entirely sure that this is the correct decision (follow-up and a rapid fall in inflammatory markers will let me know).
As the stem was therefore encased in bone (even following the ETO) the hip was still tight. The amount of distraction (i.e the jump distance) to dislocate the femoral head and attached polythene onlay would have been considerably greater than the amount of distraction required just to dislocate the femoral head. Therefore I attempted to remove the polyethylene from around the femoral head with a high speed burr.
It soon became apparent however that this was rather futile and was going to take a long time.

Therefore, still with care, a rather more concerted effort was made to dislocate the hip and as can been seen, with the metal augment removed, the hip did dislocate, notably bringing the metal shell with it.

Despite the amount of bone loss and infection the stem was fixed distally. With well fixed cases it is important to debond the implant from the bone, this can be achieved with thin osteotomes (Morelands) and K-wires. However in this case my belief was that although the implant was distally fixed the interface was sufficiently weakened to allow gentle extraction using a slap hammer.

Despite the size of the slap hammer, it is imperative to remain in control and not apply excessive force. The implant was extracted with one simple blow.

This is the only image captured from an earlier challenging part of the procedure. In the case “Exeter Trauma Stem for Intracapsular Neck of Femur Fracture” I comment on tips and tricks to reduced a tight hemiarthroplasty. One of these concerned caution when releasing structures posterosuperiorly around acetabulum. I commented that caution had to be taken regarding the superior gluteal artery.
Significant bleeding is always a risk in these revision cases and here the artery bled. The proximal artery had not retracted fully into the pelvis and it was possible to get a pair of artery forceps onto the bleeding vessel. However were not initially able to get a tie around the tip of the clip. We did however manage to get liga-clips around the vessel temporarily until vascular assistance arrived. The clip is present from slide 17.
On release of the clip, no bleeding was apparent, however we used some surgicel just to be safe. Had this failed a trip to the angiography suite would have been necessary.

The next step was to remove all infectious material from the femoral canal. We used upcuting osteotomes. and increasing size ball reamers. This was continued until the canal was clear. Flexible sterile light sources (for example flexilume) are useful tools. These are narrow enough to illuminate the inside of the femur whilst still allowing visualisation.

Once the canal had been cleared attention was then turned to the acetabulum. As you can see in this image there is considerable infected inflammatory fibrinous tissue. This all has to go.
The most superficial inflammatory mass (A) represents the starting point of the acetabulum before proximal and medial migration.

It is not a rapid process with pauses and reviews of what tissues remain and reflection of what structures are at risk. In this case the tissue (A) was targeted with care made not to hit any terminal branches of the obturator vessels which have left the pelvis via the obturator canal over the obturator membrane.

Once clear, further resection was made. Here the remaining tissues in the inferior acetabulum is seen.

This is the resultant excision of tissue.
In total five separately harvested samples are sent (each with sterile forceps / blades). As mentioned before they are processed in a Class 2 cabinet. These tissue suspensions are inoculated into ‘cooked meat’ prolonged enrichment broths. 3 or more positive samples counts as positive for infection. Samples are taken at all revisions.

The base of the floor (pelvis) is seen in this image. There still remains much work. Unfortunately with such a tight space it is impossible to get photography of the tissue dissection (hands and heads get in the way). however a combination of currettes (Volkman, Charnley spoons, ring currettes) and diathermy and sharp dissection are all used. Along the anterior aspect of this deep cavity, the internal, external and femoral arteries were all at risk.

Finally, the acetabular portion is clear. If you look closely, a vertical line is seen (at the bottom of the ‘A’ annotation). This is the suture line closure of the tissues from the intrapelvic portion of the surgery, seen from the other side.

There was no way this case was suitable for any articulating spacer therefore antibiotic delivery was via cement beads.
Prior discussion at MDT meetings is paramount to discuss what antibiotics are to be used. Vancomycin and meropenim are commonly used as well at gentamycin. The antibiotics need to be heat stable as the temperature increases as the cement cures.
A ‘heavy’ but informative review paper regarding antibiotics and bone cement has been written by Bistolfi et al.
Antibiotic-Loaded Cement in Orthopedic Surgery: A Review. A Bistolfi et al. ISRN Orthop 2011.
Here the beads are more cylindrical as they are to be used for the femur.

These beads are more discoid as they are to be used for the acetabulum. The discoid shape maximises surface area allowing maximal elution of antibiotic.

Following further cleaning of the external femur, the beads are fed into the canal. It is important to have a loop made in the end of the wire to assist extraction at second stage.

I placed a puck in the base of the acetabulum. If there was any possibility of weight bearing I make the puck have a tail which hooks over the transverse acetabular ligament into the obturator foramen. (similar to a metal reinforcement ring).

This slide shows closure of the ETO. I use Charnley wires (versatile and cheap). The passer is passed back to front (less chance of catching the sciatic nerve). It is important to keep the end of the passer in contact with bone throughout this manoeuvre as you do not want to entrap any soft tissue.

The tip of the passer is just to the right of figure A. The passer has therefore travelled behind the medial femur and anterior to the ETO but beneath vastus. (B). The wire is placed into the pusher and the pusher withdrawn.

The procedure was repeated more proximally. To stop slippage these wires can be passed through the lesser trochanter (through a 2.5mm drill hole). I tend to use these for wires that pass around the greater trochanter.

The ETO is reduced and the wires tightened. It is important to pull the wires up, away from the bone during twisting. If the wire wraps around itself it will snap.

Both wires are then cut short and twisted (as shown) to lie flat against the bone under the cover of vastus.
The remaining discoid beads were inserted into the acetabular space.

Closure in cases such as these is difficult. There are no landmarks, no normal tissue.
I closed distally, picking up obviously freshly dissected tissue and suturing it to its corresponding area on the femur.

The tissues were a complete mass of scar tissue. Identifying any planes was impossible and locating the sciatic nerve not possible. Therefore to confirm that the sciatic nerve had not been picked up in the suture I reduced the diathermy setting (10) and placed the diathermy onto the suture. This needs to be used with caution and very judiciously.
The sciatic nerve is derived from the lumbosacral plexus (L4-S3). It exits the greater sciatic foramen (boundaries are sacrotuberous and sacrospinous ligaments) with piriformis, inferior to the muscle. (in 10-12% of cases the tibial and common peroneal nerves exit the pelvis as separately, the common peroneal nerve exits superior to piriformis).
I have always speculated that as the most common nerve injury after hip replacement affects the muscles predominantly supplied by the common peroneal nerve (anterior and lateral compartments of the leg), has there been injury to a missed high dividing sciatic nerve.
The sciatic nerve descends on the posterior surface of the remaining 5 short external rotators (gemellus superior, obturator internus, gemellus inferior, obturator externus and quadriceps femoris). It then passes deep to the long head of biceps before bifurcating into the tibial and common peroneal nerves at the apex of the popliteal fossa.

Following packing with beads the femur was lateralised therefore no soft tissues could be directly attached proximally. Therefore the capsule / short external rotator mass was dissected free and mobilised to allow opposition. I still struggled to find the sciatic nerve.

This allowed a complete closure of the wound.

The facial layer was then closed.

Followed by fat closure and skin.
Finally glue was applied.

The post op CT has been uploaded to ‘Materialise’, the 3D implant manufacturing company. From this they have developed the plan for the new implant. In this image you can see, on the top row, the positions of the cement ‘puck’ and cement beads. The bottom row shows snapshots of the 3D pdf image that is emailed back before implant manufacturing.