Retrograde Stem Removal Techniques in Revision Hip Surgery

Total hip replacement has about a 50 year old past. In recent decades the number of revision hip arthroplasties in developed countries reached 15-20% of primary replacements. The average survival for both cemented and uncemented implants is about 94-95% at ten years. At the end of that period they have to be replaced. Removal of a well cemented femoral stem in revision total hip arthroplasty is a technically demanding procedure, which requires knowledge and proficiency in the usage of a multitude of surgical techniques and instruments. Experiences with traditional cement removing techniques have been published in numerous publications (Dennis et al., 1987; Ferguson, 1988; Laffargue et al., 2000; Lauer et al., 2002; Stühmer, 1987). Femoral component failure and fracture is a rarely seen complication, which presents as a difficult problem for the surgical team. Because of stress shielding in the proximal femur, osteolysis is often concentrated to the proximal part of the femur, whilst the distal portion of the stem might be stable and well fixed. This eventually leads to failure of the metal and component fracture. Although removing the proximal loose fragment is relatively straightforward, removing the distal tip is difficult, and often only feasible with techniques that result in weakening of the biomechanical properties of the femur, such as distal fenestration i.e. creating a small or larger window (in the lateral cortex or an extended trochanteric osteotomy (Wagner, 1989). The Authors use a retrograde technique in the clinical setting, in which removal of cement and stem is performed with the use of a retrograde nail passed through the knee. With this technique, removal of loose stems or broken components is both possible – when the case is appropriate for the technique. Thus further compromise can be avoided to the poor bone stock.


Introduction
Total hip replacement has about a 50 year old past.In recent decades the number of revision hip arthroplasties in developed countries reached 15-20% of primary replacements.The average survival for both cemented and uncemented implants is about 94-95% at ten years.At the end of that period they have to be replaced.Removal of a well cemented femoral stem in revision total hip arthroplasty is a technically demanding procedure, which requires knowledge and proficiency in the usage of a multitude of surgical techniques and instruments.Experiences with traditional cement removing techniques have been published in numerous publications (Dennis et al., 1987;Ferguson, 1988;Laffargue et al., 2000;Lauer et al., 2002;Stühmer, 1987).Femoral component failure and fracture is a rarely seen complication, which presents as a difficult problem for the surgical team.Because of stress shielding in the proximal femur, osteolysis is often concentrated to the proximal part of the femur, whilst the distal portion of the stem might be stable and well fixed.This eventually leads to failure of the metal and component fracture.Although removing the proximal loose fragment is relatively straightforward, removing the distal tip is difficult, and often only feasible with techniques that result in weakening of the biomechanical properties of the femur, such as distal fenestration i.e. creating a small or larger window (in the lateral cortex or an extended trochanteric osteotomy (Wagner, 1989).The Authors use a retrograde technique in the clinical setting, in which removal of cement and stem is performed with the use of a retrograde nail passed through the knee.With this technique, removal of loose stems or broken components is both possible -when the case is appropriate for the technique.Thus further compromise can be avoided to the poor bone stock.

Surgical technique
The patients were positioned supine on a radio-lucent table.Using a lateral approach, the scar tissue was excised and the hip joint was exposed.After dislocating the prosthesis, the proximal broken piece of the femoral component was easily removed.The end of the distal part became visible at approximately 4 to 6 cm below the lesser trochanter (Fig. 1.).Fig. 1.The proximal end of the distal broken piece of the femoral component -there is no sufficient gap between the broken component and the femur, therefore grasping and pulling it out with a device is not feasible (intraoperative photo).
A 2-3 cm skin incision was required through the centre of the patellar ligament as described by Moed et al. (Moed & Watson, 1995) which is generally performed for osteosynthesis.The optimal entry point for retrograde femoral nailing and defined this spot 12 mm anterior to the femoral origin of the posterior cruciate ligament in sagittal plane and at the centre of the intercondylar sulcus in coronal plane.The intramedullary canal was then opened by the awl.Under the control of a fluoroscope the 10, 12, 14 mm intramedullary nail is entered into the intramedullary canal.Reaming is not necessary, the tamp that fits best the intramedullary canal is chosen from the series.After entering the rod, it is rotated until the best alignment is obtained to the stem or the cement plug and a stable connection is achieved (Fig. 2.).To avoid deflection or perforation it is particularly important to make sure that the nail is not Fig. 2. a: After introducing, the alignment between the nail and the stem is poor, therefore the nail may slip from the tip of the stem causing perforation or fracture (photo taken from the image intensifier).b: Setting the rotation (intraoperative photo).c: After the nail is placed to the tip of the stem the femoral component is removed by gentle hammering (photo taken from the image intensifier).slipped down from the stem or the distal cement plug.When the required contact between the nail and the distal part of the stem or the cement plug is visualised by the image intensifier, the femoral implant is pushed out in proximal direction by careful hammering.Following this, the femur was reamed with a cannulated femoral reamer from the proximal end until all the granulation tissue and cement debris was removed.The intramedullary nail was pulled back during reaming without removing it from the distal femur, in order to protect the knee joint from cement particles.After thorough cleaning of the medullary canal the new femoral component was implanted.Using this method, the cortical bone is not weakened and in most cases a normal femoral stem gives sufficient stability for the revision (Fig. 3.).The perioperative protocol (thrombopropylaxis, one-shot antibiotic prophylaxis, and applying antibiotic bone cement in case of cemented refixation) does not differ from the routine procedure.Occasional knee swelling occurring in the immediate postoperative period can be treated with local ice packs and applying non-steroidal anti-inflammatory drugs.Physiotherapy and mobilisation does not alter from the routine postoperative protocol.

Discussion
When performing revision THR, removal of the distal, often inaccessible cement or a distal fragment of a broken femoral stem is a significant concern (Engh et al., 1999).Whilst intraoperative femoral fracture remains a rare but serious complication during primary THR, it is a significant problem during revisions (1% of 23,980 primary total hip arthroplasties compared with 7.8% of 6349 revisions in a study by Berry (Berry, 1999)).Subsequent studies have demonstrated similar results (Davis et al., 2003;Egan & Cesare, 1995;Mitchell et al., 2003;Sarvilinna et al., 2004;Taylor et al., 1978).Farfalli (Farfalli et al., 2007) reported a series of fifty-nine intraoperative fractures that had occurred during revision total hip arthroplasty using impaction bone-grafting.The majority of the fractures (44%) occurred during cement removal.Various new techniques have been introduced to lower the incidence of perioperative femoral fractures during cement removal, such as ultrasound (Caillouette et al., 1991;Honnart, 1996;Klapper et al., 1992;Schwaller & Elke, 2001), extra-corporal lithotripsy (Braun et al., 1992;Schmidt et al., 1998;Schreurs et al., 1991;Stranne et al., 1992;Weinstein et al., 1988), segmental cement extraction (Chin et al., 1991;Ekelund, 1992;Jingushi et al., 2000;Schurman & Maloney, 1992), application of high-energy shock waves (May et al., 1990), using a ballistically driven chiselling system (Porsch & Schmidt, 2001), or an acoustic emission-controlled milling device (Schmidt & Nordmann, 1994) or even lasers (Sherk et al., 1995).Most of these techniques had not reached widespread usage, and some authors have reported complications (Gardiner et al., 1993).There have been three clinical studies and one experimental work published about retrograde stem removal (Piatek et al., 2007;Szendroi et al., 2010;Toth et al., 2010;Tóth et al., 2011).Tóth (Toth et al., 2010) and Szendrői (Szendroi et al., 2010) reports successful, uncomplicated application of retrograde removal technique with intact stems and distal fractured components in treatment of elective hip arthroplasty revisions, while Piatek (Piatek et al., 2007) used the technique successfully in case of periprotetic fractures.The cadaver study compares the biomechanical effects of three different cement removing techniques, the distal fenestration (DF), the transfemoral approach (TFA) and the retrograde stem removal technique (RSR) using an experimental setup (Tóth et al., 2011).23 paired femora were recovered from similarly aged human cadavers and were frozen.These were later subdivided into 3 groups to provide specimens of similar age and bone quality in each group (DF, TFA, RSR).The results of the intragroup (comparison between treated and control specimens from the same cadavers from the same groups) analysis were the following: In the TFA group, the force required till fracture was significantly less than in controls (p=0.0096).Similar results were found in the DF group (p=0.068).There was no difference in the RSR group (p=0.988).Intergroup analysis showed the following: Femurs in the DF group required significantly less force to fracture than specimens in the RSR (p=0.043),whilst there was no difference in there respective controls (p=0.831).Femurs in the TFA group required highly significantly less force to fracture than specimens in the RSR group (p=0.0001),whilst there was no difference in there respective controls (p=0.178).That is, the cadaveric study supports the clinical experience that windowing the proximal femur, significantly decreases resistance against compression and rotational forces.The various windowing techniques described in the literature (Arif et al., 2004;Buehler & Walker, 1998;Cameron, 1990;Hackenbroch, 1979;Kerry et al., 1999;Klein & Rubash, 1993;Moreland et al., 1986;Nelson & Barnes, 1990;Nelson & Weber, 1980;Savvidis & Löer, 1989;Shepherd & Turnbull, 1989;Stranne et al., 1992;Tyer et al., 1987;Weber, 1981;Witt & Hackenbroch, 1976), and their modifications (Arif et al., 2004;Nelson & Barnes, 1990;Zweymüller et al., 2005) and the new instruments developed for these techniques (Brinckmann & Horst, 1985), all serve one purpose, to decrease the often inevitable weakening of the proximal femoral bone stock, preserving as much intact bone as possible.Although the more conservative windowing techniques tend to preserve more proximal bone, they still inevitably lead to decreased resistance against fracture.The studies performed by Dennis et al. (Dennis et al., 1987) on cadaveric femora, showed that when femurs without intramedullary support are stressed to failure, fractures occur through the cortical holes 90% of the time, therefore they suggested that long revision femoral stems should be used to bypass the window by at least 2 to 2.5 times the cortical diameter measured at the fenestration level (Fig. 4/b).This has been generally accepted and is widely used in orthopaedic practice (Kerry et al., 1999;Klein & Rubash, 1993;Lotke et al., 1986;Moreland et al., 1986;Nelson & Barnes, 1990;Nelson & Weber, 1980;Savvidis & Löer, 1989;Shepherd & Turbull, 1989;Tyer et al., 1987).The disadvantage of using a long stem with proximal opening of the femur instead of the RSR technique and a short stem (Fig. 3a, 3b.), is that long revision stems are more expensive, the necessary exposure requires a longer incision (Fig. 4c), with more soft tissue stripping, surgery is often much longer, blood loss can be extensive, and the local and general complication rate is higher.The RSR technique cannot always be used, and is absolutely contraindicated in the case of septic loosening or if the knee or the distal femur is affected by primary or metastatic tumours.Large distal cement plugs or the presence of oblique cement at the end of the stem are relative contraindications, because these can force the nail used for removal towards the cortical bone, which can lead to unwanted perforation or even periprosthetic fracture.
In the clinical setting depending on the quality of the proximal femoral bone, both a short a long femoral stem can be used for the revision surgery after RSR.As the cortex is not violated during component removal, a short primary stem (cemented or uncemented) is a valid option and is our preference, if the cortices are not very thin, and at least some peripheral cancellous bone is present, for cement penetration.We like to reserve long stems as a final resort, to allow a possible future revision.

Conclusion
Our cadaver experiments clearly confirm the clinical experience, that the window technique significantly weakens resistance of femur against rotation under compression forces.In contrast, with the retrograde cement removal technique this weakening does not occur.Our experience with retrograde cement removal in elective revisions and periprosthetic fractures shows that -taking into consideration indications and contraindicationsretrograde prosthesis removal technique is a viable alternative solution in case of intact as well as failed femoral components.It is associated with less intraoperative complications and faster, safer recovery compared to traditional techniques.However, longer follow-up time is required to confirm the durability of the observed clinical outcomes.

Fig. 3
Fig. 3. a: Femoral component failure 14 years following primary surgery.b: Revision with normal femoral component following retrograde removal of broken stem.c and d: Removed, broken component.

Fig. 4
Fig. 4. a: Femoral component failure and acetabular component loosening 17 years following primary surgery.b: Distal fragment of cement and stem are removed through bone window, followed by wire refixation of the window.The long revision femoral stem distally overreaches the window by the length of 2.5 times.c: Major surgical wound with extended osteotomy (intraoperative photo).