Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Review
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty: a review Expert Rev. Med. Devices 12(1), 61–72 (2015)
Eddie S Wu, Jeffrey J Cherian, Bhaveen H Kapadia, Samik Banerjee, Julio J Jauregui and Michael A Mont* Rubin Institute for Advanced Orthopedics, Center for Joint Preservation and Replacement, Sinai Hospital of Baltimore, 2401 West Belvedere Avenue, Baltimore, MD 21215, USA *Author for correspondence: Tel.: +1 410 601 8500 Fax: +1 410 601 8501 [email protected]; [email protected]
As the number of primary total hip arthroplasties increase over the next several decades so will the incidence of periprosthetic fractures around the femoral stem. Treatment can reliably be predicted using the Vancouver classification with internal fixation being indicated in fractures involving a stable implant and revision arthroplasty indicated in those with unstable prostheses. Non-displaced fractures involving the greater and lesser trochanter can generally be treated non-operatively. Extensively porous-coated stems and the use of modular uncemented revision stems to treat Vancouver B fractures have shown encouraging results. The treatment of Vancouver C periprosthetic fractures continues to follow basic AO fixation principles with an emphasis on eliminating stress risers with adequate implant overlap and length. This review will focus on the risk factors and classification of these fractures, as well as highlight the treatment options for post-operative periprosthetic femoral fractures around a total hip arthroplasty. KEYWORDS: hip • periprosthetic fracture • revision • total hip arthroplasty • Vancouver classification
Each year, more than 380,000 total hip arthroplasties are performed in the USA, with estimates expected to increase to over 1 million procedures yearly by 2030 [1,2]. As the number of primary total hip arthroplasties continues to increase, so will the number of complications and revision procedures due to periprosthetic fractures around the femoral stem. The first reports of periprosthetic femur fractures around total hip arthroplasty were described as early as 1954 [3]. Currently, it is estimated that the rate of revision total hip arthroplasty is expected to double by the year 2026, with periprosthetic fracture being an increasing cause [4]. A recent study reported that there was a 1% incidence of periprosthetic fracture in primary total hip arthroplasty, while in the revision setting the incidence rose to 4% [5]. The etiology of a post-operative periprosthetic femur fracture is multifactorial and may be influenced by the type of prosthesis used and whether a revision arthroplasty was involved. Post-operative periprosthetic fractures around the femoral stem usually occur as a
informahealthcare.com
10.1586/17434440.2015.958076
result of low energy trauma and/or undiagnosed osteolytic defects [6]. Patients who receive ingrowth prostheses may have a higher tendency to fracture within the first 6 months after implantation from iatrogenic stress risers generated from prior preparation of the canal [7]. However, patients who receive cemented stems at the time of the index hip arthroplasty have been found to fracture more commonly at the distal stem tip several years after implantation [7]. The etiology of this is likely multifactorial, but may be a result of osteolysis from the generation of wear debris. As the number of periprosthetic fractures increase, the treatment of this complication will become an increasingly important topic for discussion due to the substantial economic burden placed on the healthcare system. In a Medicare database study, Ong et al. reported that revision total hip arthroplasty consumed an average of 18.8% (US$238 million) of the entire economic burden of total hip arthroplasty [8]. With increasing procedural costs and decreasing reimbursement in total hip revision
Ó 2015 Informa UK Ltd
ISSN 1743-4440
61
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Review
Wu, Cherian, Kapadia, Banerjee, Jauregui & Mont
procedure, the economic burden in treating these complex cases will continue to grow. Therefore, identifying risk factors and establishing optimal treatment algorithms for periprosthetic fractures is imperative to minimize complications while maximizing outcomes. In this review, the authors aim to analyze the recent evidence on: risk factors, fracture classification schemes and treatment strategies that have been used for periprosthetic fractures around a total hip arthroplasty. This report aims to provide orthopedic surgeons with a concise overview of the available options and currently preferred treatment strategies. Methods
A review of the literature was performed using four electronic databases of EMBASE, CINAHL-plus, PubMed and SCOPUS, which retrieved articles between January 1976 and September 2013 by one author (EW). Periprosthetic femur fractures can occur either intra-operatively or post-operatively, with the focus of this report on fractures of the femur that occur post-operatively. Preference was given to randomized controlled trials, meta-analyses and data from national registries over the past 5 years; however, all relevant studies were included in the analysis. Search terms that were utilized consisted of (hip[title] or femur[title]) and (arthroplasty[title] or replacement[title] or periprosthetic[title]) and (fracture[title]) and single keyword search (‘Vancouver A’, ‘Vancouver B’, ‘Vancouver C’). The literature search resulted in 941 reports that addressed periprosthetic fractures around total joint arthroplasty, of which 58 specifically focused on treatment strategies for fractures around total hip arthroplasty. Risk factors
Several risk factors are described in the literature for periprosthetic femoral fracture, however, most of them are likely interrelated. Age, gender, revision surgery, implant type, medical comorbidities and osteolysis have all been reported to potentially increase the risk [7,9–11]. In general, most patients who undergo hip arthroplasty are well into their adult years, and therefore, the selection bias of studies analyzing age and periprosthetic fractures must be interpreted with caution as many cohorts involve elderly patients. Lindahl et al. reported that the risk ratio for fracture was 1.01 for every additional age year [12]. Increasing age is inevitably associated with confounding medical comorbidities such as osteopenia and the increased fall risk that may likely contribute to the multifactorial etiology of these injuries in the elderly. Interestingly, young age at the time of index surgery has also been implicated to place patients at an increased risk for future fracture [13]. Increased activity level leading to accelerated osteolysis and aseptic loosening from wear particles, as well as high-energy trauma have been postulated reasons. Female gender has been described as a risk factor for periprosthetic femoral fractures [10,14,15]. In a study of 305 postoperative periprosthetic fractures, Singh et al. [10] analyzed data from the Mayo Clinic Joint Registry and found female gender 62
to be associated with a 50% increased risk of periprosthetic femur fracture. The authors speculated that the overall higher incidence of osteoporosis among women may be a major contributing factor to their results. Similarly, Beals and Tower [7] reported that there was a prior history of vertebral or metaphyseal fractures in 38% of patients who had periprosthetic fractures. The contribution of female gender to sustaining these fractures is likely multifactorial with variables such as osteopenia and an increased proportion of female patients undergoing total hip arthroplasty in national registry data [16,17] also contributing. Conversely, it can be postulated that male patients are at an increased risk of periprosthetic fracture secondary to the increased loads placed across the implant. This increased load may lead to increased wear particle generation, osteolysis and ultimately, periprosthetic fracture. Obesity may lead to an increased stress across implant interfaces, which can lead to higher rates of osteolysis and loosening. Osteolysis and loosening both have been shown to contribute to an increased risk of periprosthetic fracture [9,11,18]. Osteolysis not only can cause decreased overall bone stock and increased fragility, but can be the cause of osteolytic defects. Also, higher loads from body weight are likely transferred through the prosthesis and bone during a fall in the obese patient. However, the evidence is inconclusive at this time [10,19]. The increased risk of post-operative periprosthetic in the revision setting has been well-documented [5,11]. Bone stock in the revision setting may be compromised for various reasons including: osteolysis, previous cortical defects, impending stress risers from previous surgery and difficulty with obtaining adequate exposure. Removing implants, as well as obtaining adequate fixation, especially with uncemented prostheses can also present challenges in the setting of decreased bone stock [9]. Interestingly, Lindahl et al. [11] noted a decreased time interval between periprosthetic fractures in those undergoing multiple revisions. The authors found that the mean time from the index hip arthroplasty to fracture was 7.4 years, and this time interval subsequently decreased to a mean of 3.9 years, 3.8 years and 2.3 years for the first, second and third revisions, respectively. The inherent complications associated with revision of the femoral stem in a total hip arthroplasty likely all contribute to the increased risk of periprosthetic femoral fractures. Classification
Over the years, there have been a number of classification schemes used to classify periprosthetic femur fractures around a femoral prosthesis, however, the most extensively used system is the Vancouver classification originally described by Duncan and Masri [20]. This classification scheme has separate variations for both intra-operative and post-operative periprosthetic fractures around the femoral stem. The post-operative version of the Vancouver classification is subdivided into types A, B and C based on fracture location, with further subdivision of Vancouver B based on implant stability and bone quality. Type A fractures involve either the greater trochanter (AG) or the lesser trochanter (AL). Type B fractures are diaphyseal fractures that Expert Rev. Med. Devices 12(1), (2015)
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty
occur at the level of, or just distal to the femoral stem. This is further divided into B1 (fractures with a stable implant), B2 (fractures with an unstable implant with adequate bone stock) and B3 (fractures with an unstable implant and inadequate bone stock). Type C fractures are diaphyseal fractures that occur distal to the tip of the femoral stem. Ideally, classification schemes should be simple to communicate, as well as be able to dictate treatment and predict prognosis to a certain degree. The Vancouver classification has been shown to be both valid and reliable in various studies [21,22]. Naqvi et al. [21] analyzed the validity of the Vancouver classification in 37 type B fractures by comparing the radiographic classification with intra-operative findings. The authors reported an 81% agreement within B1, B2 and B3 subgroups. In the same study, the inter- and intra-observer reliability was also tested using 6 observers (3 consultants, 3 trainees) in a retrospective study of 45 patients with periprosthetic proximal femur fractures. The authors reported that the mean Kappa value for inter-observer reliability was 0.69 (range, 0.63–0.72) for consultants and 0.61 for trainees (range, 0.56–0.65). Similarly, the near-perfect intra-observer kappa values for consultants and trainees were 0.81 and 0.80, respectively. The authors concluded that although the study demonstrated a high validity, a 100% agreement was not achieved and that implant stability could not be determined from plain radiographs alone. In addition to the Vancouver classification, a recent study has described a unique periprosthetic fracture pattern that has been found around a well-fixed polished, tapered, collarless femoral stem [23]. The authors described that under low trauma, the smooth stem acts as a wedge that is driven into the cement mantle resulting in splitting and fragmentation of the proximal femur with the cement bone interface remaining intact. Furthermore, the forces placed on the offset-femoral head results in a twisting mechanism of the stem resulting in a spiral fracture around the stem (this topic will be covered in the management section under the Vancouver B2 treatment heading). Management Type AG
Recent data from the Swedish Joint Registry reported that there is a 4% incidence (47 of 1049) of Vancouver type A fractures [11]. Fractures of the AG around a femoral prosthesis with 2.5 cm and non-union of the AG causing instability, pain and/or abductor weakness. Several commercially available claw-plate devices can be used along with different configurations of tension-band wiring constructs [25]. Both fixation constructs have been found to lead to successful fracture union as long as basic Arbeitsgemeinschaft fu¨r Osteosynthesefragen (AO) fixation principles are adhered to. However, the surgeon must be concerned about the high likelihood of symptomatic hardware when using bulky constructs over the AG. informahealthcare.com
Review
Type AL
These rare periprosthetic fractures involving the AL are usually inconsequential and do not result in detrimental outcomes, even if the fracture fragment is displaced [25]. Therefore, nonoperative management is usually the treatment of choice. However, when the fracture extends distally to involve the medial calcar, the stability of the stem must be considered. Fractures of the AL that extend into the calcar femorale may require the addition of cerclage wires or even revision arthroplasty with a longer femoral stem depending on prosthesis stability and fracture line propagation [25,26]. Type B1
Displaced fractures occurring around a stable femoral prosthesis usually require some form of operative fixation. Some controversy exists regarding the optimal method of fixation as several fixation constructs, including revision surgery, have been described as viable treatment options. Treatment options for Vancouver B1 fractures can include the following options: lateral dynamic compression versus locked plating with or without cortical strut allografts, cerclage wires, cables or any combination [27–30]. However, some authors prefer revision arthroplasty to a longer stem, because they believe the benefits of intramedullary fracture fixation outweigh the drawbacks of prosthesis removal [31]. More recently, in the presence of a well-fixed cement mantle that was securely attached to bone, a cement-in-cement technique to treat Vancouver B periprosthetic femoral fractures with a cemented stem has been described [32–34]. In this technique, the native femoral stem is removed and the cement mantle is carefully inspected intraoperatively with a rigid arthroscope. If the cement mantle is deemed to be well-fixed and free from defects, the fracture fragments are anatomically reduced and a smaller femoral stem is re-cemented into the existing cement mantle. Adjunctive fixation in the form of cerclage wires and cables along with plates and strut allografts should be used to supplement the construct. However, cement extrusion between fracture fragments may impede union. In a series of 23 Vancouver B fractures (3 B1, 17 B2, 3 B3) treated with the cement-in-cement technique, Briant-Evans et al. [33] reported radiographic union in 18 fractures at a mean of 4.4 months (range, 2–11 months). Only one patient in this cohort underwent revision for non-union, which the authors attributed to surgical technique. The authors concluded that this technique results in reduced operative time, is technically less demanding and is ideal in elderly patients who are unable to tolerate a more lengthy procedure. Further prospective, randomized studies are needed with long-term follow-up of this technique, however, the current short- to midterm results are encouraging. Though uncommon, component revision of the femur as well as the acetabulum should always be considered if there is implant malposition predisposing to non-union and/or re-fracture. Acetabular revision should be considered when severe osteolysis is noted in conjunction with marked polyethylene wear. O’Shea et al. reported on 19 concomitant acetabular 63
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Review
Wu, Cherian, Kapadia, Banerjee, Jauregui & Mont
revisions in 22 periprosthetic fractures around the femur. The authors noted that 7 out of the 19 patients had acetabular defects that required treatment with impaction bone grafting and a cemented cup [35]. However, distinguishing between a periprosthetic fracture with a stable implant (B1) and one with an unstable implant (B2 and B3) is critical in choosing the most appropriate treatment option. The treatment of a periprosthetic femur fracture around a well-positioned and wellfixed implant with a revision surgery may unnecessarily subject patients to a highly morbid procedure, whereas treating loose implants with internal fixation alone will inevitably lead to failure. The use of locking plate technology has been shown to be successful in achieving fracture union in the treatment of periprosthetic femur fractures around a stable implant [2,27,36]. Bryant et al. [27] reported their clinical and radiographic outcomes in 10 patients who sustained a Vancouver B1 periprosthetic femur fracture treated solely with a locked compression plate that spanned the entire femur without the use of any cortical onlay strut allografts or cerclage wires/cables. The authors found that clinical and radiographic fracture union was achieved in all 10 patients at a mean of 17 weeks (range, 12–27 weeks). However, 5 out of the 10 patients received additional interfragmentary lag screw fixation if the fracture pattern was deemed appropriate. The concept of spanning the entire length of the femur and using bicortical fixation around the proximal aspect of the femoral prosthesis was also stressed. The authors concluded that successful outcomes could be obtained with the use of a spanning lateral locked compression plate along the entire length of the femur in Vancouver B1 fractures. Using a similar fixation construct to Bryant et al. [27], Buttaro et al. [28] reported 6 failures out of 14 consecutive patients with Vancouver B1 periprosthetic femur fractures around cemented prostheses. All six failures resulted from either locking plate fracture or screw pullout. Interestingly, five out of the six failures were in patients who did not receive cortical strut allograft augmentation in addition to the lateral locked plate. Of the five patients who did not initially receive supplementary cortical struts, three were revised successfully and went on to fracture union after revision open reduction internal fixation with strut grafting. In contrast to Bryant et al., the authors concluded that treatment with an isolated lateral locked plate for Vancouver B1 fractures did not provide sufficient fracture fixation, and that the addition of cortical strut allograft to a lateral locked plate should be routinely used. The authors concluded that inadequate utilization of non-locking and locking screws in their fixation construct, in addition to a majority of the patients having undergone revision hip arthroplasty in the past (11 of 14 patients), may have contributed to the poorer outcomes compared to previous reports [27,29]. Similarly, recent evidence may suggest that there are increasing problems with the use of locking plates for the treatment of B1 fractures due to the excess rigidity from the locking fixation construct. This in turn can lead to an absence of adequate micromotion between fracture fragments, non-union, and 64
ultimately implant failure. However, these reports describe varied results [36–39]. A recent review of biomechanical testing by Moazen et al. [37] highlighted the importance of the rigidity of B1 fracture fixation with a single locking plate; moreover, the authors noted that further increasing the rigidity of the construct may be deleterious to fracture healing and callus formation from inadequate micromotion thus, ultimately placing higher mechanical stresses on the locking plate leading to implant fatigue and failure. There have been several attempts to rectify this problem by increasing the bridging length of the fixation construct, limiting the number of screws placed on each side of the fracture, to allow for adequate motion at the fracture site and placing the plate with at least 2 mm gapping from the bone. Currently, there are several approaches that can be attempted to mitigate this problem; however, the ideal fixation construct should aim to provide adequate strength for rigid internal fixation while allowing for enough micromotion for fracture callus formation. The optimal fixation construct for Vancouver B1 fractures remains controversial. Direct comparison between studies is difficult due to the heterogeneity of patient groups (age, gender, reason for revision, premorbid activity level) being studied and the variations in surgical technique based on surgeon training and preference. Variations in supplemental fixation with interfragmentary screws, cortical struts, cables and wires also add to the complexity of determining which construct is optimal. However, locking plate technology, with and without additional fixation, has shown encouraging results. Moreover, adhering to basic AO fracture principles and respecting the soft tissues around the fracture site is essential. Type B2
Periprosthetic femur fractures around an unstable prosthesis in the presence of adequate bone stock should ideally be treated with fracture reduction and fixation with revision arthroplasty to a longer cemented or cementless stem. Surgeons should aim to bypass the most distal cortical defect by at least two cortical diameters in order to obtain implant fixation away from the fracture site [25,26,40,41]. Achieving distal diaphyseal fixation allows the surgeon to gain axial and rotational control of the revision stem. Several authors have demonstrated that isolated internal fixation in this fracture subtype will lead to inadequate fixation and failure [25,26,40]. Supplementary fixation with cortical strut grafts and cerclage wires can be used, but should always be in conjunction with femoral stem revision. Currently, the use of both proximally porous-coated and extensively porous-coated stems have been described in varying reports to treat this fracture subtype, with several stems having monolithic and modular revision stem variations. More recently, modular, tapered revision stems have become popular among many surgeons due to their promising short-term survivorship and functional outcomes [42–44]. Neumann et al. [44] reported their results on 55 Vancouver B periprosthetic fractures (35 B2, 20 B3) treated with a modular cementless stem with supplementary cerclage bands. At a Expert Rev. Med. Devices 12(1), (2015)
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty
mean follow-up of 67 months, complete fracture union occurred in all 35 patients. The authors concluded that the use of a modular cementless stem in both B2 and B3 fracture subtype was a reliable technique for fracture union. Similarly, Munro et al. [43] reported 95.6% (44 out of 46) implant survival of a tapered modular fluted titanium stem at a mean follow-up of 54 months (range, 24–143 months) in the treatment of Vancouver B2 and B3 fractures. The authors found that the mean University of California Los Angles Activity score and Oxford Hip Scores for Vancouver B2 were 4 and 74, respectively. Promising results of uncemented extensively porous-coated implants have also been described in several studies [35,45]. In a study of 22 Vancouver B (10 B2, 12 B3) fractures treated with an uncemented extensively porous-coated femoral stem, O’Shea et al. [35] reported only 1 patient with an unstable subsidence of the femoral stem at a mean follow-up of 33.7 months (range, 18–55 months). The authors reported that fracture union was achieved in 20 out of 22 patients, with 17 out of the 22 patients achieving a Harris Hip Scores greater than 80 points. Additionally, the authors noted that patients had a mean University of California Los Angles Activity score activity score and total 36-item short-form score of 4.1 (range, 1–7) and 71.1 (range, 19.1–94.8), respectively. The authors concluded that for their examined patient population with a mean age of 75 years (range, 59–100 years), these were excellent results. Historically, the use of a cemented femoral stem and uncemented proximally porous-coated stem to treat Vancouver B fractures have been described in several reports. These constructs have largely fallen out of favor due to the poor outcomes seen with this treatment option [7,31]. Mont and Maar [46] reported a 31% non-union rate in cemented femoral revisions of Vancouver B2 fractures. Similarly, Springer et al. [31] treated 42 Vancouver B fractures with revision to a cemented stem. At a mean follow-up of 68 months (range, 24–185 months), 60% (25 of 42) of the stems were stable with fracture union. At a mean follow-up of 85 months (range, 30–175 months), Springer et al. [31] found evidence of radiographic loosening in 12 of 25 (48%) unrevised stems. Only 10 of 28 (36%) treated Vancouver B fractures had stems that remained well-fixed and unrevised. The authors acknowledged the overall poor results, but made mention that in their series, there was a progression from the use of cemented femoral stems to cementless proximally porous-coated stems to the currently popular cementless extensively porous-coated stem. The use of long uncemented extensively porous-coated stems to gain diaphyseal fixation is the treatment of choice among many surgeons for treating Vancouver B2 fractures. Recent advances in implant modularity have made these revision femoral stems even more appealing. Depending on the fracture pattern, many surgeons may choose to supplement with plating, cables/wires and/or struts grafts either in combination or alone. The heterogeneity of fixation constructs and variations in surgical technique make comparing the different fixation combinations difficult. However, if the femoral stem is deemed to be unstable, revision arthroplasty must be included in the treatment algorithm. informahealthcare.com
Review
A unique periprosthetic fracture subtype of Vancouver B2 facture patterns that has been found around a well-fixed femoral stems with a polished, tapered, collarless total hip arthroplasty results in a splitting, fragmentation and spiral fracture of the proximal femur with the cement bone interface remaining intact [23]. The authors note the importance of proper recognition and classification of this fracture pattern as extensive reconstruction is often required for this type of injury. Grammatopoulos et al. [23] reported on a prospectively collected series of 14 patients and described the intra-operative findings. The authors noted the proximal femoral metaphysis was characteristically split into four to six fragments rendering the prosthesis unstable. In addition, well-fixed cement-bone interfaces were routinely noted. The authors noted that all patients required complex reconstructions: nine individuals required a long prosthesis bypassing the fracture by the length of two cortical diameters while five revision cases were performed with a tri-modular stemmed prostheses with the aid of proximal metaphyseal mesh and impaction bone grafting. Type B3
The deficient bone stock present in Vancouver B3 fractures offers the unique challenge of gaining adequate prosthesis stability and fracture fixation. Periprosthetic fractures around a loose femoral stem in the presence of inadequate bone stock or comminution are typically treated with revision of the stem along with bone augmentation through the use of cortical onlay strut allograft and/or allograft-prosthetic composites (APC) [25,35,47–49]. End-stage prosthetic disease may require an endoprosthesis. More recently, modular distally fixed cementless stems have been used with encouraging results [42–44,50,51]. If there is at least 4–6 cm of remaining distal bone stock present, many of the same prostheses and techniques utilized for Vancouver B2 fractures can be used in B3 fracture types with supplementary augmentation of bone stock [52]. Using a modular fluted tapered uncemented stem with preservation of the proximal femoral bone, Berry [50] reported no cases of stem subsidence or evidence of radiographic loosening in seven out of eight patients with Vancouver B3 fractures at a mean follow-up of 1.5 years (range, 1–2 years). The one remaining patient died secondary to a medical-related condition, and was unavailable for follow-up. The authors cited that one of the major advantages of this technique was the reconstitution of proximal femoral bone and avoidance of APCs and endoprostheses. At the latest follow-up, five out of seven patients reported that they had no pain, and none had severe pain. A study by Fink et al. [42], examined fracture union and stem subsidence in 10 Vancouver B3 fractures treated with a modular, non-cemented tapered fluted femoral stem at a mean follow-up of 32 months (range, 24–60 months). The authors reported that there were no cases of stem subsidence and 100% fracture union rate in at the latest follow-up. Interestingly, the authors noted that in cases where patients had
Review
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty: a review Expert Rev. Med. Devices 12(1), 61–72 (2015)
Eddie S Wu, Jeffrey J Cherian, Bhaveen H Kapadia, Samik Banerjee, Julio J Jauregui and Michael A Mont* Rubin Institute for Advanced Orthopedics, Center for Joint Preservation and Replacement, Sinai Hospital of Baltimore, 2401 West Belvedere Avenue, Baltimore, MD 21215, USA *Author for correspondence: Tel.: +1 410 601 8500 Fax: +1 410 601 8501 [email protected]; [email protected]
As the number of primary total hip arthroplasties increase over the next several decades so will the incidence of periprosthetic fractures around the femoral stem. Treatment can reliably be predicted using the Vancouver classification with internal fixation being indicated in fractures involving a stable implant and revision arthroplasty indicated in those with unstable prostheses. Non-displaced fractures involving the greater and lesser trochanter can generally be treated non-operatively. Extensively porous-coated stems and the use of modular uncemented revision stems to treat Vancouver B fractures have shown encouraging results. The treatment of Vancouver C periprosthetic fractures continues to follow basic AO fixation principles with an emphasis on eliminating stress risers with adequate implant overlap and length. This review will focus on the risk factors and classification of these fractures, as well as highlight the treatment options for post-operative periprosthetic femoral fractures around a total hip arthroplasty. KEYWORDS: hip • periprosthetic fracture • revision • total hip arthroplasty • Vancouver classification
Each year, more than 380,000 total hip arthroplasties are performed in the USA, with estimates expected to increase to over 1 million procedures yearly by 2030 [1,2]. As the number of primary total hip arthroplasties continues to increase, so will the number of complications and revision procedures due to periprosthetic fractures around the femoral stem. The first reports of periprosthetic femur fractures around total hip arthroplasty were described as early as 1954 [3]. Currently, it is estimated that the rate of revision total hip arthroplasty is expected to double by the year 2026, with periprosthetic fracture being an increasing cause [4]. A recent study reported that there was a 1% incidence of periprosthetic fracture in primary total hip arthroplasty, while in the revision setting the incidence rose to 4% [5]. The etiology of a post-operative periprosthetic femur fracture is multifactorial and may be influenced by the type of prosthesis used and whether a revision arthroplasty was involved. Post-operative periprosthetic fractures around the femoral stem usually occur as a
informahealthcare.com
10.1586/17434440.2015.958076
result of low energy trauma and/or undiagnosed osteolytic defects [6]. Patients who receive ingrowth prostheses may have a higher tendency to fracture within the first 6 months after implantation from iatrogenic stress risers generated from prior preparation of the canal [7]. However, patients who receive cemented stems at the time of the index hip arthroplasty have been found to fracture more commonly at the distal stem tip several years after implantation [7]. The etiology of this is likely multifactorial, but may be a result of osteolysis from the generation of wear debris. As the number of periprosthetic fractures increase, the treatment of this complication will become an increasingly important topic for discussion due to the substantial economic burden placed on the healthcare system. In a Medicare database study, Ong et al. reported that revision total hip arthroplasty consumed an average of 18.8% (US$238 million) of the entire economic burden of total hip arthroplasty [8]. With increasing procedural costs and decreasing reimbursement in total hip revision
Ó 2015 Informa UK Ltd
ISSN 1743-4440
61
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Review
Wu, Cherian, Kapadia, Banerjee, Jauregui & Mont
procedure, the economic burden in treating these complex cases will continue to grow. Therefore, identifying risk factors and establishing optimal treatment algorithms for periprosthetic fractures is imperative to minimize complications while maximizing outcomes. In this review, the authors aim to analyze the recent evidence on: risk factors, fracture classification schemes and treatment strategies that have been used for periprosthetic fractures around a total hip arthroplasty. This report aims to provide orthopedic surgeons with a concise overview of the available options and currently preferred treatment strategies. Methods
A review of the literature was performed using four electronic databases of EMBASE, CINAHL-plus, PubMed and SCOPUS, which retrieved articles between January 1976 and September 2013 by one author (EW). Periprosthetic femur fractures can occur either intra-operatively or post-operatively, with the focus of this report on fractures of the femur that occur post-operatively. Preference was given to randomized controlled trials, meta-analyses and data from national registries over the past 5 years; however, all relevant studies were included in the analysis. Search terms that were utilized consisted of (hip[title] or femur[title]) and (arthroplasty[title] or replacement[title] or periprosthetic[title]) and (fracture[title]) and single keyword search (‘Vancouver A’, ‘Vancouver B’, ‘Vancouver C’). The literature search resulted in 941 reports that addressed periprosthetic fractures around total joint arthroplasty, of which 58 specifically focused on treatment strategies for fractures around total hip arthroplasty. Risk factors
Several risk factors are described in the literature for periprosthetic femoral fracture, however, most of them are likely interrelated. Age, gender, revision surgery, implant type, medical comorbidities and osteolysis have all been reported to potentially increase the risk [7,9–11]. In general, most patients who undergo hip arthroplasty are well into their adult years, and therefore, the selection bias of studies analyzing age and periprosthetic fractures must be interpreted with caution as many cohorts involve elderly patients. Lindahl et al. reported that the risk ratio for fracture was 1.01 for every additional age year [12]. Increasing age is inevitably associated with confounding medical comorbidities such as osteopenia and the increased fall risk that may likely contribute to the multifactorial etiology of these injuries in the elderly. Interestingly, young age at the time of index surgery has also been implicated to place patients at an increased risk for future fracture [13]. Increased activity level leading to accelerated osteolysis and aseptic loosening from wear particles, as well as high-energy trauma have been postulated reasons. Female gender has been described as a risk factor for periprosthetic femoral fractures [10,14,15]. In a study of 305 postoperative periprosthetic fractures, Singh et al. [10] analyzed data from the Mayo Clinic Joint Registry and found female gender 62
to be associated with a 50% increased risk of periprosthetic femur fracture. The authors speculated that the overall higher incidence of osteoporosis among women may be a major contributing factor to their results. Similarly, Beals and Tower [7] reported that there was a prior history of vertebral or metaphyseal fractures in 38% of patients who had periprosthetic fractures. The contribution of female gender to sustaining these fractures is likely multifactorial with variables such as osteopenia and an increased proportion of female patients undergoing total hip arthroplasty in national registry data [16,17] also contributing. Conversely, it can be postulated that male patients are at an increased risk of periprosthetic fracture secondary to the increased loads placed across the implant. This increased load may lead to increased wear particle generation, osteolysis and ultimately, periprosthetic fracture. Obesity may lead to an increased stress across implant interfaces, which can lead to higher rates of osteolysis and loosening. Osteolysis and loosening both have been shown to contribute to an increased risk of periprosthetic fracture [9,11,18]. Osteolysis not only can cause decreased overall bone stock and increased fragility, but can be the cause of osteolytic defects. Also, higher loads from body weight are likely transferred through the prosthesis and bone during a fall in the obese patient. However, the evidence is inconclusive at this time [10,19]. The increased risk of post-operative periprosthetic in the revision setting has been well-documented [5,11]. Bone stock in the revision setting may be compromised for various reasons including: osteolysis, previous cortical defects, impending stress risers from previous surgery and difficulty with obtaining adequate exposure. Removing implants, as well as obtaining adequate fixation, especially with uncemented prostheses can also present challenges in the setting of decreased bone stock [9]. Interestingly, Lindahl et al. [11] noted a decreased time interval between periprosthetic fractures in those undergoing multiple revisions. The authors found that the mean time from the index hip arthroplasty to fracture was 7.4 years, and this time interval subsequently decreased to a mean of 3.9 years, 3.8 years and 2.3 years for the first, second and third revisions, respectively. The inherent complications associated with revision of the femoral stem in a total hip arthroplasty likely all contribute to the increased risk of periprosthetic femoral fractures. Classification
Over the years, there have been a number of classification schemes used to classify periprosthetic femur fractures around a femoral prosthesis, however, the most extensively used system is the Vancouver classification originally described by Duncan and Masri [20]. This classification scheme has separate variations for both intra-operative and post-operative periprosthetic fractures around the femoral stem. The post-operative version of the Vancouver classification is subdivided into types A, B and C based on fracture location, with further subdivision of Vancouver B based on implant stability and bone quality. Type A fractures involve either the greater trochanter (AG) or the lesser trochanter (AL). Type B fractures are diaphyseal fractures that Expert Rev. Med. Devices 12(1), (2015)
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty
occur at the level of, or just distal to the femoral stem. This is further divided into B1 (fractures with a stable implant), B2 (fractures with an unstable implant with adequate bone stock) and B3 (fractures with an unstable implant and inadequate bone stock). Type C fractures are diaphyseal fractures that occur distal to the tip of the femoral stem. Ideally, classification schemes should be simple to communicate, as well as be able to dictate treatment and predict prognosis to a certain degree. The Vancouver classification has been shown to be both valid and reliable in various studies [21,22]. Naqvi et al. [21] analyzed the validity of the Vancouver classification in 37 type B fractures by comparing the radiographic classification with intra-operative findings. The authors reported an 81% agreement within B1, B2 and B3 subgroups. In the same study, the inter- and intra-observer reliability was also tested using 6 observers (3 consultants, 3 trainees) in a retrospective study of 45 patients with periprosthetic proximal femur fractures. The authors reported that the mean Kappa value for inter-observer reliability was 0.69 (range, 0.63–0.72) for consultants and 0.61 for trainees (range, 0.56–0.65). Similarly, the near-perfect intra-observer kappa values for consultants and trainees were 0.81 and 0.80, respectively. The authors concluded that although the study demonstrated a high validity, a 100% agreement was not achieved and that implant stability could not be determined from plain radiographs alone. In addition to the Vancouver classification, a recent study has described a unique periprosthetic fracture pattern that has been found around a well-fixed polished, tapered, collarless femoral stem [23]. The authors described that under low trauma, the smooth stem acts as a wedge that is driven into the cement mantle resulting in splitting and fragmentation of the proximal femur with the cement bone interface remaining intact. Furthermore, the forces placed on the offset-femoral head results in a twisting mechanism of the stem resulting in a spiral fracture around the stem (this topic will be covered in the management section under the Vancouver B2 treatment heading). Management Type AG
Recent data from the Swedish Joint Registry reported that there is a 4% incidence (47 of 1049) of Vancouver type A fractures [11]. Fractures of the AG around a femoral prosthesis with 2.5 cm and non-union of the AG causing instability, pain and/or abductor weakness. Several commercially available claw-plate devices can be used along with different configurations of tension-band wiring constructs [25]. Both fixation constructs have been found to lead to successful fracture union as long as basic Arbeitsgemeinschaft fu¨r Osteosynthesefragen (AO) fixation principles are adhered to. However, the surgeon must be concerned about the high likelihood of symptomatic hardware when using bulky constructs over the AG. informahealthcare.com
Review
Type AL
These rare periprosthetic fractures involving the AL are usually inconsequential and do not result in detrimental outcomes, even if the fracture fragment is displaced [25]. Therefore, nonoperative management is usually the treatment of choice. However, when the fracture extends distally to involve the medial calcar, the stability of the stem must be considered. Fractures of the AL that extend into the calcar femorale may require the addition of cerclage wires or even revision arthroplasty with a longer femoral stem depending on prosthesis stability and fracture line propagation [25,26]. Type B1
Displaced fractures occurring around a stable femoral prosthesis usually require some form of operative fixation. Some controversy exists regarding the optimal method of fixation as several fixation constructs, including revision surgery, have been described as viable treatment options. Treatment options for Vancouver B1 fractures can include the following options: lateral dynamic compression versus locked plating with or without cortical strut allografts, cerclage wires, cables or any combination [27–30]. However, some authors prefer revision arthroplasty to a longer stem, because they believe the benefits of intramedullary fracture fixation outweigh the drawbacks of prosthesis removal [31]. More recently, in the presence of a well-fixed cement mantle that was securely attached to bone, a cement-in-cement technique to treat Vancouver B periprosthetic femoral fractures with a cemented stem has been described [32–34]. In this technique, the native femoral stem is removed and the cement mantle is carefully inspected intraoperatively with a rigid arthroscope. If the cement mantle is deemed to be well-fixed and free from defects, the fracture fragments are anatomically reduced and a smaller femoral stem is re-cemented into the existing cement mantle. Adjunctive fixation in the form of cerclage wires and cables along with plates and strut allografts should be used to supplement the construct. However, cement extrusion between fracture fragments may impede union. In a series of 23 Vancouver B fractures (3 B1, 17 B2, 3 B3) treated with the cement-in-cement technique, Briant-Evans et al. [33] reported radiographic union in 18 fractures at a mean of 4.4 months (range, 2–11 months). Only one patient in this cohort underwent revision for non-union, which the authors attributed to surgical technique. The authors concluded that this technique results in reduced operative time, is technically less demanding and is ideal in elderly patients who are unable to tolerate a more lengthy procedure. Further prospective, randomized studies are needed with long-term follow-up of this technique, however, the current short- to midterm results are encouraging. Though uncommon, component revision of the femur as well as the acetabulum should always be considered if there is implant malposition predisposing to non-union and/or re-fracture. Acetabular revision should be considered when severe osteolysis is noted in conjunction with marked polyethylene wear. O’Shea et al. reported on 19 concomitant acetabular 63
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Review
Wu, Cherian, Kapadia, Banerjee, Jauregui & Mont
revisions in 22 periprosthetic fractures around the femur. The authors noted that 7 out of the 19 patients had acetabular defects that required treatment with impaction bone grafting and a cemented cup [35]. However, distinguishing between a periprosthetic fracture with a stable implant (B1) and one with an unstable implant (B2 and B3) is critical in choosing the most appropriate treatment option. The treatment of a periprosthetic femur fracture around a well-positioned and wellfixed implant with a revision surgery may unnecessarily subject patients to a highly morbid procedure, whereas treating loose implants with internal fixation alone will inevitably lead to failure. The use of locking plate technology has been shown to be successful in achieving fracture union in the treatment of periprosthetic femur fractures around a stable implant [2,27,36]. Bryant et al. [27] reported their clinical and radiographic outcomes in 10 patients who sustained a Vancouver B1 periprosthetic femur fracture treated solely with a locked compression plate that spanned the entire femur without the use of any cortical onlay strut allografts or cerclage wires/cables. The authors found that clinical and radiographic fracture union was achieved in all 10 patients at a mean of 17 weeks (range, 12–27 weeks). However, 5 out of the 10 patients received additional interfragmentary lag screw fixation if the fracture pattern was deemed appropriate. The concept of spanning the entire length of the femur and using bicortical fixation around the proximal aspect of the femoral prosthesis was also stressed. The authors concluded that successful outcomes could be obtained with the use of a spanning lateral locked compression plate along the entire length of the femur in Vancouver B1 fractures. Using a similar fixation construct to Bryant et al. [27], Buttaro et al. [28] reported 6 failures out of 14 consecutive patients with Vancouver B1 periprosthetic femur fractures around cemented prostheses. All six failures resulted from either locking plate fracture or screw pullout. Interestingly, five out of the six failures were in patients who did not receive cortical strut allograft augmentation in addition to the lateral locked plate. Of the five patients who did not initially receive supplementary cortical struts, three were revised successfully and went on to fracture union after revision open reduction internal fixation with strut grafting. In contrast to Bryant et al., the authors concluded that treatment with an isolated lateral locked plate for Vancouver B1 fractures did not provide sufficient fracture fixation, and that the addition of cortical strut allograft to a lateral locked plate should be routinely used. The authors concluded that inadequate utilization of non-locking and locking screws in their fixation construct, in addition to a majority of the patients having undergone revision hip arthroplasty in the past (11 of 14 patients), may have contributed to the poorer outcomes compared to previous reports [27,29]. Similarly, recent evidence may suggest that there are increasing problems with the use of locking plates for the treatment of B1 fractures due to the excess rigidity from the locking fixation construct. This in turn can lead to an absence of adequate micromotion between fracture fragments, non-union, and 64
ultimately implant failure. However, these reports describe varied results [36–39]. A recent review of biomechanical testing by Moazen et al. [37] highlighted the importance of the rigidity of B1 fracture fixation with a single locking plate; moreover, the authors noted that further increasing the rigidity of the construct may be deleterious to fracture healing and callus formation from inadequate micromotion thus, ultimately placing higher mechanical stresses on the locking plate leading to implant fatigue and failure. There have been several attempts to rectify this problem by increasing the bridging length of the fixation construct, limiting the number of screws placed on each side of the fracture, to allow for adequate motion at the fracture site and placing the plate with at least 2 mm gapping from the bone. Currently, there are several approaches that can be attempted to mitigate this problem; however, the ideal fixation construct should aim to provide adequate strength for rigid internal fixation while allowing for enough micromotion for fracture callus formation. The optimal fixation construct for Vancouver B1 fractures remains controversial. Direct comparison between studies is difficult due to the heterogeneity of patient groups (age, gender, reason for revision, premorbid activity level) being studied and the variations in surgical technique based on surgeon training and preference. Variations in supplemental fixation with interfragmentary screws, cortical struts, cables and wires also add to the complexity of determining which construct is optimal. However, locking plate technology, with and without additional fixation, has shown encouraging results. Moreover, adhering to basic AO fracture principles and respecting the soft tissues around the fracture site is essential. Type B2
Periprosthetic femur fractures around an unstable prosthesis in the presence of adequate bone stock should ideally be treated with fracture reduction and fixation with revision arthroplasty to a longer cemented or cementless stem. Surgeons should aim to bypass the most distal cortical defect by at least two cortical diameters in order to obtain implant fixation away from the fracture site [25,26,40,41]. Achieving distal diaphyseal fixation allows the surgeon to gain axial and rotational control of the revision stem. Several authors have demonstrated that isolated internal fixation in this fracture subtype will lead to inadequate fixation and failure [25,26,40]. Supplementary fixation with cortical strut grafts and cerclage wires can be used, but should always be in conjunction with femoral stem revision. Currently, the use of both proximally porous-coated and extensively porous-coated stems have been described in varying reports to treat this fracture subtype, with several stems having monolithic and modular revision stem variations. More recently, modular, tapered revision stems have become popular among many surgeons due to their promising short-term survivorship and functional outcomes [42–44]. Neumann et al. [44] reported their results on 55 Vancouver B periprosthetic fractures (35 B2, 20 B3) treated with a modular cementless stem with supplementary cerclage bands. At a Expert Rev. Med. Devices 12(1), (2015)
Expert Review of Medical Devices Downloaded from informahealthcare.com by Nanyang Technological University on 04/24/15 For personal use only.
Outcomes of post-operative periprosthetic femur fracture around total hip arthroplasty
mean follow-up of 67 months, complete fracture union occurred in all 35 patients. The authors concluded that the use of a modular cementless stem in both B2 and B3 fracture subtype was a reliable technique for fracture union. Similarly, Munro et al. [43] reported 95.6% (44 out of 46) implant survival of a tapered modular fluted titanium stem at a mean follow-up of 54 months (range, 24–143 months) in the treatment of Vancouver B2 and B3 fractures. The authors found that the mean University of California Los Angles Activity score and Oxford Hip Scores for Vancouver B2 were 4 and 74, respectively. Promising results of uncemented extensively porous-coated implants have also been described in several studies [35,45]. In a study of 22 Vancouver B (10 B2, 12 B3) fractures treated with an uncemented extensively porous-coated femoral stem, O’Shea et al. [35] reported only 1 patient with an unstable subsidence of the femoral stem at a mean follow-up of 33.7 months (range, 18–55 months). The authors reported that fracture union was achieved in 20 out of 22 patients, with 17 out of the 22 patients achieving a Harris Hip Scores greater than 80 points. Additionally, the authors noted that patients had a mean University of California Los Angles Activity score activity score and total 36-item short-form score of 4.1 (range, 1–7) and 71.1 (range, 19.1–94.8), respectively. The authors concluded that for their examined patient population with a mean age of 75 years (range, 59–100 years), these were excellent results. Historically, the use of a cemented femoral stem and uncemented proximally porous-coated stem to treat Vancouver B fractures have been described in several reports. These constructs have largely fallen out of favor due to the poor outcomes seen with this treatment option [7,31]. Mont and Maar [46] reported a 31% non-union rate in cemented femoral revisions of Vancouver B2 fractures. Similarly, Springer et al. [31] treated 42 Vancouver B fractures with revision to a cemented stem. At a mean follow-up of 68 months (range, 24–185 months), 60% (25 of 42) of the stems were stable with fracture union. At a mean follow-up of 85 months (range, 30–175 months), Springer et al. [31] found evidence of radiographic loosening in 12 of 25 (48%) unrevised stems. Only 10 of 28 (36%) treated Vancouver B fractures had stems that remained well-fixed and unrevised. The authors acknowledged the overall poor results, but made mention that in their series, there was a progression from the use of cemented femoral stems to cementless proximally porous-coated stems to the currently popular cementless extensively porous-coated stem. The use of long uncemented extensively porous-coated stems to gain diaphyseal fixation is the treatment of choice among many surgeons for treating Vancouver B2 fractures. Recent advances in implant modularity have made these revision femoral stems even more appealing. Depending on the fracture pattern, many surgeons may choose to supplement with plating, cables/wires and/or struts grafts either in combination or alone. The heterogeneity of fixation constructs and variations in surgical technique make comparing the different fixation combinations difficult. However, if the femoral stem is deemed to be unstable, revision arthroplasty must be included in the treatment algorithm. informahealthcare.com
Review
A unique periprosthetic fracture subtype of Vancouver B2 facture patterns that has been found around a well-fixed femoral stems with a polished, tapered, collarless total hip arthroplasty results in a splitting, fragmentation and spiral fracture of the proximal femur with the cement bone interface remaining intact [23]. The authors note the importance of proper recognition and classification of this fracture pattern as extensive reconstruction is often required for this type of injury. Grammatopoulos et al. [23] reported on a prospectively collected series of 14 patients and described the intra-operative findings. The authors noted the proximal femoral metaphysis was characteristically split into four to six fragments rendering the prosthesis unstable. In addition, well-fixed cement-bone interfaces were routinely noted. The authors noted that all patients required complex reconstructions: nine individuals required a long prosthesis bypassing the fracture by the length of two cortical diameters while five revision cases were performed with a tri-modular stemmed prostheses with the aid of proximal metaphyseal mesh and impaction bone grafting. Type B3
The deficient bone stock present in Vancouver B3 fractures offers the unique challenge of gaining adequate prosthesis stability and fracture fixation. Periprosthetic fractures around a loose femoral stem in the presence of inadequate bone stock or comminution are typically treated with revision of the stem along with bone augmentation through the use of cortical onlay strut allograft and/or allograft-prosthetic composites (APC) [25,35,47–49]. End-stage prosthetic disease may require an endoprosthesis. More recently, modular distally fixed cementless stems have been used with encouraging results [42–44,50,51]. If there is at least 4–6 cm of remaining distal bone stock present, many of the same prostheses and techniques utilized for Vancouver B2 fractures can be used in B3 fracture types with supplementary augmentation of bone stock [52]. Using a modular fluted tapered uncemented stem with preservation of the proximal femoral bone, Berry [50] reported no cases of stem subsidence or evidence of radiographic loosening in seven out of eight patients with Vancouver B3 fractures at a mean follow-up of 1.5 years (range, 1–2 years). The one remaining patient died secondary to a medical-related condition, and was unavailable for follow-up. The authors cited that one of the major advantages of this technique was the reconstitution of proximal femoral bone and avoidance of APCs and endoprostheses. At the latest follow-up, five out of seven patients reported that they had no pain, and none had severe pain. A study by Fink et al. [42], examined fracture union and stem subsidence in 10 Vancouver B3 fractures treated with a modular, non-cemented tapered fluted femoral stem at a mean follow-up of 32 months (range, 24–60 months). The authors reported that there were no cases of stem subsidence and 100% fracture union rate in at the latest follow-up. Interestingly, the authors noted that in cases where patients had
