➢ Postoperative periprosthetic fractures of the femur are likely to become an increasing clinical burden.
➢ The Vancouver classification system is a simple, valid, and reproducible means of classifying these injuries.
➢ The evidence supporting the treatment of these fractures is predominantly derived from Level-III and IV studies only.
Periprosthetic fractures at the site of a total hip replacement may occur intraoperatively or postoperatively and may involve the femur, the acetabulum, or both. The present review concerns only postoperative femoral fractures, with a brief overview of the evidence supporting the treatment of these injuries. It is not feasible to present an exhaustive description of all of the relevant studies regarding this topic, but we have searched widely to determine the levels of evidence to support our treatment recommendations. As will be seen, there is a paucity of Level-I and II research regarding periprosthetic femoral fractures, and additional study is required regarding almost every aspect of their treatment.
It is difficult to assess the true prevalence of periprosthetic femoral fractures following total hip arthroplasty. Most joint registries only record cases in which the fracture pattern necessitates revision surgery. The Swedish Hip Arthroplasty Register is designed to record all periprosthetic fractures and includes codes for internal fixation as well as revision arthroplasty. A review of Swedish registry data from 1979 to 2000 demonstrated that the cumulative prevalence of periprosthetic fracture was 0.4% after primary arthroplasty and 2.1% after revision arthroplasty1.
Vancouver Classification System
The Vancouver classification system2 is the most widely accepted system for categorizing postoperative fractures of the femur occurring around a hip prosthesis (Fig. 1). Studies have demonstrated the reliability and validity of this system when used by both experienced surgeons and junior staff alike3-5. The Vancouver classification system takes into account three relevant surgical factors: (1) the site of the fracture, (2) implant stability, and (3) available bone stock.
Type-A fractures refer to those involving the greater (AG) or lesser (AL) trochanter and may be subclassified as either stable or unstable.
Type-B fractures are those that occur adjacent or just distal to the femoral stem. Type-B fractures are further subdivided into those associated with a well-fixed prosthesis (B1), a loose prosthesis (B2), and a loose prosthesis with poor residual bone stock (B3). The determination of implant stability may not always be possible by radiographic means alone.
Type-C fractures occur distal to the femoral stem.
A study of the Swedish joint registry found that type-B2 fractures were the most common pattern after primary arthroplasty (62%) and that type-B1 fractures were the most common subtype (44%) after revision hip surgery1.
Vancouver type-A periprosthetic fractures involve either the greater (AG) or lesser (AL) trochanter.
Postoperative periprosthetic fractures of the greater trochanter should be identified as either stable or unstable. Those that show little or no radiographic displacement generally may be regarded as stable.
Hsieh et al. retrospectively reported the results of nonoperative treatment of seventeen minimally displaced type-AG fractures that had occurred through osteolytic lesions6. After a period of restricted activity and partial weight-bearing, fifteen of these fractures united whereas two displaced beyond 2 cm.
Displaced fractures of the greater trochanter can represent a difficult problem, causing pain and limitation of function as well as potential instability of the hip (Fig. 2). The pull of the abductor muscles imparts substantial force on any fixation construct.
We are not aware of any Level-I or II studies that have compared fixation techniques used for the treatment of type-AG fractures. A number of smaller, largely retrospective studies have examined the outcomes associated with various fixation constructs. While monofilament wire fixation historically has been the most common treatment method for these injuries, there is now increasing use of cable and claw-plate systems.
Wang et al. described the results of wire fixation and allogeneic bone-grafting for the treatment of fractures that had occurred through an osteolytic lesion of the greater trochanter7. Patients were managed with an abduction orthosis for three months. Union was achieved in eighteen of nineteen patients at a mean of five months.
Zarin et al. reported the combined results of claw-plate fixation for the stabilization of Vancouver type-AG periprosthetic fractures as well as intraoperative fractures and osteotomy sites8. Union was achieved in twenty-eight of thirty-one patients, although six patients had plate-related complications. McGrory and Lucas reported successful results in association with the use of a modified tibial locking plate for the treatment of three postoperative type-AG fractures9.
Whiteside et al. described a salvage procedure involving the transposition of a partial gluteus maximus flap into the defect between a greater trochanteric fragment and the lateral cortex of the femur10. The functional results for these patients were compared with those for patients in whom the greater trochanter was left unrepaired as well as with those for patients who had undergone excision of the trochanteric fragment. The patients with gluteus maximus transfers were found to have less pain and a more normal gait than those in the other two groups.
No definitive treatment has been widely accepted for the treatment of type-AG periprosthetic fractures. In our practice, fractures with <1 cm of displacement are treated nonoperatively, provided that they are not contributing to instability of the hip. Treatment typically involves partial weight-bearing with crutches and the avoidance of active abduction for at least three months.
For fractures that are displaced beyond 1 cm and those that are associated with painful nonunion, we favor fixation with use of cable hook plates. These fractures frequently occur through areas of osteolysis, and morselized allograft is used to concurrently fill any contained defects. Even in the absence of a greater trochanteric fracture, substantial greater trochanteric osteolysis may arise in association with a failed acetabular component and must be addressed as part of the overall surgical management. Postoperatively, the patient is managed with partial weight-bearing with crutches and the avoidance of active abduction for at least three months. It is important to note, however, that the use of hook plates can be associated with a substantial rate of complications, including soft-tissue irritation, which may necessitate subsequent plate removal8.
Isolated periprosthetic fractures of the lesser trochanter are uncommon1. These fractures frequently are seen in patients with preexisting osteolysis about the femoral stem, and therefore they may herald loosening of the prosthesis. Radiographs should be carefully inspected to exclude fracture extension, which may compromise implant stability.
We are not aware of any clinical studies investigating the treatment of Vancouver type-AL fractures. In our opinion, truly isolated periprosthetic fractures of the lesser trochanter do not require operative repair. The patient is managed with analgesia and activity modification until the pain abates.
Larger fractures of the lesser trochanter may involve the medial buttress and may result in implant loosening. These fractures also may result from implant subsidence, particularly in patients with tapered stems that have not shown osseous ingrowth in the femur. In such cases, we perform cerclage wire fixation of the fracture and revise the stem with a tapered, fluted prosthesis that engages the isthmus.
Vancouver type-B1 periprosthetic fractures occur around or just distal to the stem of a femoral prosthesis, with the implant remaining well fixed. The radiographic distinction between type-B1 and B2 fractures occasionally can be difficult. As such, the surgeon must be prepared to undertake either internal fixation or stem revision on the basis of intraoperative findings.
Type-B1 injuries require fracture reduction and internal fixation. However, there are a number of controversies regarding the optimum treatment of these injuries, particularly in relation to the use of allograft struts, cable systems, and minimally invasive techniques. Unfortunately, there is a lack of randomized studies examining these variables, and most available information is from retrospective case series.
Potential fixation options for type-B1 fractures include single or double locking plates, cable plate systems, and cortical allograft struts. Numerous biomechanical studies have been performed in an effort to identify the optimum fixation construct.
Lever et al. performed a cadaveric biomechanical study to compare three different screw-plate and cable-plate systems for the fixation of periprosthetic fractures located near the tip of a femoral prosthesis11. Screw-plate systems provided better mechanical stability than cable-plate systems, but the differences did not reach significance.
Wilson et al. performed a cadaveric study to compare the mechanical properties of cable plates, strut allografts, and combined plate-allograft fixation for the treatment of simulated transverse type-B1 fractures12. The most rigid fixation was achieved with use of a cable plate with an allograft. The cable plate used in isolation was particularly poor for resisting rotational forces.
Choi et al. created a model of a comminuted type-B1 fracture to test the mechanical stiffness of three different constructs and recommended dual orthogonal plating or a plate-allograft construct over single locking plate fixation13.
Although laboratory studies can yield useful information regarding fixation mechanics, caution must be taken when attempting to apply these findings directly to the clinical setting. Laboratory studies cannot simulate the biological effect of the additional soft-tissue stripping that is invariably associated with dual-fixation constructs.
Cortical struts are used with the aim of providing initial mechanical stability but also to improve bone stock through osseointegration of the graft in the longer term. Haddad et al. performed a multicenter study of forty patients to investigate the use of allograft struts (either alone or in conjunction with plate fixation) for the treatment of type-B1 fractures14. The authors reported a 98% union rate and advocated the routine use of allograft struts to augment plate fixation of type-B1 fractures. Khashan et al., in a retrospective analysis of twenty-one patients, reported superior outcomes in association with strut allograft augmentation in comparison with plate fixation alone for the treatment of type-B1 and C fractures15. In contrast, other investigators have reported excellent results in association with the use of isolated plate fixation16,17.
There is increased interest in applying minimally invasive fixation techniques to the treatment of type-B1 periprosthetic femoral fractures. This approach reduces soft-tissue dissection, preserving the blood supply about the fracture while also decreasing surgical morbidity. The concern regarding minimally invasive methods surrounds the ability to obtain an anatomical reduction as well as the difficulty of achieving adequate proximal fixation without the supplementary use of allograft struts or cables. A number of small studies have demonstrated 100% union rates in association with the use of truly minimally invasive techniques for the fixation of type-B1 and C fractures18-20. Kobbe et al. reported on sixteen such cases and documented two plate failures after the initiation of full weight-bearing within four weeks after surgery20. Both plates were revised, and both fractures subsequently united.
Newer polyaxial locking plates enable the surgeon to angle screws about an in situ stem, potentially allowing bicortical fixation in cases in which traditional systems would necessitate the use of either unicortical screws or cables. Polyaxial systems therefore may largely obviate the need for cable fixation and the additional soft-tissue stripping that is required. Although two recent studies investigated the use of such implants for a range of periprosthetic femoral fractures21,22, there is little published information regarding their use specifically for type-B1 fractures.
We recommend the fixation of type-B1 fractures with a single long locking plate. If possible, this plate should be applied with use of minimally invasive methods. If a stemmed total knee replacement is in situ, we always span this implant with a longer plate. If there is poor bone stock, we supplement the plate fixation with use of a cortical strut graft, although the use of a cortical strut necessitates a more invasive procedure.
Vancouver type-B2 periprosthetic fractures occur around or just distal to the stem of a femoral prosthesis, with loosening of the implant but maintenance of proximal bone stock (Fig. 3, A). In a study of the Swedish Joint Registry, 61% of periprosthetic femoral fractures that had occurred around primary hip replacements were of a type-B2 pattern1.
The presence of a loose stem following a periprosthetic fracture requires revision of the femoral prosthesis. Surgeons have a range of options for the treatment of these fractures, including uncemented porous-coated cylindrical stems, tapered fluted implants, and cemented stems.
Many clinical studies have combined the results of revision surgery for type-B2 fractures with those for selected type-B3 fractures in patients with a well-preserved isthmus. Munro et al., in a study of forty-six type-B2 and B3 fractures that were treated with a tapered fluted modular stem, observed a 96% rate of implant survival at a mean of fifty-four months of follow-up23. One nonunion was observed in that study. Abdel et al., in a study of forty-four type-B2 and B3 fractures that were treated with this type of prosthesis, reported a 98% rate of union at a minimum of two years of follow-up24. The authors recorded an 11% rate of dislocation and suggested that instability may be reduced in association with the use of larger-diameter femoral heads.
Neumann et al. used a curved, modular tapered uncemented stem for the treatment of fifty-three type-B2 and twenty type-B3 fractures that were reviewed at a mean of sixty-seven months25. All fractures united within six months. Two hips required revision with a larger-diameter implant following subsidence of >5 mm that had become associated with pain.
Although the majority of type-B2 fractures are treated with a long-stemmed uncemented component, some investigators have described the use of cemented revision components for highly selected subgroups of patients26,27. The theoretical concern with this approach has been that cement may be extruded into the fracture site, preventing biological union. With meticulous attention to detail, cement extrusion may be avoided.
Richards et al. described favorable short-term outcomes following cement-in-cement revision for a highly selected subgroup of patients with type-B2 fractures26. The appeal of this approach is that it is fast and requires less instrumentation of the femur, thereby limiting surgical morbidity in a population of patients who are generally elderly. In that study, the technique was reserved for low-demand individuals with a simple fracture pattern and a cement mantle that remained well attached to the constituent fragments. Provided that an anatomical fracture reduction could be achieved, a cement-in-cement revision was then performed. The stem was inserted late, such that the cement was in a more viscous state and was less likely to infiltrate the fracture site. This approach to periprosthetic fractures is certainly not consistent with the usual principles of cement-in-cement fixation and should be used with caution and only in low-demand individuals with a limited life expectancy.
Corten et al. reported on a series of type-B2 fractures that were revised with a long-stemmed cemented prosthesis27. Again, this approach was reserved for a specific group of patients who were anticipated to have a more limited life expectancy as determined on the basis of age and comorbidities. A total of thirty-one patients were managed in this manner, with additional allograft or plate fixation being used in most cases. While this approach may not necessarily have reduced the surgical morbidity of the procedure when compared with a revision without cement, the aim of the authors was rather to allow immediate weight-bearing in this carefully selected population. Of the thirty-one patients who were involved, only sixteen were alive more than one year after surgery, and none of the implants had been revised after a mean duration of follow-up of forty-six months.
Our preferred approach for the treatment of type-B2 fractures is to revise the femur with a modular tapered fluted titanium prosthesis that achieves fixation in the diaphysis. The tapered design allows for a press fit in the intact diaphysis, allowing implant stability regardless of the fracture. Modularity allows intraoperative customization of the proximal part of the segment for fine adjustments of limb length and offset once the distal portion of the implant has obtained adequate stability. An experienced surgeon can consider the use of a nonmodular implant, thereby avoiding the potential drawbacks of the modular design, such as taper corrosion and fracture (Fig. 3, B).
Our technique is to utilize the fracture site to remove the existing prosthesis and any associated cement. In some cases, an additional osteotomy may be required to facilitate cement and implant removal. Care is taken to preserve soft-tissue attachments. Before fracture reduction, the distal portion of the femur is prepared to accept the modular implant. A prophylactic wire or cable is applied to prevent distal propagation of the fracture during reaming. The proximal body of the trial implant is then assembled to verify adequate stability and limb length. Once stability is confirmed, the definitive implant is constructed and implanted. The fracture fragments are then assembled around the stem and are fixed with cables. When feasible, we favor the use of a 36 or 40-mm femoral head as a means of optimizing stability. We do not routinely use supplementary allograft struts for type-B2 fractures. The patient maintains touch weight-bearing for six to twelve weeks postoperatively.
Vancouver type-B3 fractures are uncommon, representing <5% of periprosthetic femoral injuries in the Swedish registry1. These fractures occur around or just distal to the stem of a femoral prosthesis. They are characterized by loosening of the implant as well as by the presence of poor proximal bone stock. This compromised bone stock may be the result of osteolysis or implant loosening prior to the periprosthetic fracture but also may be the end result of extensive fracture comminution. In either case, the important concept is that, without adequate proximal bone, metaphyseal-fitting prostheses cannot be employed for the treatment of type-B3 fractures. Four principal treatment options remain: (1) use of a fluted tapered titanium stem with or without allograft, (2) impaction grafting with a cemented stem, (3) use of an allograft-prosthesis composite, and (4) use of a proximal femoral prosthetic replacement.
The degree of bone loss associated with type-B3 fractures represents a continuum, and the age, comorbidities, and anticipated functional status of the patient also must be considered when planning a treatment strategy. For patients with mild bone loss, our preferred approach is to insert a modular fluted tapered titanium stem in a manner similar to that described for type-B2 fractures. Impaction grafting with the use of a cemented stem may be a viable option in selected cases, but we have largely moved away from this technique because of the superior versatility and efficiency of modular fluted tapered titanium stems. We now avoid long cylindrical stems as these have been associated with substantial stress-shielding, which further worsens the poor bone stock28. For patients with severe bone loss, an allograft-prosthesis composite or a proximal femoral replacement may be required.
Revision with a Press-Fit Uncemented Stem
For many type-B3 fractures, a long uncemented stem may be used to bypass both the fracture and the zone of compromised proximal bone, allowing stability to be achieved distally. For this treatment strategy to work, the stem needs to be firmly engaged over at least 4 cm of intact diaphysis29,30. As outlined in the discussion of type-B2 fractures, we favor the use of tapered, fluted titanium stems when the host bone allows. If a modular stem is to be employed, it is imperative that the modular junction be supported by sufficient host bone in order to avoid excessive implant fatigue and eventual breakage at this interface. For this reason, allograft struts are occasionally employed for the treatment of type-B3 fractures. The surgical technique is identical to that described for type-B2 fractures. In many cases, an anatomic reduction of fracture fragments is impossible because of bone loss or comminution. The remaining bone stock is approximated as closely as possible around the stem, which effectively acts as a scaffold. When necessary, a cortical strut is inserted either medially or laterally, as needed, to support the modular junction.
Clinical studies on the use of long uncemented stems for the treatment of periprosthetic fractures generally group together the results of treatment of type-B2 and selected type-B3 fractures. Four recent studies demonstrated favorable short to intermediate-term outcomes in association with the use of tapered modular stems for the treatment of a combined sixty-seven type-B3 fractures (Table I).
An alternative technique for the treatment of type-B3 fractures is the use of impaction grafting followed by the placement of a cemented stem. This approach offers the potential to recreate bone stock. Impaction grafting involves the use of morselized frozen allograft, which is sequentially packed against the remaining endosteal surface, and creates an allograft “neo-endosteum” into which a cemented component can then be placed. Retrieval studies have demonstrated that this graft has the capacity to partially revascularize and be replaced by autologous host bone, albeit to a variable degree31-33. While this approach is appealing for its potential to recreate bone stock, it is a technically demanding and time-consuming procedure. This technique has largely fallen out of favor in North America.
Tsiridis et al. retrospectively reviewed the outcomes for 106 patients who underwent revision with cement for the treatment of postoperative type-B2 and B3 periprosthetic fractures34. The authors reported an 88% union rate for patients who were managed with a long-stemmed revision component with impaction grafting. Those results were greatly superior to those for patients who were managed with short stems or those in whom a revision with cement was performed without any impaction grafting. Lee et al. also reported excellent results in association with the use of impaction grafting for the treatment of type-B2 and B3 fractures in which either the fracture configuration or the degree of bone loss at the isthmus precluded the use of uncemented press-fit components35.
Allograft Prosthetic Composites and Megaprostheses
With some type-B3 fractures, the femoral bone stock is so severely compromised that it requires replacement with either an allograft-prosthesis composite or a megaprosthesis.
Allograft-prosthesis composites generally have been reserved for younger patients with extreme femoral bone deficiency extending to the isthmus. This demanding technique requires an allograft femur, cut to length, into which a prosthesis is cemented. The allograft-prosthesis composite is then telescoped into the remaining distal portion of the femur or is secured at the junction with a step-cut interface and cables. Usually, the implant is not cemented into the host bone. Supplementary stability is achieved with use of autograft struts overlapping the junction between the allograft-prosthesis composite and the host bone. These struts are obtained from the resected proximal host bone and should not be stripped of soft-tissue attachments. If a substantial lateral fragment remains in continuity with the greater trochanter, it is also wrapped around the allograft and is firmly secured to it in order to improve abductor function postoperatively. The use of bulk allograft-prosthesis composites is a highly demanding procedure and, although it was commonly used at our institution ten to fifteen years ago, the versatility of the modular fluted tapered titanium stems has made the use of bulk allograft-prosthesis composites almost obsolete.
Maury et al. performed a retrospective study of twenty-five type-B3 fractures that had been treated with an allograft-prosthesis composite36. After a mean of 5.1 years of follow-up, twenty allograft-prosthesis composites had united and twenty-one patients reported little or no pain.
Proximal femoral replacement is an alternative option for the treatment of type-B3 fractures that are associated with severe bone loss (Fig. 4). These implants generally are reserved for low-demand, elderly patients. Klein et al. reported on twenty-one type-B3 fractures that were treated with megaprostheses37.The average age of the patients was 78.3 years. After an average duration of follow-up of 3.2 years, all patients but one were able to walk and had little or no pain. The authors reported a high complication rate and concluded that, given the risk of instability, there should be a low threshold for the use of constrained acetabular liners during these procedures. We continue to use this technique for low-demand, elderly patients with a limited life expectancy as it allows early mobilization with relatively low surgical morbidity.
Type-C fractures are those that occur well below the tip of the femoral prosthesis. These fractures are treated with use of standard principles of reduction and stabilization. The same techniques that are used for the treatment of type-B1 fractures may be applied, although more proximal bone is available for the fixation of type-C fractures. The available research is chiefly in the form of retrospective case series examining the results of plate fixation15,20,38. Froberg et al. reported the results of locking plate fixation for sixty consecutive type-B1 and C fractures38. Three patients required revision surgery because of loss of fixation following low-energy falls. In each of these cases, the original plate was noted to have spanned <50% of the femoral stem.
For type-C periprosthetic fractures, we recommend that an anatomical reduction be achieved, followed by fixation with use of a single lateral locking plate. If a stemmed total knee replacement is also in situ, it should be spanned with the plate.
Table II summarizes our recommendations for the treatment of periprosthetic fractures of the femur that occur after total hip arthroplasty and provides a grading of the current published evidence supporting each recommendation.
Source of Funding: No external funds were received in support of this study.
Investigation performed at the Department of Orthopaedics, University of British Columbia, Vancouver, British Columbia, Canada
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2014 by The Journal of Bone and Joint Surgery, Incorporated