Fractures around joint replacement prostheses are commonly called periprosthetic fractures, whereas fractures around plates, rods, or prostheses can be more generally termed peri-implant fractures. Nonprosthetic peri-implant fractures may be considered to represent a separate entity.[1] Interprosthetic fractures are fractures that occur between two prostheses or implants (usually a femur fracture between a total knee replacement and a total hip replacement.)
More than 300,000 hip fractures occur yearly, and almost all are treated operatively with internal fixation or prosthetic replacement. The number of orthopedic implants placed in the femur is increasing.[2] Furthermore, the number of implants placed in other bones is increasing, with expanding indications for shoulder, elbow, and ankle replacement, and internal fixation continues to be used in all the long bones, especially the tibia and the humerus.
More than 123,000 total hip arthroplasties (THAs) and 150,000 total knee arthroplasties (TKAs) are completed each year in the United States, with the numbers expected to increase as the population ages.[3] Additionally, the indications for joint replacement have expanded significantly over the past few decades to include much older and younger patients, both of which have a higher risk of periprosthetic fracture.[4]
As the number of implants placed and life expectancy increases, it is inevitable that associated fractures also become more common.[5]
The major complications of total joint arthroplasty are loosening and osteolysis. The rate of osteolysis increases with time, and osteolytic bone defects are stress risers, which predispose the patient to fractures.[6] The use of computer-assisted navigation and robotic-assisted surgery in joint replacement requires the use of tracking pins placed in the femur and tibia. Periprosthetic fractures through pin sites have been reported infrequently, but as the frequency of computer-assisted techniques increases, the incidence of these fractures will likely also increase.[7]
Mayo Clinic and Swedish registry studies found that 94% and 70% of patients with periprosthetic fracture had a loose stem before fracture.[8] The rate of periprosthetic fracture around primary total hip replacement is about 1%, with a 20-year fracture probability of 3.5%. Rates of periprosthetic fracture after revision hip replacement range from 1.5% to 7.8%, with a 20-year fracture probability of 11%. The number of periprosthetic fractures is expected to rise by 4.6% per decade over the next 30 years.[9] Abdel et al reported a 12% incidence of intraoperative fracture in revision total hip replacement.[10]
Improvements in cancer treatment also have resulted in longer life spans with increased likelihood of metastatic bone lesions and impending or actual pathologic fractures that require internal fixation. The ability of tumor to "outgrow" a fixation device and the reduced ability of irradiated or tumor-replaced bone to heal fractures also result in an increasing frequency of peri-implant fractures.
Fractures around implants pose unique fixation challenges. The original placement of the implant may predispose to later fracture, the long-term presence of the device may change the structure of the bone and increase susceptibility to fracture, and the implant itself may interfere with healing or the placement of other fixation devices.[11, 12] Common problems include malalignment, stiffness, and nonunion.[12] If malalignment occurs after a periprosthetic fracture, the abnormal joint biomechanics may cause a high rate of revision secondary to loosening.[11]
The implant may impair fracture healing because of endosteal ischemia.[13] Rates of nonunion for supracondylar fractures proximal to total knee prostheses are higher than those for supracondylar fractures without the implant.[13] Stems, rods, screws, and methylmethacrylate may block the medullary canal, preventing intramedullary fixation of fractures. Stems and rods also block screw fixation through the medullary canal to hold plates on bone. The techniques for treating peri-implant fractures may be more difficult, with more limited options and more frequent complications than the techniques used in treating fractures without the presence of an implant.
Patients with inflammatory arthritis (including rheumatoid arthritis [RA]) have a higher incidence of periprosthetic fracture due to associated osteoporosis and to bone erosion and defects resulting from the simultaneous activity of osteoclasts and inflammatory cytokines.[14] Early periprosthetic femur fracture is also associated with an age of 75 years or greater as an independent risk factor.[15]
Treatment of periprosthetic fractures requires strict adherence to the basic principles of treating any fracture. The surgeon must restore the biomechanical integrity of the bone. This requires restoration of a biologic environment in which the bone can heal and a mechanically stable construct that supports the bone to give it a chance to heal.
Biology is maintained by strict soft-tissue and indirect reduction techniques, when possible, to preserve the periosteal or endosteal blood supply. The surgeon should minimize periosteal stripping, avoid dead space, and consider bone grafting if the biologic environment is compromised. The patient's medical condition should be optimized. A patient who smokes should be encouraged to stop.
Mechanical stability is obtained by restoring the anatomic integrity of the bone and by following Arbeitsgemeinschaft für Osteosynthesefragen (AO)/Association for the Study of Internal Fixation (ASIF) principles with adequate fixation distal and proximal to the fracture.
The rate of intraoperative fracture may be lessened through careful preoperative planning and templating.[16]
Peri-implant fractures can result from the same forces that cause fractures without an implant present, but in addition, they can be caused by factors specifically related to the placement of the implant or the presence of the implant. Osteoporosis, medications, and medical comorbidities all contribute.
Of concern is the potential for low-energy subtrochanteric fractures related to long-term use of bisphosphonates to decrease the risk of fracture in postmenopausal women and others with decreased bone mass. Bisphosphonate-associated atypical femur fractures have been identified in periprosthetic hip fractures around THAs.[17, 18] Atypical femoral neck fractures have also occured in patients with RA and associated osteoporosis treated with bisphosphonates.[14]
Osteoporosis, inflammatory arthritis, corticosteroid use, increased age (>80 y), decreased age (< 60 y), female sex, and previous revision arthroplasty (especially if previously complicated by infection or fracture/nonunion) all increase the risk of periprosthetic fracture.[4, 15] Loosening of implants, RA, Paget disease, tumors, polyneuropathies, extruded cement, and varus stem position increase the risk as well.[19] A meta-analysis by Bissias et al found that female sex, RA, uncemented technique, and revision arthroplasty were risk factors, but obesity, advanced age, and poor general health were not.[14] Osteoarthritis was associated with a decreased incidence of periprosthetic fracture.
The most frequent mechanism of injury is a low-energy fall, causing 75% of primary total hip periprosthetic fractures and 56% of revision total hip periprosthetic fractures.[8]
Uncemented implants develop periprosthetic fracture three times more often than cemented implants, probably because of stabilization of weak bone by the increased mechanical strength of the cement. Additionally, failure of bone ingrowth or microfracture during insertion may predispose to later fracture.[14]
Peri-implant fractures can be caused by technical problems during their placement. Many studies have implicated notching of the anterior cortex of the femur during knee arthroplasty as the cause of supracondylar fracture,[13] with a 40% fracture rate even 8 years after surgery.[11] Other studies, however, have questioned the association.[20] Periprosthetic fractures occur significantly closer to the prosthesis (leaving less room for fracture fixation) in patients with anterior notching than in patients without notching.[21]
Different implants have different risk for peri-implant fracture. The Fakler classification of distal femur periprosthetic fractures attempted to classify and predict fracture characteristics according to implant type.[22] There is a significantly lower rate of peri-implant fracture for hip implants with collared stems and "fit-and-fill" designs.[23] There is no statistical difference between short and long cephalomedullary nails with regard to the rate of peri-implant fractures.[24]
The calcar may fracture during hip arthroplasty,[25, 26] the stem may penetrate the femoral shaft, or distal femoral fracture can occur with manipulation and preparation of the femur (see the image below).[27, 26]
An increased risk was reported by Hartford and Knowles in hip replacements using the direct anterior approach in female patients, those with morbid obesity (body mass index [BMI] >40), those with smaller implants, and those with an increased Dorr (calcar-to-canal) ratio.[10] Sershon et al confirmed the increased risk in females but also found that low BMI (< 20) and age greater than 65 years doubled the risk of periprosthetic fracture. In their review, surgical approach did not influence fracture rate.[23] Anatomic implants may decrease the risk of fracture during insertion of stems.[14]
Intraoperative fracture occurs 14 times more frequently in patients with uncemented stems, especially in female patients older than 65 years.[28] Increased use of cementless acetabular cups is predicted to increase the occurrence of future periprosthetic fractures of the acetabulum after THA.[29]
A study by Sappey-Marinier et al compared dual-mobility (DM) cups (n = 73) with single-mobility (SM) cups (n = 53) to assess their relative effects on the incidence of dislocation and periprosthetic fracture in the setting of revision THA.[30] They found that the use of DM cups increased hip stability and reduced dislocation as compared with SM cups but also led to a higher rate of periprosthetic fracture through load transfer on the femur. Further studies with a larger cohort and longer follow-up would be needed to confirm these findings.
Risk factors for cortical perforation during revision total hip replacement include "shorter [patient] stature, proximal location of the femoral isthmus, narrow femoral canal, and smaller radius of curvature" of the anterior bow of the femur.[16] In revision hip replacement, the rate of periprosthetic fracture was three times higher with uncemented stems.[31]
Fractures can occur during internal fixation when screws are placed too close or bone-holding devices crack the bone, especially in osteoporotic bone. Any drill hole up to 20% of the diameter of the bone weakens the bone by 40% of its original strength. Some 90% of fractures around fracture fixation implants occur through a drill hole (see the image below).[32]
Displacement of unrecognized femoral neck fracture or new fracture occurs in 3% of intramedullary nailings of femoral shaft fractures.[33, 34] With any implant, the end of the device becomes a stress riser in which the weaker osteoporotic bone tends to fracture first when excessive load is applied.[32]
Removal of devices is also associated with refracture. After plate removal, the cortical bone has been stress-shielded and needs to be protected. Zickel intramedullary hip nails have been associated with subtrochanteric fracture when removed,[32] and the more modern intramedullary hip screw systems may do the same. During prosthetic revisions, the rate of fracture is 17.6%, compared with 3.5% for primary procedures; osteoporotic bone or bone with osteolytic defects may fail while the prosthesis or its cement is being removed.[26]
The incidence of supracondylar fracture after total knee replacement is in the range of 0.3-2.5%.[13, 11, 4] Fracture can occur more than 10 years after joint replacement[35] ; thus, as the number of patients with replacements accumulates, more fractures occur. In data from the Mayo Clinic Joint Registry, the incidence of periprosthetic fracture after primary total hip replacement was 1.1%, and it was 4% after revision total hip replacement. Periprosthetic fracture after total hip replacement may be the second leading cause of revision, after aseptic loosening.[36]
The exact incidence and frequency of other peri-implant fractures have not been established.
A good outcome and a favorable prognosis are expected if the surgeon restores the biomechanical function of the limb. Failure to do so results in a poor outcome.[37] However, even with improved techniques and implant designs, these fractures remain challenging.
For example, Hoffmann et al, in a review of periprosthetic femur fractures proximal to a total knee implant, showed that only 69.4% healed after the initial surgery, with 8% experiencing hardware failure.[21] Nonunion rates were lower with submuscular plate insertion than with an extensive lateral approach. Range of motion (ROM) was reduced in most patients, and 13.5% had a 5° extension lag. In 23% of patients, more than mild pain was reported. Similarly, Hou et al reported only a 75% rate of union after periprosthetic fractures around total knee implants.[38]
For periprosthetic femur fracture after THA, Holley et al reported that only 74% of patients healed after the initial surgical treatment (a 26% nonunion rate) and only 84% were healed even after additional surgical interventions (a 14% nonunion rate).[36] The complication rate was 29%. Brand et al summarized the literature, reporting that only 48% of patients regained their prior walking status.[19]
When treating periprosthetic fractures, the surgeon must evaluate the stability of the implant carefully. Loose implants used for fixation allow motion at the fracture site that hinders healing and physically interferes with the placement of more stable fracture fixation. Loose prostheses used for joint replacement are painful and interfere with adequate fracture fixation.
If the implant is loose or malaligned, the implant should be revised while the fracture is fixed at the same setting. If the implant is stable and sufficient bone stock is available for fracture stabilization, the implant should be retained while the fracture is fixed according to standard treatment principles. When treating peri-implant fractures of the femur, the surgeon should have a flexible approach, using the best-fitting device, following basic fracture principles of rigid internal fixation and restoration of the anatomy and preservation of soft-tissue attachments.
For optimal results in treating periprosthetic fractures, the following are essential[39] :
By definition, all patients with a periprosthetic fracture will have a previous history either of joint replacement or of internal fixation for a fracture. The term peri-implant fracture applies to periprosthetic fractures (ie, fractures around a joint replacement) while also including fractures around an orthopedic fixation device.
The patient's acute history will usually include a traumatic event (as with any nonperiprosthetic fracture) or may present more insidiously with gradual increasing pain secondary to a stress fracture around the implant. Intraoperative periprosthetic fractures may be the result of difficulty preparing for the placement of the implant (eg, overreaming the acetabulum before placing a total hip prosthesis) or difficulty placing the prosthesis (eg, a calcar fracture during implantation of the femoral component of a hip replacement when there is a mismatch between the size of the prosthesis and the area prepared for it).
Periprosthetic occult acetabular fractures are common during press-fit acetaular insertion in primary hip replacement and may not be recognized during surgery. Surgeons should have a high index of suspicion, especially in male patients who have unexplained early postoperative groin pain after cementless acetabular placement.[40]
A prefracture history of thigh pain with weightbearing or decreased mobility at the hip (for femoral periprosthetic fractures) predicts a high likelihood that the stem is loose. A periprosthetic fracture around an uncemented prosthesis that occurs within the first few weeks of implantation should be presumed loose because of insufficient time for the stem to become integrated. This diagnosis of a loose stem is important; it may not be apparent on imaging, but it is crucial in determining classification and best treatment.[5]
The physical examination is essentially the same as for any fracture. Patients exhibit the usual signs of fracture and have a history of a previous prosthesis or implant. There is tenderness, swelling, and instability (usually) at the fracture site. There may be a limb-length discrepancy and deformity, and the patient may be unable to use the limb. The fracture can occur with minimal trauma (especially with a previously loose prosthesis or osteoporotic bone) or an obvious traumatic incident.
The examiner should check neurovascular status, evaluate for compartment syndrome, and identify any open wounds.
Diagnosis of periprosthetic fractures is generally straightforward. As with any fracture, there is usually a traumatic event, and the trauma is the cause of the fracture.
One cause of fractures around an implant without a specific traumatic event is a stress fracture. On the tibial side of total knee arthroplasties (TKAs), both Ozdemir et al[41] and Fonseca et al[42] reported stress fractures, whereas Wada et al[43] reported stress fractures on the femoral side. Similarly, stress fractures around the femoral stem of hip replacements have been reported by Eschenroeder et al[44] and by Lotke et al,[45] with insufficiency fractures of the acetabulum also reported by Kanaji et al.[46]
Stress fractures around fixation implants can also occur, as reported by Nagoshi et al[47] after internal fixation of a radial shaft with a titanium plate.
Another cause of fracture around an implant without a specific traumatic event is the atypical insufficiency fracture related to bisphosphonate use.[48, 49, 50, 51]
No special laboratory studies are required for most periprosthetic fractures. A sedimentation rate and a complete blood count (CBC) with differential are useful if infection is suspected.
Radiographs of the entire bone are required to assess the condition of the joint above and below the fracture, the condition of the implant, the presence of any deformity or lesions that may influence surgical options, and the axial alignment of the bone. Two views perpendicular to each other, most often an anteroposterior (AP) view and a lateral view of the bone, are always required.
Careful attention to the condition of the cement mantle (for cemented prostheses) is important for determining the stability of the prosthesis.[8, 5] Stem subsidence and osteolysis imply loosening.[5]
Standard computed tomography (CT) and magnetic resonance imaging (MRI) are of limited utility because of scatter artifact caused by the metallic implant; however, CT with metal artifact reduction algorithms can identify subsidence and osteolysis more reliably than standard x-rays can.[5] CT can also recognize occult acetabular fractures.[40]
Bone scans are not specific.
For uncemented stems, the location of the periprosthetic fracture predicts loosening. Different prosthetic designs achieve fixation at different sites on the implant and bone. For example, in straight stems that have proximal fixation, the stem will be loose with a metaphyseal fracture, whereas in tapered stems with distal fixation, the stem will be loose with a diaphyseal fracture. This diagnosis of a loose stem is important, in that it may not be apparent on imaging but is crucial in determining classification and best treatment.[5]
For cemented stems, the type of prosthetic design determines stability. Shape-closed (eg, composite-beam) designs have rigid fixation at the stem-cement interface that is lost with any disruption of the cement mantle. Force-closed (eg, polished-taper) stems maintain stability because the stem subsides a short distance within the cement mantle and no bond is required between the stem and cement. Therefore, the stem is considered loose in these prostheses only when the cement-bone interface is disrupted but is deemed well fixed if the cement-implant interface is preserved.[5]
Aspiration of a failed joint replacement may help if infection is suspected.
Biopsy at the time of surgery is indicated if pathologic fracture or infection is suspected.
There are different classification systems for different fracture sites, but the Unified Classification System discussed below now provides a standardized classification for all bones and implants.
For periprosthetic fractures around a total hip replacement, the most commonly used system has been the Vancouver classification of Duncan and Masri,[52] which provides information concerning the site of fracture, the stability of the prosthesis, and the quality of the bone. This classification specifies the following types:
The Vancouver classification has high reliability, validity, and usefulness in guiding treatment, according to most authors.[53, 14] However, data from a Swedish hip registry showed that as many as 47% of stems thought to be stable were found to be loose at surgery; these findings underscored the need for a careful preoperative evaluation.[5]
The Unified Classification System has been suggested for pelvis and femur fractures around a total hip replacement[54] and has been expanded to include additional sites in both the upper and the lower extremities as an extension of the Vancouver system.[55, 56] The fractured bone is specified by its Arbeitsgemeinschaft für Osteosynthesefragen (AO)/Orthopaedic Trauma Association (OTA) code number, the joint involved is specified as a modifier in square brackets, and the fracture type is then identfied on the basis of its location relative to the implant.[56]
This system specifies the following types[19, 54, 55, 56] :
As an example of the Unified Classification System, a spiral fracture around a femoral implant of a total hip arthroplasty (THA) with loosening of the implant but good bone stock would be coded 32A1[IVB2].[56] An advantage of this system is that it is phrased in such a way that it can be applied to other prostheses besides total hip replacements.[19]
De Meo et al and Fan et al reported that the Unified Classification System is as reliable and valid as the Vancouver classification for periprosthetic femoral fractures.[55, 57]
Videla-Cés et al proposed a classification system for peri-implant femoral fractures, using nomenclature similar to that of the Vancouver classification summarized above.[58] Peri-implant fractures were classified according to whether the implant was a nail, a screw, or a plate, as well as according to the location of the fracture in relation to the original implant and the affected femoral segment. Further studies would be required to determine the utility and applicability of this system.
Classifying distal femur periprosthetic fractures has proved challenging. Makaram et al evaluated seven different classification systems in an attempt to find the most reliable system that could also most accurately predict the type of surgical intervention required.[59] They found that the Fakler classification (see below) had the highest interobserver agreement and most accurately predicted which fractures required replacement instead of fixation. The Rorabeck classification had the second highest reliability and accuracy; the Neer classification of distal femur perioprosthetic fractures had the poorest.
The Fakler classification was proposed by Fakler et al in 2017 as a system that incorporated the type of implant used.[22] In this classification, the type of implant is specified as follows:
The location of the fracture is specified as follows:
The Lewis-Rorabeck classification, initially proposed in 1998, divides fractures into the following three types[60] :
Essentially all periprosthetic fractures require some treatment. Stable nondisplaced fractures may only require protected weightbearing or cast/brace immobilization (and pain medication), but most unstable peri-implant fractures require surgical stabilization, implant replacement, or both to restore function. Surgical intervention follows the same guidelines for peri-implant fractures as for other fractures. The goals of treatment include the following:
Treatment is rarely contraindicated after a periprosthetic fracture. Observation of a fracture in a paralyzed limb may be indicated, but even then, surgery is often useful for helping with nursing care. Cancer patients with widespread resistant metastases also may be treated better with hospice or pain control alone. Patients with unstable medical conditions should be in optimal condition before surgery. If an associated infection exists, its treatment should be part of the surgical plan. Peri-implant fractures usually occur in elderly patients, and a team approach is often required for treatment.
Current efforts to treat periprosthetic fractures focus on ways to avoid the fracture and new implants for improved fixation. Newer designs of replacement prostheses include changes in the shape of stems designed to share load better with the bone and to avoid the osteoporosis of stress shielding, which weakens the bone and predisposes for fracture. Newer plate designs, such as the low-contact dynamic compression plate, decrease the contact area of plates and decrease the osteoporosis of stress shielding.
Changes in materials decrease bone destruction from osteolysis. Less rigid metals (eg, titanium vs stainless steel) share the load better. Fixed-angle plate systems (eg, less invasive surgical stabilization [LISS]) allow more stable fixation with minimally invasive techniques.
Minimally invasive techniques are also being developed that may improve the biology of fracture healing and thereby result in a higher incidence of union. Percutaneous reduction of unstable B1 peri-implant fractures around total hip replacements, with percutaneous cerclage wiring combined with minimally invasive locking plates, was shown to provide satisfactory reductions and union rates in a small series of patients.[61]
Controversy exists concerning the role of retrograde intramedullary nails versus periarticular plate techniques for supracondylar femur fractures after total knee replacement. One study found that locked plating technique has an increased rate of nonunion (19% vs 9%) and twice the rate of hardware failure,[62] though another study found no significant differences in time to union, range of motion (ROM), mean Knee Society Score, and alignment measurements.[62]
A systematic review of 44 studies concluded that there was no difference in rates of secondary procedures between locked plating, conventional plating, and retrograde intramedullary nailing. However, retrograde rodding had a significantly higher rate of malunion than locked plating did, whereas locked plating had a trend toward higher rates of nonunion than retrograde rodding.[63] Overall, there does not appear to be a significant difference between treatment with locked plating and treatment with retrograde intramedullary rodding.[38, 64]
Intraprosthetic fixation (drilling holes in the prosthesis for screw fixation) that directly fixes the bone to the prosthesis has promise. Brand et al found that drilling the prosthesis did not compromise the prosthesis.[65, 66] Intraprosthetic fixation would allow stable fixed-angle bicortical screw placement in osteoporotic bone without the need to avoid the prosthesis. However, drilling the stem is associated with increased temperature, which can cause osteonecrosis and soft-tissue damage; accordingly, cooling the stem during drilling is recommended.[19]
Other newer techniques include far cortical locking and use of supplemental medial plates.[66] The Ortho-Bridge implant is another new construct for periprosthetic femur fractures.[67, 68, 69]
Casting, bracing, and protected weightbearing are indicated only for stable fractures in which the implant is not loose and alignment of both the prosthesis and the limb is acceptable for adequate function when the fracture heals. Vancouver type A fractures can be treated conservatively when displacement of the trochanteric fracture is less than 2.5 cm. If displacement exceeds 2.5 cm, then surgery may be indicated, using tension band wiring or hook plate techniques.[19]
Acromial and scapular spine periprosthetic fractures after reverse shoulder arthroplasty are usually treated nonoperatively. These commonly are stress fractures that heal with an abduction pillow with elbow support (to offload the deltoid) after about 6-8 weeks. Acute fractures are also generally treated nonoperatively; surgical treatment is unreliable because of inability to counter the pull of the deltoid and difficulty obtaining fixation in thin osteoporotic bone. Fixation has theoretical advantages that might yield improved shoulder function in select cases.[70] Although humerus fractures after reverse shoulder arthroplasty are usually treated surgically, they have also been successfully treated in a closed fashion.[71]
Surgical options include the following:
The most important factor in treating peri-implant fractures is the status of the implant. Careful assessment of preoperative x-rays and comparison with previous x-rays (when available) is essential.
When the implant is loose,[11, 26, 72] malaligned, or deformed, revision of the implant may be the best option. The potential difficulties of fixation and complications of nonunion or malunion are avoided by eliminating the fracture. Difficulties in achieving fixation because the implant is in the way also are bypassed by removing the implant. In the Vancouver classification, these tend to be B2, B3 and C fractures.[19]
Even when the implant appears stable on preoperative evaluation, the prosthesis should also be tested for stability intraoperatively.[5]
Cementless modular implants with diaphyseal anchoring are a good option for achieving optimal restitution of length, soft-tissue lever arms, and femoral offset, and they are conveniently adjustable by virtue of their modular structure. Cemented modular devices allow early weightbearing and are a good option, especially for elderly patients with osteoporotic bone.[19]
Guidelines for managing periprosthetic fractures around acetabular implants are similar to those for managing fractures around femur prostheses. If they are loose, replace them; if they are not loose, repair them if possible.[73]
Outcomes for distal femur replacement after periprosthetic distal femur fracture (see the images below) may be equivalent to those of open reduction and internal fixation (ORIF) with respect to mortality and reoperation rate; the former may be more reliable in complex fracture patterns where it is difficult to obtain adequate fixation.[74, 75] Distal femur replacement may also be the better option in patients with osteoporotic bone.[75]
When the implant is not loose, removal may be difficult, time-consuming, and complicated by further fracturing of the bone or other adverse consequences of revision surgery. When the implant is stable (as in Vancouver B1 fractures) and well aligned, it is usually possible to treat the fracture with standard fixation methods while retaining the implant or prosthesis. An exception is when the bone stock for fracture stabilization is inadequate. When stable fracture fixation cannot be achieved, even if the implant is stable, the implant (or prosthesis) must be removed, and joint replacement (or revision) is probably the best treatment.
Preoperative templating is required to ensure that adequate revision or fixation implants are available and that the goals of surgery can be achieved. If screw fixation around a medullary stem or rod is planned, careful assessment of the implant's fit in the canal is necessary to ensure that there will be room for the screws. Even unicortical screws require some space for their tip.
Johnson-Lynn et al concluded that "a delay to order necessary equipment and obtain relevant surgical expertise for the treatment of these complex fractures is safe and not associated with increased mortality or post-operative complications."[76]
The surgical approach to fixation of periprosthetic fractures depends on the site of the fracture and the local anatomy. Prior incisions should be used when possible. When additional incisions are needed, the soft-tissue envelope must be respected, with care taken to use wide skin bridges, to refrain from undermining the skin, to avoid self-retaining retractors, and to preserve the fracture hematoma.
Minimally invasive fixation has yielded improved results in comparison with open approaches, with less risk of nonunion.[9] Stoffel et al also reviewed the literature and found higher rates of nonunion (4.5% vs 0%) and refracture (3.8% vs 0.6%) after open approaches than after minimally invasive fixation.[77]
With the Vancouver classification of peri-implant fractures associated with total hip arthroplasty (THA), type A fractures are treated by nonoperative management or cerclage. B1 fractures are treated by means of ORIF with plates, cortical strut grafts, or both. An international survey of orthopedic surgeons found that for B1 fractures, ORIF with locked plating was slightly favored over ORIF with cable plating with or without cortical strut allograft (51.1% vs 45.5%).[78] B2 and B3 fractures are revised to a long-stem femoral component, possibly with additional fracture fixation with supplemental bone grafts. C fractures are treated by means of ORIF.[36]
The choice of fixation for peri-implant fractures around total knee replacements depends on the level of the fracture, the presence or absence of a stem, and the surgeon's training and preference. There appears to be no significant difference between treatment with locked plating and treatment with retrograde intramedullary rodding.[38, 64] Mean operating time, intraoperative blood loss, and time to fracture union were similar in a study by Hou et al.[38] In the locked plating group, there was a slightly higher nonunion rate; however, in the intramedullary rod group, there was a higher malunion rate.[38, 64]
A case example of a Vancouver B1 fracture at the end of a well-fixed hip replacement stem treated with a locked plate is illustrated in the images below.
Cerclage wiring alone can provide adequate fixation for fracture patterns around a well-fixed stem (Vancouver B1), but it is associated with a higher rate of stem subsidence.[79] Screws should be added when the implant will be subjected to axial and torsional loading. Cement augmentation can improve screw fixation.[80]
Locking plate constructs are more resistant to axial and torsional loads than nonlocked plates are.[9] Unicortical locked screws have improved resistance to lateral bending and torsion when compared to cables. Bicortical screws have better mechanical stability than either unicortical screws or cerlage cables,[9] but their use requires enough space to place the screw past the prosthesis.
Comparing locking plates to cable-compression plate fixation, Dehghan et al found that locking plates had a higher rate of nonunion.[81] However, locking plates have less risk of nonunion, malunion, and loss of reduction and less need for additional surgical procedures than nonlocking plates when used for periprosthetic fracture of the distal femur.[82]
Allografts have been used with and without plate fixation of periprosthetic fractures.[83] Allograft, when used alone, has inferior mechanical resistance to torsion and lateral bending as compared with plate and screws and with cerclage cables. Allografts may restore bone stock and may increase the rates of union when used with plate internal fixation. However, the additional soft-tissue damage required to place the allograft has also been associated with delayed union and increased infection rates.[9]
Use of allografts require soft-tissue stripping, which may delay bone healing or increase the risk of infection. Allografts may also transmit disease, cause immune reactions, and add to the cost.[73] However, strut allografts may prevent and fill bone defects or stress risers in patients with known risk factors such as rheumatoid arthritis (RA).[14, 84]
Dual plating with the plates placed at 90º to each other (orthogonal plating) has superior mechanical stability when compared with lateral plating alone,[9] but it does require greater exposure and soft-tissue stripping.[85] Dual plating may be indicated when bone stock or bone quality is poor in areas exposed to rotational stress.
The classic recommendation for length of fixation is two cortical diameters away from the fracture.[9] Mechanical studies support at least 6 cm,[86] and increased length (spanning the entire femur) is now recommended by many to further decrease the risk of additional peri-implant fractures.[9, 87] Drew et al showed a higher rate of reoperation when the plate extended across less than half of the length of the femur than when the plate spanned more than 75% of the femur.[88]
Promising results have been reported for the Ortho-Bridge system for femoral periprosthetic fractures. The system uses flexible titanium rods with adjustable locking screws that can be attached to the rods by clamps in optimal positions. Biomechanical parameters (including axial stiffness and torsional resistance to failure) were superior to those of locked plating.[67, 68, 69] The Ortho-Bridge system also has higher elasticity than locked plating does, possibly producing axial micromotion that stimulates callus formation.[69]
Revision long-stem prostheses have better mechanical stability than any form of internal fixation,[79] but the increased stability must be weighed against the potentially increased complexity, the surgical trauma, and the anesthetic stress expected with a long-stem revision prosthesis as compared with potentially minimally invasive internal fixation techniques.
In cases where internal fixation is not feasible, distal femoral arthroplasty may be a successful option, though complications such as loosening, patellar maltracking, knee dislocation, and additional periprosthetic fracture may occur.[89] So-called megaprostheses, such as those used in tumor situations, may have indications when adequate fixation cannot be achieved after periprosthetic fractures. Windhager et al reported satisfactory results after periprosthetic total knee replacement surgery.[90]
If the implant has failed, as in the case of a loose prosthetic replacement, surgical treatment necessarily involves removal of the failed prosthesis and repeat replacement (revision) with a new prosthesis. The stem of the new prosthesis usually must be longer than the original so that it can bypass the fracture to stabilize it.
A case example of hip replacement after failed hip replacement may be helpful. An 82-year-old woman with a preexisting loose hip replacement fell and sustained a periprosthetic femoral fracture (see the image below). Radiographic evaluation showed moderately severe osteolysis with probable subsidence of the cemented femoral component (with a gap in the stem-cement interface at the lateral aspect of the prosthesis).
Because the stem was loose, an acute revision operation with removal of the prosthesis, strut medial allograft, and long-stem femoral revision was performed.[91] The acetabular component also was revised with an uncemented component because it was found to be loose at surgery. Postoperatively, the patient did well, with partial weightbearing for 3 months and a stable prosthesis with allograft incorporation at 6 months.
If the fracture cannot be stabilized, despite a stable implant, because of inadequate bone to hold fixation devices, surgical treatment can include removal of the implant and replacement of the inadequate bone with a new prosthesis. If distal femoral bone stock (with periprosthetic total knee replacement fracture) is severely inadequate, a distal femur replacement prosthesis can be used as a salvage procedure in low-demand patients, though this is a technically demanding operation.[92]
In another case example (see the images below), the hip bipolar hemiarthroplasty prosthesis was still solidly fixed in the proximal bone, but the remaining proximal bone was inadequate for internal fixation, thus necessitating a proximal femur replacement prosthesis. Intraoperative periprosthetic splitting of the osteoporotic diaphysis was identified during surgery and treated with internal fixation. The final proximal femur replacement prosthesis with internal fixation was successful.
A case example of hip replacement after fracture at the tip of the hip lag screw may also be helpful. In this third case, an elderly man sustained an intertrochanteric hip fracture and was treated with a dynamic hip screw implant. The original fracture healed, but he had a new fracture at the tip of the lag screw after a fall (see the image below). Fixation options were few because of inadequate bone stock, and he had a good result with removal of hardware and hip hemiarthroplasty.
If fixation of the fracture is chosen instead of replacement, the usual principles of fracture fixation must be followed. Stable anatomic fixation with preservation of soft-tissue attachments through indirect reduction techniques should be achieved to obtain good results.[13, 32] After repair of periprosthetic femur fractures around total knee arthroplasties (TKAs), locked plating and intramedullary nailing had similar union rates (87% vs 84%).[93]
Whereas the surgical approach is usually open with direct reduction, minimally invasive plate fixation of periprosthetic fractures around knee replacement implants has had good results, including similar motion, alignment, and Knee Society Score postoperatively as compared with prefracture evaluations.[94] The surgeon must choose the device that fits the patient best, with careful preoperative planning and intraoperative flexibility and creativity. A wide selection of implants must be available. Options include flexible intramedullary rods, rigid intramedullary rods, and special plates, possibly with cerclage wires[95, 96, 97, 98] and external fixators.[85]
Flexible intramedullary rods (eg, Zickel supracondylar, Ender, and Rush rods) can be slipped alongside intramedullary stems. They can be placed through minimal incisions and act as an internal splint until fracture healing occurs.[32, 99] They usually require some external protection (eg, a cast or brace) and rarely allow unprotected motion or weightbearing.
Preoperative radiographs must be studied carefully to confirm that there is enough room in the medullary canal for the implant. It may be difficult to maintain axial alignment and length with these devices. Their use mainly is indicated in patients in whom surgery is especially risky and the ability to place the devices with minimal surgical trauma outweighs the risk of imperfect reduction.
A case example of distal femur fracture with proximal hip replacement demonstrates this point. An elderly woman with a solid asymptomatic previous hip hemiarthroplasty fractured her distal femur in a fall. She was treated with Zickel supracondylar devices and healed without complication (see the image below). At 3-year follow-up, the hip remained asymptomatic.
Rigid intramedullary rods (eg, antegrade, supracondylar, retrograde) are stronger than flexible rods and do not require external support. They cannot be used when a fracture has occurred around a stemmed implant (because the stem is in the way) but can provide rigid fixation for other peri-implant fractures. Advantages of intramedullary fixation include indirect reduction with less stripping of periosteal blood supply and preservation of soft tissues and the fracture hematoma with its bone-forming cells and factors. Soft-tissue protection increases the likelihood of union and decreases the likelihood of infection.[85]
Biomechanically, the intramedullary position of the nail is stronger as compared with plates because of increased resistance to torque forces and increased load transfer to the bone.[32, 99] A case example of a fracture at the end of a blade plate treated with a retrograde nail is as follows: A young man who fractured his hip in a high-speed motor vehicle accident less than 2 years previously refractured his femur at the distal end of his plate after another motor vehicle accident. Rigid fixation was obtained with retrograde rodding (see the image below).
The following is a case example of a fracture above a total knee replacement treated with an antegrade nail. An elderly woman with bilateral knee replacements sustained bilateral distal femur fractures proximal to her knee replacements. Rigid fixation and healing of both fractures was achieved with antegrade nailing (see the image below).
A case example of pathologic fracture above a plate treated with an antegrade nail follows. An elderly woman with a pathologic humerus lesion from metastatic breast cancer was treated initially with plate fixation that failed. Intramedullary fixation that was stable enough to restore function and decrease pain was required to improve quality of life (see the image below).
Intramedullary nails cannot be used in patients with severe joint contracture, those with ipsilateral joint replacement on the same bone (eg, a femur with both a total knee and a total hip replacement), or patients where the implant blocks the entry point (eg, a hinged knee replacement) or the bone is inadequate for locking screws.[85]
Plates and screws are also commonly used to repair periprosthetic fractures.[100, 101, 102] Although plates can be placed with indirect reduction techniques to minimize soft-tissue damage, and newer plate designs provide more "biologic" fixation,[99] they usually destroy at least some of the periosteal blood supply and always disrupt the fracture hematoma. Plating techniques allow direct fracture reduction. This achieves more exact anatomic alignment, which may be crucial for long-term joint function.[32]
Placement of screws through the cement mantle surrounding a cemented prosthesis does not lead to cement mantle failure, nor does it cause instability of a prosthetic stem.[103] Plates that span the whole bone have less risk of recurrent peri-implant fracture or nonunion than shorter plates do.[87] On the basis of mechanical testing, Walcher et al recommended a minimum overlap of 6 cm between a plate and an intramedullary stem to decrease the likelihood of stress risers causing fracture at the end of the implants.[86]
Plating techniques allow for interfragmentary compression more readily. This creates a more rigid construct, facilitating early motion. Although intramedullary rods act as internal splints, plates can be placed as a tension band and/or neutralize the forces acting on interfragmentary screws.[99] Special plates are usually required, allowing a combination of cerclage wires and screws to hold the plate to the bone while avoiding the intramedullary implant.
Fractures of the calcar during hip replacement can be treated with cerclage wires or Parham bands.[25, 104] Strut allografts can provide increased biomechanical advantage, with the best mechanical stability achieved when a plate is combined with a medial strut allograft.[105] However, placement of the allograft strut compromises the local biology, with increased rates of delayed union and infection.[106]
A case example of a fracture at the distal end of a hip replacement treated with a plate is as follows. An elderly woman sustained a low-energy injury to her leg, with fracture occurring at the tip of a preexisting hip replacement. She had a solid hip arthroplasty; thus, ORIF with plate, cerclage wires, and screws was performed. The fracture healed without evidence of prosthetic failure (see the image below).
A case example of fracture at the proximal end of a supracondylar nail treated with a plate follows. An elderly woman with previous supracondylar femur fracture presented with a new fracture at the proximal tip of her supracondylar rod after a motor vehicle accident. ORIF with a plate was performed, with good results (see the image below).
Newer fixed-angle locking unicortical screw plates allow less invasive fixation than was possible with older techniques, which used allografts and cerclage wires. Unicortical screws can be placed with far less periosteal stripping than cerclage wires. Mihalko et al[107] showed that cables can resist bending loads, but Schmotzer et al[108] demonstrated that cables resist torsional loads poorly as compared with screws.
The authors' cadaver research has shown that it takes six cerclage wires to equal the rotational and anteroposterior stability of a single unicortical screw with a lateral plate.[109] In another cadaver study, Lenz et al showed that bicortical locking screws provided the best resistance to failure with repetitive loading; a combination of a unicortical screw with cerclage wire was an acceptable alternative. Unicortical locking screws alone or cerclage wires alone did not provide adequate stability.[110]
In a relevant case example, a 73-year-old man with periprosthetic femur fracture distal to a well-fixed total hip replacement stem presented with a nonunion after three attempts at plate fixation using cerclage wires for proximal fixation. ORIF was accomplished with two "combi" fixed-angle locking screw plates (anterior and lateral placement to help control both anterolateral and mediolateral forces), with healing within 3 months (see the image below).
In another relevant case example, a 49-year-old woman with periprosthetic femur fracture 2 cm distal to a well-fixed total hip replacement stem presented with nonunion after three attempts at plate fixation using cerclage wires for proximal fixation and one attempt at retrograde rod fixation. ORIF was accomplished with a LISS fixator and an anterior LC-DC plate. The anterior plate included a lag screw, and the LISS was inserted with minimally invasive technique (including percutaneous proximal unicortical screw placement). The patient was clinically healed by 3 months and radiographically healed by 5 months (see the image below).
Elderly patients with periprosthetic fractures often have significant medical comorbidities and severe osteoporosis, which compromise both their ability to tolerate major surgery and the surgeon's ability to obtain adequate stability. Circular thin-wire external fixation using the Ilizarov method (see the image below) may be a solution for these patients, in that it provides rigid fixation and allows immediate full weightbearing in a minimally invasive fashion, thereby decreasing the risks of surgery and anesthesia.
Using this method, Nozaka et al achieved 100% union with return to prefracture activity level in most patients, with minor pin-tract infections as the only complication.[85]
Postoperative care varies, depending on the fracture, the implant, the method of fixation or replacement, the quality of the bone, and the ability of the patient to comply with instructions. In general, cemented prostheses and rigid intramedullary rods allow immediate weightbearing without casting or bracing. Uncemented prostheses often require protected weightbearing initially. Plate fixation and flexible intramedullary rods may require protected weightbearing and bracing or even casting. Physical or occupational therapy is often useful for maximizing function.
Elderly patients often cannot return to their prefracture ambulatory status after long-term nonweightbearing; accordingly, techniques that allow immediate full weightbearing are preferred. Circular multiplanar thin-wire external fixation does allow immediate full weightbearing.[85]
Periodic clinical and radiographic postoperative follow-up is necessary to evaluate stem stability and bone quality and thereby to identify at-risk patients and proactively attempt to prevent additional periprosthetic fracture.[14]
Although Gunther et al reported good results in their series of periprosthetic fractures,[111] other authors have reported significant complications after treatment of periprosthetic fractures, including infection, dislocation, secondary loosening, nonunion, malunion, and poor functional outcome.[5] As many as 22% of patients with periprosthetic fractures experience wound complications, with 16% requiring additional surgical treatment. Peripheral vascular disease, pulmonary disease, and bariatric surgery predispose to wound complications in this elderly population. Closed-incision negative-pressure therapy has been found to improve soft-tissue outcomes.[112]
Complications are more common in treating periprosthetic fractures than in treating fractures without an implant. Surgery is technically more difficult, and bone quality is poorer. Longer operating times and increased blood loss are expected. Failure of fixation occurs when inadequate stability is achieved. (See the images below.) Infection rates are increased because of increased soft-tissue damage from more difficult surgical dissection.[113] Deep venous thrombosis, pulmonary embolism, and systemic complications should be expected and treated early.
Drew et al reviewed patients with periprosthetic femur fractures and reported a 1-year mortality of 13% and a reoperation rate of 12%. The risk of reoperation was less with a greater span of fixation and with revision arthroplasty instead of internal fixation.[88]
Ebraheim et al reported on complications following periprosthetic total knee replacement. Locking plate fixation of these fractures had a 35% complication rate, whereas intramedullary techniques had a 53% complication rate. Complications included malunion and nonunion necessitating repeat operations.[93]
Moore et al reviewed the literature on the treatment options for periprosthetic femur fractures and found higher complication rates when allograft struts were used, with increased time to union (4.4 vs 6.6 mo) and higher deep infection rates (3.8% vs 8.3%). Plate type and use of cerclage wires did not affect these complication rates.[106]
Stoffel et al also reviewed the literature and found a 14.3% complication rate after periprosthetic femur fractures: Nonunion and refracture occurred more often after open approaches than after minimally invasive fixation. Nonlocking plates also had a higher rate of nonunion than locking plates did.[77] The risk of nonunion is 11.9 times higher with nonlocking plates than with locking fixation.[9]
A case example of a patient with three periprosthetic fractures after total hip replacement is as follows. The original fracture was at the stem of a primary total hip replacement. The second fracture occured intraoperatively during revision of the primary total hip replacement (treated with plate and cerclage wires), and the third fracture at the end of the long-stem revision prosthesis is shown in the images below. The fracture healed after minimally invasive locked plating. (To preserve the biology, the fracture site was not opened.)
A healthy diet that includes adequate calcium and vitamin D intake can slow the progression of osteoporosis, thereby potentially decreasing the risk of periprosthetic fractures. (See Osteoporosis.)
Pre- and postoperative evaluation and treatment of poor nutrition and bone health is essential to prevent additional future periprosthetic fracture.[14]
The goal in treatment of periprosthetic fractures is the same as for any fracture: early mobility and return to function using mechanically stable implants. Although osteoporosis and poor bone stock may compromise fixation, the patient should be encouraged to be as mobile and active as the fixation allows. (See Osteoporosis.)
Patients with periprosthetic fractures are frequently elderly with multiple comorbidities. Medical consultation should be obtained as needed.
The fracture should be monitored by means of radiography and clinical examination until it heals. The patient should be monitored until rehabilitated to full potential. The general recommendation for periprosthetic fractures (around a joint replacement implant) is to evaluate the prosthesis about every 1-2 years to identify loosening, subsidence, osteolysis, or other signs of progressive prosthetic failure.
Peri-implant fractures (around a fixation device) should be followed until union is established.
Bisphosphonates, including alendronate (Fosamax), risedronate (Actonel), ibandronate (Boniva), and zoledronate (Reclast) may help decrease the progression of osteoporosis. Adequate calcium and vitamin D intake is essential for bisphosphonates to be effective. Long term use has been associated with atypcial femoral shaft insufficiency fractures. Long term use is also associated with atypical proximal periprosthetic femur fractures.[18]
After menopause, estrogen therapy reduces bone loss, increases bone density in both the spine and hip, and reduces the risk of fractures. Estrogen therapy increases the risk of endometrial and possibly ovarian cancers so should be used with caution for short times only.