Recent Advances and Developments in Knee Surgery: Principles of Periprosthetic Knee Fracture Management

Epidemiology

More than 300,000 total knee arthroplasties are carried out annually in the United States [1]. The incidence of periprosthetic fractures following total knee arthroplasty is up to 5.5 % [1, 3, 4]. Supracondylar femur fractures are the most common type of periprosthetic fracture following total knee arthroplasty. The Mayo Clinic Joint Registry reported an approximately 2 % incidence of periprosthetic knee fractures, of which 0.1 % of femoral fractures occurred intraoperatively during primary surgery and 0.9 % of femoral fractures occurred during revision arthroplasty [1]. Intraoperative fractures may occur and may be undetected if only minimally displaced and may not be reported if they do not require intervention. The incidence of periprosthetic supracondylar femur fractures has been reported as ranging from 0.3 to 2.5 % following primary total knee arthroplasty and from 1.6 to 38 % after revision surgery. Periprosthetic proximal tibial fractures are less common. A series of over 17,000 total knee arthroplasties reported by the Mayo Clinic showed an incidence of periprosthetic tibial fracture following total knee arthroplasty of 0.1 % intraoperatively and 0.4 % postoperatively. The incidence of patella fracture following total knee arthroplasty has been reported as being between 0.2 and 21 % in resurfaced patellae and 0.05 % in unresurfaced patellae [1,2]. A review of 12,462 consecutive total knee arthroplasties at the Mayo Clinic reported a prevalence of patellar fracture following total knee arthroplasty as 0.68 %. 44% of the fractures were asymptomatic or minimally symptomatic. Most of these were identified on routine follow-up radiographs. 82% of patellar periprosthetic fractures were discovered within three years [5].

Risk Factors

Many periprosthetic fractures following total knee arthroplasty occur after traumatic events. Numerous patient and surgical factors predispose patients to developing these fractures. A significant patient factor that is associated with an increased risk of periprosthetic knee fracture is osteopenia [6]. Osteopenia is associated with advanced age, chronic corticosteroid use, rheumatoid arthritis and female sex. Poor bone stock may also occur secondary to stress shielding. Other patient factors include neurological disorders such as epilepsy, Parkinson’s disease, cerebellar ataxia, myasthenia gravis, poliomyelitis, cerebral palsy and other undefined neuropathic joints [1, 6, 7]. Patients with knee flexion greater than 95 degrees generate increased patellofemoral compression forces and frequently have higher activity levels, potentially predisposing to fracture [8]. Conversely, a stiff knee may be associated with mechanical stress-risers at the knee [1].

Meticulous surgical technique is essential to avoid periprosthetic fracture [7]. Intraoperative fracturing is more likely to occur during component removal, bone retraction, trial reduction or preparation for insertion of stemmed components. Screw holes or anterior notching of the distal femur increase fracture risk by creating mechanical stress-risers at or around the knee. 40 to 52% of reported fractures of the distal femur are associated with an anterior notch [8]. If an anterior notch of the distal femur is created intraoperatively, the surgeon should consider implantation of the femoral component with an attached diaphysis-engaging system to support the weakened distal femur. Axial malalignment and component malposition has been shown to be associated with an increased risk of tibial fracture. Component loosening and knee instability also increase the risk of fracture [8].

Biomechanical studies have demonstrated that contact forces substantially increase the patellofemoral malalignment. The eccentricity and magnitude of patellofemoral loads increase with patellar subluxation increasing the risk of patellar fracture. Excessive insufficient or asymmetric patellar resection, disruption of the patellar blood supply resulting in avascular necrosis increase the risk of patellar fracture. Thermal necrosis of bone due to the heat of polymerisation of cement predisposes to fracture. Additionally, an increased risk of patellar fractures has been associated with certain patellar component designs such as metal-backed uncemented patellae or components with large central pegs [1, 7, 8].

Classification

A number of classifications for periprosthetic fractures following total knee arthroplasty exist. The classification of periprosthetic supracondylar fractures by Rorabeck is commonly used and takes into account the displacement of the fracture and the stability of the prosthesis. A Type I fracture refers to an undisplaced fracture with an intact bone-prosthesis interface. A Type II fracture refers to a displaced fracture with an intact bone-prosthesis interface. A Type III fracture refers to either a displaced or undisplaced fracture with a loose or failing prosthesis [6, 8, 9].

Felix et al. classified periprosthetic tibial fractures into four types based on the anatomic location and proximity to the prosthesis as well as the status of the prosthetic fixation. Type I fractures occur at the tibial plateau. Type II fractures occur adjacent to the prosthetic stem. Type III fractures occur distal to the stem. Type IV fractures involve the tibial tubercle. The fracture types are further classified according to whether the prosthesis was well fixed (A) or loose (B) based on radiographic appearances and whether the fracture occurred intraoperatively (C) [8, 10].

The Goldberg classification of patellar fractures following total knee arthroplasty consists of four types and takes into account the stability of the implant fixation, the status of the extensor mechanism, the location of the fracture and the presence of dislocation. Type I refers to fractures through the mid-body or superior pole of the patella not involving the implant, cement or quadriceps mechanism. Type II fractures involve the implant/bone/cement composite and/or the quadriceps mechanism. Type III refers to the inferior pole fractures and is subdivided into Type IIIA, which occurs with patellar ligament rupture and Type IIIB which occurs without patellar ligament rupture. Type IV refers to fracture-dislocations [11].

Nonoperative Treatment

Nonoperative options include skeletal traction, application of a cast, pins and plaster and cast bracing followed by protective weight bearing and range of motion exercises. The advantages of nonoperative treatment are that it is noninvasive and the infection rate is negligible. Nonoperative management is generally recommended for stable, minimally displaced fractures with good bone stock and a well-fixed and well-aligned component (Type IA). Non-operative treatment is best reserved for nondisplaced fractures that do not demonstrate intercondylar extension. Disadvantages of non-operative treatment are the relatively high rate of malunion and functional loss particularly in patients with displaced fractures. Maintenance of reduction of displaced fractures is particularly difficult when there is associated osteopenia and communition. Non-operative treatment often requires immobilization which leads to loss of motion and reduced walking capacity. Radiographs must be monitored closely and frequently [7, 8].

Several studies have looked at outcomes of nonoperative treatment of supracondylar periprosthetic fractures and have shown varying results. Chen et al. reviewed the literature and reported on 195 periprosthetic supracondylar fractures in twelve published studies. They revealed satisfactory results in 83% of patients with type I fractures that were treated nonoperatively. They found satisfactory results in 64% of patients with Type II fractures treated nonoperatively [12]. A literature review by Harlow and Hoffman of 142 supracondylar periprosthetic fractures treated nonoperatively found that 29% of these fractures eventually required operatively treatment [8]. Culp et al. reviewed sixty-one supracondylar periprosthetic knee fractures. Of the thirty cases treated nonoperatively, 50% had increased pain levels or a change in ambulatory status, compared to only 13% of patients treated operatively [13]. A review by Merkel and Johnson showed that 35% of patients (9 out of 26) with supracondylar periprosthetic knee fractures treated nonoperatively required revision arthroplasty because of non-union, malunion, loosening of the components and extensor lag. Despite this, they recommend conservative management and concluded that patients who have a poor arthroplasty result after nonoperative treatment of the fracture can usually undergo a revision arthroplasty with the expectation of a satisfactory result [14].

Operative Treatment

Surgical intervention has the advantage of facilitating early mobilisation and so may be considered in healthy or active patients regardless of the fracture pattern. Displaced fractures with adequate bone stock generally require operative treatment as there is a high prevalence of progressive displacement, nonunion and joint misalignment [8].

If the prosthesis is stable, open reduction with or without bone grafting is generally indicated. If the prosthesis is loose, revision arthroplasty to a stemmed component is recommended [15]. The aim of open reduction and internal fixation is to provide anatomical reconstruction and early rehabilitation of the patient. It may be considered in healthy or active patients regardless of the fracture pattern to promote early mobilisation. A wide variety of implants are available for open reduction and internal fixation of the fracture, including angle blade plates (ABP), dynamic condylar screws (DCS), buttress plates, flexible or rigid intramedullary nails and locking periparticular plates. Locking plates such as the Less Invasive Stabilization System (LISS) plate may be implanted with minimal soft tissue dissection and periosteal stripping and allowing preservation of blood supply as well as potentially reducing intraoperative blood loss and the risk of infection. The LISS plate has virtually replaced the use of DCS and ABP plates in recent years. Locked plates have been shown in cadaver studies to be biomechanically more stable than non-locking plates such as the DCS and ABP. An advantage of the LISS plate is the ability to place multiple fixed-angled locked screws to avoid existing hardware or to capture specific fracture fragments and improve stability [1, 16]. Clinical studies have reported good outcomes with the use of LISS plates. For example, Althausen reported that LISS plates were superior to flexible intramedullary nails, plate fixation and retrograde intramedullary nails and allowed immediate postoperative mobilisation whilst being associated with low infection rates and no requirement for acute bone grafting [17].

Flexible intramedullary nails have been used as a less invasive treatment option to treat supracondylar periprosthetic knee fractures. However, this technique is associated with reduced axial and rotational stability and as a result is generally not used [18, 19]. Ritter et al. reported 100% union in a review of twenty-two cases treated with Rush rods. The average postoperative alignment was 10 degrees (compared to 7 degrees preoperatively). Two fractures healed in 15 degrees of valgus alignment [20]. This technique may be considered to treat mildly displaced fractures in patients with significant comorbidities [8]. Flexible intramedullary nails are now rarely used. Rigid retrograde femoral nails can be used instead for fractures with distal fracture fragments large enough to allow insertion of a locking screw. It is relatively minimally invasive and generally provides good axial, angular and rotational stability. However, it is limited as it cannot be used for very distal fractures, in severe comminution with loose total knee components and it should not be used in the presence of a pre-existing ipsilateral total hip replacement as it can lead to a stress riser below the femoral stem [1, 8].

External fixation is another treatment option but is rarely used due to problems with tethering of the quadriceps muscles, limitations imposed on knee range of motion of the knee and the risk of propagation of pin tract infections into the knee joint space [1, 8, 21]. Other techniques that have been used include fibula strut grafting, use of femoral allograft in combination with a prosthesis and arthrodesis. As these treatment modalities have been used much less frequently, there are limited reports on them in the literature [6].

Revision Total Knee Arthroplasty

Revision total knee arthroplasty may be used for an extremely distal and comminuted fracture where secure fixation cannot be achieved or if the fracture is associated with a loose and unstable implant or poor bone stock in the distal femur. Revision total knee arthroplasty is frequently required in cases in which nonoperative or other methods of treatment have failed. An uncemented long-stemmed prosthesis with fracture fixation, typically with interfragmentary screws and/or small plates may be used. A long-stemmed prosthesis provides stable fixation and allows patients to start early movement and weightbearing. A custom implant may be required to enable distal femoral replacement to be undertaken in cases where a loose prosthesis is coupled with metaphyseal bone stock [1, 8].

Chen et al. reported that ten out of eleven supracondylar periprosthetic fractures treated with revision total knee arthroplasty had a satisfactory result [12]. Keenan et al. reported seven cases of displaced supracondylar fractures above a total knee arthroplasty with custom made implants and reported good implant alignment and fixation as well as rapid postoperative recovery, mobilisation and good functional outcomes [22].