Liporace, Frank A. MD; Yoon, Richard S. MD; Collinge, Cory A. MD

INTRODUCTION

By 2060, in the United States, the estimated population older than 65 years will reach nearly 98 million.^1–3 Therefore, the incidence of elective arthroplasty, and those treated with fracture implants after lower extremity fractures will continue to increase as well.^4–6 Although fixation strategies exist for periprosthetic hip and knee fractures, there is no standard of care regarding the more complex interprosthetic and interimplant fractures.^7–9 Here, we review the basic fixation principles and strategies that may be applied to these more complex fractures and offer case examples to illustrate such strategies.

Biomechanical Fixation Principles of Periprosthetic Fractures

In 1970, Brooks et al¹⁰ performed a landmark study in cadaveric canine bone models to determine the ideal distance for spanning cortical incompetency in femurs to produce a stable construct. From this study arose the classical teaching that a surgeon must span a cortical defect by “2 cortical diameters” for construct stability, either by intramedullary or extramedullary means.¹⁰ The results of this study have been the basis for nearly all of the biomechanical studies subsequently performed and remain the clinically applied principle during fracture and revision surgery. The evolution of fixation constructs has coincided with advancing implant technology.¹¹ For Vancouver B1 fractures, early failures associated with Ogden plates, Mennen plates, and nylon cables have led to today's accepted recommendation using a combination of locked or compression plating with cable constructs or long-stemmed revision arthroplasty.^11–14 Similarly, for supracondylar fractures above a well-fixed femoral component, treatment has evolved where most surgeons will use distal femoral locking plates (LPs) or retrograde intramedullary nails (RIMNs).^{12,13,15–17}

Although modern technology has provided promising results, controversy still remains. For periprosthetic hip fractures, proponents of arthroplasty likely prefer bypassing Vancouver B1 fractures with long-stemmed revision components and cable stabilization; others prefer to fix the fracture, with a stable construct (which includes cable plating, with or without allograft struts), with a femoral stem that is well fixed.^9,14,18 For periprosthetic knee fractures, the debate between using lateral locked plating construct and RIMN continues.^19–21 Controversy still remains because of a lack of level 1 clinical evidence.^{7,8,10,19–25} Although there is a lack of data regarding periprosthetic fractures, there are even less data for interprosthetic fractures.^8,18

Understanding the Proximal and Distal Femurs

The proximal femur and hip joint experience some of the largest, multivector mechanical forces in the human body.²⁶ Thus, fractures in this region inherently are difficult to treat and can be even further complicated by implants. With the presence of an intramedullary implant (either IMN or femoral stem), the biomechanical properties of the native bone change.²⁷ Recent studies have noted minimal change when plates are placed about the femur; however, when a device is placed within the canal, the strength of the femur is significantly decreased.^27,28 This is an important change that leaves the femur at risk for peri-implant fracture. Bone strength decreases because of the load sharing formed by the intramedullary device.^27,28 The strain experienced by the native bone is decreased and as a result, bone quality decreases, a principle that falls in line with the Heuter-Volkmann principle.

These biomechanical changes noted in the proximal femur with intramedullary implants also apply to the diaphyseal and distal femoral regions. However, there are inherent differences in these anatomic regions, which include a change in both the deforming forces as well as the vascular perfusion, which is less abundant than in the proximal femur. In addition, in the distal, metaphyseal region of the femur, the load on an intramedullary implant is less than in the diaphyseal region secondary to a lack of ability for “fill” of this segment. The transition of stiffness (proximal to distal) and how it is altered with implant placement is an important concept, and must be considered when applying fixation either by intramedullary or extramedullary means.

Cables, Plates, and Screws

One of the earliest implants to aid in maintaining fracture reduction, cerclage cables and wires remain an important tool in an orthopaedic surgeon's armamentarium when treating periprosthetic fractures.¹¹ However, cables do not provide enough mechanical stability to be used alone in most circumstances.²⁹ Lenz et al³⁰ tested the mechanical behavior of cables by comparing their stiffness and ultimate tensile strength to unicortical and bicortical screws. Although cables were excellent when loaded in bending, they performed poorly in axial compression and torsion.³⁰ The authors concluded that cables should not be used alone and recommended that the ideal construct consisted of a combination of cables, uni-, and bicortical screws.³⁰ Recent work comparing fixation constructs has confirmed the utility of locked plate constructs about the hip.^{7,13,22,29–32} Reviewing purely biomechanical literature, bicortical locking screws are stiffer than unicortical, both of which are stronger than cables alone.^29,31,32 However, it is important to note that most of these studies were performed in the proximal femur in a periprosthetic model, which includes an implanted intramedullary stem.

The distal femur presents a different model, where the absence of an IMN stem presents a different biomechanical model. The “ideal” implants to use in a periprosthetic supracondylar fracture about a stable total knee arthroplasty (TKA) remains a controversial topic. Proponents and supporting clinical data exist for both lateral locked plating and RIMN. In a 2-part comprehensive biomechanical comparison between lateral locked plating and RIMN after periprosthetic supracondylar fracture about TKA, Salas et al reported both experimental and probability data for both fixation constructs.^33,34 Not surprisingly, when comparing locked plating to RIMN, lateral locked plating yielded a stiffer and more rotationally stable construct.^33,34 However, the authors also generated a probability model in an attempt to predict time to construct failure.³⁴ Using a computer-generated model, the engineers calculated the amount of accumulated damage to each construct with continued loading. Comparing locked plating to RIMN, the amount of damage accumulated by the lateral plate was significantly more than RIMN (21% vs. 0.019%), predicting a higher and earlier construct failure for LP alone.³⁴ Conversely, with so little damage accumulated by the RIMN, most of the energy with loading was transferred to surrounding bones, which in a periprosthetic model may be less than ideal.³⁴

Chen et al,³⁵ studied the biomechanical properties of 3 constructs: lateral plating alone, RIMN alone, or intramedullary allograft with locked plating. The authors defined each construct as lateral support (plating), middle support (RIMN), and lateral-middle support (plating + allograft).³⁵ In the osteoporotic model, the lateral plate was a stiffer construct than the RIMN but lacked rotational stability. The most stable construct was the lateral locked plate combined with the intramedullary allograft.³⁵ The authors hypothesized that the implants, when used alone, experienced too much strain, and likely were not suitable in osteoporotic and overweight patients.³⁵ However, the combined intramedullary and plate construct yielded a more balanced fixation model by providing appropriate levels of stiffness without excessive implant stress. Thus, perhaps for the distal femur, the best construct may include both an intramedullary device (ie, RIMN) combined with a lateral locked plate to provide the most biomechanically “friendly” environment.³⁵

Bypass the Fracture, Avoid Stress Risers, and Overlap Fixation

There is a paucity of literature regarding interprosthetic fractures (fracture between an implanted total hip and knee). In 1987, Brooks et al recommended bypassing the fracture by at least 2 cortical diameters. More recently, Bryant et al¹⁸ published a series of operative fixation of interprosthetic fractures with a lateral locked plate that spanned the entire femur, overlapping both femoral components proximally and distally. Although a retrospective (level IV) study, the authors reported good union rates and outcomes. Unfortunately, there was no comparison group with a plate that stopped short of the femoral component of the total knee.¹⁸ Similarly, treating Vancouver B1 fractures, Moloney et al⁹ published a cohort comparison between patients fixed with a plate that spanned the entire femur down to the condyles versus those who were fixed with a lateral plate that stopped short of the condyles. By increasing the working length, and avoiding the creation of a stress riser, the authors reported a lower nonunion and fixation failure rate, leading them to recommend spanning the entire femur in these cases.⁹

Recent biomechanical studies performed by Brand et al described an intraprosthetic technique, which links the lateral, overlapped plate to the intramedullary device.^2,3 The authors conducted a biomechanical comparison between a lateral plate with locked plating alone versus a locked plate that was linked to the intramedullary implant by creation of a threaded screw hole using a specialized drill.^2,3 The intraprosthetic construct proved significantly more stable than just a lateral locked plate additionally reporting the added potential benefit of immediate weight bearing using intraprosthetic fixation.^2,3 This must be weighed against the possibility of component loosening if the plate construct becomes unstable.

Operative Application of Fixation Principles: Case-Based Scenarios

Case 1: Interprosthetic Fracture: Overlap and Span the Femur With an LP

An 88-year-old rheumatoid arthritis female over a decade prior had a hybrid total hip arthroplasty (THA) along with subsequent ipsilateral TKA. She sustained a low interprosthetic fracture after a fall from standing (Fig. 1A). Looking closely, the fracture occurred at the junction of the TKA/bone interface at the anterior flange (Fig. 1B). The femoral component was deemed stable intraoperatively and thus, a spanning lateral LP was used (Fig. 1C). The patient was nonweight bearing for a period of 6 weeks and was advanced as tolerated. She has performed well at the 1-year follow-up, ambulating with an assistive device as per her preinjury baseline. Despite the presence of cement and extremely fragile bone, the aforementioned principles were applied—working length was maximized and adequate overlap occurred with the THA femoral stem.

Figure 1

Case 2: Interprosthetic Fracture: Restore the Middle Column, Overlap, Span, and Link the Femur

This 69-year-old morbidly obese, diabetic female sustained this interprosthetic fracture after a fall from standing, one month after her total hip replacement and 11 years after her stable total knee replacement (Fig. 2A). The decision was made to provide additional support to the middle column and restore the neutral axis with an RIMN. However, to avoid overstressing the surrounding bone, a spanning lateral LP was placed to decrease strain and balance the fixation (Fig. 2B). Finally, to construct a single unit, the LP was linked to the RIMN through locking screws in both the proximal and distal segments (Fig. 2B). Although stress shielding the lateral cortex, by spanning the entire femur, the transition of modulus was more balanced throughout the entire femur. With this construct, the patient was made full weight bearing immediately. At 2 years follow-up, the patient is healed and ambulating with a cane.

Figure 2

Case 3: Peri-Implant Nonunion

This 77-year-old-male sustained a distal femoral fracture, a sliding hip screw with side plate. Originally, he was treated with lateral locked plating. The patient went on to a nonunion and underwent three operations to promote healing. The last operation included placement of a fibular allograft in addition to revision lateral plating. When taken to the operating room, gross motion was noted (Figs. 3A, B). After ruling out infection, the implant and allograft strut were removed, and another lateral LP was placed, this time spanning the femur proximally to the tip of the sliding hip screw with side plate. To restore the neutral axis and support the middle column, an RIMN was placed with care taken to bypass the stress riser proximally and to overlap with the sliding hip screw. Finally, the hip screw and the lateral plate were linked to the RIMN to create a single construct (Fig. 3C). This patient was made full weight bearing immediately. The patient was healed by 11 weeks. The patient is without pain and doing well at 1-year follow-up.

Figure 3

Case 4: Interprosthetic Nonunion

This 79-year-old female had a surgical history significant for treatment of an interprosthetic fracture at the tip of the femoral stem of the hip prosthesis, which healed successfully. Five years after healing, she fractured distal to the lateral plate that had been used. Only bypassing the fracture, the patient went onto nonunion and was referred after 3 additional surgeries for nonunion. The patient presented having been wheelchair bound for 2.5 years with persistent pain and continued nonunion (Figs. 4A, B). The presenting construct overlapped the THA but did not span the distal femur and was fixed with mostly cerclage wires (Figs. 4A, B). Biopsy of the fracture site was positive for infection, and thus, the first stage of the surgery was resection of the atrophic nonunion and placement of an antibiotic-coated cement plate.³⁶ After a 6-week course of intravenous, culture-specific antibiotics, the entire femur was spanned with a lateral locked plate, placement of an RIMN and subsequent linkage of the construct (Figs. 4C, D). At 2.5 months, the patient was pain free at the fracture site and began to ambulate for the first time in nearly 3 years. At 1-year follow-up, the neutral axis remains restored with a healed nonunion (Figs. 4C, D).

Figure 4

Future Directions

Looking ahead, treating interprosthetic and interimplant fractures will increase in incidence as our population grows. Although we are reasonably well equipped to treat these fractures, there is much to learn and improve on as our knowledge base grows. Based on biomechanical principles and translational literature, the overall goals are to create a construct that allows for balanced distribution of strain along the neutral axis while achieving appropriate axial and rotational stability. Using known techniques, this means increasing working length and avoiding stress risers by overlapping with any existing implants. New areas for implant design and innovation are on the horizon, coinciding with growing data supporting spanning long bones and “linking” concomitant systems to create the most biomechanically friendly construct.

REFERENCES