Classifications in Brief: The Russell-Taylor Classification of Subtrochanteric Hip Fracture.

Abstract

History Subtrochanteric fractures occur in the proximal region of the femur. Anatomically speaking, the subtrochanteric region of the femur is defined as the interval between the lesser trochanter and approximately 5 cm below it, toward the isthmus of the femur [1, 13]. Subtrochanteric fractures are relatively common, accounting for approximately 10% to 30% of all hip fractures [4, 16]. The subtrochanteric region is subject to tensile and compressive stressors that are substantially greater than the patient’s body weight [11] as well as rotational and bending forces that directly influence the observed fracture patterns. These fractures commonly have short proximal fragments of comminution, which are pulled into flexion by the forces of the iliopsoas attaching on the lesser trochanter [9, 10]. Additionally, the comminution can be deformed into an abducted and externally rotated position as a result of the pull on the greater trochanter by the abductor muscles. This proximal abduction deformity can be further amplified as the distal fragment is pulled into an adducted position by the adductor’s attachment on the femoral shaft [1, 13]. These muscular-deforming forces make avoiding varus, flexion, or external rotation fracture malreduction more difficult and challenge the fixation construct. Given the complexity of the fracture pattern and anatomy within this region, many fracture classifications have been developed to create a treatment algorithm for surgeons. Subtrochanteric fractures were initially thought to be a subset of comminuted intertrochanteric fractures. A study performed by Boyd and Griffin [3] in 1949 analyzed 300 trochanteric fractures, allowing the authors to develop a classification system based on prognosis and the perceived difficulty of fracture reduction. They divided trochanteric fractures into four designations. The first, Type I, designates a fracture with a linear break extending along the intertrochanteric line from the greater to the lesser trochanter. Notably, with these fractures, the fracture is reduced and maintained in the most simple manner of those within this system [3]. A Type II fracture is comminuted with a fracture line in a similar orientation as Type I, but with points of comminution along the adjacent cortical bone. The reduction of these fractures rises in difficulty with the increasing degree of comminution. In the Boyd and Griffin classification, a Type III fracture is subtrochanteric and may have comminution. The authors noted that there may be elements of Type I and II fractures found within Type III fractures (intertrochanteric extension, cortical comminution, etc) and are accordingly the most difficult of the first three to reduce and fix. Finally, a Type IV fracture is comminuted and extends through the entire trochanteric region and usually into the diaphysis. Frequently there are fracture lines in at least two planes, and consequently fixation must be similarly oriented in two orientations [3]. Many years later, the Fielding classification, described by Dr Joseph William Fielding in 1973, was one of the first to describe fracture patterns within the subtrochanteric region. His classification focused predominantly on the anatomic location of the fracture with Type 1 fractures occurring at the lesser trochanter, Type 2 between 2.5 and 5 cm below the lesser trochanter, and Type 3 occurring 5 to 7.5 cm below the lesser trochanter [7, 8]. Although it was at the forefront of describing subtrochanteric fractures, the Fielding classification was limited as a result of its low reproducibility and its inability to describe more complex fracture patterns [1, 7, 8, 10]. In 1978, Dr Frank Seinsheimer III developed the Seinsheimer classification, a dedicated classification of subtrochanteric femur fractures that incorporated the number of fractured fragments as well as involvement of the medial or lateral cortex [1, 19]. At the time he noted that mechanical failure of pertrochanteric femur fracture fixation often exceeded 20%. Even so, few except Magliato and Fielding had presented a classification system of proximal femur fractures that attempted to describe morphology patterns that contributed to this higher incidence of treatment failure [19]. To have been included in his study, part of the fracture plane must lie between two horizontal lines: (1) at the level of the inferior aspect of the lesser trochanter; and (2) another line 5 cm below that. Notably, as a result of these parameters, this classification system included intertrochanteric femur fractures with distal extension as well as diaphyseal fractures with proximal extension provided that a fracture line crossed this zone. Fractures were classified based on postinjury radiographs and were divided into eight categories based on the number of major fragments, the shape of the fragments, and their location. The first, Type I fractures, are nondisplaced or had < 2 mm of displacement. Type II fractures are two-part fractures and are further subdivided into (II-A) transverse fractures; (II-B) spiral fractures with the lesser trochanter attached to the proximal segment; and (II-C) spiral fractures with the lesser trochanter attached to the distal segment. Type III fractures are three-part fractures and are subdivided into two subcategories: (IIIA) a spiral fracture with the lesser trochanter as part of a third fragment as well as an inferior spike of cortex; and (IIIB) a spiral fracture of the proximal one-third of the femur with the third fragment consisting of a butterfly fragment. Type IV fractures are comminuted fractures with four or more fragments. Finally, Type V fractures are any subtrochanteric fractures with extension through the greater trochanter. Through this work, Seinsheimer concludes that Type IIIA fractures have the highest incidence of fixation failure [19]. Although descriptive, the Seinsheimer classification offered little guidance on what treatment approach to use for a given fracture [1, 19]. In 1992, Dr Thomas Russell and Dr John Charles Taylor published their classification system to describe subtrochanteric fractures. The Russell-Taylor classification accounts for involvement of the fracture within the piriformis fossa and its extension into the lesser trochanter [1, 18]. The aim of this classification was to guide surgeons in treating the injury and to assist with implant selection, which differed from other, established subtrochanteric fracture classifications [1, 18]. By underscoring the unique anatomy of the proximal femur, particularly the extension of the fracture plane into the greater trochanter (lateral wall) or the presence of stability in the lesser trochanteric region (medial cortex), this descriptive schema identifies four different fracture morphologic patterns and was developed to provide the surgeon with direction in determining the proper modality of surgical fixation [5]. All in all, there are > 15 described classification schemas for fractures of the subtrochanteric region [1, 15]. Many classification systems were generated based on the anatomic location of fractures, number of fractured fragments, degree of comminution, or simply on the perceived stability of the fracture pattern [1]. Of these, the Russell-Taylor classification was one of the first to incorporate treatment options for surgeons and has proven useful in identifying the presence or absence of disruption of the lesser trochanter (medial calcar) and the greater trochanter (piriformis fossa). Although there is no predominantly accepted classification scheme for this fracture pattern, the system described by Russell and Taylor is used fairly often in the orthopaedic community as a result of its simplicity and emphasis on surgical planning and treatment [2, 11, 20, 23]. Because this classification system remains one of the most recognizable and routinely used for subtrochanteric fractures, the purpose of this article is to investigate and critically analyze the Russell-Taylor classification and its effectiveness and use in orthopaedic surgery today, particularly as a tool to guide the identification of stable or unstable fracture patterns and to predict the need for reoperation in the perioperative period [9, 17, 23]. Purpose The aim of this classification was to guide surgeons in determining the integrity of the proximal fragment and whether plating or intramedullary nailing would be the ideal mode of fixation [18]. First, the classification identified whether the piriformis fossa was disrupted and, second, evaluated the extent of lesser trochanter involvement (Fig. 1) [12]. This information was felt to be critical in deciding whether to use a standard locking intramedullary nail that extended from the greater to the lesser trochanter or to perform reconstruction nailing with a cephalomedullary device [11, 18, 20, 22]. One reason the integrity of the piriformis fossa was highlighted in this system was the previously held belief that fractures disrupting the piriformis fossa may benefit from plating rather than nailing; surgeons were concerned that fixation of the proximal fragment would be lost with use of a piriformis starting point in these situations [1, 11, 18]. As a result of disruption of the piriformis fossa, like in Type 2 fractures, Russell et al. suggested that hip screw implants were a superior option, rather than intramedullary devices, when the fracture pattern involved the piriformis fossa (Types 2-A and 2-B) [9, 10, 13]. However, with more advanced nailing techniques through the greater trochanter, the nail “falling out of the back” of the proximal fragment has become less problematic [11, 18] and ultimately diminished the utility of the Russell-Taylor classification system [1, 11, 18].Fig. 1: This image describes the four types of fractures described by the Russell-Taylor classification system. Type 1-A fractures describe fractures that do not involve the piriformis fossa. Type 1-B fractures do not involve the piriformis fossa but do involve the lesser trochanter. Type 2-A describes fractures through the piriformis fossa but not involving the lesser trochanter. Type 2-B are fractures involving both the lesser trochanter and piriformis fossa [14, 18]. Reprinted with permission from N. G. Lasanianos and N. K Kanakaris et al. [10].Classification Type 1: Fractures not involving the piriformis fossa of the femur. Subdivided into: Type 1-A: Fractures that extend below the lesser trochanter of the femur; and Type 1-B: Fractures that involve the lesser trochanter of the femur. Type 2: Fractures that do involve the piriformis fossa of the femur. Subdivided into: Type 2-A: Fracture patterns with a stable medial buttress; and Type 2-B: Fracture patterns with no medial femoral cortex stability. Validation There are multiple classification systems that have been developed to describe subtrochanteric fractures [1, 15]. Two studies have analyzed the inter- and intraobserver reliability of the Russell-Taylor classification system: Guyver et al. [10] and Imerci et al. [12]. Guyver et al. [10] analyzed various subtrochanteric descriptive classifications, one of which was the Russell-Taylor classification system. In this study, four observers analyzed 32 subtrochanteric fractures and compared the reliability of the classification system. The observers were asked to evaluate the same fracture patterns on two separate occasions to assess intraobserver reliability [7]. Their study concluded that the mean κ value for interobserver reliability and intraobserver reliability was κ = 0.35 (fair) and κ = 0.25 (fair), respectively. However, the study by Imerci et al. [12] had an alternative conclusion. Their study utilized 35 patients with subtrochanteric fractures that were classified 6 weeks apart by 16 observers, eight of whom were specialists and eight of whom were assistants [12]. Imerci et al. [12] concluded that there was substantial agreement regarding interobserver reliability of the specialists with κ = 0.695 and 0.754 on first and second evaluations, respectively. Similarly, there was also substantial agreement for the interobserver score of assistants with κ values of 0.695 and 0.749 on their first and second evaluations, respectively, as well. Furthermore, the intraobserver reliability scores among specialists were determined as nearly perfect at 0.955 (range, 0.75-1.00) and 0.915 (range, 0.82-0.97) and 0.855 (range, 0.75-0.96) and 0.90 (range, 0.79-0.97) for the assistants. Based on the results by Guyver et al. [10], the Russell-Taylor classification’s low inter- and intraobserver reliability would indicate that this system has a high risk of miscommunication, misdiagnosis, and erroneous surgical decision-making. In contrast, Imerci et al. [12] advocated that the Russell-Taylor classification was more reliable and reproducible than other classifications for subtrochanteric fractures analyzed in their study. Imerci et al. [12] stated that, as a result of a larger volume of images analyzed and more observers participating in evaluation, their study was more reliable than the Guyver et al. [10] study. Limitations According to Guyver et al. [10], the greatest limitation of the Russell-Taylor classification is that the intra- and interobserver reliability of this system is limited, typically in the range of “fair” or worse [10]. As noted, this means that the classification should be used with great care, if at all. However, Imerci et al.’s [12] results showed almost perfect repeatability and the interobserver reliability was substantial, encouraging the overall use of this classification system among surgeons. The stark contrast between these studies indicates that additional studies should be performed to better evaluate the descriptive capacity of the Russell-Taylor classification system. Regardless of the inter- and intraobserver reliability in identifying and describing these fractures, this classification system continues to have limited value with regard to treatment approach [1, 10]. As stated, the Russell-Taylor classification was created primarily to guide surgeons with surgical treatment. When the system was developed, fracture patterns that did not extend into the piriformis fossa (mainly Russell-Taylor Type 1 fractures) were best treated with intramedullary nailing, typically using a cephalomedullary nail with a piriformis entry point [15]. However, with newer orthopaedic implant designs and expanding surgical indications, most subtrochanteric fractures now are treated with intramedullary nails regardless of their Russell-Taylor classification, further limiting the system’s utility [1, 10, 11, 14]. Although the Russell-Taylor classification’s reproducibility and reliability are debated among authors, the addition of trochanteric-starting intramedullary nails have diminished this system’s overall clinical relevance. Starr et al. [21] performed a randomized study that found no difference in outcome regarding a trochanteric versus piriformis fossa entry portal in fixation of this fracture pattern. They suggested that the quality of the fracture reduction was more important than whether the piriformis fossa was disrupted [21]. Another limitation of this classification system is that the degree of comminution is often not fully appreciated on radiographic imaging, especially in elderly patients with osteoporotic bone [23]. As such, intraoperative findings often reveal additional complexity to the injury, further complicating surgery. Finally, the lack of well-defined parameters encompassing the subtrochanteric anatomic region adds an element of vagueness to the already complex issue at hand. Loizou et al. [15] identified 15 different classification methods for subtrochanteric fractures with only eight classification systems defining which area of the bone was the subtrochanteric region. Although generally accepted as 5 cm distal to the lesser trochanter, Loizou et al. [15] found that there was no anatomic agreement among the different classifications on the distal border of the subtrochanteric region, particularly in those fractures that crossed anatomic boundaries [15]. Conclusion An ideal classification system needs to be reliable, reproducible, and able to accurately communicate between surgeons involved in care [3]. The Russell-Taylor classification, although not perfect with regard to reliability and reproducibility, remains one of the most widely known classifications for subtrochanteric femur fractures [2, 11, 20, 23]. With advancements in reduction techniques, nailing procedures, and improved versatility of implants, there is decreased clinical utility of the Russell-Taylor classification system. Although the classification system aids in describing the fracture’s pattern, at this point, this system is of limited clinical relevance. Recognizing the variability in describing these fracture patterns, as concluded by Guyver et al. [10] and Imerci et al. [12], the authors of this study recommend caution while using the Russell-Taylor classification to describe fracture patterns. Furthermore, we recommend against its use to guide treatment of patients with subtrochanteric femur fractures. With such a commonly used classification system, additional studies must be done to resolve and conclude whether the reproducibility and reliability of this classification system are consistent and efficacious to maximize the potential to treat patients safely [2, 11, 20]. Given the complex nature of this fracture pattern and the detrimental effect on quality of life after injury, one can infer that an all-inclusive and fully descriptive classification system would also be equally complex [6]. Nonetheless, because of its historic significance to the study, orthopaedic trauma, and its relative familiarity, the Russell-Taylor classification may serve a role in describing the initial fracture pattern, even as the advent of trochanteric starting intramedullary nails and advanced orthopaedic implants have limited its clinical relevance.