Femoral bone defect classifications in revision total hip arthroplasty: a comprehensive review and proposal of a new algorithm of management

Gross classification (1985)

The first description of femoral defects by Gross et al. included only two categories: the calcar and the proximal femoral defect. The proposed management of both lesions included the use of allografts (17, 18). Subsequently, a new version has been proposed distinguishing the defect in intraluminal and cortical (5) (Table 1). The intraluminal defect was characterized by an intact cortex which is supportive enough to allow the insertion of the implant. The cortical defect can be non-circumferential or circumferential and requires the use of allografts.

A modified version of Gross classification (15) (Table 2) has been suggested by other authors evaluating the localization and extension of the defect in five different types of bony lesion. The extension of the defect in the axial plane was described as intraluminal, cortical, or circumferential. The longitudinal length of the defect was expressed as more or less than 5 cm extending distally from the lesser trochanter. The periprosthetic fracture associated with the bony defect characterized the most severe lesion (type V). The modified Gross classification (15) demonstrated very good intra-observer reliability (K = 0.75) and validity (K = 0.75).

Gustilo and Pasternak classification (1988)

The Gustilo and Pasternak classification (18) was developed to evaluate failed cemented stems (Table 3) including four categories based to the extent of the defect. The type I was characterized by a minimal endosteal or inner cortical bone loss. The type II delineated a proximal canal enlargement with an intact circumferential wall. In the most severe cases, the defect involved the lesser trochanter (type III) or it determined a total proximal circumferential loss below the lesser trochanter (type IV). The authors suggested a management mainly based on cementless stems eventually associated to bone grafts.

Endo-Klinik classification (1988)

The Endo-Klinik classification evaluated the femoral bony defects associated with cemented stems (19) and demonstrated an excellent intra-observer (K = 0.83) and inter-observer (K = 0.84) agreement (20).

The classification focused on the location and the proximal to distal extension of radiolucent lines taking into account the involvement of femoral isthmus considered a critical anatomical structure to obtain a proper distal fixation (Table 4). In type I defects, the authors suggested to perform femoral osteotomies to prevent fractures during the removal of fixed femoral stems. In type II defects, they proposed to fill large defects with fracture plates embedded in the cement (‘reinforced concrete’ technique) and to use standard or long cemented stems basing on the extension of the femoral defect. In type III defects, several strategies have been suggested from long cemented stems to combined constructs including bone grafts, cement embedded hardware, and revision prostheses. The type IV defect was managed with total femoral replacement or resection arthroplasty.

Mallory classification (1988)

The Mallory classification (21) described the proximal femoral defects in cemented stems (Table 5) evaluating the type of the bone affected (cortical and/or cancellous) and the longitudinal extension of the lesion. The authors identified three different zones according to anatomical bony landmarks: zone 1, proximal to lesser trochanter; zone 2, between lesser trochanter and femoral isthmus; and zone 3, distal to the isthmus. The management was based on the residual bone stock, with a preference for the long cementless stems to obtain a distal fixation. The optional use of proximal allografts has been also considered if the proximal part did not provide the required support for the new implant.

Engh classification (1988)

The Engh classification (22) provided a qualitative description of the defect including three main patterns: minimal, moderate, and severe damage (Table 6). The proximal bone deficiency affecting the femoral neck and metaphysis corresponded to a moderate damage, whereas the distal deficiency involving both femoral metaphysis and diaphysis characterized the severe damage.

The most common defects were represented by moderate-to-severe bone loss (types II and III) characterized by the irreversible sclerosis with loss of osteogenic potential of the femoral cortical bone. The authors suggested the use of stem with proximal or full porous coating instead of cemented stems to achieve a stable primary implant fixation. Whether distal fixation was not possible, the use of large implants to replace the portion of bone loss, bipolar components, and total acetabular allograft has been proposed.

Chandler and Penenberg classification (1989)

The Chandler and Penenberg classification (23) described the femoral bony defect evaluating the location and the extension of the lesion, and the type of bone involved (Table 7). In the types I and II, the location was defined by anatomical structures such as the calcar and greater trochanter. In the types III, IV, and VI, the general features of cortical thinning, cortical perforation, and circumferential deficiencies were taken into account. The type V was characterized by the presence or not of periprosthetic fractures. The authors identified different causes for each of the six types: stress shielding for types I and II, stem loosening for type III, stem drifting to varus for type IV, host bone fractures for type VA, or previous allograft–prosthetic composite fracture for type VB. The authors proposed three main therapeutic strategies: the strut or napkin-ring allografts for all types of defect except for the type VI, the allografts and/or long stems for the type VI, and internal plates and screws for bone fixation in allograft fractures.

American Academy of Orthopaedic Surgeons classification (1993)

The American Academy of Orthopaedic Surgeons (AAOS) classification was characterized by a three-step assessment of the bony defect including both preoperative and intraoperative findings (Table 8) (24). It demonstrated an excellent intra-observer (K = 0.81) and a very good inter-observer (K = 0.65) agreement (19).

The first step was focused on the evaluation of the morphology of the defect and the shape of the femur describing six categories: segmental, cavitary, combined (segmental and cavitary), malaligment, femoral stenosis, and femoral discontinuity or femoral fractures. The segmental defect was defined as any loss of bone in the cortices; the cavitary defect was characterized by a medullary-contained lesion or also an ectasia, if thinning of the diaphyseal cortex is present. The femoral malalignment is characterized by a rotational or angular distortion of the diaphysis. The femoral stenosis was defined as the narrowing of the femoral canal. The femoral discontinuity described the lack of bony integrity resulting from a femoral fracture regardless the presence of the implant. The second step focused on the defect extension distinguishing three levels: level I, proximal to the inferior portion of the lesser trochanter; level II, up to 10 cm from the lesser trochanter; and level III, more than 10 cm from the lesser trochanter. Finally, the third step focused on the intraoperative findings according with the degree of prosthesis–bone contact.

The three-step evaluation could provide a standard classification system to describe preoperatively the defect and to report the result of the surgical reconstruction. Basing on the current state of art procedures, the author suggested different surgical strategies from bone grafting to cementless stems.

Paprosky classification (1993)

The Paprosky classification (25, 26, 27, 28) has been developed for the assessment of preoperative radiographs to classify the location and the extent of femoral bone loss (Table 9 and 10).

The classification evaluated the type of bone affected (cancellous and cortical) and the involved portion of the femur (metaphysis and diaphysis). In the most severe cases, characterized by diaphyseal bone defect, the authors focused on the integrity of the isthmus to achieve or not 4 cm scratch fit (Table 10). For each type of the defect type defined, the authors reported the most common etiology. The type I defect resulted by the removal of a cementless femoral component without a biological ingrowth surface. The type II defect occurred after the removal of a cemented or cementless femoral component, in the earlier stages of loosening. The type IIIA defect was commonly seen after the removal of a grossly loose femoral component that was inserted with first-generation cementing techniques. The type IIIB defect often resulted by the failure of a cemented femoral component that was inserted with a cement restrictor or a cementless femoral component that was associated with substantial distal osteolysis. No specific causes have been described for the type IV defect.

The management of type I defect was the same for primary replacement with the use of cemented or cementless stems. In type II and III defects, the authors suggested to achieve a diaphyseal fixation with proximally or fully porous-coated cementless stems. The most severe cases were characterized by a non-supportive isthmus with unavailable 4 cm scratch fit. Therefore, the authors proposed the use of modular, tapered, cementless stems or impaction bone grafting or femoral allograft–prosthesis composite. For selected patients, such as elderly or patients with low physical demand, a long femoral cemented or uncemented component or a proximal femoral replacing prosthesis can be used. Finally, the authors recommended the use of the extended trochanteric osteotomy for the removal of a well-fixed cemented or cementless stem. This procedure can also facilitate the insertion of revision stems in the presence of complicated femoral deformities.

Käfer et al. (29) demonstrated that the classification does not provide a valid and reliable assessment of femoral bone stock loss, showing an inter-observer reliability of r _s = 0.45–0.67 and r _s = 0.59–0.68 in the preoperative and intraoperative setting respectively. The authors also reported that the defects were grossly underestimated preoperatively and the level of experience did not influence the validity of measurements. On the other hand, Parry et al. (19) demonstrated a very good intra-observer (K = 0.77) and inter-observer (K = 0.71) agreement.

Deutschen Gesellschaft für Orthopädie und Traumatologie classification (1997)

The Deutschen Gesellschaft für Orthopädie und Traumatologie (DGOT) classification (30) focused on the type of bone involved (cancellous and cortical) and the proximal-to-distal extension of the lesion (intramedullary, trochanteric, calcar, metaphysis, diaphysis) (Table 11). No management has been included. Käfer et al. (29) reported that the classification provides a poor reliable assessment of femoral bone stock loss, showing an inter-observer reliability of r _s = 0.38–0.59 and r _s = 0.44–0.60 in the preoperative and intraoperative setting, respectively.

Modified Mallory classification (1999)

The modified Mallory classification (31) (Table 12) focused on the type of bone involved (cancellous and cortical) and the proximal-to-distal extension of the lesion (metaphysis and diaphysis). Moreover, the authors described the quality of the remaining bone reporting the ability or not to support the femoral implant. In type II defect, the metaphyseal support is intact but compromised, whereas the metaphysis is not supportive in type III defect. Finally, the type IV defect is characterized by the involvement of the diaphysis with a non-supportive isthmus. The proposed management ranges form primary implants to impaction bone grafting or fully porous-coated stems or cemented stems combined with structural femoral allograft for proximal reconstruction.

Italian Society for Revision Arthroplasty classification (2000)

The Italian Society for Revision Arthroplasty (GIR) classification has been developed to correlate the femoral bone defects with surgical strategies for the reconstruction (32) (Table 13). A new concept introduced by this classification was the dynamic grading of the defect. Because the bone is continuously reshaping, emphasizing the possible evolution not only overtime but even intra-operatively of the defect, the authors decided to not distinguish bone loss into types but they used different grades to describe the evolution of the defect.

The classification assessed quality, extension, and location of the defect, distinguishing between medullary loss and cortical thinning. The description of the lesion aimed to identify the ability of the remaining bone to support femoral implant. In grades I and II, the bone allows the proximal fixation through larger and longer stems. In grades III and IV, the distal fixation and bone reconstruction represent the management of choice.

Saleh classification (2001)

The Saleh classification (33) (Table 14) focused on the extension of the lesion in both axial and longitudinal axes. In terms of the axial extension, the authors distinguished between non-circumferential and circumferential lesion; in terms of the longitudinal extension, they distinguished between lesions involving the femoral diaphysis more or less 5 cm distal to the lesser trochanter. According with the severity of the defect, the management included since cemented and uncemented stems as in primary THA to porous-coated stems, modular stems, impaction bone grafting, and custom implants. This classification demonstrated an excellent inter-observer reliability (K = 0.88) (34).

Parry classification (2010)

The Parry classification (19) (Table 15) focused on the femoral region involved (metaphysis and diaphysis) and the continuity of the supporting structures. The authors distinguished between contained and uncontained defect, both potentially occurring in metaphysis or diaphysis. The main surgical strategies included implants with distal fixation to bypass the defect and bone grafts. This classification demonstrated an excellent intra-observer agreement (K = 0.87) and a good inter-observer agreement (K = 0.59).

Functional bone stock classification

The functional bone stock classification (FBSC) is a new classification proposed by the authors (Table 16) that subdivides femoral bone defects in three stages based on two different features: the ‘functional’ residual bone stock and the extension of the defect. The ‘functional’ residual bone stock has been defined as the remaining cortical and medullary femoral bone available to obtain a primary press-fit fixation of the stem and a secondary biological fixation by the osteointegration. The quantitative assessment of the bone stock was performed with the cortico-femoral index (CFI), which is defined as the ratio of the femoral diaphyseal diameter minus the intramedullary canal diameter to the femoral diaphyseal diameter (35) (Fig. 1). The ‘functional’ residual bone stock is characterized by a CFI ≥0.5 since the average healthy CFI is ≥0.5 (36). Indeed Nguyen et al. (36) found a correlation between the decrease in CFI below 0.5 and a decrease in bone mineral density and the fracture risk assessment tool. In terms of the extension of the lesion, the authors distinguished in minimal proximal defects, and medium-to-very large defects localized within or over 150 mm from the apex of the great trochanter with subsequent impairment of the femoral isthmus.

View Full Size

A: Stage 3 according to the FBSC, characterized by CFI: 0.29. B: Stage 2 according to the FBSC, characterized by CFI: 0.65.

Citation: EFORT Open Reviews 9, 3; 10.1530/EOR-21-0088

• Download Figure
• Download figure as PowerPoint slide

In the management algorithm, the authors also considered the use of trochanteric osteotomy (short or extensive) as technique to remove the stem and also to preserve the residual trochanteric bone stock.

The stage 1 is characterized by good quality of the bone, absent or minimal bone defect with proximal localization, and a possible economic removal of the stem. The revision can be performed with cementless stem with rectangular or conical section. Small bone defects can be filled with autologous or synthetic bone transplants. In patients with previous cemented stem with a continuous and stable mantle, the ‘cement-within-cement’ technique (34) can be performed.

The stage 2 is characterized by poor quality of the bone, cavitary or segmentary bone defect with medium-to-large size, localized within 150 mm from the apex of the great trochanter. Therefore, with a CFI at 150 mm ≥0.5 (Fig. 1A). The removal of the stem is not simple with consequent but limited provoked bone defect. In this stage, the residual functional bone stock still remains at less than 150 mm from the apex of the great trochanter without impairment of the anatomical femoral isthmus. For this reason, an eventual trochanteric osteotomy should not be extensive. The revision can be performed with a monoblock straight stem with length less than 250 mm (Wagner Cone and Wagner SL 190 and 225). The Wagner Cone and Wagner SL stems are cylindrical implants with diaphyseal fixation, respectively proximal and distal, with longitudinal fins providing the rotational stability. Whether a significant thinning of the cortices with a widening of the femoral canal occurs, the proximal diaphyseal fixation of conical stems cannot be obtained. However, a proximal primary fixation is still possible basing on the functional residual bone stock. Therefore, porous modular stems shorter than 250 mm allow to obtain a diaphyseal press-fit combining a distal component of larger diameter with an appropriate proximal component for the reconstruction of the correct length and offset.

The stage 3 is characterized by poor quality of the bone, extended cavitary or segmentary bone defect, localized more than 150 mm from the apex of greater trochanter. Therefore, with a CFI at 150 mm <0.5 (Fig. 1B). The removal of the stem is probably difficult and destructive. In this stage, the femoral isthmus is partially or completely non-supportive and the residual functional bone stock is localized distally at more than 150 mm from the apex of the greater trochanter. Therefore, the trochanteric osteotomy should be extended, if required. The revision can be performed with monoblock cylindrical stems (e.g. Wagner), straight or curved modular stems, with or without the locking system longer than 250 mm. The choice between the different types of stem depends on the type of filling needed for the cavity, because they have taper of different degrees. The modular cementless, conical, distal fixating, porous titanium stems have a more conical distal geometry than the Wagner stem, which allows the isthmus to be engaged more effectively providing stability of the stem also with a diaphyseal contact of 2–3 cm. The curved stems aim to maximize the cortical contact and primary stability following the native femoral shape. Finally, the stems with locking systems can be useful if it is not possible to obtain the primary stability by press-fit. In any case, the secondary biological stability is obtained through the osteointegration of the porous surface.

The inter-reliability evaluation between the two orthopedic surgeons expressed as kappa coefficient was 0.504 on occasion 1 and on occasion 2, 2 weeks later 0.694. Instead the intra-reliability evaluation between occasion 1 and occasion 2 was 0.84 in the first orthopedic surgeon and 0.64 in the second one.