Field testing the Unified Classification System for peri-prosthetic fractures of the pelvis and femur around a total hip replacement

Transcription

1 M. H. Vioreanu, M. C. Parry, F. S. Haddad, C. P. Duncan From Department of Orthopaedics, University of British Columbia, Canada M. H. Vioreanu, MCh, MD, FRCS, Fellow in Oncology and Reconstructive Orthopaedic Surgery M. C. Parry, BSc (Hons), MBChB, MD, FRCS, Fellow in Oncology and Reconstructive Orthopaedic Surgery C. P. Duncan, MD, MSc, FRCS(C), Professor University of British Columbia, Department of Orthopaedics, Vancouver, British Columbia, Canada. F. S. Haddad, BSc MD (Res), FRCS (Tr&Orth), Professor of Orthopaedic Surgery University College London Hospitals, 235 Euston Road, London NW1 2BU, UK. Correspondence should be sent to Mr M. C. Parry; The British Editorial Society of Bone & Joint Surgery doi: / x.96b $2.00 Bone Joint J 2014;95-B: Received 29 April 2014; Accepted 4 July 2014 HIP Field testing the Unified Classification System for peri-prosthetic fractures of the pelvis and femur around a total hip replacement AN INTERNATIONAL COLLABORATION The Unified Classification System (UCS) emphasises the key principles in the assessment and management of peri-prosthetic fractures complicating partial or total joint replacement. We tested the inter- and intra-observer agreement for the UCS as applied to the pelvis and femur using 20 examples of peri-prosthetic fracture in 17 patients. Each subtype of the UCS was represented by at least one case. Specialist orthopaedic surgeons (experts) and orthopaedic residents (pre-experts) assessed reliability on two separate occasions. For the pelvis, the UCS showed inter-observer agreement of (95% confidence intervals (CI) to 0.876) for the experts and (95% CI to 0.767) for the preexperts. The intra-observer agreement for the experts was (95% CI to 0.963) and (95% to 0.918) for the pre-experts. For the femur, the UCS showed an interobserver kappa value of (95% CI to 0.845) for the experts and a value of (95% CI to 0.773) for the pre-experts. The intra-observer agreement was (95% CI to 0.973) for the experts, and (95% CI to 0.892) for the pre-experts. This corresponds to a substantial and almost perfect inter- and intra-observer agreement for the UCS for peri-prosthetic fractures of the pelvis and femur. We hope that unifying the terminology of these injuries will assist in their assessment, treatment and outcome. Cite this article: Bone Joint J 2014;96-B: Peri-prosthetic fracture about a total hip replacement (THR) is uncommon and has an estimated incidence of 1.7% within ten years of implantation. 1 The one-year mortality is 11% and is comparable with that of fractures of the hip. 2 Risk factors include female gender, patient age and revision procedures, where the incidence increases to 4.2%. 1,3 The anticipated increase of 137% in the demand for THR over the next 15 years, 4 coupled with an ageing population and a decreasing age at which primary THR may be considered, 5 is likely to result in an increased incidence of periprosthetic fractures of the hip, a trend that has been recognised for a number of years. 6 Peri-prosthetic fractures of the acetabulum in association with THR occur less frequently than those of the femur. Generally, they occur intraoperatively, particularly at the time of impaction of a large diameter uncemented acetabular component. 7 They can also occur as a late complication due to trauma or osteolysis. 1,8 The incidence of pelvic fracture in association with the acetabular component of a THR has been estimated at 0.4%, 2,7 however, it rarely occurs in association with a cemented acetabular component. In most cases, acetabular fractures which occur intra-operatively have little or no effect on component stability. 2,3,9-11 The principles of management are based on the location of the fracture, the stability of the implant and the adequacy of the remaining bone stock. A number of classification systems have been proposed to assist in the preoperative analysis and management of periprosthetic fractures of the acetabulum. 4,8,12-14 They vary in the description of the important principles of management. Furthermore, they have not included peri-prosthetic fractures which affect the pelvis beyond the acetabulum. Many of the historical classification systems for peri-prosthetic fractures of the femur have been superseded by the Vancouver classification system (VCS). 6,20 This system is based on the key principles of management of peri-prosthetic fractures, has demonstrated 1472 THE BONE & JOINT JOURNAL

2 FIELD TESTING THE UNIFIED CLASSIFICATION SYSTEM FOR PERI-PROSTHETIC FRACTURES OF THE PELVIS AND FEMUR AROUND A THR 1473 Fig. 1 Anteroposterior radiograph of the pelvis showing a B3 type peri-prosthetic fracture of the right femur (arrows). The prosthesis is loose and there is deficient bone stock in the proximal femur. Fig. 2 Anteroposterior radiograph of the left proximal femur and hemipelvis showing a type E fracture of the proximal femur with an associated fracture of the acetabulum (arrows). substantial inter- and intra-observer reliability and has become widely accepted. 7,21-23 The Unified Classification System (UCS), as applied to THR, has been proposed to expand the VCS to include three other types of fracture that may occur in combination or in isolation, and to deal with the pelvis as a whole (not just the acetabulum). The Unified Classification System (UCS) The purpose of the (UCS) is three-fold: 1) To expand the VCS to include three fracture types which have unique principles of management; 2) To apply it in a principles-based manner to all periprosthetic fractures regardless of the bone broken and the joint involved; 3) To introduce a simple and common language to assist with the evaluation and treatment of these injuries, and perhaps assist with outcomes measurement. The following mnemonic is offered to assist understanding and recall of the UCS fracture types: - Type A: Apophyseal. - Type B: Bed of the implant or close to it. - Type C: Clear of the implant bed. - Type D: Dividing one bone which supports two joint replacements. - Type E: Each of two bones supporting one joint replacement. - Type F: Facing or articulating with an implant. As in the original VCS, 20 Type B fractures are subdivided on the basis of implant stability and remaining bone stock. In subtype B1, the implant is well fixed; B2 fractures describe a loose implant within a well-preserved bone stock and B3 fractures are those in which the implant is loose and the remaining bone stock is poor or deficient. In relation to the pelvis, all six types are applicable (A, B, C, D, E, F) and Type B subdivided (B1, B2, B3). Furthermore, the UCS is applied to the whole pelvis, not just the acetabulum. When applied to the femur, the UCS incorporates the previous VCS 20 but is expanded to include two new fracture patterns, types D and E. The type F fracture pattern applies only to the acetabulum when using the UCS for fractures in association with THR. Description of the Unified Classification System (UCS): femur. Type A (apophyseal): greater or lesser trochanter; Type B (bed of the implant): around or close to the femoral stem (stem well fixed if B1, stem loose if B2, loose and poor bone if B3 (Fig. 1)); Type C (well clear of the implant): the femur distant from the implant; Type D (dividing the femur between two implants): between a hip and knee replacement; Type E (each of two bones supporting a joint replacement): both the femur and acetabulum (Fig. 2); Type F: does not apply to the femur. Description of the Unified Classification System (UCS): pelvis. Type A (apophyseal): anterior superior or inferior iliac spine, or ischial tuberosity; Type B (bed of the implant): the acetabulum (subtype description below); Type C (clear of the implant): ilium, superior and/or inferior rami; Type D (dividing the pelvis between implants): a fracture of the pelvis complicating bilateral replacements; Type E (each of two bones supporting one joint replacement): acetabulum and femur; Type F (facing a joint replacement): fracture of the acetabulum after hemiarthroplasty. The Type B (acetabular) subtypes are as follows: - B1: Fracture of the acetabular lip, wall or floor, which does not affect the stability of the component which is still well fixed. VOL. 96-B, No. 11, NOVEMBER 2014

3 1474 M. H. VIOREANU, M. C. PARRY, F. S. HADDAD, C. P. DUNCAN Fig. 3 Anteroposterior radiograph of the pelvis showing an intraoperative B2 type peri-prosthetic fracture of the acetabulum (arrow), which became obvious three days later when the acetabular component became displaced. The acetabular prosthesis is loose and there is sufficient bone stock to support fixation of the posterior column with exchange of the acetabular component. Fig. 4 Anteroposterior radiograph of the pelvis showing a late B3 type peri-prosthetic pelvic fracture with pelvic discontinuity. - B2: Fracture of the acetabulum (such as the posterior column) or pelvic discontinuity, with a loose acetabular component but adequate bone stock to support an uncomplicated acetabular component revision; with or without bone graft; with or without plated fixation of the fracture (Fig. 3). - B3: Fracture of the acetabulum, with a loose component, and severe bone loss such that complex reconstruction or a salvage procedure is required (Fig. 4). The purpose of these distinctions, similar to Type B subtyping in the femur, is to emphasise the importance of implant loosening and degree of bone loss (volume or strength) when planning treatment. The most common fracture of the pelvis is Type B, within which the most common sub-type is B1 in the acute setting during operation, and B2 or B3 in the delayed setting of revision hip replacement. For each fracture pattern, the success of treatment is dependent on the stability of the component and the available bone stock for reconstruction, both important factors that are incorporated in the UCS. While each fracture pattern is more common at differing times, 24 they can occur early or late, at the time of implantation or following injury, and during primary or revision surgery. The UCS emphasises the important factors that should guide treatment and predict successful outcome: these are implant stability and the available bone stock for reconstruction. The cause of the fracture and its time of occurrence (early or late, acute or chronic) are considered of secondary importance, although they do feature in the treatment plan. It is acknowledged that a B2 or B3 fracture as a late or delayed presentation, secondary to implant migration and osteolysis, such as chronic pelvic discontinuity or dissociation, is more challenging to treat successfully because the potential for fracture healing is compromised. However, it is believed that the quantity and quality of the remaining bone stock is dominant in choosing the most appropriate treatment plan, hence the proposed distinction between B2 and B3. Therapeutic application of the UCS: femur and pelvis Type A. Depends on the importance of the structure attached to the apophysis to the health and function of the adjacent joint replacement, especially if displaced. Some can be observed, such as the lesser trochanter, iliac spines and ischial tuberosity. 24 By contrast, this type would require intervention, such as a fracture of the greater trochanter. Type B. Depends on the subtype. If B1 (acetabular component or stem are well fixed), then observation for the pelvis and reduction with fixation for the femur, ideally using the principles of indirect reduction and minimally-invasive plate osteosynthesis, 25 will suffice. If B2 (acetabular component or stem loose) then revision is required. If B3 (implant loose and poor bone health), then complex reconstruction is required. Type C. Standard management of the pelvic fracture, and standard management for the femoral fracture, would apply. In most cases this would be reduction and fixation. Type D. This requires the consideration of each joint replacement separately, which will lead to a rational treatment plan involving revision of one, both or neither joint replacement. Type E. Again each component (at the pelvis and the femur) would need to be considered separately to plan logical treatment. Type F. This will depend on the degree of fracture displacement. If mildly displaced, non-operative management with protected weight-bearing would be reasonable, with revision total hip replacement later, once the fracture has united, if indicated. If displacement was more marked, then THE BONE & JOINT JOURNAL

4 FIELD TESTING THE UNIFIED CLASSIFICATION SYSTEM FOR PERI-PROSTHETIC FRACTURES OF THE PELVIS AND FEMUR AROUND A THR 1475 Table I. International experts panel participating in the study. North America Australasia Africa Europe Asia C. Della Valle (United States) J. Munro (New Zealand) R. McLennan-Smith (South Africa) K. Gunther (Germany) P. Chiu (Hong Kong) J. Parvizi (United States) D. Howie (Australia) D. Biau (France) M. Dunbar (Canada) E. Masterson (Ireland) N. Ohly (United Kingdom) G. Biring (United Kingdom) Table II. Inter- and intra-observer reliability for the expert and pre-expert panel when assessing the Unified Classification System as applied to both the pelvis and femur. Values expressed are Kappa values with 95% confidence intervals in brackets (without bootstrap analysis). Inter-observer Intra-observer Expert Pre-expert Expert Pre-expert Pelvis (0.798 to 0.876) (0.689 to 0.767) (0.760 to 0.963) (0.688 to 0.918) Femur (0.765 to 0.845) (0.690 to 0.773) (0.867 to 0.973) (0.652 to 0.892) immediate intervention would need consideration, unless contraindicated because of pre-existing co-morbidity. Patients and Methods We retrospectively reviewed all peri-prosthetic fractures of the hip, including the acetabulum, treated at our hospital over a ten-year period (2002 to 2012), totalling approximately 400 fractures. From these, we identified ten pelvic fractures in nine patients, and ten femoral fractures in eight patients, which were associated with THRs. Each subtype of the UCS classification was represented at least once. Each patient was described with a simple clinical scenario and an appropriate radiograph showing the fracture. Two different groups of observers were used to test the reliability of the UCS. The expert group comprised 11 international experts recruited from tertiary referral arthroplasty centres in ten countries (Canada, United States, Australia, New Zealand, France, United Kingdom, Ireland, Germany, South Africa, Hong Kong) on five continents (Table I). The pre-expert group included 17 orthopaedic residents from our hospital in their last two years of training. One week before receiving the survey containing the cases for review, each observer received an information pack outlining the principles of the UCS. The survey was sent to each observer on an electronic survey platform ( Each observer reviewed the cases on two separate occasions with a minimum of six weeks between the reviews. The first was for measurement of inter-observer reliability. For the intra-observer assessment, the same cases were used, though assembled in a different order. There was no communication between observers. All collected data were analysed using the weighted Kappa statistic test for multiple observers to measure interand intra-rater reliability within the experts and preexperts. 26 The weighted Kappa was calculated to measure the intra-rater reliability within the same groups. Kappa values of < 0 indicated poor agreement; 0.00 to 0.20 slight agreement, 0.21 to 0.40 fair; 0.41 to 0.60 moderate; 0.61 to 0.80 substantial and 0.81 to 0.99 almost perfect agreement. 27,28 An independent statistician completed the statistical analysis calculating 95% confidence intervals (CI) for the available data. A bootstrap based analysis was applied to overcome any possible dependence encountered between readings, because for some questions the same radiograph was used. However, for the questions where the same radiograph was used, a block-out analysis was used to compose the different questions thus allowing 95% CI to be applied. The results of the bootstrap analysis are included at least as a sensitivity analysis. Results When applied to the pelvis, the UCS gave an inter-observer reliability for the expert panel of (95% CI to 0.876) which corresponds to an almost perfect agreement (Table II). With respect to the pre-expert panel, the UCS as applied to the pelvis gave a Kappa value of (95% CI to 0.767), which is considered to be substantial agreement. The bootstrap analysis, applied to address any potential concerns of dependence between sets of readings, resulted in an increase in the 95% CI but the agreement remained almost perfect for the expert panel (95% CI to 0.955) and substantial for the pre-expert panel (95% CI to 0.818). The intra-observer weighted Kappa value when the UCS was applied to the pelvis by the expert panel was (95% CI to 0.963), which represents almost perfect agreement, and in the pre-expert group (95% CI to 0.918), which is substantial agreement. The bootstrap analysis resulted in an increase in the 95% CI for both the expert (0.685 to 0.926) and pre-expert panel (0.437 to 0.871), though this had no effect on the degree of agreement. When applied to the femur, the UCS showed almost perfect inter-observer reliability for the expert panel (Kappa value 0.805, 95% CI to 0.845) and substantial inter-observer reliability for the pre-expert panel (Kappa value 0.732, 95% CI to 0.773). As for the pelvis, the VOL. 96-B, No. 11, NOVEMBER 2014

5 1476 M. H. VIOREANU, M. C. PARRY, F. S. HADDAD, C. P. DUNCAN bootstrap analysis resulted in a widening of the 95% CIs (expert panel, to 0.865; pre-expert panel to 0.796), but no effect on the degree of agreement. The intra-observer agreement for the UCS when applied to the femur was almost perfect for the expert panel (Kappa value 0.920, 95% CI to 0.973) and substantial for the pre-expert panel (Kappa value 0.772, 95% CI to 0.892). As previously, the bootstrap analysis broadened the 95% CIs (0.820 to for the expert panel, to for the pre-expert panel) although it did not alter the degree of agreement. Discussion The purpose of any fracture classification system is to allow the identification of comparable cases and to guide treatment. Consequently, a robust, reliable classification of periprosthetic fractures is needed. A number of these have been proposed for peri-prosthetic fractures about the hip. In respect of the femur, Parrish and Jones 15 classified peri-prosthetic fractures in relation to their location, as: trochanteric; proximal; middle; or distal third fractures. Johansson et al 16 classified fractures in relation to the implant as proximal to the tip of the stem; extending beyond the tip of the stem; or entirely beyond the tip. Bethea et al 17 expanded on these classifications to incorporate position and pattern: type A occurring distal to the tip of the stem; type B as spiral fractures around the stem; and type C as comminuted fractures around the stem. Cooke and Newman 18 described four types of peri-prosthetic fracture and emphasised the need for early revision in the comminuted type 1 fracture, while the type 2 transverse fracture around the stem could be managed either conservatively or by revision. They proposed that fractures at or distal to the tip of the stem (type 3 and 4 respectively) required internal fixation. Jensen et al 19 emphasised the importance of implant stability at the time of injury as a predictor for revision of the prosthesis. Fractures were classified as type 1 when located around the proximal two-thirds of the femoral component, type 2 when the fracture extended proximally and distally from the stem tip, and type 3 when the fracture extended distal to the tip of the stem. The VCS built on previous work and proposed an analysis which combined three important variables which should guide treatment, with emphasis on the stability of the implant. 20 This classification system has shown substantial inter-observer agreement in a number of studies Therefore it seems logical that it should form the basis of an expanded, UCS incorporating fractures between prostheses and those that involved both the femur and acetabulum. A number of classification systems have been proposed to address fractures around the acetabular component of a THR and to assist in their pre-operative analysis and subsequent management. Intra-operative acetabular fractures were classified on the basis of in vitro investigation. 12,13 On the basis of the anatomical location four types of fracture were identified: anterior wall, transverse, inferior lip, and posterior wall. Based on 11 cases, Peterson and Lewallen 8 classified post-operative peri-prosthetic acetabular fractures into Type I, with a radiologically and clinically stable acetabular component and Type II, with an unstable acetabular component. Della Valle et al 14 classified acetabular fractures and included both intra-operative and postoperative types. Type-I fractures occurred intra-operatively at the time of insertion of the acetabular component and were sub-classified as type A if they were diagnosed intraoperatively, were undisplaced, and associated with a stable component; as type B if they were diagnosed intra-operatively but were associated with a displaced fracture; and as type C if they were diagnosed post-operatively. Type-II injuries occurred at the time of component removal and were further classified as type A if there was < 50% loss of acetabular bone stock and as type B if it was > 50%. To our knowledge none of these previously described acetabular fracture classification systems has been tested for reliability. The purpose of the UCS is to apply the principles of management of peri-prosthetic fractures more broadly to any joint replacement, regardless of the joint that is involved and the bone that is broken, though the system has only been tested for reliability at the hip. The system expands on, and augments, the broadly accepted VCS to incorporate other fracture patterns, the incidence of which is likely to increase in the future. When applied to the acetabulum, we have shown a substantial inter-observer and intra-observer reliability for the pre-expert panel, and an almost perfect inter-observer and intra-observer reliability for the expert panel. When applied to the femur, this finding was reproduced. To put this in context, the universally accepted classification system for open fractures of the tibia described by Gustilo and Anderson 29 and refined by Gustilo, Gruninger and Davis, 30 shows only a moderate inter-observer agreement depending on the experience of the surgeon. 31 The commonly used classification of proximal humeral fractures described by Neer 32 demonstrates a kappa value of only 0.5, equating to a moderate level of agreement between observers. 33 The high kappa values for the UCS, both between and within observers of different levels of experience, highlights the simplicity and ease of application of the system, which, it is hoped, will translate into a more general application. Further reliability testing of this system is required to assess its application to other prosthetic joints. However, the system should allow greater comparison of cases and ease of decision-making in the treatment of peri-prosthetic fractures about the hip. We thank E. Sayre PhD, for his assistance with the statistical analysis and D. Savoy BA, for her assistance in the preparation of this manuscript. We would like to acknowledge the pre-expert panel of orthopaedic residents in the fourth and fifth years of training at the University of British Columbia: C. Hiller, P. Grunau, P. Voorhoeve, M. Zec, K. Taunton, R. Ghag, L. Dielwart, A. Simmonds, P. Soswa, L. Leveille, J. Strelzow, M. Pelletier, R. Petretta, J. Potter, C. Kong, N. Levy and H. Sadr. THE BONE & JOINT JOURNAL

6 FIELD TESTING THE UNIFIED CLASSIFICATION SYSTEM FOR PERI-PROSTHETIC FRACTURES OF THE PELVIS AND FEMUR AROUND A THR 1477 Each author (CPD) certifies that he has or may receive payments or benefits [e.g., serve as a consultant (CPD), or speakers bureaus (CPD)] from a commercial entity (Zimmer, Inc.); or speakers bureaus (CPD) from a commercial entity (DePuy, A Johnson & Johnson Company) related to this work. The institution of the authors (MHV, MP, CPD) has received funding from Zimmer Inc., DePuy Synthes, A Johnson & Johnson Company, Stryker Orthopaedics and Bayer. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by G. Scott and first proof edited by A. Ross. References 1. Meek RM, Norwood T, Smith R, Brenkel IJ, Howie CR. The risk of peri-prosthetic fracture after primary and revision total hip and knee replacement. J Bone Joint Surg [Br] 2011;93-B: Bhattacharyya T, Chang D, Meigs JB, Estok DM 2nd, Malchau H. Mortality after periprosthetic fracture of the femur. J Bone Joint Surg [Am] 2007;89-A: Singh JA, Jensen MR, Lewallen DG. Patient factors predict periprosthetic fractures after revision total hip arthroplasty. J Arthroplasty 2012;27: Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to J Bone Joint Surg [Am] 2007;89-A: Mota RE, Tarricone R, Ciani O, Bridges JF, Drummond M. Determinants of demand for total hip and knee arthroplasty: a systematic literature review. BMC Health Serv Res 2012;12: Younger AS, Dunwoody J, Duncan CP. Periprosthetic hip and knee fractures: the scope of the problem. Instr Course Lect 1998;47: Haidukewych GJ, Jacofsky DJ, Hanssen AD, Lewallen DG. Intraoperative fractures of the acetabulum during primary total hip arthroplasty. J Bone Joint Surg [Am] 2006;88-A: Peterson CA, Lewallen DG. Periprosthetic fracture of the acetabulum after total hip arthroplasty. J Bone Joint Surg [Am] 1996;78-A: Adler E, Stuchin SA, Kummer FJ. Stability of press-fit acetabular cups. J Arthroplasty 1992;7: Curtis MJ, Jinnah RH, Wilson VD, Hungerford DS. The initial stability of uncemented acetabular components. J Bone Joint Surg [Br] 1992;74-B: Sharkey PF, Hozack WJ, Callaghan JJ, et al. Acetabular fracture associated with cementless acetabular component insertion: a report of 13 cases. J Arthroplasty 1999;14: Callaghan JJ. Periprosthetic fractures of the acetabulum during and following total hip arthroplasty. Instr Course Lect 1998;47: Callaghan JJ, Kim YS, Pedersen DR, Brown TD. Periprosthetic fractures of the acetabulum. Orthop Clin North Am 1999;30: Della Valle CJ, Momberger NG, Praposky WG. Periprosthetic fractures of the acetabulum associated with a total hip arthroplasty. Instr Course Lect 2003;52: Parrish TF, Jones JR. Fracture of the femur following prosthetic arthroplasty of the hip: report of nine cases. J Bone Joint Surg [Am] 1964;46-A: Johansson JE, McBroom R, Barrington TW, Hunter GA. Fractures of the ipsilateral femur in patients with total hip replacement. J Bone Joint Surg [Am] 1981;63- A: Bethea JS 3rd, DeAndrade JR, Fleming LL, Lindenbaum SD, Welch RB. Proximal femoral fractures following total hip arthroplasty. Clin Orthop Relat Res 1982;170: Cooke PH, Newman JH. Fractures of the femur in relation to cemented hip prostheses. J Bone Joint Surgery [Br] 1988;70-B: Jensen JS, Barfod G, Hansen D, et al. Femoral shaft fracture after hip arthroplasty. Acta Orthop Scand 1988;59: Duncan CP, Masri BA. Fractures of the femur after hip replacement. Instr Course Lect 1995;44: Brady OH, Garbuz DS, Masri BA, Duncan CP. The reliability of validity of the Vancouver classification of femoral fractures after hip replacement. J Arthroplasty 2000;15: Rayan F, Dodd M, Haddad FS. European validation of the Vancouver classification of periprosthetic proximal femoral fractures. J Bone Joint Surg [Br] 2008;90-B: Naqvi GA, Baig SA, Awan N. Interobserver and intraobserver reliability and validity of the Vancouver classification system of periprosthetic femoral fractures after hip arthroplasty. J Arthroplasty 2012;27: Masri BA, Meek RM, Duncan CP. Periprosthetic fractures evaluation and treatment. Clin Orthop Relat Res 2004;420: Moore RE, Baldwin K, Austin MS, Mehta S. A systematic review of open reduction and internal fixation of periprosthetic femur fractures with or without allograft strut, cerclage, and locked plates. J Arthroplasty 2014;29: Cohen J. Weighted kappa: nominal scale agreement provision for scaled disagreement or partial credit. Psychol Bull 1968;70: Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med 2005;37: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33: Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand and twenty-five open fractures of long bones: retrospective and prospective analyses. J Bone Joint Surg [Am] 1976;58-A: Gustilo RB, Gruninger RP, Davis T. Classification of type III (severe) open fractures relative to treatment and results. Orthopedics 1987;10: Horn BD, Rettig ME. Interobserver reliability in the Gustilo and Anderson classification of open fractures. J Orthop Trauma 1993;7: Neer CS 2nd. Displaced proximal humeral fractures part II. Treatment of three-part and four-part displacement. J Bone Joint Surg [Am] 1970;52-A: Sidor ML, Zuckerman JD, Lyon T, et al. The Neer classification system for proximal humeral fractures: an assessment of interobserver reliability and intraobserver reproducibility. J Bone Joint Surg [Am] 1993;75-A: VOL. 96-B, No. 11, NOVEMBER 2014