Injury, Int. J. Care Injured 46 (2015) 213–217

The biomechanical effect of bone quality and fracture topography onlocking plate fixation in periprosthetic femoral fractures

Andreas Leonidou a,*, Mehran Moazen b, Panagiotis Lepetsos a,c, Simon M. Graham a,George A. Macheras c, Eleftherios Tsiridis a,d

a Academic Department of Orthopaedics and Trauma, Division of Surgery, Aristotle University Medical School, University Campus, 54 124 Thessaloniki, Greeceb Medical and Biological Engineering, School of Engineering, University of Hull, Hull HU6 7RX, UKc 4th Department of Trauma & Orthopaedics, KAT Hospital, Nikis 2, Kifissia, 14561 Athens, Greeced Department of Surgery and Cancer, Division of Surgery, Imperial College London, B-block Hammersmith Hospital, Du-Cane Road, London W12 0HS, UK

A R T I C L E I N F O

Article history:

Accepted 22 October 2014

Keywords:

Periprosthetic femoral fracture

Bone quality

Fracture angle

Fracture level

Fracture topography

Finite element analysis

A B S T R A C T

Optimal management of periprosthetic femoral fractures (PFF) around a well fixed prosthesis

(Vancouver B1) remains controversial as adequate fixation needs to be achieved without compromising

the stability of the prosthesis. The aim of this study was to highlight the effect of bone quality i.e. canal

thickness ratio (CTR), and fracture topography i.e. fracture angle and its position in relation to the stem,

on the biomechanics of a locking plate for a Vancouver B1 fracture. A previously corroborated simplified

finite element model of a femur with a cemented total hip replacement stem was used in this study.

Canal thickness ratio (CTR) and fracture topography were altered in several models and the effect of

these variations on the von Mises stress on the locking plate as well as the fracture displacement was

studied. Increasing the CTR led to reduction of the von Mises stress on the locking plate as well as the

fracture movement. In respect to the fracture angle with the medial cortex, it was shown that acute

angles resulted in lower von Mises stress on the plate as opposed to obtuse angles. Furthermore, acute

fracture angles resulted in lower fracture displacement compared to the other fractures considered here.

Fractures around the tip of the stem had the same biomechanical effect on the locking plate. However,

fractures more distal to the stem led to subsequent increase of stress, strain, and fracture displacement.

Results highlight that in good bone quality and acute fracture angles, single locking plate fixation is

perhaps an appropriate management method. On the contrary, for poor bone quality and obtuse fracture

angles alternative management methods might be required as the fixation might be under higher risk of

failure. Clinical studies for the management of PFF are required to further support our findings.

� 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Injury

jo ur n al ho m epag e: ww w.els evier . c om / lo cat e/ in ju r y

Introduction

Total hip arthroplasty (THA) is complicated by periprostheticfemoral fractures (PFFs), with incidence varying from about 1% forcemented and 5.4% for uncemented primary prostheses [1,2]. Themajority of the PFFs are located around the tip of the stem and aresubdivided as B1 with the stem stable, B2 with the stem unstableand B3, with significant bone loss, according to the Vancouverclassification [3]. Optimal management of Vancouver B1 peripros-thetic femoral fractures around a well fixed prosthesis remainscontroversial as adequate fixation needs to be achieved withoutcompromising the stability of the prosthesis [2].

* Corresponding author. Tel.: +44 7707 695139.

E-mail address: leonidou@doctors.org.uk (A. Leonidou).

http://dx.doi.org/10.1016/j.injury.2014.10.060

0020–1383/� 2014 Elsevier Ltd. All rights reserved.

Locking plates have been frequently used in the management ofB1 PFFs with variable published results [2,4,5]. Over recent yearsthere has been an increase in number of PFF fixation failure reportsincluding locking plates [2,4,5]. While various studies areinvestigating development of new fixation methods for thesefractures, patient bone quality, stem stability and fracture level arealso considered as contributing factors to the success or failure ofPFF fixations [1,2].

A recent clinical study by Leonidou et al. highlighted thatfracture angle may also need to be considered as an independentparameter when contemplating a treatment method for Vancouvertype B1 fractures [1]. Regardless, the definition of bone quality andfracture level in B1 fractures is also not clear [1].

Therefore the underlying hypothesis of this study was that,bone quality, fracture angle and level (fracture topography) canchange the biomechanics of PFF fixation in B1 fractures where

A. Leonidou et al. / Injury, Int. J. Care Injured 46 (2015) 213–217214

some cases might be at higher risk of failure. The specific aim ofthis study was to illustrate the effect of bone quality, fracture angleand level in a simplified finite element (FE) model replicating aVancouver B1 fracture. It must be mentioned that the simplifiedmodel has been widely used in literature to understand thebiomechanics of fracture fixation and was validated against aclinical case study in one of our recent studies [6]. Nevertheless,due to its simplified nature our results here have intended as apreliminary investigation where the emphasis has been put on thepattern of FE results rather than their exact predictions.

Materials and methods

Model description

A simplified parametric finite element model of a cementedtotal hip replacement (Fig. 1) was developed previously by Moazenet al., and was adopted in this study [7]. The model was validatedagainst a clinical case study [6]. The bone, stem and the cementwere modelled as concentric cylinders. A transverse fracture belowthe tip of the stem was created and fixed laterally with a ten holesplate. In order to achieve an appropriate bridging length, two holesin the centre of the plate (i.e. across the fracture line) were leftempty [8]. The fracture was held in place with four unicorticalscrews proximally and four bicortical screws distally [9,10]. Thescrews were modelled as cylinders with a diameter of 4.5 mmcorresponding to common diameter of locking cortical screws usedin the femur [11]. The plate was placed directly on bone, as it hasbeen shown that direct contact of the locking plate with the bone incombination with two holes of working length resulted in

Fig. 1. Simplified model as created in Abaqus CAE Software.

decreased stresses on the construct [8]. All parts of the modelwere assigned isotropic materials properties with a Youngmodulus of 2 GPa for cement, 20 GPa for bone and 200 GPa forthe metal plate and screws [12]. A Poisson’s ratio of 0.3 was usedfor all materials.

Boundary conditions and loads

The stem–cement, cement–bone, and screw head–plate inter-faces were tied together. Interaction at the plate–bone interfacedwas modelled with contact elements with coefficient of friction of0.3 [13]. Interfaces at the fracture site were also modelled withcontact elements with coefficient of friction of 0.04, correspondingto early stages of fracture healing, where no callus has formedbetween the two fragments [7]. The distal part of the model wasrigidly fixed and the proximal part of the stem was loaded with atransverse force equal to P = 5W � sin u, where W was the bodyweight and u was the loading angle between the line of action ofgravity and the long axis of femur. Body weight of 600 N andloading angle of 118 were used in this study [12].

Mesh sensitivity

The model was meshed with Tetrahedral (C3D4) elements.Convergence was tested by increasing the number of elementsfrom 42,000 to 1,600,000 in five steps. The solution converged onthe parameters of the interest (�5%-for von Mises stress across thetwo empty screw holes where the plate is at high risk of failure[7,12,13]) with approximately 300,000 elements. Models with thisnumber of elements or more were used for each of the casespresented.

Bone quality and fracture topography

Three models with different canal thickness ratio (CTR) weredeveloped representing poor, average and best bone quality withrespective CTRs of 0.44, 0.88 and 1.46 (Fig. 2). This was done basedon our previous study, which indicated that bone quality can varywithin Vancouver B1 fractures [1]. Further three models weredeveloped with angle fractures varying from the unstabletransverse (08) and short oblique (1468) to the stable long obliqueconfiguration (768) (Fig. 3). Finally, three models were developedwith the fracture at the tip of the stem, 4 mm and 14 mm below thetip of the stem (Fig. 4).

Simulations and measurements

The models were solved and analysed using a finite elementsimulation package (ABAQUS v. 6.9, Simulia Inc., Providence, RI,USA). The pattern of von Mises stress on the plate and fracturemovement was compared across the cases. Fracture movementwas quantified as the relative displacement of the most distal pointof the proximal fragment and the most proximal point of the distalfragment on the medial side of the bone.

Fig. 2. Sagittal view of the model showing different bone quality configurations. (A)

Worst bone quality (CTR = 0.44), (B) average bone quality (CTR = 0.88), (C) best bone

quality (CTR = 1.46).

Fig. 3. Finite element models and the relevant angles of the PFF. A is a transverse

fracture, B equals to 768 and C is a fracture with a 1468 angle.

Fig. 4. Finite element models and the relevant level of the PFF. (A) Fracture at the tip

of the stem, (B) 4 mm below the tip and (C) 14 mm below the tip.

A. Leonidou et al. / Injury, Int. J. Care Injured 46 (2015) 213–217 215

Results

Increasing the bone quality (CTR) led to reduction of von Misesstresses on the locking plate. The fracture displacement was alsodecreased. In respect to the fracture angle with the medial cortex, itwas shown that acute angles resulted in lower level of von Misesstress on the plate as opposed to obtuse angles. Furthermore, acutefracture angles resulted in lower fracture movement. Fracturesaround the tip of the stem (0 and 4 mm) had the samebiomechanical effect on the locking plate. However, as the distanceof the fracture from the tip was increased at 13 mm the von Misesstresses and fracture movement were also increased. Numericalresults of the FE are summarized in Table 1. Pattern of von Misesstress around the empty screw holes of the plate for all the modelsis shown in Figs. 5–7.

Discussion

The results of this study confirm our hypothesis that poor bonequality, unstable patterns of fractures, and fractures below the tipof the stem in Vancouver B1 fractures increase the mechanicalstress on the locking plate as well as fracture movement underloading.

Finite element analysis results from this study highlighted thatthe worse bone quality model exhibited increased levels of stressaround the empty screw holes (i.e. potentially leading to fixationfailure [7,8,14]), as well as increased fracture displacement (i.e.potentially leading to non-union). Considering that Vancouverclassification does not take into account the variation that exists inthe bone quality in B1 fractures, the preliminary results of thisstudy suggest that perhaps B1 fractures with low CTR need to betreated with cautions. Nevertheless, what classifies as a low CTRbased on a plan X-ray requires further investigation [1].

The fracture angle was shown to have a considerable effect overthe distribution of the von Mises stress on the locking plate. Theworse results were recorded with the short oblique (1468) modeland the best ones with the long oblique (768) model with thetransverse fracture results being within the range of long and shortoblique fractures. These results confirm the inherent instability ofthe transverse and short oblique PFF and question the suitability ofsingle locking plate alone for the management of these injuries.The orthopaedic surgeon might therefore consider that shortoblique and transverse fractures are unstable and require strongerand more secure fixation in all planes [15,16].

The level of the PFF is an important parameter of its personality.In this study it was shown that von Mises stresses and fracturemovement were increased as the fracture occurred more distal tothe tip of the stem (Table 1). More specifically, for a fracture 14 mm

Table 1Results of FE for the different models.

von Mises (MPa) Fracture displacement (mm)

Bone qualityModel 1 (worst) 168 0.53

Model 2 (average) 111 0.37

Model 3 (best) 104 0.27

Fracture angleModel 1 (08) 111 0.37

Model 2 (768) 60 0.10

Model 3 (1468) 342 0.72

Fracture levelModel 1 (tip) 114 0.30

Model 2 (4 mm below) 111 0.37

Model 3 (14 mm below) 139 0.75

Fig. 5. Visual Distribution of von Mises stress around the empty screw holes of the

plate. A demonstrates worst bone quality, B equals to average and C to best bone

quality.

Fig. 6. Visual Distribution of von Mises stress around the empty screw holes of the

plate. A demonstrates the transverse fracture, B equals to 768 (long oblique fracture)

and C is the fracture with a 1468 angle (short oblique fracture).

A. Leonidou et al. / Injury, Int. J. Care Injured 46 (2015) 213–217216

away from the tip of the stem the fracture displacement was twicehigher as opposed to a fracture at the tip of the stem. This highlightthe importance of identifying a clear border between Vancouver Band C type fractures. Regardless, our results suggest that singlelocking plate in distal femoral fractures can be under high level ofstress [5]. The importance of fracture level was identified by Bryantet al., as they claimed that the B1 fractures around the tip of thestem are the most challenging to treat [17]. They further advocatedthe use of long locking plates spanning the whole femur withmultiple screws for the management of this type of fracture[17]. Choi et al. further suggested that single locking plate fixationmay not be adequate for comminuted B1 PFF occurring at or nearthe tip of the stem [18]. They suggested supplementation of thelocking plate fixation either with an allograft or with an anteriorplate [18]. Results obtained in this study further suggest that distalfractures can increase the stress level on the single locking platefixation. This could potentially explain reported cases of lockingplate failure and pullout when used for the management of distalPFF [5,19].

Fracture displacement was studied in the FE models and is ofparamount importance as it disturbs the continuity of the medialcortex of the femur and can potentially affect the healing process.The importance of the medial cortex was outlined by Corten et al.,as they suggested that a lateral plate alone was suitable for thefixation of a B1 PFF only if the medial cortex was anatomically

reduced and not comminuted [20]. They further advised biplanarfixation if the medial cortex was disrupted, as they thought thatthis fracture configuration would be more unstable [20]. Further-more, fracture displacement and movement are known risk factorsfor delayed healing, which can subsequently interpret thedocumented delayed and non-unions of B1 PFF treated withlocking plates [4,11,21].

This study is to the best of authors’ knowledge the first one toinvestigate the effect of fracture topography and bone quality onlocking plate fixation for a Vancouver B1 PFF. One of the keyadvantages of the simplified model used in this study was that theeffect of different parameters on the stress level of the fixationcould be isolated. At the same time, one of the main limitations ofthis FE model is that it has not been corroborated with abiomechanical experimental model yet was validated against aclinical case study [6]. Nevertheless, similar simplified femoralmodels have been extensively used in the literature and have beenvalidated both on theoretical and on biomechanical grounds[8,22,23]. More specifically, both Iesaka et al. and Stoffel et al.corroborated their simplified femoral model against experimentalmodel [8,23]. Furthermore, published data from biomechanicaland clinical studies – with different experimental setup – revealedincreased stresses and subsequent failure through the empty holesof the locking plate, which agree with the findings of the currentstudy [4,8,24,25]. Therefore, despite the limitations of the

Fig. 7. Visual distribution of von Mises stress around the empty screw holes of the

plate. A is the plate with the fracture at the tip of the stem, B at 4 mm below the tip

and C at 14 mm below the tip.

A. Leonidou et al. / Injury, Int. J. Care Injured 46 (2015) 213–217 217

validation process, relative comparisons that were made in thisstudy between the models remain valid.

This study suggests that the orthopaedic surgeon shouldcontemplate on the PFF topography and bone quality and notrely merely on the Vancouver Classification in order to formulate atreatment plan. More specifically, in Vancouver B1 fractureswith unstable patterns, away from the tip of the stem withpoor underlying bone stock, the surgeon might consider alterna-tive methods to enhance the single locking plate fixation.This can include increasing the plate length, revision to longstem or applying bone allograft. Further clinical, biomechanicaland computational studies need to be conducted focusing onthe fracture personality and not only on the locking plateconfiguration.

Conflict of interest

None declared.

References

[1] Leonidou A, Moazen M, Skrzypiec DM, Graham SM, Pagkalos J, Tsiridis E.Evaluation of fracture topography and bone quality in periprosthetic femoralfractures: a preliminary radiographic study of consecutive clinical data. Injury2013;44(12):1799–804.

[2] Graham SM, Moazen M, Leonidou A, Tsiridis E. Locking plate fixation forVancouver B1 periprosthetic femoral fractures: a critical analysis of 135 cases.J Orthop Sci 2013;18(3):426–36.

[3] Duncan CP, Masri BA. Fractures of the femur after hip replacement. InstrCourse Lect 1995;44:293–304.

[4] Buttaro MA, Farfalli G, Paredes Nunez M, Comba F, Piccaluga F. Lockingcompression plate fixation of Vancouver type-B1 periprosthetic femoral frac-tures. J Bone Joint Surg Am 2007;89(9):1964–9.

[5] Chakravarthy J, Bansal R, Cooper J. Locking plate osteosynthesis for VancouverType B1 and Type C periprosthetic fractures of femur: a report on 12 patients.Injury 2007;38(6):725–33.

[6] Moazen M, Mak JH, Etchels LW, Jin Z, Wilcox RK, Jones AC, et al. The effect offracture stability on the performance of locking plate fixation in periprostheticfemoral fractures. J Arthroplasty 2013;28(9):1589–95.

[7] Moazen M, Jones AC, Leonidou A, Jin Z, Wilcox RK, Tsiridis E. Rigid versusflexible plate fixation for periprosthetic femoral fracture-computer modellingof a clinical case. Med Eng Phys 2012;34(8):1041–8.

[8] Stoffel K, Dieter U, Stachowiak G, Gachter A, Kuster MS. Biomechanical testingof the LCP – how can stability in locked internal fixators be controlled? Injury2003;34(Suppl. 2):B11–9.

[9] Cronier P, Pietu G, Dujardin C, Bigorre N, Ducellier F, Gerard R. The concept oflocking plates. Orthop Traumatol Surg Res 2010;96(4):S17–36.

[10] Ehlinger M, Bonnomet F, Adam P. Periprosthetic femoral fractures: the mini-mally invasive fixation option. Orthop Traumatol Surg Res 2010;96(3):304–9.

[11] Campbell WC, Canale ST, Beaty JH. Campbell’s operative orthopaedics. 11th ed.Philadelphia, PA: Mosby/Elsevier; 2008.

[12] Yoon YS, Jang GH, Kim YY. Shape optimal design of the stem of a cemented hipprosthesis to minimize stress concentration in the cement layer. J Biomech1989;22(11-12):1279–84.

[13] Cordey J, Borgeaud M, Perren SM. Force transfer between the plate and thebone: relative importance of the bending stiffness of the screws frictionbetween plate and bone. Injury 2000;31(Suppl. 3):C21–8.

[14] Pappas CA, Young PG, Lee AJ. Development of the mennen 3 peripro fixationplate for the treatment of periprosthetic fractures of the femur. Proc Inst MechEng H 2006;220(7):775–85.

[15] Parvizi J, Rapuri VR, Purtill JJ, Sharkey PF, Rothman RH, Hozack WJ. Treatmentprotocol for proximal femoral periprosthetic fractures. J Bone Joint Surg Am2004;86-A(Suppl 2):8–16.

[16] Moazen M, Mak JH, Etchels LW, Jin Z, Wilcox RK, Jones AC, et al. Periprostheticfemoral fracture – a biomechanical comparison between Vancouver type B1and B2 fixation methods. J Arthroplasty 2014;29(3):495–500.

[17] Bryant GK, Morshed S, Agel J, Henley MB, Barei DP, Taitsman LA, et al. Isolatedlocked compression plating for Vancouver Type B1 periprosthetic femoralfractures. Injury 2009;40(11):1180–6.

[18] Choi JK, Gardner TR, Yoon E, Morrison TA, Macaulay WB, Geller JA. The effect offixation technique on the stiffness of comminuted vancouver B1 peripros-thetic femur fractures. J Arthroplasty 2010;25(6 Suppl.):124–8.

[19] Wood GC, Naudie DR, McAuley J, McCalden RW. Locking compression platesfor the treatment of periprosthetic femoral fractures around well-fixed totalhip and knee implants. J Arthroplasty 2011;26:886–92.

[20] Corten K, Vanrykel F, Bellemans J, Frederix PR, Simon JP, Broos PL. An algorithmfor the surgical treatment of periprosthetic fractures of the femur around awell-fixed femoral component. J Bone Joint Surg Br 2009;91(11):1424–30.

[21] Fulkerson E, Tejwani N, Stuchin S, Egol K. Management of periprosthetic femurfractures with a first generation locking plate. Injury 2007;38(8):965–72.

[22] Rybicki EF, Simonen FA, Mills EJ, Hassler CR, Scoles P, Milne D, et al. Mathe-matical and experimental studies on the mechanics of plated transversefractures. J Biomech 1974;7(4):377–84.

[23] Iesaka K, Kummer FJ, Di Cesare PE. Stress risers between two ipsilateralintramedullary stems: a finite-element and biomechanical analysis. J Arthro-plasty 2005;20(3):386–91.

[24] Ahmad M, Nanda R, Bajwa AS, Candal-Couto J, Green S, Hui AC. Biomechanicaltesting of the locking compression plate: when does the distance betweenbone and implant significantly reduce construct stability? Injury 2007;38(3):358–364.

[25] Wilkens KJ, Curtiss S, Lee MA. Polyaxial locking plate fixation in distal femurfractures: a biomechanical comparison. J Orthop Trauma 2008;22(9):624–8.