Finite element study of the proximal femurwith retained trochanteric gamma nail andafter removal of nail

B. Mahaisavariya a,*, K. Sitthiseripratip b, J. Suwanprateeb b

Injury, Int. J. Care Injured (2006) 37, 778—785

www.elsevier.com/locate/injury

a Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, ThailandbNational Metal and Materials Technology Centre, Pathumthani, Thailand

Accepted 12 January 2006

KEYWORDSFracture;Trochanteric fracture;Intramedullary nail;Gamma nail;Finite element study

Summary This study aims to evaluate the stress and strain distributions in thehealed proximal femur after fixation with a trochanteric gamma nail (TGN) and afterTGN removal, using the finite element method. The stress distributions in theproximal femur with retained TGN and after TGN removal were very similar. Thestrain and the strain energy density in the femoral neck region with retained TGNweremuch higher than in the lag screw hole at the subtrochanter and the distal lockingscrew hole at the proximal femur, and even higher after TGN removal. Stair climbingresulted in higher strain and higher strain energy density at the femoral neck thannormal walking. The conclusion can be drawn that removal of the TGN may result inhigh risk of femoral neck fracture.# 2006 Elsevier Ltd. All rights reserved.

Introduction

Proximal femoral nailing using the trochantericgamma nail (TGN) (Stryker—Howmedica—Osteonics)is now widely accepted as one of the treatmentmethods for unstable trochanteric fractures.17 Themethod has been found effective and less invasivethan using the compression hip screw or angle-bladeplate osteosynthesis.13 Our previous study, using thefinite element method of trochanteric fracture sta-bilisation with a TGN and the one-legged stance,

* Corresponding author. Tel.: +662 4197968; fax: +662 4128172.E-mail address: sibmh@mahidol.ac.th (B. Mahaisavariya).

0020–1383/$ — see front matter # 2006 Elsevier Ltd. All rights resedoi:10.1016/j.injury.2006.01.019

showed that even when the fracture had healed highstress levels still occurred at the implant structure.This may lead to the fatigue failure if the implant isretained for a long period.16 However, there havebeen two recent reports of femoral neck fractureafter removal of the gamma nail.7 The authorsconsidered that the fracture resulted from boneweakening where there was a relatively large bonydefect at the femoral neck after implant removal.Although removal of the TGN is not routine, it may benecessary when there are clinical symptoms ofmechanical irritation by the TGN. In such cases, afterimplant removal, the stress—strain distribution maybe altered and may result in high susceptibility to

rved.

Finite element study of the proximal femur 779

fracture in the area of bony defect resulting fromimplant removal. To our knowledge, there are nostudies of alteration of stress distribution, and wetherefore conducted this study using a finite elementmethod to evaluate the stress and strain distributionsin the healed proximal femur with retained TGN andafter TGN removal.

Materials and methods

All finite element models presented here were con-structed using the digital CAD technique based on CT

Figure 1 Loading condition for the pre-clinical testing.4

data, and all analyses were performed using MSCPatran/Mentat/Marc finite element softwarepackages.

The finite element model

A three-dimensional CAD model of the intact femurwas created from theaverage geometry derived fromCT of 108 Thai cadaveric femora11 using a Philipsspiral CT scanner (Tomoscan AV). In the proximaland distal regions of the femur, CT scan acquisitionwas performedwith 3-mm slice thickness, and recon-structionwith 1-mm interpolated slice thickness. Forthe femoral shaft, CTscan acquisition was performedwith 10-mm slice thickness, and reconstruction with5-mm interpolated slice thickness.11 The TGNemployed in the model had a proximal diameter of17 mm, distal diameter of 11 mm, length of 180 mm,1308 neck shaft angle and 48 valgus angle, with onlyone transverse distal locking screw. The TGN wasvirtually inserted into the intramedullary canal ofthe intact femur. The lag screw was inserted throughthe femoral head centre and away from the outerboundary of femoral headby 10 mm,according to thesurgical technique for TGN fixation.

Four-noded tetrahedral elements based on STLautomatic mesh generation technique (Magic RP,Materialise N.V., Belgium) were used to build upthe mesh of the intact femur and the TGN. Differentregions were introduced into the model (Fig. 1),enabling the definition of different material proper-ties (Table 1) and contact conditions in the fractureplane. The femur model had a total of 17,059 nodesand 102,052 elements, whereas the TGN model hada total of 15,441 nodes and 104,668 elements.

Material properties

Linear elastic isotropic material properties wereassigned to all materials involved in the model.

Table 1 Properties applied for the FE model20

Region modeled Young’smodulus (MPa)

Poisson’sratio

Cortical boneFemoral cortex 17000 0.28Femoral neck cortex 2000 0.3

Cancellous boneFemoral head 600 0.3Femoral neck 1000 0.3Femur 600 0.3

ImplantTrochanteric gamma nail 200000 0.3

MPa: equivalent von Mises stress.

780 B. Mahaisavariya et al.

Table 2 Loading conditions under walking and stair climbing activities4

Force applied by Walking Stair climbing Act at point

X Y Z X Y Z

Hip contact �54.0 �32.8 �229.2 �59.3 �60.6 �263.3 P0Intersegmental resultant �8.1 �12.8 �78.2 �13.0 �28.0 �70.1 P0Abductor (1) 58.0 4.3 86.5 70.1 28.8 84.9 P1Proximal ileotibial tract (2a) — — — 10.5 3.0 12.8 P1Distal ileotibial tract (2b) — — — �0.5 �0.8 �16.8 P1Proximal tensor fasciae latae (3a) 7.2 11.6 13.2 3.1 4.9 2.9 P1Distal tensor fasciae latae (3b) �0.5 �0.7 19.0 �0.2 �0.3 6.5 P1Vastus lateralis (4) �0.9 18.5 �92.9 �2.2 22.4 �135.1 P2Vastus medialis (5) — — — �8.8 39.6 �267.1 P3

Different material properties were attributed todifferent femoral regions: cortical bone in thefemoral shaft, femoral head and femoral neck,trabecular bone in the femoral head, femoral neckand trochanteric region.20 Corresponding elasticconstants are given in Table 1. For the TGN, stainlesssteel was used for analysis purposes.

Loading conditions

In order to simulate physiological loading in theproximal femur, we used the loading conditionsfor preclinical testing investigated by Helleret al.4 This loading condition focused on two com-mon activities of daily life, i.e. walking and stairclimbing. Walking is the most common, whereas thestair climbing generates the highest force and thehighest torsion in the femur. The loading consists ofjoint reaction force and related muscle forces, asshown in Fig. 1 and Table 2. The femoral model wasfully constrained (zero displacement) at the distalfemur.

Four cases were examined in the study. The firsttwo cases were designed to investigate the stress/strain distributions in the proximal femur with TGNfixation after fracture healing and after TGNremoval, while walking. The latter two cases weredesigned to evaluate the stress/strain distributions

Table 3 Equivalent von Mises stress in the femur (MPa)

Area Walk

TGN

Proximal cortex in the femoral neck region 6.1Proximal cancellous bone in the femoral neck region 3.1Inside the lag screw hole in the femoral neck region 4.4Distal locking screw hole 10.5

TGN: trochanteric gamma nail.

in the proximal femur with TGN fixation and afterTGN removal, while climbing stairs.

Results

The results focused on the critical area of theproximal femur with screw-hole defects in bothcortical and cancellous bony layers.

Stress distribution

As shown in Table 3, the TGN produced high stressesaffecting surrounding bone, which were caused bythemain function of the fracture fixation device, i.e.to achieve primary stability. Therefore, the stressesmainly affected cancellous bone in the proximalfemur and, when the TGN was removed, the stresschange was higher in the proximal cancellous bonethan in the cortical bone. The stresses while climbingstairs were higher than while walking.

Strain distribution

As shown in Table 4, during fixation with the TGN thehigher strain was at the femoral neck inside thecancellous bone; this softer material tends toadapt and resist bending and torsion forces under

ing Stair climbing

fixation TGN removed TGN fixation TGN removed

9 6.54 8.30 8.266 1.46 3.91 1.838 1.45 5.49 1.613 9.59 14.03 14.33

Finite element study of the proximal femur 781

Table 4 Equivalent total strain in the femur (mstrain)

Area Walking Stair climbing

TGN fixation TGN removed TGN fixation TGN removed

Proximal cortex in the femoral neck region 325 1509 731 1991Proximal cancellous bone in the femoral neck region 2389 1297 2795 1608Inside the lag screw hole in the femoral neck region 2446 1281 3759 1429Distal locking screw hole 549 561 1826 747

loading. The higher strain was at the outer corticallayer in the femoral neck region after TGN removal,when the cortical bone is the main bony structureresisting bending moment and torque. The strainsof stair climbing were greater than the strains ofwalking.

Total strain energy density

The total strain energy density (SED) is an importantparameter in evaluating the risk of bony fracture. Asthe shown in Table 5, there was high SED inside thecancellous bone in the femoral neck region duringfixation with TGN. After TGN removal, the high SEDwas in the cortical bone in the femoral neck region(Fig. 2), which is related to the strain distributiondescribed previously. The SED demonstrated thecritical results, which may lead to high risk offracture at the femoral neck (Figs. 3 and 4). Usuallyfracture caused by over-bending and torsion beginsin the cortical bone, particularly the thin corticallayer.12 The SED while stair climbing was two to fourtimes higher than while walking.

Discussion

Stress/strain distribution analysis using the finiteelement method is widely accepted as a usefultechnique to evaluate or predict the biomechanicalbehaviour of orthopaedic implants under certainload conditions.16,14,19 In simulating the loadingconditions of the hip joint, many authors havedemonstrated the importance of including muscle

Table 5 Total strain energy density in the femur (MPa/100

Area Walk

TGN

Proximal cortex in the femoral neck region 1233Proximal cancellous bone in the femoral neck region 3653Inside the lag screw hole in the femoral neck region 6025Distal locking screw hole 3488

MPa: equivalent von Mises stress.

loading.3,8,15 This study was carried out using themodel which includes muscle loading as describedby Heller et al.4 This model of loading conditions hasbeen found to simulate very closely in vivo hip jointforces.4

The TGN used in this study was a modified form ofthe standard gamma nail (SGN) and has been widelyused in the Asia—Pacific region. The TGN measuresonly 11 mm in diameter distally, with one distallocking screw hole, whereas the proximal diameteris 17 mm to accept the 12-mm lag screw (similar tothe SGN). Although the TGN was inserted into anappropriate position, with the lag screw placedalong the axis of the femoral neck, it was observedthat the proximal part of the TGN protruded outsidethe greater trochanter (Fig. 1). This was caused bythemismatch of the TGN to the geometry of the Thaifemur, as described in our previous report.10

The results of the present study show that, whilethe TGN is retained, the higher stresses are trans-ferred through the nail to the adjacent contact boneat the area of the femoral neck and around distallocking hole (Fig. 2), as also shown in previousreports.13,19 The high stresses in the bone adjacentto the lag screw tip may contribute to lag screw cut-out, which may easily occur if the lag screw tip isclose to the superior portion of the femoral head6 orif bone strength is diminished by avascular necro-sis.2,18

The stress occurring in the studied areas with aretained TGN was higher than after TGN removal,during both walking and stair climbing. This can beexplained by the fact that with a retained TGN thestress transfers mostly through the stiffer material,

0)

ing Stair climbing

fixation TGN removed TGN fixation TGN removed

3255 4925 62731234 11770 18031045 18720 13173645 6194 6456

782 B. Mahaisavariya et al.

Figure 2 Strain energy density in the femoral cortex during TGN fixation and after TGN removal under walking (top) andstair climbing (bottom) activities.

Finite element study of the proximal femur 783

Figure 3 Strain energy density in the proximal cancellous during TGN fixation and after TGN removal under walking(top) and stair climbing (bottom) activities.

which is the implant material, and also through thebone directly in contact, resulting in high stress andminimal strain in this area.

Strain and SED inform the biomechanical changein bone, and in the femoral neck were found to bemuch higher after TGN removal than with a retainedTGN (Table 4). Therefore there is a higher risk offemoral neck fracture after TGN removal. The find-ings from this finite element study correspond tothose of Kukla et al.,7 who showed that fractureresistance of the femoral neck was weakened by 40%after the SGN was removed. This is because therelatively large diameter of the lag screw (12 mm)of the SGN leaves a large bony defect at the femoralneck after removal.7

Our study has also shown that the femoral neckhas much higher strain and SED than the entry pointof the lag screw and the distal locking screw hole.This means that during walking or stair climbingafter implant removal, the femoral neck would bemore likely to fail and fracture than subtrochantericportion at the entry point of lag screw or distallocking screw hole. This conclusion differs fromprevious reports on subtrochanteric fracture fixa-tion with various implant systems such as Gardenscrew or cannulated screw.1,5,9

When comparing the risks for neck fracture asso-ciated with normal walking and with stair climbing,it has been shown that strain and SED are higherduring stair climbing.

784 B. Mahaisavariya et al.

Figure 4 Strain energy density inside the bone defect during TGN fixation and after TGN removal under walking (top)and stair climbing (bottom) activities.

Conclusion

Stress distributions in the proximal femur with aretained TGN and after fracture healing andremoval of the TGN are very similar. However,the strain and SED at the femoral neck aremuch higher after TGN removal thanwith a retainedTGN, and higher than at the screw hole for the lagscrew or the distal locking screw. Stair climbing islikely to create higher strain and strain energydensity than normal walking. If implant removalis unavoidable, care should be taken to avoidfemoral neck fracture during the early postopera-tive period. Stair climbing should be avoided duringthis phase.

Acknowledgement

The authors would like to thank the National Metaland Materials Technology Centre (MTEC), Thailand,for financial support and the use of facilities.

References

1. Andew TA, Thorogood M. Subtrochanteric fracture afterGarden screw fixation: a review of predisposing factorsand management in nine cases. Injury 1984;16:169—77.

2. Chen CM, Chiu FY, Lo WH. Avascular necrosis of femoral headafter gamma nailing for unstable intertrochanteric fractures.Arch Orthop Trauma Surg 2001;21:505—7.

Finite element study of the proximal femur 785

3. Duda GN, Heller M, Albinger J, et al. Influence of muscleforce on femoral strain distribution. J Biomech 1998;31:841—6.

4. Heller MO, Bergmann G, Kassi JP, et al. Determination ofmuscle loading at the hip joint for use in pre-clinical testing.J Biomech 2005;38:1155—63.

5. Karagkevrekis CB, Rahman H. Subtrochanteric fracture fol-lowing removal of screw for slipped capital femoral epiphy-sis: case report. Injury 2003;34:320—1.

6. Kawaguchi S, Sawada K, Nabeta Y. Cutting-out of the lagscrew after internal fixation with the Asiatic gamma nail.Injury 1998;29:47—53.

7. Kukla C, Pichl W, Prokesch R, et al. Femoral neck fractureafter removal of the standard gamma interlocking nail: acadaveric study to determine factors influencing the biome-chanical properties of the proximal femur. J Biomech2001;34:1519—26.

8. Lu TW, Taylor SJ, O’Connor JJ, Walker PS. Influence of muscleactivity on the forces in the femur: an in vivo study. J Biomech1997;30:1101—6.

9. MacEachern AG, Heyse-Moore GH, Jones RN. Subtrochantericfractures of the femur through the track of the lower Gardenscrew: treatment with a Richards sliding screw. Injury1984;15:337—40.

10. Mahaisavariya B, Sitthiseripratip K, Oris P, et al. Fit-and-fillanalysis of trochanteric gamma nail for the Thai proximalfemur: a virtual simulation study. J Med Assoc Thai2004;87:1315—20.

11. Mahaisavariya B, Sitthiseripratip K, Tongdee T, et al.Morphological study of the proximal femur: a newmethod of geometrical assessment using three-dimen-sional reverse engineering. Med Eng Phys 2002;24:617—22.

12. Mayhew PM, Thomas CD, Clement JG, et al. Relation betweenage, femoral neck cortical stability, and hip fracture risk.Lancet 2005;366:129—35.

13. Schipper IB, Marti RK, Werken CVD. Unstable trochantericfractures: extramedullary or intramedullary fixation. Rev LitInjury 2004;35:142—51.

14. Seral B, Garcia JM, Cegonino J, et al. Finite element study ofintramedullary osteosynthesis in the treatment of trochan-teric fractures of the hip: gamma and PFN. Injury2004;35:130—5.

15. Sitthiseripratip K, Mahaisavariya B, Suwanprateeb J. Influ-ence of lateral muscle loading in the proximal femur afterfracture stabilisation with a trochanteric gamma nail (TGN).JSME Int J Ser C 2004;47:1070—4.

16. Sitthiseripratip K, Oosterwyck HV, Sloten JV, et al. Finiteelement study of trochanteric gamma nail for trochantericfracture. Med Eng Phys 2003;25:99—106.

17. Utrilla AL, Reig JS, Munos FM, Tufanisco CB. Trochantericgamma nail and compression hip screw for trochantericfractures: a randomisation prospective, comparative studyin 210 elderly patients with a new design of the gamma nail. JOrthop Trauma 2005;19:229—33.

18. Vicario C, Marco F, Ortega L, Alcobendas M, Dominguez I,Lopez-Duran L. Necrosis of the femoral head after fixation oftrochanteric fractures with gamma locking nail: a cause ofthe late mechanical failure. Injury 2003;34:129—34.

19. Wang CJ, Yettram AL, Yao MS, Procter P. Finite elementanalysis of a gamma nail within a fractured femur. MedEng Phys 1998;20:677—83.

20. Wei HW, Sun SS, Jao SHE, et al. The influence of mechanicalproperties of subchondral plate, femoral head and neck ondynamic stress distribution of the articular cartilage. Med EngPhys 2005;27:295—304.