) 455–459www.elsevier.com/locate/knee

The Knee 13 (2006

Biomechanical testing of hybrid hamstring graft tibialfixation in anterior cruciate ligament reconstruction

Jae Chul Yoo a,⁎, Jin Hwan Ahn a, Jae Hoon Kim a, Byung Kwan Kim b,Kui Won Choi c, Tae Soo Bae c, Chang Yang Lee c

a Department of Orthopaedic Surgery, Sungkyunkwan University School of Medicine, Samsung Medical Center,50 Ilwon-Dong, Kangnam-Ku Seoul, Korea 135-710

b Department of Orthopaedic Surgery, Eulji University School of Medicine, Eulji Medical Center, Seoul, Koreac Korea Institute of Science and Technology, Biomedical Research Center, Seoul, Korea

Received 5 December 2005; received in revised form 9 August 2006; accepted 16 August 2006

Abstract

Hamstring tendon using quadrupled semitendinosus and gracilis autografts is a well-established technique for ACL reconstruction.However, several methods have been used for tibial fixation of the tendon graft. The purpose of this study was to compare the biomechanicalcharacteristics of quadrupled hamstring graft tibial fixation using three different fixation methods. Nine matched pairs (18 specimens) ofcadaver tibias were divided into three groups of six specimens. The first group was fixed with only a tapered 30-mm bioabsorbable screw(BIS), the second group was fixed first with a BIS and then the remaining tendon portion was additionally fixed with a titanium cortical screwand spike washer, and the third group was fixed with only a cortical screw and spike washer. A custom-made probe hook was mounted on aload cell (Interface, MFG, Scottsdale, AZ) to measure the ACL tension before and after the final tibial fixation. Group 2 displayed greatermean maximum load at failure than both groups 1 and 3 ( pb0.05). The stiffness of the graft nearly doubled in group 2 compared to groups 1and 3 ( pb0.05). All specimens failed by slippage and pullout. Biomechanical testing with cadavers showed that a BIS and additional corticalscrew and spike washer fixation to the distal hamstring tendon resulted in higher load at failure and stiffness compared to either BIS orcortical screw and spike washer fixation alone.© 2006 Elsevier B.V. All rights reserved.

Keywords: Anterior cruciate ligament reconstruction; Bioabsorbable interference screw; Hamstring tendon; Tibial fixation; Biomechanical testing

1. Introduction

Recently, hamstring tendon reconstruction, particularlyquadrupled semitendinosus and gracilis autografts (QSGT)has gained popularity over patella bone–tendon–bone(PBTB) reconstruction due to the improvement in soft tissuefixation methods and nearly equivalent clinical results [1–14]. The femoral fixation of hamstring tendon grafts using across-pin device has been reported to be stronger than anyother femoral fixation, regardless of graft type [15,16].However, the tibial fixation is considered to be the weak link

⁎ Corresponding author. Tel.: +82 2 3410 3501; fax: +82 2 3410 0061.E-mail addresses: coolshoulder@hotmail.com, jaemidoc@yahoo.com

(J.C. Yoo).

0968-0160/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.knee.2006.08.001

when a quadrupled hamstring graft and a bioabsorbableinterference screw (BIS) are used in ACL reconstruction[17–22].

During most daily living activities, the ACL is reported tobe repetitively loaded to approximately 450 N [18,23–26].During postoperative rehabilitation after hamstring graftACL reconstruction, the graft should withstand such forces,regardless of the tibial fixation construct. Recent studieshave shown that in quadrupled hamstring graft ACLreconstruction, single BIS fixation, also known as aperturefixation [13], could provide sufficient initial fixation [27].However, after the final fixation with BIS, we havesometimes observed graft tension loss, which was main-tained prior to tibial fixation with a 20-lb tension appliedusing a tensioner (Tie Tensioner, Mitek, Somerville, NJ).

Fig. 1. Schematic drawing of the three different fixation method in tibia.

Fig. 3. The specimen was mounted on servohydraulic testing machine with aprobe-like load cell applied horizontally.

456 J.C. Yoo et al. / The Knee 13 (2006) 455–459

This raised a concern that an immediate accelerated rehabili-tation protocol would result in laxity in the ACL reconstruc-tion. For this reason, we have used additional fixation inhamstring ACL reconstructions consisting of a titaniumcortical screw and spike washer (CSSW) applied to the distalend of the hamstring graft. We questioned whether thisadditional fixation would improve the tibial fixation and if itwould result in a change in ACL tension before and after thetibial fixation.

The purpose of this study was to compare and evaluate thebiomechanical characteristics (failure mode, maximum loadat failure, displacement at failure and stiffness) of a quad-rupled hamstring graft tibial fixation using three differentfixation methods tapered 30-mm bioabsorbable interferencescrew (BIS, Linvatec, Largo, FL) only, BIS with additionalcortical screw and spike washer fixation, and only cortical

Fig. 2. Schematic drawing of the mechanical testing construct. Femoralportion of the jig was made to loop around the tendon.

screw and spike washer fixation. An additional purpose wasto report the changes in ACL tension before and after tibialfixation using a digitized load cell probe.

2. Method

The hamstring tendon grafts (semitendinosus and gracilis) wereharvested from an initial sample of nine paired cadaveric knees (total of18 specimens). Prior to specimen selection, bone mineral density (BMD) ofthe proximal tibia was assessed using a dual energy X-ray absorptionmetry(DEXA) scanner (Hologic QDR-2000 Whole-Body X-ray Bone Denit-ometer, Hologic, Bedford, MA). A previous study indicated that a BMDgreater than 0.6 g/cm2 (mean=0.78 g/cm2; range=0.78 to 1.08 g/cm2) wasdeemed acceptable for testing specimens [27]. The mean cadaver age was 54years (range=38 to 60). The qualified specimens were randomly assignedinto three groups of six specimens each; however, if both left and rightspecimens from the same cadaver were assigned to one group, one wasreassigned to a different group.

Group 1 received 30-mm bioabsorbable interference screw fixation only;group 2 was fixed with a titanium cortical screw and spike washer after 30-mm BIS fixation, and group 3 was fixed with only a titanium cortical screwand spike washer (Fig. 1). Bone mineral density was compared between thethree groups. Insertion torque was measured for BIS and cortical screwfixation with a torque measuring device (STC, Tohnichi, Japan).

The quadrupled semitendinosus and gracilis tendons were prepared toachieve maximal length without the end strand too thin. During graftpreparation, the grafts were folded over a No. 5 Ethibond (Ethicon,Somerville, NJ) suture. A No. 2 Ethibond suture was used to first sewapproximately 4 cm of the distal ends of each tendon using a crisscrossingsuture and connect the distal ends of the tendons together at approximately10 cm to enable equal tensioning. The proximal end of the graft, which thequadrupled loop had formed, was not sutured since we used a custom jig to

Fig. 4. Maximum load at failure between groups.

Table 1ACL tension measurement results before and after tibial fixation

Tensionmanual

Tensiontensioner

Final tension after fixation

G1 G2 G3 Total

Mean 15.6 8.5 3.6 8.2 4.9 5.6S.D. 4.6 2.3 1.8 2.2 0.8 2.6

Reconstructed ACL tension was measured by a probe hooked on to a loadcell, which pulls the ligament horizontally.

Fig. 5. The displacement at failure in three groups. Note that the BIS-onlygroup showed large standard deviations.

457J.C. Yoo et al. / The Knee 13 (2006) 455–459

loop around the two tendons (Figs. 2 and 3). This configuration wassomewhat similar to our femoral fixation method with the RIGIDfix system(Mitek, Johnson & Johnson, Somerville, NJ). All grafts were sprayed withnormal saline during preparation and testing to prevent tissue desiccation.

The grafts were sized to the nearest 0.5 mm after suturing. The tibialtunnel was prepared with an ACL tibial guide set at 50 mm and the guide-pinlocated in the mid-portion of the ACL footprint. The tibial tunnel length wasmeasured to be approximately 40 mm. The diameter of the tibial tunnel wasdrilled first with a 6.0 mm reamer and then enlarged to match the diameter ofthe quadrupled hamstring graft, which ranged from 7.0 to 9.0 mm. Tunneldilators (Linvatec, Largo, FL) were used to sequentially dilate the tunnel by0.5 mm to the desired diameter.

The prepared graft was passed from the tibial side to loop the proximalportion of the graft to the proximal jig. The graft length was maintained atapproximately 30 mm in the femoral tunnel portion, 30 mm in the intra-articular portion, 40 mm in the tibial tunnel, and 20 mm was used for distaltibial extra-tunnel fixation. For groups 1 and 2, a whipstitch was placed inthe ends of the graft and also in the portion where the BIS directly contacts inthe tibial tunnel. The portion of the tendon in which the BIS was in contactwas measured roughly with a probe prior to fixation and a whipstitch wasperformed through that length.

Biomechanical testing was done on a servohydraulic testing machine(Instron 8511, MTS, Minneapolis, MN). The proximal hamstring graft wassecured by looping both tendons around a custom-made round bar. The tibialspecimens were also mounted in a custom-made jig which enabled thedisplacement force vector to be aligned at an angle of 20° in line with the tibialtunnel (not worst case scenario) (Fig. 3). In all specimens, the suture tails fromeach tendon that exited the tibial tunnel were tied together and placed aroundthe two arms of a tensioning device (Tie Tensioner,Mitek, Somerville, NJ). Theconstructs were then cycled 10 times with a force of approximately 30 N forpretensioning. In group 1, the hamstring tendon graft was loaded with a 20-lbforce and secured to the tibia by using a 30-mmBIS. In group 2, the 30-mmBISwas first secured then the distal end of the graft was secured using a corticalscrew and spike washer with 20-lb tensioner still being applied. In group 3,only cortical screw and spike washer fixation was performed while applying20-lb tensioning. A single surgeon did all tensioning. Load was then applied at20 mm/min to failure. The maximum load at failure, displacement at failure,stiffness, and mode of failure were determined.

Prior to the tibial fixation, the ACL tension was measured by pulling theintra-articular portion of the graft horizontally with a custom-made hook,

Fig. 6. The stiffness of the reconstructed graft in three groups.

which was mounted on the load cell digital sensor (Interface, Scottsdale, AZ;Figs. 2 and 3). This was intended to simulate manual tensioning of the ACLusing a probe during routine knee arthroscopy. We first measured the ACLtension with the use of rough manual tension (manual tension applied to thesutures by an assistant, using maximal strength without rupturing the suturematerial or the hamstring). Afterwards, we measured the ACL tension whena tensioner was applied at 20 lb. Finally, the ACL tension was measured aftereach type of tibial fixation was completed.

Kruskal–Wallis analysis of variance and Mann–WhitneyU tests for posthoc comparisons were used to assess the mean differences between groups.All statistical analyses were performed with SPSS software (SPSS forWindows Release 11.0, SPSS Inc, Chicago, IL). All analyses were set at a95% confidence interval for statistical significance.

3. Results

Group 2 (bioabsorbable interference screw plus distal titaniumcortical screw and spike washer fixation) had twice the stiffness andmean maximal load at failure compared to the other two groups.The mean bone mineral density was similar for each of the groups(group 1, 0.91±0.11 g/cm2; group 2, 0.93±0.13 g/cm2; group 3,0.90±0.76 g/cm2; pN0.05). The mean BIS insertion to torques forgroup 1 was 1.44±53 Nm, and for group 2, it was 1.57±0.20 Nm.Insertion torques for cortical screw fixation of the distal end of thehamstring tendon were 1.93±0.24 Nm for group 2, and 1.75±0.32 Nm for group 3. Insertion torqueses were also comparablebetween each fixation method.

The biomechanical test results are summarized in Figs. 4, 5, and 6.Group 2 (bioabsorbable interference screw plus distal titanium corticalscrew and spike washer fixation) displayed a significantly greater meanmaximum load at failure than both group 1 (30-mm BIS fixation,p=0.005) and group 3 (cortical screw and spike washer fixation only,p=0.05) (Fig. 4). Themean displacement at failurewas not significantlydifferent between the groups. However group 1 had a relatively largestandard deviation compared to groups 2 and 3 (Fig. 5). The stiffness ofthe graft was nearly doubled in group 2 compared to groups 1( p=0.003) and 2 (p=0.006) (Fig. 6). This suggests that only singlefixation, either 30-mmBISonly or cortical screwand spikewasher only,provide less stiffness than the combined fixation. All specimens failedby slippage and pullout.

The results of ACL tension measurements are summarized inTable 1. The rough manual tensioning displayed the greatesthorizontal tension (mean=14.9±4.5 N) followed by the manualtensioner (mean=8.0±3.18 N). After the final tibial fixation themean ACL tension decreased to a value of 3.69 N (S.D.=1.36 N).The final tension for group 2 was significantly higher than bothgroup 1 ( p=0.001) and group 3 ( p=0.005).

4. Discussion

Tibial fixation with only bioabsorbable screw fixation forquadrupled hamstring tendon grafts have been reported to

458 J.C. Yoo et al. / The Knee 13 (2006) 455–459

result in satisfactory clinical outcomes [13]. However thisstudy indicates that additional cortical screw washer fixationor ‘hybrid fixation’, will nearly double the load to failure.Furthermore, the stiffness of the graft will also doublecompared to a single fixation device.

The majority of tendon fixation constructs are less stiffthan an interference screw against a bone plug. However, inthis study, the stiffness of a quadrupled hamstring tendon graftwas doubled by use of a cortical screw and washer in additionto a BIS compared to a single fixation device. Graft fixationremains the weak link in the early postoperative period afterligament reconstruction [21,28–31]. Presently the mostpopular options for soft tissue tibial fixation are interferencescrews with or without graft sleeves and screw and washerfixation. However, controversy remains as to the suitability ofsoft tissue fixation for early progressive rehabilitation.

The most common current hamstring graft tibial fixationtechniques utilize bioabsorbable interference screw fixation[11,12,14,32,33]. The screw type, size, and fixation methodvary among surgeons. In a biomechanical study, the four-quadrant sleeve and screw device showed greater resistancestrength in 1 and 2mmgraft laxity, defined as the separation ofthe femur and the tibia at the points of graft fixation [15]. In astudy comparing the effect of screw size, a 35-mmBIS showedsuperior pullout resistance to either a single 28-mm BIS or abicortical method of fixation (single, 20-mm and 17-mmBISs), and the authors recommended 35-mm BIS fixationparticularly for thosewho have less than ideal tibial BMD [27].

Although several authors have suggested that the amountand direction of initial graft tension may be of criticalimportance for the clinical outcome, there is no clear consensusfor what the magnitude of force or the position of the kneeduring tensioning should be [34–38]. Furthermore, biologicstudies have not been conclusive about the optimum tension forrevascularizationwhileminimizing stress relaxation. Clinically,the risk of undertensioning the graft, and thereby not correctingknee laxity, must be balanced with the risk of overtensioningwhichmay lead to pathologic stresses on the joint cartilage [39],graft failure [40], or infrapatellar contracture syndrome [41]. Inthis study, an interesting finding was that with submaximalrough manual force applied, the ACL tension was highest bypulling the graft horizontally with a load cell probe. Also,tension was maintained with a 20-lb manual tensioner, and thenlost after final tibial fixation of the hamstring tendon.

There are several weaknesses to our study: There is inherentvariability in cadaveric tibia BMD that may have affected theresults. Our study resulted in a slightly lower mean load atfailure and stiffness than previous studies [18], whichmight bedue to the older cadaveric specimens we used. The measure-ment of the ACL during rough manual tension was verysubjective since the force was applied by manual traction froman assistant. Since this is a time 0 study, the strength of eachfixation method may also change over time in vivo.

In conclusion, biomechanical testing with cadavers showedthat a BIS with additional cortical screw and spike washer forfixation of the distal hamstring tendon to the tibia resulted in

higher load at failure and stiffness compared to either a singleBIS or cortical screw and spike washer fixation alone.

References

[1] Laxdal G, Kartus J, Hansson L, Heidvall M, Ejerhed L, Karlsson J.A prospective randomized comparison of bone–patellar tendon–boneand hamstring grafts for anterior cruciate ligament reconstruction.Arthroscopy 2005;21:34–42.

[2] Herrington L, Wrapson C, MatthewsM, Matthews H. Anterior cruciateligament reconstruction, hamstring versus bone–patella tendon–bonegrafts: a systematic literature review of outcome from surgery. Knee2005;12:41–50.

[3] Goldblatt JP, Fitzsimmons SE, Balk E, Richmond JC. Reconstructionof the anterior cruciate ligament: meta-analysis of patellar tendonversus hamstring tendon autograft. Arthroscopy 2005;21:791–803.

[4] Forster MC, Forster IW. Patellar tendon or four-strand hamstring? Asystematic review of autografts for anterior cruciate ligamentreconstruction. Knee 2005;12:225–30.

[5] Webster KE, Gonzalez-Adrio R, Feller JA. Dynamic joint loadingfollowing hamstring and patellar tendon anterior cruciate ligamentreconstruction. Knee Surg Sports Traumatol Arthrosc 2004;12:15–21.

[6] Spindler KP, Kuhn JE, Freedman KB, Matthews CE, Dittus RS, HarrellJr FE. Anterior cruciate ligament reconstruction autograft choice: bone–tendon–bone versus hamstring: does it really matter? A systematicreview. Am J Sports Med 2004;32:1986–95.

[7] JanssonKA,LinkoE, Sandelin J,HarilainenA.Aprospective randomizedstudy of patellar versus hamstring tendon autografts for anterior cruciateligament reconstruction. Am J Sports Med 2003;31:12–8.

[8] Feller JA, Webster KE. A randomized comparison of patellar tendonand hamstring tendon anterior cruciate ligament reconstruction. Am JSports Med 2003;31:564–73.

[9] Ejerhed L, Kartus J, Sernert N, Kohler K, Karlsson J. Patellar tendon orsemitendinosus tendon autografts for anterior cruciate ligamentreconstruction? A prospective randomized study with a two-yearfollow-up. Am J Sports Med 2003;31:19–25.

[10] Radford MJ, Noakes J, Read J, Wood DG. The natural history of abioabsorbable interference screw used for anterior cruciate ligamentreconstruction with a 4-strand hamstring technique. Arthroscopy2005;21:707–10.

[11] Hill PF, Russell VJ, Salmon LJ, Pinczewski LA. The influence ofsupplementary tibial fixation on laxity measurements after anteriorcruciate ligament reconstruction with hamstring tendons in femalepatients. Am J Sports Med 2005;33:94–101.

[12] Hayes DA, Watts MC, Tevelen GA, Crawford RW. Central versusperipheral tibial screw placement in hamstring anterior cruciateligament reconstruction: in vitro biomechanics. Arthroscopy2005;21:703–6.

[13] Ma CB, Francis K, Towers J, Irrgang J, Fu FH, Harner CH. Hamstringanterior cruciate ligament reconstruction: a comparison of bioabsorb-able interference screw and endobutton-post fixation. Arthroscopy2004;20:122–8.

[14] Giron F, Aglietti P, Cuomo P, Mondanelli N, Ciardullo A. Anteriorcruciate ligament reconstruction with double-looped semitendinosusand gracilis tendon graft directly fixed to cortical bone: 5-year results.Knee Surg Sports Traumatol Arthrosc 2004.

[15] Starch DW, Alexander JW, Noble PC, Reddy S, Lintner DM.Multistranded hamstring tendon graft fixation with a central four-quadrant or a standard tibial interference screw for anterior cruciateligament reconstruction. Am J Sports Med 2003;31:338–44.

[16] Howell SM, Hull ML. Aggressive rehabilitation using hamstringtendons: graft construct, tibial tunnel placement, fixation properties,and clinical outcome. Am J Knee Surg 1998;11:120–7.

[17] Selby JB, Johnson DL, Hester P, Caborn DN. Effect of screw length onbioabsorbable interference screw fixation in a tibial bone tunnel. Am JSports Med 2001;29:614–9.

459J.C. Yoo et al. / The Knee 13 (2006) 455–459

[18] Brand Jr J, Weiler A, Caborn DN, Brown Jr CH, Johnson DL. Graftfixation in cruciate ligament reconstruction. Am J Sports Med2000;28:761–74.

[19] Howell SM, Wallace MP, Hull ML, Deutsch ML. Evaluation of thesingle-incision arthroscopic technique for anterior cruciate ligamentreplacement. A study of tibial tunnel placement, intraoperative grafttension, and stability. Am J Sports Med 1999;27:284–93.

[20] Hamner DL, Brown Jr CH, Steiner ME, Hecker AT, Hayes WC.Hamstring tendon grafts for reconstruction of the anterior cruciateligament: biomechanical evaluation of the use of multiple strands andtensioning techniques. J Bone Joint Surg Am 1999;81:549–57.

[21] Giurea M, Zorilla P, Amis AA, Aichroth P. Comparative pull-out andcyclic-loading strength tests of anchorage of hamstring tendon grafts inanterior cruciate ligament reconstruction. Am J Sports Med1999;27:621–5.

[22] Scranton Jr PE, Lanzer WL, Ferguson MS, Kirkman TR, Pflaster DS.Mechanisms of anterior cruciate ligament neovascularization andligamentization. Arthroscopy 1998;14:702–16.

[23] Morrison JB. The mechanics of the knee joint in relation to normalwalking. J Biomech 1970;3:51–61.

[24] Morrison JB. Function of the knee joint in various activities. BiomedEng 1969;4:573–80.

[25] Noyes FR, Butler DL, Grood ES, Zernicke RF, Hefzy MS.Biomechanical analysis of human ligament grafts used in knee–ligament repairs and reconstructions. J Bone Joint Surg Am1984;66:344–52.

[26] Noyes FR, Butler DL, Paulos LE, Grood ES. Intra-articular cruciatereconstruction: I. Perspectives on graft strength, vascularization, andimmediate motion after replacement. Clin Orthop 1983:71–7.

[27] Caborn DN, Nyland J, Selby J, Tetik O. Biomechanical testing ofhamstring graft tibial tunnel fixation with bioabsorbable interferencescrews. Arthroscopy 2003;19:991–6.

[28] Bellemans J, Eid T, Fabry G. A modified technique for tibialinterference screw fixation of hamstring anterior cruciate ligamentgrafts. Arthroscopy 1999;15:669–71.

[29] Clark R, Olsen RE, Larson BJ, Goble EM, Farrer RP. Cross-pinfemoral fixation: a new technique for hamstring anterior cruciateligament reconstruction of the knee. Arthroscopy 1998;14:258–67.

[30] Howell SM, Deutsch ML. Comparison of endoscopic and two-incisiontechniques for reconstructing a torn anterior cruciate ligament usinghamstring tendons. Arthroscopy 1999;15:594–606.

[31] Steiner ME, Hecker AT, Brown Jr CH, Hayes WC. Anterior cruciateligament graft fixation. Comparison of hamstring and patellar tendongrafts. Am J Sports Med 1994;22:240–6 discussion 6–7.

[32] Mahirogullari M, Kuskucu M, Kiral A, Pehlivan O, Akmaz I, TirmikU. Early results of reconstruction of chronic anterior cruciate ligamentruptures using four-strand hamstring tendon autografts. Acta OrthopTraumatol Turc 2005;39:224–30.

[33] Nurmi JT, Kannus P, Sievanen H, Jarvela T, Jarvinen M, Jarvinen TL.Interference screw fixation of soft tissue grafts in anterior cruciateligament reconstruction: Part 2. Effect of preconditioning on grafttension during and after screw insertion. Am J Sports Med2004;32:418–24.

[34] Amis AA, Jakob RP. Anterior cruciate ligament graft positioning,tensioning and twisting. Knee Surg Sports Traumatol Arthrosc 1998;6(Suppl 1):S2–S12.

[35] Corsetti JR, Jackson DW. Failure of anterior cruciate ligamentreconstruction: the biologic basis. Clin Orthop 1996:42–9.

[36] Vergis A, Gillquist J. Graft failure in intra-articular anterior cruciateligament reconstructions: a review of the literature. Arthroscopy1995;11:312–21.

[37] Jaureguito JW, Paulos LE. Why grafts fail. Clin Orthop 1996:25–41.[38] Andersen HN, Amis AA. Review on tension in the natural and

reconstructed anterior cruciate ligament. Knee Surg Sports TraumatolArthrosc 1994;2:192–202.

[39] Schabus R, Fuchs M, Kwasny O. The effect of ACL graft preload onthe static pressure distribution in the knee joint. Orthop Trans1990;14:431–2.

[40] Yoshiya S, Andrish JT, Manley MT, Bauer TW. Graft tension inanterior cruciate ligament reconstruction. An in vivo study in dogs. AmJ Sports Med 1987;15:464–70.

[41] Paulos LE, Wnorowski DC, Greenwald AE. Infrapatellar contracturesyndrome. Diagnosis, treatment, and long-term follow-up. Am J SportsMed 1994;22:440–9.