Double-BundleReconstruction of theAnterior Cruciate Ligament:Anatomic andBiomechanical Rationale

AbstractPatients continue to suffer residual pain and instability followinganterior cruciate ligament reconstruction. Although overalloutcomes of anterior cruciate ligament reconstruction arefavorable, improved outcomes can be achieved. Recentbiomechanical studies have questioned the ability of conventionalsingle-bundle anterior cruciate ligament constructs to adequatelyrestore normal knee kinematics. Consequently, the use of double-bundle anterior cruciate ligament constructs has beenrecommended to restore knee stability more effectively. Recentbiomechanical data indicate that double-bundle anterior cruciateligament reconstruction may provide better anteroposterior androtational knee stability than do conventional single-bundletechniques. Studies are needed to evaluate the clinical impact ofdouble-bundle reconstruction techniques on long-term functionaloutcomes.

Surgical management of anteriorcruciate ligament (ACL) defi-ciency has been the subject of in-creased interest for the past severaldecades. Surgical techniques haveadvanced from direct repair and extra-articular reconstructions to mini-mally invasive arthroscopic tech-niques, with various graft choices,sophisticated fixation options, andadvanced rehabilitation protocols.

Most current ACL reconstructiveprocedures have focused on arthro-scopic reconstruction of the antero-medial (AM) bundle of the ACL,with little regard for the posterolat-eral (PL) bundle. Although most pa-tients have no detectable residual in-

stability following single-bundleACL reconstruction, a subset of pa-tients has residual instability andpain. Recent reports of ACL recon-struction have demonstrated that aconsiderable number of patientshave persistent anteroposterior lax-ity and a persistent pivot shift; manypatients will not return to their pre-vious level of athletic activity afterACL reconstruction.1-4 Moreover, re-cent in vivo studies have shown thatsingle-bundle ACL reconstructionsufficiently restores the anteroposte-rior tibial translation but is associat-ed with increased internal tibial ro-tation during running activities.5

These results prompted knee sur-

Boris A. Zelle, MD

Armando F. Vidal, MD

Peter U. Brucker, MD, MSc

Freddie H. Fu, MD, DSc (Hon),

DPs (Hon)

Dr. Zelle is Resident, Department ofOrthopaedic Surgery, University ofPittsburgh School of Medicine,Pittsburgh, PA. Dr. Vidal is AssistantProfessor, Department of OrthopaedicSurgery, University of Colorado atDenver and Health Science Centers,Denver, CO. Dr. Brucker is ResearchFellow, Department of OrthopaedicSurgery, Division of Sports Medicine,University of Pittsburgh. Dr. Fu isChairman and David Silver Professor,Department of Orthopaedic Surgery,Division of Sports Medicine, Universityof Pittsburgh.

None of the following authors or thedepartments with which they areaffiliated has received anything of valuefrom or owns stock in a commercialcompany or institution related directly orindirectly to the subject of this article:Dr. Zelle, Dr. Vidal, Dr. Brucker, andDr. Fu.

Reprint requests: Dr. Fu, Department ofOrthopaedic Surgery, University ofPittsburgh School of Medicine,Kaufmann Building, Suite 1011, 3471Fifth Avenue, Pittsburgh, PA 15213.

J Am Acad Orthop Surg 2007;15:87-96

Copyright 2007 by the AmericanAcademy of Orthopaedic Surgeons.

Perspectives on Modern Orthopaedics

Volume 15, Number 2, February 2007 87

geons to devise a double-bundle ACLreconstruction technique with thegoals of achieving a more anatomicrestoration of the ACL and its twobundles and of improving knee sta-bility and patient outcomes.6-16

Historical Perspective

Initial reports of double-bundle ACLreconstruction appeared in the liter-ature in the 1980s.14,17 In 1983,Mott17 described a double-bundleACL reconstruction technique usingsemitendinosus autograft throughtwo separate tibial and femoral tun-nels. In 1987, Zaricznyj14 described adouble-bundle ACL reconstructiontechnique using a single femoraltunnel and two tibial tunnels. He re-ported good to excellent results in 12of 14 patients at an average of 3.6years postoperatively.14

Radford and Amis18 studied thebiomechanical significance of recon-structing both bundles of the ACL.They believed that the normal func-tion of the ACL was too complex tobe recreated using standard single-bundle techniques. They hypothe-sized that ACL behavior might be

modeled more closely by implantswhich seek to duplicate its complex-ity.18 These authors compared thebiomechanical characteristics of bothover-the-top and through-the-condylesingle-bundle reconstructions with adouble-bundle reconstruction con-sisting of a posterolateral graft passedthrough a femoral drill hole and ananteromedial graft in the over-the-top position. The double-bundle re-construction was the only constructthat reproduced the function of theintact ACL at both 20 and 90 ofknee flexion.

Radford et al19 subsequently stud-ied the in vivo effects of double-bundle reconstruction in a sheepmodel using synthetic polyestergrafts. In the double-bundle group,the authors reported increased boneresorption at the bone tunnels, asubsequent increase in anteriortranslation, and more pronounceddegenerative changes. Their conclu-sions were limited by the animalmodel and choice of graft. Despitethese limitations, the authors rec-ommended that double-bundle re-constructions in ACL surgery werenot indicated. Subsequently, double-

bundle reconstruction for ACL in-sufficiency received limited atten-tion in the literature for manyyears.

Clinical reports and more sophis-ticated biomechanical studies ofdouble-bundle ACL reconstructionsappeared in the literature 5 yearslater.6-13,15,16,20 Several studies de-scribed different techniques and graftoptions. As techniques developed,sophisticated methods of studyingthe biomechanical influences of graftplacement and geometry furtherdeepened understanding of the influ-ence of both conventional single-and double-bundle reconstructionson knee kinematics.13,20-22 Specifi-cally, the ability to evaluate the in-fluence of graft placement on bothanterior translation and coupled ro-tation greatly advanced the under-standing of the PL bundle and its rolein both knee kinematics and ACL re-construction.

Anatomy andBiomechanics

Anatomy of the AnteriorCruciate Ligament

The ACL consists of dense con-nective tissue enveloped in a syn-ovial membrane, which places theligament in an intra-articular but ex-trasynovial position.23,24 The liga-ment originates from a fossa on theposterior aspect of the lateral femo-ral condyle and runs in an obliquecourse distally, anteriorly, and medi-ally to insert between the medialand lateral spines of the tibial pla-teau. The ACL attaches to the femurand the tibia, not as a singular cord,but rather as a collection of fasciclesthat fan out as they approach theirinsertion sites.23-27 The area of inser-tion is 3 to 3.5 times larger than thecross-sectional area of the ligamentmidsubstance.27 The cross-sectionalarea of the midsubstance ACL is ap-proximately 36 to 44 mm2,28 whilethe cross-sectional area of the femo-ral and tibial insertion sites is be-tween 113 and 136 mm2.27 This

Figure 1

Location of the anteromedial bundle (AMB) and posterolateral bundle (PLB) of theanterior cruciate ligament in a left knee. The AMB attaches more proximally to thelateral femoral condyle (LFC) than does the PLB. The PLB runs more nearlyhorizontally and attaches more distally to the LFC. MFC = medial femoral condyle,PCL = posterior cruciate ligament

Double-Bundle Reconstruction of the Anterior Cruciate Ligament

88 Journal of the American Academy of Orthopaedic Surgeons

anchor-shaped design of the relative-ly broad ACL footprint may serve tominimize stresses on the ligament-bone interface.27

The ACL consists of individual fi-ber bundles (Figure 1). Palmer29 firstprovided anatomic descriptions ofthe ACL bundles in the 1930s. Re-searchers have since confirmed thatthe ACL consists of two separateand functionally distinct anatomicbundles.23,26,27 These bundles dem-onstrate clearly different biome-chanical properties and tensioningpatterns with varying degrees ofknee flexion.30,31

Each bundle is named accordingto its insertion on the tibial foot-print. On the tibial side, the AMbundle inserts anteromedially andthe PL bundle inserts posterolateral-ly. On the femoral side, the AM bun-dle originates more proximally thandoes the PL bundle24,26,27 (Figure 2).On the whole, the AM and PL bun-dles comprise relatively equal pro-portions of the overall ligament in-sertion on both the femoral andtibial footprints.27

Physiologic Function of theAnterior Cruciate LigamentBundles

Biomechanical investigationshave demonstrated that the ACL fi-ber bundles are not isometricthroughout knee flexion and exten-sion. The AM bundle tightens inflexion and the PL bundle relaxes,whereas in extension, the PL bundletightens and the AM bundle relax-es.23,32,33 Cadaveric investigationshave demonstrated that both the AMand PL bundles show their maxi-mum shortening peak at 30 of kneeflexion. In extension from this posi-tion, the PL bundle elongates. Con-versely, with increasing flexion, theAM bundle elongates while thelength of the PL bundle essentiallyremains constant.33

In response to anterior and com-bined rotatory loads, forces are dis-tributed differentially between theAM and the PL bundles. Sakane et

al31 evaluated the magnitude of thein situ forces within the individualbundles of the ACL in response toapplied anterior loads at varyingknee flexion angles. The highest insitu force within the PL bundle in re-sponse to an anterior load occurredbetween 0 and 45 of knee flexion,with a maximum in situ forcereached at 15. Between 0 and 15 ofknee flexion, the in situ forceswithin the PL bundle in response toan anterior load were significantlyhigher (P < 0.05) than those observedin the AM bundle. In contrast, themagnitude of the in situ forceswithin the AM bundle in responseto an anterior load was relativelyconstant throughout knee flexionand extension.

Similar results were obtained byGabriel et al,30 who investigated thedistribution of the in situ forces inthe ACL in response to both anteriorand combined rotatory loads. Theydiscovered that in response to an an-terior tibial load of 134 N, the in situforce in the PL bundle increased withknee extension and decreased withknee flexion. In extension, the in

situ force in the AM bundle in re-sponse to anterior loads was lowerthan in the PL bundle. However, theforces in the AM bundle increasedwith knee flexion, reaching a maxi-mum at 60. In response to a com-bined rotatory load of 10 Nm valgustorque and 5 Nm internal tibialtorque, the AM bundle experiencedgreater in situ forces than did the PLbundle at 15 and 30 of knee flexion.However, both bundles demon-strated significant (P < 0.05) in situforces in response to a combined ro-tatory load, especially at 15.

Evaluations of the length changepatterns and in situ forces in the in-dividual bundles of the ACL indicatethat both the AM and PL bundles areimportant to the stability of the kneejoint. The AM bundle is an impor-tant stabilizer of the knee in flexion,whereas the PL bundle acts as a sta-bilizer against anterior loads whenthe knee is in extension. Both theAM and the PL bundles provide sig-nificant contributions to the stabil-ity of the knee in response to com-bined rotatory loads.

Figure 2

Tibial and femoral insertion sites of the left knee. A, On the femoral side, theanteromedial (AM) bundle inserts more proximally than the posterolateral (PL)bundle. B, On the tibial side, the AM bundle inserts anteromedially and the PLbundle inserts posterolaterally. ANT = anterior, DIST = distal, POST = posterior,PROX = proximal

Boris A. Zelle, MD, et al

Volume 15, Number 2, February 2007 89

Biomechanics ofSingle-Bundle VersusDouble-Bundle ACLReconstruction

In 1994, Radford et al19 succinct-ly stated that single-bundle ACL re-constructions are a great simplifi-cation of the complexity of thenatural ligament structure and thatit would therefore be surprising ifsuch a reconstruction were capableof recreating all the subtleties of nor-mal knee behavior. The advent ofnewer biomechanical testing meth-ods has resulted in a better appreci-ation of knee kinematics followingACL reconstruction. The use ofrobotic/universal force-moment sen-sor testing systems has allowed re-searchers to assess knee kinematicsin cadaveric models with multipledegrees of freedom. Additionally,these systems allow for assessmentof the in situ forces within recon-structed ligaments in response toknown loads and translations.13,20-22

Recently, Woo et al22 demon-strated in vitro that conventionalsingle-bundle ACL reconstructionsare relatively successful at limitinganterior translation of the tibia in re-sponse to anterior-directed tibialloads, but they are ineffective at re-sisting combined rotatory loads.These investigators believed thatconventional single-bundle tech-niques place the graft too close to thecentral axis of the knee, thus limit-ing the ability of the graft to resist ro-tatory loads.

Several biomechanical studieshave since been published docu-menting the superiority of double-bundle techniques at limiting notonly anterior translation but, moreimportantly, translation in responseto combined rotatory loads. In a ca-daveric study, Mae et al34 demon-strated that double-bundle ACLreconstruction using hamstring ten-don grafts and two femoral tunnelsprovides greater anterior stabilitythan do current single-bundle tech-niques.

Yagi et al13 observed that double-

bundle ACL reconstruction is moresuccessful at restoring knee kine-matics to normal than is conven-tional single-bundle ACL recon-struction, which simply reconstructsthe AM bundle. In an in vitro study,these authors demonstrated that, onapplication of an anterior tibial loadof 134 N, anterior tibial translationwas significantly (P < 0.05) closer tonormal with double-bundle ACLreconstruction than with single-bundle ACL reconstruction. Addi-tionally, with application of a com-bined rotatory load of 10 Nm valgustorque and 5 Nm internal tibialtorque at 30 of knee flexion, the ki-nematics of the double-bundle re-construction were significantly (P