CLINICAL ORTHOPAEDICS AND RELATED RESEARCHNumber 402, pp. 9–20© 2002 Lippincott Williams & Wilkins, Inc.

9

This review of the literature assessed what isknown about the biomechanics of the normal an-terior cruciate ligament during rehabilitation ex-ercises, the biomechanical behavior of the ante-rior cruciate ligament graft during healing, andclinical studies of rehabilitation after anteriorcruciate ligament replacement. After anteriorcruciate replacement, immobilization of the knee,or restricted motion without muscle contraction,leads to undesired outcomes for the ligamentous,articular, and muscular structures that surroundthe joint. It is clear that rehabilitation that incor-porates early joint motion is beneficial for reduc-ing pain, minimizing capsular contractions, de-creasing scar formation that can limit jointmotion, and is beneficial for articular cartilage.There is evidence derived from randomized con-trolled trials that immediately after anterior cru-ciate ligament reconstruction, weightbearing ispossible without producing an increase of ante-rior knee laxity and is beneficial because it lowersthe incidence of patellofemoral pain. Rehabilita-

tion with a closed kinetic chain program results inanteroposterior knee laxity values that are closerto normal, and earlier return to normal daily ac-tivities, compared with rehabilitation with anopen kinetic chain program. This review re-vealed that more randomized, controlled trials ofrehabilitation are needed. These should includethe clinicians’ and patients’ perspective of theoutcome, and biomarkers of articular cartilagemetabolism.

From a clinical standpoint, the long-term out-come of anterior cruciate ligament reconstruc-tion is dependent on surgical and rehabilitationvariables and although a substantial amount ofclinical and biomechanical investigations havefocused on the former, much less research hasbeen directed at the latter. This is not surpris-ing because successful anterior cruciate liga-ment reconstruction is dependent on many sur-gical variables that only recently have becomewell understood (selection of an appropriategraft material, proper positioning of the graft,adequate tensioning of the graft, reliant fixa-tion); however, this presents a concern becauseoptimal healing of an anterior cruciate liga-ment graft and knee also is dependent on reha-bilitation. Although most would agree that thestrains applied to an anterior cruciate ligamentgraft by body weight, muscle activity, and jointmotion affect its healing response, there is lit-

The Science of Anterior CruciateLigament Rehabilitation

Bruce D. Beynnon, PhD; Robert J. Johnson, MD; and Braden C. Fleming, PhD

From the McClure Musculoskeletal Research Center,Department of Orthopaedics & Rehabilitation, Univer-sity of Vermont, Burlington, VT.This study was funded by a grant from the ArthritisFoundation.Reprint requests to Bruce D. Beynnon, PhD, McClureMusculoskeletal Research Center, Department of Ortho-paedics & Rehabilitation, University of Vermont, 438AStafford Hall, Burlington, VT 05405.DOI: 10.1097/01.blo.0000026072.30435.95

tle consensus regarding how these factors in-fluence the biomechanical behavior of the heal-ing graft and, in turn, how this modulates itshealing response.

The objective of the current study was to re-view the current knowledge regarding the ef-fect of rehabilitation activities on the normalanterior cruciate ligament (data that are usefulfor establishing the basis of rehabilitation pro-grams), the basic science of the healing responseof anterior cruciate ligament grafts, and clini-cal investigations of the healing anterior cru-ciate ligament graft and knee.

Biomechanics of the Normal AnteriorCruciate LigamentRecognizing that it is important to understandthe biomechanical behavior of the normal an-terior cruciate ligament before developing cri-teria for reconstruction and rehabilitation of ahealing anterior cruciate ligament graft, a tech-nique to measure the strain behavior of thenormal anterior cruciate ligament in vivo wasdeveloped.10 The primary objective of thiswork was to develop clinical criteria for re-construction and rehabilitation of this impor-tant structure that are based on normal musclefunction and include the loads produced bybody weight. Study participants were patientvolunteers with normal anterior cruciate liga-ments and no history of ligament trauma whowere candidates for diagnostic arthroscopicsurgery done under local anesthesia, allowingthem full control of their leg muscles. Aftercompletion of the surgical procedure that typ-ically involved treatment of a meniscus tear, aDifferential Variable Reluctance Transducer(DVRT, MicroStrain Inc, Burlington, VT) wasattached to the anterior cruciate ligament tomeasure its strain behavior.10 The strain be-havior of the anterior cruciate ligament duringcommon rehabilitation activities such as iso-metric contraction of the quadriceps,11 iso-tonic contraction of the quadriceps,11,13 squat-ting,13 bicycling,25 stair climbing,26 and whilevarious loads were applied to the knee fittedwith a functional brace12,16,27 have been stud-ied. Rehabilitation exercises that involve iso-

metric contraction of the hamstring muscles,do not strain the anterior cruciate ligament atany knee position or magnitude of muscle con-traction. In contrast, exercises that engage thepowerful extensor mechanism produce ante-rior cruciate ligament strain values that de-pend on knee flexion angle and the magnitudeof muscle contraction. Specifically, isometricexercises that strain the anterior cruciate liga-ment involve contraction of the dominantquadriceps muscle group with the knee be-tween extension and 60� flexion, or involveisotonic contraction of the quadriceps betweenextension and 50� flexion. The largest anteriorcruciate ligament strain magnitudes that havebeen measured are produced by isometric andisotonic contraction of the quadriceps muscleswith the knee near extension. Increasing resis-tance during active flexion and extension mo-tion of the leg, and open kinetic chain exercisethat does not incorporate body weight loadingand does not involve appreciable cocontrac-tion of the muscles spanning the knee, gener-ates a significant increase in anterior cruciateligament strain values. In contrast, increasingresistance during squatting, a closed kineticchain exercise that involves body weight load-ing and substantial cocontraction of the mus-cles spanning the knee, does not create an ap-preciable change in anterior cruciate ligamentstrain values. A similar finding was shown forstationary bicycling, another closed kineticchain exercise; during this activity, differentlevels of resistance (power settings of 75, 125,and 175 W) and pedal speeds (cadences of 60and 90 rpms) produced similar anterior cruci-ate ligament strain values.25 These findings in-dicate that rehabilitation after anterior cruciateligament reconstruction with closed kineticchain exercises such as squatting and bicy-cling permits increased muscle activity with-out subjecting the anterior cruciate ligament toincreased strain values.

The previously mentioned anterior cruciateligament strain data are a portion of a compre-hensive database that includes a ranking of ex-ercises based on the peak strain values pro-duced during the respective activities, which

Clinical Orthopaedics10 Beynnon et al and Related Research

can be used as the basis for rehabilitation ofhealing anterior cruciate ligament grafts (Table1). It is important to emphasize that the limitsof strain that are safe for an injured ligamentor healing graft currently are unknown. Al-though the current authors cannot identify whichexercises are either safe or harmful, these datahave been used to design anterior cruciate lig-ament rehabilitation programs that are beingcompared in a prospective, randomized, dou-ble blinded trial. With adequate followup, thisshould help identify those exercise programsthat are safe for a healing anterior cruciate lig-ament graft.

Biomechanical Studies of Healing AnteriorCruciate Ligament GraftsStudies of healing anterior cruciate ligamentgrafts that have been done using animal modelsprovide insight into the biologic remodelingand biomechanical behavior of the graft during

healing14; however, application of the findingsfrom these studies to the clinical environmentand rehabilitation of a healing graft must bedone with caution because knees in animals aredifferent than those in humans, and they havean uncontrolled rehabilitation regimen.

Butler et al20 used the primate model to showthat an anterior cruciate ligament graft under-goes a large decrease of its ultimate failurestrength and linear stiffness during the first 6weeks of healing, and with time the structuralbehavior of the graft improves, although itnever returns to that of the normal anteriorcruciate ligament. Studies of anterior cruciateligament reconstruction using the patellar ten-don autograft in canine, goat, rabbit and pri-mate models have reported ultimate failureloads that range between 11% and 50% of thecontrol anterior cruciate ligament, and stiff-ness values that range between 13% and 57%of normal after 1 year or more of healing.6,43

Number 402September, 2002 Anterior Cruciate Ligament Rehabilitation 11

TABLE 1. Rank Comparison of Mean Peak Anterior Cruciate Ligament StrainValues During Various Knee Loading Conditions and Commonly PrescribedRehabilitation Activities

Number of Rehabilitation Activity Peak Strain Subjects

Isometric quadriceps contraction at 15� (30 N-m of extension torque) 4.4 (0.6)% 8Squatting with Sport Cord 4.0 (1.7)% 8Lachman test (150 N of anterior shear load; 30� flexion) 3.7 (0.8)% 10Squatting 3.6 (1.3)% 8Isometric gastrocnemius contraction at 15� (15 N-m of plantar flexion torque) 3.5 (0.9)% 6Active flexion and extension (no weight boot) of the knee 2.8 (0.8)% 18Simultaneous quadriceps and hamstrings contraction at 15� 2.8 (0.9)% 8Isometric gastrocnemius contraction at 5� (15 N-m of plantar flexion torque) 2.8 (2.3)% 6Isometric quadriceps contraction at 30� (30 N-m of extension torque) 2.7 (0.5)% 18Stair climbing 2.7 (2.9)% 5Weightbearing at 20� knee flexion 2.1 (1.7)% 11Anterior drawer (150 N of anterior shear load; 90� flexion) 1.8 (0.9)% 10Stationary bicycling 1.7 (1.9)% 8Isometric hamstrings contraction at 15� (to �10 N-m of flexion torque) 0.6 (0.9)% 8Simultaneous quadriceps and hamstrings contraction at 30� 0.4 (0.5)% 8Isometric gastrocnemius contraction at 30� (15 N-m of plantar flexion torque) 0.2 (1.4)% 6Passive flexion and extension of the knee 0.1 (0.9)% 10Isometric quadriceps contraction at 60� and 90� (30 N-m of extension torque) 0.0% 8Isometric gastrocnemius contractions at 45� (15 N-m of plantar flexion torque) 0.0% 6Simultaneous quadriceps and hamstrings contraction at 60� 0.0% 8Simultaneous quadriceps and hamstrings contraction at 90� 0.0% 8Isometric hamstrings contraction at 30�, 60�, and 90� (�10 N-m of flexion) 0.0% 8

Likewise, investigations of anterior cruciateligament reconstruction with an iliotibial tractautograft have revealed ultimate failure loadvalues that range between 23% and 40% of thenormal anterior cruciate ligament althoughstiffness was reported to be 45% of normal.43

Recognizing that an anterior cruciate liga-ment graft must have adequate structural prop-erties and control anterior displacement of thetibia relative to the femur, several investigatorshave used animal models to measure the an-teroposterior (AP) load-displacement responseof the knee (AP laxity) during graft heal-ing.6,33,43 Hulse et al33 revealed that anteriorcruciate ligament graft healing is not only asso-ciated with a decrease in the structural proper-ties of the graft, but also is accompanied by sub-stantial increases of anterior displacement oftibia relative to the femur. Investigations of thehealing patella tendon autograft done in ani-mals have shown that AP joint laxity rangesbetween 156% and 269% of the contralateralnormal side after 1 year of healing.6,14,33,43

Therefore, investigations of anterior cruciateligament grafts that have been done in animalsindicate that they lose their ultimate failurestrength and undergo a decrease of stiffness,and the knees have an increase in anterior lax-ity develop during healing. The exact cause ofthese changes and the application of these datato humans are unclear. For example, Rougraffand Shelbourne49 reported that between 3 and 8weeks after transplantation, a substantial por-tion of the graft remained histologically similarto patellar tendon tissue. This led the investiga-tors to suggest that a large proportion of theoriginal tendon survives and that anterior cruci-ate ligament graft healing in humans may notundergo the same complete necrotic stage thathas been reported to occur in animals.3,44 Addi-tional evidence of dissimilarity of the healingresponse of anterior cruciate ligament grafts be-tween human and animals can be seen in a casestudy of a 37-year-old man who had anteriorcruciate ligament reconstruction with a bone-patella tendon-bone autograft and 8 months ofhealing.15 The ultimate failure load and linearstiffness properties of the graft approached that

of the contralateral normal anterior cruciate lig-ament, and AP laxity of the reconstructed kneewas somewhat larger than the normal knee.Compared with the aforementioned animalstudies of the biomechanical behavior of heal-ing anterior cruciate ligament grafts, the bio-mechanical behavior of the reconstructed kneeand bone-patellar tendon-bone graft that weremeasured in the human seem to be superior.

Increased anterior knee laxity is a concernbecause it is associated with altered contactloading of the tibiofemoral articulation, andthis was the motivation for a study of the rela-tionship between AP knee laxity and the struc-tural properties of healing patellar tendon au-tografts used to reconstruct the anterior cruciateligament in a canine model.14 An inverse cor-relation was established between the structuralproperties of the anterior cruciate ligament graft(characterized by the ultimate failure strength,and linear stiffness values) and increased an-terior knee laxity. Increased anterior knee lax-ity was correlated with a decrease in the struc-tural properties of the anterior cruciate ligamentgraft. This finding indicates that clinical stud-ies that show increased anterior displacementof the tibia relative to the femur during graftremodeling may reflect a graft that has inferiorstructural properties. Unfortunately, the mech-anism that created the inverse relationship be-tween increased anterior knee laxity and thestructural properties of the graft was not de-scribed. It may be that this was produced bysurgical variables that were not controlled. Al-ternatively, it may be that the increase of ante-rior knee laxity was produced by excessivegraft strain during healing that led to perma-nent elongation of the graft.

Studies of tendon repairs have revealed thatcontrolled loading can enhance the quality andrate of healing. For example, matrix collagenand repair cells become aligned with the axisof load applied to a healing tendon repair,whereas in the absence of load the matrix andrepair cells become disorganized.1,5 In addi-tion, controlled loading after medial collateralligament injury also enhances the healing re-pair site by increasing the wet and dry weight

Clinical Orthopaedics12 Beynnon et al and Related Research

of the injured ligament, and inducing the rapidreturn of normal tissue deoxyribonucleic acidcontent, collagen synthesis, and strength.4,28,40,54

In contrast, excessive loading has the potentialto disrupt a healing ligament and may inhibitrepair, particularly in a knee with combinedinjuries. For example, a study of healing me-dial collateral ligaments in rats with otherwisestable knees revealed that forced exercise in-creased the strength and stiffness of the liga-ment repair and those animals with unstablejoints did not suffer a change in structuralproperties of the ligament, but instead had in-creased joint instability develop.19 Tendon re-pair combined with 3 weeks of mobilizationresults in a construct with a twofold increasein failure strength compared with the same re-pair with 3 weeks of immobilization whenevaluated 12 weeks after the index injury.29

Most of what is known about the biology ofhealing connective tissue has come from stud-ies done on injured medial collateral ligamentsand tendons, both of which are extraarticularstructures, and it is unclear how the findingsfrom these studies relate to healing intraartic-ular tissues such as an anterior cruciate liga-ment graft.

Retrospective and Observational Studiesof Anterior Cruciate LigamentRehabilitationBasic science and clinical investigations ofsubjects after anterior cruciate reconstructionhave revealed that immobilization of the knee,or limited motion without muscle activity, re-sults in an unwanted outcome (inferior struc-tural and material properties) for the structuresthat surround the knee (ligaments, cartilage,bone, and musculature).2,5,31,35–37,39,45,55 Earlyjoint motion after anterior cruciate ligament re-construction certainly is beneficial, it leads to areduction in pain, lessens adverse changes inarticular cartilage, and helps prevent the for-mation of scar and capsular contractions thathave the potential to limit joint motion.18,38

A retrospective investigation of rehabilita-tion after anterior cruciate ligament recon-struction was done by Shelbourne and Nitz.52

They showed that aggressive rehabilitation thatincluded immediate walking with full weight-bearing, and return to sports by 8 weeks wasmore effective than a conservative rehabilita-tion program. One quarter of the individualsin the aggressive and conservative treatmentgroups had a 3-mm increase in anterior kneelaxity of the reconstructed side in comparisonwith the normal side. An increase of this mag-nitude is a concern from a biomechanical per-spective because it is greater then 2.7 mm, orthe side-to-side difference in AP knee laxitymeasured from subjects with normal knees.22

However, it is unclear whether this increase isa concern from a biologic perspective; it maybe that increases of knee laxity that are withincertain limits of normal do not result in al-tered metabolism of the articular cartilage andmeniscus, or produce additional intraarticularinjury. There currently are no data that de-scribe this relationship.

A retrospective study of early (2–6 months)versus late (7–14 months) return to activitiesthat involve cutting was done by Glasgow etal.30 Early return to activity did not produce anincreased prevalence of reinjury. The subjects’report of knee function and KT-1000 mea-surements of anterior knee laxity were similarbetween the two treatments.

Barber-Westin and Noyes7 did a retrospec-tive study of rehabilitation after anterior cru-ciate ligament reconstruction. Rehabilitationconsisted of four separate phases, the first ofwhich was an assisted ambulatory phase in-volving immediate continuous passive kneemotion and partial weightbearing with crutchsupport until postoperative Weeks 7 to 9. Thesecond phase was from postoperative Week 9through Week 16 and consisted of early strengthtraining. The third phase was from postopera-tive Week 16 through Week 52 and was com-prised of intensive strength training. The fourthphase was the return to sport-specific activity.Fifty-four percent of the subjects who had an-terior cruciate ligament reconstruction with abone-patellar tendon-bone autograft experi-enced an increase in anterior knee laxity of 3mm or greater as evaluated with the KT-1000

Number 402September, 2002 Anterior Cruciate Ligament Rehabilitation 13

arthrometer. Half of these subjects had in-creased anterior knee laxity greater than nor-mal within 1 year of the operation, whereas theremaining 1⁄2 experienced increases 1 year orlater after reconstruction. Twenty-eight per-cent of subjects who had anterior cruciateligament reconstruction with a bone-patellartendon-bone autograft combined with an ili-otibial band extraarticular procedure had anincrease in anterior knee laxity that was greaterthan 3 mm. Again, 1⁄2 of these subjects sus-tained abnormal increases of laxity within thefirst year of surgical reconstruction, whereasthe other 1⁄2 had abnormal increases in laxity af-ter the first year of healing was complete. Thisinvestigation emphasizes the importance offollowing up subjects for a minimum of 2 to 4years after reconstruction before establishingthe merits of a particular reconstruction andrehabilitation protocol. Barber-Westin et al8did a subsequent observational study of reha-bilitation after anterior cruciate ligament re-construction with the bone-patella tendon-bone autograft using the same four phases ofrehabilitation previously described. This wascomprised of the assisted ambulatory phasethat lasted until postoperative Weeks 4 to 8and included partial weightbearing with theuse of a cane or a crutch and closed kineticchain exercises such as minisquats and toeraises, isometric quadriceps contraction, rangeof motion (ROM) exercises, and straight legraises. The second phase (early strength train-ing) occurred from Weeks 4 to 8 throughWeeks 12 to 16, and added proprioceptive,balance, and gait training. The third phase var-ied depending on the subjects’ needs and tookplace between Weeks 12 to 16 and Weeks 24to 52. This included running, stair climbing,and bicycling programs combined with pro-gressive resistive exercises. Successful com-pletion of the aforementioned phase qualifiedthe patient for the fourth and final phase, re-turn to sports. At 2-year followup, 85% of thesubjects had KT-1000 arthrometer measure-ments of knee laxity that were less than 3 mm(injured minus normal differences), 10% haddifferences between 3 and 5.5 mm, and 5%

had differences greater than 5.5 mm. Thisfinding led the authors8 to conclude that ante-rior cruciate ligament reconstruction with abone-patella tendon-bone graft and the four-phase rehabilitation described results in an ac-ceptable failure rate of 5%.

Howell and Taylor32 did an observationalstudy of anterior cruciate ligament reconstruc-tion with the double-looped semitendinosusand gracilis graft. Rehabilitation included con-tinuous passive motion for the first 1 to 2 daysafter surgery, followed by toe touch weight-bearing for 3 weeks, unrestricted exercises af-ter 4 weeks, running in a straight line after 8 to10 weeks, and then return to sports after 4months of healing. Evaluation of AP knee lax-ity with the KT-1000 manual maximum exam-ination at the 4-month and 2-year followup re-sulted in differences in laxity that were 3 mm orgreater in 18% and 11% of the subjects, re-spectively. At these same intervals, the inci-dence of a positive pivot test was 11% and 10%,respectively. Thigh girth and range of joint mo-tion were similar at 4 month and at 2 years.From these findings, the authors concluded thatunlimited return to sport and work activities 4months after anterior cruciate ligament recon-struction with a double-looped semitendinosus-gracilis graft is safe and effective.32

The findings from the retrospective and ob-servational studies of anterior cruciate liga-ment rehabilitation should be considered withappreciation for the design characteristics as-sociated with these approaches. For example,because inclusion and exclusion admissioncriteria were not established a priori, there is apotential for a susceptibility bias to be intro-duced. This bias could arise from comparingthe outcomes of rehabilitation between groupsof subjects that differ prognostically. Compar-ing patients with multiple ligament injuries(anterior cruciate ligament and medial collat-eral ligament injury, or anterior cruciate liga-ment and posterior lateral ligament complexinjury), or combined articular injuries (menis-cus or chondral lesions) with those who havean isolated anterior cruciate ligament injury allwould yield a susceptibility bias. The findings

Clinical Orthopaedics14 Beynnon et al and Related Research

from retrospective studies of rehabilitationalso may be influenced by a performance biasthat results from changes in the level of skill indoing a particular surgical procedure, or exe-cuting a rehabilitation program. A transferbias also may exist when comparisons aremade to a historic rehabilitation group that hada substantially longer followup. For example,it may be that as time progresses, anterior kneelaxity increases for reasons other than the re-habilitation program or the subject’s activitylevel; other factors intrinsic to the biology ofthe healing graft that currently are unknownmay influence the graft during an extendedtime. Control of the treatment being studiedalso is important if the eventual goal is to un-derstand its effect on outcome. This is evi-denced by the findings from a recent prospec-tive, randomized, double blinded investigationof rehabilitation after anterior cruciate liga-ment reconstruction that revealed compliancewith a clinic-based program varies betweensubjects and declines substantially with time(Fig 1). (Uh BS, Beynnon BD, Johnson RJ, etal: A prospective, randomized study of patient

compliance with two rehabilitation programsfollowing anterior cruciate ligament recon-struction. Presented at the Eighth Congress ofthe European Society of Sports Traumatology,Knee Surgery and Arthroscopy, Nice, France,1998.) Subjects with acute, isolated anteriorcruciate ligament disruptions had reconstruc-tion with a bone-patellar-tendon graft and thenwere randomized to either an accelerated orconservative rehabilitation program. After 16weeks of rehabilitation, there was no differ-ence between the programs regarding the sub-ject’s total compliance with the exercises thatwere prescribed (Fig 1). After this period,there was a significant decrease in compliancefor both programs, and subjects attributed thisto the length of the program. This finding isimportant because compliance with rehabilita-tion is a measure of the cumulative strains im-parted to a healing graft, and in view of the factthat this biomechanical dose is directly linkedto the healing response of the graft and knee,it is important for establishing the type of re-habilitation the subject had. For example, asubject could be considered to be in an accel-

Number 402September, 2002 Anterior Cruciate Ligament Rehabilitation 15

Fig 1. Patients in the accelerated and nonaccelerated rehabilitation programs used a standardized logto document the number of sets and repetitions of each exercise that was done on a daily basis. Com-pliance then was calculated as the proportion of exercise completed in relation to the number of exer-cises that were prescribed. These data are plotted as mean values for all exercises for patients in theaccelerated and nonaccelerated programs (vertical axis), with time (horizontal axis). Some subjectsovercomplied, whereas others undercomplied. Compliance with the accelerated and nonacceleratedrehabilitation programs was similar; however, compliance decreased with time.

erated rehabilitation program, and after partic-ipating for a week, have patellofemoral paindevelop that limits his or her participation withaggressive exercises such as isolated contrac-tion of the quadriceps with the knee near exten-sion. Subsequently, he or she may self-selectout of the program by altering the frequencyand magnitude of the extension exercises, oreven immobilize the limb. In effect, this wouldresult in a program that was conservative.Conversely, disregarding the limitations of aconservative program could make it equiva-lent to a more accelerated program. Unlesscompliance is monitored during a clinicalstudy of rehabilitation, it is very difficult to as-sess what a subject actually completes through-out the program, it is even more complex toevaluate how the graft is loaded during heal-ing, and therefore, it may become unclear whateffect the rehabilitation treatment being stud-ied has on outcome. This is very difficult torecreate with a retrospective study design.

Prospective, Randomized Studies ofAnterior Cruciate LigamentRehabilitationNoyes et al46 did a prospective, randomizedinvestigation of immediate versus delayed mo-tion rehabilitation after anterior cruciate liga-ment reconstruction and revealed that contin-uous passive knee motion immediately afteranterior cruciate ligament reconstruction didnot lead to an increase in anterior knee laxityduring healing. Subjects in the immediate mo-tion program had continuous passive motionof the knee on the second postoperative day,whereas those in the delayed motion grouphad their knees placed in a brace at 10� flexion,and began continuous passive motion on theseventh postoperative day. Subjects in both re-habilitation programs reported similar inci-dences of joint effusion, hemarthrosis, soft tis-sue swelling, flexion and extension limits ofthe knee, use of pain medications, and time ofstay in the hospital.

Rosen et al48 did a prospective, randomizedstudy of rehabilitation after arthroscopically-assisted anterior cruciate ligament reconstruc-

tion with a bone-patellar tendon-bone auto-graft. They extended the work of Noyes et al,46

by showing that continuous passive motionduring the first month after anterior cruciateligament reconstruction compared with earlyactive motion produced similar range of jointmotion, and KT-1000 measurements of APknee laxity.

Richmond et al47 did a prospective, ran-domized study that compared the effects ofcontinuous passive knee motion for 4 daysversus 14 days after arthroscopically-assistedanterior cruciate ligament reconstruction witha bone-patellar tendon-bone autograft. Theyreported similar values of knee ROM andlower limb girth between treatment groups.

Jorgensen et al34 did a prospective, random-ized investigation of the effect of weightbearingafter anterior cruciate ligament reconstructionwith the iliotibial band. After surgery, subjectswere randomized to rehabilitation with eitherimmediate weightbearing or rehabilitation withnonweightbearing for 5 weeks followed by agradual return to full weightbearing during thefirst 9 weeks of healing. At the 2-year followup,there was no difference between the groups re-garding AP knee laxity, and patient activitylevel (evaluated with the Tegner and Interna-tional Knee Documentation Committee scores).Similar findings were reported by Tyler et al,53

who did a prospective, randomized investiga-tion of anterior cruciate ligament reconstructionwith a central third bone-patellar tendon-boneautograft followed by rehabilitation with eitherimmediate weightbearing, or the same programbut with a 2-week delay of weightbearing. Im-mediate weightbearing did not effect AP kneelaxity (evaluated by clinical examination andKT-1000), and resulted in a decreased inci-dence of patellofemoral pain compared with re-habilitation with a 2-week delay of weight-bearing. The findings from these investigationsindicate that weightbearing immediately afteranterior cruciate ligament reconstruction doesnot seem to produce excessive loads across ahealing graft that permanently deform the graftor its fixation and is beneficial because it lowersthe incidence of patellofemoral pain.

Clinical Orthopaedics16 Beynnon et al and Related Research

Bynum et al21 did a prospective, random-ized study that compared open kinetic chainversus closed kinetic chain rehabilitation afteranterior cruciate ligament reconstruction witha bone-patella tendon-bone autograft. Imme-diately after surgery, the subject’s knees wereplaced in a rehabilitation brace that was ad-justed to allow 0� through 90� motion, andstarted continuous passive motion from 0� to60� flexion. Rehabilitation started on the firstpostoperative day and included passive andactive motion of the knee without external re-sistance for all subjects. Partial weightbearingwith the use of crutches was permitted, andsubjects progressed to full weightbearing astolerated. Subjects randomized to the open ki-netic chain treatment group had rehabilitationwith introduction of exercises in the followingsequence: Weeks 0 through 3 (cocontractionisometrics, hamstring concentric and eccentricisotonics); Week 3 (straight leg raises with theknee at 30� flexion); Week 6 (isotonic quadri-ceps contraction with low resistance, biking,and proprioception training); Week 8 (isoki-netic hamstrings); Week 12 (unrestricted iso-tonic contraction of the quadriceps); Week 16( jogging, forward and backward running, andsingle-leg deep knee bends); Week 24 (unre-stricted progressive resistance training, andisokinetic quadriceps contractions); Weeks 24through 28 (progressive running and sport-specific activity); Week 32 (noncutting, pivot-ing, and jumping); and Week 52 (return to un-restricted sports). Those subjects randomizedto the closed kinetic chain rehabilitation treat-ment had the following sequence: Weeks 0through 8 (1⁄3 knee bends during double-legstance, seated leg presses, and hamstring curls);Week 6 (stationary bicycling and propriocep-tion training); Week 8 (1⁄3 knee bends duringsingle-leg stance, forward and backward walk-ing against resistance produced by the SportCord (Sport Cord Inc, Sandy, UT) with pro-gression to slow jogging); Week 12 (continu-ation of previous exercises with the progressiveuse of a stiffer Sport Cord with the addition ofside–to-side jumping against the resistance ofthe Sport Cord); Week 16 (sport-specific exer-

cises against the elastic resistance produced bythe Sport Cord, with addition of free weightleg presses and squats); Week 24 (progressiverunning and sport-specific activity); Week 32(noncutting, pivoting, and jumping sports); andWeek 52 (return to unrestricted sports). At afollowup of 1 year or greater, the closed kineticchain rehabilitation program produced KT-1000 arthrometer measurements of AP kneelaxity that were closer to normal, and earlierreturn to normal daily activities, comparedwith the open kinetic chain rehabilitation pro-gram. The AP laxity values of the closed ki-netic chain group that were closer to normal,compared with the open kinetic chain group,were attributed to lower strains on the healinggraft. Subsequent study of the anterior cruci-ate ligament in vivo13,25 confirmed this obser-vation by showing that increasing resistanceduring closed kinetic chain activities such assquatting or bicycling does not significantlyincrease anterior cruciate ligament strain val-ues compared with the baseline performanceof the activity, whereas increasing resistanceduring open kinetic chain exercises such asisotonic contraction of the quadriceps pro-duces significant increases of anterior cruciateligament strain values.10

Mikkelsen et al41 did a prospective, ran-domized study that compared closed kineticchain versus combined closed and open ki-netic chain rehabilitation initiated 6 weeksafter anterior cruciate ligament surgery. Six-month followup revealed that the addition ofopen kinetic chain exercises produced a sig-nificant improvement in quadriceps strength(evidenced by moderate improvements in ex-tension torque), an earlier return to sport at thepreinjury level, and did not effect KT-1000arthrometer measurements of AP knee laxity.

More recently, two prospective, random-ized studies were reported that compared re-habilitation with and without the use of abrace.17,42 Rehabilitation with the use of abrace during the first 3 weeks after surgery re-sulted in fewer problems with swelling, alower prevalence of hemarthrosis and woundleakage, and less pain throughout the early re-

Number 402September, 2002 Anterior Cruciate Ligament Rehabilitation 17

covery period in comparison with rehabilita-tion without the use of a brace.

Three randomized controlled trials have beenreported that showed that anterior cruciateligament reconstruction with a bone-patellatendon-bone graft followed by a home-basedrehabilitation program produces a similar out-come compared with reconstruction with thesame graft material and clinic-based rehabili-tation.9,24,50 The most recent of these studiesshowed that subjects randomized to home-based rehabilitation had an average of fivephysical therapy visits (range, 3–7 visits)whereas those treated with clinic-based reha-bilitation had 20 visits (range, 10–28 visits).Details were not presented regarding the spe-cific activities and restrictions that were asso-ciated with both programs; instead a goalbased approach was used with the followingtemporal sequence of activities: restorationof ROM, beginning strengthening, advancedstrengthening, and then improvement of agilityand speed. There was no difference betweenthe treatments regarding physical examination(including ROM, thigh atrophy, Lachman test,anterior drawer and pivot shift examinations,and KT-1000 testing), one-leg hop, and thehealth status questionnaire.

The current review of the literature did notidentify a consensus regarding what variablesshould be used to characterize a rehabilitationprogram, nor did it reveal how this informa-tion should be used to compare different pro-grams, for example, to distinguish betweenan accelerated and conservative rehabilitationprogram. Consequently, for the purpose of thisreview, the authors considered anterior cruci-ate ligament rehabilitation as a series of activ-ities (or restrictions) that a subject (knee) is di-rected to do, the time when the activities arerecommended and the duration of the activi-ties (the number or sets and repetitions perday, week, month), and the point when a sub-ject is advised to return to sport or at risk ac-tivity. Most of the prospective studies thatwere reviewed clearly described the activities(restrictions) a subject was advised to do andthe time that the activities were recommended.

There was, however, little information pre-sented regarding the frequency and duration ofthe activities and how well subjects compliedwith the program that was specified. There-fore, the authors’ interpretation of the litera-ture is that there is some information availablethat derives from prospective, randomizedclinical trials regarding how much loading andmotion a knee with a healing anterior cruciateligament graft can sustain without perma-nently stretching the graft (as evidenced by ab-normal increases of anterior knee laxity), dis-rupting the graft, or creating failure of graftfixation, but that additional research is needed.For example, currently, it has become almostuniversally accepted that rehabilitation of ahealing anterior cruciate ligament graft withrapid return to sports that have the potential togenerate high stresses in the graft and kneedoes not lead to any deleterious effects, yet thecurrent review revealed that relatively fewstudies actually have been published that ad-vocate return to sports between 4 and 6 monthsafter an anterior cruciate ligament reconstruc-tion.32,51,52 Six months after anterior cruciateligament reconstruction patients walk withnormal kinematics, but do so with dramatic al-terations in torque and power about the hip andknee that have the potential to influence thegraft and articular cartilage metabolism.23 Fromthis perspective there is little evidence in theliterature that early return to sport is safe orefficacious. In contrast, the well-designedprospective study of closed versus open ki-netic chain rehabilitation reported by Bynumet al21 recommended initiation of noncutting,pivoting, and jumping activities at 8 months,followed by return to unrestricted sports at 1year. No randomized clinical trial that com-pares accelerated rehabilitation with the moreconservative rehabilitation ever has been pub-lished that indicates that there are no signifi-cant risks to what currently is termed acceler-ated rehabilitation. Although a prospective,randomized clinical study with blinding of thepatient and the physician to the rehabilitationregimen is difficult to do, it is the only way thatthe biases that plague many of the studies that

Clinical Orthopaedics18 Beynnon et al and Related Research

are now in the literature can be avoided. Untilsuch a study or group of studies becomes avail-able, there is reason to be concerned about theassumption that accelerated rehabilitation isappropriate. The authors encourage others touse prospective, randomized, blinded clinicaltrials that can lead to validation of the methodsthat are appropriate and safe. Only until this isaccomplished can individuals speculate as tothe effectiveness of any rehabilitation protocolor program.

References1. Arem AJ, Madden JW: Effects of stress on healing

wounds: I: Intermittent noncyclical tension. J SurgRes 20:93–102, 1976.

2. Akeson WH, Amiel D, Woo SL-Y: Immobility ef-fects on synovial joints: The pathomechanics of jointcontracture. Biorheology 17:95–110, 1980.

3. Amiel D, Kleiner JB, Roux RD, et al: The phenom-enon of “ligamentization”: Anterior cruciate liga-ment reconstruction with autogenous patellar ten-don. J Orthop Res 4:162–172, 1986.

4. Andriacchi T, Sabiston P, DeHaven K, et al: Liga-ment: Injury and Repair. In Woo SL-Y, BuckwalterJA (eds). Injury and Repair of the MusculoskeletalSoft Tissues. Park Ridge, IL, American Academy ofTrans Orthopaedic Surgeons 103–128, 1988.

5. Bair GR: The effect of early mobilization versuscasting on anterior cruciate ligament reconstruction.Trans Orthop Res Soc 5:108, 1980.

6. Ballock RT, Woo SL-Y, Lyon RM, et al: Use ofpatellar tendon autograft for anterior cruciate liga-ment reconstruction in the rabbit: A long term histo-logical and biomechanical study. J Orthop Res7:474–485, 1989.

7. Barber-Westin SD, Noyes FR: The effect of rehabil-itation and return to activity on anterior-posteriorknee displacements after anterior cruciate ligamentreconstruction. Am J Sports Med 21:264–270, 1993.

8. Barber-Westin SD, Noyes FR, Heckmann TP, Shaf-fer BL: The effect of exercise and rehabilitation onanterior-posterior knee displacements after anteriorcruciate ligament autograft reconstruction. Am JSports Med 27:84–92, 1999.

9. Beard DJ, Dodd CA: Home or supervised rehabilita-tion following anterior cruciate ligament reconstruc-tion: A randomized controlled trial. J Orthop SportsPhys Ther 27:134–143, 1998.

10. Beynnon BD, Fleming BC: Anterior cruciate liga-ment strain in-vivo: A review of previous work. JBiomech 31:519–525, 1998.

11. Beynnon BD, Fleming BC, Johnson RJ, et al: Anteriorcruciate ligament strain behavior during rehabilitationexercises in-vivo. Am J Sports Med 23:24–34, 1995.

12. Beynnon BD, Johnson RJ, Fleming BC, et al: The ef-fect of functional knee bracing on the anterior cruci-ate ligament in the weight-bearing and non-weight-bearing knee. Am J Sports Med 25:353–359, 1997.

13. Beynnon BD, Johnson RJ, Fleming BC, et al: The

strain behavior of the anterior cruciate ligament dur-ing squatting and active flexion-extension: A com-parison of an open and closed kinetic chain exercise.Am J Sports Med 25:823–829, 1997.

14. Beynnon BD, Johnson RJ, Tohyama H, et al: The re-lationship between anterior-posterior knee laxity andthe structural properties of the patellar tendon graft:A study in canines. Am J Sports Med 22:812–820,1994.

15. Beynnon BD, Risberg MA, Tjomsland O, et al: Eval-uation of knee joint laxity and the structural proper-ties of the anterior cruciate ligament graft in humans.Am J Sports Med 25:203–206, 1997.

16. Beynnon BD, Wertheimer CM, Johnson RJ, et al:The effect of functional knee braces on anterior cru-ciate ligament strain in-vivo. J Bone Joint Surg74A:1298–1312, 1992.

17. Brandsson S, Faxen E, Kartus J, Eriksson BI, Karls-son J: Is a knee brace advantageous after anteriorcruciate ligament surgery? A prospective, random-ized study with a two-year follow-up. Scand Med SciSports 11:110–114, 2001.

18. Buckwalter JA: Activity vs rest in the treatment ofbone, soft tissue and joint injuries. Iowa Orthop J15:29–42, 1995.

19. Burroughs P, Dahners LE: The effect of enforced ex-ercise on the healing of ligament injuries. Am JSports Med 18:376–378, 1990.

20. Butler DL, Grood ES, Noyes FR, et al: Mechanicalproperties of primate vascularized vs nonvascular-ized patellar tendon grafts: Changes over time. J Or-thop Res 7:68–79, 1989.

21. Bynum EB, Barrack RL, Alexander AH: Open ver-sus closed kinetic chain exercises after anterior cru-ciate ligament reconstruction. Am J Sports Med23:401–406, 1995.

22. Daniel DM, Stone ML, Sachs R, Malcom LL: In-strumented measurement of anterior knee laxity inpatients with acute anterior cruciate ligament disrup-tion. Am J Sports Med 13:401–407, 1985.

23. DeVita P, Hortobagyi T, Barrier J: Gait biomechan-ics are not normal after anterior cruciate ligament re-construction and accelerated rehabilitation. Med SciSports Exerc 30:1481–1488, 1998.

24. Fischer DA, Tewes DP, Boyd JL, Smith JP, QuickDC: Home based rehabilitation for anterior cruciateligament reconstruction. Clin Orthop 347:194–199,1998.

25. Fleming BC, Beynnon BD, Renstrom PA, et al: Thestrain behavior of the anterior cruciate ligament dur-ing bicycling: An in-vivo study. Am J Sports Med26:109–118, 1998.

26. Fleming BC, Beynnon BD, Renstrom PA, et al: Thestrain behavior of the anterior cruciate ligament dur-ing stair climbing: An in-vivo study. J Arthroscopy15:185–191, 1999.

27. Fleming BC, Renstrom PA, Engstrom B, Peura GD,Beynnon BD: The influence of functional knee brac-ing on the ACL strain biomechanics in weightbear-ing and nonweightbearing knees. Am J Sports Med28:815–824, 2000.

28. Fronek J, Frank C, Amiel D, et al: The effects of in-termittent passive motion in the healing of medialcollateral ligament. Trans Orthop Res Soc 8:31, 1983.

Number 402September, 2002 Anterior Cruciate Ligament Rehabilitation 19

29. Gelberman RH, Woo SL-Y, Lothringer K, AkesonWH, Amiel D: Effects of early intermittent passivemobilization on healing canine flexor tendons. JHand Surg 7A:170–175, 1982.

30. Glasgow SG, Gabriel JP, Sapega AA, et al: The ef-fect of early versus late return to vigorous activitieson the outcome of anterior cruciate reconstruction.Am J Sports Med 21:243–248, 1993.

31. Haggmark T, Erickson E: Cylinder or mobile castbrace after knee ligament surgery: A clinical analysisand morphological and enzymatic study of changes inquadriceps muscle. Am J Sports Med 7:48–56, 1979.

32. Howell SM, Taylor MA: Brace-free rehabilitation,with early return to activity, for knees reconstructedwith a double-looped semitendinosus and gracilisgraft. J Bone Joint Surg 78A:814–825, 1996.

33. Hulse DA, Butler DL, Kay MD, et al: Biomechanics ofthe cranial cruciate ligament reconstruction in the dog:In-vitro laxity testing. Vet Surg 12:109–112, 1983.

34. Jorgensen U, Jensen CM, Scavenius M, Bak K: Re-habilitation with or without initial weightbearing: A prospective randomized study. Proceedings ofSports Medicine, Stockholm, Sweden 76, 1995.

35. Jozsa L, Jarvinen M, Kannus P, et al: Fine structuralchanges in the articular cartilage of the rat’s knee fol-lowing short-term immobilization in various posi-tions: A scanning electron microscopical study. IntOrthop 11:129–133, 1987.

36. Jozsa L, Reffy A, Jarvinen M, et al: Cortical and tra-becular osteopenia after immobilization: A quantita-tive histological study in rats. Int Orthop 12:169–172,1988.

37. Jozsa L, Thöring J, Jarvinen M, et al: Quantitative al-terations in intramuscular connective tissue follow-ing immobilization: An experimental study in ratcalf muscle. Exp Mol Pathol 49:267–278, 1988.

38. Kennedy JC: Symposium: Current concepts in themanagement of knee instability. Contemp Orthop5:59–78, 1982.

39. Leppala J, Kannus P, Natri A, et al: The effect of an-terior cruciate ligament injury of the knee on bonemineral density of the spine and affected lower ex-tremity: A prospective one-year follow-up study.Calcif Tissue 64:357–63, 1999.

40. Long M, Frank C, Schachar N, et al: The effects ofmotion on normal and healing ligaments. Trans Or-thop Res Soc 7:43, 1982.

41. Mikkelsen C, Werner S, Eriksson E: Closed kineticchain alone compared to combined open and closedkinetic chain exercises for quadriceps strengtheningafter anterior cruciate ligament reconstruction withrespect to return to sports: A prospective matchedfollow-up study. Knee Surg Sports Traumatol Arthrosc8:337–342, 2000.

42. Moller E, Forssblad M, Hansson L, Wange P, Wei-denhielm L: Bracing versus nonbracing in rehabilita-tion after anterior cruciate ligament reconstruction: A

randomized prospective study with 2-year follow-up.Knee Surg Sports Traumatol Arthrosc 9:102–108,2001.

43. Newton PO, Horibe S, Woo SL-Y: ExperimentalStudies of Anterior Cruciate Ligament Autograftsand Allografts. In Daniel DM, Akeson WH, O’Con-ner JJ (eds). Knee Ligaments: Structure, Function,Injury and Repair. New York, Raven Press 389–399,1990.

44. Nikolaou PK, Seaber AV, Glisson RR, et al: Anteriorcruciate ligament allograft transplantation: Long-termfunction, histology, revascularization, and operativetechnique. Am J Sports Med 14:348–360, 1986.

45. Noyes FR: Functional properties of knee ligamentsand alterations induced by immobilization. Clin Or-thop 123:210–242, 1977.

46. Noyes FR, Mangine RE, Barber S: Early knee mo-tion after open and arthroscopic anterior cruciate lig-ament reconstruction. Am J Sports Med 15:149–160,1987.

47. Richmond JC, Gladstone J, MacGillivray J: Contin-uous passive motion after arthroscopically assistedanterior cruciate ligament reconstruction: Compari-son of short-versus long-term use. J Arthroscopy7:39–44, 1991.

48. Rosen MA, Jackson DW, Atwell EA: The efficacy ofcontinuous passive motion in the rehabilitation ofanterior cruciate ligament reconstruction. Am JSports Med 20:122–127, 1992.

49. Rougraff BT, Shelbourne KD: Early histologic ap-pearance of human patellar tendon autografts usedfor anterior cruciate ligament reconstruction. KneeSurg Sports Traumatol Arthrosc 7:9–14, 1999.

50. Schenck RC, Blaschak MJ, Lance ED, Turturro TC,Holmes CF: A prospective outcome study of reha-bilitation programs and anterior cruciate ligament re-construction. Arthroscopy 13:285–290, 1997.

51. Shelbourne KD, Klootwyk TE, Wilckens JH, et al:Ligament stability two to six years after anterior cru-ciate ligament reconstruction with autogenous patel-lar tendon graft and participation in accelerated re-habilitation program. Am J Sports Med 23:575–579,1995.

52. Shelbourne KD, Nitz P: Accelerated rehabilitation af-ter ACL reconstruction. Am J Sports Med 18:292–299,1990.

53. Tyler TF, McHugh MP, Gleim GW, Nicholas SJ: Theeffect of immediate weightbearing after anterior cruci-ate ligament reconstruction. Clin Orthop 357:141–148,1998.

54. Vailas AC, Tipton CM, Matthes RD, Gart M: Phys-ical activity and its influence on the repair process ofmedial collateral ligaments. Connect Tissue Res9:25–31, 1981.

55. Woo S-LY, Vogrin TM, Abramowitch SD: Healingand repair of ligament injuries in the knee. J AmAcad Orthop Surg 8:364–372, 2000.

Clinical Orthopaedics20 Beynnon et al and Related Research