Arthroscopic Reconstruction of the Posterior Cruciate Ligament by Using a Quadriceps Tendon Autograft: A Minimum 5Year Follow-up

oC

Cc

7

Arthroscopic Reconstruction of the Posterior Cruciate LigamentWith Use of a Quadruple Hamstring Tendon Graft With

3- to 5-Year Follow-up

Yi-Sheng Chan, M.D., Shih-Chieh Yang, M.D., Chung-Hsun Chang, M.D.,Alvin Chao-Yu Chen, M.D., Li-Jen Yuan, M.D., Kuo-Yao Hsu, M.D.,

and Ching-Jen Wang, M.D.

Purpose: This study prospectively evaluated 20 patients treated consecutively to determine patientoutcomes, efficacy, and complication potential of arthroscopically assisted posterior cruciate liga-ment (PCL) reconstruction performed with hamstring tendon grafts. Methods: Twenty patients (15men and 5 women), each with an isolated PCL injury, underwent PCL reconstruction with hamstringtendon autograft and were enrolled in this prospective study. Average age at time of surgery was 29years (range, 20 to 57 years). Average time from injury to surgery was 4 months (range, 3 to 12months). Average follow-up period was 40 months (range, 36 to 50 months). Patients underwentregular follow-up after clinical and radiographic preoperative and postoperative evaluation. Fol-low-up examinations comprised the Lysholm Knee Score, the Tegner Activity Score, the Interna-tional Knee Documentation Committee (IKDC) score, thigh muscle assessment, and radiographicevaluation. Results: Mean preoperative Lysholm score for 20 knees was 63 � 10 (range, 48-73);mean postoperative Lysholm score was 93 � 9 (range, 77-100). Eighteen of 20 patients (90%)showed good or excellent results at final assessment. Mean preinjury and preoperative Tegner scoreswere 7 � 1.5 (range, 5-9) and 3 � 1.9 (range, 2-5), respectively; mean postoperative Tegner scorefor 20 knees was 6.3 � 2.4 (range, 4-9). In final IKDC ratings, 85% of patients (17 of 20) wereassessed as normal or near normal (grade A or B). A statistically significant improvement was seenin thigh girth difference, extensor strength ratio, and flexor strength ratio before and after recon-struction at a minimum of 3 years of follow-up. Conclusions: After follow-up for longer than 36months, analytical results showed satisfactory function after PCL reconstruction with the use ofhamstring tendon autografts. We suggest that the hamstring tendon autograft is a safe, effective, andacceptable choice for PCL reconstruction, and that it affords good ligament reconstruction. Level ofEvidence: IV, therapeutic case series. Key Words: Arthroscopy—Posterior cruciate ligament—Hamstring tendon autograft.

From the Department of Orthopaedic Surgery, Division of Sports Medicine, Chang-Gung Memorial Hospital, Chang Gung University Collegef Medicine (Y-S.C., S-C.Y., C-H.C., A.C-Y.C., L-J.Y., K-Y.H.), Taoyuan; Department of Orthopaedic Surgery, Division of Sports Medicine,hang-Gung Memorial Hospital, Chang Gung University College of Medicine (C-J.W.), Kauohsing, Taiwan.Presented at the 24th Annual Meeting of the Arthroscopy Association of North America, Vancouver, Canada, May 12-15, 2005.Address correspondence and reprint requests to Yi-Sheng Chan, M.D., Department of Orthopaedic Surgery, Division of Sports Medicine,

hang Gung Memorial Hospital, Chang Gung University College of Medicine, 5 Fu-Hsing St., Kweishan, Taoyuan 333, Taiwan. E-mail:[email protected]

© 2006 by the Arthroscopy Association of North America0749-8063/06/2207-5193$32.00/0doi:10.1016/j.arthro.2006.03.020

62 Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 22, No 7 (July), 2006: pp 762-770

Takecltott(ibtbtIicvorfcg

oiq(Ceseoapohiaa

swap2

immiawadnnr

fiiegttoopppdskPcctcaworrtvr(i

S

otncfla

763ARTHROSCOPIC PCL RECONSTRUCTION

he posterior cruciate ligament (PCL) is the stron-gest ligament and acts as the primary restraint

gainst straight posterior translation of the tibia at allnee positions; it is critical in maintaining posterolat-ral knee stability.1,2 Level of pain and instabilityaused by PCL injury varies from no interference withifestyle to severe impairment of daily activities. Pa-ients with minimal PCL injury may achieve goodutcomes with nonoperative treatment. However, pa-ients with grade 3 (tibial plateau displaced posterioro the femoral condyle between 10 and 15 mm) or 4posterior displacement greater than 15 mm) PCLnjuries are at high risk of functional disability causedy recurrent pain, instability, and knee degenera-ion.3,4 In the past, the significance of PCL injury haseen overly simplified and the functional disability ofhe knee with PCL injury has been underestimated.nitially, some studies on the natural history of PCLnjury reported satisfactory functional outcomes withonservative treatment.5,6 Later studies, however, fa-ored surgical treatment because long-term follow-upf patients who were given conservative treatmentevealed a high incidence of osteoarthritis and poorunctional outcomes.7-9 Currently, early surgical re-onstruction of the knee is generally recommended forrade 3 or 4 PCL injury.10-13

Reconstruction failure may be caused by a numberf factors, including selection of a graft substitute withnadequate strength. Both autografts (patellar tendon,uadriceps tendon, hamstring tendon) and allograftsAchilles tendon, patellar tendon) are frequently used.ontroversy is ongoing about which graft is mostffective. On the bases of availability of tissue andurgical simplicity, hamstring tendon autografts weremployed in this series for arthroscopic reconstructionf the isolated PCL. This study prospectively evalu-ted 20 patients treated consecutively to determineatient outcomes, efficacy, and complication potentialf arthroscopically assisted PCL reconstruction withamstring tendon grafts. The hypothesis of the studys that hamstring tendon autograft is a safe, effective,nd acceptable choice for PCL reconstruction, and itffords good ligament reconstruction.

METHODS

From August 1999 through August 2001, 20 con-ecutive patients (15 men and 5 women) who under-ent PCL reconstruction with a hamstring tendon

utograft at the authors’ hospital were enrolled in therospective study. Average age at time of surgery was

njury to surgery was 4 months (range, 3 to 12onths). The average follow-up period was 40onths (range, 36 to 50 months). Mechanisms of

njury involved participation in sports (3 of 20, 15%)nd motorcycle accidents (17 of 20, 85%). Patientsith a history of any of the following were excluded:

ssociated ligament injury, avulsion fracture, chondralamage, previous meniscectomy (excision �1 of 3 me-iscus), malposition of tibia and femoral tunnels, ab-ormal preoperative radiographs, previous failed PCLeconstruction, or abnormal contralateral knee joints.

In this study, all PCL rupture diagnoses were con-rmed by physical examination, magnetic resonance

maging, and arthroscopic examination. The physicalxamination comprised assessment for posterior sag-ing sign, posterior drawer test, reverse pivot-shiftest, varus angulation test, and posterolateral drawerest. A posterior drawer examination was performedn all knees in neutral rotation for grading of 1, 2, 3,r 4. Knees were graded as follows: less than 5 mm ofosterior tibial displacement, grade 1; 6 to 10 mm ofosterior tibial displacement, grade 2; 11 to 15 mm ofosterior tibial displacement, grade 3; and posteriorisplacement greater than 15 mm, grade 4.3,4 Theurgical indication was functional disability of thenee due to pain and instability caused by high-energyCL injury and lack of satisfaction with 3 months ofonservative treatment. Conservative treatment in-luded a vigorous quadriceps strengthening programhat was begun as the patient’s symptoms allowed,losed-chain exercises such as squats and leg press,nd hamstring strengthening that was initiated 6eeks after injury. All patients were graded as 3 or 4n the posterior drawer test and had a positive poste-ior sag test. Knees were examined for posterolateralotatory instability. No subject had posterolateral ro-atory instability or a positive extension varus recur-atum test. Twenty arthroscopic procedures for PCLeconstruction were performed by the same surgeonY-S.C.), who employed the same surgical techniquen all cases.

urgical Technique

The patient is placed in the supine position on theperating table. The surgical leg hangs over the side ofhe operating table and is fixed by the leg holder. Theonsurgical leg is supported by the leg post. Surgery isonducted under tourniquet control. An arthroscopicuid pump is used for fluid inflow. After adequatenesthetization has been achieved, a complete physi-

fat

A

uct

G

lthrsonaitEatsdttt

ttEwpp5aapt

T

t8wMw

lP

pprfiatmtiftosfbotamlsmoig

G

tftBttwmeatbtgs9a(s

764 Y-S. CHAN ET AL.

ormed to determine the condition of all relevant an-tomic structures and to identify the extent of ligamentear and associated injuries.

rthroscopic Portals

Standard anterolateral and anteromedial portals aresed for arthroscopy, and a posteromedial portal isreated under direct arthroscopic vision for PCL tibialunnel preparation.

raft Harvest and Preparation

A short vertical incision approximately 2 cm inength is made from a point 1.5 cm medial to the tibialubercle and is extended distally so the graft can bearvested. The sartorial fascia is palpated from supe-ior to inferior for identification of the underlyingemitendinosus and gracilis tendons; it is then incisedver the interval between the gracilis and semitendi-osus tendons and the pes anserinus and both tendonsre exposed. The semitendinosus is subsequentlydentified and a Penrose drain tube is inserted aroundhe tendon; a running baseball whipstitch with No. 2thibond sutures (Ethicon, Somerville, NJ) is placedlong 1.5 cm of the tendon. The tendon is then de-ached from its insertion point. While tension on theutures is maintained, blunt dissection around the ten-on is performed. The tendon is then harvested by aendon stripper with the patient’s knee flexed andraction applied with the grasping suture. The gracilisendon is harvested through a similar technique.

Harvested tendons are prepared by a surgical assis-ant at the side table. Each tendon is made into aendon graft that is approximately 24 cm in length.ach tendon end is sutured in a running baseballhipstitch mode with No. 2 Ethibond. The graft com-rises a double strand of each tendon, thus a quadru-le-strand graft that is 12 cm in length is the result. A-mm width of Mersilene tape (Ethicon) is passedround the looped section of the graft at its diameternd is sutured together with 3-0 Vicryl (Ethicon) at aoint 3 cm from the looped end and the end with freeendons.

unnel Preparation

After thorough and careful arthroscopic examina-ion of the knee, the knee joint is flexed to roughly0°. Arthroscopic PCL reconstruction is performedith the use of Acufex PCL guides (Acufex, Andover,A). The ruptured PCL is removed from the medial

eft attached to the medial wall to mark the originalCL footprint.The goal of single-bundle reconstruction is to re-

roduce the anterolateral component of a PCL. Theosterior tibial tunnel opening should exit the poste-ior tibia within the PCL footprint near the distolateralbers. The transtibial PCL tunnel is drilled from thenteromedial aspect of the proximal tibia 1 cm belowhe tibial tubercle and exits posteriorly at approxi-ately 1 cm below the articular surface of the medial

ibial plateau slightly lateral to the midline at the PCLnsertion site. The tip of the guide pin must be care-ully protected with a curved curette to prevent injuryo the neurovascular structures when it exits posteri-rly. On the basis of measurements provided by graftizers, an 8- to 10-mm-diameter tunnel is constructedor the graft. A femoral tunnel for the anterolateralundle of the PCL reconstruction is created with anutside-in, 2-incision method. Briefly, the femoralunnel opening should start in the femoral epicondylet a minimum of 1 cm from the articular cartilageargin and approximately midway between the troch-

ea and the epicondyle. The entry point in the jointhould extend 10 mm from the articular cartilageargin and should encompass the anterolateral fibers

f the footprint.10,11 An 8- or 10-mm-diameter tunnels constructed for the quadruple hamstring tendonraft.

raft Passage and Fixation

The Mersilene tape–looped end of the hamstringendon graft is connected by passing No. 5 Ethibondrom the tibial tunnel through the joint to the femoralunnel. On the femoral side, an 8 or 9 � 25 mmioScrew (Linvatec, Largo, FL) is fixed at the inner

unnel aperture of the femoral tunnel. The Mersileneape–looped end of the hamstring tendon graft is tiedith a washer for extra fixation. After the graft has beenanually tensioned and at least 10 cyclic loads have

xhibited full range of movement, the knee is flexed at70° angle, and an anterior drawer force is applied to

he proximal tibial aspect. Normal step-off is palpableetween the medial femoral condyle and the medialibial plateau after reduction. The tendon end of theraft is fixed to the anteromedial tibia by tying theutures to a biocortical screw and washer. An 8 or� 25 mm BioScrew (Linvatec) is fixed to the inner

pertures of the tibia tunnel for additional fixationFig 1). Knee stability is then retested. The graft isubsequently examined arthroscopically to ensure

ivik

P

roPmcsbpbpsmt0tw

daaeppmishuwoml

F

rtmtpdkKliwtmcldtteAtdpmatPdfwrM

Fql

765ARTHROSCOPIC PCL RECONSTRUCTION

ntra-articular impingement. A cold compression de-ice, a Cryo/Cuff (Aircast, Summit, NJ), is appliedmmediately after surgery. The knee is placed in anee brace and is locked in a fully extended position.

ostoperative Rehabilitation

All patients with isolated PCL reconstruction wereehabilitated according to the standard protocol. Theriginal goals of postoperative management of suchCL reconstruction were (1) to decrease pain, inflam-ation, and swelling, (2) to re-establish quadriceps

ontrol, and (3) to restore normal gait. After such aurgical procedure, a fully extended functional kneerace was employed at the operative limb for the firstostoperative week. Ambulation with partial weightearing on the surgical leg was initiated on the secondostoperative day. Quadriceps isometric exercise,traight leg raising exercises, and passive range ofotion should be initiated as early as possible. Pro-

ected range of motion was gradually increased from° to 90° during the first postoperative month, andhen from 90° to 120° during the second month. At 4

IGURE 1. Arthroscopic technique in PCL reconstruction withuadruple hamstring tendon autograft. (A) Anteroposterior and (B)ateral views.

eeks, a series of kinetic closed-chain exercises were s

emonstrated for patients, who were strongly encour-ged to cogently participate. At 6 weeks, patients werellowed full weight bearing with hamstring strength-ning exercises. The brace was unlocked, and eachatient was advised to establish a normal gait and aassive range of motion. At 8 weeks, active range ofotion in the knee should progress to complete flex-

on and extension. Quadriceps and hamstring muscletrength training was especially emphasized through aome rehabilitation program. At 3 months, patientssually returned to their normal daily activity andere allowed to exercise on a stationary bike or standn a single leg. Light sports activities began at 6onths. After 9 months, full activity, including ath-

etic activity, was permitted.

ollow-up Evaluation

Each patient underwent follow-up for clinical andadiographic evaluation at regular periods. Each pa-ient was evaluated preoperatively, at 3 months, 6onths, 9 months, and 1 year postoperatively, and

hen annually thereafter. Follow-up examination em-loyed the following scoring systems: Lysholm,14 toocument subjective symptoms; Tegner activity score15;nee ligament standard evaluation of the Internationalnee Documentation Committee (IKDC)16; and single-

eg hop test. Preinjury Tegner activity scores and activ-ty level by IKDC guidelines were calculated fromhat patients reported their activities to be before the

ime of PCL injury. At each follow-up visit, range ofotion of the involved knee in relation to that of the

ontralateral, normal knee was measured with use of aong-arm goniometer. Extension deficit, which wasetermined with the subject lying in the prone posi-ion, was measured as the difference in heel height ofhe involved limb compared with the passively fullyxtended posture of the contralateral, uninjured limb.

manual posterior drawer test and a KT-1000 Ar-hrometer (MEDmetric, San Diego, CA) were used toetermine knee stability; according to the methodsresented by Daniel et al.,17 side-to-side differences ataximal manual test were used preoperatively and

nnually postoperatively to evaluate anteroposteriorranslation and to discern degree of ligament laxity.atients self-assessed anterior knee pain to measureonor site morbidity. A visual analogue scale (rangingrom 0 to 10, with 0 � no pain and 10 � severe pain)as used to measure the intensity of knee pain.18 Knee

adiographs in standing anteroposterior, lateral, anderchant’s views were examined for alignment, joint

pace narrowing, and degenerative changes in the

kttccAatpbetisaotac

S

dgc(alrscAncS

L

aL4�focactw

T

2(sps

I

tgoBelaatmaosutp

At

L

T

t

766 Y-S. CHAN ET AL.

nee, as well as for bone tunnel enlargement. Boneunnel enlargement was defined as the percentage ofunnel width on postoperative radiographs that ex-eeded width at follow-up. Severity of degenerativehanges of the knee was graded according to thehlback classification.19 Thigh atrophy was defined assmaller thigh circumference in the involved knee

han in the contralateral knee at a position 10 cmroximal to the superior pole of the patella. The Cy-ex 340 Dynamometer (Cybex, New York, NY) wasmployed preoperatively and annually postoperativelyo identify postoperative residual thigh muscle deficitsn reconstructed and contralateral knees. Peak exten-ion and flexion torques were isokinetically measuredt 180°/sec. The side-to-side ratio (peak muscle torquef the involved side/peak muscle torque of the con-ralateral side � 100) in peak muscle torque wasdopted as the representative parameter for thigh mus-le strength.

tatistics

Statistical analysis was conducted by an indepen-ent statistician who was not associated with the sur-ical team. The �2 test was used for comparisons ofategorical IKDC data with A and B versus C and Dnormal or nearly normal versus abnormal or severelybnormal) from preoperative assessment to final fol-ow-up. The Mann-Whitney U test was applied foranked continuous data (thigh atrophy and Lysholmcores), and the unpaired Student t test was used forontinuous data (KT-1000 Arthrometer comparison).

value of P � .05 was considered statistically sig-ificant. Statistical analysis was performed with theomputer program SigmaStat, version 2.0 (Systatoftware, Port Richmond, CA).

RESULTS

ysholm Knee Scores

The Lysholm knee scoring system was used tonalyze subjective symptoms. The mean preoperativeysholm score for 20 knees was 63 � 10 (range,8-73); the mean postoperative Lysholm score was 93

9 (range, 77-100). After a minimum of 3 years ofollow-up, 6 of 20 patients (30%) achieved excellentutcomes, and 12 patients (60%) achieved good out-omes. Of the remaining patients, 2 (70 and 72 points)chieved fair outcomes. No patient had a poor out-ome. A significant difference in Lysholm scores be-ween preoperative and final follow-up evaluations

egner Activity Level

Mean preinjury and preoperative Tegner scores for0 knees were 7 � 1.5 (range, 5-9) and 3 � 1.9range, 2-5), respectively. Mean postoperative Tegnercore was 6.3 � 2.4 (range, 4-9). Improvement fromreoperative to postoperative values was statisticallyignificant (P � .037) (Table 1).

KDC EvaluationActivity: The IKDC assessment evaluates symp-

oms and signs. Each category receives an overallrade of A (normal), B (nearly normal), C (abnormal),r D (severely abnormal). The final assignment of A,, C, or D is made on the basis of the worst score inach category. IKDC categories were applied to ana-yze patient activity levels (preinjury, preoperative,nd at final examination). Preinjury activity levels forll patients ranged from strenuous to moderate. At theime of surgery, only 6 (30%) of 20 patients hadoderate activity levels, 13 (65%) of 20 had light

ctivity levels, and 1 patient reported a sedentary levelf activity (5%). Of 20 patients, 18 (90%) achievedtrenuous to moderate activity levels at final follow-p. Two (10%) of 20 patients performed light activi-ies. Evaluation revealed significant improvement inatient activity levels after surgery (P � .05).Knee Function According to Patient Subjectivessessment: Seventeen of 20 patients (85%) subjec-

ively rated their knee function as normal or nearly

TABLE 1. Comparison of Preoperative LysholmKnee Scores and Tegner Activity Level With Final

Follow-up Scores

PreoperativeFinal

Follow-upP

Value*No. % No. %

ysholm knee scoreExcellent (95–100) 0 0 6 30Good (84–94) 0 0 12 60Fair (65–83) 6 30 2 10Poor (�65) 14 70 0 0Mean � SD 63 � 10 93 � 9 .005Range 48–73 77–100

egner activity level0-3 15 75 0 04-6 5 25 12 607-10 0 0 8 40Mean � SD 3 � 1.9 6.3 � 2.4 .037Range 2–5 4–9

*Comparison between preoperative and final follow-up evalua-ion scores by Mann-Whitney U test.

ormal when compared with preinjury status.

n

(ph13nse(rai

deAwutmtp2pe

ua(o

pnuaytra

tpubi(mft

ahttTt

pmp(aIm

P

K

o

ABCD

at

767ARTHROSCOPIC PCL RECONSTRUCTION

Symptom: Of 20 patients, 95% (n�19) reportedo pain during moderate or strenuous activities.Range of Motion: Before surgery, 3 patients

15%) had a flexion defect of greater than 15° com-ared with the contralateral side. Two patients (10%)ad an extension defect of greater than 10°. At review,7 (85%) patients with a difference at full extension of° or less, or at full flexion of 5° or less, betweenormal and reconstructed limbs were rated as normaltatus. One patient (5%) with 3° to 5° difference inxtension was rated as nearly normal status. Two10%) patients with 16° to 25° deficit in flexion wereated as abnormal status. No patient had a severelybnormal rating (an extension deficit �10° or a flex-on deficit �25°).

Ligament Examination: KT-1000 Arthrometerata at 89 N were available for all 20 patients. Preop-ratively, 100% (20 of 20) scored more than 6 mm.verage preoperative posterior displacement assessedith the KT-1000 was 12 � 3.4 mm. At final follow-p, KT-1000 examination revealed a 0- to 2-mm an-erior-posterior translation in 10 (50%) patients, 3 to 5m of posterior laxity in 7 (35%) patients, and more

han 5 mm translation in 3 (15%) patients. Averageosterior displacement at final follow-up was 3.8 �.5 mm. A statistically significant improvement inostoperative KT-1000 scores was seen versus preop-rative data (P � .007) (Table 2).

Patellofemoral Joint Findings: At final follow-p, moderate patellofemoral crepitation on extensiongainst slight resistance was recorded in 2 patients10%), and patellofemoral crepitus with mild painccurred in 2 patients (10%).

TABLE 2. Comparison of Posterior Displacement byKT-1000 Arthrometer Findings (side-to-side difference inmillimeters) at 89 N of Force Between Preoperative and

Final Follow-up Conditions

Difference PreoperativeFinal

Follow-up

osterior Drawer TestGrade I (0–5 mm) 0 16Grade II (6–10 mm) 0 3Grade III (11–15 mm) 15 1Grade IV (�15 mm) 5 0

T-1000 Test for IKDC RatingNormal (0–2 mm) 0 10Nearly normal (3–5 mm) 0 7Abnormal (6–10 mm) 8 3Severely abnormal (�10 mm) 12 0Mean � SD* 12 � 3.4 3.8 � 2.5

(*P � .007 (unpaired Student t test) when compared with pre-

perative data.

Harvest Site Disease: Self-assessed anterior kneeain, used to determine donor-site morbidity, wasoted in 3 (15%) of 20 patients at first year follow-p. The average visual analogue scale score fornterior knee pain was 2.3 (range, 2-3). At secondear follow-up, 1 (5%) of 20 patients reported an-erior knee pain. The visual analogue scale wasated at 1 point. However, no patient describednterior knee pain at final follow-up.Radiographic Assessment: Eighteen (90%) pa-

ients had no radiographic deterioration and 2 (10%)atients showed stage I degeneration at final follow-p, according to the Ahlback classification.19 Averageone tunnel enlargement at final follow-up was notedn 8 patients (40%) for femoral tunnel and 10 patients50%) for tibial tunnel. Average bone tunnel enlarge-ent was 0.9 � 1.4 mm (10.2% � 14.5%) for the

emoral tunnel and 1.4 � 2.1 mm (16.2% � 20%) forhe tibial tunnel.

Functional Test: Functional 1-leg hop test resultst final follow-up showed that 11 (55%) patientsopped at least 90% of the total distance hopped onheir healthy leg. Seven (35%) patients achieved 76%o 89% of the distance hopped with their healthy limb.wo (10%) patients attained 50% to 75% of the dis-

ance achieved with their healthy limb.Overall Rating: Preoperatively, 20 of 20 (100%)

atients were assessed as abnormal or severely abnor-al (grade C or D). At final follow-up, 17 (85%)

atients were assessed as normal or nearly normalgrade A or B) (Table 3). Three patients (15%) weressigned an abnormal or severely abnormal rating.KDC grade evaluation showed significant improve-ent in the grades of patients after surgery (P � .027)

TABLE 3. Numbers of Patients With Preinjury,Preoperative, and Final Follow-up IKDC Grades and

Overall Percentages

Rating

Preinjury PreoperativeFinal

Follow-up*

No. % No. % No. %

(normal) 10 50 0 0 5 25(nearly normal) 9 45 0 0 12 60(abnormal) 1 5 6 30 2 10(severely abnormal) 0 0 14 70 1 5

IKDC, International Knee Documentation Committee.*P � .027 (�2 test) for final rating in the normal–nearly normal

gainst the abnormal–severely abnormal rating (comparison be-ween preoperative and final follow-up data).

Table 3).

T

mmabpsek2krksnfeby

C

ctstcsmp

wewttpoira(pwtrflfi

aaHnmbf(rst

T

E

F

Mann-

768 Y-S. CHAN ET AL.

high Atrophy and Muscle Strength

Table 4 shows a comparison of preoperative thighuscle parameters with final follow-up measure-ents. At final follow-up, 19 (95%) patients exhibiteddifference of less than 10 mm in thigh circumferenceetween their reconstructed and normal limbs. Only 1atient had a difference greater than 10 mm. A Cybextudy revealed that 18 (90%) patients achieved recov-ry of extensor muscle strength in the reconstructednee that was 90% or greater of normal knee strength;(10%) patients recovered 80% to 90% of normal

nee strength. Moreover, 17 (85%) patients achievedecovery of flexor muscle strength in the reconstructednee that was 90% or greater of normal kneetrength, and 2 (10%) recovered to 80% to 90% oformal knee strength. Statistically significant dif-erences were detected in thigh girth difference,xtensor strength ratio, and flexor strength ratioefore and after reconstruction at a minimum of 3ears of follow-up (Table 4).

omplications

No complications directly associated with arthros-opy occurred in any patient. No deep infection,hrombophlebitis, or vascular injury was noted in thiseries. One patient (5%) developed stitch abscess inhe tibial site 1 month after surgery, which was suc-essfully treated by local debridement. Because ofcrew head protrusion, 1 patient (5%) underwent re-oval of hardware at the site of tibial fixation 1 year

TABLE 4. Thigh Atrophy and Muscle Strength DifferenceKnee and t

Thigh MuscleParameter

Patient Number,Preoperative/Postoperative

high Girth Difference�10 mm 5/1910–20 mm 7/1�20 mm 8/0

xtensor Strength Ratio�90% 4/1880%–90% 5/2�80% 11/0

lexor Strength Ratio�90% 5/1780%–90% 6/2�80% 9/1

*Comparison between preoperative and final follow-up data by

ostoperatively. b

DISCUSSION

Most PCL ruptures may be managed successfullyith supervised conservative treatment.5-7 Shelbourne

t al.6 studied 133 patients with isolated PCL injuryho were treated through an unsupervised rehabilita-

ion program and found that 42% consistently ratedhe knee as good or excellent. Shino et al.7 studied 15atients with PCL injury and reported that 53% had anverall IKDC of A or B, and 73% were participatingn moderate to strenuous activities. Arthroscopic PCLeconstruction has recently increased in frequency,nd satisfactory results with an IKDC of A or Brange, 68% to 82%) have been documented for mostatients in whom surgical principles and techniquesere adequate.10-13 The following controversies con-

inue: choice of graft tissue, use of 1- or 2-bundleeconstruction, location of tunnel placements, kneeexion angle when a graft is secured, and methods ofxation.The Achilles tendon allograft is large in diameter,

nd its bone block on a single end enables potentiallydequate interference screw fixation in the tunnel.owever, allograft tissues are not readily available inumerous countries, and concern remains about trans-ission of unknown diseases.20,21 Patellar tendon–

one autograft is considered the preferred procedureor reconstruction of the anterior cruciate ligamentACL)20,22 and is widely used as a graft for PCLeconstruction.23,24 This graft technique has severalhortcomings. The patellar tendon autograft may beoo weak to adequately substitute for the PCL and can

peratively and at Final Follow-up Between the Operativelthy Knee

Preoperativeeans � SD, mm

Final Follow-upMeans � SD, mm P Value*

16.35 � 7.54 7.21 � 3.25 �.001

76.25 � 11.26 90.26 � 9.65 .005

80.65 � 9.54 93.68 � 8.23 .027

Whitney U test.

Preohe Hea

M

e difficult to pass through the tunnel when the tran-

sdtftaadb

psccTimtma

ctcslcvepwsmuebif

lta3isl1btt

r

aalarat

acpdstsc

rns9RaceIfwdhip

aaf(tpFlowc

iftha

769ARTHROSCOPIC PCL RECONSTRUCTION

tibial tunnel procedure is used. The quadriceps ten-on autograft is thicker and wider than the patellarendon, thereby providing an ample source of tendonsor ligament reconstruction purposes. The quadricepsendon–bone construction may result in a versatilelternative graft for use in primary and revision ACLnd PCL reconstruction.11,25,26 However, a high inci-ence of donor-site pain and graft-site morbidity haseen reported when both tendon grafts are used.27-29

Hamstring tendon autograft has recently gainedopularity. The quadrupled graft of double loops ofemitendinosus and gracilis tendons has greater me-hanical strength than the bone–patellar tendon–boneomplex commonly used in ACL reconstruction.30,31

he hamstring tendon graft method provides greaternitial fixation strength than the patellar tendon graftethod,32 allowing for accelerated rehabilitation pro-

ocols. Moreover, because a hamstring tendon graft isultistranded and has a larger surface area, it can be

dvantageous in promoting revascularization.33

Considering the availability of tissue and the surgi-al simplicity of the technique, hamstring tendon au-ografts were used in this series for arthroscopic re-onstruction of the isolated PCL. This study identifiedignificant improvement in knee function, activityevel, IKDC classification, Lysholm scores, and mus-le strength. No morbidity was associated with har-esting of the hamstring tendon. However, the IKDCvaluation system showed that only 25% (n�5) ofatients were classified as normal and 60% (n�12)ere rated nearly normal at final follow-up. Many

tudies have also reported a lower percentage of nor-al or nearly normal results when the IKDC system is

sed than with other rating systems. It seems, how-ver, that most patients lowered their activity leveletween reconstruction and follow-up, likely reflect-ng the realization that complete restoration of kneeunction after reconstructive surgery is extremely rare.

In all, 17 patients (85%) revealed ligament laxity ofess than 5 mm as measured by KT-1000 Arthrometerests. Average posterior displacement preoperativelynd postoperatively was measured at 12 � 3.4 and.8 � 2.5 mm, respectively. Significant improvementn laxity may be achieved with this technique. In thiseries, 20 patients showed a trend toward increasedigament laxity after 3 years as measured by the KT-000 Arthrometer; however, no correlation was seenetween laxity values and outcome scores. It seemshat a certain degree of anteroposterior laxity can beolerated as long as knees do not give way.

In this series, 3 (15%) patients had problems with

3° to 5° difference in extension, and 2 (10%) exhibited16° to 25° deficit in flexion compared with a normal

imb. Postoperative limitation in range of motion may beproblem when hamstring tendon graft is used for PCL

econstruction. More aggressive training for knee flexionnd squatting ability should be strongly emphasized sohat this problem can be avoided.

Numerous studies have reported that PCL injury isssociated with an increased incidence of degenerativehanges in the knee, primarily involving the medial,atellofemoral, and lateral compartments in that or-er.3,7,8 This rate increased with duration of injury,everity of ligament laxity, and length of follow-upime.34 In this study, only 2 (10%) patients showedtage I degenerative change according to the Ahlbacklassification at the last postoperative visit.

At final follow-up, extensor muscle strength in theeconstructed knee recovered to at least 90% that of theormal knee in 18 (90%) patients. Flexor muscletrength in the reconstructed knee recovered to at least0% that of the normal knee in 17 (85%) patients.ecovery of extensor strength and flexor strength asssessed preoperatively and postoperatively was statisti-ally significant. Thigh muscle training may be moreffective when hamstring tendon reconstruction is used.sometric hamstring muscle strength after graft harvestrom the uninjured limb may achieve complete recoveryithin 3 months.35 As compared with the patellar ten-on–bone graft or quadriceps tendon graft technique, theamstring tendon graft approach avoids an additionalncision and extensor mechanism problems, with whichatellar pain has been a major discomfort.11,27

This prospective study is unique for 2 reasons: (1)ll patients had an isolated PCL injury; those withssociated injuries were excluded to control for con-ounding factors that may have affected outcome; and2) the arthroscopic surgical technique was identical inerms of autograft type, performing surgeon, graftlacement, graft fixation, and rehabilitation program.urthermore, all 20 patients received complete fol-

ow-up and outcome assessment. However, limitationsf this work must be acknowledged. No control groupas studied. Principal disadvantages were the small

ase number and the limited observation period.After follow-up for longer than 36 months, analyt-

cal results showed that patients achieved satisfactoryunction after PCL reconstruction was performed withhe hamstring tendon autograft. We suggest that theamstring tendon autograft is a safe, effective, andcceptable choice for PCL reconstruction with good

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

770 Y-S. CHAN ET AL.

REFERENCES

1. Harner CD, Livesay GA, Kashiwaguchi S, Fujie H, Choi NY,Woo SL. Comparative study of the size and shape of humananterior and posterior cruciate ligaments. J Orthop Res 1995;13:429-434.

2. Gollehon DL, Torzilli PA, Warren RF. The role of the pos-terolateral and cruciate ligaments in the stability of the humanknee. A biomechanical study. J Bone Joint Surg Am 1987;69:233-242.

3. Keller PM, Shelbourne KD, McCarroll JR, Rettig AC. Non-operatively treated isolated posterior cruciate ligament inju-ries. Am J Sports Med 1993;21:132-136.

4. Rubinstein RA Jr, Shelbourne KD, McCarroll JR, et al. Theaccuracy of the clinical examination in the setting of pos-terior cruciate ligament injury. Am J Sports Med 1994;22:550-557.

5. Fowler PJ, Messieh SS. Isolated posterior cruciate ligamentinjuries in athletes. Am J Sports Med 1987;9:107-113.

6. Shelbourne KD, Davis TJ, Patel DV. The natural history ofacute, isolated, non operatively treated posterior cruciate lig-ament injuries. A prospective study. Am J Sports Med 1999;27:276-283.

7. Shino K, Horibe S, Nakata K, et al. Conservative treatment ofisolated injuries to the posterior cruciate ligament in athletes.J Bone Joint Surg Br 1995;77:895-900.

8. Dandy DJ, Pusey RJ. The long-term results of unrepaired tearsof the posterior cruciate ligament. J Bone Joint Surg Br 1982;64:92-94.

9. Cross MJ, Powell JF. Long-term follow-up of posterior cruci-ate ligament rupture: A study of 116 cases. Am J Sports Med1984;12:292-297.

0. Kim SJ, Kim HK, Kim HJ. Arthroscopic posterior cruciateligament reconstruction using a one-incision technique. ClinOrthop 1999;359:156-166.

1. Chen CH, Chen WJ, Shih CH. Arthroscopic reconstruction ofthe posterior cruciate ligament: A comparison of quadricepstendon autograft and quadruple hamstring tendon graft. Ar-throscopy 2002;18:603-612.

2. Pinczewski LA, Thuresson P, Otto D, Nyquist F. Arthroscopicposterior cruciate ligament reconstruction using four-strandhamstring tendon graft and interference screws. Arthroscopy1997;13:661-665.

3. Deehan DJ, Salmon LJ, Russell VJ, Pinczewski LA. Endo-scopic single-bundle cruciate ligament reconstruction: Resultsat minimum 2-year follow-up. Arthroscopy 2003;19:955-962.

4. Lysholm J, Gillquist J. Evaluation of knee ligament surgeryresults with a special emphasis on use of a scoring scale. AmJ Sports Med 1982;10:150-154.

5. Tegner Y, Lysholm J. Rating system in the evaluation of kneeligament injuries. Clin Orthop 1985;198:43-49.

6. Hefti F, Muller W, Jakob RP, Staubli HU. Evaluation of kneeligament lesions with the IKDC form. Knee Surg Sports Trau-matol Arthrosc 1993;1:226-234.

7. Daniel DM, Akeson W, O’Conner J, eds. Knee ligaments—structure, function, injury, and repair. New York: Raven,

8. Flandry F, Hunt JP, Terry GC, Hughston JC. Analysis ofsubjective knee complaints using visual analog scales. AmJ Sports Med 1991;19:112-118.

9. Ahlback S. Osteoarthritis of the knee. A radiographic investi-gation. Acta Radiol 1968;277:7-72.

0. Shelton WR, Papendick L, Dukes AD. Autograft versus allo-graft anterior ligament reconstruction. Arthroscopy 1997;13:446-449.

1. Buck BE, Malinin TI, Brown MD. Bone transplantation andhuman immunodeficiency virus: An estimated risk of acquiredimmunodeficiency syndrome (AIDS). Clin Orthop 1989;240:129-136.

2. Peterson RK, Shelton WR. Allograft versus autograft patellartendon anterior cruciate ligament reconstruction: A 5-yearfollow-up. Arthroscopy 2001;17:9-13.

3. Kim SJ, Min BH. Arthroscopic intraarticular interferencescrew technique of posterior cruciate ligament reconstruction:One-incision technique. Arthroscopy 1994;10:319-323.

4. Clancy WG Jr, Pandya RD. Posterior cruciate ligament recon-struction with patellar tendon autograft. Clin Sports Med 1994;13;561-570.

5. Stäubli HU. The quadriceps tendon–patellar bone construct forACL reconstruction. Sports Med Arthrosc Rev 1997;5:59-67.

6. Fulkerson IP, Langeland R. An alternative cruciate reconstruc-tion graft: The central quadriceps tendon. Arthroscopy 1995;11:252-254.

7. Aglietti P, Buzzi R, Zaccherotti G, DeBiase P. Patellar tendonversus double semitendinosus and gracilis tendons for anteriorcruciate ligament reconstruction. Am J Sports Med 1994;22:211-218.

8. Christen B, Jakob R. Fractures associated with patellar liga-ment grafts in cruciate ligament surgery. J Bone Joint Surg Br1992;74:617-619.

9. DeLee JC, Craviotto DF. Rupture of the quadriceps tendonafter a central third patellar tendon anterior cruciate ligamentreconstruction. Am J Sports Med 1991;19:415-416.

0. Hamner DL, Brown CH Jr, Steiner ME, Hecker AT, HayesWC. Hamstring tendon grafts for reconstruction of the anteriorcruciate ligament: Biomechanical evaluation of the use ofmultiple strands and tensioning techniques. J Bone Joint SurgAm 1999;81:549-557.

1. Lipscomb AB, Johnston RK, Snyder RB, et al. Evaluation ofhamstring strength following use of semitendinosus and gra-cilis tendons to reconstruct the anterior cruciate ligament. AmJ Sports Med 1982;10:340-342.

2. Rowden NJ, Sher D, Rogers GJ, Schindhelm K. Anteriorcruciate ligament graft fixation. Initial comparison of patellartendon and semitendinosus autografts in young fresh cadavers.Am J Sports Med 1997;25:472-478.

3. Pinczewski LA, Clingeleffer AJ, Otto DD, Bonar SF, Corry IS.Integration of hamstring tendon graft with bone in reconstruc-tion of the anterior cruciate ligament. Arthroscopy 1997;13:641-643.

4. Wang CJ, Chen HS, Huang TW. Outcome of arthroscopicsingle bundle reconstruction for complete posterior cruciateligament tear. Injury 2003;34:747-751.

5. Yasuda K, Tsujino J, Ohkoshi Y. Graft site morbidity with