ORIGINAL ARTICLE

Patrick Hunt Æ Sven U. Scheffler

Frank N. Unterhauser Æ Andreas Weiler

A model of soft-tissue graft anterior cruciate ligament reconstructionin sheep

Received: 10 July 2003 / Published online: 16 March 2004� Springer-Verlag 2004

Abstract Introduction Since there is to our knowledgeno clinically valid and reproducible animal model ofsoft-tissue anterior cruciate ligament (ACL) recon-struction currently available, we developed one in sheep,in terms of graft suitability, postsurgical recovery, andknee stability. Materials and methods To find a suitablegraft, anatomical dissections of the hind limbs of 7 sheepwere performed. After a pilot study in 3 sheep, wereconstructed the ACL with an ipsilateral, longitudinallysplit, superficial digital flexor tendon autograft andanatomic graft fixation in 42 sheep (study 1) and with afull, superficial digital flexor tendon autograft and ex-tracortical graft fixation in 48 sheep (study 2). Follow-upexaminations ranged from 6 to 104 weeks (study 1) and3 to 24 weeks (study 2). Results All animals toleratedthe graft harvest well and returned to physiologicalmovement after about 4 weeks. Only 1 out of 93 ACLreconstructions failed. At final follow-up, the antero-posterior (AP) drawer displacement in both studies hadalmost regained the value of the intact contralateralknee. Maximum load-to-failure improved over time inboth studies but was significantly lower at all time pointscompared with the intact ACL and the graft tissues.Tensile stress was significantly lower at final follow-up inboth studies compared with the intact ACL and grafttissues. It attained 43.3% of the intact ACL and 58.3%of the graft tissue in study 1 and 28.9% and 22.8% in

study 2, respectively. Conclusion The flexor tendon issuitable, and sheep appear to be an appropriate animalmodel for soft-tissue graft ACL reconstruction. Theytolerate the graft harvest well and quickly return to fullweight-bearing and physiological movement. Theirknees become stable without showing signs of macro-scopically evident osteoarthritis.

Keywords Anterior cruciate ligament reconstruction ÆSoft-tissue graft Æ Superficial flexor tendon Æ Sheepmodel Æ Clinical relevance

Introduction

Healing of anterior cruciate ligament (ACL) replace-ment surgery has been extensively studied in the past insmall and large animals. The graft tissue used wasmainly of patellar tendon origin [3, 32, 38, 39, 43, 51,68]. Recently in humans, the use of soft-tissue grafts hasbecome more popular, mainly the hamstring tendonsand special allograft tissue for ACL reconstruction, dueto the reduced graft site morbidity compared with thepatellar tendon graft [1, 8, 27, 28, 30, 44, 66]. Soft-tissuegrafts may present a different healing process, andespecially their osseous integration is of great interest.

Even though there have been various animal studieson graft healing, most studies used small animals, extra-articular models, or questionable graft tissues [11, 18, 19,29, 31, 35, 41, 42, 53, 62, 64]. To our knowledge, there isno clinically relevant animal model of soft-tissue graftACL reconstruction currently available. We thereforedeveloped a new animal model allowing for the study ofthe healing of a soft-tissue graft after ACL reconstruc-tion in a reproducible way with regard to the surgicalprocedure. The central questions of the study were: (1) Isthe utilized graft tissue suitable for ACL reconstructionin sheep? (2) Do they tolerate the graft harvest satis-factorily? (3) Do the sheep regain a stable knee afterACL reconstruction?

P. Hunt Æ S. U. Scheffler Æ F. N. Unterhauser Æ A. WeilerSports Traumatology and Arthroscopy Service,Trauma & Reconstructive Surgery, Charite,Campus Virchow Clinic,Humboldt University,Berlin, Germany

A. Weiler (&)Unfall- & Wiederherstellungschirurgie,Charite, Campus Virchow-Klinikum,Humboldt-Universitat zu Berlin,Augustenburger Platz 1,13353 Berlin, GermanyE-mail: andreas.weiler@charite.deTel.: +49-30-450552012Fax: +49-30-450552901

Arch Orthop Trauma Surg (2005) 125: 238–248DOI 10.1007/s00402-004-0643-z

Materials and methods

We used female sheep, 2 years of age, with skeletallymature hind limbs obtained from a licensed local dealer.The sheep used in study 1 were merino sheep (meanweight 51.4±8.7 kg), and the ones used in study 2 wereGerman moorland sheep (mean weight 34.1±3.5 kg).Before surgery, all animals were examined for generalhealth, lameness, and good physical conditions by a li-censed veterinarian. All procedures were performed withthe approval of the local governmental animal rightsprotection authorities in accordance with the animalrights protection laws for the use of laboratory animals(Reg. no. 0249/96, Reg. no. 0325/98).

We performed two studies in order to investigatetendon-to-bone healing and graft healing following in-tra- and extra-articular graft fixation as well as growthfactor application [56, 57, 58, 59, 61].

Pilot study A Suitability of the ovine semitendinosustendon as a soft-tissue graft, its accessability, and graftharvest effort were assessed. In case of its inapplicability,an appropriate substitute for ACL reconstruction had tobe found. Therefore, preliminary anatomical dissectionsof the hindlegs of 7 sheep were performed. For thispurpose, skin and fat were removed from the legs, andthe individual muscles and tendons were displayed(Figs. 1 and 2).

Pilot study B To determine the appropriateness of thenew animal model in vivo, we reconstructed the ACL in3 sheep in an open technique as will be described belowfor study 1. The sheep were put down 6 weeks aftersurgery, and standard biomechanical tests were per-formed followed by histological evaluations [57].

Study 1

In 42 animals, the ACL was replaced with a superficialflexor digitorum tendon split graft in an open technique.

We called it split graft because the tendon was longitu-dinally split. The grafts were directly fixed with biode-gradable interference screws (Fig. 3a). The sheep weresubdivided into six groups with 7 animals each. Theywere put down at 6, 9, 12, 24, 52 and 104 weeks aftersurgery. Twelve contralateral knees with intact ACLsserved as controls. Ten contralateral reconstructed kneesand 12 flexor tendon split grafts were tested for time zerocontrol [56, 57, 58]. As there was no information in theliterature on the recovery of animals after harvesting ofthe superficial flexor tendon, we only used a superficialflexor tendon split graft in pilot study B and study 1.

Study 2

In 48 animals, the ACL was replaced using the entiresuperficial flexor digitorum tendon from the belly of themuscle up to the calcaneal bone. The grafts were fe-morally fixed with an Endobutton and tibially over a

Fig. 1 Tweezer positioned under the ovine semitendinosus tendon

Fig. 2 Graft tissue (brace) and proximally released gastrocnemiusmuscle (arrow)

Fig. 3 a Schematic drawing of the ACL reconstruction techniqueused in study 1. Reprinted from [58].b Schematic drawing of theACL reconstruction technique used in study 2

239

bony bridge (Fig. 3b). The sheep were subdivided intofour groups with 12 animals each. They were killed at 3,6, 12, and 24 weeks after surgery. Since it is known thatgrowth factors enhance the healing of the medial col-lateral ligament in vivo [20, 34], we treated the grafts ofsix animals of each group with platelet-derived growthfactor BB (PDGF-BB). Yet, the effects of growth factortreatment on graft healing were not the purpose of thepresent report. We included this information solely toshow the feasibility and appropriateness of the animalmodel as such.

Ten contralateral knees with intact ACLs served ascontrols. Six contralateral reconstructed knees and 11flexor tendon grafts were tested for time zero control(unpublished data) [60, 61].

Surgical techniques

In both studies, the sheep received an intravenous doseof thiopental-sodium (20 mg/kg) as short-term anaes-

thesia. They were intubated, and anaesthesia wasmaintained using a mix of isoflurane (1.5%), oxygen,and nitrous oxide. Before surgery, all sheep received abroad spectrum antibiotic as antimicrobial prophylaxis.The left hind leg was shaved and prepared in the stan-dard sterile way.

Using an aseptic surgical technique, the left Achillestendon was approached via a posterolateral skin incision(Fig. 4a). Longitudinal splitting of the gastrocnemiustendon in the fiber direction revealed the superficialflexor tendon which was surrounded by it (Fig. 4b).After harvesting of the graft, the gastrocnemius tendonand the skin incision were sutured in layers with po-lyglactin and polyamidic sutures.

Study 1

Approximately two-thirds of the tendon of the leftsuperficial flexor digitorummuscle were sharply dissectedin the longitudinal fiber direction. Its length was 7.5 cm,and its diameter was approximately 4 mm (Fig. 4c).

As passing sutures, nonresorbable polyester sutures(2 Ethibond-Excel; Ethicon, Norderstedt, Germany)were passed through either end in a modified Bunnell

Fig. 4 a Posterolateral skin incision exposing the Achilles tendon.b Splitting of the gastrocnemius tendon. c Superficial flexor tendonafter dissection in fiber direction. d Display of the entire flexortendon before harvest. e Display of the entire flexor tendon afterharvest

240

technique. The prepared graft was wrapped in gauze andmoistened with 0.9% saline. The harvested graft lengthallowed for an interference fit fixation of at least 2.5 cmon both sides, leaving an intra-articular distance ofapproximately 1.5 cm.

Study 2

The entire tendon of the left superficial flexor digitorummuscle was harvested by dissecting it just distal fromthe muscle belly and proximal from its insertion at thecalcaneal tubercule (Fig. 4d). Graft length wasapproximately 7 cm, and the diameter was 6 mm(Fig. 4e). Graft preparation was equivalent to that instudy 1.

Arthrotomy

Surgery was performed with the sheep lying on theirright side. With the left leg in extension, the knee jointwas approached through an anteromedial skin incision.The joint capsule was opened by an anteromedialincision with release of the medial femoropatellar liga-ment [2]. The patella was placed laterally, and the kneewas brought to deep flexion. The Hoffa fat pad wasincised longitudinally, and the underlying ACL wascompletely excised at its tibial and femoral attach-ments. While protecting the posterior cruciate ligament(PCL), all remaining soft tissue was removed from theinsertion sites of the ACL. In study 1, a small wallplasty of approximately 2 mm was performed at thelateral face of the intercondylar notch, using a smallosteotome.

Tunnel preparation

With the knee maximally flexed, a 6 mm femoral tunnelwas drilled in the 1 o’clock position, leaving approxi-mately 2 mm of the dorsal femur cortex (Fig. 5a). Thetunnel was then enlarged by dilation with a 7 mm dilator

Fig. 5 a Femoral tunnel. b Drilling of the tibial tunnel. c Femoralinsertion of the graft. d Tibial insertion of the graft and Beath pin. eAnatomic position of the graft. f Femoral application of the screwin study 1. g Tying of the Endobutton in study 2. h Tibial bony postin study 2

241

(Instrument Makar, Okemos, MI, USA). The corticalbone at the tunnel entrance was then carefully cham-fered to a diameter of approximately 8 mm. The femoraltunnel wall was notched at the 12 o’clock position alongthe tunnel wall as a guide for screw insertion to preventclockwise rotation of the screw around the graft. A6 mm tibial tunnel was drilled from inside in an outwarddirection, in contrast to the human situation, at theanatomic tibial insertion site of the ACL (Fig. 5 b). Atibial drill-guide was used.

Tunnel preparation in study 2 was performed in thesame way as in study 1, with the difference that thediameter of the tibial tunnel was created according to thesize of the graft diameter previously measured with aspecial template in increments of 0.5 mm.

Graft insertion and fixation

The graft was inserted via the passing sutures using aBeath pin. It was inserted in an inside-out direction intothe femoral tunnel (Fig. 5c), followed by inside-outinsertion into the tibial tunnel (Fig. 5d). By that point,the graft had assumed the anatomic position of the na-tive ACL (Fig. 5e).

Study 1

The biodegradable interference screws used for graftfixation consisted of pure poly-(D,L-lactide) (Sysorb,Centerpulse Orthopedics, Baar, Switzerland). Screwparameters were 23 mm in length and 8 mm in threaddiameter. Under bilateral graft tensioning, the firstinterference screw was inserted into the femoral tunnel(Fig. 5f). The knee was then taken through several fullranges of motion under 90 N of graft pretensioningusing a tensiometer (Centerpulse). The tibial screw wasinserted in an inside-out direction in 90� of flexion underpretension with 90 N. We chose the inside-out applica-tion as the subchondral cancellous bone lamella in sheepis very thin (Fig. 6), and in a preliminary sheep cadaverstudy, a mechanically more secure graft fixation wasachieved by inside-out application of the screw than withthe outside-in application performed in humans [58].

Study 2

Femorally, the graft was fixed by openly tying the at-tached sutures to an Endobutton (Acufex; Smith &Nephew Endoscopy, MA, USA) (Fig. 5g), in contrast tothe human situation, where it is passed through the bonetunnel and then flipped. After taking the knee throughseveral full ranges of motion as described above, thegraft was fixed at the tibia in 60� of flexion with thesutures tied over a bony bridge (Fig. 5h).

After graft fixation, its tension throughout the fullrange of motion and the knee motility were tested byrepeatedly flexing and extending the knee. Thereafter,the separated anterior fat pad was sutured with polygl-

actin sutures, and the patella was repositioned. Themedial femoropatellar ligament was repaired with astrong polyglactin suture. The joint capsule, soft tissue,and skin were sutured individually, using polyglactinand polyamidic sutures.

Postoperative measures

Each wound was covered with dressing for a period ofapproximately 3 postoperative days. After a postopera-tive 12- to 24-hour stay in a special wake-up cage, thesheep were reunited with their flock at the laboratorypen. They were allowed to bear full weight with nolimitation of range of motion. As pain relief medication,we used an intramuscular injection of tramadol (50 mg)and metamizole (1.25 g) for a period of 3–4 days.

During the first 14 postoperative days and prior tosacrifice, gait was recorded on a daily basis. Weight-bearing was visually monitored for 4 weeks and gradedon a scale of 0–4 (Table 1). Normal gait was defined asno visually detectable abnormalities during running andpronking. Two weeks after surgery, the animals weretransferred to an external animal housing institution tomove without restriction.

Mechanical testing and statistical analysis

Animals were put down using thiopental-sodium (20 mg/kg) followed by an overdose of potassium chloride (60 ml;1 mmol/ml) intravenously. Knee harvesting was per-formed, leaving the joint capsule, the collateral ligaments,and intra-articular structures and tissues untouched.Specimens were wrapped in saline-soaked gauze andstored at –20�Cuntil 12 h prior to testing, when they werethawed at 4�C. After removal of all soft tissue from the

Fig. 6 Ovine tibia with borderline between proximal subchondralbone lamella and the medullar cavity marked (dashed line)

242

tibia and femur, the free bone ends were embedded inaluminum cylinders using polymethylmethacrylate.Dehydration of the specimens was prevented using saline(0.9%) spray. With the knee mounted on a material test-ingmachine (Zwick 1455; Zwick,Ulm,Germany), drawertestingwas donewith the knee in 90�of flexion and neutralrotation with an anteroposterior (AP) load application of±50 N and a displacement rate of 1 mm/s. FollowingAPdrawer testing of the entire knee, macroscopic inspectionof possible chondral pathology was undertaken. Subse-quently, all soft tissue but the ACLwas removed from thebones, and the AP drawer test was repeated. Failure testswere performedwith the specimensmounted on a custom-made adjustment device, allowing for free rotation of theconstruct in different knee flexion angles. With a preloadof 1 N and a load displacement rate of 1 mm/s, specimenswere loaded until failure. Further details of knee har-vesting and mechanical testing have been explicitly de-scribed before [56, 57, 58]. Data were analyzed for equaldistribution using theKolmogorov-Smirnov test. Becauseequal distribution testing failed, unpaired comparisonsbetween the control groups and changes between thegroups over time were performed using the Mann-Whit-ney U Wilcoxon rank sum test. Overall differences be-tween the study and the control groups were tested usingKruskal-Wallis analysis of variance. Failure modes dur-ing mechanical testing were compared using chi-squareanalysis. The differences were considered to be significantat a probability level of p £ 0.05.

Since in both studies there were different questions toanswer, not all results refer to both groups alike.

Results

Pilot study A

The anatomical dissections of the hindlegs revealed thesemitendinosus tendon of sheep to be a fascia-like,transparent structure. It is too short and thin to be usedas a solid soft-tissue graft (Fig. 1). The superficial flexordigitorum tendon was surrounded by the gastrocnemiustendon, and together they formed the Achilles tendon,which was easy to access surgically. After splitting of thelatter, the flexor tendon could be displayed as a solid,

approximately 6–8 cm long structure with a diameter ofabout 6 mm (Fig. 2).

Pilot study B

In all 3 animals, the arthrotomy and tendon harvest siteshealed without complications. Prior to sacrifice they hadreturned to full weight-bearing without visible signs oflameness during running or pronking. Gross inspectionof the knees showed that all grafts were intact andcovered by a synovial membrane. Mild inflammatorychanges with hypervascularization and thickened syno-vial tissue were found. Biomechanical tests revealedneither tibial nor femoral graft pullout. Histologicalevaluations did not show any signs of tissue damage atthe screw-graft interface [57, 58].

Studies 1 and 2

In study 1, five animals had to be excluded, 2 due to sur-gical failure (dorsal femoral wall break-out) and 3 due tononsurgery-related complications (pneumonia, pericar-ditis, intestinal obstruction) after 5 days, 5 and 27 weeks.In the 52-week-group one graft failed and showed a thinfibrous graft residuum adjacent to the PCL [58]. At 6–24and 104 weeks, 6 specimens, and in the 52-week-group, 5specimens were included in the final evaluation. In study2, 1 animal of the 24-week group had suffered a patelladislocation which was detected at sacrifice and wastherefore excluded from the final evaluation.

Arthrotomy and graft harvest sites healed withoutcomplications. Gross inspection of the harvested kneesshowed slight inflammatory changes 6 weeks after sur-gery. At all other timepoints, no pathological changeswithin the knee joint could be detected. In some animals,a chondromalacia grade II to III was found in thetrochlear groove, but did not differ from the intactcontralateral knee.

Recovery and weight-bearing

All animals recovered without complications from theanaesthesia. Arthrotomy and graft harvest sites healedsmoothly. By 2–3 days after surgery, most of the ani-mals developed a slight, sterile effusion within the knee,which was resorbed without complications or treatmentafter about 10 days. The sheep returned to normal gaitwith no visible lameness during motion. Mean grade ofweight-bearing of the animals was 1.4±0.8 in study 1and 1.9±0.8 in study 2 at 1 week after surgery andsteadily improved to 3.9±2.1 (study 1) and 3.9±0.49(study 2) after 4 weeks. Prior to sacrifice, all animalsshowed a normal gait pattern as described above.

Gross inspection of the stifle joints proved all grafts tobe intact and covered with a synovial membrane as de-scribed above for pilot studyB. In particular, there was no

Table 1 Classification of postoperative weight-bearing of the sheep

Grade ofweight-bearing

Gait pattern

0 Effective three-leggedwalking

1 Clear lameness duringwalking and running

2 Bare lameness during walkingand clear lameness duringrunning or pronking

3 Bare lameness duringrunning or pronking

4 No visible lameness duringrunning or pronking

243

macroscopically visible meniscus lesion or cartilagedegeneration at the femoral condyles or the tibial plateau.

ACL and graft properties

AP drawer displacement of the intact knee was2.04±0.69 mm in study 1 and 2.8±0.3 mm in study 2(Table 2). Stiffness at failure of the intact ACL was143.9±16.1 N/mm standard in study 1 and 104.2±15.2 N/mm in study 2 (Table 3). Maximum load-to-failure of the intact ACL was 1513.3±180.3 N in study 1and 888.2±139.4 N in study 2. The split graft in study 1exhibited 74% (1120.1±223.4 N) of the ACL, and theentire flexor tendon graft in study 2 exhibited 172.3%(1530.6±523.1 N) of the ACL. The value of maximumload-to-failure of the intact ACL in study 1 was signifi-cantly higher (p £ 0.05) than the value of the split graftand the same as the value of the entire flexor tendon graftused in study 2. Maximum load-to-failure of the intactACL in study 2 was significantly lower (p £ 0.05) than ofthe entire flexor tendon graft (Table 4). Concerningtensile stress, the split graft in study 1 exhibited 74.3%(39.8±7.8 MPa) of the intact ACL (53.6±13.6 MPa),and the entire flexor tendon graft in study 2 exhibited116% (53.0±13.0 MPa) of the intact ACL(41.9±4.5 MPa). Tensile stress of the intact ACL instudy 1 was significantly higher (p £ 0.05) compared withthe split graft and the same as the value of the entireflexor tendon graft in study 2 (Table 5).

Reconstruction properties

In study 1, AP drawer displacement for time zeroreconstruction was 134% (2.74±0.86) of the intactcontralateral knee. It was significantly higher (p=0.03)than the intact contralateral knee, whereas in study 2there was no significant difference (Table 2). In study 1,stiffness at failure for time zero reconstruction was28.6% (41.2±13 N/mm) of the ACL (143.9±16.1 N/mm). In study 2, stiffness at failure for time zeroreconstruction was 15.6% (16.2±2.8 N/mm) of theACL (104.2±15.2 N/mm). In both studies, stiffness at

failure for time zero reconstruction was significantlylower (p £ 0.05) compared with the intact ACL(Table 3). Maximum load-to-failure for time zeroreconstruction was 17.4% (267±82 N) of the ACL and23.8% of the split graft in study 1. In study 2, it was24.4% (217.4±54.2 N) of the ACL and 14.2% of theentire flexor tendon graft. Maximum load-to-failure fortime zero reconstruction was significantly lower(p £ 0.05) in studies 1 and 2 compared with the intactACL and both grafts (Table 4).

Reconstruction properties over time

In study 1, AP drawer displacement improved over time,but it was significantly higher (p £ 0.05) at all postsur-gical timepoints than the intact contralateral knee. Up to24 weeks after surgery, AP drawer displacement wassignificantly higher (p £ 0.05) compared with time zeroreconstruction. However, at 52 weeks, there was nodifference at all, and at 104 weeks the difference wasinsignificant. In study 2, there was no significant changein AP drawer displacement over the groups or time(Table 2).

In study 1, stiffness at failure substantially improvedover time but was significantly lower (p £ 0.05) up to52 weeks after surgery compared with the intact ACL. At52 and 104 weeks postoperatively, it was significantly

Table 3 Results showing stiffness data (N/mm) of studies 1 and 2.Values are given as mean±standard deviation

Time aftersurgery

Study 1 Study 2

Control PDGF

Intact ACL 143.9±16.1 104.2±15.20 weeks 41.2±13 16.2±2.83 weeks - 45.8±19.4 39±106 weeks 18.3±4.4 36±13.4 46±109 weeks 24.7±7 - -12 weeks 33.5±7.2 57.4±18.4 44.9±924 weeks 58.6±25.9 50.1±4.9 69.8±23.152 weeks 77.9±12.9 - -104 weeks 99.5±50 - -

Table 2 Results showing the AP drawer displacement (mm) ofstudies 1 and 2. Values are given as mean±standard deviation

Time aftersurgery

Study 1 Study 2

Control PDGF

Intact knee 2.04±0.69 2.8±0.3Time zero reconstruction 2.74±0.86 2.4±0.63 weeks - 2.6±1.1 2.6±0.76 weeks 8.25±2.31 3.9±2.8 3.1±0.99 weeks 5.93±1.67 - -12 weeks 3.91±1 2.9±0.8 3.2±0.924 weeks 3.81±1.21 3.0±0.8 2.6±0.752 weeks 2.78±0.5 - -104 weeks 3.5±0.98 - -

Table 4 Results showing the maximum load-to-failure (N) ofstudies 1 and 2. Values are given as mean±standard deviation

Time aftersurgery

Study 1 Study 2

Control PDGF

Intact ACL 1513.3±180.3 888.2±139.4Graft 1120.1±223.4 1530.6±523.1Time zero reconstruction 267±82 217.4±54.23 weeks - 232.5±46 228.1±44.66 weeks 44.8±4 135.5±48.4 192.8±24.19 weeks 105±43 - -12 weeks 237±59.8 302.2±75.1 244±58.624 weeks 313.8±164.4 465.5±97.3 468.3±190.652 weeks 684.9±252.7 - -104 weeks 669.7±283.3 - -

244

higher (p £ 0.05) than at time zero reconstruction. At104 weeks, it was not significantly lower (p £ 0.05) thanthe intact ACL. In study 2, stiffness at failure improvedover time but was significantly lower (p £ 0.05) comparedwith the intact ACL at all time points. Stiffness of theanatomic graft fixation in study 1was considerably higherat time zero reconstruction but considerably lower at6 weeks compared with the extracortical graft fixation instudy 2.At 24 weeks stiffness at failure in both studies wassimilar, and in study 2 significantly higher (p £ 0.05)compared with time zero reconstruction (Table 3). Max-imum load-to-failure improved over time in both studiesbut was significantly lower (p £ 0.05) at all time pointscompared with the intact ACL and the graft tissues. Instudy 1,maximum load-to-failure was significantly higherat 52 and 104 weeks postoperatively, and in study 2, it wassignificantly higher (p £ 0.05) at 24 weeks compared withtime zero reconstruction. There was no improvement ofmaximum load-to-failure between 52 and 104 weekspostoperatively (Table 4). Tensile stress for both studiesimproved considerably over time but remained signifi-cantly lower (p £ 0.05) compared with the ACL and thegraft tissues (Table 5).

Discussion

The aim of this study was to develop a clinically validand reproducible animal model of soft-tissue graft ACLreconstruction in terms of graft suitability, tolerance ofgraft harvest by the animals, and regaining of postsur-gical knee stability, using current intra- and extra-articular fixation methods.

Many animal studies investigating tendon-to-bonehealing have been performed in extra-articular settingsand/or small animals [11, 18, 19, 29, 31, 35, 41, 42, 53,62, 64]. These models have certain advantages over in-tra-articular models concerning the evaluation of thegeneral mechanical conditions of the tendon-boneinterface due to there being fewer variables such as graftremodeling or synovial fluid inflow. Nevertheless, inorder to study the healing process of the grafted tissueafter reconstruction of the ACL in a real situation,

extra-articular models may be of lesser value as thetendon-to-bone healing and graft remodeling process,mechanical and biological boundary conditions arefundamentally different within the stifle joint than extra-articularly.

Recently, publications on the outcomes of cruciateligament reconstruction in animal models have been onthe increase. There are multiple reports on ACL recon-struction using small animals such as rats, rabbits, dogsand rhesus monkeys [3, 5, 6, 10, 12, 17, 33, 45, 54, 65,67]. Although these models are surely more appropriatein order to study graft healing after ACL reconstructionthan extra-articular set-ups, it is questionable for variousreasons whether small animals resemble clinically validmodels for state-of-the-art ACL reconstruction. Theirstifle joints are too small to adequately place the bonetunnels which are crucial for the final outcome, and theuse of state-of-the-art intra-articular fixation methods.As intra-articular graft fixation using interference screwsis next to impossible in small animals, comparisons ofdifferent graft and tendon-to-bone healing processessuch as intra-articular vs. extra-cortical graft fixationcannot be adequately studied. Due in part to funda-mentally different gait patterns (e.g. rabbit vs. dog), re-sults may differ significantly and therefore lackcomparability and clinical relevance. Additionally, dogsseem to develop cartilage defects after ACL recon-struction [12], and rabbits are always limited to cageactivity, and thus unrestricted postsurgical movement isimpossible. Therefore, an easily available and manage-able animal with stifle joint properties anatomicallysimilar to the human knee joint is required.

Pigs are not used as in vivo models for ACL recon-struction in the literature despite the mechanical simi-larities to the human knee [63], most likely due to thelack of manageability. Sheep as well as goats are to ourknowledge the only large animal models utilized forACL reconstruction which approach these requirements[2, 40] and have been frequently used for studies of ACLreplacement healing [9, 14, 15, 16, 23, 24, 25, 26, 37, 38,43, 47, 56, 58, 59, 68]. However, even these animals haveanatomical limitations regarding clinical relevance, sincetheir stifle joint size is still far from the human situation,and precise graft tunnel placement is difficult to perform[56]. Even though the sheep has proven to be a suitableanimal for ACL reconstruction models, using bone-patellar tendon-bone grafts, still a soft-tissue graft withsimilar mechanical properties to the intact ACL had tobe found.

Different soft-tissue grafts have been used in the past.Shino et al. utilized the central part of the patellar tendonwithout attached bone blocks in dogs [51, 52]. Some au-thors report ACL reconstruction using an extensortendon soft-tissue autograft in sheep (unpublished data)[7, 55]. However, which specific extensor tendon had beenused was not stated by these authors. Goradia et al. de-scribed the use of the semitendinosus tendon in sheep [14,15, 16]. Our own dissections of ovine hindlegs revealedtheir semitendinosus tendon to be too thin and short to be

Table 5 Results showing tensile stress (MPa) of studies 1 and 2.Values are given as mean±standard deviation

Time aftersurgery

Study 1 Study 2

Control PDGF

Intact ACL 53.6±13.6 41.9±4.5Graft 39.8±7.8 53.0±13.0Time zero reconstruction - -3 weeks - 4.7±1.3 7.5±2.16 weeks 2.7±0.9 2.7±1.1 5.1±2.59 weeks 3.8±1.3 - -12 weeks 6.3±1.1 8.4±2.3 12.6±5.724 weeks 10.1±2.8 12.1±1.7 15.2±4.952 weeks 25.4±10.8 - -104 weeks 23.2±8.7 - -

245

used as a soft-tissue graft (Fig. 1), therefore it was notclear to us which tissue exactly was used in the studies ofGoradia et al. In order to harvest a graft which ismechanically suitable as a free soft-tissue ACL substitute,easy to access, and does not compromise wound healingnor gait, nor require excessive postoperative woundtreatment, we chose to use the superficial flexor digitorumtendon. It proved to be a solid and sufficiently long graftand biomechanically similar to the intact ACL. Since toour knowledge we were the first ones to use this tendon asgraft tissue, there was no information in the literatureconcerning its biomechanical properties, postsurgicalrecovery of the animals, or graft healing. Hence we usedonly a flexor tendon split graft in pilot studyBand study 1.The satisfying postoperative recovery of the animals inthese two studies regarding gait and weight-bearingencouragedus to use the entire flexor tendon in study 2.Asall of the sheep returned to full weight-bearing and novisible lameness approximately 4 weeks after surgery,even though the operated leg was by no means protectedand the animals were not restricted in their motion, weassume that the harvest of the superficial flexor tendondoes not remarkably impair their long-term ability towalk and pronk.

As regards graft remodeling, graft fixation in allstudies did not prove to be the weak link during the earlypostoperative period, even though the sheep were neverrestricted in their motion. Just one of the reconstructionsfailed in study 1. Therefore, we assume that concerningearly biomechanical stability, both fixation methods aresafe to be performed in an animal study.

The significance of regaining knee stability for a posi-tive long-term outcome of ACL reconstruction is undis-puted. Even though the stability of the knees did notregain the values of the intact knee, especially in study 1,we could show that after a period of relative instability upto week 24, the knees obtained values comparable to timezero reconstruction. In man, the mean side-to-side dif-ference after successful reconstruction of the ACL aver-ages 2 mm [49]. In study 2, there were no significantdifferences inAP drawer displacement compared not onlywith the time zero control but even compared with theintact ACL. This showed that the knee did not lose sta-bility during the entire healing phase.

Studies on cadaveric human knees during cyclicloading with loads of 200 N and more, comparing intra-vs. extra-articular graft fixations, showed the intra-articular fixation methods to be superior with regard tomechanical and biological results. Especially in the earlypostoperative period, it is generally mechanically supe-rior with respect to AP drawer displacement, stiffness,energy loss, and laxity increase [46]. Biologically, intra-articular graft fixation promotes osseous integration ofthe graft and development of a direct ligament insertion,as it prevents graft motion within the tunnel and tunnelenlargement, resulting from the so-called ‘windshield-wiper effect’ [21, 22, 60]. Concerning stiffness at failure,both studies never regained the values of the intact ACL.However, the intra-articular graft fixation proved to be

superior to the extra-articular fixation method in termsof tendon-to-bone healing and tunnel enlargement [60].The loss of range of motion which is often referred to as‘arthrofibrosis’ is a commonly known complication fol-lowing ACL reconstruction [49, 50]. None of the sheepin our studies ever developed a visible or measurablerestriction of range of motion. Nevertheless, Amis et al.found significant loss of extension and measurable lossof flexion 6, 12, and 24 months after reconstruction ofthe ACL in sheep using an artificial ligament [4]. Webelieve that the increase in knee stability of the sheepmay not have been a result of fibrotic degeneration, butsolely of biomechanical improvement of the autograft,as signs of capsular thickening could not be detected.

Previous studies have demonstrated the significanceof sufficient space in the intercondylar notch in order toprevent graft impingement [13]. The ovine knee joint isconsiderably smaller and the intercondylar notch nar-rower compared with the human knee [63]. Therefore, agraft tunnel with a diameter greater than 7 mm bears therisk of trauma to the PCL during tunnel preparation.We performed a wall plasty in all animals of pilot studyB and study 1 in order to prevent graft impingement andallow for an easier placement without violating the PCL.In contrast to humans, the stifle joint of sheep neverattains full extension but is maximally extended at anangle 35±5� below a fully straightened human leg [2].Thus, the risk of roof impingement in sheep is generallyless than in humans, if the tunnels are positionedproperly. We did not perform a wall plasty in any of theanimals in study 2 and could not find any signs of graftimpingement thereafter.

The differences in the values of the ACL and the grafttissues in studies 1 and 2 are most likely due to thedifferent breeds mentioned above. Generally, all animalmodels for ACL reconstruction have significant limita-tions regarding knee anatomy and gait pattern alone.There are very few animals which have continuous (os-trich) or occasional (bear) bipedal gait. Biomechanicalcomparisons between ovine, caprine, and porcine kneesshowed the porcine knee to be closest to the humansituation [63]. Histological comparisons between younghuman, canine, bovine, and ovine ACLs showed thecanine ACL to be closest to the human ACL (unpub-lished data) [36]. However, other factors than the onesdirectly related to the study and its results have to beconsidered, such as availability, postsurgical treatmentof the animals, and practicability. Sheep are comparablyeasy to obtain and handle and have anatomic kneecharacteristics similar to those of humans [40].

In conclusion, we can say that the tissue used issuitable for soft-tissue graft ACL reconstruction, due toits similar mechanical properties to the native ovineACL. The sheep appears to be a good model forautogenous soft-tissue graft healing with respect tocurrent surgical techniques and biomechanical results,since the results following surgery are quite similar to thehuman situation. They tolerate the graft harvest wellwith return to full weight-bearing approximately

246

4 weeks after surgery, and they regain considerable kneestability without visible restrictions of motion or mac-roscopically evident development of osteoarthritis.

References

1. Aglietti P, Buzzi R, D’Andria S, Zaccherotti G (1993) Patel-lofemoral problems after intraarticular anterior cruciate liga-ment reconstruction. Clin Orthop:195–204

2. Allen MJ, Houlton JE, Adams SB, Rushton N (1998) Thesurgical anatomy of the stifle joint in sheep. Vet Surg 27:596–605

3. Amiel D, Kleiner JB, Roux RD, Harwood FL, Akeson WH(1986) The phenomenon of ‘ligamentization’: anterior cruciateligament reconstruction with autogenous patellar tendon. JOrthop Res 4:162–172

4. Amis AA, Camburn M, Kempson SA, Radford WJ, Stead AC(1992) Anterior cruciate ligament replacement with polyesterfibre. A long-term study of tissue reactions and joint stability insheep. J Bone Joint Surg Br 74:605–613

5. Aune AK, Hukkanen M, Madsen JE, Polak JM, Nordsletten L(1996) Nerve regeneration during patellar tendon autograftremodelling after anterior cruciate ligament reconstruction: anexperimental and clinical study. J Orthop Res 14:193–199

6. Blickenstaff KR, Grana WA, Egle D (1997) Analysis of asemitendinosus autograft in a rabbit model. Am J Sports Med25:554–559

7. Bliss J, Svehla M, Yu Y, Bruce W, Waller C, Arm DM, Ball C,GiffordK,WalshW (2002) Early effect of autologous platelets atthe tendon-bone interface. In: Proceedings of the 48th AnnualMeeting of the Orthopaedic Research Society, Dallas, TX, USA

8. Breitfuss H, Frohlich R, Povacz P, Resch H, Wicker A (1996)The tendon defect after anterior cruciate ligament reconstruc-tion using the midthird patellar tendon—a problem for thepatellofemoral joint? Knee Surg Sports Traumatol Arthrosc3:194–198

9. Bush-Joseph CA, Cummings JF, Buseck M, Bylski-AustrowDI, Butler DL, Noyes FR, Grood ES (1996) Effect of tibialattachment location on the healing of the anterior cruciateligament freeze model. J Orthop Res 14:534–541

10. Clancy WG Jr, Narechania RG, Rosenberg TD, Gmeiner JG,Wisnefske DD, Lange TA (1981) Anterior and posterior cru-ciate ligament reconstruction in rhesus monkeys. J Bone JointSurg Am 63:1270–1284

11. Forward AD, Cowan RJ (1963) Tendon suture to bone. J BoneJoint Surg Am 45:807–823

12. Goertzen M, Dellmann A, Gruber J, Clahsen H, Burrig KF(1993) Homologous cruciate ligament transplantation as intra-articular ligament replacement (in German). Z Orthop131:179–186

13. Good L, Odensten M, Gillquist J (1991) Intercondylar notchmeasurements with special reference to anterior cruciate liga-ment surgery. Clin Orthop:185–189

14. Goradia VK, Rochat MC, Grana WA, Egle DM (1998)Strength of ACL reconstructions using semitendinosus tendongrafts. J Okla State Med Assoc 91:275–277

15. Goradia VK, Rochat MC, Grana WA, Rohrer MD, Prasad HS(2000) Tendon-to-bone healing of a semitendinosus tendonautograft used for ACL reconstruction in a sheep model. Am JKnee Surg 13:143–151

16. Goradia VK, Rochat MC, Kida M, Grana WA (2000) Naturalhistory of ahamstring tendonautograft used for anterior cruciateligament reconstruction in a sheep model. Am J Sports Med28:40–46

17. Grana WA, Egle DM, Mahnken R, Goodhart CW (1994) Ananalysis of autograft fixation after anterior cruciate ligamentreconstruction in a rabbit model. Am J Sports Med 22:344–351

18. Grassman SR, McDonald DB, Thornton GM, Shrive NG,Frank CB (2002) Early healing processes of free tendon graftswithin bone tunnels is bone-specific: a morphological study in arabbit model. Knee 9:21–26

19. Greis PE, Burks RT, Bachus K, Luker MG (2001) The influ-ence of tendon length and fit on the strength of a tendon-bonetunnel complex. A biomechanical and histologic study in thedog. Am J Sports Med 29:493–497

20. Hildebrand KA, Woo SL, Smith DW, Allen CR, Deie M,Taylor BJ, Schmidt CC (1998) The effects of platelet-derivedgrowth factor-BB on healing of the rabbit medial collateralligament. An in vivo study. Am J Sports Med 26:549–554

21. Hoher J, Moller HD, Fu FH (1998) Bone tunnel enlargementafter anterior cruciate ligament reconstruction: fact or fiction?Knee Surg Sports Traumatol Arthrosc 6:231–240

22. Hoher J, Livesay GA, Ma CB, Withrow JD, Fu FH, Woo SL(1999) Hamstring graft motion in the femoral bone tunnelwhen using titanium button/polyester tape fixation. Knee SurgSports Traumatol Arthrosc 7:215–219

23. Holden JP, Grood ES, Butler DL, Noyes FR, Mendenhall HV,Van Kampen CL, Neidich RL (1988) Biomechanics of fascialata ligament replacements: early postoperative changes in thegoat. J Orthop Res 6:639–647

24. Jackson DW, Grood ES, Arnoczky SP, Butler DL, Simon TM(1987) Cruciate reconstruction using freeze dried anterior cru-ciate ligament allograft and a ligament augmentation device(LAD). An experimental study in a goat model. Am J SportsMed 15:528–538

25. Jackson DW, Grood ES, Cohn BT, Arnoczky SP, Simon TM,Cummings JF (1991) The effects of in situ freezing on theanterior cruciate ligament. An experimental study in goats. JBone Joint Surg Am 73:201–213

26. Jackson DW, Simon TM, Lowery W, Gendler E (1996) Bio-logic remodeling after anterior cruciate ligament reconstructionusing a collagen matrix derived from demineralized bone. Anexperimental study in the goat model. Am J Sports Med24:405–414

27. Jansson KA, Harilainen A, Sandelin J, Karjalainen PT, AronenHJ, Tallroth K (1999) Bone tunnel enlargement after anteriorcruciate ligament reconstruction with the hamstring autograftand endobutton fixation technique. A clinical, radiographicand magnetic resonance imaging study with 2 years follow-up.Knee Surg Sports Traumatol Arthrosc 7:290–295

28. Jarvela T, Paakkala T, Kannus P, Jarvinen M (2001) Theincidence of patellofemoral osteoarthritis and associated find-ings 7 years after anterior cruciate ligament reconstruction witha bone-patellar tendon-bone autograft. Am J Sports Med29:18–24

29. Jones JR, Smibert JG, McCullough CJ, Price AB, Hutton WC(1986) Tendon implantation into bone: an experimental study.J Hand Surg Br 12:306–312

30. Kartus J, Magnusson L, Stener S, Brandsson S, Eriksson BI,Karlsson J (1999) Complications following arthroscopic ante-rior cruciate ligament reconstruction. A 2–5-year follow-up of604 patients with special emphasis on anterior knee pain. KneeSurg Sports Traumatol Arthrosc 7:2–8

31. Kernwein GA (1942) A study of tendon implantations intobone. Surg Gynecol Obstet 75:794–796

32. Kuroda R, Kurosaka M, Yoshiya S, Mizuno K (2000) Local-ization of growth factors in the reconstructed anterior cruciateligament: immunohistological study in dogs. Knee Surg SportsTraumatol Arthrosc 8:120–126

33. Labs K, Perka C, Schneider F (2002) The biological and bio-mechanical effect of different graft tensioning in anterior cru-ciate ligament reconstruction: an experimental study. ArchOrthop Trauma Surg 122:193–199

34. Letson AK, Dahners LE (1994) The effect of combinations ofgrowth factors on ligament healing. Clin Orthop 308:207–212

35. Liu SH, Panossian V, al-Shaikh R, Tomin E, Shepherd E,Finerman GA, Lane JM (1997) Morphology and matrix com-position during early tendon to bone healing. Clin Orthop339:253–260

36. Murray M, Weiler A, Spindler K (2002) Histologic com-parisons of the ACL in humans, dogs, cows and sheep. In:Proceedings of the 48th Annual Meeting of the OrthopaedicResearch Society, Dallas, TX, USA

247

37. Ng GY, Oakes BW, Deacon OW, McLean ID, Lampard D(1995) Biomechanics of patellar tendon autograft for recon-struction of the anterior cruciate ligament in the goat: three-year study. J Orthop Res 13:602–608

38. Ng GY, Oakes BW, Deacon OW, McLean ID, Eyre DR(1996) Long-term study of the biochemistry and biome-chanics of anterior cruciate ligament-patellar tendon auto-grafts in goats. J Orthop Res 14:851–856

39. Park MJ, Lee MC, Seong SC (2001) A comparative study of thehealing of tendon autograft and tendon-bone autograft usingpatellar tendon in rabbits. Int Orthop 25:35–39

40. Radford WJP, Amis AA, Stead AC (1996) The ovine stifle asa model for human cruciate ligament surgery. Vet CompOrthop Traumatol 9:134–139

41. Rodeo SA, Arnoczky SP, Torzilli PA, Hidaka C, Warren RF(1993) Tendon-healing in a bone tunnel. A biomechanicaland histological study in the dog. J Bone Joint Surg Am75:1795–1803

42. Rodeo SA, Suzuki K, Deng XH, Wozney J, Warren RF (1999)Use of recombinant human bone morphogenetic protein-2 toenhance tendon healing in a bone tunnel. Am J Sports Med27:476–488

43. Roth JH, Mendenhall HV, McPherson GK (1988) Theeffect of immobilization on goat knees following reconstruc-tion of the anterior cruciate ligament. Clin Orthop 229:278–282

44. Sachs RA, Daniel DM, Stone ML, Garfein RF (1989) Pa-tellofemoral problems after anterior cruciate ligament recon-struction. Am J Sports Med 17:760–765

45. Sakai H, Fukui N, Kawakami A, Kurosawa H (2000) Bio-logical fixation of the graft within bone after anterior cru-ciate ligament reconstruction in rabbits: effects of theduration of postoperative immobilization. J Orthop Sci 5:43–51

46. Scheffler SU, Sudkamp NP, Gockenjan A, Hoffmann RF,Weiler A (2002) Biomechanical comparison of hamstring andpatellar tendon graft anterior cruciate ligament reconstruc-tion techniques: the impact of fixation level and fixationmethod under cyclic loading. Arthroscopy 18:304–315

47. Schiavone Panni A, Denti M, Franzese S, Monteleone M(1993) The bone-ligament junction: a comparison betweenbiological and artificial ACL reconstruction. Knee SurgSports Traumatol Arthrosc 1:9–12

48. Shelbourne KD, Davis TJ (1999) Evaluation of knee stabilitybefore and after participation in a functional sports agilityprogram during rehabilitation after anterior cruciate ligamentreconstruction. Am J Sports Med 27:156–161

49. Shelbourne KD, Johnson G (1994) Evaluation of knee exten-sion following anterior cruciate ligament reconstruction.Orthopedics 17:205–206

50. Shelbourne KD, Patel DV (1999) Treatment of limited mo-tion after anterior cruciate ligament reconstruction. KneeSurg Sports Traumatol Arthrosc 7:85–92

51. Shino K, Horibe S (1991) Experimental ligament recon-struction by allogeneic tendon graft in a canine model. ActaOrthop Belg 57 [Suppl 2]:44–53

52. Shino K, Kawasaki T, Hirose H, Gotoh I, Inoue M, Ono K(1984) Replacement of the anterior cruciate ligament by anallogeneic tendon graft. An experimental study in the dog. JBone Joint Surg Br 66:672–681

53. Shoemaker SC, Rechl H, Campbell P, Kram HB, Sanchez M(1989) Effects of fibrin sealant on incorporation of autograftand xenograft tendons within bone tunnels. A preliminarystudy. Am J Sports Med 17:318–324

54. Tohyama H, Yasuda K (1998) Significance of graft tension inanterior cruciate ligament reconstruction. Basic backgroundand clinical outcome. Knee Surg Sports Traumatol Arthrosc 6[Suppl 1]:S30–S37

55. Walsh WR, Huckle J, Bliss J, Evans R, Yu Y, Hughes P, BruceW, Carter A (2003) Can low-intensity ultrasound be used toaugment tendon-to-bone healing? In: Proceedings of the 49thAnnual Meeting of the Orthopaedic Research Society, NewOrleans, LA, USA

56. Weiler A, Peters G, Maurer J, Unterhauser FN, Sudkamp NP(2001) Biomechanical properties and vascularity of an anteriorcruciate ligament graft can be predicted by contrast-enhancedmagnetic resonance imaging. A two-year study in sheep. Am JSports Med 29:751–761

57. Weiler A, Hoffmann RF, Bail HJ, Rehm O, Sudkamp NP(2002) Tendon healing in a bone tunnel. Part II. Histologicanalysis after biodegradable interference fit fixation in a modelof anterior cruciate ligament reconstruction in sheep.Arthroscopy 18:124–135

58. Weiler A, Peine R, Pashmineh-Azar A, Abel C, Sudkamp NP,Hoffmann RF (2002) Tendon healing in a bone tunnel. Part I.Biomechanical results after biodegradable interference fit fixa-tion in a model of anterior cruciate ligament reconstruction insheep. Arthroscopy 18:113–123

59. Weiler A, Unterhauser FN, Bail H-J, Huning M, Haas NP(2002) alpha-Smooth muscle actin is expressed by fibroblasticcells of the ovine anterior cruciate ligament and its free tendongraft during remodeling. J Orthop Res 20:310–317

60. Weiler A, Unterhauser FN, Faensen B, Hunt P, Bail HJ, HaasNP (2002) Comparison of tendon-to-bone healing using ex-tracortical and anatomic interference fit fixation of soft tissuegrafts in a sheep model of acl reconstruction. In: Proceedings ofthe 48th Annual Meeting of the Orthopaedic Research Society,Dallas, TX, USA

61. Weiler A, Forster C, Hunt P, Falk R, Jung T, Unterhauser FN,Bergmann V, Schmidmaier G, Haas NP (2003) The influence oflocally applied platelet-derived growth factor-BB on free ten-don graft remodeling after anterior cruciate ligament recon-struction—the use of a biodegradable drug delivery tool in asheep model. Am J Sports Med (in press)

62. Whiston TB, Walmsley R (1960) Some observations on thereaction of bone and tendon after tunneling of bone andinsertion of tendon. J Bone Joint Surg Br 42:377–386

63. Xerogeanes JW, Fox RJ, Takeda Y, Kim HS, Ishibashi Y,Carlin GJ, Woo SL (1998) A functional comparison of animalanterior cruciate ligament models to the human anterior cru-ciate ligament. Ann Biomed Eng 26:345–352

64. Yamakado K, Kitaoka K, Yamada H, Hashiba K, NakamuraR, Tomita K (2002) The influence of mechanical stress on grafthealing in a bone tunnel. Arthroscopy 18:82–90

65. Yamazaki S, Yasuda K, Tomita F, Minami A, Tohyama H(2002) The effect of graft-tunnel diameter disparity on intra-osseous healing of the flexor tendon graft in anterior cruciateligament reconstruction. Am J Sports Med 30:498–505

66. Yasuda K, Tsujino J, Ohkoshi Y, Tanabe Y, Kaneda K (1995)Graft site morbidity with autogenous semitendinosus andgracilis tendons. Am J Sports Med 23:706–714

67. Yoshiya S, Nagano M, Kurosaka M, Muratsu H, Mizuno K(2000) Graft healing in the bone tunnel in anterior cruciateligament reconstruction. Clin Orthop 376:278–286

68. Zimmerman MC, Contiliano JH, Parsons JR, Prewett A,Billotti J (1994) The biomechanics and histopathology ofchemically processed patellar tendon allografts for anteriorcruciate ligament replacement. Am J Sports Med 22:378–386

248