TECHNIQUE OF ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION USING PES ANSERINUS TENDONS

S. PLAWESKI

A. Michallon University Hospital - Grenoble, France

A number of techniques have been found to work well in the reconstruction of a torn anterior cruciate ligament (ACL). We agree with JY Dupont that the chief outcome criterion in the long term is an anatomical one, with "complete restoration of the knee joint." However, in addition to the modified Marshall-MacIntosh and similar techniques for ACL reconstruction, which are known to provide long-term stability, several arthroscopic techniques have been developed, which have the benefit of being minimally invasive and of causing less iatrogenic damage. This, obviously, is a major consideration in ACL surgery. With the development of these techniques, the choice of graft harvesting site has also become an important factor. It should be borne in mind that, in many athletes, the patellofemoral pain following graft harvesting from the extensor mechanism will rule out a return to sports at the previous level. Using the pes anserinus as a donor site appears to be a safer alternative. While the long-term stability outcome of semitendinosus-gracilis (STG) grafts is not yet fully known, there is a current trend to use these pes tendons. With the advent of better arthroscopic aiming and graft fixation systems, an increasing proportion of the ACL reconstruction "market" is being catered for by STG grafts. However, there are pes tendon grafts and pes tendon grafts, as may be seen from the history of the use of these tendons in ACL reconstruction surgery (P Colombet). The 1999 Symposium of the French Society for Arthroscopy also stressed the fact that the many ways in which pes tendon grafts may be performed make it difficult to assess the value of the concept in a multicentre study. In the interest of evaluation, ACL reconstruction with pes anserinus tendons should be performed following certain rules.

PATIENT POSITIONING

This is a crucial aspect of surgical technique, since faulty positioning may result in major technical difficulties.

Care should be taken to ensure that the tourniquet is placed as proximally as possible on the thigh, so as to leave the knee fully mobile. It must be possible to take the knee from full extension into nearly-full flexion. A thigh-holder placed outside the tourniquet, or a foot-rest, will allow the knee to be positioned in an intermediate position of 70° of flexion. A lateral brace is placed at the level of the greater trochanter, to allow valgus stress to be applied during the arthroscopic inspection/treatment of the menisci (Fig. 1).

	Fig. 1 Knee positioning. 1 - foot on foot rest; 2 - thigh holder; 3 - tourniquet on proximal thigh; 4 - lateral brace

ARTHROSCOPY

Once the diagnosis of ACL deficiency has been made (by such means as clinical examination, arthrometer measurements, MRI), the procedure is started by the harvesting of the graft, followed by the arthroscopic inspection of the menisci and treatment of any meniscal and/or cartilaginous lesions encountered. The anterolateral portal for the scope is at the lateral border of the patellar tendon, just distal and lateral to the apex of the patella. This provides a good view of the notch, and in particular of the medial aspect of the lateral condyle; preferably, a scope with a 30° lens should be used.

The anteromedial portal for the instruments should be more distal and more anterior, so as to provide the ideal angle for drilling the femoral tunnel and for inserting the interference screw, if arthroscopic aiming is being used. With unduly posterior placement of this portal, the medial condyle would obstruct the view; with excessively anterior placement, the ligamentum mucosum would get in the way (Fig. 2a and b).

Fig. 2a and b Arthroscopic portals. L = lateral site; M = medial site. Arthroscopic aiming over the medial meniscus.

HARVESTING THE PES TENDONS

This stage of the procedure is a common feature, regardless of the fixation technique to be used. A skin incision is made between 6 and 7 cm distal to the tibiofemoral joint line, and 2 cm medial to the tibial tubercle. The incision is ca. 25 mm long, and requires meticulous haemostasis. Great care is taken not to damage the sensory branch of the saphenous nerve (Fig. 3).

	Fig. 3 Skin incision on anteromedial aspect, 2 cm from the tibial tubercle and 6 cm below the medial joint line

The tendons of of semitendinosus and gracilis (Figs. 4 and 5) are identified under the fascia thus exposed, working from anterior to posterior; once the tendons have been exposed, they are removed using a stripper.

Fig. 4 Knee flexors (source: P. Kamina)	Fig. 5 Pes anserinus tendons (medial view) (Source: P. Kamina)

Preserving the insertions of the tendons makes harvesting easier. In order correctly to identify the course of the tendons at the deep surface of the fascia, a small longitudinal incision should be made very far forward, to partially detach the tibial insertion of sartorius along the anterior aspect of the tibia. The sartorius tendon will be found just in front of and over the tendons of semitendinosus and gracilis, with which it forms a final common tendinous insertion. This allows the tendon sheath of the semitendinosus to be identified. The fascial incision can then be very readily extended horizontally along the upper margin of the semitendinosus. The two tendons will then be visible deep to the fascia. The gracilis, in front of the semitendinosus, is taken on, and firmly pulled upwards with, a right-angled clamp. This will allow part of the tendon to be pulled out of the incision, and to insert the stripper over a distance of ca. 22 cm, to the musculotendinous junction. The semitendinosus is harvested in like manner. This technical detail is of great importance: if the stripper is pushed in without the tendons having been pulled out of the incision, there will be a mjaor risk of the tendon being prematurely amputated half-way along its length. The correct procedure, therefore, consists in "pulling" on the tendon, rather than in pushing on the stripper (Fig. 6). The distal portion of the tendon is detached from its tibial insertions (care being taken to ensure good haemostasis). The two tendons can then be readily removed distally (Fig. 7).

Fig. 6 Stripping of gracilis, with firm countertraction to deliver the tendon out of the incision	Fig. 7 The two tendons have been stripped, and will subsequently be detached from their distal insertions.

Even with the precautions described above, things can go wrong, and part of a tendon may be lost. While this is obviously not ideal, some extra material may be gained by going further down along the tibial periosteal insertion; three strands may be sufficient to fashion a graft with a minimum diameter of 7 mm. If an entire tendon is lost, the intended technique will have to be abandoned. This is why a longitudinal incision is so important: if need be, it can be extended proximally, to allow, for instance, a Jones procedure to be performed. Once the graft has been harvested, it is prepared on the workbench. Any muscle remnants are removed, and the two tendons are laced together with a No. 2 Ethibond suture (Fig. 8). The two tendons thus united are then folded in half, so as to produce four parallel strands at least 9 cm in length (Figs. 9 and 10).

Fig. 8 The tendons have been sutured together.
Fig. 9 Schematic diagram of the graft
Fig. 10 Measuring graft length	Fig. 11 Sizing the four-strand graft

Traction sutures are inserted at either end of the graft. The four-strand graft is sized to a uniform diameter of 7-10 mm (Fig. 11). The total length of the graft should not be less than 85 mm; ideally, it should be 110 mm. The length must be taken into account when fashioning the femoral socket. Once the graft has been prepared in this way, it is wrapped in a moistened sponge, to await further use. On no account should the graft be submerged in saline, since doing so would cause swelling. Some surgeons stitch the tendons together only over the last 2 cm at either end. The question arises whether this will allow better positioning of the bundles as the knee is being mobilized. We think that, given the great elasticity of the pes tendons, stretching during intraoperative mobilization should be avoided. Therefore

- the graft should be cyclically stressed, either on the workbench, after preparation, or following insertion into the knee, prior to tibial fixation; and
- the compliance of the graft should be reduced, by suturing the two limbs together over their entire length. This basting with a nonabsorbable suture will precondition the bundles and thus "stiffen" the graft.

IV ARTHROSCOPIC SURGERY

A - Intra-articular preparation

The remnants of the torn ACL are debrided with a shaver or with basket forceps. At the tibial attachment site, care should be taken to ensure that there is not any bulky ligamentous tissue left standing. However, preserving a little tissue at the native footprint of the ACL may be useful, not only to mark the intra-articular exit site of the tibial tunnel, but also in order to place the graft against tissue containing sensory neurons, in the interest of proprioception. Graft sizing to 8 or 9 mm will allow the tibial attachment fibres of the torn ACL to be preserved. This is an advantage of pes grafts over a Jones procedure, which involves the use of a much wider graft. The notch is debrided with a shaver. We never perform a notchplasty. However, it helps to straighten the medial aspect of the lateral condyle, in a strictly sagittal plane, so as to expose the over-the-top position on the posterior cortical rim at the intercondylar notch. Attention should be paid to the "resident's ridge" two thirds of the way posteriorly, which may cause faulty placement of the femoral socket entry point. The correct site is beyond this ridge, towards the over-the-top position, into which the aiming device is hooked.

The "black hole" of the notch should appear on the screen. Aiming can then be performed to suit the pattern of the patient's knee, at a distance of ca. 5 mm forward from the over-the-top position This so-called isometric point should be at 11 o'clock in the right knee, and at 1 o'clock in the left knee (Fig. 14).

Fig. 14a Frontal view	Fig. 14b Sagittal view
Fig. 14c Femoral aimer (5-7 mm anterior to the over-the-top position)	Fig. 14a-c Femoral socket entry point: isometric position in the notch

Tibial tunnel

The tibial tunnel exit hole inside the joint is not always easy to establish. This is why it is useful to leave the native footprint of the ACL standing. The guide pin cannula of the drill guide is applied to a point ca. 3 to 4 cm below the joint line, medial to the tibial tubercle. The tip of the tibial drill guide is hooked in the posterior fibres of the footprint. The back of the tip should be ca. 2 to 3 mm anterior to the forward-most fibres of the PCL. The device is placed in such a way as to form an angle of about 45 degrees with the long axis of the tibia (Figs. 12 and 13a). Regardless of the device used, the angle between the tibial tunnel and the horizontal arm of the guide should always be set at 55°. This produces optimal guide pin orientation to allow drilling the femoral tunnel through the tibial tunnel, if desired. The pin should protrude a few centimetres into the joint, to mark the femoral socket entry point. An arthroscopic check is made to ensure that the tip of the pin is at the correct site on the femur (Fig. 13b). Some surgeons insist on checking the pin position on the image intensifier. With the knee fully extended, the position of the pin relative to Blumensaat's line is checked, to rule out trochlear impingement. The intersection of Blumensaat's line with the tibial plateau must be at the anterior border of the tibial tunnel. If the tunnel is excessively anterior, there will be graft impingement, which may result in the graft being guillotined or in loss of extension postoperatively.

	m
Fig. 12 Tibial aiming (using the Acufex Direktor)	Fig. 13a Tibial aiming. Angle between horizontal arm and guide pin cannula = 55°. broche guide = guide pin
	Fig. 13b Establishing the femoral socket entry point (knee flexed to 80°)

The tibial tunnel is created using a cannulated reamer matching the diameter of the graft. It is essential to have matching tunnel and graft diameters, so as to prevent graft micromotion inside the bony tunnels.

C - Femoral socket

The blind-end femoral tunnel may be created in two ways - either through the tibial tunnel, which has the advantage of providing an excellent view of the femoral entry site, with the knee flexed to 90 degrees; or independently of the tibial tunnel, with the knee in hyperflexion. In the latter case, the femoral socket would be on a different axis from that of the tibial tunnel. This approach has the disadvantage of giving a less good view of the so-called isometric position on the femur, since the ligamentum mucosum obstructs the drill guide at the back.

A - Arthroscopic aiming The knee is placed in hyperflexion. The femoral aimer is hooked into the over-the-top position. The device comes in a range of offsets (mean, 5-6 mm). The choice will depend on the anatomical pattern of the notch. The guide pin is introduced into the femur, and pushed until the lateral cortex has been pierced. The socket is then reamed to match the diameter of the tibial tunnel. The depth of the socket will be a function of the graft length measured earlier on (mean depth, 35 mm). A traction wire is placed in the slot of the guide wire, which will allow a wire loop to be brought out of the tibial tunnel. This loop will then be used to pull the graft through the tibial tunnel. The graft should move easily, but fit snugly throughout the tibial tunnel. The proximal end of the graft engages into the blind end of the femoral socket (Figs. 21-23).

	Fig. 21 Arthroscopic aiming for femoral socket, through the medial portal, with knee fully flexed
	Fig. 23 Arthroscopic insertion of femoral interference screw (RCI technique)

B - Concentric aiming The axis of the femoral socket is in line with the tibial tunnel axis. For this to be achieved, the knee will need to be flexed to between 70 and 90 degrees. The femoral aimer is introduced through the tibial tunnel, and positioned as described above (ca. 5 mm anterior to the over-the-top position). The reaming of the socket, and the depth attained (35-40 mm), can be readily observed through the arthroscope (Fig. 15).

	Fig. 15 Creating the femoral socket through the tibial tunnel

GRAFT FIXATION

Many different ways of graft fixation have been devised. Rosenberg has shown that a four-strand STG graft that is fixed inside the bone at each end and subjected to tensile stress along its axis will tear in mid-substance in 5% of cases; in 40%, the tear will be at the bony tibial insertion; and in 55%, at the bony femoral insertion. The calculated (theoretical) isometry of the graft cannot be obtained with current ligament reconstruction techniques. Unfortunately, the construct is not very forgiving: if the tunnels are only slightly off the correct isometric position, the substitute ligament may slacken secondarily. The way in which it stretches will be a function of the initial graft fixation strength.

Where the tensile stress exceeds the the yield strength of the fixation system, the graft will loosen. This early laxity is directly related to the fixation system used. If, on the other hand, the systerm is not mechanically stressed beyond its initial failure strength, further stress may cause gradual stretching of the graft fibres. Some authors have shown that initial graft micromotion in the tunnels occurs in 90% of the reconstructions; this is thought to be due to the fact that ideal isometry was not obtained. These findings show the importance of the two fundamental features of knee ligament reconstructions: isometry, and graft fixation.

Femoral fixation:Three main meansof fixation (Fig. 16)

	Fig. 16 Arthrex femoral guide and marking hook

(1) - Systems with purely cortical support, the best-known of which is the EndoButton®. This system consists of a metal device that is placed flat against the anterolateral cortex of the lateral condyle. The graft is attached to the EndoButton by a 5-mm polyester tape or by a braided suture of the surgeon's choice. Two-limb tendon grafts may be fixed with one EndoButton, or have each limb fixed with a separate button. The tensile strength of these devices has been tested and found to be reasonably satisfactory; however, correct positioning of the button on the lateral cortex is not easy to achieve, and requires accurate calculation of the total length of the femoral tunnel.

The theoretical advantage of cortical fixation is offset by the technical problems at surgery, and by worries about the strength of the linkage material between the graft and the button. Another mechanical disadvantage is the increase in the total length of the graft that will be subject to strain (since the proximal fixation site will be 8-10 cm away from the joint surface). This is why surgeons using an EndoButton would be well advised to fully stabilize the graft in the tunnel by inserting a supplementary interference screw into the femoral socket.

	Fig. 20 TransFix® (Arthrex) femoral fixation system and tibial interference screw

(2) - Transfixion devices, such as the TransFix® (Fig. 20). The TransFix technique involves the insertion of a metal implant into the lateral condyle, strictly perpendicular to the axis of the femoral socket. The device passes between the two limbs of the graft. Once the femoral socket has been fashioned with concentric reaming, a marking hook, which is firmly mounted on the femoral guide, is introduced. The hook (which matches the socket diameter) is pushed home in the socket. A guide pin is drilled through the distal femur, from lateral to medial. This pin passes through an opening in the marking hook, in the depth of the femoral socket (Figs. 16 and 17a). The guide pin is replaced by a guide wire, which is brought out through the tibial tunnel (Fig. 17b). The graft is loaded into the wire loop. Pulling on the two ends of the wire will take the graft through the tibial tunnel and into the femoral socket (Fig. 18). The TransFix implant is slid over the wire, through the axilla of the graft, and impacted into the lateral femoral cortex (Figs. 19 and 20). The implant must not sit proud laterally, since hardware protrusion may cause soft-tissue irritation. The advantage of this mode of fixation is that it produces ideal tensile-strength conditions. However, there are two disadvantages: the technique is complex, and requires a learning period; and - as with the technique described above - the distance of the fixation construct from the joint surface increases the total length of the graft that may undergo strain.

Fig. 16 Arthrex femoral guide and marking hook	Fig. 17a A guide pin is drilled through the distal femur.
Fig. 17b The pin is replaced by a wire, which is brought out through the tibial tunnel.	Fig. 18 The graft is pulled up into the tibial tunnel.
Fig. 19 The TransFix implant is impacted. fil de traction du transplant = graft traction sutures

(3) - Interference systems (interference screws). Many interference screw patterns have been devised and are commercially available.
The RCI screw has a blunt thread, which protects against graft damage, and a round head, which minimizes the stress on the graft fibres at the femoral socket entrance (Fig. 22). The screw has a 2.4 mm cannulation, which accommodates the largest-diameter guide pin, thus facilitating arthroscopic screw insertion (Fig. 23).

Fig. 22 Femoral tunnel positioned independently of tibial tunnel. Graft pulled through in two stages

	Fig. 23 Arthroscopic insertion of femoral interference screw (RCI technique)

It comes in a range of diameters, from 7 mm to 9 mm, several lengths (from 25 mm to 40 mm), and with normal or reverse thread. Pinczewski recommends a standard diameter of 7 mm for the femoral screw (or of 8 mm, in poor bone stock or where the socket has been made too wide in relation to the graft diameter). For a socket diameter of 7 mm or 8 mm, a 7-mm screw should be used; for a 9-mm or 10-mm socket, the preferred screw diameter would be greater (8 mm or 9 mm). Since the interference screw must be placed anterior to the graft, with the graft residing against the posterior wall, Pinczewski has developed a reverse-thread screw (reverse RCI), for use in the right knee of patients. With this hardware, he observed significantly less graft stretching. Currently, the screw is made of titanium; it will soon be available in a polylactide PLA 100 version. There are many other metal or bioabsorbable screws on the market. At our present state of knowledge, and in the light of our clinical experience with PLA 98 screws, we would recommend the use of polylactide screws. However, these results were reported in bone-to-bone fixation studies (Jones procedure). The long-term behaviour of bone-tendon fixation constructs using polylactide interference screws is as yet unkown. This aspect will need to be closely monitored; studies of direct interference screw fixation on the femoral side using PLA interference screws are ongoing.

Tibial fixation

	Fig. 24 Femoral and tibial interference screws fil de traction du transplant = graft traction sutures

(a) - Interference screws For the tibial side, too, many screw patterns are on offer (Fig. 24). We prefer a PLA 98 screw of a diameter matching that of the tibial tunnel. PLA 98 is resorbed slowly, which permits excellent primary fixation to be obtained, and ensures sound secondary fixation up to and beyond the stage of ingrowth. In our study, no granuloma formation or resorption products were seen at 3 years' follow-up. Complementary fixation (with a cortical staple) may be considered. In a study of female patients with poor bone stock, Pinczewski showed significantly less residual stretch in the patients managed with a screw and staple, as compared with those who had only an interference screw inserted (Figs. 25-27).

Fig. 25 Radiographic check of correct graft positioning, using radiological criteria (Blumensaat's line, Aglietti index, etc.)	Fig. 26 Check radiograph (4-strand STG graft fixed with TransFix and a Phusis tibial screw)
Fig. 27a and b Postoperative check radiographs (TransFix + tibial RCI screw)

(b) - Tibial anchors, isolated tibial staples These devices have the disadvantage of lesser pull-out strength and a greater graft length subject to strain (bony fixation site well away from the exit of the tibial tunnel in the joint).

Technical variants: The femoral tunnel may be created using an outside-in technique; the tendons may be left attached at their distal insertion. Other femoral and tibial fixation systems have been devised: Rigid-fix®, Endofix®, LinX HT®, spiked washers, belt-buckle, anchors, etc.).

The choice of fixation technique will depend upon a number of considerations:
- Fixation should be as close as possible to the joint line, to prevent a windshield wiper effect causing gradual tunnel expansion, and a bungee effect , as a result of longitudinal graft-tunnel motion. Correctly sized interference screws may be a solution. In the tibia, these devices are commonly used; in fact, some manufacturers have developed new systems using more than one screw (Tri-Cortical™). In the femur, finding a good fixation system is more of a problem: interference screws meet the requirement for fixation close to the joint line; however, mechanical studies have shown these screws to have less pull-out strength on their own, as compared with their use in conjunction with other fixation devices. A hybrid construct may be ideal (e.g. EndoButton + interference screw, or TransFix + interference screw).

Graft isometry is tested , prior to tibial graft fixation, by measuring the change in length of the substitute ACL while ranging the knee from extension to 90 degrees of flexion. The graft is mulitply cycled, to prevent subsequent stretching. The tibial fastener is inserted with the knee in slight (30°-40°) flexion; the tibial approach is closed in layers over a suction drain; and the arthroscopic portals are closed.

POSTOPERATIVE MANAGEMENT

In the immediate postoperative period, the patient will be on an analgesic and anti-inflammatory regimen (femoral nerve block, drugs). A splint may be used, for pain relief only. Active extension exercises are started the day after surgery, as is walking with weight-bearing, using crutches that will be discontinued as soon as good knee locking has been obtained (between two weeks to a month postoperatively).

REHABILITATION

The rehabilitation protocol is aimed at

- Full active extension: electrostimulation, relief of graft harvest site pain (which may cause an early postoperative flexion contracture).

- During the healing phase, the hamstrings should not be worked with active concentric resisted exercises (which may cause muscle sprains).

- Patellar mobilization is performed immediately. In patients with excessive knee laxity, exercises involving hyperextension may cause a recurvatum deformity, and should be avoided.
Proprioceptive training is started in the third week, using closed kinetic chain exercises, with isometric cocontraction of the quadriceps, hamstrings, and gastrocnemius.

- The gradual return to dynamic work is achieved through lower-limb closed-chain exercises. Restoration of exercise capacity starts at 6-7 weeks, with aquatic exercises and cycling on firm ground. A return to non-pivot, non-contact sports is allowed at between 2 and 3 months.

CONCLUSION

Four-strand STG grafts are being increasingly used in clinical practice. At the 1999 Symposium of the French Society for Arthroscopy, the following indications were established:

- quadriceps-dominant sports;
- kneeling jobs;
- revision of failed reconstructions using patellar tendon;
- anterior knee pain;
- patients > 40 years of age.
The recognized contraindications are:

- prior surgery of the medial soft tissues;
- contact sports;
- major anterior instability (more than 10-mm difference in anterior translation between the affected and the well knee);
- excessive knee laxity;
- female patients practising little or no sport.

Other clinical studies are ongoing. The important aspect of this evaluation will be the collection of so-called anatomical results, and the creation of homogeneous series. With strict attention to the details of surgical technique, and with the use of a fixation system of proven efficacy, more patients should be allowed to benefit from this form of ACL reconstruction.

(Transl KRMB)

References

- AGLIETTI P., BUZZI R., D'ANDRIA S., ZACCHEROTTI G. : Patellofemoral problems after intraarticular anterior cruciate ligament reconstruction. Clin Orthop. 1993. 288 195-204. - BAKER L. : Effets d'une fixation tibiale résorbable supplémentaire dans la reconstruction du LCA au DIDT 4 faisceaux : propriétés biomécaniques initiales. Hughston sports medicine Foundation, Inc - CHRISTEL P. : Ligamentoplastie du LCA avec les tendons du droit interne et du demi tendineux. AOLF (7ème Congrès) - Beyrouth - Mai 2000 - CLEMENS M. MERJANEH G. : The transfix ACL reconstruction, a new procedure. E.Cavell Medical Institute. Bruxelles - COLLOMBET PH. Et coll. L'histoire de la chirurgie du LCA. Maîtrise Orthopédique N87 - Octobre 1999 - DUPONT J.Y. Résultats à 7 ans de l'opération de Marshall Mac Intosh modifiée dans le traitement des laxités antérieures du genou. (débat sur les ligamentoplasties antérieures du genou). Annales Orthopédiques de l'ouest 1998. - FRANCESCHI J.P., POTEL et coll. : Reconstruction du LCA : Tendon rotulien versus patte d'oie. Une étude multicentrique de la Société Francaise d'Arthroscopie. Paris, Déc .99 - ISHIBASHI Y ET COLL. : The effect of anterior cruciate ligament graft fixation site at the tibia on knee stability : evaluation using a robotic testing system, Arthroscopy 13(2): 177-182,1997 - HOWELL M. et HULL M. : Structural properties of six tibial fixation méthods for anterieur cruciate ligament tissu graft. Am J Sports, volume 27, N°1, 1999. - MORGAN CD, KALMAN VR, GRAWL DM : isometry testing for anterior cruciate ligament reconstruction revisited, Arthroscopy 11(6) :647-659, 1995 - NOYES FR et coll. : Biomechanical analysis of human ligament grafts used in knee ligament repairs and reconstructions. J Bone Joint Surg 66A : 344-352, 1984 - PINCZEWSKI L. et coll. : Integration of hamstrings tendon graft with bone in reconstruction of antérieur cruciate ligament. Arthroscopy.13 : 641-643 ;1997 - PINCZEWSKI L. and coll : Results of femoral fixation in ACL reconstruction with hamstrings graft using a reversed interference screw. (à paraître : Am J Sports Med ; octobre 2000) - STÄHELIN AC et coll. : Clinical degradation and biocompatibility of different bioabsorbable interference screws : A report of six cases, Arthroscopy 13(2) :238-244,1997. - WEILER A. et coll. : Anatomic, semianatomic and extraarticular hamstring tendon graft fixation techniques for ACL reconstruction under incremental cyclic loading conditions. 8th Congress of ESSKA, April 29-May 2, 1998 - WEILER A. et coll. : Biodegradable interference screw fixation exhibits pull-out force and stiffness similar to titanium screws. Am J Sports Med 26 : 119-128, 1998 - WEILER A. et coll. : Hamstring tendon fixation using interference screws : A biomechanical study in calf tibial bone. Arthroscopy 14 : 29-37,1998 - WEILER A. : Direct tendon graft fixation in ACL reconstruction using interference screws - experimental study in sheep - 8th Congress of ESSKA, April 29-May 2, 1998 - WEILER A. et coll. : Semitendinosus graft fixation with bioabsorbable interference screws, 43rd annual meeting, Orthopaedic Research Society, San Francisco, California, February 9-13, 1997