The Knee 12 (2005) 370 – 376 www.elsevier.com/locate/knee

The effect of screw taper on interference fit during load to failure at the soft tissue/bone interface Charles J. Manna,T, John J. Costib, Richard M. Stanleyb, Peter J. Dobsona b

a Wakefield Orthopaedic Clinic, Wakefield Street, Adelaide 5000, South Australia, Australia Department Of Orthopaedics, Repatriation General Hospital And Flinders University, Adelaide 5041, South Australia, Australia

Received 19 October 2004; accepted 21 December 2004

Abstract The effect of screw geometry on the pullout strength of an anterior cruciate ligament reconstruction is well documented. The effect of a truly tapered screw has not been previously investigated. Thirty bovine knees in right and left knee pairs were collected. Superficial digital flexors from the hind legs of sheep were harvested to form a quadruple tendon graft. For each knee pair, one tendon graft was fixed using a tapered screw (n=15) and the other with a non-tapered screw (n=15). Interference screws were manufactured from stainless steel, and apart from the tapered or non-tapered profile were identical. The screws were inserted into a tibial tunnel already containing the tendon graft. The interference fit was tested by extensile load to failure tests. The insertion torque of the screws and first sign of load to failure (by pullout) of the interference fit were recorded. Results were analysed using paired t-tests. The results indicated that tapered screws have significantly higher resistance to interference failure ( p=0.007) and insertion torque ( pb0.001) than non-tapered screws. The improved biomechanical performance of tapered screws demonstrated in this study may translate into superior clinical results, particularly at the tibial attachment of hamstring anterior cruciate ligament reconstruction, and also of hamstring fixation to the medial femoral condyle for patella instability. D 2005 Elsevier B.V. All rights reserved. Keywords: Cruciate reconstruction; Interference screw; Pullout strength; Insertion torque; Load to failure

1. Introduction The long term results and relative merits of hamstring and patella tendon grafts for anterior cruciate ligament (ACL) reconstruction are well documented [1–10]. ACL reconstruction failures are typically due to interference failure rather than fixation failure or failure of the screw itself. Any design feature which improves the strength of the interference fit should reduce the number of clinical failures not only of ACL reconstruction, but also of patella T Corresponding author. The Norfolk and Norwich University Hospital, Colney Lane, Norwich NR4 7UY, UK. Tel.: +44 1603 289106. E-mail addresses: [email protected], [email protected] (C.J. Mann). 0968-0160/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.knee.2004.12.007

stabilisation procedures where interference screws are used to anchor a semitendinosus graft in a bony tunnel in the medial femoral condyle. It has been shown for hamstring grafts used for ACL reconstruction that fixation is weakest at the tibial tunnel [11]. There are many different types of screw available for hamstring and patella tendon graft fixation, with wide variations in core diameter, thread depth, pitch and taper. Differences in screw geometry, particularly length and diameter can have a significant effect on performance, as has been shown in several studies [12–14]. Tapered screws are associated with higher insertion torque than non-tapered screws, and higher insertion torque has been shown to be predictive of higher ACL graft load to failure resistance [15,16].

C.J. Mann et al. / The Knee 12 (2005) 370–376

A truly tapered screw has a constant taper along its entire length. However, it is rare for an interference screw to have a true taper as the btaperedQ interference screws currently used in clinical practice are only tapered at the tip to facilitate insertion. If tapered screws were associated with increased pullout strength, this may support the use of these screws for ACL reconstruction in humans. Therefore, the aim of this study was to determine whether screw taper is an important factor affecting both load to failure and insertion torque of interference screws.

2. Materials and methods 2.1. Screw design and manufacture Tapered and non-tapered screws (Fig. 1) were designed based on commercially available interference screws. The length (30 mm), pitch (2 mm) and thread depth (1 mm) were identical for both non-tapered and tapered screws. The nontapered screw had an outer diameter of 8 mm. The tapered screw had a consistent taper along the entire length, starting from a diameter of 9 mm and reducing to 7 mm. This gradient along the tapered screw is equivalent to a taper angle of 3.88 along both sides of the screw (tan 12/30). Screws were manufactured by Austofix (Australian Orthopaedic Fixations Pty, Adelaide, Australia) from 316L surgical grade stainless steel. They were polished in a vibratory machine with ceramic media to round-off sharp edges on the thread profile and produce a standardised surface finish. 2.2. Specimen collection All bovine specimens were collected from an abattoir. Thirty knees (in pairs) were dissected from bovine carcasses within 30 min of sacrifice. Animals were at least 18 months old. The distal femur and proximal tibia, including the knee joint, were stripped of soft tissue at harvest, wrapped in saline soaked gauze and frozen at 20 8C.

371

One hundred superficial digital flexor tendons were harvested from the hind limbs of freshly killed sheep using a tendon stripper and stored as above. Only tendons of at least 12 cm in length were accepted. Before testing, knee pairs and tendons were defrosted overnight. All soft tissue surrounding the knees was removed, leaving the femur–ACL–tibia complex intact. Knees and tendons were then placed and subsequently tested in a 0.9% saline bath at 37 8C. All 15 knee pairs underwent distraction testing of the intact bovine ACL followed by load to failure tests of the ovine tendons fixed by a tapered or non-tapered screw in a bovine tibial tunnel. Knees were randomised in pairs to receive either the non-tapered or tapered screw to anchor the ovine tendon graft in the bovine tibial tunnel. The knees were harvested in pairs so that each sided knee would act as a control for the other side to reduce any effect of differences in bone density between non-paired knees. 2.3. Intact ACL failure strength All biomechanical testing was conducted in an Instron model 8511 servo-hydraulic materials testing machine (Instron, High Wycombe, UK). To measure the failure strength of the intact ACL an apparatus was manufactured from stainless steel to allow rigid fixation of the distal femur and proximal tibia. A ball joint was fixed distally between the load cell and the tibial mounting cup and a horizontal x– y bearing table was placed proximally between the femoral cup and actuator to remove all shear forces and allow the ligament to self align along its long axis (Fig. 2). Each bone was fixed to the apparatus using transfixing pins and set screws. The construct was preconditioned with loads of 10– 110 N applied sinusoidally at 0.1 Hz over 20 cycles. The intact bovine knee was then tested to failure at a distraction rate of 20 mm/min. Load and displacement data were recorded on a personal computer and the mode of failure determined by visual inspection. The distraction of the intact bovine ACL would provide additional confirmation that there was no difference between the right and left knee pairs. 2.4. Preparation of ACL reconstruction and measurement of screw insertion torque

Fig. 1. Photo of the non-tapered (top) and tapered (bottom) screws.

After the intact ACL failure tests, tendon reconstructions were performed on the proximal tibia. Tendons were combined in pairs to give a graft diameter of 7 mm, measured using an Arthrex graft sizer (Arthrex, Naples, FL, USA), and whip-stitched using 0/vicryl (Ethicon, Piscataway, NJ, USA). The entry point for the tibial tunnel was stripped of periosteum to allow graft and screw insertion. A natural depression adjacent to the tibial tubercle in bovine knees was used for the entry point. An Arthrex tibial tunnel jig was set to allow the exit point to be on the intercondylar

372

C.J. Mann et al. / The Knee 12 (2005) 370–376

Instron Actuator

X-Y bearing table

femur

tibia

saline bath

steel bar, used during distraction testing. The distal end of the graft exiting the entry point of the tibial tunnel was attached to a hand held tensioner which allowed for easier screw insertion and allowed the tendons to be evenly tensioned whilst screw insertion was performed. The tensioner also maintained the separation between each of the four tendons in the quadruple graft which facilitated screw insertion. A tapered or non-tapered screw was subsequently inserted into the bony tunnel containing the quadruple tendon graft. Screw insertion torque was measured using an instrumented hand-held torque screwdriver (Rumul, Germany) calibrated from 12 to +12 Nm with a precision of F0.05 Nm. Signal conditioning and excitation were provided by a Unimeter XQL (Autoplex International Pty, Adelaide, Australia) with an analog output of 0–10 V. The analog output signal was collected using a 12-bit data acquisition card (DAQ 6024E, National Instruments, Austin, TX, USA) and personal computer, allowing real-time display of the generated torque.

Fig. 2. Test set-up for intact bovine ACL distraction.

2.5. ACL reconstruction pullout strength area of the tibial plateau. A guide wire was then drilled retrograde through the jig and the proximal tibia. The angle selector on the jig was set to 458, as this allowed for easier orientation of the proximal tibia in the Instron machine. The jig was then disassembled. A cannulated drill was placed over the guide wire and the tunnel drilled. One knee in each pair was randomly allocated either a tapered or non-tapered screw. This was done by blind selection. The other knee in the pair received the alternative screw type to be tested. Thus each time a screw was randomly selected the odds of choosing either a tapered or non-tapered screw were the same. This was because the number of non-tapered and tapered screws remaining was always the same. For non-tapered screws the tibial tunnel was created initially using a 7 mm diameter drill. This was followed by a 9 mm diameter drill. This resulted in a 1 mm clearance between the screw and tunnel. For tapered screws, the same 7 mm starting drill was used which was followed by an 8 mm drill. A tapered punch (Fig. 3) was then inserted to dilate the tunnel to 10 mm at the entry point and 8 mm at the end of the punch. The punch was marked to allow the punched hole to be exactly the same length as the tapered screw. The punch was used so that there was a clearance of 1 mm between screw and tunnel over the length of the tapered screw in order to match the same screw/tunnel clearance created for the non-tapered screw. Paired tendons were combined and looped together to create a four-strand graft of 7 mm diameter. A suture was looped around the apex of the four-strand graft, allowing the graft to be pulled retrograde through the tunnel by hand. The apex of the loop of tendon which had emerged from the tibial plateau was placed around a cylindrical

Grafted specimens were placed in a saline bath for at least half an hour before testing. All testings were performed in-line with the long axis of the tibial tunnel to simulate a worst-case scenario (Fig. 4). The 13 mm diameter cylindrical bar was fixed to an x–y bearing system, which in turn was fixed to the Instron actuator. This ensured that there were no horizontal loads passing through the graft during testing. All grafts were preconditioned and tested to failure in the same manner as the intact ACL failure tests. Graft failure was defined as the first sign of graft pullout. The pullout strength was calculated using linear regression and was defined as the load at which there was a significant deviation from linearity in the elastic region of the loading curve. This deviation was based on the change in the R 2 value of the linear regression. The load where R 2 decreased from 0.99 to 0.98 was defined as the pullout strength. The pullout strength was a more accurate reflection of graft fixation failure when compared to the

Fig. 3. Tapered punch.

C.J. Mann et al. / The Knee 12 (2005) 370–376 Instron Actuator

373

forearm software in 15 cm of water. One third of the sample size was considered sufficient to test as the source, age and nutrition of all of the animals was the same. Two separate 1.0-cm2 areas on either side of the tibial spine, distal to the epiphyseal plate, were selected. The midline was avoided, as this was where the tibial tunnel was drilled.

X-Y bearing table Proximal graft fixation over cylindrical bar

2.7. Statistical analysis All data are shown as meanFstandard deviation. Statistical analysis was performed using paired samples ttests, using SPSS software (SPSS, Chicago, IL, USA) on a personal computer. PV0.05 was considered significant.

Fixation pin

Ball joint

Instron load cell

3. Results

Fig. 4. Test setup for reconstruction distraction testing. The water bath has been omitted for clarity.

maximum failure strength and provided a standardized and reliable method across all specimens (Fig. 5). The maximum failure strength was not used as it may be artificially increased by tendon bunching and does not represent the point at which interference fixation fails. This point is more accurately represented by the point at which the loading curve deviates from linearity. 2.6. Bone densitometry The bone mineral density of five pairs of bovine knees was determined using a Lunar densitometer (Model #1107, GE Medical Systems, Waukesha, WI, USA) using

Table 1 summarises the results for each knee pair. The mode of failure for the intact ACL tests was due to tibial avulsion in all cases. Although the maximum load to failure was recorded, it was evident from the load versus displacement graphs and visual inspection of the ligaments that some of the fibres of the ACL were failing before the maximum load was reached. The failure strength of left knees was 3263F596 N compared to 3034F593 N for right knees (n=14 pairs), which were not significantly different ( p=0.14). Bone mineral density scans of left and right knees were not significantly different (2.02F0.17 g cm 2 and 1.97F0.18 g cm 2 respectively. p=0.58. n=5 pairs). The insertion torque for tapered screws was significantly higher than non-tapered screws (2.8F0.6 Nm and 1.7F0.6 Nm respectively, pb0.001, n=13 pairs). The mode of graft failure was by interference failure in all cases. The grafts using tapered screws failed at significantly higher loads compared to the grafts using

1000 Maximum failure strength 900 800 700

Load (N)

600 500 400 300 200

First sign of graft pullout (R-squared = 0.98)

100 Elastic region (R-squared = 0.99) 0 0

2

4

6

8

10

12

14

Displacement (mm)

Fig. 5. ACL reconstruction failure graph.

16

18

20

374

C.J. Mann et al. / The Knee 12 (2005) 370–376

Table 1 Summary of results for each knee pair for intact ACL load to failure, bone mineral density (BMD), insertion torque and pullout strengths for the non-tapered and tapered screw ACL reconstructions Specimen

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Mean S.D. P value

Left intact ACL (N)

Right intact ACL (N)

2118 3914 2876 4093 3633 2697 3770 2986 3431 3532 2661 2677 3925 3374

2297 3388 2489 3533 3579 3190 3824 3255 2307 3868 2018 3112 2830 2785

3263 596 0.14

3034 593

Left BMD (g/cm2)

Right BMD (g/cm2)

1.673 2.160 1.964 1.995 2.042

1.817 1.945 2.149 1.930 2.235

1.967 0.180 0.58

2.015 0.172

Non-tapered screw insertion torque (Nm)

T T 0.8 0.7 2.6 1.9 1.9 1.4 2.9 1.8 1.6 1.8 1.7 1.0 1.4 1.7 0.6 b0.001TT

Tapered screw insertion torque (Nm)

Non-tapered graft pullout (N)

Tapered graft pullout (N)

3.2 4.1 4 2.2 2.7 3.2 3.0 2.3 3.2 2.4 2.6 2.1 2.6 2.1 3.0 2.8 0.6

266.4 364.0 224.0 210.8 181.8 378.5 393.5 335.9 290.1 359.7 184.5 318.8 212.3 132.7 271.8 275.0 81.8 0.007TT

190.9 639.8 372.2 290.6 446.0 496.2 345.3 254.6 374.0 483.9 400.7 225.3 556.1 349.7 385.8 387.4 123.2

T Data not recorded due to data acquisition technical problem. TT Significant difference at the 0.05 level.

non-tapered screws (387.4F123.2 N versus 275.0F81.8 N respectively, p=0.007, n=15 pairs).

4. Discussion The load to failure of the intact bovine ACL was high, up to 4000 N. This is markedly higher than the load to failure of the human ACL, which is 1730–2200 N [17,18]. The mechanism of failure in each case was avulsion at the ligament/bone interface at the tibial end, which happened as sequential failure. This mechanism of failure may reflect the significant resistance to distraction of the bovine ACL itself, despite anchorage in bone which is significantly denser than human bone, and particularly cadaveric human bone. However, age is known to have an effect on the biomechanical properties of human ligaments; younger donors demonstrate ligaments that are stiffer and have increased ultimate load to failure [19]. The tests indicated that there was no difference in distraction testing between right and left knees, validating the use of the paired knee design for the subsequent load to failure tests of the tendon grafts. The paired design was further supported by the similarity in bone density between the right and left knees (1.97 g cm 2 and 2.02 g cm 2 respectively). The significantly increased load to failure strength of the tendon graft with the tapered screw indicated that tapered

screws show improved biomechanical performance in the laboratory compared to non-tapered screws in a bovine model. The results of ACL reconstruction are generally very satisfactory using non-tapered interference screws [20]. Hamstring as opposed to patella tendon reconstruction has many advantages, but the integrity of tibial fixation has often been questioned. Commercially available screws lead to slightly increased pullout strengths when compared to our tapered screws [21–23], but this is not likely to be significant, particularly when some of these studies have used the maximum failure strength as opposed to the first deviation from linearity during the load to failure tests as used in this study. The maximum failure strength may be artificially high due to progressive gathering and bunching of the tendons in the tunnel once slippage has occurred. In addition, the deviation from the linear elastic region is a well-recognised method for testing those materials that do not exhibit a clear failure point. Both tapered and nontapered screws, on average, failed at loads significantly above 250 N, a load estimated to simulate aggressive rehabilitation [23] which would permit the use of these screws in a clinical setting as part of a clinical trial. The higher pull-out strengths and insertion torques for the tapered screws occurred despite identical clearances between tunnel and screw. This may be due to the depth of insertion of the screws not being standardised, although both sets of screws were inserted until they were at least flush with the bone surface and tightened to a degree similar

C.J. Mann et al. / The Knee 12 (2005) 370–376

to that in a clinical setting. The same person inserted all the screws, and was blinded to the insertion torque measurement during insertion of the screw. While the depth of insertion may not be critical for a non-tapered screw in a non-tapered drilled tunnel, it is critical for a tapered screw whether in a tapered or non-tapered tunnel. Advancing a tapered screw further would result in higher insertion torque due to the wedging action of the taper. This problem was circumvented to some extent by having a tapered punch of exactly the same length as the tapered screw. Sectioning of some tendons was a problem in early pilot studies when a non-tapered hole was used for the tapered screws. Thus tapered screws were used with a tapered tunnel. The only difference between the tapered and non-tapered screws, other than the screws themselves, is the additional use of the punch when completing the tapered tunnel. It is possible that by using the punch, the bone adjacent to the tunnel was partially compacted, providing superior purchase for the screw threads. Recent studies have shown that compaction made no difference to pullout strengths in porcine and human bone, although this has not been shown in denser bovine bone or for punched or tapered tunnels [24,25]. The use of a bovine model is a limitation of this study as is the use of any animal model. Bovine bone, being much denser than human bone, may provide superior purchase for interference fixation. However the bovine bone used in this study provided a uniform medium to investigate the initial fixation effects of both screw types, and the bone was equally dense in right and left knees. The density of cadaveric human bone may not reflect the density of bone of those in whom ACL reconstruction is usually performed. We tested our specimens using a one-off pullout test, which may not represent how reconstructed ACL grafts fail in vivo. However, as a first step in testing tapered screws we wished to investigate purely the biomechanical performance of tapered screws. Cyclical distraction testing was not considered as the mechanism of interference failure of hamstring tendon fixation may not be the same for patella instability surgery as for ACL reconstructions in vivo. For future tests, it may be appropriate to cycle the graft to failure over a clinically representative number of cycles with gradually increasing tensile load, as this may more accurately reproduce the failure mechanism in vivo. The results of this study show that tapered screws, using the technique described above provides superior initial fixation and insertion torque compared to non-tapered screws. The improved results for tapered screws may reflect the technique itself rather than the screw. The effect of bone compaction using tapered screws in bovine bone is the subject of future research.

Acknowledgements The authors would like to thank the following for their assistance: Mr. Richard Clarnette, Mr. David Campbell and

375

Mr. Kevin Angel of Wakefield Orthopaedic Clinic, Adelaide, South Australia. Dr. Chris Schultz of the Nuclear Medicine and Bone Densitrometry Department Royal Adelaide Hospital, South Australia. Dr. Wayne Rankin of the Department of Orthopaedics of the Repatriation General Hospital, Adelaide, South Australia. The staff of the Institute of Medical and Veterinary Science, Adelaide, South Australia.

References [1] Bartlett RJ, Clatworthy MG, Nguyen TN. Graft selection in reconstruction of the anterior cruciate ligament. J Bone Joint Surg Br 2001;83:625 – 34. [2] Pinczewski LA, Deehan DJ, Salmon LJ, Russell VJ, Clingeleffer A. A five-year comparison of patellar tendon versus four-strand hamstring tendon autograft for arthroscopic reconstruction of the anterior cruciate ligament. Am J Sports Med 2002;30:523 – 36. [3] Witvrouw E, Bellemans J, Verdonk R, Cambier D, Coorevits P, Almqvist F. Patellar tendon vs. doubled semitendinosus and gracilis tendon for anterior cruciate ligament reconstruction. Int Orthop 2001;25:308 – 11. [4] Beynnon BD, Johnson RJ, Fleming BC, Kannus P, Kaplan M, Samani J, et al. Anterior cruciate ligament replacement: comparison of bone– patellar tendon–bone grafts with two-strand hamstring grafts. A prospective, randomized study. J Bone Joint Surg Am 2002;84A:1503 – 13. [5] Yunes M, Richmond JC, Engels EA, Pinczewski LA. Patellar versus hamstring tendons in anterior cruciate ligament reconstruction: a metaanalysis. Arthroscopy 2001;17:248 – 57. [6] Corry IS, Webb JM, Clingeleffer AJ, Pinczewski LA. Arthroscopic reconstruction of the anterior cruciate ligament. A comparison of patellar tendon autograft and four-strand hamstring tendon autograft. Am J Sports Med 1999;27:444 – 54. [7] Anderson AF, Snyder RB, Lipscomb Jr AB. Anterior cruciate ligament reconstruction. A prospective randomized study of three surgical methods. Am J Sports Med 2001;29:272 – 9. [8] Aglietti P, Buzzi R, Zaccherotti G, De Biase P. Patellar tendon versus doubled semitendinosus and gracilis tendons for anterior cruciate ligament reconstruction. Am J Sports Med 1994;22:211 – 7. [9] Eriksson K, Anderberg P, Hamberg P, Lofgren AC, Bredenberg M, Westman I, et al. A comparison of quadruple semitendinosus and patellar tendon grafts in reconstruction of the anterior cruciate ligament. J Bone Joint Surg Br 2001;83:348 – 54. [10] Barrett GR, Noojin FK, Hartzog CW, Nash CR. Reconstruction of the anterior cruciate ligament in females: a comparison of hamstring versus patellar tendon autograft. Arthroscopy 2002;18:46 – 54. [11] Bach BR. Knee and leg: soft tissue trauma. In: Kasser JR, editor. Orthopaedic Knowledge Update. Rosemont7 AAOS, 1996. p. 463 – 76. [12] Weiler A, Hoffmann RF, Siepe CJ, Kolbeck SF, Sudkamp NP. The influence of screw geometry on hamstring tendon interference fit fixation. Am J Sports Med 2000;28:356 – 9. [13] Selby JB, Johnson DL, Hester P, Caborn DN. Effect of screw length on bioabsorbable interference screw fixation in a tibial bone tunnel. Am J Sports Med 2001;29:614 – 9. [14] Pomeroy G, Baltz M, Pierz K, Nowak M, Post W, Fulkerson JP. The effects of bone plug length and screw diameter on the holding strength of bone–tendon–bone grafts. Arthroscopy 1998;14:148 – 52. [15] Zdeblick TA, Kunz DN, Cooke ME, McCabe R. Pedicle screw pullout strength. Correlation with insertional torque. Spine 1993;18:1673 – 6. [16] Brand Jr JC, Pienkowski D, Steenlage E, Hamilton D, Johnson DL, Caborn DN. Interference screw fixation strength of a quadrupled hamstring tendon graft is directly related to bone mineral density and insertion torque. Am J Sports Med 2000;28:705 – 10.

376

C.J. Mann et al. / The Knee 12 (2005) 370–376

[17] Noyes FR, Butler DL, Paulos LE, Grood ES. Intra-articular cruciate reconstruction: I. Perspectives on graft strength, vascularization, and immediate motion after replacement. Clin Orthop 1983:71 – 7. [18] Miller MD. Sports medicine. In: Miller MD, Brinker MR, editors. Review of Orthopaedics. Philadelphia7 W.B. Saunders Company; 2000. p. 195 – 240. [19] Spindler KP, Wright RW. Soft tissue physiology and repair. In: Beaty JH, editor. Orthopaedic Knowledge Update. Rosemont7 AAOS; p. 3 – 18. [20] Albright JC, Carpenter JE, Graf BK, Richmond JC. Knee and leg: soft tissue trauma. In: Beaty JH, editor. Orthopaedic Knowledge Update. Rosemont7 AAOS; 1999. p. 533 – 57. [21] Weiler A, Hoffmann RF, Stahelin AC, Bail HJ, Siepe CJ, Sudkamp NP. Hamstring tendon fixation using interference screws: a biomechanical study in calf tibial bone. Arthroscopy 1998;14:29 – 37.

[22] Rupp S, Krauss PW, Fritsch EW. Fixation strength of a biodegradable interference screw and a press-fit technique in anterior cruciate ligament reconstruction with a BPTB graft. Arthroscopy 1997;13:61 – 5. [23] Seil R, Rupp S, Krauss PW, Benz A, Kohn DM. Comparison of initial fixation strength between biodegradable and metallic interference screws and a press-fit fixation technique in a porcine model. Am J Sports Med 1998;26:815 – 9. [24] Nurmi JT, Kannus P, Sievanen H, Jarvinen M, Jarvinen TL. Compaction drilling does not increase the initial fixation strength of the hamstring tendon graft in anterior cruciate ligament reconstruction in a cadaver model. Am J Sports Med 2003;31:353 – 8. [25] Nurmi JT, Jarvinen TL, Kannus P, Sievanen H, Toukosalo J, Jarvinen M. Compaction versus extraction drilling for fixation of the hamstring tendon graft in anterior cruciate ligament reconstruction. Am J Sports Med 2002;30:167 – 73.