Graft-bone motion and tensile properties of hamstring and patellar tendon anterior cruciate ligament femoral graft fixation under cyclic loading

Graft-Bone Motion and Tensile Properties of Hamstring and Patellar Tendon Anterior Cruciate Ligament Femoral Graft Fixation Under Cyclic Loading Charles H. Brown, Jr., M.D., David R. Wilson, D.Phil., Aaron T. Hecker, M.S., and Mike Ferragamo, B.Sc.

Purpose: To assess longitudinal graft-bone motion and tensile properties of the femur–anterior cruciate ligament (ACL) graft fixation–ACL graft complex based on the hypothesis that there is little difference in graft-bone motion between suspensory and aperture hamstring ACL femoral graft fixation techniques, and between hamstring and patellar tendon ACL femoral graft fixation techniques. Type of Study: In vitro biomechanical study using human cadavers. Methods: The distal femur–ACL graft fixation–ACL graft complex was cyclically loaded between 50 and 250 N at 1 Hz for 1,000 cycles with the direction of the load applied parallel to the axis of the femoral bone tunnel. Graft-bone motion was measured indirectly using retroreflective markers and a video motion-analysis system. Tensile testing to failure was performed at 1 mm/sec for fixation techniques completing 1,000 cycles without fixation failure. Results: Among the hamstring fixation techniques, 4 of 13 Bio-Interference screws (Arthrex, Naples, FL), 2 of 12 LinX HT fasteners (DePuy Mitek, Norwood, MA), and 1 of 11 TransFix cross-pins (Arthrex) failed before completing 1,000 cycles. Five of 13 patellar tendon grafts fixed with metal interference screws, and 2 of 12 patellar tendon grafts fixed with a plastic button and No. 5 sutures failed before completing 1,000 cycles. Suture/button fixation of patellar tendon grafts resulted in significantly more graft-bone motion than hamstring tendon grafts fixed using the Bone Mulch Screw (Arthrotek, Warsaw, IN), or interference screw fixation of patellar tendon and hamstring grafts. Otherwise, there was no significant difference in graft-bone motion among the various hamstring fixation techniques or the various hamstring fixation techniques and interference screw fixation of patellar tendon grafts. Maximum graft-bone displacement after cyclic loading was significantly greater for hamstring grafts fixed with the EndoButton and EndoButton Tape (Smith & Nephew Endoscopy, Andover, MA) compared with the other hamstring fixation techniques and interference screw fixation of patellar tendon grafts. All fixation techniques except hamstring tendon grafts fixed with the Bio-Interference screw achieved at least 59% of maximum graft-bone displacement after 20 cycles. Hamstring tendon grafts fixed with the EndoButton CL were significantly stronger than all other hamstring and patellar tendon fixation methods. Patellar tendon grafts fixed with interference screws and hamstring tendon grafts fixed with interference screws and the Bone Mulch Screw and TransFix were significantly stiffer than hamstring tendon grafts fixed with the EndoButton CL. Conclusions: There is no significant difference in graft-bone motion between aperture and suspensory femoral fixation methods when the stiffness of the femur–ACL graft fixation–ACL graft complex is similar. Clinical Relevance: The small differences in graft-bone motion reported in our study provide further evidence that graft-tunnel motion or the so-called bungee effect is unlikely to be the primary cause of radiographic bone tunnel enlargement following ACL reconstruction. Key Words: ACL graft fixation—ACL graft tunnel motion—Bungee effect—Tunnel enlargement.

From the Orthopaedic Biomechanics Laboratory, Department of Orthopaedic Surgery, Beth Israel Deaconess Medical Center (C.H.B., D.R.W., A.T.H.), and the Department of Orthopaedic Surgery, Brigham and Women’s Hospital (C.H.B.), Boston; Smith & Nephew Endoscopy, Inc (A.T.H., M.F.), Mansfield, Massachusetts, U.S.A.; and the Department of Orthopaedics, The University of British Columbia and the Vancouver Hospital & Health Science Centre (D.R.W.), Vancouver, British Columbia, Canada. Supported by research grants from Aircast Incorporated, and

922

the Research and Development Department, Smith & Nephew Endoscopy. Address correspondence and reprint requests to Charles H. Brown, Jr., M.D., Department of Orthopaedic Surgery, Brigham and Women’s Hospital, 850 Boylston St, Suite 130, Chestnut Hill, MA 02467, U.S.A. E-mail: [email protected] © 2004 by the Arthroscopy Association of North America 0749-8063/04/2009-3770$30.00/0 doi:10.1016/j.arthro.2004.06.032

Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 20, No 9 (November), 2004: pp 922-935

ACL GRAFT TUNNEL MOTION

R

adiographic enlargement of the femoral and tibial bone tunnels has been reported following anterior cruciate ligament (ACL) reconstruction using bone–patellar tendon– bone and hamstring tendon autografts, and bone–patellar tendon– bone and Achilles tendon allografts.1-12 At the present time, the etiology and clinical significance of radiographic bone tunnel enlargement following ACL reconstruction remains unknown. It has been hypothesized that radiographic bone tunnel enlargement after ACL reconstruction results from local bone resorption or osteolysis.4,5,8,11 Various biologic and mechanical factors have been proposed as the cause of bone tunnel enlargement. Biologic factors include use of allograft tissue, toxic products, a nonspecific cytokine-mediated inflammatory response, and cell necrosis from drilling.5 Mechanical factors which have been hypothesized to cause bone tunnel enlargement include stress shielding, aggressive rehabilitation, high graft forces due to nonanatomic graft placement, and graft tunnel motion.5 None of these factors has been definitively established as the primary cause of bone tunnel enlargement.1,5 Motion of the ACL graft in the bone tunnel has been one of the most commonly proposed causes of bone tunnel enlargement.1,5,8,13,14 It has been hypothesized that fixation of the ACL graft away from the anatomic insertion sites of the normal ACL (suspensory fixation) may result in increased amounts of graft-tunnel motion compared with fixation techniques that fix the graft closer to the anatomic insertion site of the normal ACL (aperture fixation).1,5,8,13-15 Longitudinal grafttunnel motion along the axis of the bone tunnel under cyclic loading conditions has been referred to as the “bungee effect,” and is believed by some investigators to be one of the major causes of bone tunnel enlargement.1,5 It has been suggested that longitudinal grafttunnel motion under cyclic loading conditions may impair the biologic incorporation of the ACL graft into the bone tunnels, leading to bone tunnel enlargement.5 The magnitude of graft-tunnel motion under defined cyclic loading conditions for femoral and tibial ACL graft fixation techniques commonly used in clinical practice is presently unknown. The purpose of this in vitro biomechanical study was to measure longitudinal graft-tunnel motion, graft slippage, and the load-tofailure properties of hamstring and patellar tendon ACL femoral fixation techniques under defined cyclical loading conditions in human cadavers. It is our hypothesis that there is little difference in graft-bone motion between aperture and suspensory hamstring

923

ACL femoral graft fixation techniques, and between hamstring and patellar tendon ACL femoral graft fixation techniques. METHODS Eighteen unembalmed human cadaveric knee specimens from 9 donors (6 male and 3 female) were obtained through the LifeLegacy Foundation, Tucson, Arizona. Specimens from donors with a history or evidence of fractures, tumors, or previous surgery were excluded from the study. The mean age of the 9 donors was 46 ⫾ 2.5 years, with a range of 42 to 50 years. All specimens were stored at ⫺20°C, and thawed at room temperature for 24 hours before dissection. The gracilis and semitendinosus tendons were harvested as free grafts in an open fashion. The ends of each tendon graft were sutured with 5 throws of a baseball-type whipstitch using a No. 2 Ethibond suture (Special order D-5757; Ethicon, Somerville, NJ). The gracilis and semitendinosus tendons were looped around a No. 5 Ethibond suture creating a doubled gracilis and semitendinosus graft (DGST). The diameter of the DGST graft was measured to the nearest half-millimeter by passing the 4 tendon strands through a commercially available sizing block (Arthrex, Naples, FL). The diameter of the DGST grafts in our study ranged from 6.5 to 8.5 mm. Ten-millimeter wide patellar tendon (PT) grafts with rectangular patellar bone blocks measuring 10 mm wide and 25 mm long were harvested from the medial and lateral half of the patellar ligament from each knee using a sagittal saw. The distal end of the PT and periosteum was sharply dissected from their insertion into the tibial tubercle and the anterior cortex of the proximal tibia. The rectangular patellar bone block was trimmed into a cylindrical shape using a bone rongeur until it passed snugly through a 10-mm sizing tube. Although the gap distance between the patellar bone block and the sizing tube was not measured, the cylindrical shape of the bone plug minimized the gap distance between the bone block and sizing tube. Normal saline at room temperature was applied with a spray bottle to keep the specimens moist during dissection and mechanical testing. To conserve scarce, young, fresh-frozen human specimens, we reused the distal femur and DGST and PT grafts from each knee to test all fixation techniques within that knee. However, if the DGST or PT graft showed any sign of injury or structural damage, the untested DGST or PT graft from the same knee was used to complete the remaining biomechanical tests in that

924

C. H. BROWN, JR., ET AL.

knee. We speculate that the tensile properties of the femur– graft fixation–ACL graft complex are unlikely to be significantly affected by reusing the same ACL graft for multiple tests because the ACL grafts were cyclically loaded to less than 7% of their maximum failure load, and the failure loads for the various graft fixation techniques were expected to fall far below the maximum failure loads of the DGST (4,590 ⫾ 674 N) and PT grafts (3,899 ⫾ 298 N).16-19 Five of the 8 ACL femoral graft fixation techniques were tested in the lateral femoral condyle of each knee, and the 3 remaining graft fixation techniques were tested in the medial femoral condyle. Behrens et al.,20 in a human cadaveric knee model, found that the bone on the medial side of the distal femur is approximately 20% to 27% stronger than on the lateral side. However, they found no correlation between strength and bone material density. Brand et al.21 have shown that bone mineral density best correlates with interference screw fixation of soft-tissue grafts. Given the lack of correlation of strength and bone mineral density between bone in the medial and lateral distal femur, we speculate that using the medial femoral condyle to test some ACL graft fixation techniques is unlikely to affect the results of kinematic and tensile testing. Testing of the remaining fixation techniques in a given specimen was also suspended if there was evidence of structural damage to the specimen. Because some graft fixation techniques failed before completing 1,000 cycles, kinematic and tensile data were not available for all fixation techniques in every specimen. In an attempt to minimize potential carry-over effects among the various graft fixation techniques, the following testing sequence was used in the lateral femoral condyle of each specimen (Fig 1): DGST graft fixed with an 8 ⫻ 23-mm Bio-Interference screw (Arthrex, Naples, FL); DGST graft fixed with an EndoButton and a 20-mm loop of EndoButton Tape secured with 5 squareknot throws (Smith & Nephew Endoscopy, Andover, MA); DGST graft fixed with an EndoButton and a 20-mm continuous polyester loop, EndoButton CL (Smith & Nephew Endoscopy); DGST graft fixed with a LinX HT polymer hamstring tendon fastener (DePuy Mitek, Norwood, MA); DGST graft fixed with a Bone Mulch Screw (Arthrotek, Warsaw, IN). The rationale for using the above testing sequence is as follows: fixation of hamstring tendon grafts with a Bio-Interference screw is dependent on compression of the hamstring tendon graft against the cancellous

FIGURE 1. Hamstring and PT fixation techniques (see Methods section for a description of each technique).

bone of the femoral tunnel.21 This fixation technique was tested first because we speculated that it was most likely to be adversely affected by prior drilling of bone tunnels. A 2.4-mm diameter drill-tip guidewire (Smith & Nephew Endoscopy) was drilled in an open fashion at the femoral anatomic attachment site of the ACL. A 5-mm diameter bone tunnel was drilled in the lateral femoral condyle using a cannulated endoscopic reamer. As recommended by the manufacturer, the bone tunnel was progressively dilated using half-millimeter incremental, cannulated, smooth tunnel dilators (Arthrex) up to the measured size of the DGST graft. The DGST grafts in our study ranged from 6.5 to 8.5 mm in diameter. The femoral tunnel was notched at the 12-o’clock position using a Bio-Interference screw tunnel notcher (Arthrex). The DGST graft was marked 25 mm from the axilla, and pulled into the bone tunnel using a No. 5 Ethibond suture until the 25-mm mark was flush with the entrance of the tunnel. A 1.1-mm Nitinol guidewire (Arthrex) was inserted into the notched area of the bone tunnel and passed between the DGST graft and the bone tunnel wall, parallel to the long axis of the bone tunnel. Based on the manufacturer’s recommendations at the time this study was performed, an 8 ⫻ 23-mm BioInterference screw was used the fix the graft in the lateral femoral condyle in all specimens. The BioInterference screw was advanced over the guidewire, parallel to the axis of the bone tunnel, until the round head of the screw was flush with the entrance of the bone tunnel. After testing to failure, the DGST graft

ACL GRAFT TUNNEL MOTION and Bio-Interference screw were removed from the knee. The 2 EndoButton fixation methods were tested next because prior testing of the specimen using a Bio-Interference screw would be unlikely to interfere with the integrity of the cortical bone on which the EndoButton implant is anchored.19 A 2.7-mm drilledtip guidewire (Smith & Nephew Endoscopy) was drilled down the center of the existing bone tunnel and through the lateral femoral cortex. A 4.5-mm cannulated EndoButton Drill (Smith & Nephew Endoscopy) was used to drill a tunnel through the lateral cortex of the distal femur. The DGST graft was fixed in the lateral femoral condyle by looping the 4 limbs of the graft around a 20-mm loop of 5-mm wide woven polyester EndoButton Tape, which was passed through the 2 central holes of the EndoButton. The Endobutton Tape was tied to itself using 5 square-knot throws creating a loop. The EndoButton and DGST graft were passed into the femoral bone tunnel and the EndoButton was flipped and anchored on the lateral femoral cortex. After testing to failure, the EndoButton and DGST graft were removed from the lateral femoral condyle and fixation of the DGST graft was performed in a similar fashion using an EndoButton with a 20-mm continuous loop. The loop length for the 2 EndoButton fixation techniques was held constant in order to keep the stiffness of the femur–ACL graft fixation–ACL graft complex constant in each specimen. This resulted in different amounts of hamstring tendon graft being inserted into the femoral tunnel of each specimen. However, To et al.19 showed, using a strings-in-series mathematical model, that the length of the ACL graft does not significantly affect the stiffness of the femur–ACL graft fixation–ACL graft complex unless the stiffness of the ACL graft fixation method is similar to that of the ACL graft. Because the stiffness of the DGST graft (861 ⫾ 186 N/mm) is approximately 3.4 times that of the 20-mm EndoButton CL (251 N/mm; unpublished data, Research and Development Department, Smith & Nephew Endoscopy), and 13.4 times that of EndoButton Tape (64 N/mm; unpublished data, Research and Development Department, Smith & Nephew Endoscopy), changes in the amount of DGST graft inserted into the femoral tunnel will not significantly affect the stiffness of the femur–EndoButton– DGST graft complex.16,19 The LinX HT fastener was tested after the 2 EndoButton techniques because its fixation is achieved by screwing a threaded sleeve into the cortical bone of the femur. We speculate that the integrity of the cor-

925

tical bone into which the LinX HT threaded sleeve was inserted would not be significantly affected by prior testing of the Bio-Interference screw and EndoButton implants. After removing the DGST graft and EndoButton implant, a 2.7-mm guidewire was passed into the previously drilled 4.5-mm cortical bone tunnel and the cortical bone was over-reamed using a 5-mm cannulated drill (DePuy Mitek). The lateral femoral cortex was tapped using a cannulated bone tap (DePuy Mitek) and the LinX HT threaded sleeve was inserted until the threads engaged into the lateral femoral cortex. The DGST graft was threaded through the eyelet tab, and the eyelet tab inserted into the threaded sleeve. The final implant tested in the lateral femoral condyle was the Bone Mulch Screw. The DGST graft was removed and the U-Shaped Drill Guide (Arthrotek) was inserted a distance of 25 mm into the femoral tunnel. The aiming bullet was inserted into the U-Shaped Drill Guide and the tip of the bullet was positioned anterior to the lateral epicondyle. A 2.4-mm drill-tipped guidewire was drilled through the aiming bullet until it stopped against the lateral side of the insertion rod. The U-Shaped Drill Guide was removed from the femoral tunnel, and the 2.4-mm drilltipped guidewire was directed across the femoral tunnel, dividing it in half. An 8-mm reamer was used to drill a tunnel for the Bone Mulch Screw in the lateral femoral condyle. A 10.5 ⫻ 25-mm Bone Mulch Screw was inserted into the lateral femoral condyle so that the tip of the implant bisected the femoral tunnel. Under direct vision, 25 mm of the DGST graft was inserted into the femoral tunnel and the graft looped around the tip of the Bone Mulch Screw. The Bone Mulch Screw was advanced into the lateral femoral condyle until the head of the screw was flush with the cortex of the lateral femoral condyle. After testing to failure, the Bone Mulch Screw was removed from the lateral femoral condyle, and a 2.4-mm diameter drill-tipped guidewire was drilled into the inner wall of the medial femoral condyle, 7 mm anterior to the posterior edge of the medial femoral condyle. A 5-mm diameter bone tunnel was drilled in the medial femoral condyle, and the bone tunnel was progressively dilated to 10 mm using cannulated, smooth tunnel dilators. In an attempt to minimize carry-over effects, 3 fixation techniques were tested in the medial femoral condyle in the following sequence (Fig 1): PT graft fixed with a 7 ⫻ 25-mm interference screw (Smith & Nephew);

926

C. H. BROWN, JR., ET AL.

PT graft fixed with 3 No. 5 Ethibond sutures tied to a 17-mm plastic ligament fixation button (Smith & Nephew); DGST graft fixed with the TransFix cross-pin (Arthrex). The 10-mm PT graft was inserted into the bone tunnel until the end of the bone block was flush with the entrance of the femoral tunnel. The cancellous bone surface of the PT bone block was oriented superiorly in the femoral tunnel. A 1.5-mm Nitinol guidewire was inserted anterior to the bone block and parallel to the axis of the bone tunnel. A 7 ⫻ 25-mm metal interference screw was inserted over the guidewire, parallel to the axis of the bone tunnel until the head of the screw was flush with the edge of the bone tunnel. After testing to failure, the PT graft and metal interference screw were removed from the knee, and a 2.7-mm drill-tipped guidewire was drilled through the center of the bone tunnel and out through the medial cortex of the distal femur. A 4.5-mm tunnel was drilled through the distal medial femoral cortex using a 4.5-mm EndoButton drill. Three evenly spaced 2-mm drill holes were drilled in the PT bone block, and a No. 5 Ethibond suture was passed through each drill hole. The 3 No. 5 sutures were passed through the bone tunnel in the medial femoral cortex and the bone block was advanced until flush with the inner edge of the bone tunnel. The No. 5 Ethibond sutures were passed through the holes of a 17-mm plastic ligament fixation button (Smith & Nephew, Andover, MA), the fixation button was placed flush against the medial cortex, and each suture tied individually to itself over the plastic button with 5 square-knot throws. After testing to failure, the PT graft and fixation button were removed from the knee, a Long Adapter Drill Guide C-Ring (Arthrex) was advanced 25 mm into the femoral tunnel, and the Guide Pin Sleeve was positioned against the medial femoral epicondyle. A 2.0-mm TransFix Drill Pin (Arthrex) was drilled across the bone tunnel bisecting it. The guide pin was removed, and the TransFix implant impacted into the medial femoral condyle under direct vision, bisecting the 10-mm bone tunnel. Under direct vision, 25 mm of DGST graft were inserted into the femoral bone tunnel, and both limbs of the graft were looped around the TransFix cross-pin. The TransFix cross-pin was impacted flush with the cortex of the medial femoral condyle.

Biomechanical Testing Graft-bone motion, ultimate failure load, linear stiffness, and elongation to failure of the femur– graft fixation–ACL graft complex were measured using a Servohydraulic materials testing system (Model 1331; Instron, Canton, MA). The distal femur was mounted to the actuator of the materials testing machine using a custom-designed clamp that allowed the axis of the bone tunnels in the medial and lateral femoral condyle to be positioned parallel to the axis of the applied load. The free ends of the grafts were secured to the load cell of the materials testing machine using a custom-built tendon-freezing grip.16,19 The distance from the entrance of the bone tunnel to the tendon-freezing grip was held constant at 30 mm to simulate the intra-articular length of the ACL. Retroreflective 7-mm diameter spherical markers were attached to the medial and lateral femoral condyles using screws, on the ACL graft near the tendon-freezing grip, and at the exit point from the bone tunnel using 2-0 Vicryl sutures. An additional set of markers was placed on the body of the tendon-freezing grip and the custom clamp attaching the distal femur to the actuator of the materials testing machine to monitor slippage. A video motion-analysis system (PC Reflex; Qualysis, Glastonbury, CT) was used to determine graft-bone motion by measuring the relative displacement between the markers on the ACL graft and femoral condyles (Fig 2). We assessed the mean error of the video motionanalysis system by moving retroreflective markers a known distance using a sliding table actuated by lead screws scaled in 0.01-mm increments. We displaced 4 different markers by 5, 10, 15, and 20 mm in the vertical direction and measured their displacement with the camera positioned and oriented as it was during the in vitro cyclic loading experiments. Measurement error was determined by calculating the absolute value of the difference between known displacement and displacement measured using the video system at each position for each marker. The mean measurement error of the video motion-analysis system was determined to be 0.0967 mm. Each fixation technique was first subjected to uniaxial cyclic loading between 50 and 250 N at 1 Hz for 1,000 cycles. These loads approximate the forces that have been measured in the ACL during terminal passive extension of the knee.22 Loading was applied parallel to the axis of the bone tunnel with all other

ACL GRAFT TUNNEL MOTION

927

FIGURE 2. Testing apparatus, and placement of retroreflective markers. Graft-bone motion was measured by tracking the difference in displacement between markers placed on the femoral condyles and the ACL graft as it exited the femoral tunnel.

translation and rotation of the specimen constrained. A loading frequency of 1 cycle per second (1 Hz) was chosen because this approximates the frequency of normal walking.23 Based on the recommendations of Beynnon and Amis,24 we cycled each fixation technique for 1,000 cycles. One thousand cycles approximates a week of flexion-extension loading on an ACL reconstruction.25 Displacement of the retroreflective marker on the ACL graft near the exit of the bone tunnel relative to the markers on the femoral condyles was calculated in each cycle and averaged over the full 1,000 cycles to determine steady-state graft-bone motion (Fig 3). The overall maximum displacement of the ACL graft marker relative to the markers on the femoral condyles after 1,000 cycles and the percentage of maximum graft-bone displacement reached after 20 cycles were

FIGURE 3.

Representative graft-bone displacement curve.

also calculated (Fig 3). Fixation techniques that survived 1,000 cycles were then tested to failure at 1 mm/sec with the tensile load applied parallel to the axis of the bone tunnel with all other motions constrained. Load-displacement curves were recorded using data acquisition software (Labtech Notebook, Wilmington, MA) in addition to an X-Y plotter (Model 3025; Yokogawa Corporation of America, Newnan, GA). The ultimate failure load, linear stiffness, and displacement to failure of the femur–ACL graft fixation–ACL graft complex were determined using in-house data-analysis software. The failure mechanism for each fixation technique was also recorded. Statistical Analysis A 1-way analysis of variance followed by a Tukey test was used to test the following statistical hypotheses: (1) The failure load for each fixation technique is not higher than the failure load for other techniques. (2) The stiffness of each fixation technique is not higher than the stiffness of other techniques. (3) The mean steady-state graft-bone motion for each fixation technique is not greater than the mean steady-state graft-bone motion for other fixation techniques. (4) The maximum graftbone displacement for each fixation technique is not greater than the maximum graft-bone displacement for other techniques. Significance was set at P ⬍ .05.

928

C. H. BROWN, JR., ET AL. RESULTS

ation failure in both specimens occurred as a result of the sutures cutting through the bone plug.

Failure During Cyclic Loading DGST grafts fixed with Bio-Interference screws failed in 4 of 13 specimens before completing 1,000 cycles. Fixation failure occurred in all cases by the DGST graft pulling out of the bone tunnel with the Bio-Interference screw remaining firmly fixed in the bone tunnel. DGST grafts fixed with the LinX HT polymer fastener failed in 2 of 12 specimens before completing 1,000 cycles. Both specimens failed owing to improper placement of the LinX HT implant. In 1 specimen, fixation failure occurred as a result of the failure of the LinX HT threaded sleeve to be inserted into the cortex of the distal femur. Failure occurred in the second specimen when the LinX HT eyelet was not fully engaged into the threaded sleeve. One of 11 DGST grafts fixed with the TransFix technique failed during cyclic loading when the implant toggled into the cancellous bone of the medial femoral condyle. PT grafts fixed with a 7 ⫻ 25-mm interference screw failed in 5 of 13 specimens before completing 1,000 cycles. Fixation failure occurred in 2 specimens due to fracture of the bone plug, in 2 specimens the bone plug and screw pulled out of the bone tunnel, and in 1 specimen the PT graft pulled out from the tendonfreeze grip. PT grafts fixed with 3 No. 5 Ethibond sutures tied around a 17-mm plastic button failed in 2 of 12 specimens before completing 1,000 cycles. FixTABLE 1.

Graft-Bone Motion Graft-bone motion data were not available for all tests because of the retroreflective markers being obscured and the failure of the computer system to capture and record kinematic data at the beginning of the test. Graft-bone motion data were also excluded when the fixation method failed to complete 1,000 cycles. The mean steady-state graft-bone motion, maximum graft-bone displacement, and percent maximum graft-bone displacement after 20 cycles are reported in Table 1. The mean steady-state graft-bone motion for all fixation techniques ranged from 0.34 ⫾ 0.15 mm for PT grafts fixed with metal interference screws to 0.67 ⫾ 0.17 mm for PT grafts fixed with sutures and button (Fig 4). Interference screw fixation of PT and hamstring tendon grafts, and hamstring tendon grafts fixed with the Bone Mulch Screw, resulted in the least amount of graft-bone motion. PT grafts fixed with sutures and button had significantly greater steady-state graft-bone motion compared with PT grafts fixed with interference screws, and hamstring tendon grafts fixed with Bio-Interference screws and the Bone Mulch Screw. Otherwise, there was no significant difference in graft-bone motion among the other hamstring fixation techniques or the hamstring

Mean (⫾ SD) Steady-State Graft-Bone Motion, Maximum Graft-Bone Displacement, Graft-Bone Displacement After 20 Cycles, Ultimate Failure Load, Linear Stiffness, and Displacement to Failure

Variable

EndoButton, BioInterference EndoButton Tape Screw 20 mm 8 ⫻ 23 mm

EndoButton Continuous Loop 20 mm

Steady-state graft0.35 ⫾ 0.15 0.55 ⫾ 0.17 0.51 ⫾ 0.14 bone motion n⫽9 n⫽7 n⫽7 (mm) Maximum graft4.34 ⫾ 3.16 5.82 ⫾ 1.81 2.13 ⫾ 0.26 bone displacement n⫽7 n⫽7 n⫽6 after 1,000 cycles (mm) Graft-bone 42% 79% 72% displacement after n⫽7 n⫽7 n⫽6 20 cycles (% of max) Ultimate failure load 562 ⫾ 69 644 ⫾ 91 1,345 ⫾ 179 (N) n⫽9 n ⫽ 10 n ⫽ 11 Linear stiffness (N/ 257 ⫾ 37 182 ⫾ 20 179 ⫾ 39 mm) n⫽9 n ⫽ 10 n ⫽ 11 Displacement to 3.00 ⫾ 0.66 6.27 ⫾ 2.16 9.89 ⫾ 2.41 failure (mm) n⫽9 n ⫽ 10 n ⫽ 11

LinX HT

Bone Mulch Screw

TransFix

PT Screw 7 ⫻ 25 mm

PT Suture Button

0.54 ⫾ 0.27 0.36 ⫾ 0.08 0.44 ⫾ 0.23 0.34 ⫾ 0.15 0.67 ⫾ 0.17 n⫽7 n⫽8 n⫽9 n⫽7 n ⫽ 10 2.20 ⫾ 0.95 2.24 ⫾ 0.63 2.37 ⫾ 1.43 1.53 ⫾ 0.42 4.42 ⫾ 1.53 n⫽7 n⫽7 n⫽7 n⫽5 n⫽8 71% n⫽7

70% n⫽7

59% n⫽7

687 ⫾ 129 977 ⫾ 238 934 ⫾ 296 n ⫽ 10 n ⫽ 10 n ⫽ 10 230 ⫾ 32 257 ⫾ 50 240 ⫾ 74 n ⫽ 10 n ⫽ 10 n ⫽ 10 3.74 ⫾ 1.05 6.49 ⫾ 2.66 7.37 ⫾ 371 n ⫽ 10 n ⫽ 10 n ⫽ 10

62% n⫽5 710 n 298 n 3.17 n

⫾ ⫽ ⫾ ⫽ ⫾ ⫽

75% n⫽9

224 664 ⫾ 132 8 n ⫽ 10 36 207 ⫾ 36 8 n ⫽ 10 0.87 6.02 ⫾ 2.47 8 n ⫽ 10

ACL GRAFT TUNNEL MOTION

929

FIGURE 4. Steady-state graft-bone motion (⫾SD) for hamstring and PT femoral fixation techniques.

FIGURE 5. Maximum graft-bone displacement (⫾SD) for hamstring and PT femoral fixation techniques after 1,000 cycles of uniaxial loading between 50 and 250 N at 1 Hz.

fixation techniques and interference screw fixation of PT grafts. The mean maximum graft-bone displacement after 1,000 cycles ranged from 1.53 ⫾ 0.42 mm for PT grafts fixed with interference screws to 5.82 ⫾ 1.81 mm for hamstring tendon grafts fixed with an EndoButton and EndoButton Tape (Fig 5). PT grafts fixed with interference screws showed the least amount of displacement after 1,000 cycles. However, there was no significant difference in maximum graftbone displacement between PT grafts fixed with interference screws and all hamstring fixation techniques except for hamstring tendon grafts fixed with an EndoButton and EndoButton Tape. Among hamstring tendon fixation techniques, the EndoButton CL, Bone Mulch Screw, and LinX HT techniques showed the least amount of displacement after cyclic loading. Maximum graft-bone displacement was significantly greater for hamstring tendon grafts fixed with an EndoButton and EndoButton Tape compared with all hamstring fixation techniques except Bio-Interference screws. PT grafts fixed with interference screws had significantly less maximum graft-bone displacement than PT grafts fixed with sutures and button. The percentage of maximum graft-bone displacement reached after 20 cycles varied from a low of 42% for hamstring tendon grafts fixed with Bio-Interference screws to 85% for PT grafts fixed with suture/ button (Fig 6). Hamstring tendon grafts fixed with Bio-Interference screws achieved a significantly lower percentage of maximum graft-bone displacement

compared with the other fixation techniques except for hamstring tendon grafts fixed with the TransFix crosspin, and PT grafts fixed with interference screws. Otherwise, there was no significant difference among the other hamstring and PT techniques. Load-to-Failure Properties The load to failure, stiffness, and displacement to failure of the femur–ACL graft fixation–ACL graft

FIGURE 6. Percentage of maximum graft-bone displacement reached after 20 cycles of uniaxial loading between 50 and 250 N at 1 Hz for hamstring and PT femoral fixation techniques.

930

C. H. BROWN, JR., ET AL.

complex for the various hamstring and PT fixation techniques are shown in Table 1. The mean failure load for the different fixation techniques ranged from 562 ⫾ 69 N for hamstring tendon grafts fixed with Bio-Interference screws to 1,345 ⫾ 179 N for hamstring tendon grafts fixed with the EndoButton CL (Fig 7). Hamstring tendon grafts fixed with the EndoButton CL were significantly stronger than all other hamstring and PT fixation techniques. Hamstring tendon grafts fixed with Bio-Interference screws were significantly weaker than all hamstring fixation techniques except for the EndoButton and EndoButton Tape, and the LinX HT fastener. There was no significant difference in the maximum failure load of the 2 PT fixation techniques or between PT grafts fixed with interference screws and the other hamstring fixation techniques. The mean linear stiffness of the femur–ACL graft fixation–ACL graft complex for the different fixation techniques ranged from 179 ⫾ 39 N/mm for hamstring tendon grafts fixed with an EndoButton CL to 299 ⫾ 36 N for PT grafts fixed with interference screws (Fig 8). Hamstring tendon grafts fixed with the 2 EndoButton techniques and the LinX HT, and PT grafts fixed with sutures and button, were significantly less stiff than PT grafts fixed with interference screws. There was no significant difference in stiffness between PT grafts fixed with interference screws, and hamstring tendon grafts fixed with Bio-Interference screws, the Bone Mulch Screw, and TransFix crosspin. Hamstring tendon grafts fixed with the Bone

FIGURE 8. Linear stiffness (⫾SD) for hamstring and PT femoral fixation techniques.

Mulch Screw were significantly stiffer than hamstring tendon fixation using the 2 EndoButton techniques. There was no significant difference in the stiffness of hamstring tendon grafts fixed with Bio-Interference screws, LinX HT, Bone Mulch Screw, and TransFix cross-pin. Displacement to failure ranged from 3.00 ⫾ 0.66 mm for hamstring tendon grafts fixed with BioInterference screws to 9.89 ⫾ 2.41 mm for hamstring tendon grafts fixed with the EndoButton CL. Interference screw fixation of hamstring and PT grafts resulted in the shortest displacement to failure. Hamstring tendon grafts fixed with the Bio-Interference screw resulted in significantly shorter displacement to failure compared with all hamstring tendon fixation techniques except for the LinX HT. Displacement to failure for hamstring tendon grafts fixed with the EndoButton CL was significantly greater than all other hamstring and PT fixation techniques except for hamstring tendon grafts fixed with the TransFix cross-pin. Failure Modes During Tensile Testing

FIGURE 7. Maximum failure load (⫾SD) for hamstring and PT femoral fixation techniques.

Hamstring tendon grafts fixed with the Bio-Interference screw failed in all cases by the tendon pulling out of the bone tunnel with the Bio-Interference screw remaining firmly fixed in the bone tunnel. Hamstring tendon graft fixation with the 2 EndoButton techniques failed in all cases by rupture of the polyester tape and continuous loop. The LinX HT polymer fastener failed in all cases by breaking at the junction of the eyelet pin. Hamstring tendon graft fixation with

ACL GRAFT TUNNEL MOTION the Bone Mulch Screw failed in all cases by the tip bending at the junction with the body of the screw. The TransFix cross-pin failed by the implant toggling into the cancellous bone of the medial femoral condyle, or by the cross-pin bending and the tendons sliding off. PT grafts fixed with metal interference screws failed by the bone plug pulling out of the bone tunnel, fracture of the bone plug, or bony avulsion at the PT insertion site. PT grafts fixed with sutures and button failed as a result of the sutures cutting through the bone, fracture of the bone plug, or the sutures breaking. DISCUSSION It has been hypothesized that radiographic bone tunnel enlargement following ACL reconstruction results from motion of the ACL graft in the bone tunnel.1,5,8,14 Longitudinal motion of the ACL graft in the bone tunnel has been referred to as the bungee effect, and this particular type of motion has been hypothesized as being a major cause of bone tunnel enlargement.1,5,14 The objective of this biomechanical study was to measure and compare the amount of graft-bone motion and the load-to-failure properties of commonly performed hamstring and PT ACL graft femoral fixation techniques under defined cyclical and tensile loading conditions. Our data show that the mean steady-state graft-bone motion of the femur– graft fixation–ACL graft complex under cyclical loading from 50 to 250 N ranged from 0.34 to 0.67 mm for the hamstring and PT ACL graft femoral fixation techniques tested in this study. Interference screw fixation of PT grafts and hamstring tendon graft fixation using Bio-Interference screws and the Bone Mulch Screw resulted in the least amount of graft-bone motion. However, we found no significant difference in graft-bone motion between PT and hamstring tendon grafts fixed using aperture fixation with interference screws, and among hamstring tendon grafts fixed using suspensory fixation techniques such as the EndoButton, LinX HT, Bone Mulch Screw, and TransFix. Our results support our hypothesis that there are small differences in graftbone motion between aperture and suspensory hamstring ACL femoral fixation techniques, and among the various hamstring and PT ACL femoral fixation techniques. PT grafts fixed with sutures tied around a button was the only fixation technique to demonstrate significantly greater graft-bone motion. Hamstring tendon grafts fixed using the EndoButton CL, LinX HT, Bone Mulch Screw, and TransFix

931

cross-pin techniques, and the PT grafts fixed with interference screws, resulted in the least amount of graft-bone displacement (slippage) after 1,000 cycles. Hamstring tendon grafts fixed with Bio-Interference screws, EndoButton with EndoButton Tape, and PT grafts fixed with sutures tied around a plastic fixation button resulted in approximately twice the amount of slippage compared with other fixation techniques. However, in the case of hamstring tendon grafts fixed with an EndoButton and EndoButton Tape and PT grafts fixed with sutures and button, over 75% of the maximum displacement (slippage) was achieved within the first 20 cycles. To minimize slippage, our findings suggest that cycling of suspensory fixation techniques using a connecting material that is tied is advisable to allow the knots to slip and tighten, and the connecting material to stretch. The low percentage of maximum displacement achieved after 20 cycles with Bio-Interference screw fixation of hamstring tendon grafts, combined with the high (31%) failure rate during cyclic loading suggest that cycling is unlikely to reduce slippage of this fixation technique. Hamstring tendon grafts fixed with the EndoButton CL showed superior fixation strength compared with all other hamstring and PT fixation methods. There was no significant difference in strength between PT grafts fixed with interference screws and the other hamstring and PT fixation techniques except for the EndoButton CL. Hamstring tendon fixation using the Bone Mulch Screw and TransFix cross-pin demonstrated superior fixation strength compared with all other hamstring fixation techniques except for the EndoButton CL. Interference screw fixation of PT grafts resulted in the highest stiffness of any hamstring or PT fixation technique. This fixation technique was significantly stiffer than suture/button fixation of PT grafts, and hamstring tendon grafts fixed using the 2 EndoButton techniques and LinX HT. The Bio-Interference screws and the Bone Mulch Screw were the stiffest fixation methods for hamstring tendon grafts. Suspensory fixation techniques using connecting materials (EndoButton and EndoButton Tape, EndoButton CL, sutures and button) resulted in lower stiffness values compared with aperture fixation methods, and suspensory fixation techniques which did not rely on a connecting linkage (cross-pins and LinX HT). Our results for PT grafts fixed with interference screws compare favorably with those reported by Honl et al.23 In that study, 10-mm wide bone-PT grafts were fixed in the lateral femoral condyle of human cadaveric knees (mean age, 46 ⫾ 11 years) using 9 ⫻

932

C. H. BROWN, JR., ET AL.

25-mm metal interference screws inserted using a rear-entry technique. Specimens were cyclically loaded between 30 to 300 N at 1 Hz for 60,000 cycles. Three of 8 (38%) PT bone grafts fixed with interference screws failed during cyclic loading. We experienced a similar 38% failure rate for PT grafts fixed with interference screws during cyclic loading. In the study of Honl et al.23 the ultimate failure load for PT grafts fixed with interference screws after cyclic loading was 837 ⫾ 137 N, with a stiffness of 220 ⫾ 28 N/m. Their results compare favorably with our values for ultimate failure load and stiffness of 710 ⫾ 224 N, and 298 ⫾ 36 N/m. We speculate that the higher failure load reported by Honl et al.23 was due to their use of a 9-mm diameter screw. Our result of 0.34 mm for mean steady-state graft-bone motion compares favorably with their value of 0.2 mm. The load to failure (977 ⫾ 238 N) and stiffness (257 ⫾ 50 N/mm) values we report for DGST grafts fixed with the Bone Mulch Screw compare favorably with those reported by To et al.19 of 1,126 ⫾ 80 N, and 225 ⫾ 23 N/mm. The findings of that study are particularly relevant since fresh-frozen human distal femoral specimens similar in age to the specimens in our study were used, and tensile loading of the femur– graft fixation–ACL graft complex was also performed parallel to the axis of the femoral bone tunnel. Recently, Kousa et al.26 measured the tensile properties of human cadaveric gracilis and semitendinosus tendon grafts fixed in porcine femurs using 6 different commercially available hamstring tendon fixation techniques. In that study, displacement of the fixation techniques after cyclic loading followed by a load-tofailure test was performed. The specimens were loaded from 50 to 200 N for 1,500 cycles with the load applied parallel to the axis of the femoral bone tunnel. One 7 ⫻ 25-mm BioScrew (Linvatec, Largo, FL), and one 7 ⫻ 25-mm RCI screw (Smith & Nephew) failed under cyclic loading. The yield loads for the Bone Mulch Screw and BioScrew were reported to be 925 ⫾ 280 N and 565 ⫾ 137 N, respectively. The results for these fixation techniques are almost identical to the values we report in our study. However, the value we report for the EndoButton CL fixation technique (1,345 ⫾ 179 N) was much higher than their reported value of 781 ⫾ 252 N. A possible explanation for this difference is that in our study, the failure mode for the femur–EndoButton CL–DGST graft complex was rupture of the continuous loop in all specimens. In the study of Kousa et al.26 6 of 10 specimens failed by rupture of the continuous loop and 4 of 10 specimens failed by rupture of the hamstring tendon grafts. The

lower mean failure load reported by Kousa et al.26 probably reflects the lower failure load of specimens where failure occurred as a result of rupture of the hamstring tendon graft. Höher et al.14 measured graft-bone motion of DGST human hamstring tendon grafts fixed in a single human cadaveric lateral femoral condyle with an EndoButton using 15-, 25-, and 35-mm loops of 5-mm Mersilene tape. The femur–ACL graft fixation–ACL graft complex was cyclically loaded between a lower limit of 3 N and upper limits of 50, 100, 200, and 300 N. In that study, graft-bone motion was defined as the maximum displacement of the graft marker relative to the bone after 10 cycles. Graft-bone motion in that study corresponds to maximum graft-bone displacement in our current study, except we report displacement after 1,000 cycles rather than 10 cycles. Höher et al.14 reported graft-bone motion of the femur– graft fixation– hamstring tendon graft complex ranged from 1.2 ⫾ 0.3 to 5.8 ⫾ 0.2 mm under the above loading conditions. For DGST grafts fixed in the lateral femoral condyle using an EndoButton and a 20-mm loop of EndoButton Tape, we reported an average maximum displacement of 5.8 ⫾ 1.81 mm after 1,000 cycles with cyclic loading from 50 – 250 N, with 79% of the maximum displacement occurring within the first 20 cycles. There were several limitations of our study. First, because of the limited availability of young human cadaveric specimens, multiple biomechanical tests were conducted using the same femoral specimen, and hamstring and PT grafts. To minimize carry-over effects, we used a specific testing sequence. Fixation techniques dependent on the bone quality around the femoral tunnel were performed first, followed by suspensory fixation techniques where the fixation was dependent on the cortical bone of the distal femur, and finally suspensory techniques where the fixation was dependent on bone quality in the lateral and medial femoral condyles. The sequence of testing the suspensory techniques was designed such that techniques that required a smaller transcortical bone tunnel were tested first. Suspensory techniques that required the implant to be placed through the cancellous bone of the medial or lateral femoral condyle perpendicular to the axis of the bone tunnel were tested last. We believe that this testing sequence minimized the carryover effects of multiple tests on a single femoral specimen. During specimen preparation and biomechanical testing, the ACL grafts and bones were kept moist with saline solution. If damage to the femur or

ACL GRAFT TUNNEL MOTION ACL graft occurred, further tests using those specimens were suspended. Among the hamstring fixation techniques we experienced 1 failure of the TransFix technique during cyclic loading due to the implant toggling through the cancellous bone of the medial femoral condyle. Because the TransFix implant was the last fixation technique to be tested, it is possible that there may have been some carry-over effect from the previously drilled bone tunnels in the lateral and medial femoral condyles that compromised the quality of the cancellous bone, resulting in the implant toggling. Therefore, the tensile properties we report for hamstring tendon fixation with the TransFix cross-pin probably represent the lower limit for this fixation technique. The cross-over effect from previously drilled bone tunnels may have also contributed to the high incidence of fixation failure under cyclic loading we experienced for PT grafts fixed with interference screws. A second limitation of our study relates to the use of human cadaveric specimens. We attempted to minimize this effect by using middle-aged specimens. However, the mean age of the specimens used in our study is still older than the patient population in which ACL reconstruction is commonly performed. The mechanical properties of cancellous bone are known to deteriorate with specimen age.27,28 Fixation techniques in which the fixation strength of the implant is dependent on the bone density of cancellous bone, such as interference screws, cross-pins, and sutures placed through a bone block, would probably be more adversely affected by specimen age when compared with fixation techniques such as the EndoButton and LinX HT, in which the implant is anchored on the cortical bone of the distal femur.21 Additionally, the bone quality of the patellar bone blocks, and the bone-PT insertion sites are more likely to be affected by specimen age when compared with hamstring tendon grafts whose tensile properties are unaffected by specimen age.29 These factors may have contributed to the high rate of failure we experienced during cyclic loading of hamstring tendon grafts and PT grafts fixed with interference screws. A third limitation of our study relates to the direction of applied loading. In our biomechanical testing protocol, the femur-ACL graft complex was tested to failure with the load applied parallel to the long axis of the bone tunnel. This orientation aligns the ACL graft and femoral bone tunnel, subjecting the ACL graft and femoral fixation device to pure tensile forces. Clinically, maximum loading of the ACL graft occurs near full extension. At full knee extension, the ACL graft

933

and the axis of the femoral bone tunnels are no longer lined up. The oblique orientation of the bone tunnel and the direction of ACL graft loading in extension decrease the tension force on the ACL femoral fixation device. Therefore, our testing protocol represents a worst-case loading scenario. Another limitation of our study is the uniaxial direction of applied loading; our experimental design is capable of producing only longitudinal motion of the ACL graft in the bone tunnel. Thus our experimental design does not produce or allow us to measure the “windshield wiper effect” or graft-bone motion in the sagittal plane. Our study defines steady-state graft-bone motion as the displacement of the graft marker relative to the bone marker in each cycle, averaged over the full 1,000 cycles, and represents the amount of longitudinal motion of the ACL graft relative to the bone during each cycle. Longitudinal motion of the ACL graft relative to the bone tunnel has been referred to as the bungee effect, and has been one of the most commonly cited explanations for bone tunnel enlargement after ACL reconstruction.1,5 It has been suggested that fixation techniques that secure the ACL graft closer to the joint line (aperture fixation) will eliminate the bungee effect, decrease graft-tunnel motion, and reduce the incidence of bone tunnel widening compared with fixation methods that fix the ACL graft distant from the joint line (suspensory fixation).2,5,8,13,15 However, our study does not support this hypothesis. We found no significant difference in steady-state graft-bone motion between PT grafts fixed with interference screws (aperture fixation) and aperture or suspensory hamstring graft fixation techniques. Our results also show that suspensory or anatomic fixation techniques with similar stiffness values resulted in similar magnitudes of steady-state graftbone motion. These results suggest that it is the stiffness of the femur–ACL graft fixation–ACL graft technique rather than the point of graft fixation that is the most important determinant of graft-bone motion. L’Insalata et al.8 reported that both femoral and tibial tunnel enlargement is significantly greater following ACL reconstruction using hamstring tendon autografts compared with PT autografts. Tunnel enlargement in the hamstring tendon group was greater on the femoral versus the tibial side. The authors speculated that the greater incidence of tunnel widening following ACL reconstruction with hamstring tendon grafts was due to the fact that, compared with PT grafts, the point of fixation for hamstring tendon grafts was at a greater distance from the anatomic insertion of the normal ACL.

934

C. H. BROWN, JR., ET AL.

In their review of bone tunnel enlargement after ACL reconstruction, Höher et al.5 have suggested that bone tunnel enlargement might be prevented by using more anatomic graft fixation methods. However the prospective study of Clatworthy et al.1 suggests that graft-tunnel motion secondary to graft fixation distant from the joint line is unlikely to be the primary cause of bone tunnel enlargement after ACL reconstruction. This study evaluated radiographic bone tunnel widening in patients who had undergone ACL reconstruction using DGST grafts fixed on the femoral side with a Mitek ligament anchor and staples on the tibia, and patients who underwent reconstruction using a PT graft fixed with a Mitek ligament anchor on the femoral side and interference screw fixation on the tibia. Clatworthy et al.1 reported a mean increase in femoral tunnel widening of 100% in the hamstring reconstructions versus a 25% increase for the PT reconstructions. Since a suspensory type of fixation method was used for hamstring and PT grafts, the bungee cord effect was present in both groups of patients. If longitudinal motion or the bungee cord effect were primarily responsible for tunnel widening, one would have expected similar amounts of bone tunnel widening in both groups. The authors were also unable to find a correlation between a longer tunnel (bungee cord) and increased bone tunnel widening. Clatworthy et al.2 have also shown that bone tunnel enlargement cannot be avoided by fixing the ACL graft closer to the joint line, which minimizes or eliminates the so-called bungee effect. They reported an increase in mean tunnel area for DGST grafts fixed with BioScrews to be 122%, 89% when metal interference screws were used, 76% when the Bone Mulch Screw was used for femoral fixation, and 36% when the grafts where fixed in the femur using an EndoButton technique. If the bungee effect were primarily responsible for tunnel widening, one would have expected the greatest amount of tunnel widening to have occurred when the EndoButton technique was used and the least amount when metal or bioabsorbable interference screws were utilized. Based on the radiographic findings in their study, Clatworthy et al.2 concluded that graft-tunnel motion is not the primary cause of tunnel widening following ACL reconstruction. Clinical studies have found that the incidence and magnitude of bone tunnel enlargement after ACL reconstruction is greater for hamstring tendon autografts compared with PT autografts.8 It has been hypothesized that this difference is the result of increased longitudinal graft-tunnel motion caused by fixation

methods for hamstring tendon grafts that fix the graft distant from the anatomic insertion site of the normal ACL.5 However, our results show small differences in graft-bone motion between the most commonly used hamstring and PT femoral ACL graft fixation techniques. When the stiffness of the femur–ACL fixation–ACL graft technique was similar, we found no significant difference in graft-bone motion between aperture or suspensory hamstring tendon graft fixation techniques. We also found no significant difference in graft-bone motion between suspensory hamstring fixation techniques and aperture fixation of PT grafts using interference screws when the stiffness of the femur–ACL graft fixation–ACL graft complex was similar. Our findings provide further evidence that longitudinal graft-tunnel motion or the bungee effect is unlikely to be the primary cause of radiographic bone tunnel enlargement following ACL reconstruction. REFERENCES 1. Clatworthy MG, Annear P, Bulow JU, Bartlett RJ. Tunnel widening in anterior cruciate ligament reconstruction: a prospective evaluation of hamstring and patellar tendon grafts. Knee Surg Sports Traumol Arthrosc 1999;7:138-145. 2. Clatworthy MG, Bulow J-U, Pinczewski LA, et al. Tunnel widening in hamstring ACL reconstruction: A prospective clinical and radiographic evaluation of four different fixation techniques. Presented at the 68th Annual Meeting of the American Academy of Orthopaedic Surgeons, San Francisco, CA, 2001. Proceedings, Book of Abstracts of Poster Exhibits, Papers and Scientific Exhibits, p. 369. 3. Dyer CR, Elrod BF. Tibial and femoral bone tunnel enlargement following allograft replacement of the anterior cruciate ligament. Arthroscopy 1995; 11:353-354 (abstr). 4. Fahey M, Indelicato PA. Bone tunnel enlargement after anterior cruciate ligament replacement. Am J Sports Med 1994;22: 410-414. 5. Höher J, Möller HD, Fu F. Bone tunnel enlargement after anterior cruciate ligament reconstruction: Fact or fiction? Knee Surg Sports Traumatol Arthrosc 1998;6:231-240. 6. Jansson KA, Harilainen A, Sandelin J, et al. Bone tunnel enlargement with anterior cruciate ligament reconstruction with hamstring autograft and EndoButton fixation technique. Knee Surg Sports Traumatol Arthrosc 1999;7:290-295. 7. Linn RM, Fischer DA, Smith JP, et al. Achilles tendon allograft reconstruction of the anterior cruciate ligament-deficient knee. Am J Sports Med 1993;21:825-831. 8. L’Insalata JC, Klatt B, Fu FH, Harner CD. Tunnel expansion following ACL reconstruction: A comparison of hamstring and patellar tendon autografts. Knee Surg Sports Traumatol Arthrosc 1997;5:234-238. 9. Nebelung W, Becker R, Merkel M, Röpke M. Bone tunnel enlargement after anterior cruciate ligament reconstruction with semitendinosus tendon using EndoButton fixation on the femoral side. Arthroscopy 1998;14:810-815. 10. Peyrache MD, Dijam P. Christel P, Witvoet J. Tibial tunnel enlargement after anterior cruciate ligament reconstruction by autogenous bone–patellar tendon– bone graft. Knee Surg Sports Traumatol Arthrosc 1996;4:2-8.

ACL GRAFT TUNNEL MOTION 11. Schulte K, Majewski M, Irrgang JJ, Fu FH, Harner CD. Radiographic tunnel changes following arthroscopic ACL reconstruction: Autograft versus allograft. Arthroscopy 1996;11: 372-373. 12. Zijl JAC, Kleipool AEB, Willems WJ. Comparison of tibial tunnel enlargement after anterior cruciate ligament reconstruction using patellar tendon autograft or allograft. Am J Sports Med 2000;28:547-551. 13. Brand J, Weiler A, Caborn DN, et al. Current Concepts: Graft fixation in cruciate ligament reconstruction. Am J Sports Med 2000;28:761-774. 14. Höher J, Möller HD, Fu F. Hamstring graft motion in the femoral bone tunnel when using titanium button/polyester tape fixation. Knee Surg Sports Traumatol Arthrosc 1999;7:215219. 15. Ishibashi Y, Rudy TW, Liveday GA, Stone JD, et al. The effect of anterior cruciate ligament graft fixation site at the tibia on knee stability: Evaluation using a robotic testing system. Arthroscopy 1997;13:177-182. 16. Hamner DL, Brown CH, Steiner ME, et al. Hamstring tendon grafts for reconstruction of the anterior cruciate ligament: Biomechanical evaluation of the use of multiple strands and tensioning techniques. J Bone Joint Surg Am 1999;81:549557. 17. Cooper DE, Deng XH, Burstein AL, Warren RF. The strength of the central third patellar tendon graft: A biomechanical study. Am J Sports Med 1993;21:818-824. 18. Muellner T, Reihsner R, Mrkonjic W, et al. Twisting of patellar tendon grafts does not reduce their mechanical properties. J Biomechanics 1998;31:311-315. 19. To JT, Howell SM, Hull ML. Contributions of femoral fixation methods to the stiffness of anterior cruciate ligament replacement at implantation. Arthroscopy 1999;15:379-387. 20. Behrens JC, Walker PS, Shoji H. Variations in strength and structure of cancellous bone at the knee. J Biomechanics 1974;7:201-207.

935

21. Brand JC, Pienkowski D, Steenlage E, Hamilton D, Johnson DL, Caborn DNM. 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; 28:705-710. 22. Markolf KL, Gorek JF, Kabo M, Shapiro MS. Direct measurement of resultant forces in the anterior cruciate ligament. An in vitro study performed with a new experimental technique. J Bone Joint Surg Am 1990;72:557-567. 23. Honl M, Carrero V, Hille E, Schneider E, Morlock MM. Bone–patellar tendon– bone grafts for anterior cruciate ligament reconstruction. An in vitro comparison of mechanical behavior under failure tensile loading and cyclic submaximal tensile loading. Am J Sports Med 2002;30:549-557. 24. Beynnon BD, Amis AA. In vitro testing protocols for the cruciate ligaments and ligament reconstructions. Knee Surg Sports Traumatol Arthrosc 1998;6:S7-76 (suppl 1). 25. Weiss JA, Paulos LE. Mechanical testing of ligament fixation devices. Tech Orthop 1999;14:14-21. 26. Kousa T, Järvinen TLN, Vihavainen M, Kannus P, Järvinen M. The fixation strength of six hamstring tendon graft fixation devices in anterior cruciate ligament reconstruction. Part I: Femoral Site. Am J Sports Med 2003;31:174-181. 27. Burstein AH, Reilly DT, Martens M. Aging of bone tissue: Mechanical properties. J Bone Joint Surg Am 1976;58:82-86. 28. Weaver JK, Chalmers J. Cancellous bone: Its strength and changes with aging and an evaluation of some methods for measuring its mineral content. Part I: Age changes in cancellous bone. J Bone Joint Surg Am 1966;48:289-299. 29. Hecker AT, Brown CH, Deffner KT, Rosenberg TD. Tensile properties of young multiple stranded hamstring tendon grafts. In: Book of Abstracts and Outline Specialty Day. San Francisco, CA: American Orthopaedic Society for Sport Medicine, 1997;8.