Clinical Biomechanics 29 (2014) 230–234
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
A biomechanical comparison of anterior cruciate ligament suspensory fixation devices in a porcine cadaver model Lucas Rylander a, Jeffrey Brunelli a, Michal Taylor a, Todd Baldini a,⁎, Byron Ellis a, Monica Hawkins b, Eric McCarty a a b
University of Colorado—Denver, Aurora, CO, USA Stryker Joint Preservation, Mahwah, NJ, USA
a r t i c l e
i n f o
Article history: Received 7 May 2013 Accepted 1 November 2013 Keywords: Anterior cruciate ligament reconstruction surgery Knee arthroscopy
a b s t r a c t Background: Suspensory fixation use during anterior cruciate ligament reconstruction has increased due to ease of use and high pullout strength. We hypothesize that there are no significant differences in biomechanical performance among four types of suspensory fixation devices: Stryker VersiTomic G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton. Methods: Forty fresh frozen porcine femurs and flexor digitorum profundus tendons were obtained. Each tendon graft was sized to 8.5 mm or 9.0 mm. Ten of each device were used to fix the grafts in the femur at the 2 o'clock (left) or 10 o'clock (right) position. The graft–femur complex was secured to a servohydraulic test machine in line with the femoral tunnel. The graft was cyclically loaded from 50 to 250 N for 1000 cycles at 1 Hz then loaded to failure at 20 mm/min. Actuator load and displacement were recorded. Data were analyzed with multiple oneway ANOVA and Tukey HSD post-hoc tests. Bonferroni correction was applied resulting in P ≤ 0.005 considered statistically significant for ANOVA, P ≤ 0.05 for Tukey. Findings: There were no significant differences in cyclic displacement among any of the groups (P = 0.43). The only significant difference in failure properties is the Endobutton exhibited at least 50% greater displacement at failure than the other three devices. Interpretation: Suspensory femoral soft tissue fixation devices are biomechanically similar with respect to failure load but differ in failure displacement. However, there was no significant difference in displacement after cyclic loading. All four fixation devices should withstand the forces associated with daily activities without failure. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction A majority of anterior cruciate ligament (ACL) injuries are treated with ACL reconstruction surgery. Unfortunately not all ACL reconstructions are successful. The failure/reoperation rate has been reported to be as low as 3% and as high as 27% (Allen et al., 2003; Bach, 2003; Fox et al., 2004; Miyasaka et al., 1991; Strand et al., 2005; Weiler et al., 2007). Failure may be attributed to recurrent instability, graft failure, or arthrofibrosis (Miyasaka et al., 1991). There are multiple potential causes of mechanical failure with respect to graft fixation on the femoral side, including graft slippage, loss of fixation, and graft rupture. Failure can occur with repetitive sub-threshold failure loads or as a result of one catastrophic event (Miyasaka et al., 1991). Daily activities such as standing from a seated position place forces on the ACL ranging from 59 to 100 N, and up to 445 N of force is generated when walking down an incline (Kvist and Gillquist, 2001; Noyes et al., 1984). Clinical graft failure has been defined as greater than 5 mm of graft lengthening which can lead to meniscus and cartilage damage (Conner et al., 2010). ⁎ Corresponding author at: 13001 E. 17th Place, MS F432, Aurora, CO 80045, USA. E-mail address: [email protected] (T. Baldini). 0268-0033/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2013.11.001
Ideal fixation techniques would withstand daily activity forces without graft lengthening or failure until the graft is fully incorporated into the femoral tunnel (Beynnon et al., 2005; Conner et al., 2010; Quatman et al., 2011; Rodeo et al., 2006; Wilson et al., 2004). There are multiple options available for femoral-side graft fixation. Recent femoral tunnel drilling techniques including retrograde and medial portal drilling have made it more challenging to use transfixation pins. Suspension fixation techniques have been shown to have higher pullout strength in soft tissue grafts when compared to interference screws resulting in more frequent use of suspensory fixation devices (Ahmad et al., 2004). With several suspensory fixation devices currently available it is important to have a direct biomechanical comparison among the devices to help the surgeon decide which device to use. This study aims to determine if there is a significant difference in biomechanical performance in an in-vitro porcine model among four types of suspensory fixation devices: Stryker VersiTomic G-Lok (Stryker Joint Preservation, Mahwah, NJ, USA), Smith & Nephew Endobutton (Smith & Nephew, Andover, MA, USA), Biomet ToggleLoc (Biomet Sports Medicine, Warsaw, IN, USA), and Arthrex RetroButton (Arthrex, Naples, FL, USA). The hypothesis is that there are no significant differences in biomechanical performance among the four devices tested.
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
2. Methods Forty fresh frozen 2-year-old Yorkshire-cross porcine hind limbs (10 per group) were obtained from an abattoir (Thomas D. Morris, Inc., Reisterstown, MD, USA). No IRB approval is needed by our institution for testing of dead animal tissue. The femurs were removed of all flesh and their flexor digitorum profundus (FDP) tendons were harvested for grafts. Porcine femora were chosen due to the ability to select specimens with a BMD (bone mineral density) that is similar to that of healthy, active, young adult humans (Conner et al., 2010; EspejoBaena et al., 2006; Miyata et al., 2000). Young human cadaver specimens are difficult to obtain, whereas there is a high availability of porcine tissue available from abattoirs. Porcine flexor digitorum profundus tendons were also chosen based on their high availability and due to their consistent size and mechanical properties similar to a human hamstring tendon (Conner et al., 2010). The FDPs were folded in half to make double bundled grafts and at least 70 mm in length. The graft size was measured with a standard measuring block ranging from 4.5 mm to 12.5 mm in 0.5 mm increments (Stryker Joint Preservation, Mahwah, NJ, USA). Of the forty FDPs tested, 33 were 9.0 mm and 7 were 8.5 mm. Of the seven 8.5 mm FDPs, 1 was tested with the Endobutton, 1 with the G-Lok, 2 with the ToggleLoc, and 3 with the RetroButton. The device used was randomly assigned to a specimen resulting in unequal numbers of specimens tested with 8.5 mm grafts. The loose ends of the grafts were whip stitched with number 2 Force Fiber (Stryker Endoscopy, San Jose, CA, USA). The femoral tunnel was prepared by first drilling a 2.4 mm guide pin through the lateral cortex. The 9 mm diameter socket tunnel was then reamed to the appropriate depth for a 20 mm suspensory fixation device by over drilling the guide pin. After the socket tunnel was reamed a 4.5 mm reamer was used over the guide pin and drilled through the lateral cortex. The double bundled
231
grafts were then placed in the femur at the 2 o'clock (left) or 10 o'clock (right) position using a Stryker G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, or Arthrex RetroButton (Fig. 1). All suspensory devices had 20 mm loops. Suspensory fixation was placed through the femoral tunnel onto the lateral cortex, and the button was toggled flush against the cortex in order to place maximum force perpendicular to the bone surface. The femur was secured to the base of an Instron Model 1351 (Instron Inc., Canton, MA, USA) servohydraulic test machine with a custom built test fixture (Fig. 2). The distance from the bottom edge of the fixture to the graft was 3 cm. The femoral tunnel was aligned with the axis of the actuator to maximize the tension on the suspensory fixation device and reduce the load sharing at the graft–tunnel interface. The graft was fixed in a custom built cryo-clamp. The distance from the edge of the cryoclamp to the cortex of the femur was 4 cm. The graft was cyclically loaded from 50 to 250 N for 1000 cycles at 1 Hz then loaded to failure at 20 mm/min (Ahmad et al., 2004). Actuator load and displacement were recorded at 100 Hz on a PC equipped with a Keithley 1802HC (Keithley Instruments Inc., Cleveland, OH, USA) analog to digital board and TestPoint (Capital Equipment Corp., Billerica, MA, USA) data acquisition software. The mode of failure was noted. The results reported from the cyclic data are displacement after 1, 10, 50, 100, and 1000 cycles. The results reported from the single load to failure are the failure load and displacement at failure (defined as the first peak in load), yield load and displacement at yield defined at 0.2% offset, and stiffness. To determine the yield point the load and deflection data were converted to stress and strain. Stress was calculated by dividing the load by the cross sectional area of the graft that was either 9.0 mm or 8.5 mm in diameter. The strain was calculated as the displacement of the actuator divided by the initial length of the graft
Fig. 1. Suspensory fixation devices: A. Stryker VersiTomic G-Lok, B. Smith & Nephew Endobutton, C. Biomet ToggleLoc, D. Arthrex RetroButton.
232
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234 Table 1 Load to failure data: Mean (standard deviation). Device
Load to failure (N)
G-Lok
613.6 (176.0) Endobutton 716.7 (128.2) ToggleLoc 559.7 (101.3) RetroButton 525.5 (160.2) ANOVA P 0.04 F 3.14 F critical 4.46
Fig. 2. Test set-up showing the femur and graft in the test fixtures.
from the femur to the clamp, 40 mm. A line parallel to the linear section of the stress strain curve offset by 0.2% was plotted on the stress vs. strain curve. The point of intersection of the two graphs was defined as the yield point. A one-way ANOVA was used to detect significant differences in each of the following parameters, displacement at 1, 10, 50, 100, 1000 cycles, failure load and displacement at failure, yield load and displacement at yield, and stiffness, as a function of implant type using JMP 9 statistical analysis software (SAS Institute Inc., Cary, NC, USA). Because 10 multiple comparisons were made on both the cyclic data and the failure data the Bonferroni correction was used for both data sets with n = 10. Tukey's HSD post-hoc test was performed for all pair-wise comparisons upon rejection of the null hypothesis. Differences were considered significant at P ≤ 0.005 for ANOVA, P ≤ 0.05 for Tukey. 3. Results A small number of specimens failed during cyclic testing. All specimens that failed during cyclic loading failed by the device pulling through the cortex. One G-Lok failed after 51 cycles, no Endobuttons failed during cyclic testing, one ToggleLoc failed at 523 cycles, and two RetroButtons failed at 11 and 719 cycles. There were no significant differences in displacement among the four devices after 1, 10, 50, 100, and 1000 cycles with P values of 0.16, 0.10, 0.10, 0.13, and 0.43, respectively. Plotted data are shown in Fig. 3. All the specimens that survived the cyclic testing were loaded with a single load to failure. The only significant difference in failure properties is the Endobutton that had significantly greater failure displacement than the G-Lok by 50% (P = 0.005), the ToggleLoc by 50% (P = 0.005), and the RetroButton by 56% (P = 0.003). There were no other significant
Displacement Yield load at failure (N) (mm)
Stiffness Yield displacement (N/mm) (mm)
9.78⁎ (3.75)
5.51 (1.48)
14.67 (2.89) 9.77⁎ (1.65) 9.36⁎ (2.98) b0.001 7.18 4.46
386.0 (37.8) 406.9 (72.1) 359.9 (51.0) 345.9 (39.1) 0.09 2.38 4.46
7.21 (1.61) 5.86 (0.90) 5.90 (0.65) 0.03 3.41 4.46
99.7 (16.1) 79.1 (11.7) 94.0 (12.4) 88.3 (14.1) 0.02 3.99 4.46
n
9 10 9 8
Degrees of freedom: Between groups = 3, within groups = 32, total = 35. ⁎ Denotes significant difference from Endobutton (P b 0.005).
differences in failure displacement. There were no significant differences in failure load, yield load, displacement at yield, and stiffness. The failure data are shown in Table 1. Two G-Lok specimens failed by the tendon tearing at the loop. All other specimens failed by the device pulling through the cortex. 4. Discussion Successful ACL reconstruction surgery must maintain graft fixation in the femur (and tibia) long enough for it to be biologically incorporated into the bone (Beynnon et al., 2005; Conner et al., 2010; Quatman et al., 2011; Rodeo et al., 2006; Wilson et al., 2004). Failures can occur with repetitive sub-threshold loads or from one catastrophic event. To increase the success of ACL reconstruction surgery, suspensory fixation devices are being used more frequently due to clinical ease and superior mechanical test results when compared to other fixation methods including interference screws (Ahmad et al., 2004). The purpose of this study was to determine if significant differences exist in the biomechanical performance of 4 different suspensory fixation devices, the Stryker VersiTomic G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton in cyclic loading and single load to failure in an in-vitro porcine model. The data showed no significant differences in displacement during cyclic loading to 250 N for 1000 cycles among the 4 different devices tested. However, only the Endobuttons had 100% survival during cyclic testing. One G-Lok, one ToggleLoc, and two RetroButtons failed during cyclic loading. Shelburne et al. calculated that the force on the ACL during normal gait can reach 300 N (Shelburne et al., 2004). This was a theoretical model with ideal conditions and estimations that vary with normal gait patterns. Therefore, as a model of ligamentous strain,
Fig. 3. Cyclic data — Graft elongation.
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
it may apply, but only in ideal conditions. Also, limitations are placed on rehab protocols to limit the number of times the graft is strained while being incorporated into ACL tunnels. There was not a significant difference in the single load to failure performance of the devices and all devices had failure loads greater than the 300 N calculated load that the ACL must withstand during gait (Shelburne et al., 2004). This is important but probably less important than withstanding cyclic loads because the ACL shouldn't be significantly loaded until it has time to properly heal, typically 8–12 weeks, unless a traumatic event occurs in the early postoperative period (Conner et al., 2010). Kamelger et al. conducted a similar study on 3 suspensory fixation devices (ToggleLoc, RetroButton, and Endobutton) in an in vitro 16week-old porcine model (Kamelger et al., 2009). They found the Endobutton to have a significantly greater load to failure than the RetroButton. This study found a similar result. The two studies differed in the failure mode. Kamelger et al. had 50% of the RetroButtons and 50% of the Endobuttons failing at the loop material but none of the ToggleLocs failing at the loop material (Kamelger et al., 2009). None of the suspensory fixation devices failed at the loop material in the current study. However, the current study used a soft tissue tendon graft. Kamelger at al. used a metal hook to load the suspensory fixation device. This could explain the difference in failure modes. The loading protocol in this study was similar to that used by Ahmad et al. to test several fixation devices, including the Endobutton (Ahmad et al., 2004). Both tests used porcine femurs but Ahmad et al. used 10 mm bovine grafts. This study used 9 mm porcine grafts. The different graft size and different loading rate, 200 mm/min vs. 20 mm/min (current study), may account for the difference in failure load, 864 N for the 10 mm graft and 716 N for the 9 mm grafts. Ahmad et al. reported no Endobutton failures during cyclic loading and all failures during single load to failure were by the device pulling through the cortex, which was the same for the Endobutton in this study (Ahmad et al., 2004). The graft displacement between the two studies also cannot be directly compared. Ahmad et al. measured graft slippage between the edge of the femoral tunnel and a suture placed in the graft 2 mm from the edge of the femur (Ahmad et al., 2004). This study measured the overall displacement of the graft, the fixation device, and the compliance of the system with the displacement of the actuator. Conner et al. conducted cyclic loading and load to failure tests in a porcine model with ToggleLoc and Endobutton fixation devices (Conner et al., 2010). Their grafts were all 8 mm in diameter and their testing protocol differed. They loaded the specimens cyclically from 50 N to 450 N and single load to failure at 1800 mm/min. Eight of 16 Endobuttons and 7 of 16 ToggleLocs failed during cyclic loading. Conner et al. found significantly greater total displacement in the ToggleLoc than the Endobutton during cyclic loading. The current study found no significant differences in displacement between these two devices during cyclic testing. However, both studies used different loading protocols. Conner et al. did not find a significant difference in load to failure or stiffness between the ToggleLoc and EndoButton, which is similar to the current study (Conner et al., 2010). The authors of this study chose to use a porcine model because specimens of the same age, sex, and similar size were available which would reduce the variability in tissue quality among specimens (Conner et al., 2010; Espejo-Baena et al., 2006; Miyata et al., 2000). The specimens are presumed to have a bone mineral density (BMD) similar to that of young adult humans, although no bone density measurements were obtained in the current study to verify this similarity, or variability, among individual porcine specimens. Further, most human cadaver models vary in BMD because of age and sex differences and are usually from an older population. It is important to note that the failure mechanism of nearly all specimens was via device pullout through the femoral cortex. Only two graft ruptures occurred, and no specimens failed due to suture loop rupture. These results suggest that the most common mechanism of failure with
233
suspensory fixation is a problem at the bone–button interface. Future suspensory device designs may be well advised to focus on ways to improve this particular interface in an effort to minimize device failure. Increasing the length of the device would increase the contact area with the cortex without requiring a larger hole to pass the device through. The current study has several limitations. This study did not directly measure graft elongation. It measured actuator displacement. By measuring actuator displacement it is not possible to determine percentages due to graft elongation, loop elongation, or the device migration into the bone. The authors consider actuator displacement to more accurately represent what is happening in situ because graft elongation, loop elongation, and device migration into the bone all contribute to knee laxity. In addition to being able to withstand the cyclic loads and not failing catastrophically the graft should not elongate a significant amount that could result in knee laxity. Clinical graft failure has been defined as greater than 5 mm of graft lengthening (Noyes et al., 1984). All 4 of the devices had greater than 5 mm of actuator displacement after 1000 cycles but we were unable to determine how much of the displacement was due to graft lengthening. However, the grafts were not pre-conditioned before cyclic testing as done in the operating room to remove most permanent visco-elastic elongation that may occur in vivo (Nurmi et al., 2004). Because the visco-elastic elongation was not removed prior to testing they could have exhibited greater elongation than would be seen clinically. To compare the results to a more clinically relevant situation, the first ten cycles can be considered to be preconditioning cycles. When the displacement after 10 cycles is removed from the displacement at 1000 cycles the results would range from 3.32 to 3.85 mm, less than the clinical graft failure definition of 5 mm. Another potential limitation is the usage of both 8.5 mm and 9 mm grafts sizes in 9 mm diameter femoral tunnels; i.e., seven of 40 grafts were 8.5 mm in diameter but tested in 9 mm diameter tunnels. The effect of different graft sizes was minimized by loading the graft in line with the tunnel, potentially decreasing graft contact with the wall of the tunnel. Another limitation, as noted above, is that bone density measurements of individual porcine specimens were not obtained in the current study. Finally, this was a cadaver model and the effect of healing and bone tunnel enlargement on the biomechanical properties of the fixation devices could not be evaluated.
5. Conclusions Recent increases in suspensory fixation device use make it vital to know their biomechanical performance. This study tested the biomechanical performance of 4 different suspensory fixation devices (Stryker G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton) in cyclic loading followed by a single load to failure. There were no significant differences in displacement during cyclic loading. There was only a significant difference in displacement at failure, but all grafts were able to support loads greater than that seen during normal gait. Therefore, device selection is not critical if a surgeon wants a device that will withstand normal loads in the post-op rehab period.
Conflict of interest statement Dr. Monica Hawkins is employed by Stryker Corp. and owns stock/stock options in Stryker Corp. Dr. Eric McCarty, Dr. Lucas Rylander, Dr. Jeffrey Brunelli, Dr. Michal Taylor, Todd Baldini, MS, and Byron Ellis, BS, have no financial and personal relationships with other people or organizations that could inappropriately influence (bias) this work.
234
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
Acknowledgments This study was funded by Stryker Joint Preservation. Dr. Monica Hawkins is employed by Stryker Joint Preservation and was involved in the design of the study and editing the manuscript. However, Dr. Hawkins was not involved in the decision to submit the manuscript.
References Ahmad, C.S., Gardner, T.R., Groh, M., Arnouk, J., Levine, W.N., 2004. Mechanical properties of soft tissue femoral fixation devices for anterior cruciate ligament reconstruction. Am. J. Sports Med. 32, 635–640. Allen, C.R., Giffin, J.R., Harner, C.D., 2003. Revision anterior cruciate ligament reconstruction. Orthop. Clin. N. Am. 34, 79–98. Bach Jr., B.R., 2003. Revision anterior cruciate ligament surgery. Arthroscopy 19 (Suppl. 1), 14–29. Beynnon, B.D., Johnson, R.J., Abate, J.A., Fleming, B.C., Nichols, C.E., 2005. Treatment of anterior cruciate ligament injuries, part 2. Am. J. Sports Med. 33, 1751–1767. Conner, C.S., Perez, B.A., Morris, R.P., Buckner, J.W., Buford, W.L., Ivey, F.M., 2010. Three femoral fixation devices for anterior cruciate ligament reconstruction: comparison of fixation on the lateral cortex versus the anterior cortex. Arthroscopy 26, 796–807. Espejo-Baena, A., Ezquerro, F., de la Blanca, A.P., Serrano-Fernandez, J., Nadal, F., Montanez-Heredia, E., 2006. Comparison of initialmechanical properties of 4 hamstring graft femoral fixation systems using nonpermanent hardware for anterior cruciate ligament reconstruction: an in vitro animal study. Arthroscopy 22, 433–440. Fox, J.A., Pierce, M., Bojchuk, J., Hayden, J., Bush-Joseph, C.A., Bach Jr., B.R., 2004. Revision anterior cruciate ligament reconstruction with nonirradiated fresh-frozen patellar tendon allograft. Arthroscopy 20, 787–794.
Kamelger, F.S., Onder, U., Schmoelz, W., Tecklenburg, K., Arora, R., Fink, C., 2009. Suspensory fixation of grafts in anterior cruciate ligament reconstruction: a biomechanical comparison of 3 implants. Arthroscopy 25, 767–776. Kvist, J., Gillquist, J., 2001. Sagittal plane knee translation and electromyographic activity during closed and open kinetic chain exercises in anterior cruciate ligament-deficient patients and control subjects. Am. J. Sports Med. 29, 72–82. Miyasaka, K.C., Daniel, D.M., Stone, M.L., 1991. The incidence of knee ligament injuries in the general population. Am. J. Knee Surg. 4, 43–48. Miyata, K., Yasuda, K., Kondo, E., Nakano, H., Kimura, S., Hara, N., 2000. Biomechanical comparisons of anterior cruciate ligament: reconstruction procedures with flexor tendon graft. J. Orthop. Sci. 5, 585–592. Noyes, F.R., Butler, D.L., Grood, E.S., Zernicke, R.F., Hefzy, M.S., 1984. Biomechanical analysis of human ligament grafts used in knee-ligament repairs and reconstructions. J. Bone Joint Surg. Am. 66, 344–352. Nurmi, J.T., Kannus, P., Sievänen, H., Järvelä, T., Järvinen, M., Järvinen, T.L., 2004. Interference screw fixation of soft tissue grafts in anterior cruciate ligament reconstruction: part 2: effect of preconditioning on graft tension during and after screw insertion. Am. J. Sports Med. 32, 418–424. Quatman, C.E., Paterno, M.V., Wordeman, S.C., Kaeding, C.C., 2011. Longitudinal anterior knee laxity related to substantial tibial tunnel enlargement after anterior cruciate ligament revision. Arthroscopy 27, 1160–1163. Rodeo, S.A., Kawamura, S., Kim, H.J., Dynybil, C., Ying, L., 2006. Tendon healing in a bone tunnel differs at the tunnel entrance versus the tunnel exit: an effect of graft–tunnel motion? Am. J. Sports Med. 34, 1790–1800. Shelburne, K.B., Pandy, M.G., Anderson, F.C., Torry, M.R., 2004. Pattern of anterior cruciate ligament force in normal walking. J. Biomech. 37, 797–805. Strand, T., Mølster, A., Hordvik, M., Krukhaug, Y., 2005. Long-term follow-up after primary repair of the anterior cruciate ligament: clinical and radiological evaluation 15–23 years postoperatively. Arch. Orthop. Trauma Surg. 125, 217–221. Weiler, A., Schmeling, A., Stöhr, I., Kääb, M.J., Wagner, M., 2007. Primary versus single-stage revision anterior cruciate ligament reconstruction using autologous hamstring tendon grafts: a prospective matched-group analysis. Am. J. Sports Med. 35, 1643–1652. Wilson, T.C., Kantaras, A., Atay, A., Johnson, D.L., 2004. Tunnel enlargement after anterior cruciate ligament surgery. Am. J. Sports Med. 32, 543–549.
Contents lists available at ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
A biomechanical comparison of anterior cruciate ligament suspensory fixation devices in a porcine cadaver model Lucas Rylander a, Jeffrey Brunelli a, Michal Taylor a, Todd Baldini a,⁎, Byron Ellis a, Monica Hawkins b, Eric McCarty a a b
University of Colorado—Denver, Aurora, CO, USA Stryker Joint Preservation, Mahwah, NJ, USA
a r t i c l e
i n f o
Article history: Received 7 May 2013 Accepted 1 November 2013 Keywords: Anterior cruciate ligament reconstruction surgery Knee arthroscopy
a b s t r a c t Background: Suspensory fixation use during anterior cruciate ligament reconstruction has increased due to ease of use and high pullout strength. We hypothesize that there are no significant differences in biomechanical performance among four types of suspensory fixation devices: Stryker VersiTomic G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton. Methods: Forty fresh frozen porcine femurs and flexor digitorum profundus tendons were obtained. Each tendon graft was sized to 8.5 mm or 9.0 mm. Ten of each device were used to fix the grafts in the femur at the 2 o'clock (left) or 10 o'clock (right) position. The graft–femur complex was secured to a servohydraulic test machine in line with the femoral tunnel. The graft was cyclically loaded from 50 to 250 N for 1000 cycles at 1 Hz then loaded to failure at 20 mm/min. Actuator load and displacement were recorded. Data were analyzed with multiple oneway ANOVA and Tukey HSD post-hoc tests. Bonferroni correction was applied resulting in P ≤ 0.005 considered statistically significant for ANOVA, P ≤ 0.05 for Tukey. Findings: There were no significant differences in cyclic displacement among any of the groups (P = 0.43). The only significant difference in failure properties is the Endobutton exhibited at least 50% greater displacement at failure than the other three devices. Interpretation: Suspensory femoral soft tissue fixation devices are biomechanically similar with respect to failure load but differ in failure displacement. However, there was no significant difference in displacement after cyclic loading. All four fixation devices should withstand the forces associated with daily activities without failure. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction A majority of anterior cruciate ligament (ACL) injuries are treated with ACL reconstruction surgery. Unfortunately not all ACL reconstructions are successful. The failure/reoperation rate has been reported to be as low as 3% and as high as 27% (Allen et al., 2003; Bach, 2003; Fox et al., 2004; Miyasaka et al., 1991; Strand et al., 2005; Weiler et al., 2007). Failure may be attributed to recurrent instability, graft failure, or arthrofibrosis (Miyasaka et al., 1991). There are multiple potential causes of mechanical failure with respect to graft fixation on the femoral side, including graft slippage, loss of fixation, and graft rupture. Failure can occur with repetitive sub-threshold failure loads or as a result of one catastrophic event (Miyasaka et al., 1991). Daily activities such as standing from a seated position place forces on the ACL ranging from 59 to 100 N, and up to 445 N of force is generated when walking down an incline (Kvist and Gillquist, 2001; Noyes et al., 1984). Clinical graft failure has been defined as greater than 5 mm of graft lengthening which can lead to meniscus and cartilage damage (Conner et al., 2010). ⁎ Corresponding author at: 13001 E. 17th Place, MS F432, Aurora, CO 80045, USA. E-mail address: [email protected] (T. Baldini). 0268-0033/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2013.11.001
Ideal fixation techniques would withstand daily activity forces without graft lengthening or failure until the graft is fully incorporated into the femoral tunnel (Beynnon et al., 2005; Conner et al., 2010; Quatman et al., 2011; Rodeo et al., 2006; Wilson et al., 2004). There are multiple options available for femoral-side graft fixation. Recent femoral tunnel drilling techniques including retrograde and medial portal drilling have made it more challenging to use transfixation pins. Suspension fixation techniques have been shown to have higher pullout strength in soft tissue grafts when compared to interference screws resulting in more frequent use of suspensory fixation devices (Ahmad et al., 2004). With several suspensory fixation devices currently available it is important to have a direct biomechanical comparison among the devices to help the surgeon decide which device to use. This study aims to determine if there is a significant difference in biomechanical performance in an in-vitro porcine model among four types of suspensory fixation devices: Stryker VersiTomic G-Lok (Stryker Joint Preservation, Mahwah, NJ, USA), Smith & Nephew Endobutton (Smith & Nephew, Andover, MA, USA), Biomet ToggleLoc (Biomet Sports Medicine, Warsaw, IN, USA), and Arthrex RetroButton (Arthrex, Naples, FL, USA). The hypothesis is that there are no significant differences in biomechanical performance among the four devices tested.
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
2. Methods Forty fresh frozen 2-year-old Yorkshire-cross porcine hind limbs (10 per group) were obtained from an abattoir (Thomas D. Morris, Inc., Reisterstown, MD, USA). No IRB approval is needed by our institution for testing of dead animal tissue. The femurs were removed of all flesh and their flexor digitorum profundus (FDP) tendons were harvested for grafts. Porcine femora were chosen due to the ability to select specimens with a BMD (bone mineral density) that is similar to that of healthy, active, young adult humans (Conner et al., 2010; EspejoBaena et al., 2006; Miyata et al., 2000). Young human cadaver specimens are difficult to obtain, whereas there is a high availability of porcine tissue available from abattoirs. Porcine flexor digitorum profundus tendons were also chosen based on their high availability and due to their consistent size and mechanical properties similar to a human hamstring tendon (Conner et al., 2010). The FDPs were folded in half to make double bundled grafts and at least 70 mm in length. The graft size was measured with a standard measuring block ranging from 4.5 mm to 12.5 mm in 0.5 mm increments (Stryker Joint Preservation, Mahwah, NJ, USA). Of the forty FDPs tested, 33 were 9.0 mm and 7 were 8.5 mm. Of the seven 8.5 mm FDPs, 1 was tested with the Endobutton, 1 with the G-Lok, 2 with the ToggleLoc, and 3 with the RetroButton. The device used was randomly assigned to a specimen resulting in unequal numbers of specimens tested with 8.5 mm grafts. The loose ends of the grafts were whip stitched with number 2 Force Fiber (Stryker Endoscopy, San Jose, CA, USA). The femoral tunnel was prepared by first drilling a 2.4 mm guide pin through the lateral cortex. The 9 mm diameter socket tunnel was then reamed to the appropriate depth for a 20 mm suspensory fixation device by over drilling the guide pin. After the socket tunnel was reamed a 4.5 mm reamer was used over the guide pin and drilled through the lateral cortex. The double bundled
231
grafts were then placed in the femur at the 2 o'clock (left) or 10 o'clock (right) position using a Stryker G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, or Arthrex RetroButton (Fig. 1). All suspensory devices had 20 mm loops. Suspensory fixation was placed through the femoral tunnel onto the lateral cortex, and the button was toggled flush against the cortex in order to place maximum force perpendicular to the bone surface. The femur was secured to the base of an Instron Model 1351 (Instron Inc., Canton, MA, USA) servohydraulic test machine with a custom built test fixture (Fig. 2). The distance from the bottom edge of the fixture to the graft was 3 cm. The femoral tunnel was aligned with the axis of the actuator to maximize the tension on the suspensory fixation device and reduce the load sharing at the graft–tunnel interface. The graft was fixed in a custom built cryo-clamp. The distance from the edge of the cryoclamp to the cortex of the femur was 4 cm. The graft was cyclically loaded from 50 to 250 N for 1000 cycles at 1 Hz then loaded to failure at 20 mm/min (Ahmad et al., 2004). Actuator load and displacement were recorded at 100 Hz on a PC equipped with a Keithley 1802HC (Keithley Instruments Inc., Cleveland, OH, USA) analog to digital board and TestPoint (Capital Equipment Corp., Billerica, MA, USA) data acquisition software. The mode of failure was noted. The results reported from the cyclic data are displacement after 1, 10, 50, 100, and 1000 cycles. The results reported from the single load to failure are the failure load and displacement at failure (defined as the first peak in load), yield load and displacement at yield defined at 0.2% offset, and stiffness. To determine the yield point the load and deflection data were converted to stress and strain. Stress was calculated by dividing the load by the cross sectional area of the graft that was either 9.0 mm or 8.5 mm in diameter. The strain was calculated as the displacement of the actuator divided by the initial length of the graft
Fig. 1. Suspensory fixation devices: A. Stryker VersiTomic G-Lok, B. Smith & Nephew Endobutton, C. Biomet ToggleLoc, D. Arthrex RetroButton.
232
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234 Table 1 Load to failure data: Mean (standard deviation). Device
Load to failure (N)
G-Lok
613.6 (176.0) Endobutton 716.7 (128.2) ToggleLoc 559.7 (101.3) RetroButton 525.5 (160.2) ANOVA P 0.04 F 3.14 F critical 4.46
Fig. 2. Test set-up showing the femur and graft in the test fixtures.
from the femur to the clamp, 40 mm. A line parallel to the linear section of the stress strain curve offset by 0.2% was plotted on the stress vs. strain curve. The point of intersection of the two graphs was defined as the yield point. A one-way ANOVA was used to detect significant differences in each of the following parameters, displacement at 1, 10, 50, 100, 1000 cycles, failure load and displacement at failure, yield load and displacement at yield, and stiffness, as a function of implant type using JMP 9 statistical analysis software (SAS Institute Inc., Cary, NC, USA). Because 10 multiple comparisons were made on both the cyclic data and the failure data the Bonferroni correction was used for both data sets with n = 10. Tukey's HSD post-hoc test was performed for all pair-wise comparisons upon rejection of the null hypothesis. Differences were considered significant at P ≤ 0.005 for ANOVA, P ≤ 0.05 for Tukey. 3. Results A small number of specimens failed during cyclic testing. All specimens that failed during cyclic loading failed by the device pulling through the cortex. One G-Lok failed after 51 cycles, no Endobuttons failed during cyclic testing, one ToggleLoc failed at 523 cycles, and two RetroButtons failed at 11 and 719 cycles. There were no significant differences in displacement among the four devices after 1, 10, 50, 100, and 1000 cycles with P values of 0.16, 0.10, 0.10, 0.13, and 0.43, respectively. Plotted data are shown in Fig. 3. All the specimens that survived the cyclic testing were loaded with a single load to failure. The only significant difference in failure properties is the Endobutton that had significantly greater failure displacement than the G-Lok by 50% (P = 0.005), the ToggleLoc by 50% (P = 0.005), and the RetroButton by 56% (P = 0.003). There were no other significant
Displacement Yield load at failure (N) (mm)
Stiffness Yield displacement (N/mm) (mm)
9.78⁎ (3.75)
5.51 (1.48)
14.67 (2.89) 9.77⁎ (1.65) 9.36⁎ (2.98) b0.001 7.18 4.46
386.0 (37.8) 406.9 (72.1) 359.9 (51.0) 345.9 (39.1) 0.09 2.38 4.46
7.21 (1.61) 5.86 (0.90) 5.90 (0.65) 0.03 3.41 4.46
99.7 (16.1) 79.1 (11.7) 94.0 (12.4) 88.3 (14.1) 0.02 3.99 4.46
n
9 10 9 8
Degrees of freedom: Between groups = 3, within groups = 32, total = 35. ⁎ Denotes significant difference from Endobutton (P b 0.005).
differences in failure displacement. There were no significant differences in failure load, yield load, displacement at yield, and stiffness. The failure data are shown in Table 1. Two G-Lok specimens failed by the tendon tearing at the loop. All other specimens failed by the device pulling through the cortex. 4. Discussion Successful ACL reconstruction surgery must maintain graft fixation in the femur (and tibia) long enough for it to be biologically incorporated into the bone (Beynnon et al., 2005; Conner et al., 2010; Quatman et al., 2011; Rodeo et al., 2006; Wilson et al., 2004). Failures can occur with repetitive sub-threshold loads or from one catastrophic event. To increase the success of ACL reconstruction surgery, suspensory fixation devices are being used more frequently due to clinical ease and superior mechanical test results when compared to other fixation methods including interference screws (Ahmad et al., 2004). The purpose of this study was to determine if significant differences exist in the biomechanical performance of 4 different suspensory fixation devices, the Stryker VersiTomic G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton in cyclic loading and single load to failure in an in-vitro porcine model. The data showed no significant differences in displacement during cyclic loading to 250 N for 1000 cycles among the 4 different devices tested. However, only the Endobuttons had 100% survival during cyclic testing. One G-Lok, one ToggleLoc, and two RetroButtons failed during cyclic loading. Shelburne et al. calculated that the force on the ACL during normal gait can reach 300 N (Shelburne et al., 2004). This was a theoretical model with ideal conditions and estimations that vary with normal gait patterns. Therefore, as a model of ligamentous strain,
Fig. 3. Cyclic data — Graft elongation.
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
it may apply, but only in ideal conditions. Also, limitations are placed on rehab protocols to limit the number of times the graft is strained while being incorporated into ACL tunnels. There was not a significant difference in the single load to failure performance of the devices and all devices had failure loads greater than the 300 N calculated load that the ACL must withstand during gait (Shelburne et al., 2004). This is important but probably less important than withstanding cyclic loads because the ACL shouldn't be significantly loaded until it has time to properly heal, typically 8–12 weeks, unless a traumatic event occurs in the early postoperative period (Conner et al., 2010). Kamelger et al. conducted a similar study on 3 suspensory fixation devices (ToggleLoc, RetroButton, and Endobutton) in an in vitro 16week-old porcine model (Kamelger et al., 2009). They found the Endobutton to have a significantly greater load to failure than the RetroButton. This study found a similar result. The two studies differed in the failure mode. Kamelger et al. had 50% of the RetroButtons and 50% of the Endobuttons failing at the loop material but none of the ToggleLocs failing at the loop material (Kamelger et al., 2009). None of the suspensory fixation devices failed at the loop material in the current study. However, the current study used a soft tissue tendon graft. Kamelger at al. used a metal hook to load the suspensory fixation device. This could explain the difference in failure modes. The loading protocol in this study was similar to that used by Ahmad et al. to test several fixation devices, including the Endobutton (Ahmad et al., 2004). Both tests used porcine femurs but Ahmad et al. used 10 mm bovine grafts. This study used 9 mm porcine grafts. The different graft size and different loading rate, 200 mm/min vs. 20 mm/min (current study), may account for the difference in failure load, 864 N for the 10 mm graft and 716 N for the 9 mm grafts. Ahmad et al. reported no Endobutton failures during cyclic loading and all failures during single load to failure were by the device pulling through the cortex, which was the same for the Endobutton in this study (Ahmad et al., 2004). The graft displacement between the two studies also cannot be directly compared. Ahmad et al. measured graft slippage between the edge of the femoral tunnel and a suture placed in the graft 2 mm from the edge of the femur (Ahmad et al., 2004). This study measured the overall displacement of the graft, the fixation device, and the compliance of the system with the displacement of the actuator. Conner et al. conducted cyclic loading and load to failure tests in a porcine model with ToggleLoc and Endobutton fixation devices (Conner et al., 2010). Their grafts were all 8 mm in diameter and their testing protocol differed. They loaded the specimens cyclically from 50 N to 450 N and single load to failure at 1800 mm/min. Eight of 16 Endobuttons and 7 of 16 ToggleLocs failed during cyclic loading. Conner et al. found significantly greater total displacement in the ToggleLoc than the Endobutton during cyclic loading. The current study found no significant differences in displacement between these two devices during cyclic testing. However, both studies used different loading protocols. Conner et al. did not find a significant difference in load to failure or stiffness between the ToggleLoc and EndoButton, which is similar to the current study (Conner et al., 2010). The authors of this study chose to use a porcine model because specimens of the same age, sex, and similar size were available which would reduce the variability in tissue quality among specimens (Conner et al., 2010; Espejo-Baena et al., 2006; Miyata et al., 2000). The specimens are presumed to have a bone mineral density (BMD) similar to that of young adult humans, although no bone density measurements were obtained in the current study to verify this similarity, or variability, among individual porcine specimens. Further, most human cadaver models vary in BMD because of age and sex differences and are usually from an older population. It is important to note that the failure mechanism of nearly all specimens was via device pullout through the femoral cortex. Only two graft ruptures occurred, and no specimens failed due to suture loop rupture. These results suggest that the most common mechanism of failure with
233
suspensory fixation is a problem at the bone–button interface. Future suspensory device designs may be well advised to focus on ways to improve this particular interface in an effort to minimize device failure. Increasing the length of the device would increase the contact area with the cortex without requiring a larger hole to pass the device through. The current study has several limitations. This study did not directly measure graft elongation. It measured actuator displacement. By measuring actuator displacement it is not possible to determine percentages due to graft elongation, loop elongation, or the device migration into the bone. The authors consider actuator displacement to more accurately represent what is happening in situ because graft elongation, loop elongation, and device migration into the bone all contribute to knee laxity. In addition to being able to withstand the cyclic loads and not failing catastrophically the graft should not elongate a significant amount that could result in knee laxity. Clinical graft failure has been defined as greater than 5 mm of graft lengthening (Noyes et al., 1984). All 4 of the devices had greater than 5 mm of actuator displacement after 1000 cycles but we were unable to determine how much of the displacement was due to graft lengthening. However, the grafts were not pre-conditioned before cyclic testing as done in the operating room to remove most permanent visco-elastic elongation that may occur in vivo (Nurmi et al., 2004). Because the visco-elastic elongation was not removed prior to testing they could have exhibited greater elongation than would be seen clinically. To compare the results to a more clinically relevant situation, the first ten cycles can be considered to be preconditioning cycles. When the displacement after 10 cycles is removed from the displacement at 1000 cycles the results would range from 3.32 to 3.85 mm, less than the clinical graft failure definition of 5 mm. Another potential limitation is the usage of both 8.5 mm and 9 mm grafts sizes in 9 mm diameter femoral tunnels; i.e., seven of 40 grafts were 8.5 mm in diameter but tested in 9 mm diameter tunnels. The effect of different graft sizes was minimized by loading the graft in line with the tunnel, potentially decreasing graft contact with the wall of the tunnel. Another limitation, as noted above, is that bone density measurements of individual porcine specimens were not obtained in the current study. Finally, this was a cadaver model and the effect of healing and bone tunnel enlargement on the biomechanical properties of the fixation devices could not be evaluated.
5. Conclusions Recent increases in suspensory fixation device use make it vital to know their biomechanical performance. This study tested the biomechanical performance of 4 different suspensory fixation devices (Stryker G-Lok, Smith & Nephew Endobutton, Biomet ToggleLoc, and Arthrex RetroButton) in cyclic loading followed by a single load to failure. There were no significant differences in displacement during cyclic loading. There was only a significant difference in displacement at failure, but all grafts were able to support loads greater than that seen during normal gait. Therefore, device selection is not critical if a surgeon wants a device that will withstand normal loads in the post-op rehab period.
Conflict of interest statement Dr. Monica Hawkins is employed by Stryker Corp. and owns stock/stock options in Stryker Corp. Dr. Eric McCarty, Dr. Lucas Rylander, Dr. Jeffrey Brunelli, Dr. Michal Taylor, Todd Baldini, MS, and Byron Ellis, BS, have no financial and personal relationships with other people or organizations that could inappropriately influence (bias) this work.
234
L. Rylander et al. / Clinical Biomechanics 29 (2014) 230–234
Acknowledgments This study was funded by Stryker Joint Preservation. Dr. Monica Hawkins is employed by Stryker Joint Preservation and was involved in the design of the study and editing the manuscript. However, Dr. Hawkins was not involved in the decision to submit the manuscript.
References Ahmad, C.S., Gardner, T.R., Groh, M., Arnouk, J., Levine, W.N., 2004. Mechanical properties of soft tissue femoral fixation devices for anterior cruciate ligament reconstruction. Am. J. Sports Med. 32, 635–640. Allen, C.R., Giffin, J.R., Harner, C.D., 2003. Revision anterior cruciate ligament reconstruction. Orthop. Clin. N. Am. 34, 79–98. Bach Jr., B.R., 2003. Revision anterior cruciate ligament surgery. Arthroscopy 19 (Suppl. 1), 14–29. Beynnon, B.D., Johnson, R.J., Abate, J.A., Fleming, B.C., Nichols, C.E., 2005. Treatment of anterior cruciate ligament injuries, part 2. Am. J. Sports Med. 33, 1751–1767. Conner, C.S., Perez, B.A., Morris, R.P., Buckner, J.W., Buford, W.L., Ivey, F.M., 2010. Three femoral fixation devices for anterior cruciate ligament reconstruction: comparison of fixation on the lateral cortex versus the anterior cortex. Arthroscopy 26, 796–807. Espejo-Baena, A., Ezquerro, F., de la Blanca, A.P., Serrano-Fernandez, J., Nadal, F., Montanez-Heredia, E., 2006. Comparison of initialmechanical properties of 4 hamstring graft femoral fixation systems using nonpermanent hardware for anterior cruciate ligament reconstruction: an in vitro animal study. Arthroscopy 22, 433–440. Fox, J.A., Pierce, M., Bojchuk, J., Hayden, J., Bush-Joseph, C.A., Bach Jr., B.R., 2004. Revision anterior cruciate ligament reconstruction with nonirradiated fresh-frozen patellar tendon allograft. Arthroscopy 20, 787–794.
Kamelger, F.S., Onder, U., Schmoelz, W., Tecklenburg, K., Arora, R., Fink, C., 2009. Suspensory fixation of grafts in anterior cruciate ligament reconstruction: a biomechanical comparison of 3 implants. Arthroscopy 25, 767–776. Kvist, J., Gillquist, J., 2001. Sagittal plane knee translation and electromyographic activity during closed and open kinetic chain exercises in anterior cruciate ligament-deficient patients and control subjects. Am. J. Sports Med. 29, 72–82. Miyasaka, K.C., Daniel, D.M., Stone, M.L., 1991. The incidence of knee ligament injuries in the general population. Am. J. Knee Surg. 4, 43–48. Miyata, K., Yasuda, K., Kondo, E., Nakano, H., Kimura, S., Hara, N., 2000. Biomechanical comparisons of anterior cruciate ligament: reconstruction procedures with flexor tendon graft. J. Orthop. Sci. 5, 585–592. Noyes, F.R., Butler, D.L., Grood, E.S., Zernicke, R.F., Hefzy, M.S., 1984. Biomechanical analysis of human ligament grafts used in knee-ligament repairs and reconstructions. J. Bone Joint Surg. Am. 66, 344–352. Nurmi, J.T., Kannus, P., Sievänen, H., Järvelä, T., Järvinen, M., Järvinen, T.L., 2004. Interference screw fixation of soft tissue grafts in anterior cruciate ligament reconstruction: part 2: effect of preconditioning on graft tension during and after screw insertion. Am. J. Sports Med. 32, 418–424. Quatman, C.E., Paterno, M.V., Wordeman, S.C., Kaeding, C.C., 2011. Longitudinal anterior knee laxity related to substantial tibial tunnel enlargement after anterior cruciate ligament revision. Arthroscopy 27, 1160–1163. Rodeo, S.A., Kawamura, S., Kim, H.J., Dynybil, C., Ying, L., 2006. Tendon healing in a bone tunnel differs at the tunnel entrance versus the tunnel exit: an effect of graft–tunnel motion? Am. J. Sports Med. 34, 1790–1800. Shelburne, K.B., Pandy, M.G., Anderson, F.C., Torry, M.R., 2004. Pattern of anterior cruciate ligament force in normal walking. J. Biomech. 37, 797–805. Strand, T., Mølster, A., Hordvik, M., Krukhaug, Y., 2005. Long-term follow-up after primary repair of the anterior cruciate ligament: clinical and radiological evaluation 15–23 years postoperatively. Arch. Orthop. Trauma Surg. 125, 217–221. Weiler, A., Schmeling, A., Stöhr, I., Kääb, M.J., Wagner, M., 2007. Primary versus single-stage revision anterior cruciate ligament reconstruction using autologous hamstring tendon grafts: a prospective matched-group analysis. Am. J. Sports Med. 35, 1643–1652. Wilson, T.C., Kantaras, A., Atay, A., Johnson, D.L., 2004. Tunnel enlargement after anterior cruciate ligament surgery. Am. J. Sports Med. 32, 543–549.
