- Email: [email protected]
Biomechanical comparison of femoral fixation devices for anterior cruciate ligament reconstruction using a novel testing method
Clinical Biomechanics 28 (2013) 193–198
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Biomechanical comparison of femoral fixation devices for anterior cruciate ligament reconstruction using a novel testing method☆ Mark Ehrensberger a, b,⁎, Donald W. Hohman Jr. b, Kurt Duncan b, Craig Howard b, Leslie Bisson b a b
State University of New York at Buffalo, Department of Biomedical Engineering, Buffalo, NY, USA State University of New York at Buffalo, Department of Orthopaedic Surgery, Buffalo, NY, USA
a r t i c l e
i n f o
Article history: Received 9 March 2012 Accepted 11 December 2012 Keywords: Anterior cruciate ligament (ACL) Biomechanics Femoral anchor Graft fixation
a b s t r a c t Background: A novel biomechanical test method was implemented to compare the mechanical performance of two femoral fixation anchors (AperFix(r), Cayenne Medical, Scottsdale, AZ, USA or the AppianFx(r), KFx Medical, Carlsbad, CA, USA) that were utilized in anterior cruciate ligament reconstruction. Methods: Anterior cruciate ligament reconstructions were performed in 20 porcine femurs by using bovine extensor tendon grafts secured with 9 mm femoral anchors (AperFix(r) or AppianFx(r)). 10 specimens were tested for each anchor type. Infrared position sensors determined the repair construct displacements during conditioning (20 cycles at 5–50 N at 0.25 Hz), cyclic loading (1500 cycles at 50–200 N at 1 Hz), and ultimate loading (150 mm/min). Outcomes included tendon elongation, anchor displacement, stiffness, maximum load, yield load, and load at 5mm of anchor displacement. It was hypothesized that there would be no differences in the outcomes of these two devices. Independent measure t-tests compared the performance of the devices (p b 0.05). Findings: The performance of the two anchors was comparable during the cyclic loading. During ultimate loading, a statistically higher yield load (p b 0.01) and a load at 5 mm of anchor displacement (p b 0.01) were demonstrated for the AppianFx(r) as compared to AperFix(r). Maximum load and stiffness were not significantly different. Interpretation: Given the good clinical track record of the AperFix(r), the comparable, and in some cases superior, the biomechanical data presented here for the AppianFx(r) are encouraging for their clinical implementation. This study also introduced a novel test method that directly tracks the relevant construct displacements during cyclic and ultimate loading tests of the anterior cruciate ligament reconstructions. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction The goal of anterior cruciate ligament (ACL) reconstruction surgery is to restore normal knee biomechanics by placing and fixing a graft that recreates the native ligament's mechanical properties (Harvey et al., 2005). During the immediate post-operative phase the stability of the repair is primarily dependent on the mechanical properties provided by the fixation method (Giurea et al., 1999). Furthermore, eventual biological healing is also dependent on the initial graft stability provided by the mechanical fixation (Mayr et al., 2012). Stable graft-tunnel healing is the goal with the hope that graft tissue will be incorporated into the bone tunnel. Surgical variables such as graft/tunnel position, type of graft used (auto versus allograft), method of graft fixation, and
☆ All authors were fully involved in the study and preparation of the manuscript and that the material within has not been and will not be submitted for publication elsewhere. ⁎ Corresponding author at: 162 Farber Hall, 3435 Main Street, Buffalo, NY 14214-3000, USA. E-mail address: [email protected] (M. Ehrensberger). 0268-0033/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2012.12.007
graft tensioning are likely to have an effect on healing, as are postoperative variables such as graft motion. The mechanical fixation methods are categorized as either direct or indirect. Indirect mechanical fixation involves suspending the graft within the bone tunnel using cross-pin fixation or extra-articular buttons (Harvey et al., 2005). Direct mechanical fixation involves compressing the tendon graft against the walls of the bone tunnel (Harvey et al., 2005). Interference screws are a common type of direct mechanical fixation device. The AperFix® (Cayenne Medical, Scottsdale, AZ, USA) and the AppianFx® (KFx Medical, Carlsbad, CA, USA) ACL reconstruction systems provide additional approaches to direct mechanical fixation for hamstring grafts. Both of these devices are composed of polyetheretherketone (PEEK), which is a high strength and biocompatible polymer (Sagomonyants et al., 2008). The AperFix anchor (Fig. 1A) has two sets of fixation arms, one proximal and one distal, and a hole in the central region of the device. Two tendon bundles are looped distally through the central hole (Fig. 1C). When the AperFix is fully deployed, the proximal arms are expanded out to the left and the right in the plane of the photograph (Fig. 1A) and directly engage the cancellous bone, while the distal
194
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Fig. 1. Photographs showing the design features (A, B) and tendon attachment positions (C, D) of the AperFix® and the AppianFx® femoral ACL reconstruction anchors, respectively.
arms open perpendicularly to the plane of the photograph (1A) and circumferentially compress the tendon bundles against the walls of the bone tunnel. The AppianFx anchor (Fig. 1B) has a single set of proximal fixation arms that when the device is deployed are expanded out to the left and the right in the plane of the photograph and directly engage the cancellous bone. Tendon grafts are looped around the most proximal end of the AppianFx anchor (Fig. 1D) and press fit into the bone tunnel prior to deployment. The primary goal of this study was to evaluate and compare the biomechanical properties of the femoral Aperfix and AppianFx direct mechanical fixation anchors under both cyclic and ultimate loading. Many biomechanical studies of ACL reconstructions have utilized the displacement of the load frame crosshead or actuator as a measure of repair construct integrity during cyclic and ultimate load testing (Kousa et al., 2003; Nurmi et al., 2002; Seil et al., 1998; Shen et al., 2009; Zantop et al., 2004, 2006). However, the overall integrity of the repair construct is a function of tendon elongation, tendon slippage from the anchor, and movement of the anchor within the bone. All of these individual displacements cannot be accurately resolved by only tracking the overall actuator displacement. Efforts have been made to indirectly assess the individual displacement components by performing separate mechanical tests of the tendon graft alone in addition to testing the entire repair construct (bone plus anchor plus tendon). These studies (Milano et al., 2006; Speirs et al., 2010) showed that during cyclic loading to 150 N, a porcine flexor tendon graft alone will elongate approximately 1.5 mm (gage length of 25 mm) while the total construct displacements ranged between 2 and 11 mm. Based on this data, it was proposed that most of the construct displacement was occurring at the anchor/bone interface (Speirs et al., 2010). However, it is not possible to verify this conclusion without directly measuring either the tendon slippage at the
anchor or the anchor motion relative to the bone. Other authors (Kleweno et al., 2009) attempted to measure tendon slippage from the anchor by attaching a differential variable reluctance transducer to the tendon where it exits the bone tunnel. However, because the motion of the anchor was not directly measured, the authors concluded that any movement of the anchor was included within their measurement of graft slippage (Kleweno et al., 2009). To our knowledge, there is no accepted methodology published in the literature that allows for accurate assessment of all the individual displacement components of ACL repair constructs. Therefore, a secondary goal of this study was to develop a novel test method that would allow for tracking of all relevant displacements during cyclic and ultimate loading tests of ACL reconstructions. These details may provide valuable insights on how to improve future construct designs. 2. Methods 2.1. Specimen preparation A total of 20 fresh-frozen mature porcine femurs were used in this study. The femurs were stored at −20 °C and thawed at room temperature for 12 h prior to testing. All muscle and soft tissues were removed and the femur was transected mid-shaft. Fresh bovine extensor digitorum communis tendons were used as tendon grafts. The tendons were sharply trimmed parallel to the fiber orientation to yield an appropriate size graft. In previous ACL biomechanical studies these animal tissues have been utilized as representative models of human tissues (Ahmad et al., 2004; Kleweno et al., 2009). Per manufacturer's protocol, a single tendon was doubled over and sized through a 7.5 mm sizing hole for use with the 9 mm AppianFx anchor. Graft sizing for the 9 mm AperFix anchor was achieved by splitting
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
a single tendon longitudinally into two equally sized tendon segments, aligning the two segments into a single composite graft, doubling over, and sizing through a 7.5 mm sizing hole. The bone tunnels (10 mm for AppianFx and 9 mm for AperFix) were drilled to a 30 mm depth at the midpoint of the femoral intracondylar notch and then the devices were deployed by the same physician. Subsequent to deployment, a thin custom T-bar was passed down the tunnel and was rigidly secured to the femoral anchors by utilizing the connection mechanism on the backside of the anchor where it had been attached to the delivery handle prior to deployment (see Fig. 2). This T-bar was thin enough so as not to obstruct or interfere with the tendon graft or the sides of the bone tunnel. The transected femur was then securely mounted within the base
195
grip of an 858 Mini Bionix load frame (MTS Systems Corporation, Eden Prairie, MN, USA) such that the entrance of the bone tunnel was oriented superiorly. The free ends of the tendon were whip-stitched to aid in the loading of tendon into the thermal-electric cooled tissue grips (Bose Corporation, Eden Prairie, MN, USA) that were attached to the MTS actuator. These grips are specially designed to prevent tissue slippage by quickly freezing the tissue within the clamp. The tendon gage length between the tunnel aperture and the tissue grip was 39 mm. The tendons were marked with a permanent marker at the location where they exited the bone tunnel and at the edge of the tissue grip. Tracking the relative position of these markings provided an assessment of tendon slippage during testing.
Fig. 2. (A) Schematic of the experimental setup showing the orientation of the graft fixation construct, mechanical test system, and the array of infrared position sensors (1–5). (B) Photograph showing the details of the experimental setup. Please note that the only infrared position sensors visible in the photo are those attached to the t-bar.
196
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Once the specimen was securely fixed in the load frame an array of infrared sensors were attached to the experimental setup. The 3-D positions of these sensors were measured with 0.01 mm resolution via an Optotrak 3020 infrared system (Northern Digital Inc., Waterloo, Ontario, CA). As shown in Fig. 2A, one sensor was attached to the MTS actuator, two sensors were placed on the T-bar attached to the anchor, a fourth sensor was secured to the femur, and a fifth sensor was placed on the base grip. This array of sensors allowed for the real-time tracking of the tendon elongation and anchor displacement relative to the bone and actuator positions. The fixation of the bone within the base grips was verified by tracking the displacement of sensor 4 relative to sensor 5. It was determined that the base grip rigidly held the bone during cyclic and ultimate loading. The tendon elongation was calculated by subtracting the vertical displacement of the anchor (determined from the sensors on the T-bar) from the vertical displacement of the MTS actuator. A close up view of the experimental setup is shown by the photograph in Fig. 2B. 2.2. Mechanical testing The mechanical testing protocol consisted of three phases. The construct was initially conditioned by cycling between 5 and 50 N at 0.25 Hz for 20 cycles. The position of the construct after the conditioning was designated as the initial position. The construct was then loaded for 1500 cycles between 50 N and 200 N at 1 Hz. It has been reported that up to 200 N are applied to the ACL during active full extension of the knee (Engebretsen et al., 1989; Rupp et al., 1999) and therefore our cyclic loading represented a clinically relevant load range. Following cyclic testing the construct was pulled to failure at 150 mm/min. All loads were applied in line with the longitudinal axis of the bone tunnel to simulate worst-case loading. The specimens were sprayed with saline every 5 min to prevent drying effects. 2.3. Data analysis The measureable outcomes included tendon elongation and anchor displacement during cyclic loading, and the maximum load, yield load, and stiffness during the ultimate loading. The mode of failure (tendon rupture or anchor pull out) was also reported for each specimen. In addition, by syncing the load data from the MTS system and the displacements from the Optotrak system we are able to report the load at 5 mm of anchor displacement, which we considered clinical failure. The 5 mm of anchor displacement was measured relative to the initial position (post conditioning) and could occur either during the cyclic or failure loading phase. A total of 10 samples were tested for each type of anchor. Independent measures t-tests were used to compare the outcomes between the two anchor groups. A statistically significant difference was defined as P b 0.05. 3. Results
approximately 0.7 mm and decreased to 0.5 mm at the end of the cyclic test (cycles 1490–1500). 3.2. Ultimate loading The load and displacement data obtained from the ultimate loading were analyzed to determine the maximum load, load at 5 mm of anchor displacement, yield load and the stiffness displayed in Table 2. A statistically higher yield load (P =0.001) and load at 5 mm of anchor displacement (P =0.007) was demonstrated for the AppianFx as compared to AperFix. The results for stiffness and max load were comparable for the two types of anchors tested. 3.3. Mode of failure There were no repair constructs that failed during cyclic loading. The tendon/anchor complex was completely pulled out of the bone tunnel in six of the AppianFx devices and in nine of the AperFix devices. In the remaining specimens the repair construct failed by tendon graft rupture. 4. Discussion Previous studies (Beynnon and Amis, 1998; Kousa et al., 2003; Nurmi et al., 2002; Zantop et al., 2004) have highlighted the importance of assessing the mechanical properties of ACL anchors during both cyclic loading and load to failure. While other studies have applied cyclic elongation to evaluate ACL reconstructions (Nakano et al., 2000; Yamanaka et al., 1999), we utilized a cyclic loading protocol that represented the loads placed on the reconstructions during active knee extension in early rehabilitation programs (Engebretsen et al., 1989; Rupp et al., 1999; Seil et al., 1998). In our cyclic loading tests we observed that all specimens had an asymptotic increase in the construct displacements with the majority of the migration occurring during the first 100 cycles. Furthermore, we documented that the overall construct displacement was a combination of tendon elongation and anchor displacement. These results imply that, despite the 20 cycles of 5 N–50 N conditioning, there was still tendon creep and seating of the anchor that occurred during the initial cyclic loading from 50 N–200 N. However, after this initial migration the constructs were quite stable with very little continued migration. This emphasizes the need for proper conditioning of the tendon graft prior to tibial fixation. We found no difference in the amount of tendon elongation or anchor displacement during cycling when comparing the AperFix and the AppianFx. On average, after 1500 cycles (50–200 N at 1 Hz) these devices both displayed 1.6 mm tendon elongation, 2.1 mm of anchor movement, and 3.8 mm of total construct displacement. These results are comparable to the results obtained in similar studies of other types of ACL anchors. For example, tests for a variety of interference screws, cross-pins, and cortical suspension devices with cyclic loads of 50 N– 200 N for 1000–1500 cycles produces total construct displacements of
3.1. Cyclic loading Fig. 3 displays a representative plot of the construct displacements obtained by the Optotrak system during the cyclic loading. Throughout the cycling the bone is rigidly fixed in the base grips. However, the displacement of the actuator, anchor, and tendon all appear to asymptotically increase during the cyclic loading. In addition, Fig. 3 shows that the total motion of the repair construct (actuator displacement) is a composite of anchor displacement and tendon elongation. Table 1 reports the tendon elongation and anchor displacement after 1500 loading cycles and indicates that there were no significant differences when comparing the two types of anchors. The repairs also showed similar total construct cyclic amplitudes (peak to valley displacement within a load cycle) throughout the testing. The cyclic amplitudes for both constructs at the start of the cyclic test (cycles 20–30) were
Fig. 3. A representative plot of the construct displacements during the 1500 cycles of loading from 50 to 200 N at 1 Hz. Most of the displacement occurred early in the cycling phase for all specimens tested.
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198 Table 1 Results obtained during the cyclic loading tests. AppianFx®
AperFix®
P-value
Tendon elongation (mm) Mean 1.85 (SD 1.14) Mean 1.30 (SD 0.53) 0.123 Anchor displacement (mm) Mean 2.32 (SD 0.88) Mean 1.79 (SD 1.69) 0.329 All data presented as mean (SD)
2–5 mm (Kousa et al., 2003; Zantop et al., 2004) while protocols that cycled up to 150 N for 100–2000 cycles showed total construct displacements of 2–11 mm with tendon elongations of 1.5 mm (Milano et al., 2006; Speirs et al., 2010; Zhang et al., 2007). Furthermore, the cyclic amplitudes reported in the current study are in line with those reported in the literature for ACL reconstructions using similar test specimens and protocols (Speirs et al., 2010). However, these cyclic amplitudes are smaller than the 2 mm cyclic elongations used in displacement control studies of ACL reconstructions (Nakano et al., 2000; Yamanaka et al., 1999). These cyclic elongation studies were conducted in porcine knees with both femoral and tibial fixation sites and graft lengths that were much longer than our gage length and it is therefore likely that these contrasts in test protocol can explain the differences in the cyclic displacements. This study has shown that similar maximum loads are reached for both devices during load to failure testing. However, maximum loading may have little relevance in the clinical setting. It has been shown by others (Ishibashi et al., 1997) that the stiffness of the ACL repair is of more importance than the ultimate strength when attempting to restore normal knee kinematics. The stiffness values reported in the current study for both repair constructs were comparable to each other and were higher than those values reported for other fixation devices in porcine femurs (Kousa et al., 2003; Milano et al., 2006; Shen et al., 2009; Speirs et al., 2010). This study also demonstrated a higher yield load for the AppianFx as compared to the AperFix. This result, coupled with the finding that these devices have similar stiffness values, implies the AppianFx repair construct will maintain its high stiffness character for greater loads as compared to the AperFix. Further examination of the ultimate loading data revealed several important findings. Fig. 4 shows a plot of the load, actuator displacement, tendon elongation, and anchor displacement versus time that is representative of all ultimate loading data sets obtained in this study. The load versus time plot shows an initial linear region (~0.5 s for data set shown in Fig. 4) of constant slope that transitions to a second region that is approximately linear with a lower slope. In the initial linear load region it is noted that the anchor does not move and that all construct displacement is due to tendon elongation. We noted no slippage of the graft from either the tissue grips or from the bone tunnel during the testing protocol. If the tendon and the anchor/bone interface are mechanically modeled as two springs in series with different spring constants (stiffness values) (To et al., 1999), then under tension the spring with the lower stiffness will displace and the other spring will not. Since, as shown in Fig. 4, the anchor was not initially displacing, the stiffness of the AperFix and AppianFx must be greater than that of the tendon. In a previous study (To et al., 1999) it was highlighted that the mechanical limitation of ACL reconstructions was the low stiffness of the anchor fixation and that to improve biomechanical performance efforts should be made to increase the stiffness of the anchor fixation. Given the results of this current
197
study, both the AperFix and the AppianFx may represent designs that provide appropriate stiffness to ACL reconstructions. As further shown by the representative data in Fig. 4, as the load is increased, eventually the anchor breaks free and begins to move. The transition point where the anchor begins to move also corresponds to where the initial linear region of the load plot ends. If the load was plotted versus the actuator displacement, then the slope of the initial linear region would be the construct stiffness and the transition point would indicate the yield load. Therefore, the yield loads reported for these tests are actually reflective of the static fixation strength of the anchors and not due to mechanical yielding or slippage of the tendon grafts. In this regard, it will take greater loads to disrupt the initial static fixation of the AppianFx anchor (~431 N) as compared to the AperFix anchor (~324 N). This biomechanical advantage provided by the AppianFx may be clinically important, since it has been advocated that the initial fixation strength of the ACL repair construct should be greater than 450 N (Zantop et al., 2004, 2006) in order to withstand loads associated with activities of daily living (Noyes et al., 1984) and stair decent (Morrison, 1969). Our findings that the yield load of these constructs is coupled to the initiation of anchor motion provides further support to the preferential reporting in the literature of yield load (Kousa et al., 2003; Nurmi et al., 2002) as compared to ultimate load. The load at 5 mm of anchor motion, which we feel may represent the critical threshold for construct clinical failure, was significantly higher for the AppianFx (667±144 N) as compared to the AperFx at (316±162 N). The lower load at 5 mm for the AperFix was likely influenced by a bimodal load–displacement curve noted in 5 out of 10 specimens during the ultimate loading test phase (Fig. 5C). In these bimodal cases, after the yield load was reached there was a decrease in the load carrying capacity and the construct displaced several millimeters under low loads. This was followed by a second rise in load to reach the maximum load. This was in contrast to the behavior of the other 5 AperFix and all of the AppianFx, which after reaching their yield load showed no appreciable decrease in the load carrying capacity until after reaching its ultimate tensile load (Fig. 5A, B). The bimodal curve could be related to the fact that the AperFix has two sets of arms that provide fixation. The proximal arms engage the cancellous bone while the distal arms compress the tendons against the aperture of the bone tunnel. The overall fixation for the AperFix is a composite of the fixation at each of these sites. In the bimodal cases, the drop in the load after the yield point may indicate the transition from fixation with the two sets of arms initially to fixation with just one set. This transition may possibly occur after the anchor begins to move and the distal arms are pulled out of the bone tunnel. The subsequent rise to a second higher peak load could reflect that the proximal arms are engaging the stronger cortical bone as they are pulled towards the aperture of the bone tunnel. It is remarkable that half of the AperFix samples displayed this bimodal curve as this may reflect an important variability in the biomechanical performance of this device. However, this concern stemming from variability in laboratory testing has not
Table 2 Results obtained during the ultimate loading tests. AppianFx® Stiffness (N/mm) Yield load (N) Load 5 mm (N) Max load (N)
AperFix®
Mean 217 (SD 31) Mean 242 Mean 431 (SD 79) Mean 324 Mean 667 (SD 144) Mean 316 Mean 1021 (SD 149) Mean 912 All data presented as mean (SD)
P-value (SD (SD (SD (SD
43) 45) 162) 116)
0.162 0.001 0.007 0.086
Fig. 4. A representative plot of the load, actuator displacement, anchor displacement, and tendon elongation versus time that was collected during the ultimate loading test.
198
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Fig. 5. Representative plots of load versus displacement during the ultimate loading tests. All AppianFx ® anchors displayed consistent load–displacement relationships and were similar data presented in (A). AperFix® anchors displaced two distinct types of load–displacement relationships. Five AperFix® specimens displayed data similar to (B) and five AperFix ® specimens displayed a bimodal load–displacement relationship noted in (C).
been established in the clinical literature as the AperFix has reported very good clinical outcomes (Connor, 2009; Uzumcugil et al., 2010). In light of these good clinical results for the AperFix, the higher load at 5 mm of AppianFx displacement and less variable ultimate loading performance may indicate similar or enhanced clinical performance for the AppianFx anchor. This in vitro study was conducted to assess the simulated in vivo biomechanical performance of ACL reconstructions and as such there are some limitations to consider. First, we performed these studies utilizing porcine femurs instead of human femurs. Human cadavers are usually elderly and the harvested tissues can display great sample-to-sample variation in their mechanical properties. Porcine femurs have been utilized in numerous ACL reconstruction studies (Ahmad et al., 2004; Kleweno et al., 2009; Kousa et al., 2003; Milano et al., 2006; Nakano et al., 2000; Nurmi et al., 2002; Seil et al., 1998; Shen et al., 2009; Speirs et al., 2010; Zhang et al., 2007) as an appropriate substitute for young human bone. We also used bovine extensor tendons as our simulated human ACL grafts. These tendons have been used as grafts in other biomechanical studies of ACL reconstructions (Ahmad et al., 2004; Kleweno et al., 2009) and have been shown to have equivalent viscoelastic and stiffness properties of human hamstring tendons (Donahue et al., 2001). In this study we also applied the load directly in line with the axis of the bone tunnel, which does not accurately represent the loading pattern experienced by the ACL in vivo when the tibia translates anteriorly. Therefore, the values of construct displacements and strengths reported in this study might not directly translate into clinical application. However, this method is widely used in the literature as it does allow for the direct biomechanical comparison of ACL reconstructions under “worst-case” conditions. To our knowledge, this is the first report of a mechanical test method that can accurately resolve and directly quantify the individual displacement components of ACL reconstructions under controlled cyclic and ultimate loading. Furthermore, by syncing the load data and the displacement data of the individual construct components, this novel method provides the means to determine exactly what load causes a specific construct component to displace. This method could find great utility in the development and evaluation of new ACL anchor designs because it allows for the direct assessment of the anchor motion under relevant loading. 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. Beynnon, B.D., Amis, A.A., 1998. In vitro testing protocols for the cruciate ligaments and ligament reconstructions. Knee Surg. Sports Traumatol. Arthrosc. 6, S70–S76. Connor, G., 2009. Analysis Of Outcome In ACL Reconstruction Using The Cayenne Medical AperFix System. Cayenne Medical Inc., White Paper. Donahue, T.L.H., Gregersen, C., Hull, M.L., Howell, S.M., 2001. Comparison of viscoelastic, structural, and material properties of double-looped anterior cruciate ligament grafts made from bovine digital extensor and human hamstring tendons. J. Biomech. Eng. 123, 162–169.
Engebretsen, L., Lew, W.D., Lewis, J.L., Hunter, R.E., 1989. Knee mechanics after repair of the anterior cruciate ligament. A cadaver study of ligament augmentation. Acta Orthop. 60, 703–709. Giurea, M., Z. P., Amis, A.A., Aichroth, P., 1999. Comparative pull-out and cyclic-loading strength tests of anchorage of hamstring tendon grafts in anterior cruciate ligament reconstruction. Am. J. Sports Med. 27, 621–625. Harvey, A., Thomas, N.P., Amis, A.A., 2005. Fixation of the graft in reconstruction of the anterior cruciate ligament. J. Bone Joint Surg. Br. 87-B, 593–603. Ishibashi, Y., R. T., Livesay, G.A., Stone, J.D., Fu, F.H., Woo, S.L., 1997. The effect of anterior cruciate ligament graft fixation site at the tibia on knee stability: evaluation using a robotic testing system. Arthroscopy 13, 177–182. Kleweno, C.P., Jacir, A.M., Gardner, T.R., Ahmad, C.S., Levine, W.N., 2009. Biomechanical evaluation of anterior cruciate ligament femoral fixation techniques. Am. J. Sports Med. 37, 339–345. Kousa, P., Jarvinen, T.L.N., Vihavainen, M., Kannus, P., Jarvinen, M., 2003. The fixation strength of six hamstring tendon graft fixation devices in anterior cruciate ligament reconstruction-part I: femoral site. Am. J. Sports Med. 31, 174–181. Mayr, R., Rosenberger, R., Agraharam, D., Smekal, V., El Attal, R., 2012. Revision anterior cruciate ligament reconstruction: an update. Arch. Orthop. Trauma Surg. 132, 1299–1313. Milano, G., Mulas, P.D., Ziranu, F., Piras, S., Manunta, A., Fabbriciani, C., 2006. Comparison between different femoral fixation devices for ACL reconstruction with doubled hamstring tendon graft: a biomechanical analysis. Arthroscopy 22, 660–668. Morrison, J.B., 1969. Function of the knee joint in various activities. Biomed. Eng. 4, 573–580. Nakano, H., Yasuda, K., Tohyama, H., Yamanaka, M., Wada, T., Kaneda, K., 2000. Interference screw fixation of doubled flexor tendon graft in anterior cruciate ligament reconstruction — biomechanical evaluation with cyclic elongation. Clin. Biomech. 15, 188–195. 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., Järvinen, T.L.N., Kannus, P., Sievänen, H., Toukosalo, J., Järvinen, M., 2002. Compaction versus extraction drilling for fixation of the hamstring tendon graft in anterior cruciate ligament reconstruction. Am. J. Sports Med. 30, 167–173. Rupp, S., Hopf, T., Hess, T., Seil, R., Kohn, D.M., 1999. Resulting tensile forces in the human bone–patellar tendon–bone graft: direct force measurement in vitro. Arthroscopy 15, 179–184. Sagomonyants, K.B., Jarman-Smith, M.L., Devine, J.N., Aronow, M.S., Gronowicz, G.A., 2008. The in vitro response of human osteoblasts to polyetheretherketone (PEEK) substrates compared to commercially pure titanium. Biomaterials 29, 1563–1572. Seil, R., Rupp, S., Krauss, P.W., Benz, A., Kohn, D.M., 1998. Comparison of initial fixation strength between biodegradable and metallic interference screws and a press-fit fixation technique in a porcine model. Am. J. Sports Med. 26, 815–819. Shen, P.-H., Lien, S.-B., Shen, H.-C., Wang, C.-C., Huang, G.-S., Chao, K.-H., et al., 2009. Comparison of different sizes of bioabsorbable interference screws for anterior cruciate ligament reconstruction using bioabsorbable bead augmentation in a porcine model. Arthroscopy 25, 1101–1107. Speirs, A., Simon, D., Lapner, P., 2010. Evaluation of a new femoral fixation device in a simulated anterior cruciate ligament reconstruction. Arthroscopy 26, 351–357. To, J.T., Howell, L.C.S.M., Hull, M.L., 1999. Contributions of femoral fixation methods to the stiffness of anterior cruciate ligament replacements at implantation. Arthroscopy 15, 379–387. Uzumcugil, O., Dogan, A., Dalyaman, E., Yalcinkaya, M., Akman, E., Ozturkmen, Y., et al., 2010. AperFix versus TransFix in reconstruction of anterior cruciate ligament. J. Knee Surg. 23, 029,035. Yamanaka, M., Yasuda, K., Tohyama, H., Nakano, H., Wada, T., 1999. The effect of cyclic displacement on the biomechanical characteristics of anterior cruciate ligament reconstructions. Am. J. Sports Med. 27, 772–777. Zantop, T., Weimann, A., Rümmler, M., Hassenpflug, J., Petersen, W., 2004. Initial fixation strength of two bioabsorbable pins for the fixation of hamstring grafts compared to interference screw fixation. Am. J. Sports Med. 32, 641–649. Zantop, T., Weimann, A., Schmidtko, R., Herbort, M., Raschke, M.J., Petersen, W., 2006. Graft laceration and pullout strength of soft-tissue anterior cruciate ligament reconstruction: in vitro study comparing titanium, poly-D, L-lactide, and poly-D, L-lactide–tricalcium phosphate screws. Arthroscopy 22, 1204–1210. Zhang, A.L., Lewicky, Y.M., Oka, R., Mahar, A., Pedowitz, R., 2007. Biomechanical analysis of femoral tunnel pull-out angles for anterior cruciate ligament reconstruction with bioabsorbable and metal interference screws. Am. J. Sports Med. 35, 637–642.
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Biomechanical comparison of femoral fixation devices for anterior cruciate ligament reconstruction using a novel testing method☆ Mark Ehrensberger a, b,⁎, Donald W. Hohman Jr. b, Kurt Duncan b, Craig Howard b, Leslie Bisson b a b
State University of New York at Buffalo, Department of Biomedical Engineering, Buffalo, NY, USA State University of New York at Buffalo, Department of Orthopaedic Surgery, Buffalo, NY, USA
a r t i c l e
i n f o
Article history: Received 9 March 2012 Accepted 11 December 2012 Keywords: Anterior cruciate ligament (ACL) Biomechanics Femoral anchor Graft fixation
a b s t r a c t Background: A novel biomechanical test method was implemented to compare the mechanical performance of two femoral fixation anchors (AperFix(r), Cayenne Medical, Scottsdale, AZ, USA or the AppianFx(r), KFx Medical, Carlsbad, CA, USA) that were utilized in anterior cruciate ligament reconstruction. Methods: Anterior cruciate ligament reconstructions were performed in 20 porcine femurs by using bovine extensor tendon grafts secured with 9 mm femoral anchors (AperFix(r) or AppianFx(r)). 10 specimens were tested for each anchor type. Infrared position sensors determined the repair construct displacements during conditioning (20 cycles at 5–50 N at 0.25 Hz), cyclic loading (1500 cycles at 50–200 N at 1 Hz), and ultimate loading (150 mm/min). Outcomes included tendon elongation, anchor displacement, stiffness, maximum load, yield load, and load at 5mm of anchor displacement. It was hypothesized that there would be no differences in the outcomes of these two devices. Independent measure t-tests compared the performance of the devices (p b 0.05). Findings: The performance of the two anchors was comparable during the cyclic loading. During ultimate loading, a statistically higher yield load (p b 0.01) and a load at 5 mm of anchor displacement (p b 0.01) were demonstrated for the AppianFx(r) as compared to AperFix(r). Maximum load and stiffness were not significantly different. Interpretation: Given the good clinical track record of the AperFix(r), the comparable, and in some cases superior, the biomechanical data presented here for the AppianFx(r) are encouraging for their clinical implementation. This study also introduced a novel test method that directly tracks the relevant construct displacements during cyclic and ultimate loading tests of the anterior cruciate ligament reconstructions. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction The goal of anterior cruciate ligament (ACL) reconstruction surgery is to restore normal knee biomechanics by placing and fixing a graft that recreates the native ligament's mechanical properties (Harvey et al., 2005). During the immediate post-operative phase the stability of the repair is primarily dependent on the mechanical properties provided by the fixation method (Giurea et al., 1999). Furthermore, eventual biological healing is also dependent on the initial graft stability provided by the mechanical fixation (Mayr et al., 2012). Stable graft-tunnel healing is the goal with the hope that graft tissue will be incorporated into the bone tunnel. Surgical variables such as graft/tunnel position, type of graft used (auto versus allograft), method of graft fixation, and
☆ All authors were fully involved in the study and preparation of the manuscript and that the material within has not been and will not be submitted for publication elsewhere. ⁎ Corresponding author at: 162 Farber Hall, 3435 Main Street, Buffalo, NY 14214-3000, USA. E-mail address: [email protected] (M. Ehrensberger). 0268-0033/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinbiomech.2012.12.007
graft tensioning are likely to have an effect on healing, as are postoperative variables such as graft motion. The mechanical fixation methods are categorized as either direct or indirect. Indirect mechanical fixation involves suspending the graft within the bone tunnel using cross-pin fixation or extra-articular buttons (Harvey et al., 2005). Direct mechanical fixation involves compressing the tendon graft against the walls of the bone tunnel (Harvey et al., 2005). Interference screws are a common type of direct mechanical fixation device. The AperFix® (Cayenne Medical, Scottsdale, AZ, USA) and the AppianFx® (KFx Medical, Carlsbad, CA, USA) ACL reconstruction systems provide additional approaches to direct mechanical fixation for hamstring grafts. Both of these devices are composed of polyetheretherketone (PEEK), which is a high strength and biocompatible polymer (Sagomonyants et al., 2008). The AperFix anchor (Fig. 1A) has two sets of fixation arms, one proximal and one distal, and a hole in the central region of the device. Two tendon bundles are looped distally through the central hole (Fig. 1C). When the AperFix is fully deployed, the proximal arms are expanded out to the left and the right in the plane of the photograph (Fig. 1A) and directly engage the cancellous bone, while the distal
194
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Fig. 1. Photographs showing the design features (A, B) and tendon attachment positions (C, D) of the AperFix® and the AppianFx® femoral ACL reconstruction anchors, respectively.
arms open perpendicularly to the plane of the photograph (1A) and circumferentially compress the tendon bundles against the walls of the bone tunnel. The AppianFx anchor (Fig. 1B) has a single set of proximal fixation arms that when the device is deployed are expanded out to the left and the right in the plane of the photograph and directly engage the cancellous bone. Tendon grafts are looped around the most proximal end of the AppianFx anchor (Fig. 1D) and press fit into the bone tunnel prior to deployment. The primary goal of this study was to evaluate and compare the biomechanical properties of the femoral Aperfix and AppianFx direct mechanical fixation anchors under both cyclic and ultimate loading. Many biomechanical studies of ACL reconstructions have utilized the displacement of the load frame crosshead or actuator as a measure of repair construct integrity during cyclic and ultimate load testing (Kousa et al., 2003; Nurmi et al., 2002; Seil et al., 1998; Shen et al., 2009; Zantop et al., 2004, 2006). However, the overall integrity of the repair construct is a function of tendon elongation, tendon slippage from the anchor, and movement of the anchor within the bone. All of these individual displacements cannot be accurately resolved by only tracking the overall actuator displacement. Efforts have been made to indirectly assess the individual displacement components by performing separate mechanical tests of the tendon graft alone in addition to testing the entire repair construct (bone plus anchor plus tendon). These studies (Milano et al., 2006; Speirs et al., 2010) showed that during cyclic loading to 150 N, a porcine flexor tendon graft alone will elongate approximately 1.5 mm (gage length of 25 mm) while the total construct displacements ranged between 2 and 11 mm. Based on this data, it was proposed that most of the construct displacement was occurring at the anchor/bone interface (Speirs et al., 2010). However, it is not possible to verify this conclusion without directly measuring either the tendon slippage at the
anchor or the anchor motion relative to the bone. Other authors (Kleweno et al., 2009) attempted to measure tendon slippage from the anchor by attaching a differential variable reluctance transducer to the tendon where it exits the bone tunnel. However, because the motion of the anchor was not directly measured, the authors concluded that any movement of the anchor was included within their measurement of graft slippage (Kleweno et al., 2009). To our knowledge, there is no accepted methodology published in the literature that allows for accurate assessment of all the individual displacement components of ACL repair constructs. Therefore, a secondary goal of this study was to develop a novel test method that would allow for tracking of all relevant displacements during cyclic and ultimate loading tests of ACL reconstructions. These details may provide valuable insights on how to improve future construct designs. 2. Methods 2.1. Specimen preparation A total of 20 fresh-frozen mature porcine femurs were used in this study. The femurs were stored at −20 °C and thawed at room temperature for 12 h prior to testing. All muscle and soft tissues were removed and the femur was transected mid-shaft. Fresh bovine extensor digitorum communis tendons were used as tendon grafts. The tendons were sharply trimmed parallel to the fiber orientation to yield an appropriate size graft. In previous ACL biomechanical studies these animal tissues have been utilized as representative models of human tissues (Ahmad et al., 2004; Kleweno et al., 2009). Per manufacturer's protocol, a single tendon was doubled over and sized through a 7.5 mm sizing hole for use with the 9 mm AppianFx anchor. Graft sizing for the 9 mm AperFix anchor was achieved by splitting
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
a single tendon longitudinally into two equally sized tendon segments, aligning the two segments into a single composite graft, doubling over, and sizing through a 7.5 mm sizing hole. The bone tunnels (10 mm for AppianFx and 9 mm for AperFix) were drilled to a 30 mm depth at the midpoint of the femoral intracondylar notch and then the devices were deployed by the same physician. Subsequent to deployment, a thin custom T-bar was passed down the tunnel and was rigidly secured to the femoral anchors by utilizing the connection mechanism on the backside of the anchor where it had been attached to the delivery handle prior to deployment (see Fig. 2). This T-bar was thin enough so as not to obstruct or interfere with the tendon graft or the sides of the bone tunnel. The transected femur was then securely mounted within the base
195
grip of an 858 Mini Bionix load frame (MTS Systems Corporation, Eden Prairie, MN, USA) such that the entrance of the bone tunnel was oriented superiorly. The free ends of the tendon were whip-stitched to aid in the loading of tendon into the thermal-electric cooled tissue grips (Bose Corporation, Eden Prairie, MN, USA) that were attached to the MTS actuator. These grips are specially designed to prevent tissue slippage by quickly freezing the tissue within the clamp. The tendon gage length between the tunnel aperture and the tissue grip was 39 mm. The tendons were marked with a permanent marker at the location where they exited the bone tunnel and at the edge of the tissue grip. Tracking the relative position of these markings provided an assessment of tendon slippage during testing.
Fig. 2. (A) Schematic of the experimental setup showing the orientation of the graft fixation construct, mechanical test system, and the array of infrared position sensors (1–5). (B) Photograph showing the details of the experimental setup. Please note that the only infrared position sensors visible in the photo are those attached to the t-bar.
196
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Once the specimen was securely fixed in the load frame an array of infrared sensors were attached to the experimental setup. The 3-D positions of these sensors were measured with 0.01 mm resolution via an Optotrak 3020 infrared system (Northern Digital Inc., Waterloo, Ontario, CA). As shown in Fig. 2A, one sensor was attached to the MTS actuator, two sensors were placed on the T-bar attached to the anchor, a fourth sensor was secured to the femur, and a fifth sensor was placed on the base grip. This array of sensors allowed for the real-time tracking of the tendon elongation and anchor displacement relative to the bone and actuator positions. The fixation of the bone within the base grips was verified by tracking the displacement of sensor 4 relative to sensor 5. It was determined that the base grip rigidly held the bone during cyclic and ultimate loading. The tendon elongation was calculated by subtracting the vertical displacement of the anchor (determined from the sensors on the T-bar) from the vertical displacement of the MTS actuator. A close up view of the experimental setup is shown by the photograph in Fig. 2B. 2.2. Mechanical testing The mechanical testing protocol consisted of three phases. The construct was initially conditioned by cycling between 5 and 50 N at 0.25 Hz for 20 cycles. The position of the construct after the conditioning was designated as the initial position. The construct was then loaded for 1500 cycles between 50 N and 200 N at 1 Hz. It has been reported that up to 200 N are applied to the ACL during active full extension of the knee (Engebretsen et al., 1989; Rupp et al., 1999) and therefore our cyclic loading represented a clinically relevant load range. Following cyclic testing the construct was pulled to failure at 150 mm/min. All loads were applied in line with the longitudinal axis of the bone tunnel to simulate worst-case loading. The specimens were sprayed with saline every 5 min to prevent drying effects. 2.3. Data analysis The measureable outcomes included tendon elongation and anchor displacement during cyclic loading, and the maximum load, yield load, and stiffness during the ultimate loading. The mode of failure (tendon rupture or anchor pull out) was also reported for each specimen. In addition, by syncing the load data from the MTS system and the displacements from the Optotrak system we are able to report the load at 5 mm of anchor displacement, which we considered clinical failure. The 5 mm of anchor displacement was measured relative to the initial position (post conditioning) and could occur either during the cyclic or failure loading phase. A total of 10 samples were tested for each type of anchor. Independent measures t-tests were used to compare the outcomes between the two anchor groups. A statistically significant difference was defined as P b 0.05. 3. Results
approximately 0.7 mm and decreased to 0.5 mm at the end of the cyclic test (cycles 1490–1500). 3.2. Ultimate loading The load and displacement data obtained from the ultimate loading were analyzed to determine the maximum load, load at 5 mm of anchor displacement, yield load and the stiffness displayed in Table 2. A statistically higher yield load (P =0.001) and load at 5 mm of anchor displacement (P =0.007) was demonstrated for the AppianFx as compared to AperFix. The results for stiffness and max load were comparable for the two types of anchors tested. 3.3. Mode of failure There were no repair constructs that failed during cyclic loading. The tendon/anchor complex was completely pulled out of the bone tunnel in six of the AppianFx devices and in nine of the AperFix devices. In the remaining specimens the repair construct failed by tendon graft rupture. 4. Discussion Previous studies (Beynnon and Amis, 1998; Kousa et al., 2003; Nurmi et al., 2002; Zantop et al., 2004) have highlighted the importance of assessing the mechanical properties of ACL anchors during both cyclic loading and load to failure. While other studies have applied cyclic elongation to evaluate ACL reconstructions (Nakano et al., 2000; Yamanaka et al., 1999), we utilized a cyclic loading protocol that represented the loads placed on the reconstructions during active knee extension in early rehabilitation programs (Engebretsen et al., 1989; Rupp et al., 1999; Seil et al., 1998). In our cyclic loading tests we observed that all specimens had an asymptotic increase in the construct displacements with the majority of the migration occurring during the first 100 cycles. Furthermore, we documented that the overall construct displacement was a combination of tendon elongation and anchor displacement. These results imply that, despite the 20 cycles of 5 N–50 N conditioning, there was still tendon creep and seating of the anchor that occurred during the initial cyclic loading from 50 N–200 N. However, after this initial migration the constructs were quite stable with very little continued migration. This emphasizes the need for proper conditioning of the tendon graft prior to tibial fixation. We found no difference in the amount of tendon elongation or anchor displacement during cycling when comparing the AperFix and the AppianFx. On average, after 1500 cycles (50–200 N at 1 Hz) these devices both displayed 1.6 mm tendon elongation, 2.1 mm of anchor movement, and 3.8 mm of total construct displacement. These results are comparable to the results obtained in similar studies of other types of ACL anchors. For example, tests for a variety of interference screws, cross-pins, and cortical suspension devices with cyclic loads of 50 N– 200 N for 1000–1500 cycles produces total construct displacements of
3.1. Cyclic loading Fig. 3 displays a representative plot of the construct displacements obtained by the Optotrak system during the cyclic loading. Throughout the cycling the bone is rigidly fixed in the base grips. However, the displacement of the actuator, anchor, and tendon all appear to asymptotically increase during the cyclic loading. In addition, Fig. 3 shows that the total motion of the repair construct (actuator displacement) is a composite of anchor displacement and tendon elongation. Table 1 reports the tendon elongation and anchor displacement after 1500 loading cycles and indicates that there were no significant differences when comparing the two types of anchors. The repairs also showed similar total construct cyclic amplitudes (peak to valley displacement within a load cycle) throughout the testing. The cyclic amplitudes for both constructs at the start of the cyclic test (cycles 20–30) were
Fig. 3. A representative plot of the construct displacements during the 1500 cycles of loading from 50 to 200 N at 1 Hz. Most of the displacement occurred early in the cycling phase for all specimens tested.
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198 Table 1 Results obtained during the cyclic loading tests. AppianFx®
AperFix®
P-value
Tendon elongation (mm) Mean 1.85 (SD 1.14) Mean 1.30 (SD 0.53) 0.123 Anchor displacement (mm) Mean 2.32 (SD 0.88) Mean 1.79 (SD 1.69) 0.329 All data presented as mean (SD)
2–5 mm (Kousa et al., 2003; Zantop et al., 2004) while protocols that cycled up to 150 N for 100–2000 cycles showed total construct displacements of 2–11 mm with tendon elongations of 1.5 mm (Milano et al., 2006; Speirs et al., 2010; Zhang et al., 2007). Furthermore, the cyclic amplitudes reported in the current study are in line with those reported in the literature for ACL reconstructions using similar test specimens and protocols (Speirs et al., 2010). However, these cyclic amplitudes are smaller than the 2 mm cyclic elongations used in displacement control studies of ACL reconstructions (Nakano et al., 2000; Yamanaka et al., 1999). These cyclic elongation studies were conducted in porcine knees with both femoral and tibial fixation sites and graft lengths that were much longer than our gage length and it is therefore likely that these contrasts in test protocol can explain the differences in the cyclic displacements. This study has shown that similar maximum loads are reached for both devices during load to failure testing. However, maximum loading may have little relevance in the clinical setting. It has been shown by others (Ishibashi et al., 1997) that the stiffness of the ACL repair is of more importance than the ultimate strength when attempting to restore normal knee kinematics. The stiffness values reported in the current study for both repair constructs were comparable to each other and were higher than those values reported for other fixation devices in porcine femurs (Kousa et al., 2003; Milano et al., 2006; Shen et al., 2009; Speirs et al., 2010). This study also demonstrated a higher yield load for the AppianFx as compared to the AperFix. This result, coupled with the finding that these devices have similar stiffness values, implies the AppianFx repair construct will maintain its high stiffness character for greater loads as compared to the AperFix. Further examination of the ultimate loading data revealed several important findings. Fig. 4 shows a plot of the load, actuator displacement, tendon elongation, and anchor displacement versus time that is representative of all ultimate loading data sets obtained in this study. The load versus time plot shows an initial linear region (~0.5 s for data set shown in Fig. 4) of constant slope that transitions to a second region that is approximately linear with a lower slope. In the initial linear load region it is noted that the anchor does not move and that all construct displacement is due to tendon elongation. We noted no slippage of the graft from either the tissue grips or from the bone tunnel during the testing protocol. If the tendon and the anchor/bone interface are mechanically modeled as two springs in series with different spring constants (stiffness values) (To et al., 1999), then under tension the spring with the lower stiffness will displace and the other spring will not. Since, as shown in Fig. 4, the anchor was not initially displacing, the stiffness of the AperFix and AppianFx must be greater than that of the tendon. In a previous study (To et al., 1999) it was highlighted that the mechanical limitation of ACL reconstructions was the low stiffness of the anchor fixation and that to improve biomechanical performance efforts should be made to increase the stiffness of the anchor fixation. Given the results of this current
197
study, both the AperFix and the AppianFx may represent designs that provide appropriate stiffness to ACL reconstructions. As further shown by the representative data in Fig. 4, as the load is increased, eventually the anchor breaks free and begins to move. The transition point where the anchor begins to move also corresponds to where the initial linear region of the load plot ends. If the load was plotted versus the actuator displacement, then the slope of the initial linear region would be the construct stiffness and the transition point would indicate the yield load. Therefore, the yield loads reported for these tests are actually reflective of the static fixation strength of the anchors and not due to mechanical yielding or slippage of the tendon grafts. In this regard, it will take greater loads to disrupt the initial static fixation of the AppianFx anchor (~431 N) as compared to the AperFix anchor (~324 N). This biomechanical advantage provided by the AppianFx may be clinically important, since it has been advocated that the initial fixation strength of the ACL repair construct should be greater than 450 N (Zantop et al., 2004, 2006) in order to withstand loads associated with activities of daily living (Noyes et al., 1984) and stair decent (Morrison, 1969). Our findings that the yield load of these constructs is coupled to the initiation of anchor motion provides further support to the preferential reporting in the literature of yield load (Kousa et al., 2003; Nurmi et al., 2002) as compared to ultimate load. The load at 5 mm of anchor motion, which we feel may represent the critical threshold for construct clinical failure, was significantly higher for the AppianFx (667±144 N) as compared to the AperFx at (316±162 N). The lower load at 5 mm for the AperFix was likely influenced by a bimodal load–displacement curve noted in 5 out of 10 specimens during the ultimate loading test phase (Fig. 5C). In these bimodal cases, after the yield load was reached there was a decrease in the load carrying capacity and the construct displaced several millimeters under low loads. This was followed by a second rise in load to reach the maximum load. This was in contrast to the behavior of the other 5 AperFix and all of the AppianFx, which after reaching their yield load showed no appreciable decrease in the load carrying capacity until after reaching its ultimate tensile load (Fig. 5A, B). The bimodal curve could be related to the fact that the AperFix has two sets of arms that provide fixation. The proximal arms engage the cancellous bone while the distal arms compress the tendons against the aperture of the bone tunnel. The overall fixation for the AperFix is a composite of the fixation at each of these sites. In the bimodal cases, the drop in the load after the yield point may indicate the transition from fixation with the two sets of arms initially to fixation with just one set. This transition may possibly occur after the anchor begins to move and the distal arms are pulled out of the bone tunnel. The subsequent rise to a second higher peak load could reflect that the proximal arms are engaging the stronger cortical bone as they are pulled towards the aperture of the bone tunnel. It is remarkable that half of the AperFix samples displayed this bimodal curve as this may reflect an important variability in the biomechanical performance of this device. However, this concern stemming from variability in laboratory testing has not
Table 2 Results obtained during the ultimate loading tests. AppianFx® Stiffness (N/mm) Yield load (N) Load 5 mm (N) Max load (N)
AperFix®
Mean 217 (SD 31) Mean 242 Mean 431 (SD 79) Mean 324 Mean 667 (SD 144) Mean 316 Mean 1021 (SD 149) Mean 912 All data presented as mean (SD)
P-value (SD (SD (SD (SD
43) 45) 162) 116)
0.162 0.001 0.007 0.086
Fig. 4. A representative plot of the load, actuator displacement, anchor displacement, and tendon elongation versus time that was collected during the ultimate loading test.
198
M. Ehrensberger et al. / Clinical Biomechanics 28 (2013) 193–198
Fig. 5. Representative plots of load versus displacement during the ultimate loading tests. All AppianFx ® anchors displayed consistent load–displacement relationships and were similar data presented in (A). AperFix® anchors displaced two distinct types of load–displacement relationships. Five AperFix® specimens displayed data similar to (B) and five AperFix ® specimens displayed a bimodal load–displacement relationship noted in (C).
been established in the clinical literature as the AperFix has reported very good clinical outcomes (Connor, 2009; Uzumcugil et al., 2010). In light of these good clinical results for the AperFix, the higher load at 5 mm of AppianFx displacement and less variable ultimate loading performance may indicate similar or enhanced clinical performance for the AppianFx anchor. This in vitro study was conducted to assess the simulated in vivo biomechanical performance of ACL reconstructions and as such there are some limitations to consider. First, we performed these studies utilizing porcine femurs instead of human femurs. Human cadavers are usually elderly and the harvested tissues can display great sample-to-sample variation in their mechanical properties. Porcine femurs have been utilized in numerous ACL reconstruction studies (Ahmad et al., 2004; Kleweno et al., 2009; Kousa et al., 2003; Milano et al., 2006; Nakano et al., 2000; Nurmi et al., 2002; Seil et al., 1998; Shen et al., 2009; Speirs et al., 2010; Zhang et al., 2007) as an appropriate substitute for young human bone. We also used bovine extensor tendons as our simulated human ACL grafts. These tendons have been used as grafts in other biomechanical studies of ACL reconstructions (Ahmad et al., 2004; Kleweno et al., 2009) and have been shown to have equivalent viscoelastic and stiffness properties of human hamstring tendons (Donahue et al., 2001). In this study we also applied the load directly in line with the axis of the bone tunnel, which does not accurately represent the loading pattern experienced by the ACL in vivo when the tibia translates anteriorly. Therefore, the values of construct displacements and strengths reported in this study might not directly translate into clinical application. However, this method is widely used in the literature as it does allow for the direct biomechanical comparison of ACL reconstructions under “worst-case” conditions. To our knowledge, this is the first report of a mechanical test method that can accurately resolve and directly quantify the individual displacement components of ACL reconstructions under controlled cyclic and ultimate loading. Furthermore, by syncing the load data and the displacement data of the individual construct components, this novel method provides the means to determine exactly what load causes a specific construct component to displace. This method could find great utility in the development and evaluation of new ACL anchor designs because it allows for the direct assessment of the anchor motion under relevant loading. 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. Beynnon, B.D., Amis, A.A., 1998. In vitro testing protocols for the cruciate ligaments and ligament reconstructions. Knee Surg. Sports Traumatol. Arthrosc. 6, S70–S76. Connor, G., 2009. Analysis Of Outcome In ACL Reconstruction Using The Cayenne Medical AperFix System. Cayenne Medical Inc., White Paper. Donahue, T.L.H., Gregersen, C., Hull, M.L., Howell, S.M., 2001. Comparison of viscoelastic, structural, and material properties of double-looped anterior cruciate ligament grafts made from bovine digital extensor and human hamstring tendons. J. Biomech. Eng. 123, 162–169.
Engebretsen, L., Lew, W.D., Lewis, J.L., Hunter, R.E., 1989. Knee mechanics after repair of the anterior cruciate ligament. A cadaver study of ligament augmentation. Acta Orthop. 60, 703–709. Giurea, M., Z. P., Amis, A.A., Aichroth, P., 1999. Comparative pull-out and cyclic-loading strength tests of anchorage of hamstring tendon grafts in anterior cruciate ligament reconstruction. Am. J. Sports Med. 27, 621–625. Harvey, A., Thomas, N.P., Amis, A.A., 2005. Fixation of the graft in reconstruction of the anterior cruciate ligament. J. Bone Joint Surg. Br. 87-B, 593–603. Ishibashi, Y., R. T., Livesay, G.A., Stone, J.D., Fu, F.H., Woo, S.L., 1997. The effect of anterior cruciate ligament graft fixation site at the tibia on knee stability: evaluation using a robotic testing system. Arthroscopy 13, 177–182. Kleweno, C.P., Jacir, A.M., Gardner, T.R., Ahmad, C.S., Levine, W.N., 2009. Biomechanical evaluation of anterior cruciate ligament femoral fixation techniques. Am. J. Sports Med. 37, 339–345. Kousa, P., Jarvinen, T.L.N., Vihavainen, M., Kannus, P., Jarvinen, M., 2003. The fixation strength of six hamstring tendon graft fixation devices in anterior cruciate ligament reconstruction-part I: femoral site. Am. J. Sports Med. 31, 174–181. Mayr, R., Rosenberger, R., Agraharam, D., Smekal, V., El Attal, R., 2012. Revision anterior cruciate ligament reconstruction: an update. Arch. Orthop. Trauma Surg. 132, 1299–1313. Milano, G., Mulas, P.D., Ziranu, F., Piras, S., Manunta, A., Fabbriciani, C., 2006. Comparison between different femoral fixation devices for ACL reconstruction with doubled hamstring tendon graft: a biomechanical analysis. Arthroscopy 22, 660–668. Morrison, J.B., 1969. Function of the knee joint in various activities. Biomed. Eng. 4, 573–580. Nakano, H., Yasuda, K., Tohyama, H., Yamanaka, M., Wada, T., Kaneda, K., 2000. Interference screw fixation of doubled flexor tendon graft in anterior cruciate ligament reconstruction — biomechanical evaluation with cyclic elongation. Clin. Biomech. 15, 188–195. 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., Järvinen, T.L.N., Kannus, P., Sievänen, H., Toukosalo, J., Järvinen, M., 2002. Compaction versus extraction drilling for fixation of the hamstring tendon graft in anterior cruciate ligament reconstruction. Am. J. Sports Med. 30, 167–173. Rupp, S., Hopf, T., Hess, T., Seil, R., Kohn, D.M., 1999. Resulting tensile forces in the human bone–patellar tendon–bone graft: direct force measurement in vitro. Arthroscopy 15, 179–184. Sagomonyants, K.B., Jarman-Smith, M.L., Devine, J.N., Aronow, M.S., Gronowicz, G.A., 2008. The in vitro response of human osteoblasts to polyetheretherketone (PEEK) substrates compared to commercially pure titanium. Biomaterials 29, 1563–1572. Seil, R., Rupp, S., Krauss, P.W., Benz, A., Kohn, D.M., 1998. Comparison of initial fixation strength between biodegradable and metallic interference screws and a press-fit fixation technique in a porcine model. Am. J. Sports Med. 26, 815–819. Shen, P.-H., Lien, S.-B., Shen, H.-C., Wang, C.-C., Huang, G.-S., Chao, K.-H., et al., 2009. Comparison of different sizes of bioabsorbable interference screws for anterior cruciate ligament reconstruction using bioabsorbable bead augmentation in a porcine model. Arthroscopy 25, 1101–1107. Speirs, A., Simon, D., Lapner, P., 2010. Evaluation of a new femoral fixation device in a simulated anterior cruciate ligament reconstruction. Arthroscopy 26, 351–357. To, J.T., Howell, L.C.S.M., Hull, M.L., 1999. Contributions of femoral fixation methods to the stiffness of anterior cruciate ligament replacements at implantation. Arthroscopy 15, 379–387. Uzumcugil, O., Dogan, A., Dalyaman, E., Yalcinkaya, M., Akman, E., Ozturkmen, Y., et al., 2010. AperFix versus TransFix in reconstruction of anterior cruciate ligament. J. Knee Surg. 23, 029,035. Yamanaka, M., Yasuda, K., Tohyama, H., Nakano, H., Wada, T., 1999. The effect of cyclic displacement on the biomechanical characteristics of anterior cruciate ligament reconstructions. Am. J. Sports Med. 27, 772–777. Zantop, T., Weimann, A., Rümmler, M., Hassenpflug, J., Petersen, W., 2004. Initial fixation strength of two bioabsorbable pins for the fixation of hamstring grafts compared to interference screw fixation. Am. J. Sports Med. 32, 641–649. Zantop, T., Weimann, A., Schmidtko, R., Herbort, M., Raschke, M.J., Petersen, W., 2006. Graft laceration and pullout strength of soft-tissue anterior cruciate ligament reconstruction: in vitro study comparing titanium, poly-D, L-lactide, and poly-D, L-lactide–tricalcium phosphate screws. Arthroscopy 22, 1204–1210. Zhang, A.L., Lewicky, Y.M., Oka, R., Mahar, A., Pedowitz, R., 2007. Biomechanical analysis of femoral tunnel pull-out angles for anterior cruciate ligament reconstruction with bioabsorbable and metal interference screws. Am. J. Sports Med. 35, 637–642.