Abstract
Background:
Adjustable-loop cortical buttons for femoral fixation of bone-tendon-bone grafts have potential advantages over interference screw fixation; however, these devices have not been benchmarked biomechanically against interference screws.
Purpose/Hypothesis:
The purpose was to compare the time zero biomechanical properties of commercially available, adjustable-loop cortical button and metallic interference screws for femoral fixation of bone-tendon-bone grafts. It was hypothesized that no significant differences would be found in biomechanical properties between fixation techniques.
Study Design:
Controlled laboratory study.
Methods:
Adjustable-loop cortical buttons (n = 8) and metallic interference screws (n = 8) were used to fix matched pairs of human bone-tendon-bone allografts in porcine distal femurs. These constructs were preconditioned (10 N to 50 N at 1 Hz, 10 cycles), subjected to cyclic loading (50 N to 250 N at 1 Hz, 500 cycles), and then pulled to failure at 20 mm/min.
Results:
The loads to failure (mean ± SD, 700 ± 256 N vs 688 ± 215 N, P = .92) and linear stiffnesses (219 ± 48 N/mm vs 218 ± 49 N/mm, P = .97) for the adjustable-loop cortical button and metallic interference screws, respectively, were not significantly different. Cyclic displacement was higher in the adjustable-loop cortical button group (2.1 ± 0.6 mm vs 1.3 ± 0.4 mm, P = .01). The mechanism of failure was different between groups, with bone block slippage occurring most commonly in the interference screw group (n = 5) and fracture of the bone block through the suture hole occurring most commonly in the adjustable-loop cortical button group (n = 6).
Conclusion:
Adjustable-loop cortical buttons and interference screws have similar time zero failure loads, although cyclic displacement was higher with the adjustable-loop cortical buttons. The mean difference in displacement was less than 1 mm compared with the interference screw.
Clinical Relevance:
Adjustable-loop cortical buttons may be an acceptable alternative to an interference screw for femoral fixation of bone-tendon-bone grafts in anterior cruciate ligament reconstruction. The clinical relevance of the observed differences in cyclic displacement is unknown and should be evaluated in future studies.
Keywords
Recent literature on anterior cruciate ligament (ACL) reconstruction has placed a strong emphasis on “anatomic” reconstruction, with a particular focus on the placement of the femoral tunnel. As a result, many surgeons place the femoral tunnel lower on the lateral wall of the intercondylar notch. 17 When a bone-tendon-bone (BTB) graft is used, anatomic femoral tunnel placement can lead to unique challenges including difficult graft passage and an increased risk of graft-tunnel mismatch. 3 Therefore, several device companies have developed adjustable-loop cortical buttons for femoral fixation. While adjustable-loop cortical buttons have been examined for soft tissue grafts,12,18 no biomechanical studies have compared these buttons with interference screw fixation for grafts with bone blocks.
Femoral suspensory fixation for BTB grafts is used less often than interference screw fixation, the former being used in approximately 10% of cases. 20 Only 2 biomechanical studies have evaluated suspensory fixation with BTB grafts, and both used first-generation implants including an Endobutton with Mersilene tape 10 or a suture button with No. 5 Ethibond. 5 Both studies showed increased graft-tunnel motion with these constructs, compared with interference screw fixation, but similar ultimate loads to failure. Nonetheless, clinical success has been reported with this technique when a fixed-length button construct or polypropylene buttons and suture have been used. 19 It cannot be assumed, however, that adjustable-loop devices will perform similarly, as later generation sutures exhibit higher failure loads and stiffness.2,14 Adjustable-loop devices intended for use with soft tissue grafts have been carefully evaluated in the recent literature, after early concerns about increased cyclic displacement compared with fixed-length devices.1,18 More recent biomechanical studies suggested variability in cyclic displacement between different adjustable-loop devices.12,16 In addition, it has been shown that retensioning and knot tying after initial reduction of the graft with adjustable-loop devices dramatically reduce cyclic displacement. 15 This finding likely explains the equivalent clinical outcomes of soft tissue adjustable-loop and fixed-length devices.4,11 Nonetheless, biomechanical validation of adjustable-loop cortical buttons for femoral fixation of BTB grafts is clearly needed.
The purpose of this study was to compare the ultimate load and cyclic displacement of an adjustable-loop cortical button for femoral BTB fixation with the standard metallic interference screw in an animal cadaveric model using established loading protocols for ACL fixation devices. 18 We hypothesized that there would be no significant differences in time zero biomechanical properties between these two fixation methods.
Methods
Study Design and Specimen Preparation
This study was designed at the authors’ institution (Duke University) and was funded by an investigator-initiated grant from Arthrex. All surgical procedures and biomechanical testing were performed at the Arthrex research facility (Naples, Florida). Eighteen 6-month-old porcine distal femurs, cleaned of all soft tissue, were potted in fiberglass resin and mounted for femoral graft fixation. The medial femoral condyle was removed for improved access to the ACL footprint. A 2.4-mm Beath pin was passed under direct visualization through the lateral wall of the notch within the center of the native ACL footprint and out the proximal lateral cortex. An ACL guide was used such that all guide pins exited 15 mm proximal and 5 mm anterior to the lateral epicondyle (the established mean exit point using anteromedial portal drilling in hyperflexion 9 ). A 10-mm cannulated reamer was used to create a 30-mm–deep socket over the guide pin. Eighteen paired human allograft, patellar tendon BTB grafts (Joint Restoration Foundation, Centennial, Colorado) were prepared by trimming the patellar bone blocks to 10 mm in diameter and 20 mm in length. The higher quality bone block was chosen for fixation into the femur, and the opposite bone block on the graft was removed for testing. A 2-mm hole was drilled in the remaining bone block, 10 mm from the terminal end, allowing placement of a traction suture or adjustable-loop fixation device. Half of the paired BTB allografts (n = 9) were fixed in the femoral tunnel by use of a metallic interference screw, while the other half (n = 9) were fixed by use of an adjustable-loop cortical button (BTB TightRope; Arthrex).
Interference Screw Fixation Technique
A high tensile strength traction suture (No. 2 Fiberwire; Arthrex) was passed through the 2-mm hole in the allograft bone block, and the tails were shuttled through the 2.4-mm Beath pin out the lateral cortex. Before graft passage, a notch was placed at the anterior aperture of the femoral socket to ease insertion of the interference screw. The graft was then advanced into the socket by use of the traction suture until the bone plug was flush with the aperture of the tunnel. A nitinol guide wire was placed in the notch, and a 7 × 25–mm titanium interference screw was inserted anteriorly against the cancellous surface of the bone block under direct visualization while the graft was held under tension. The screw was advanced until the head was flush with the femoral tunnel aperture.
Cortical Button Fixation Technique
An adjustable-loop cortical button was assembled through the 2-mm hole in the allograft bone block in standard fashion (Figure 1). The lateral cortex was overdrilled with a 4-mm cannulated reamer to allow passage of the cortical button. Then, a shuttling suture was used to advance the button and associated sutures out the lateral femoral cortex, and the button was firmly seated against cortical bone. The adjustable loop was then shortened, thereby pulling the graft maximally into the femoral socket. A single author (J.C.R.) performed all graft tensioning using maximum manual tension. The tensioning sutures were left in place for retensioning after preconditioning during biomechanical testing.
Mechanical testing setup. The porcine femur was affixed so the direction of pull was at 30° of flexion. The soft tissue was secured to the cross-head with a custom interdigitizing freeze clamp with dry ice.
Biomechanical Testing
Mechanical testing was performed with an E10000 (Instron Corp), with a 10-kN load cell. The porcine femurs were oriented on the testing surface such that the direction of pull approximated that of the ACL at 30° of flexion, a more functional vector of pull compared with pulling directly in line with the femoral tunnel. The soft tissue of the graft was secured to the cross-head with a custom interdigitizing freeze clamp with dry ice (Figure 1). Each specimen was preconditioned by loading sinusoidally between 10 N and 50 N, for 10 cycles, at 1 Hz. For the adjustable-loop cortical button specimens, the graft was then held at 30 N to allow for retensioning by use of maximum manual force, and 5 alternating half-hitches were tied over the button. Testing then continued with 500 sinusoidal cycles between 50 N and 250 N, at 1 Hz. Testing was terminated with a pull to failure performed at 20 mm/min. The specimens were visually inspected after testing to determine modes of failure. Load and displacement data were recorded at 1000 Hz. A digital video camera was used to record the vertical positions of tissue marks placed on the graft and the femur, and these images were compared to determine the plastic displacement of the graft at the femoral socket (Figure 1). The ultimate load and stiffness were determined from the load-displacement curves generated during the pull-to-failure portion of the testing. Stiffness was determined by calculating the slope of the linear portion of the load-displacement curve immediately after the cyclic loading regimen. Any tightening or shortening of the adjustable-loop cortical button construct achieved during the retensioning step was measured from the load-displacement curve during the 30-N tension holding step. Plastic displacement that occurred during cyclic loading was determined from analysis of the digital video by use of MaxTRAQ Software with subpixel resolution (Innovision Systems, Inc). Metrics for comparison of the fixation devices included preconditioning displacement (mm), plastic cyclic displacement from digital video tracking data (mm), stiffness (N/mm), and ultimate load (N).
Statistical Analysis
A priori power analysis was performed with G*Power 3.0 software from published studies of adjustable-loop cortical fixation devices.12,18 Six samples per group were found sufficient to detect a 1-mm difference in cyclic displacement between groups with 80% power. A final sample size of 9 per group was chosen to account for any sample attrition during testing. All data were confirmed to have a normal distribution through use of the D’Agostino-Pearson omnibus test. Therefore, for each outcome variable, a 2-tailed nonpaired t test was performed to compare the adjustable-loop cortical button and interference screw groups. Means and SDs were reported for each group, and a P value less than .05 was considered statistically significant.
Results
Complete biomechanical data were available for all 9 specimens in the interference screw group and for 8 of the 9 specimens in the adjustable-loop cortical button group. One sample in the cortical button group failed during the first cycle of cyclic loading by the tendon’s avulsing off the bone block due to poor allograft quality. Therefore, data from this sample and its matched pair were excluded from the analysis, resulting in 8 samples in each group for final analysis.
Both groups showed minimal displacement during preconditioning, with no significant difference between groups (Figure 2D). During retensioning of the adjustable-loop cortical button, a mean ± SD 0.9 ± 0.6 mm of loop shortening was achieved. No difference was found between groups in ultimate load to failure (Figure 2A) or construct stiffness (Figure 2B). However, increased cyclic displacement was noted in the adjustable-loop cortical button group (Figure 2C). The mechanisms of failure were different between groups, with bone block slippage occurring most commonly in the interference screw group (n = 5) and bone block fracture through the suture hole occurring most commonly in the cortical button group (n = 6) (Table 1).
Biomechanical testing results between bone-tendon-bone metallic interference screw fixation and adjustable-loop cortical button fixation: (A) ultimate load to failure, (B) construct stiffness, (C) cyclic displacement, and (D) displacement during preconditioning.
Mode of Failure: Adjustable-Loop Cortical Button Versus Interference Screw
Discussion
The primary finding of this study is that an adjustable-loop cortical button had similar time zero failure loads and stiffness values compared with metallic interference screws for femoral fixation of BTB grafts in a cadaveric model. However, the cortical button had increased maximum cyclic displacement, although the difference on average was less than 1 mm.
Adjustable-loop cortical button devices have gained widespread adoption for the fixation of soft tissue grafts, with demonstrated clinical success, 6 although some biomechanical studies have shown increased cyclic displacement compared with fixed-length devices. 18 First-generation adjustable-loop devices came loaded with a closed loop of suture, precluding their use with grafts including bone blocks. Several manufacturers have now developed systems in which the suture can first be loaded through a drill hole in a bone block and then secondarily loaded into a locking mechanism within the button. Given the concerns with adjustable-loop devices raised in the biomechanics literature, as well as the added variable of the surgeon’s loading the suture into a locking mechanism in the operating room, biomechanical validation of these novel fixation devices is critical.
We compared a single, adjustable-loop cortical button with metallic interference screws for femoral fixation of clinical grade human BTB allografts within porcine femurs. No difference was observed in ultimate load to failure between these two groups. The ultimate load was between 650 N and 700 N for both constructs, which exceeds the highest anticipated forces during early rehabilitation (up to 500 N during isokinetic knee extension 8 ). Therefore, it is likely that both devices would be suitable for use with accelerated ACL rehabilitation protocols. 19 Similarly, no difference in construct stiffness was seen between the adjustable-loop cortical button and the interference screw. This is a critical point, since anterior knee laxity after ACL reconstruction is primarily determined by construct stiffness, not ultimate load, provided that excessive cyclic displacement does not occur over time. 7 The ultimate loads reported in this cadaveric study for the adjustable-loop cortical button (~650 N) are lower than those reported for the first generation of cortical buttons in isolated biomechanical testing (750-1000 N). 12 This can be explained by the fact that all failures in our study occurred by tissue failure (suture cutting through bone block or tendon avulsion) rather than device failure. One would expect that with the use of a BTB autograft, bone and tendon quality would be better, and the ultimate load would likely be higher than in our human cadaveric grafts. Regardless, this study suggests that clinical success with use of adjustable-loop cortical buttons is likely to be influenced by graft and specifically bone quality.
Six of the 9 specimens in the adjustable-loop cortical button group failed by catastrophic fracture of the bone block at the suture hole, and 1 additional sample failed by the sutures cutting through bone block. This finding has several technical implications for the use of such devices. The surgeon needs to carefully place the drill hole in the femoral bone block, ensuring a sufficient bone bridge (10-15 mm) between the drill hole and the end of the graft, as a short bone bridge might not provide sufficient resistance to catastrophic fracture. In addition, there is a risk of partial suture cutout, which could lead to pistoning of the suture loop through an elongated drill hole; however, this occurred in only a single specimen in our study. To reduce these complications, the graft should be carefully sized and shaped to allow smooth insertion into the femoral socket, since vigorous shortening of the adjustable-loop sutures to overcome poor graft fit, or excessive friction during insertion, may increase the risk of failure at or near the suture hole. An alternative solution is to pull the femoral bone block all the way into the femoral socket by use of the lead sutures and then secondarily shorten the adjustable loop while guiding the cortical button back under the iliotibial band. Finally, surgeons should carefully evaluate the bone quality of a BTB allograft if intending to use an adjustable-loop cortical button, since allograft preparation methods have been shown to have variable effects on graft biomechanical properties. 13
Our study showed a statistically significant difference in cyclic displacement between the adjustable-loop cortical button and interference screw fixation. The mean difference between groups was 0.8 mm, favoring the interference screw group. The performance of the adjustable-loop cortical button in our study (2.1 mm of mean cyclic displacement) was consistent with prior biomechanical analyses of traditional cortical buttons (cyclic displacement 1.8-2.2 mm). 12 This provides evidence that the added step wherein the surgeon constructs the BTB-specific adjustable loop does not significantly affect the properties of the device. In addition, the mean cyclic displacement of 2.1 mm falls below the accepted 3-mm threshold for clinical success.12,18 What cannot be concluded from our study is whether a less than 1-mm difference in cyclic displacement between the adjustable-loop cortical button and interference screws would translate into a clinically meaningful difference in outcome. A randomized controlled trial comparing an adjustable-loop cortical button with interference screws for fixation of soft tissue allografts showed no difference in KT-1000 arthrometer results or International Knee Documentation Committee objective scores at 24 months, despite biomechanical data showing approximately 2 mm of cyclic displacement with the same cortical button device.12,16,18 Similarly, a clinical study in which radio-opaque markers were implanted in soft tissue and BTB grafts showed that small amounts of graft-tunnel motion persist even 1 year after ACL reconstruction; however, these findings do not correlate with knee kinematics or patient outcomes. 11 These studies suggest that biomechanical data cannot always be correlated with clinical metrics of stability. Further clinical studies comparing adjustable-loop cortical buttons and interference screw fixation using BTB grafts are warranted. However, considering the available clinical and biomechanical literature, the authors believe that a less than 1-mm difference in cyclic displacement is unlikely to result in meaningful differences in clinical stability or outcome after BTB ACL reconstruction.
In this study, the biomechanical properties of the adjustable-loop cortical button were tested after retensioning and knot tying over the cortical button, following a preconditioning protocol. This is a critical point, since prior work has shown that this step significantly improves the fixation properties of this device. 15 On average, in our study, 0.8 mm of additional “tightening” was achieved during retensioning. We performed this step after “cycling” the knee in the operating room. Therefore, readers should be careful extrapolating the results of our testing to a clinical scenario in which retensioning and knot tying are not performed, as it is possible that the maximum cyclic displacement in this scenario could exceed the 3-mm threshold. Considering the published data, 15 as well as our own results, we recommend routinely including retensioning and knot tying when using an adjustable-loop cortical button to optimize fixation properties.
We recognize that this study has limitations. Porcine rather than human femurs were used; however, the porcine model has been shown to closely approximate bone quality and biomechanics of young adult bone. 18 Only clinical grade human BTB allografts were included in this study. Additionally, an isolated femur-graft construct was tested in this study, as described in multiple published studies.15,16,18 It is possible that use of a femur-graft-tibia construct may have altered the findings. Finally, biomechanical testing was performed at a single graft-to-femur angle of 30°, which was chosen to simulate the position of the ACL graft during complex loading conditions. With respect to external validity, the results of this study cannot be extrapolated to other companies’ cortical suspensory devices or sutures or to the use of screws of other materials (polyetheretherketone or composite), length, or diameter.
Conclusion
Adjustable-loop cortical buttons and metallic interference screws have similar time zero loads to failure and stiffness for femoral fixation of BTB grafts. However, a statistically significant increase in cyclic displacement was seen with the adjustable-loop cortical button group, although this increase was on average less than 1 mm compared with interference screw displacement.
Footnotes
One or more of the authors has declared the following potential conflict of interest or source of funding: This study was sponsored by Arthrex Inc (Naples, Florida). J.C.R. and D.T.M. receive fellowship support from Arthrex. J.C.R. has received support from Stryker Corporation for travel, lodging, and food during an educational session. K.G. has received payment for legal patent consulting services from Arthrex unrelated to the device in this study.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
