ORTHOPEDICS October 2010;33(10):722.

Mechanical Evaluation of Cross Pins Used for Femoral Fixation of Hamstring Grafts in ACL Reconstructions

by Gregory E. Bellisari, MD; Christopher C. Kaeding, MD; Alan S. Litsky, MD, ScD

Abstract

The goal of this study was to test the mechanical strength of 4 different cross pins currently available for femoral fixation by loading each cross pin to failure as received and determine the effect of 1 million cycles of fatigue loading. Additionally, the strength of resorbable pins was tested after prolonged exposure to biologic conditions. Six implants each of the Arthrotek LactoSorb (Biomet, Warsaw, Indiana), Mitek RigidFix (DePuy Mitek Inc, Raynham, Massachusetts), Arthrotek Bone Mulch Screw (Biomet), cortical allograft, and control were tested for 3-point failure without prior loading and after cyclic loading between 50 to 200 N at 10 Hz for 1 million cycles. The bioabsorbable pins were placed in sterile water at 37°C and tested after 2, 4, and 6 months for 3-point failure strength. All implants tested without antecedent loading demonstrated adequate strength for initial fixation for hamstring grafts. During fatigue testing, RigidFix implants (n=6) failed at 18,893±8365 cycles (with a central deformation of 0.48±0.11 mm prior to fracture). All of the other implants tested endured 1 million cycles of loading (50-200 N) without fracture or 1.5 mm central deformation. Neither of the bioabsorbable pins demonstrated a significant change in yield strength after prolonged exposure to water.

All implants tested demonstrated adequate strength for initial fixation of hamstring grafts. The metal and bone implants far exceed the strength required to sustain mechanical fixation until biological fixation occurs; both polymeric implants demonstrated that they maintained enough mechanical strength to achieve this goal.

Epidemiologic studies estimate an increasing incidence of anterior cruciate ligament (ACL) injuries because of an increased emphasis on sporting lifestyles.¹ Good outcomes associated with ACL reconstruction rely on factors such as graft choice, surgical technique, and method of fixation. Significant morbidity has been associated with bone–patellar tendon–bone grafts such as patellofemoral pain, patellar tendonitis, quadriceps atrophy, and loss of extensor mechanism.²

When hamstring grafts are harvested, overall hamstring performance with regard to knee flexion was not affected,³ and recovery of hamstring strength to 95% of preoperative values 3 years after hamstring tendon graft ACL reconstruction have been reported.⁴ With less harvest morbidity^2,4 and higher ultimate tensile load, the hamstring graft is becoming an increasingly popular choice.^5,6 Both bone–patellar tendon–bone and quadrupled hamstring tendon autografts are viable options for ACL reconstruction. Although graft choice is still debated, both demonstrate adequate strength to be used for ACL reconstruction. Bone–patellar tendon–bone has an ultimate tensile strength of approximately 2300 N⁷ and the quadruple semitendinosus/gracilis tendon graft⁵ an ultimate tensile strength as high as 4108 N.⁵ Both of these far exceed the 500 N load reported by Grood and Noyes⁸ for strenuous activity.

Hamstring grafts are an attractive choice only if fixation of the graft is optimal. Graft fixation is usually the weakest link in ACL reconstructions. Graft fixation needs to be strong enough to protect the reconstruction from failure during activities of daily living and early rehabilitation until the graft is biologically healed in the bone tunnels.⁹ The demand for higher fixation strength has also increased with the advent of aggressive rehabilitation, including early mobilization, after ACL reconstruction.^10-12

Cross-pin femoral fixation has shown good results with a 2-year follow-up.⁵ In a study that quantified the amount of permanent elongation in response to cyclic loading, femoral fixation with a cross pin showed no significant permanent elongation, while the button/tape femoral fixation along with its sutures (whipstitches) had significantly greater permanent elongation.¹³ Clinically, permanent elongation could translate to a loss of graft function. Initial biomechanical properties of cross pins have been shown to be superior to interference screws and soft tissue fixation.¹⁴ In an in vivo sheep study, cross pins were superior to interference screw fixation both initially and at 6 weeks postoperatively.⁹

The goals of this study were (1) to test the initial mechanical strength of 4 different cross-pin products currently available for femoral fixation by loading each cross pin to failure, and (2) to determine the effect of 1 million cycles of fatigue loading on the mechanical strength. A degradation study was also conducted with the bioabsorbable implants to test the effects of being subjected to a biologic environment. Our hypothesis was that all products tested would demonstrate adequate mechanical strength to provide adequate fixation for hamstring grafts in ACL reconstructions.

Materials and Methods

The cross pins used in our study were the Arthrotek LactoSorb (50 mm long×4 mm diameter polylactic/polyglycolic acid copolymer; Biomet, Warsaw, Indiana), the Mitek RigidFix (42×3.3 mm polylactic acid; DePuy Mitek Inc, Raynham, Massachusetts), the Arthrotek Bone Mulch Screw (25×10.5 mm Ti6Al4V alloy measuring 2 mm width where the pin was tested; Biomet), cortical allograft (50×4 mm freeze dried, ethylene oxide sterilized bone), and a control consisting of 3.2 mm (1/8”) diameter stainless steel rod (Figure). Each Bone Mulch Screw was prepared for 3-point bending by separating the central pin portion from the threaded portion so the pin would lie flat across the cylindrical supports. For the RigidFix cross pins, only 1 of the 2 pin systems was used for mechanical testing, as there was no means to test both pins simultaneously. Six implants of each design were randomly tested in 3-point bending and fatigue testing.

Figure: Cross pins tested (from left): LactoSorb, RigidFix, Bone Mulch Screw, cortical allograft, control.

For the 3-point bending, each implant was placed across a 12-mm span supported by 6.3-mm-diameter cylindrical supports. A 12-mm bone tunnel was chosen to represent the upper limits of bone tunnel diameter and worst-case scenario if widening of the tunnel were to occur postoperatively. Using a servohydraulic materials testing system (Bionix 858; MTS Systems, Eden Prairie, Minnesota), each implant was centrally loaded in a linear ramp at 1 mm/min until fracture or 1.5 mm central deformation. In our study, 1.5 mm of deformation was chosen as failure of the implant. Although strict, it is important to point out that the implant is only 1 part of the construct and deformation of 3 mm with EndoButton (Smith & Nephew, Memphis, Tennessee) fixation has led to clinically unacceptable results.¹⁵

Three-point fatigue testing was accomplished by cyclic loading between 50 to 200 N at 10 Hz for 1 million cycles followed by 3-point bending to failure as described above if the implant has not previously failed. Two hundred Newtons is representative of the literature values for ACL tension in normal walking. Five hundred Newtons is the highest reported value for ACL tension during normal walking, reported by Kuster et al.¹⁶ Grood and Noyes⁸ and Noyes et al,¹⁷ however, reported values close to 500 N to occur only during strenuous athletic activities. One million cycles was chosen to approximate 1 year of in vivo loading, sufficient time for the graft to be biologically secured in the femoral bone tunnel.

The bioabsorbable implants were tested after prolonged exposure to a biological environment. Both the RigidFix and LactoSorb implants were immersed in sterile water at 37°C for 6 months to measure the effect of hydrolysis on their mechanical properties. Six of each implant were pulled out at 2, 4, and 6 months for 3-point bending as described for the time-zero specimens.

Statistical analysis was carried out on yield load, load at 1.5 mm deformation, and structural stiffness. They were compared across the different designs using analysis of variance (ANOVA) and post-hoc analysis.

Results

The results of the 3-point bending test without antecedent cyclic fatigue testing are summarized in Table 1. The mean yield load for the Bone Mulch Screw was significantly higher than the other implants (P<.001). The mean yield loads in the LactoSorb implant and the allograft were similar and significantly higher than the RigidFix implant (P<.001). The stiffness of the cortical allograft, however, was more than twice that of the LactoSorb implant. The greatest stiffness was observed in the Bone Mulch Screw.

The mode of failure for the RigidFix implants and cortical allografts was by brittle fracture. The allografts fractured at 0.3±0.03 mm central deformation; the RigidFix implants after 1.5 mm deflection. The other 3 implants bent past 1.5 mm for failure without evidence of fracture. Load at 1.5 mm central deformation was 28% to 45% greater than the yield load.

Fatigue testing revealed that the RigidFix implants (n=6) fail at 18,893±8365 cycles (with a central deformation of 0.48±0.11 mm prior to fracture). All the other implants tested endured 1 million cycles of loading (50-200 N) without fracture or 1.5 mm central deformation. Post-fatigue 3-point bending showed no changes to the Bone Mulch Screw; however, there was a statistically significant (P<.05) increase in stiffness for both the LactoSorb and the cortical allograft. The fracture load of the allograft also increased significantly with fatigue loading, which was conducted dry and at room temperature.

After exposure to a biologic environment, neither of the bioabsorbable implants showed a statistically significant change in mean yield or stiffness from their zero time values under these experimental conditions. The RigidFix implants became slightly less compliant with time and more of the specimens broke before reaching 1.5 mm central displacement (Table 2).

Discussion

Femoral fixation of quadrupled hamstring grafts is a key element to a durable ACL reconstruction, and there are many implants designed to accomplish this. There is no consensus as to which method is superior. Our study focused on the mechanical characteristics of some of the more popular cross-pin fixation devices. Fixation of the device in the femoral metaphysis was specifically not examined, nor did we compare our results to other methods of fixation on the femoral side. Clinically, the biomechanics of the final graft construct will be determined by multiple factors: tibial fixation, femoral fixation, graft characteristics, and surgical technique. None of these factors was evaluated or accounted for in our study.

Clinically, the benefits of cross-pin femoral fixation appear to outweigh the drawbacks. The benefits of cross-pin fixation are rigid fixation and its proximity to the joint line, possibly decreasing the bungee effect and preventing tunnel widening.¹⁸ Soft tissue graft fixation methods are chosen on their strength and ability to resist deformation under cyclic loading. In our study, the strength of the Bone Mulch Screw and the LactoSorb implant/pin/device would be sufficient to withstand the reported values of normal walking. The Bone Mulch Screw displayed superior mechanical properties to the other cross pins. The RigidFix values for yield load are likely not representative of the final construct values because of its 2-pin configuration. The LactoSorb implant yield load (449±23 N) approaches the 454 N needed for daily activities reported by Grood and Noyes⁸ and Noyes et al.¹⁷ There are limitations to cross-pin fixation and reported complications. Marx and Spock¹⁹ reported a series of 2 cases with medial and lateral irritation of proud cross pins that required reoperation for hardware removal.

Several recent studies^20-23 have investigated the biomechanics of femoral fixation for ACL grafts. These studies have mostly compared cross-pin fixation to other forms of femoral fixation such as interference screws, considered the gold standard by many, and the EndoButton.

Kousa et al,¹³ in an ex vivo study, found no statistically significant difference between the Bone Mulch Screw and RigidFix cross pins when tested under a 50 mm/min bending load to failure before or after 1500 cycles of fatigue loading (50-200 N). Their test was performed with cadaveric quadrupled hamstring grafts looped over the femoral fixation devices in porcine femora. They concluded that the aforementioned cross pins provided greater fixation strength than 3 types of interference screws included in their protocol and that “the rigidity of the device itself improves the fixation characteristics of the implant.”

Our results show lower mean yield strength (187 N) compared to the findings of Kousa et al¹³ (868 N) in the RigidFix cross pin. Again, we tested 1 cross pin in a 2-pin system. Our loading rate was also much slower. For the Bone Mulch Screw, the mean yield strength (1146 N) found in this study was comparable to Kousa et al’s¹³ data (1112 N). The stiffness values in our study were different as a function of our decision to test the cross pin itself as opposed to testing the hamstring graft, the cross pin, and its fixation in the femoral bone.

Espejo-Baena et al¹⁴ compared, in porcine femurs, the biomechanical properties of the Bio-Transfix (Arthrex, Naples, Florida) and Biosteon (Stryker Endoscopy, San Jose, California) cross pins to an interference screw and to wrapping the graft around the femoral condyle. They tested these constructs by a single load until failure without any cyclic fatigue data. The strength of the cross pins (Bio-Transfix, 905 N; Biosteon, 684 N) and their stiffness values were significantly greater than the other 2 methods of fixation. They concluded that the biomechanical properties of the 2 cross-pin fixation systems were superior to the other 2 systems studied.

Ahmad et al²⁴ studied femoral fixation devices for the ACL by measuring fixation slippage. In porcine femora, the Endobutton, RigidFix, Bio-Transfix, and interference screw were subjected to cyclic loading from 50 N to 250 N for 1000 cycles. The load was held at 50 N for 45 seconds while slippage was measured at regular intervals. After cyclic fatigue, the grafts were loaded until failure. They found the graft slippage was greater in the RigidFix and interference screw constructs. The failure load of the interference screw was significantly less than the other 3 methods. The yield load of the cross pins were 737 N for the RigidFix and 746 N for the Bio-Transfix.

Our study evaluated only the cross pins and not the final construct seen in clinical practice. To study the implants in our idealized femoral tunnel, some modifications had to be made. For the RigidFix system, only 1 pin could be tested for accurate data across our cylindrical supports. The Bone Mulch Screw was detached from its proximal threaded portion so it would lie across the supports. This provides data relevant to the clinical application but differs from other studies. Specifically for the cortical allograft, its failure by brittle fracture likely may not occur in vivo or in solution.

Conclusion

The cross-pin femoral fixation with the implants tested in this study would provide adequate initial fixation for hamstring grafts. If one assumes that mechanical fixation is needed for 8 to 12 weeks until biologic fixation occurs, both polymeric implants showed gradual enough loss of mechanical properties to achieve this goal, and the other 2 implants far exceed the strength needed. Considering that 2 RigidFix pins are used clinically, the mechanical properties of either bioabsorbable cross pin appear to meet clinical demands and maintain their strength for a sufficient length of time to permit biologic fixation.

For further information: http://www.orthosupersite.com/view.aspx?rid=70286