- Email: [email protected]
In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation
THEKNE-02064; No of Pages 5 The Knee xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
The Knee
In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation John Nyland a,b,⁎, Ryan Krupp a, Joe Greene a, Richard Bowles a, Robert Burden a, David N.M. Caborn a a b
Division of Sports Medicine, Department of Orthopaedic Surgery, University of Louisville, 550 South Jackson Street, First Floor ACB, Louisville, KY 40202, United States Athletic Training Program, Kosair Charities College of Health and Natural Sciences, Spalding University, 901 South 4th Street, Louisville, KY 40203-2188, United States
a r t i c l e
i n f o
Article history: Received 6 March 2014 Received in revised form 2 March 2015 Accepted 17 March 2015 Available online xxxx Keywords: Fixation Soft tissue graft Interference screw Bioresorbable Implantation
a b s t r a c t Purpose: This mechanical study using an in vitro porcine model compared composite interference screw fixation of soft tissue ACL grafts in tibial tunnels. Methods: Forty-eight porcine profundus tendons and tibiae were divided into four groups of 12 closely matched specimens. Equivalent diameter grafts were assigned to each group. Tibial bone tunnels were drilled to 0.5 mm greater than graft diameter. Grafts were fixed in tunnels using one 10 × 35 mm composite interference screw designed by four different manufacturers. Maximal insertion torque and perceived within group mechanical testing outcome predictions were recorded. Constructs were potted and loaded into a six degrees of freedom clamp that placed the servohydraulic device tensile loading vector in direct tunnel alignment. Constructs were pre-loaded to 25 N, pre-conditioned between 0 and 50 N for 10 cycles (0.5 Hz), submaximally tested between 50 and 250 N for 500 cycles (one hertz) and load to failure tested at 20 mm/min. Results: Statistically significant differences were not observed between groups for displacement during submaximal cyclic loading, yield load, displacement at yield load, stiffness, ultimate load at failure and displacement at ultimate load. One composite screw group displayed a slightly greater proportion of specimens that required use of more than one screw during insertion. Conclusions: Under highly controlled conditions groups displayed comparable fixation. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The gold standard interference screw would be non-metallic, easy to use, and able to provide strong fixation until the graft incorporates, and then undergoes full resorption being replaced by bone [1]. Degradation kinetics differ substantially among different bioabsorbable polymers and numerous factors affect degradation rates, including molecular weight, sterilization, implant size, self-reinforcement, copolymer or stereocopolymer ratios, and processing techniques [2]. Postoperative radiographic evaluation of early generation poly-L-lactic acid (PLLA), polyglycolide and poly-D,L-lactide-co-glycolide screws revealed slow resorption rates and minimal, if any, osteoconductivity, despite good to excellent clinical results [3–11]. The addition of β-tricalcium phosphate (βTP) or hydroxyapatite (HA) helps buffer PLLA acidic breakdown providing a scaffold for bony ingrowth [12,13]. The addition of βTP or HA, can accelerate the incorporation of tendon grafts into bone tunnels and provide better mechanical properties [12–15]. Use of composite interference screws may lead to earlier and stronger graft incorporation, replacement of the screws with cancellous bone, and easier ⁎ Corresponding author at: Athletic Training Program, Kosair Charities College of Health and Natural Sciences, Spalding University, 901 South 4th Street, Louisville, KY 40203–2188. E-mail address: [email protected] (J. Nyland).
revision surgery. When ease of use and initial soft tissue graft fixation is comparable, composite screw selection should be based more on tissue remodeling and osteoconductive properties during resorption and the completeness of resorption. The purpose of this mechanical study using an in vitro porcine model was to compare composite interference screw fixation of soft tissue ACL grafts in tibial tunnels. Under strict controls the following 10 × 35 mm composite interference screw groups were compared: (Group 1) DePuy Milagro, (Group 2) Arthrex BioComposite, (Group 3) Stryker Biosteon, and (Group 4) Smith & Nephew Biosure HA (Fig. 1). The study hypothesis was that significant group differences would not exist. 2. Methods An a priori power analysis based on pilot testing revealed that a minimum of 10 specimens/group were needed to attain a statistical power of 0.80 at an alpha level of P = 0.05. To accommodate for possible methodological difficulties 12 specimens per study group were used. Fortyeight porcine profundus tendons and tibiae were divided into four groups of 12 closely matched specimens. From a group of 100 tibiae, pre-screening for bone mineral density (BMD) was performed using anterior–posterior and mediolateral dual-energy X-ray absorptiometry (DXA) (QDR 4500, Hologic, Inc., Bedford, MA, USA) scans to only
http://dx.doi.org/10.1016/j.knee.2015.03.009 0968-0160/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
2
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
Fig. 1. Composite 10 × 35 mm interference screw comparison (top-to-bottom): Arthrex Biocomposite, DePuy Milagro, Smith & Nephew Biosure HA, and Stryker Biosteon.
select specimens that simulated human tibial bone mineral density (approximately 1.09–1.30 g/cm2). Tibiae with comparable BMD were randomly assigned to a composite interference screw group. Equivalent diameter whip-stitched soft tissue tendon grafts were prepared using manufacturer recommended suture material: (Group 1) DePuy Mitek Orthocord, (Group 2) Arthrex FiberWire, (Group 3) Stryker Force Fiber, and (Group 4) Smith Nephew UltraBraid and were also assigned to their respective group. After identification of the tibial ACL footprint, using a drill guide fixed to the footprint, a guide wire was inserted at 55° from the tibial plateau through the native ACL insertion. Tibial tunnels 0.5 mm greater than graft diameter were then drilled. Grafts were fixed in tunnels using one of four different 10 × 35 mm composite interference screws: (Group 1) DePuy Milagro consisting of 30% βTP and 70% poly-L lactide-co-glycolide, (Group 2) Arthrex BioComposite consisting of 30% biphasic calcium phosphate and 70% poly-D -lactide, (Group 3) Stryker Biosteon consisting of 25% HA and 75% PLLA, or (Group 4) Smith & Nephew Biosure HA consisting of 25% HA and 75% PLLA [14]. The same fellowship trained surgeon (RK) performed all graft implantations and composite interference screw insertions. Maximal insertion torque (AccuForce Torque-chek, Ametek, Largo, FL) and perceived within group mechanical testing outcome predictions (0 to 10 visual analog scale, end range descriptors 0 = extremely poor, 10 = excellent) were recorded. Following preparation, constructs were potted in 7.62 cm (three inches) diameter, 17.78 cm (seven inches) long PVC tubes and loaded into a six degrees of freedom clamp. The specially designed clamp enabled the servohydraulic device (MTS 858, Eden Prairie, MN) tensile loading vector to be aligned directly with the tunnel (Fig. 2) [16–18]. This direct tensile load on the looped soft tissue tendon graft provided a “worst case” loading scenario. Constructs were pre-loaded to 25 N, followed by a pre-conditioning phase (0 to 50 N, 0.5 Hz, 10 cycles). After pre-conditioning, constructs underwent 500 submaximal loading cycles between 50 and 250 N at one hertz. Lastly, the constructs underwent load to failure testing at 20 mm/min with load (N) and displacement (mm) data recorded at 10 Hz. Groups were compared for independent variables of displacement during submaximal cyclic loading, yield load, ultimate load, stiffness, and displacement during ultimate load testing. Groups were also compared for the frequency of needing more than one composite interference screw to achieve fixation. Yield load represented the point in the stress–strain curve where a nonlinear relationship was observed. Stiffness was determined by recording peak load along the linear portion of the stress–strain curve, subtracting
Fig. 2. ACL graft–tibia construct mechanical testing in servohydraulic device using custom clamp.
this from the minimum load, and dividing this value by the construct displacement difference between the points. Ultimate load represented the maximum load observed prior to construct failure. 3. Statistical analysis Kolmogorov–Smirnov tests revealed that data displayed a normal distribution for yield and ultimate load at failure and for displacement at yield and ultimate failure loads; therefore parametric statistical analysis was performed. Normality was determined for both the complete dataset and for only those constructs that scored at least satisfactory for the perceived mechanical test outcome prediction. A series of oneway ANOVA were used to confirm that groups displayed comparable construct preparation characteristics and to evaluate group mechanical test differences. A Fisher's Exact Test was used to determine frequency between groups for the number of specimens that required more than one screw due to breakage. An alpha level of P b 0.05 was used to indicate statistical significance. All statistical procedures were performed using SPSS version 21.0 software (IBM Corporation, Armonk, NY). 4. Results Groups displayed comparable tibial BMD, graft diameter, graft length, composite interference screw insertion torque, perceived mechanical test outcome prediction, graft loop distance, and tibial tunnel length (Table 1). Group 3 had more specimens that required more than one screw (4/12 specimens, 33.3%) because of insertion breakage (Fisher's Exact Test = 6.9, P = 0.043). Statistically significant differences were not observed between groups for displacement during submaximal cyclic loading, yield load, displacement at yield load, stiffness, ultimate load at failure, and displacement at ultimate load (Table 2).
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
3
Table 1 Construct preparation data (ML = mediolateral, AP = anteroposterior). NA = not applicable. Group construct characteristics did not display statistically significant differences (mean ± standard deviation) [95% confidence interval]. Variable
DePuy (n = 12)
Arthrex (n = 12)
Stryker (n = 12)
Smith Nephew (n = 12)
Mean square
F
P
Screw dimensions Material properties
10 × 35 mm 30% βTP and 70% poly-L-lactide-co-glycolide
10 × 35 mm 25% HA and 75% PLLA
10 × 35 mm 25% HA and 75% PLLA
NA NA
NA NA
NA NA
Tibia BMD (average of ML, AP, g/cm2)
1.12 ± 0.09 [1.06–1.18]
10 × 35 mm 30% biphasic calcium phosphate and 70% poly-D-lactide 1.12 ± 0.1 [1.05–1.18]
Graft diameter (mm)
9.1 ± 0.3 [8.9–9.3]
9.1 ± 0.4 [8.8–9.4]
Graft length (mm)
77.3 ± 3.5 [75.0–79.5]
77.4 ± 3.7 [75.0–79.8]
Composite screw insertion torque (in-lbs)
41.1 ± 16.3 [30.8–51.4]
36.6 ± 12.1 [29.0–44.3]
Perceived outcome (0–10, modified VAS)
8.2 ± 2.4 [6.6–9.7]
7.2 ± 1.9 [6.0–8.3]
Graft loop distance (mm)
16.2 ± 3.2 [14.1–18.2]
17.3 ± 3.9 [14.8–19.7]
Tunnel length (mm)
44.4 ± 6.0 [40.6–48.3]
42.8 ± 4.7 [39.8–45.7]
Specimens requiring more than 1 screw due to breakage during insertion
0/12
1/12
5. Discussion Under highly controlled surgical and mechanical test conditions the four different composite interference screw groups displayed comparable ACL soft tissue graft tibial tunnel fixation. This finding suggests that determination of potential differences in the capacity for each composite interference screw type to positively influence intra-tunnel tissue remodeling and osteoconduction may be of greater importance when deciding upon which screw to use rather than on time zero resistance to repetitive submaximal and failure loads. Due to breakage during insertion, more Group 3 specimens required more than one screw to be used. Whether this was related to methodological or material issues such as porcine tissue use, inappropriate guide wire placement and tracking, tapered screw design, graft preparation method, and surgeon technique, cannot be confirmed from this study. With the high insertion torques generated during screw insertion, even subtle changes in screw angulation or surgeon-induced compressive forces may have contributed to the increased screw breakage observed for this group. In a study that used a combined bovine tibia, porcine soft tissue graft construct in vitro model and tested the same composite screws for ACL graft fixation as in the current study, Heard et al. [19] also reported comparable biomechanical properties during load-to-failure testing. However, this study did not include submaximal cyclic loading prior to load to failure testing, did not screen specimens for bone density, occasionally augmented grafts to increase diameter using an extra porcine flexor tendon section, created serially dilated bone tunnels, did not report interference screw insertion torque, and used tibial specimens that had been split sagittally, thereby not replicating the surgical
1.12 ± 0.08 [1.07–1.18] 1.13 ± 0.09 [1.07–1.19] Between groups = 0.000 Within groups = 0.008 9.1 ± 0.2 [8.9–9.3] 9.0 ± 0.4 [8.8–9.3] Between groups = 0.021 Within groups = 0.123 78.1 ± 4.1 [75.5–80.7] 77.3 ± 4.7 [74.3–80.2] Between groups = 1.89 Within groups = 16.24 31.9 ± 13.9 [23.0–40.7] 35.5 ± 10.3 [29.0–42.1] Between groups = 174.0 Within groups = 177.5 7.3 ± 2.5 [5.8–8.9] 7.4 ± 1.9 [6.2–8.6] Between groups = 2.35 Within groups = 4.70 16.9 ± 2.2 [15.5–18.3] 16.9 ± 2.8 [15.2–18.7] Between groups = 2.52 Within groups = 9.54 43.3 ± 5.1 [40.0–46.5] 44.0 ± 3.6 [41.7–46.3] Between groups = 6.69 Within groups = 24.35 4/12 0/12 Fisher's Exact Test = 6.86
0.026 0.99
0.169 0.92
0.116 0.95
0.98
0.41
0.50
0.68
0.26
0.85
0.28
0.84
0.043
condition. In a study that tested eight different soft tissue ACL graft fixation devices using a combined porcine tibia, bovine soft tissue graft construct in vitro model, Aga et al. [20] reported insignificant group differences for ultimate failure load and cyclic displacement. This study however did not screen porcine tibia for bone mineral density, tested only select constructs following serial tunnel dilation, and did not report screw insertion torque. Given the findings of the current study and the previous report by Heard et al. [19], potential differences in the capacity for each composite interference screw to positively influence intratunnel tissue remodeling and osteoconduction may be of greater importance than time zero resistance to repetitive submaximal and failure loads. 6. Limitations These in vitro test results provide only time zero tibial tunnel soft tissue graft fixation mechanical characteristics as would be present immediately following surgery. Therefore, important information regarding graft–tibial tunnel healing, osteoconductivity, tunnel widening prevention, soft tissue graft–bone tunnel integration, and remodeling cannot be determined. Additionally, use of a porcine model may not be representative of young human tibiae and soft tissue grafts. Porcine tibiae were however screened for BMD using DXA scanning to better simulate human tibia. Previous soft tissue ACL graft fixation studies have found porcine soft tissue grafts [17,18] and tibiae [16–18] to be more representative of young human tissue than cadaveric tissues. Five of 48 total constructs (10.4%) tested provided a less than satisfactory perceived mechanical test outcome prediction score from the
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
4
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
Table 2 Construct mechanical test group comparisons for constructs that displayed a minimum satisfactory (outcome prediction ≥ 5) on the VAS measure of perceived outcome. NA = not applicable. Statistically significant group differences were not evident (mean ± standard deviation) [95% confidence interval]. Variable
DePuy (n = 11)
Arthrex (n = 11)
Stryker (n = 10)
Smith Nephew (n = 11)
Screw dimensions Material properties
10 × 35 mm 30% βTP and 70% poly-L-lactide-co-glycolide
10 × 35 mm 25% HA and 75% PLLA
Cyclic displacement (mm)
1.1 ± 0.2 [0.9–1.2]
10 × 35 mm 30% biphasic calcium phosphate and 70% poly-D-lactide 1.2 ± 0.4 [0.9–1.4]
Yield load (N)
845.1 ± 243.7 [681.3–1008.8]
Displacement at yield load (mm)
F
P
10 × 35 mm NA 25% HA and 75% PLLA NA
NA NA
NA NA
1.5 ± 0.6 [1.1–1.9]
1.4 ± 0.4 [1.1–1.6]
2.6
0.06
795.2 ± 157.5 [689.4–901.0]
797.6 ± 293.3 [587.8–1007.4]
684.1 ± 163.9 [574.0–794.2]
7.4 ± 1.2 [6.6–8.2]
7.8 ± 1.7 [6.7–8.9]
8.2 ± 1.1 [7.4–9.0]
7.6 ± 0.5 [7.2–7.9]
Stiffness (N/mm)
196.6 ± 23.8 [180.6–212.6]
185.3 ± 26.7 [167.4–203.2]
185.0 ± 25.2 [166.9–203.0]
182.2 ± 12.9 [173.5–190.8]
Ultimate load at failure (N)
1113.2 ± 362.2 [869.9–1356.6]
1051.2 ± 244.5 [887.0–1215.5]
1073.8 ± 378.7 [802.8–1344.7]
920.3 ± 283.5 [729.8–1110.7]
Displacement at ultimate load (mm)
11.1 ± 4.1 [8.4–13.9]
10.6 ± 2.7 [8.8–12.4]
13.9 ± 1.9 [12.5–15.2]
12.0 ± 2.8 [10.1–13.9]
surgeon. Possible explanations for this are that DXA BMD screening does not provide a true volumetric trabecular bone density measurement specifically in the ACL reconstruction tibial region of interest, therefore some specimens may have had trabecular densities that differed from what the DXA BMD screening estimate suggested. Another possible explanation is that interference screw insertion integrity naturally varies and in the surgical setting this often leads the knee surgeon to use supplemental or “back-up” extra-tunnel fixation [21]. The final possibility is that despite careful construct preparation and guide wire positioning, some less than optimal composite interference screw placements may have occurred. Because of occasional interference screw breakage during insertion some constructs also required the use of more than one composite interference screw prior to achieving perceived secure soft tissue graft–tibial tunnel fixation. Similar to polymer-based interference screws, breakage during insertion is an issue for composite interference screws [14]. 7. Conclusions Under rigidly-controlled surgical and mechanical test conditions, the four composite interference screw groups tested displayed comparable ACL soft tissue graft fixation characteristics. Acknowledgment We would like to thank the Stryker Corporation (Project: No. S08007), Mahwah, New Jersey, USA for sponsoring this study. References [1] Bourke HE, Salmon LJ, Waller A, Winalski CS, Williams HA, Linklater JM, et al. Randomized controlled trial of osteoconductive fixation screws for anterior cruciate ligament reconstruction: a comparison of the Calaxo and Milagro screws. Arthroscopy 2013;29(1):74–82.
Mean square
Between groups = 0.45 Within groups = 0.17 Between groups = 51,149.1 Within groups = 48,322.8 Between groups = 1.26 Within groups = 1.39 Between groups = 445.2 Within groups = 517.5 Between groups = 76,490.4 Within groups = 102,679.5 Between groups = 21.3 Within groups = 9.1
1.06 0.38
0.91 0.45
0.86 0.47
0.75 0.53
2.35 0.09
[2] Weiler A, Hoffman RF, Stahelin AC, Helling HJ, Sudkamp NP. Biodegradable implants in sports medicine: the biological base. Arthroscopy 2000;16(3):305–21. [3] Johnston M, Morse A, Arrington J, Pliner M, Gasser S. Resorption and remodeling of hydroxyapatite-poly-L-lactic acid composite anterior cruciate ligament interference screws. Arthroscopy 2011;27(12):1671–8. [4] Lind M, Feller J, Webster KE. Tibial bone widening is reduced by polylactate/ hydroxyapatite interference screws compared to metal screws after ACL reconstruction with hamstring grafts. Knee 2009;16(6):447–51. [5] Macarini L, Milillo P, Mocci A, Vinci R, Ettorre GC. Poly-L-lactic acid–hydroxyapatite (PLLA–HA) bioabsorbable interference screws for tibial graft fixation in anterior cruciate ligament (ACL) reconstruction surgery: MR evaluation of osteo-integration and degradation features. Radiol Med 2008;113(8):1185–97. [6] Macarini L, Murrone M, Marini S, Mocci A, Ettorre GC. MRI in ACL reconstructive surgery with PDLLA bioabsorbable interference screws: evaluation of degradation and osteointegration processes of bioabsorbable screws. Radiol Med 2004;107(1–2): 47–57. [7] Martinek V, Seil R, Lattermann C, Watkins SC, Fu FH. The fate of the poly-L-lactic acid interference screw after anterior cruciate ligament reconstruction. Arthroscopy 2001;17(1):73–6. [8] McGuire DA, Barber FA, Milchgrub S, Wolchok JC. A postmortem examination of poly-L-lactic acid interference screws 4 months after implantation during anterior cruciate ligament reconstruction. Arthroscopy 2001;17(9):988–92. [9] Radford MJ, Noakes J, Read J, Wood DG. The natural history of a bioabsorbable interference screw used for anterior cruciate ligament reconstruction with a 4-strand hamstring technique. Arthroscopy 2005;21(6):707–10. [10] Warden WH, Friedman R, Teresi LM, Jackson DW. Magnetic resonance imaging of bioabsorbable polylactic acid interference screws during the first 2 years after anterior cruciate ligament reconstruction. Arthroscopy 1999;15(5):474–80. [11] Barber FA, Dockery WD. Long-term absorption of poly-L-lactic acid interference screws. Arthroscopy 2006;22(8):820–6. [12] Barber FA, Dockery WD. Long-term absorption of beta-tricalcium phosphate poly-Llactic acid interference screws. Arthroscopy 2008;24(4):441–7. [13] Bodde EW, Wolke JG, Kowalski RS, Jansen JA. Bone regeneration of porous betatricalcium phosphate (Conduit TCP) and of biphasic calcium phosphate ceramic (Biosel) in trabecular defects in sheep. J Biomed Mater Res A 2007;82(3):711–22. [14] Suchenski M, McCarthy MB, Chowaniec D, Hansen D, McKinnon W, Apostolakos J, et al. Material properties and composition of soft-tissue fixation. Arthroscopy 2010;26(6):821–31. [15] Huangfu X, Zhao J. Tendon–bone healing enhancement using injectable tricalcium phosphate in a dog anterior cruciate ligament reconstruction model. Arthroscopy 2007;23(5):255–462. [16] Chang HC, Nyland J, Nawab A, Burden R, Caborn DN. Biomechanical comparison of the bioabsorbable RetroScrew system, BioScrew XtraLok with stress equalization tensioner, and 35-mm Delta Screws for tibialis anterior graft–tibial tunnel fixation in porcine tibia. Am J Sports Med 2005;33(7):1057–64.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
J. Nyland et al. / The Knee xxx (2015) xxx–xxx [17] Duffee AR, Brunelli JA, Nyland J, Burden R, Nawab A, Caborn D. Bioabsorbable screw divergence angle, not tunnel preparation method influences soft tissue tendon graft–bone tunnel fixation in healthy bone. Knee Surg Sports Traumatol Arthrosc 2007;15(1):17–25. [18] Dunkin BS, Nyland J, Duffee AR, Brunelli JA, Burden R, Caborn D. Soft tissue tendon graft fixation in serially dilated or extraction-drilled tibial tunnels: a porcine model study using high-resolution quantitative computerized tomography. Am J Sports Med 2007;35(3):448–57. [19] Heard WM, Paller DJ, Christino MA, Behrens SB, Biercevicz A, Fadale PD, et al. Effect of a single interference screw on the mechanical properties of porcine anterior
5
cruciate ligament reconstruction grafts. Am J Orthop (Belle Mead NJ) 2013;42(4): 168–72. [20] Aga C, Rasmussen MT, Smith SD, Jansson KS, LaPrade RF, Engebretsen L, et al. Biomechanical comparison of interference screws and combination screw and sheath devices for soft tissue anterior cruciate ligament reconstruction on the tibial side. Am J Sports Med 2013;41(4):841–8. [21] Eisen SH, Davidson PA, Rivenburgh DW. Supplemental tibial fixation for anterior cruciate reconstruction. Arthroscopy 2008;24(9):1078–80.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
Contents lists available at ScienceDirect
The Knee
In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation John Nyland a,b,⁎, Ryan Krupp a, Joe Greene a, Richard Bowles a, Robert Burden a, David N.M. Caborn a a b
Division of Sports Medicine, Department of Orthopaedic Surgery, University of Louisville, 550 South Jackson Street, First Floor ACB, Louisville, KY 40202, United States Athletic Training Program, Kosair Charities College of Health and Natural Sciences, Spalding University, 901 South 4th Street, Louisville, KY 40203-2188, United States
a r t i c l e
i n f o
Article history: Received 6 March 2014 Received in revised form 2 March 2015 Accepted 17 March 2015 Available online xxxx Keywords: Fixation Soft tissue graft Interference screw Bioresorbable Implantation
a b s t r a c t Purpose: This mechanical study using an in vitro porcine model compared composite interference screw fixation of soft tissue ACL grafts in tibial tunnels. Methods: Forty-eight porcine profundus tendons and tibiae were divided into four groups of 12 closely matched specimens. Equivalent diameter grafts were assigned to each group. Tibial bone tunnels were drilled to 0.5 mm greater than graft diameter. Grafts were fixed in tunnels using one 10 × 35 mm composite interference screw designed by four different manufacturers. Maximal insertion torque and perceived within group mechanical testing outcome predictions were recorded. Constructs were potted and loaded into a six degrees of freedom clamp that placed the servohydraulic device tensile loading vector in direct tunnel alignment. Constructs were pre-loaded to 25 N, pre-conditioned between 0 and 50 N for 10 cycles (0.5 Hz), submaximally tested between 50 and 250 N for 500 cycles (one hertz) and load to failure tested at 20 mm/min. Results: Statistically significant differences were not observed between groups for displacement during submaximal cyclic loading, yield load, displacement at yield load, stiffness, ultimate load at failure and displacement at ultimate load. One composite screw group displayed a slightly greater proportion of specimens that required use of more than one screw during insertion. Conclusions: Under highly controlled conditions groups displayed comparable fixation. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The gold standard interference screw would be non-metallic, easy to use, and able to provide strong fixation until the graft incorporates, and then undergoes full resorption being replaced by bone [1]. Degradation kinetics differ substantially among different bioabsorbable polymers and numerous factors affect degradation rates, including molecular weight, sterilization, implant size, self-reinforcement, copolymer or stereocopolymer ratios, and processing techniques [2]. Postoperative radiographic evaluation of early generation poly-L-lactic acid (PLLA), polyglycolide and poly-D,L-lactide-co-glycolide screws revealed slow resorption rates and minimal, if any, osteoconductivity, despite good to excellent clinical results [3–11]. The addition of β-tricalcium phosphate (βTP) or hydroxyapatite (HA) helps buffer PLLA acidic breakdown providing a scaffold for bony ingrowth [12,13]. The addition of βTP or HA, can accelerate the incorporation of tendon grafts into bone tunnels and provide better mechanical properties [12–15]. Use of composite interference screws may lead to earlier and stronger graft incorporation, replacement of the screws with cancellous bone, and easier ⁎ Corresponding author at: Athletic Training Program, Kosair Charities College of Health and Natural Sciences, Spalding University, 901 South 4th Street, Louisville, KY 40203–2188. E-mail address: [email protected] (J. Nyland).
revision surgery. When ease of use and initial soft tissue graft fixation is comparable, composite screw selection should be based more on tissue remodeling and osteoconductive properties during resorption and the completeness of resorption. The purpose of this mechanical study using an in vitro porcine model was to compare composite interference screw fixation of soft tissue ACL grafts in tibial tunnels. Under strict controls the following 10 × 35 mm composite interference screw groups were compared: (Group 1) DePuy Milagro, (Group 2) Arthrex BioComposite, (Group 3) Stryker Biosteon, and (Group 4) Smith & Nephew Biosure HA (Fig. 1). The study hypothesis was that significant group differences would not exist. 2. Methods An a priori power analysis based on pilot testing revealed that a minimum of 10 specimens/group were needed to attain a statistical power of 0.80 at an alpha level of P = 0.05. To accommodate for possible methodological difficulties 12 specimens per study group were used. Fortyeight porcine profundus tendons and tibiae were divided into four groups of 12 closely matched specimens. From a group of 100 tibiae, pre-screening for bone mineral density (BMD) was performed using anterior–posterior and mediolateral dual-energy X-ray absorptiometry (DXA) (QDR 4500, Hologic, Inc., Bedford, MA, USA) scans to only
http://dx.doi.org/10.1016/j.knee.2015.03.009 0968-0160/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
2
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
Fig. 1. Composite 10 × 35 mm interference screw comparison (top-to-bottom): Arthrex Biocomposite, DePuy Milagro, Smith & Nephew Biosure HA, and Stryker Biosteon.
select specimens that simulated human tibial bone mineral density (approximately 1.09–1.30 g/cm2). Tibiae with comparable BMD were randomly assigned to a composite interference screw group. Equivalent diameter whip-stitched soft tissue tendon grafts were prepared using manufacturer recommended suture material: (Group 1) DePuy Mitek Orthocord, (Group 2) Arthrex FiberWire, (Group 3) Stryker Force Fiber, and (Group 4) Smith Nephew UltraBraid and were also assigned to their respective group. After identification of the tibial ACL footprint, using a drill guide fixed to the footprint, a guide wire was inserted at 55° from the tibial plateau through the native ACL insertion. Tibial tunnels 0.5 mm greater than graft diameter were then drilled. Grafts were fixed in tunnels using one of four different 10 × 35 mm composite interference screws: (Group 1) DePuy Milagro consisting of 30% βTP and 70% poly-L lactide-co-glycolide, (Group 2) Arthrex BioComposite consisting of 30% biphasic calcium phosphate and 70% poly-D -lactide, (Group 3) Stryker Biosteon consisting of 25% HA and 75% PLLA, or (Group 4) Smith & Nephew Biosure HA consisting of 25% HA and 75% PLLA [14]. The same fellowship trained surgeon (RK) performed all graft implantations and composite interference screw insertions. Maximal insertion torque (AccuForce Torque-chek, Ametek, Largo, FL) and perceived within group mechanical testing outcome predictions (0 to 10 visual analog scale, end range descriptors 0 = extremely poor, 10 = excellent) were recorded. Following preparation, constructs were potted in 7.62 cm (three inches) diameter, 17.78 cm (seven inches) long PVC tubes and loaded into a six degrees of freedom clamp. The specially designed clamp enabled the servohydraulic device (MTS 858, Eden Prairie, MN) tensile loading vector to be aligned directly with the tunnel (Fig. 2) [16–18]. This direct tensile load on the looped soft tissue tendon graft provided a “worst case” loading scenario. Constructs were pre-loaded to 25 N, followed by a pre-conditioning phase (0 to 50 N, 0.5 Hz, 10 cycles). After pre-conditioning, constructs underwent 500 submaximal loading cycles between 50 and 250 N at one hertz. Lastly, the constructs underwent load to failure testing at 20 mm/min with load (N) and displacement (mm) data recorded at 10 Hz. Groups were compared for independent variables of displacement during submaximal cyclic loading, yield load, ultimate load, stiffness, and displacement during ultimate load testing. Groups were also compared for the frequency of needing more than one composite interference screw to achieve fixation. Yield load represented the point in the stress–strain curve where a nonlinear relationship was observed. Stiffness was determined by recording peak load along the linear portion of the stress–strain curve, subtracting
Fig. 2. ACL graft–tibia construct mechanical testing in servohydraulic device using custom clamp.
this from the minimum load, and dividing this value by the construct displacement difference between the points. Ultimate load represented the maximum load observed prior to construct failure. 3. Statistical analysis Kolmogorov–Smirnov tests revealed that data displayed a normal distribution for yield and ultimate load at failure and for displacement at yield and ultimate failure loads; therefore parametric statistical analysis was performed. Normality was determined for both the complete dataset and for only those constructs that scored at least satisfactory for the perceived mechanical test outcome prediction. A series of oneway ANOVA were used to confirm that groups displayed comparable construct preparation characteristics and to evaluate group mechanical test differences. A Fisher's Exact Test was used to determine frequency between groups for the number of specimens that required more than one screw due to breakage. An alpha level of P b 0.05 was used to indicate statistical significance. All statistical procedures were performed using SPSS version 21.0 software (IBM Corporation, Armonk, NY). 4. Results Groups displayed comparable tibial BMD, graft diameter, graft length, composite interference screw insertion torque, perceived mechanical test outcome prediction, graft loop distance, and tibial tunnel length (Table 1). Group 3 had more specimens that required more than one screw (4/12 specimens, 33.3%) because of insertion breakage (Fisher's Exact Test = 6.9, P = 0.043). Statistically significant differences were not observed between groups for displacement during submaximal cyclic loading, yield load, displacement at yield load, stiffness, ultimate load at failure, and displacement at ultimate load (Table 2).
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
3
Table 1 Construct preparation data (ML = mediolateral, AP = anteroposterior). NA = not applicable. Group construct characteristics did not display statistically significant differences (mean ± standard deviation) [95% confidence interval]. Variable
DePuy (n = 12)
Arthrex (n = 12)
Stryker (n = 12)
Smith Nephew (n = 12)
Mean square
F
P
Screw dimensions Material properties
10 × 35 mm 30% βTP and 70% poly-L-lactide-co-glycolide
10 × 35 mm 25% HA and 75% PLLA
10 × 35 mm 25% HA and 75% PLLA
NA NA
NA NA
NA NA
Tibia BMD (average of ML, AP, g/cm2)
1.12 ± 0.09 [1.06–1.18]
10 × 35 mm 30% biphasic calcium phosphate and 70% poly-D-lactide 1.12 ± 0.1 [1.05–1.18]
Graft diameter (mm)
9.1 ± 0.3 [8.9–9.3]
9.1 ± 0.4 [8.8–9.4]
Graft length (mm)
77.3 ± 3.5 [75.0–79.5]
77.4 ± 3.7 [75.0–79.8]
Composite screw insertion torque (in-lbs)
41.1 ± 16.3 [30.8–51.4]
36.6 ± 12.1 [29.0–44.3]
Perceived outcome (0–10, modified VAS)
8.2 ± 2.4 [6.6–9.7]
7.2 ± 1.9 [6.0–8.3]
Graft loop distance (mm)
16.2 ± 3.2 [14.1–18.2]
17.3 ± 3.9 [14.8–19.7]
Tunnel length (mm)
44.4 ± 6.0 [40.6–48.3]
42.8 ± 4.7 [39.8–45.7]
Specimens requiring more than 1 screw due to breakage during insertion
0/12
1/12
5. Discussion Under highly controlled surgical and mechanical test conditions the four different composite interference screw groups displayed comparable ACL soft tissue graft tibial tunnel fixation. This finding suggests that determination of potential differences in the capacity for each composite interference screw type to positively influence intra-tunnel tissue remodeling and osteoconduction may be of greater importance when deciding upon which screw to use rather than on time zero resistance to repetitive submaximal and failure loads. Due to breakage during insertion, more Group 3 specimens required more than one screw to be used. Whether this was related to methodological or material issues such as porcine tissue use, inappropriate guide wire placement and tracking, tapered screw design, graft preparation method, and surgeon technique, cannot be confirmed from this study. With the high insertion torques generated during screw insertion, even subtle changes in screw angulation or surgeon-induced compressive forces may have contributed to the increased screw breakage observed for this group. In a study that used a combined bovine tibia, porcine soft tissue graft construct in vitro model and tested the same composite screws for ACL graft fixation as in the current study, Heard et al. [19] also reported comparable biomechanical properties during load-to-failure testing. However, this study did not include submaximal cyclic loading prior to load to failure testing, did not screen specimens for bone density, occasionally augmented grafts to increase diameter using an extra porcine flexor tendon section, created serially dilated bone tunnels, did not report interference screw insertion torque, and used tibial specimens that had been split sagittally, thereby not replicating the surgical
1.12 ± 0.08 [1.07–1.18] 1.13 ± 0.09 [1.07–1.19] Between groups = 0.000 Within groups = 0.008 9.1 ± 0.2 [8.9–9.3] 9.0 ± 0.4 [8.8–9.3] Between groups = 0.021 Within groups = 0.123 78.1 ± 4.1 [75.5–80.7] 77.3 ± 4.7 [74.3–80.2] Between groups = 1.89 Within groups = 16.24 31.9 ± 13.9 [23.0–40.7] 35.5 ± 10.3 [29.0–42.1] Between groups = 174.0 Within groups = 177.5 7.3 ± 2.5 [5.8–8.9] 7.4 ± 1.9 [6.2–8.6] Between groups = 2.35 Within groups = 4.70 16.9 ± 2.2 [15.5–18.3] 16.9 ± 2.8 [15.2–18.7] Between groups = 2.52 Within groups = 9.54 43.3 ± 5.1 [40.0–46.5] 44.0 ± 3.6 [41.7–46.3] Between groups = 6.69 Within groups = 24.35 4/12 0/12 Fisher's Exact Test = 6.86
0.026 0.99
0.169 0.92
0.116 0.95
0.98
0.41
0.50
0.68
0.26
0.85
0.28
0.84
0.043
condition. In a study that tested eight different soft tissue ACL graft fixation devices using a combined porcine tibia, bovine soft tissue graft construct in vitro model, Aga et al. [20] reported insignificant group differences for ultimate failure load and cyclic displacement. This study however did not screen porcine tibia for bone mineral density, tested only select constructs following serial tunnel dilation, and did not report screw insertion torque. Given the findings of the current study and the previous report by Heard et al. [19], potential differences in the capacity for each composite interference screw to positively influence intratunnel tissue remodeling and osteoconduction may be of greater importance than time zero resistance to repetitive submaximal and failure loads. 6. Limitations These in vitro test results provide only time zero tibial tunnel soft tissue graft fixation mechanical characteristics as would be present immediately following surgery. Therefore, important information regarding graft–tibial tunnel healing, osteoconductivity, tunnel widening prevention, soft tissue graft–bone tunnel integration, and remodeling cannot be determined. Additionally, use of a porcine model may not be representative of young human tibiae and soft tissue grafts. Porcine tibiae were however screened for BMD using DXA scanning to better simulate human tibia. Previous soft tissue ACL graft fixation studies have found porcine soft tissue grafts [17,18] and tibiae [16–18] to be more representative of young human tissue than cadaveric tissues. Five of 48 total constructs (10.4%) tested provided a less than satisfactory perceived mechanical test outcome prediction score from the
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
4
J. Nyland et al. / The Knee xxx (2015) xxx–xxx
Table 2 Construct mechanical test group comparisons for constructs that displayed a minimum satisfactory (outcome prediction ≥ 5) on the VAS measure of perceived outcome. NA = not applicable. Statistically significant group differences were not evident (mean ± standard deviation) [95% confidence interval]. Variable
DePuy (n = 11)
Arthrex (n = 11)
Stryker (n = 10)
Smith Nephew (n = 11)
Screw dimensions Material properties
10 × 35 mm 30% βTP and 70% poly-L-lactide-co-glycolide
10 × 35 mm 25% HA and 75% PLLA
Cyclic displacement (mm)
1.1 ± 0.2 [0.9–1.2]
10 × 35 mm 30% biphasic calcium phosphate and 70% poly-D-lactide 1.2 ± 0.4 [0.9–1.4]
Yield load (N)
845.1 ± 243.7 [681.3–1008.8]
Displacement at yield load (mm)
F
P
10 × 35 mm NA 25% HA and 75% PLLA NA
NA NA
NA NA
1.5 ± 0.6 [1.1–1.9]
1.4 ± 0.4 [1.1–1.6]
2.6
0.06
795.2 ± 157.5 [689.4–901.0]
797.6 ± 293.3 [587.8–1007.4]
684.1 ± 163.9 [574.0–794.2]
7.4 ± 1.2 [6.6–8.2]
7.8 ± 1.7 [6.7–8.9]
8.2 ± 1.1 [7.4–9.0]
7.6 ± 0.5 [7.2–7.9]
Stiffness (N/mm)
196.6 ± 23.8 [180.6–212.6]
185.3 ± 26.7 [167.4–203.2]
185.0 ± 25.2 [166.9–203.0]
182.2 ± 12.9 [173.5–190.8]
Ultimate load at failure (N)
1113.2 ± 362.2 [869.9–1356.6]
1051.2 ± 244.5 [887.0–1215.5]
1073.8 ± 378.7 [802.8–1344.7]
920.3 ± 283.5 [729.8–1110.7]
Displacement at ultimate load (mm)
11.1 ± 4.1 [8.4–13.9]
10.6 ± 2.7 [8.8–12.4]
13.9 ± 1.9 [12.5–15.2]
12.0 ± 2.8 [10.1–13.9]
surgeon. Possible explanations for this are that DXA BMD screening does not provide a true volumetric trabecular bone density measurement specifically in the ACL reconstruction tibial region of interest, therefore some specimens may have had trabecular densities that differed from what the DXA BMD screening estimate suggested. Another possible explanation is that interference screw insertion integrity naturally varies and in the surgical setting this often leads the knee surgeon to use supplemental or “back-up” extra-tunnel fixation [21]. The final possibility is that despite careful construct preparation and guide wire positioning, some less than optimal composite interference screw placements may have occurred. Because of occasional interference screw breakage during insertion some constructs also required the use of more than one composite interference screw prior to achieving perceived secure soft tissue graft–tibial tunnel fixation. Similar to polymer-based interference screws, breakage during insertion is an issue for composite interference screws [14]. 7. Conclusions Under rigidly-controlled surgical and mechanical test conditions, the four composite interference screw groups tested displayed comparable ACL soft tissue graft fixation characteristics. Acknowledgment We would like to thank the Stryker Corporation (Project: No. S08007), Mahwah, New Jersey, USA for sponsoring this study. References [1] Bourke HE, Salmon LJ, Waller A, Winalski CS, Williams HA, Linklater JM, et al. Randomized controlled trial of osteoconductive fixation screws for anterior cruciate ligament reconstruction: a comparison of the Calaxo and Milagro screws. Arthroscopy 2013;29(1):74–82.
Mean square
Between groups = 0.45 Within groups = 0.17 Between groups = 51,149.1 Within groups = 48,322.8 Between groups = 1.26 Within groups = 1.39 Between groups = 445.2 Within groups = 517.5 Between groups = 76,490.4 Within groups = 102,679.5 Between groups = 21.3 Within groups = 9.1
1.06 0.38
0.91 0.45
0.86 0.47
0.75 0.53
2.35 0.09
[2] Weiler A, Hoffman RF, Stahelin AC, Helling HJ, Sudkamp NP. Biodegradable implants in sports medicine: the biological base. Arthroscopy 2000;16(3):305–21. [3] Johnston M, Morse A, Arrington J, Pliner M, Gasser S. Resorption and remodeling of hydroxyapatite-poly-L-lactic acid composite anterior cruciate ligament interference screws. Arthroscopy 2011;27(12):1671–8. [4] Lind M, Feller J, Webster KE. Tibial bone widening is reduced by polylactate/ hydroxyapatite interference screws compared to metal screws after ACL reconstruction with hamstring grafts. Knee 2009;16(6):447–51. [5] Macarini L, Milillo P, Mocci A, Vinci R, Ettorre GC. Poly-L-lactic acid–hydroxyapatite (PLLA–HA) bioabsorbable interference screws for tibial graft fixation in anterior cruciate ligament (ACL) reconstruction surgery: MR evaluation of osteo-integration and degradation features. Radiol Med 2008;113(8):1185–97. [6] Macarini L, Murrone M, Marini S, Mocci A, Ettorre GC. MRI in ACL reconstructive surgery with PDLLA bioabsorbable interference screws: evaluation of degradation and osteointegration processes of bioabsorbable screws. Radiol Med 2004;107(1–2): 47–57. [7] Martinek V, Seil R, Lattermann C, Watkins SC, Fu FH. The fate of the poly-L-lactic acid interference screw after anterior cruciate ligament reconstruction. Arthroscopy 2001;17(1):73–6. [8] McGuire DA, Barber FA, Milchgrub S, Wolchok JC. A postmortem examination of poly-L-lactic acid interference screws 4 months after implantation during anterior cruciate ligament reconstruction. Arthroscopy 2001;17(9):988–92. [9] Radford MJ, Noakes J, Read J, Wood DG. The natural history of a bioabsorbable interference screw used for anterior cruciate ligament reconstruction with a 4-strand hamstring technique. Arthroscopy 2005;21(6):707–10. [10] Warden WH, Friedman R, Teresi LM, Jackson DW. Magnetic resonance imaging of bioabsorbable polylactic acid interference screws during the first 2 years after anterior cruciate ligament reconstruction. Arthroscopy 1999;15(5):474–80. [11] Barber FA, Dockery WD. Long-term absorption of poly-L-lactic acid interference screws. Arthroscopy 2006;22(8):820–6. [12] Barber FA, Dockery WD. Long-term absorption of beta-tricalcium phosphate poly-Llactic acid interference screws. Arthroscopy 2008;24(4):441–7. [13] Bodde EW, Wolke JG, Kowalski RS, Jansen JA. Bone regeneration of porous betatricalcium phosphate (Conduit TCP) and of biphasic calcium phosphate ceramic (Biosel) in trabecular defects in sheep. J Biomed Mater Res A 2007;82(3):711–22. [14] Suchenski M, McCarthy MB, Chowaniec D, Hansen D, McKinnon W, Apostolakos J, et al. Material properties and composition of soft-tissue fixation. Arthroscopy 2010;26(6):821–31. [15] Huangfu X, Zhao J. Tendon–bone healing enhancement using injectable tricalcium phosphate in a dog anterior cruciate ligament reconstruction model. Arthroscopy 2007;23(5):255–462. [16] Chang HC, Nyland J, Nawab A, Burden R, Caborn DN. Biomechanical comparison of the bioabsorbable RetroScrew system, BioScrew XtraLok with stress equalization tensioner, and 35-mm Delta Screws for tibialis anterior graft–tibial tunnel fixation in porcine tibia. Am J Sports Med 2005;33(7):1057–64.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009
J. Nyland et al. / The Knee xxx (2015) xxx–xxx [17] Duffee AR, Brunelli JA, Nyland J, Burden R, Nawab A, Caborn D. Bioabsorbable screw divergence angle, not tunnel preparation method influences soft tissue tendon graft–bone tunnel fixation in healthy bone. Knee Surg Sports Traumatol Arthrosc 2007;15(1):17–25. [18] Dunkin BS, Nyland J, Duffee AR, Brunelli JA, Burden R, Caborn D. Soft tissue tendon graft fixation in serially dilated or extraction-drilled tibial tunnels: a porcine model study using high-resolution quantitative computerized tomography. Am J Sports Med 2007;35(3):448–57. [19] Heard WM, Paller DJ, Christino MA, Behrens SB, Biercevicz A, Fadale PD, et al. Effect of a single interference screw on the mechanical properties of porcine anterior
5
cruciate ligament reconstruction grafts. Am J Orthop (Belle Mead NJ) 2013;42(4): 168–72. [20] Aga C, Rasmussen MT, Smith SD, Jansson KS, LaPrade RF, Engebretsen L, et al. Biomechanical comparison of interference screws and combination screw and sheath devices for soft tissue anterior cruciate ligament reconstruction on the tibial side. Am J Sports Med 2013;41(4):841–8. [21] Eisen SH, Davidson PA, Rivenburgh DW. Supplemental tibial fixation for anterior cruciate reconstruction. Arthroscopy 2008;24(9):1078–80.
Please cite this article as: Nyland J, et al, In situ comparison of varying composite tibial tunnel interference screws used for ACL soft tissue graft fixation, Knee (2015), http://dx.doi.org/10.1016/j.knee.2015.03.009