Dimensional Assessment of Continuous Loop Cortical SuspensionDevices and Clinical Implications for Intraoperative Button Flippingand Intratunnel Graft Length

Travis Lee Turnbull,1 Christopher M. LaPrade,1 Sean D. Smith,1 Robert F. LaPrade,1,2 Coen A. Wijdicks1

1Steadman Philippon Research Institute, BioMedical Engineering, 181 W. Meadow Drive, Suite 1000, Vail, Colorado 81657,2The Steadman Clinic, Orthopaedic Surgery, 181 W. Meadow Drive, Suite 400, Vail, Colorado 81657

Received 9 January 2015; accepted 10 March 2015

Published online 25 May 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.22904

ABSTRACT: Continuous loop cortical suspension devices have been demonstrated to be more consistent and biomechanically superiorcompared to adjustable loop devices; however, continuous loop devices present unique challenges compared to adjustable loop devices,especially in short tunnel reconstruction applications. Specifically, adjustable loop devices have the advantage of a “one size fits all”approach, and the ability to tension these devices following button flipping allows for the intratunnel graft length to be maximized.Nevertheless, the reliability of continuous loop devices has sustained their widespread use. We hypothesized that continuous loopcortical suspension devices from different manufacturers would exhibit equivalent 15mm loop lengths, as advertised. Loop length wasmeasured using a tensile testing machine. Contrary to our hypothesis, continuous loop cortical suspension devices with equivalentadvertised lengths exhibited different loop lengths (up to 27% discrepancy). Inconsistencies with regards to manufacturers’ reportedloop lengths for continuous loop devices could have serious clinical implications and additionally complicate technique transferal amongdevices. Consequently, the manufacturers’ accurate and complete disclosure of the dimensions and specifications associated with eachcontinuous loop device is critical. Furthermore, surgeon awareness of true loop length dimensions and inconsistencies among devices isneeded to ensure optimal implantation and resultant clinical outcomes. � 2015 Orthopaedic Research Society. Published by WileyPeriodicals, Inc. J Orthop Res 33:1327–1331, 2015.

Keywords: cortical button; flipping distance; loop length; ACL reconstruction; ligament reconstruction

Recent biomechanical studies have identified continu-ous loop cortical suspension devices to be more consis-tent and biomechanically superior to adjustable loopdevices, especially with regard to cyclic displacement attime-zero.1,2 Nevertheless, there remain challengesunique to the proper implementation of continuous loopcompared to adjustable loop devices. Specifically, thefixed loop length of continuous loop devices necessitatescareful surgical planning and execution to achieve thedesired graft fixation and minimize the potential devel-opment of time-zero and/or future problems (e.g.,inability to flip/deploy the button, cortical blowout orbone bridge collapse, insufficient intratunnel graftlength, amount of unused tunnel, and tunnel widen-ing).3–6 These concerns are magnified for anteromedialfemoral tunnel reaming for modern anterior cruciateligament (ACL) reconstructions, in which anatomicfemoral tunnel placement requires shorter femoraltunnels.5,7,8 In contrast, adjustable loop devices havethe advantage of a “one size fits all” approach, andthe ability to tension these devices following buttonflipping allows for the intratunnel graft length to bemaximized.

Given the inherent challenges associated with properimplementation of continuous loop cortical suspensiondevices, especially in cases of short reconstructiontunnels where intratunnel graft length needs to be

maximized, surgeons depend on the manufacturers’accurate and precise disclosure of pertinent dimensionsand specifications. As such, one might reasonably as-sume that continuous loop cortical suspension devicesfrom different manufacturers that are advertised ashaving an identical loop length would be correspondinglyidentical in loop length and have a high accuracy to theadvertised loop length. However, as the present studydemonstrates, these assumptions are inaccurate andcould have serious clinical implications.

During pilot testing for another study,2 we inadver-tently identified differences between the loop lengthsof continuous loop devices from different manufac-turers that were advertised as having equivalent15mm loop lengths. To our knowledge, no otherstudies have reported on the differences in loop lengthbetween devices and the concomitant clinical implica-tions. Therefore, our objective was to investigatedifferences in the loop length among advertised equiv-alent continuous loop cortical suspension devices. Wehypothesized that continuous loop cortical suspensiondevices from different manufacturers would exhibitequivalent loop lengths, as advertised.

METHODSDevicesContinuous loop cortical suspension devices from three manu-facturers, each with a manufacturer reported loop length of15mm, were evaluated in this study (Fig. 1). The devices werethe ENDOBUTTON CL ULTRA (Smith & Nephew, Inc.,Andover, MA), XO BUTTON (ConMed Linvatec, Inc., Largo,FL), and RIGIDLOOP (DePuy Mitek, Inc., Raynham, MA).Manufacturing lot numbers were identical for each device fora given manufacturer. All devices were donated gratis, andthe study was funded internally to avoid any perceived bias.

Investigation performed at the Department of BioMedical Engi-neering, Steadman Philippon Research Institute, Vail, Colorado,USA.Correspondence to: Travis L. Turnbull, (T: 970-479-9797 ext.5252; E-mail: tturnbull@sprivail.org)

# 2015 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2015 1327

Mechanical TestingTwenty-four continuous loop devices from three differentmanufacturers (n¼ 8/each) were tested using an identicalprotocol. Testing was performed using a clinically relevantcustom fixture (Fig. 2) and a tensile testing machine (Electro-Puls E10000, Instron Systems, Norwood, MA). Measurementerror of the testing machine was certified by Instron to be lessthan or equal to � 0.01mm and � 0.3% of the indicated force.Two, 5mm thick steel inserts, each with a single holecorresponding to different manufacturers’ recommended drillhole diameter (4.5mm for the Endobutton and Rigidloop, and5.0mm for the XO Button), simulated the femoral cortex.Device loops were placed around a 4.5mm diameter hardened,precision ground steel rod, simulating half the diameter of a9mm soft tissue graft,1 which was rigidly attached to thetesting machine actuator with a custom clevis. The corticalsuspension devices were then passed through the hole of thesteel insert using the passing sutures and secured againstthe inferior surface of the steel insert. The force vector was in-

line with the suture loop and perpendicular to the button, asdescribed previously.1,2 Note that, prior to device insertion ortesting, the actuator of the testing machine was independentlypositioned such that the steel rod was in 1N of compressionwith each steel insert and these positions were recorded forsubsequent loop length calculations.

Devices were subjected to a single tensile ramp loadingprofile from 1 to 75N over 10 s. Loop lengths at 5 and 75Nwere calculated as the sum of the steel insert thickness, therespective displacement of the testing machine actuatorrelative to the rod in contact with the top of the steel insert(�1N), and the diameter of the steel rod (Fig. 2). Thethickness of each steel insert and diameter of the rod weremeasured with digital calipers (Swiss Precision Instruments,Inc., Garden Grove, CA) with a manufacturer reportedaccuracy of 0.03mm.

Statistical AnalysisAnalysis metrics included loop lengths at 5 and 75N and thedifference between them, as well as the difference betweenthe loop length at 75N and advertised loop length.The primary loop length was defined as the length of thecontinuous loop suture under a tensile load of 75N to berepresentative of the load and corresponding length at whichthe device would be fixed intraoperatively. Loop length wascompared to the manufacturer advertised loop length viaone-sample t-tests using IBM SPSS Statistics, Version 20(Armonk, NY). Intra-device elongation was assessed usingpaired-sample t-tests. Inter-device elongation and differencesin loop length were assessed using Welch’s analysis ofvariance (ANOVA) with a Games–Howell test for post hocgroup comparisons.

RESULTSIntra-Device Dimensional Accuracy and PrecisionDevice loop lengths (mm) are reported here (mean�SD)and in Table 1. The Rigidloop loop length at 75N(15.1�0.3) was accurate and not significantly different

Figure 1. Photographs of continuous loop cortical suspension devices evaluated in this study and their respective product packagingdescriptions with equivalent advertised lengths. Note that devices were photographed and are displayed at an equivalent scale.

Figure 2. Photograph (left) and section view schematic (right)depicting the fixtures and methodology, respectively, used tomeasure the loop length of continuous loop cortical suspensiondevices. Left: A¼ testing machine actuator; B¼ custom clevisand steel rod; C¼ steel insert with a single hole. Right: looplengths at 5 and 75N (Step 2) were calculated as the sum of the(C) steel insert thickness, (D) displacement of the testingmachine actuator relative to the rod in contact with the top ofthe steel insert (Step 1), and diameter of the (B) steel rod.

1328 TURNBULL ET AL.

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2015

from the manufacturer advertised loop length (p¼ 0.264)(Fig. 3). In contrast, the Endobutton and XO Button looplengths at 75N (17.7� 0.3, 19.0�0.7, respectively) werenot accurate and were significantly different fromthe manufacturer advertised loop length (p<0.001,p<0.001, respectively).

Intra-Device ElongationIncreased loading of the devices from 5 to 75Nresulted in a significant increase in loop length (mm)(mean�SD) for the Endobutton (0.5�0.1, p< 0.001),Rigidloop (0.4�0.2, p¼0.001), and XO Button(1.3�0.4, p< 0.001). However, compared to both theEndobutton and Rigidloop, the XO Button loop lengthincreased significantly more when loaded from 5 to75N (p¼0.003 and p¼ 0.001, respectively).

Inter-Device Dimensional VariabilityIn addition to the significant differences observed formeasured device loop lengths at 75N in comparison tothe manufacturer advertised loop length, significantdifferences were also observed between devices(Fig. 3). The mean differences between loop lengths forthe Endobutton compared to the Rigidloop (2.6mm,p< 0.001) and XO Button (1.3mm, p¼ 0.002) deviceswere significant. Similarly, the mean difference in looplength between the Rigidloop and XO Button (3.9mm,p< 0.001) was significant.

DISCUSSIONContrary to our hypothesis, continuous loop corticalsuspension devices from different manufacturersexhibited significantly different loop lengths, in spiteof being advertised as equivalent in loop length.Deviations from the advertised loop length, whetheras-manufactured and/or upon initial loading of thecontinuous loop, could negatively affect the outcome ofsurgical reconstructions. For example, in the setting ofACL reconstructions, several concerns may arise fromdeviations in the advertised loop length and includean inability to flip/deploy the button, cortical blowoutor bone bridge collapse, insufficient intratunnel graftlength, amount of unused tunnel, and tunnel widen-ing.3–6 Notably, several of these potential problems arecompounding; e.g., in an attempt to facilitate buttonflipping, the socket must be drilled deeper whichincreases the risk of cortical blowout or bone bridgecollapse5 and the amount of unused tunnel correspond-ingly increases which could promote tunnel widening.6

Similarly, a longer-than-advertised loop length mayfacilitate button flipping but may not provide sufficientintratunnel graft length9,10 and would increase theamount of unused tunnel and concomitant tunnelwidening.

The above concerns are all worthy of consideration;however, the amount of unused tunnel (related to buttonflipping) and graft insertion distance have been shown to

Table 1. Lengths of Advertised Equivalent Continuous Loop Cortical Suspension Devices Under 5 and 75N of Force,and the Differences Between These States and Between the Manufacturer Reported and 75N Loop Length

DeviceManufacturer reported

loop length (mm)

5NLoop length

(mm)a

75NLoop length(mm)a,c,d

D5–75N(mm)a

DReported &75N(mm)a

Endobutton 15 17.2 (0.3) 17.7 (0.3)b 0.5 (0.1) 2.7 (0.3)Rigidloop 15 14.7 (0.4) 15.1 (0.3) 0.4 (0.2) 0.1 (0.3)XO Button 15 17.8 (0.8) 19.0 (0.7)b 1.3 (0.4)e 4.0 (0.7)aData reported as: mean (standard deviation).bLoop length was significantly different from the manufacturer reported loop length(p< 0.001, one-sample t-test).cSignificant differences existed between the 5 and 75N loop length of each device (p�0.001 for all posthoc comparisons, paired-sample t-test).dSignificant differences existed between all pairings of devices (p�0.002 for all post hoccomparisons, Games-Howell).eLoop length increased significantly more compared to the other devices (p�0.003 for all post hoccomparisons, Games-Howell).

Figure 3. Illustration highlighting differences in the loop lengths and resultant graft insertion distances for continuous loop corticalsuspension devices deployed in a clinically relevant, short tunnel representative of anatomically placed femoral tunnels for ACLreconstruction. In this example, the total tunnel length is 25mm, and advertised 15mm continuous loop devices with manufacturerrecommended socket depths for device flipping distance are shown with the expected graft insertion distance of 10mm (dashed line).*The XO Button packaging literature explicitly advises against the deployment of the device in tunnels shorter than 30mm.

LENGTH DIFFERENCES OF CONTINUOUS LOOP DEVICES 1329

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2015

affect the quality and success of a reconstruction. First,femoral ACL reconstruction sockets are drilled deeperthan the graft insertion distance (related to the adver-tised loop length) by necessity to provide additionaldistance to flip the button (i.e., the flipping distance). Assuch, the principal concern is the inherent tradeoffbetween the ease of button flipping (longer tunnels) andresultant effects of longer tunnels (e.g., increasedamount of unused tunnel leading to tunnel widening6

and risk of cortical blowout or bone bridge collapse5).The current literature is limited and does not contain aconsensus regarding the effect of tunnel widening ongraft function and knee laxity; nevertheless, tunnelexpansion is generally undesirable and may lead toincreased laxity.6,11 Second, a minimum of approximate-ly 15mm of graft insertion has been reported to benecessary for optimal reconstruction outcomes,9 andgraft insertions of only several millimeters less (e.g., 5, 9,and 14mm compared to 17mm) resulted in reducedstrength and stiffness, especially in the critically impor-tant early postoperative healing period.10 However, noneof the devices achieved 15mm of graft insertion inclinically relevant short tunnels, and only one deviceachieved the expected 10mm insertion distance based onits advertised loop length in this study (Fig. 3). Addition-ally, current graft preconditioning protocols may notoptimally reduce initial graft viscoelastic lengthening(approximately 2mm using current clinical techni-ques).12 Therefore, the concomitant effect of less-than-expected graft insertion and viscoelastic lengtheningcould have serious clinical implications.

The above concerns are magnified for anatomic ACLreconstructions, which utilize an anteromedial arthro-scopic portal for femoral tunnel reaming and subse-quently create shorter femoral tunnels, relative to thetranstibial technique.5,7,8 Recent studies have demon-strated an ability to create femoral tunnels greater than30mm in length;13,14 however, Golish et al. reported themedian length to be 25mm for anteromedial reamedtunnels7 which is consistent with the technique andexperience of the senior surgeon. Therefore, the currentstudy results are schematically represented in 25mmfemoral tunnels for ACL reconstruction (Fig. 3). Notethat in this clinically relevant example, the XO Buttonshould technically not be deployed according to packag-ing literature which explicitly advises against deploy-ment of the device in tunnels shorter than 30mm.Moreover, upon inspection of the advertised loop length(15mm), recommended flipping distance (15mm), andrequirement of a minimum 5mm bone bridge, the XOButton packaging literature seems incorrect becausetunnels shorter than 35mm would violate the dimen-sional constraints. Additionally, the Endobutton was theonly device with a 10mm continuous loop option;therefore, although not tested in this study, the 10mmdevice may be better suited to this short tunnel applica-tion and may provide an additional 5mm of graftinsertion (12.3mm, assuming all else equal) compared toits 15mm loop counterpart.

This study had some limitations. Only three manu-facturers were represented; however, the continuousloop devices chosen were consistent with those of otherstudies and among the most popular.1–3,8,9 Additionally,loop length was measured to the end of the suture loop;however, other measurement references (e.g., to themiddle or top of the doubled-over graft) could haveresulted in shorter loop lengths and larger resultantgraft insertion distances. Nevertheless, in measuringloop length (and corresponding graft insertion distance)to the end of the suture loop, the results of this studyare conservative and not subject to unconstrainedvariables (e.g., graft size, graft spreading and thinningon the loop, etc.). Furthermore, differences in measure-ment references may be the underlying and previouslyunderappreciated inconsistency among manufacturers.

The results of this study demonstrate inconsistencieswith regards to manufacturers’ reported loop lengthsfor continuous loop cortical suspension devices. Asnoted above, these inconsistencies could have potential-ly serious clinical implications. Consequently, we rec-ommend the manufacturers’ accurate and completedisclosure of the dimensions and specifications associat-ed with each continuous loop device. Furthermore,surgeon awareness of true loop length dimensions andinconsistencies among devices is necessary to ensureoptimal implantation and resultant clinical outcomes.

AUTHORS’ CONTRIBUTIONSAll authors have made substantial contributions to allof the following: (1) the conception and design of thestudy, or acquisition of data, or analysis and interpre-tation of data, (2) drafting the article or revising itcritically for important intellectual content, (3) finalapproval of the version to be submitted. Each of theauthors has read and concurs with the content in themanuscript.

ACKNOWLEDGMENTSThe authors thank Grant Dornan, MSc for his assistancewith statistical analysis and Angelica Wedell for her assis-tance with medical photography. We have no direct conflictof interest surrounding this work. All devices were donatedgratis, and the study was funded internally by the SteadmanPhilippon Research Institute. RFL is a consultant forArthrex, €Ossur, and Smith & Nephew and receives royaltiesfrom Arthrex and Smith & Nephew. CML has a relationshipwith a family member that is a paid consultant for Arthrex,€Ossur, and Smith & Nephew. CAW is currently an employeeof Arthrex; however, from the time of study inception tocompletion, CAW was an employee of the Steadman Philip-pon Research Institute. All other authors and acknowledgedindividuals have no conflicts of interest.

REFERENCES1. Petre BM, Smith SD, Jansson KS, et al. 2013. Femoral

cortical suspension devices for soft tissue anterior cruciateligament reconstruction: a comparative biomechanical study.Am J Sports Med 41:416–422.

2. Johnson JS, Smith SD, LaPrade CM, et al. 2015. Abiomechanical comparison of femoral cortical suspension

1330 TURNBULL ET AL.

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2015

devices for soft tissue anterior cruciate ligament reconstruc-tion under high loads. Am J Sports Med 43:154–160.

3. Brand J, Weiler A, Caborn DN, et al. 2000. Graft fixation incruciate ligament reconstruction. Am J Sports Med 28:761–774.

4. Beynnon BD, Johnson RJ, Abate JA, et al. 2005. Treatmentof anterior cruciate ligament injuries, part 2. Am J SportsMed 33:1751–1767.

5. Harner CD, Honkamp NJ, Ranawat AS. 2008. Anteromedialportal technique for creating the anterior cruciate ligamentfemoral tunnel. Arthroscopy 24:113–115.

6. Weber AE, Delos D, Oltean HN, et al. 2015. Tibial andfemoral tunnel changes after ACL reconstruction: a prospec-tive 2-year longitudinal MRI study. Am J Sports Med. http://dx.doi.org/10.1177/0363546515570461

7. Golish SR, Baumfeld JA, Schoderbek RJ, et al. 2007. Theeffect of femoral tunnel starting position on tunnel length inanterior cruciate ligament reconstruction: a cadaveric study.Arthroscopy 23:1187–1192.

8. Takeda Y, Iwame T, Takasago T, et al. 2013. Comparison oftunnel orientation between transtibial and anteromedialportal techniques for anatomic double-bundle anterior cruci-ate ligament reconstruction using 3-dimensional computedtomography. Arthroscopy 29:195–204.

9. Zantop T, Ferretti M, Bell KM, et al. 2008. Effect of tunnel-graft length on the biomechanics of anterior cruciate liga-ment-reconstructed knees: intra-articular study in a goatmodel. Am J Sports Med 36:2158–2166.

10. Yuan F, Zhou W, Cai J, et al. 2013. Optimal graft length foranterior cruciate ligament reconstruction: a biomechanicalstudy in beagles. Orthopedics 36:e588–e592.

11. J€arvel€a T, Moisala A-S, Paakkala T, et al. 2008. Tunnelenlargement after double-bundle anterior cruciate ligamentreconstruction: a prospective, randomized study. Arthrosco-py 24:1349–1357.

12. Jaglowski JR, Williams BT, Turnbull TL, et al. 2014. High-load preconditioning of soft tissue grafts: an in vitrobiomechanical bovine tendon model. Knee Surg SportsTraumatol Arthrosc. http://dx.doi.org/10.1007/s00167-014-3410-x

13. Tompkins M, Milewski MD, Carson EW, et al. 2013. Femo-ral tunnel length in primary anterior cruciate ligamentreconstruction using an accessory medial portal. Arthroscopy29:238–243.

14. Iyyampillai G, Raman ET, Rajan DV, et al. 2013. Determi-nants of femoral tunnel length in anterior cruciate ligamentreconstruction: CT analysis of the influence of tunnelorientation on the length. Knee Surg Relat Res 25:207–214.

LENGTH DIFFERENCES OF CONTINUOUS LOOP DEVICES 1331

JOURNAL OF ORTHOPAEDIC RESEARCH SEPTEMBER 2015