Remplissage Using Interconnected Knotless Anchors:Superior Biomechanical Properties to a Knotted

Technique?Tadanao Funakoshi, M.D. Ph.D., Robert Hartzler, M.D., M.S., Eduardo Stewien, M.D., and

Stephen Burkhart, M.D.

Purpose: To evaluate the biomechanical fixation strength and gap formation of 2 different remplissage fixation methods(double pulley knotted construct and interconnected knotless repair construct) in cadaver specimens. Methods: Sevenmatched pairs of human cadaveric shoulders were used for testing (mean age, 56 ! 10 years). A shoulder from each matchedpair was randomly selected to receive a Hill-Sachs remplissage using either a knotted (No. 2 FiberWire double pulley with 3.0-mm SutureTak anchors) or knotless (coreless No. 2 FiberWire interconnected between 3.9-mm knotless CorkScrew anchors)double mattress construct. The tendon was cycled between 10 and 100 N at 1 Hz for 100 cycles, followed by a single-cycle pullto failure at 33 mm/s. Cyclic displacement, load to clinical failure (5 mm), yield load, and mode of failure were recorded.Results: Neither construct demonstrated clinical failure under cyclic loading. Load to clinical failure was higher for the knotlessrepair than that of the knotted repair (788! 162 N vs 488! 227 N; P¼ .003). The yield load was higher for the knotless repairthan that of the knotted repair (1,080 ! 298 N vs 591 ! 265 N; P ¼ .008). The most common failure mode for the knottedrepair was knot failure or tendon tearing, whereas the failure mode for the knotless repair was by anchor pull-out or tendontear with no failures occurring via the interconnected suture construct mechanism. Conclusions: In this biomechanical studycomparing cyclic and ultimate loading for 2 double mattress remplissage repairs, the construct using interconnected, knotlesssutures outperformed the knotted construct. No failure of the interconnected suture construct mechanism by slippage orbreakage was observed in the knotless group. Clinical Relevance: The use of the interconnected knotless suture techniquemight improve the biomechanical strength of arthroscopic remplissage repairs in treating shoulder instability.

See commentary on page 2960

The procedure of arthroscopic remplissage has beenused successfully to improve shoulder stability

in cases where a Hill-Sachs lesion of the humerus

engages1 the glenoid or is biomechanically at risk forengagement (“off-track” lesion).2 Insetting the posteriorrotator cuff into a significant Hill-Sachs lesion representsa nonanatomic reconstruction of the shoulder, but issupported by both clinical3 and biomechanical4 studiesthat demonstrate the effectiveness of this procedure.Successful outcomes of arthroscopic procedures such

as remplissage require secure, primary fixation at thesite of tendon-to-bone interface to allow biologichealing to occur. Although knot tying is an essentialskill for proficiency in arthroscopic surgery,5-9 knotsremain a potential site of failure6,9,10 of knotted repairconstructs.11 Evidence has shown that considerablevariation exists in knot tying consistency and knotstrength among surgeons.6

Knotless suture anchors of varying design havebeen an important advance in arthroscopic shouldersurgery. Knotless anchors have the theoretical advan-tages of decreased operative time, simplicity of suturemanagement, and the elimination of knots as a source

From the University of Department of Orthopaedic Surgery (T.F.), Hok-kaido University School of Medicine, Sapporo, Japan; The San Antonio Or-thopaedic Group and Burkhart Research Institute for Orthopaedics (R.H.,S.B.), San Antonio, Texas, U.S.A.; and Hospital Adventista de Manaus (E.S.),Manaus, Brazil.

The authors report the following potential conflict of interest or source offunding: Study funding was provided by Arthrex. R. H. receives supportfrom Arthrex and receives textbook royalties from Wolters Kluwer HealtheLippincott Williams & Wilkins. S. B. receives support from Arthrex, has apatent pending, has a patent, and receives textbook royalties from WoltersKluwer HealtheLippincott Williams & Wilkins. Full ICMJE author disclo-sure forms are available for this article online, as supplementary material.

Received November 21, 2017; accepted June 10, 2018.Address correspondence to Robert Hartzler, M.D., M.S., TSAOG Ortho-

paedics and Burkhart Research Institute for Orthopaedics, 2829 Babcock Rd,Ste 700, San Antonio, TX 78229, U.S.A. E-mail: rhartzler@tsaog.com

! 2018 by the Arthroscopy Association of North America0749-8063/171381/$36.00https://doi.org/10.1016/j.arthro.2018.06.030

2954 Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 34, No 11 (November), 2018: pp 2954-2959

of cartilage wear or postoperative pain.12 Additionally,one knotless anchor design (Knotless SutureTak andKnotless CorkScrew; Arthrex, Naples, FL) has an in-ternal splice locking mechanism in the anchor body(Fig 1) that has the purported advantages of beingtensionable and self-cinching (self-reinforcing).13

The purpose of this study was to evaluate thebiomechanical fixation strength and gap formation of 2different remplissage fixation methods (double pulleyknotted construct11 and interconnected knotless repairconstruct14) in cadaver specimens. Our hypothesis isthat, in a cyclic loading and load-to-failure biome-chanical study, the interconnected knotless constructwill be equivalent to the knotted construct comparingload at 5 mm of displacement and yield load.

MethodsLocal institutional review board approval was

granted for this biomechanical study. Seven matchedpairs of fresh frozen cadaveric shoulders (all malespecimens) with a mean age of 56 ! 10 yearsunderwent simulated knotted and knotless Hill-Sachsremplissage followed by biomechanical testing. Thesample size was determined arbitrarily, but was inkeeping with similar previous studies and the reason-able constraints of time, availability, and cost of spec-imens. The specimens were thawed overnight beforepreparation. Each shoulder was dissected such thatonly the humerus with the rotator cuff tendons andcapsule remained. The distal humeral shaft was pottedin PVC pipe using fiberglass epoxy resin. Specimenswere prepared in a pairwise manner with care toensure even allocation of repairs between left andright specimens. Specimen preparation, testing, andassessment was done by the authors, who all arefellowship trained in shoulder surgery and are expe-rienced in conducting biomechanical studies.

Simulated Hill-Sachs RemplissageBecause the specimens were intact with regard to the

humeral head, the remplissage was simulated by fixingthe rotator cuff and capsule onto the intact cartilage.Creation of an experimental Hill-Sachs lesion was

avoided to maintain the integrity of the subchondralbone with the goal of minimizing anchor failure suchthat the mechanical properties of the suture constructs(rather than anchor properties) could be compared. Thelocation of the remplissage was centered over theinfraspinatus tendon by calculating half the distancebetween the suprainfraspinatus raphe and the superioraspect of the teres minor insertion (Fig 2). The remplis-sage was thus centered at approximately the 2 o’clock(right shoulder) or 10 o’clock (left shoulder) position4

(Fig 2). The 2 suture anchors were placed 15 mmapart and 15 mm onto the humeral head articularcartilage measured from the infraspinatus capsularinsertion (Fig 2), thus, simulating a small to moderatesized Hill-Sachs lesion by width.4 The anchors wereplaced and retrieved through the cuff for all specimens.

Knotted GroupIn the knotted group, the PEEK (polyether ether

ketone) 3.0-mm SutureTak anchors were placed asabove leaving 1 strand of No. 2 FiberWire (Arthrex) ineach anchor. The sutures were tied as a double-pulleymattress11 after being retrieved through the rotatorcuff just lateral to the musculotendinous junction usinga Penetrator suture passer (Arthrex; Fig 1B) with 1 passfor each suture. The first knot of the double pulley wastied by hand using a 6 throw surgeon’s knot. Thesecond knot was tied using a Surgeon’s 6th Finger knotpusher (Arthrex)11 to simulate an arthroscopic repair(Fig 3A).

Knotless GroupIn the knotless group, the 2 PEEK 3.9-mm Knotless

CorkScrew anchors were placed as above with respectto location on the humeral head. The sutures werepassed through the infraspinatus tendon in the samemanner as in the knotted group. The remplissageconstruct in the knotless group was created by thread-ing the coreless No. 2 FiberWire repair suture from eachanchor into the suture splice locking mechanism13 ofthe other anchor. The threaded repair sutures werethen tensioned sequentially such that 2 suture limbscompressed tendon to bone (Fig 3B).

Fig 1. A novel knotless sutureanchor (A) with coreless No. 2FiberWire repair and shuttlingsutures. (B) An internal splicelocking mechanism (bracket)in the anchor body makes theanchor tensionable and self-cinching as the repair sutureis pulled tight (arrow).

INTERCONNECTED KNOTLESS ANCHORS 2955

Biomechanical TestingThe potted humeral shaft was positioned in a custom-

made fixture that was mounted vertically on the tableof a servohydraulic materials testing machine (MTSSystems, Minneapolis, MN). Using an adjustable anglefixture, the humerus specimen was fixed at an angleand position of rotation such that the line of force onthe infraspinatus tendon simulated that which wouldoccur in the apprehension position of the shoulder (60#

of abduction in the coronal plane and approximately

30# posterior to the scapular plane) but midrangerotation (60#).4 The infraspinatus musculotendinousunit (Fig 4) was then fixed in the cross-head of themachine using a cryo clamp, and visual verification wasdone to ensure that the vector of pull was normal to therepair construct (Fig 4).The tendon was cycled between 10 and 100 N at 1 Hz

for 100 cycles, followed by a single-cycle pull to failureat 33 mm/s. Forceedisplacement curves were gener-ated for each specimen. Clinical failure was defined as

Fig 2. Left shoulder specimenduring preparation. Two sutureanchors (PEEK SutureTak,3.0 mm) were inserted 15 mmapart (A) centered at aboutthe 10 o’clock position of thehumerus (left shoulder). Thesutures were passed just lateralto the musculotendinous junc-tion using a suture retriever (B).The remplissage was centeredover the infraspinatus tendon asdeterminedby thebordersof thesuprainfraspinatus raphe andthe characteristic upper bordermuscle fibers of the teres minor.

Fig 3. Left shoulder specimenduring preparation. In bothknotted (A) and knotless (B)groups, a double mattresssuture construct was used tocreate a simulated remplissageconstruct. In the knottedgroup, No. 2 FiberWire, anultraehigh-molecular-weightpolyethylene (UHMWPE) corestrand jacketed with braidedpolyester-UHMWPE, wassecured with 2 surgeon’s knots(6 throws). In the knotlessgroup, the suture was No. 2FiberWire coreless (core strandremoved).

2956 T. FUNAKOSHI ET AL.

5 mm of displacement. For each specimen, load atclinical failure7,15,16 and yield load (yield point definedas sudden drop in tensile forces signified by change inslope of forceedisplacement curve) were recorded.A digital video camera (Handycam, SONY, Tokyo,

Japan) recorded the testing procedure to document themode of failure at the yield point. The mode of failurewas reported for each specimen.

Statistical AnalysisResults are reported as mean ! standard deviation.

Statistical analysis for yield load and load at 5 mm ofdisplacement was done with (SigmaPlot 11 software;Systat Software, San Jose, CA). A paired t test was usedto for statistical comparison between the 2 groups.Statistical significance was set at P < .05. No a prioripower analysis was performed because the sample sizewas constrained by available resources, as stated.

ResultsNo clinical failure (more than 5 mm of displacement)

was observed in any specimen after the first 100 cycles.The mean loading force to create a 5-mm displacementin the knotless group (788 ! 162 N) was significantlyhigher than that in the knotted group (488 ! 227 N;P ¼ .003). The load to failure (yield load) for theknotless group (1,080 ! 298 N) was higher than thatfor the knotted group (591 ! 265 N; P ¼ .008).Displacement at the yield point in the knotless group(6.6 ! 1.3 mm) was not significantly different from thatin the knotted group (6.0 ! 1.7 mm) (Table 1).

In the knotted group, there were no failures involvingpullout of the anchor (Table 2). In the knotted group,failure occurred via the suture (breakage or knot loos-ening) or tendon tearing in all cases. In the knotlessgroup, failure occurred by pullout of the anchor in 4cases and tendon tear in 3 cases without any observedfailure of the suture construct.

DiscussionIn the current study, the biomechanical testing of

knotless versus knotted double mattress remplissagesuture constructs confirmed the hypothesis that thespecific knotless method using anchors with an internalsplice locking mechanism13 had at least equivalentstrength to the knotted method when tested undercyclic loading and load to failure. In fact, under load tofailure testing, the knotless construct had far betterstrength than the knotted construct.Previous studies have demonstrated mixed results

regarding the biomechanical superiority of repairs

Fig 4. Lateral (A) and medial(B) views of the biome-chanical testing setup (leftshoulder shown). Theadjustable-angle fixture waspositioned to simulate theline of pull of the infra-spinatus tendon against theconstruct with the shoulderin the apprehension positionat midrange rotation.

Table 1. Cyclic Loading and Load-to-Failure Testing ofKnotted and Knotless Remplissage Constructs

KnottedGroup

KnotlessGroup P Value

Load to Clinical Failure,* n 488 ! 227 788 ! 162 .003Yield Load, n 591 ! 265 1080 ! 298 .008Displacement at YieldLoad, mm

6.0 ! 1.7 6.6 ! 1.3 .5

*5 mm of displacement.

INTERCONNECTED KNOTLESS ANCHORS 2957

performed using knotless anchors.15-20 Burkhart andAthanasiou21 previously outlined and validated themechanisms for improved security of 1 knotless fixationmethod (twist-lock) over a knotted construct.Conversely, Uggen et al17 noted suture slippage arounda similar style of knotless anchor as the mode of failurein biomechanical testing of SLAP repairs in 5 of 6 cases(83%). In the current study, the specific knotless an-chor mechanism is a novel, self-reinforcing suturelocking system similar to a “Chinese finger trap,” inwhich the system grips more tightly as a tensile longi-tudinal force increases against the suture loop.Self-reinforcement is very desirable22 and this partic-ular characteristic of the knotless anchor design mayhave been responsible for the results seen in this study.In the current study, the mechanical integrity of the

suture fixation in the knotless group was excellent, asdemonstrated by (1) the superior load to failure and (2)the fact that the failure mode for the knotless group wasusually tendon tear or anchor pullout rather thansuture breakage or slippage. In addition to the self-reinforcing nature of the knotless anchors (increasedfrictional forces against suture slippage as the load in-creases tightens22), for the knotless construct to fail bysuture slippage or breakage, each suture would have tofail because they are each fixed to the opposite anchorindependently. In the knotted construct, only 1 knothas to slip or break to result in failure of both suturesbecause they are tied together. Furthermore, this typeof biomechanical experiment provides the best casescenario for knot integrity because these knots weretied under close scrutiny under controlled laboratoryconditions.

LimitationsThis study does have limitations. The very specific

design mechanism of the knotless anchor (internalsuture splice mechanism in the anchor body) is uniqueto the manufacturer (Arthrex), and these results shouldnot be assumed to extrapolate to other anchors of thesame general kind (“knotless”). Some variables besidesthe suture fixation method (knots vs knotless), werenot able to be held constant between groups, primarilydue to anchor design. In the knotted group, a 3.0-mm“push-in” anchor was used, whereas the knotlessgroup had a 3.9-mm threaded anchor. Additionally, thesuture itself was different (No. 2 FiberWire vs corelessNo. 2 FiberWire), also owing to differences in anchor

design. This limitation of anchor differences did notseem to be experimentally relevant, because the smalleranchor did not fail in any case by anchor pull-out.Additionally, the suture material differences did notseem to be experimentally relevant, because the core-less suture did not fail in any case by breakage orloosening. The bone density of the cadaver specimenswas unknown, which could have potentially affectedthe results had different modes of failure beenobserved. Last, this study has the weakness inherent ina biomechanical study performed on cadaveric speci-mens with the variability in tissue quality betweenspecimens.

ConclusionsIn this biomechanical study comparing cyclic and

ultimate loading for 2 double mattress remplissagerepairs, the construct using interconnected, knotlesssutures outperformed the knotted construct. No failureof the interconnected suture construct mechanism byslippage or breakage was observed in the knotlessgroup.

AcknowledgmentsThe authors express their appreciation to Josh

Karnes, M.S., for his assistance in setting up and oper-ating the testing apparatus.

References1. Burkhart SS, De Beer JF. Traumatic glenohumeral bone

defects and their relationship to failure of arthroscopicBankart repairs: Significance of the inverted-pear glenoidand the humeral engaging Hill-Sachs lesion. Arthroscopy2000;16:677-694.

2. Di Giacomo G, Itoi E, Burkhart SS. Evolving concept ofbipolar bone loss and the Hill-Sachs lesion: From"engaging/non-engaging" lesion to "on-track/off-track"lesion. Arthroscopy 2014;30:90-98.

3. Wolf EM, Arianjam A. Hill-Sachs remplissage, an arthro-scopic solution for the engaging Hill-Sachs lesion: 2- to10-year follow-up and incidence of recurrence. J ShoulderElbow Surg 2014;23:814-820.

4. Hartzler RU, Bui CN, Jeong WK, et al. Remplissage of anoff-track Hill-Sachs lesion is necessary to restore biome-chanical glenohumeral joint stability in a bipolar bone lossmodel. Arthroscopy 2016;32:2466-2476.

5. Lo IK, Burkhart SS, Chan KC, Athanasiou K. Arthroscopicknots: Determining the optimal balance of loop securityand knot security. Arthroscopy 2004;20:489-502.

6. Hanypsiak BT, DeLong JM, Simmons L, Lowe W,Burkhart S. Knot strength varies widely among expertarthroscopists. Am J Sports Med 2014;42:1978-1984.

7. Burkhart SS, Diaz Pagan JL, Wirth MA, Athanasiou KA.Cyclic loading of anchor-based rotator cuff repairs:Confirmation of the tension overload phenomenon andcomparison of suture anchor fixation with transosseousfixation. Arthroscopy 1997;13:720-724.

Table 2. Observed Modes of Failure During BiomechanicalTesting of Knotted and Knotless Remplissage Constructs

Knotted Group Knotless GroupTendon Tear 1 3Suture Breakage 2 0Suture Loosening 4 0Anchor Pull-out 0 4

2958 T. FUNAKOSHI ET AL.

8. Swan KG Jr, Baldini T, McCarty EC. Arthroscopic suturematerial and knot type: An updated biomechanical anal-ysis. Am J Sports Med 2009;37:1578-1585.

9. Pedowitz RA, Nicandri GT, Angelo RL, Ryu RK,Gallagher AG. Objective assessment of knot-tying profi-ciency with the fundamentals of arthroscopic surgerytraining program workstation and knot tester. Arthroscopy2015;31:1872-1879.

10. Kim SH, Crater RB, Hargens AR. Movement-induced knotmigration after anterior stabilization in the shoulder.Arthroscopy 2013;29:485-490.

11. Koo SS, Burkhart SS, Ochoa E. Arthroscopic double-pulley remplissage technique for engaging Hill-Sachslesions in anterior shoulder instability repairs. Arthros-copy 2009;25:1343-1348.

12. Park JG, Cho NS, Kim JY, Song JH, Hong SJ, Rhee YG.Arthroscopic knot removal for failed superior labrumanterior-posterior repair secondary to knot-induced pain.Am J Sports Med 2017;45:2563-2568.

13. Sullivan D. Tensionable knotless anchors with splice andmethods of tissue repair, https://patents.google.com/patent/US9107653. Published 2015. Accessed July 5, 2018.

14. Burkhart SS. The cowboy’s conundrum: Complex andadvanced cases in shoulder arthroscopy. Philadelphia: Lip-pincott Williams & Wilkins, 2017.

15. Barber FA, Herbert MA. Cyclic loading biomechanicalanalysis of the pullout strengths of rotator cuff andglenoid anchors: 2013 update. Arthroscopy 2013;29:832-844.

16. Barber FA, Drew OR. A biomechanical comparison oftendon-bone interface motion and cyclic loading betweensingle-row, triple-loaded cuff repairs and double-row,suture-tape cuff repairs using biocomposite anchors.Arthroscopy 2012;28:1197-1205.

17. Uggen C, Wei A, Glousman RE, et al. Biomechanicalcomparison of knotless anchor repair versus simple suturerepair for type II SLAP lesions. Arthroscopy 2009;25:1085-1092.

18. Sileo MJ, Ruotolo CR, Nelson CO, Serra-Hsu F,Panchal AP. A biomechanical comparison of the modifiedMason-Allen stitch and massive cuff stitch in vitro.Arthroscopy 2007;23:235-240:240.e1-2.

19. Zheng N, Harris HW, Andrews JR. Failure analysis of ro-tator cuff repair: A comparison of three double-rowtechniques. J Bone Joint Surg Am 2008;90:1034-1042.

20. Funakoshi T, Suenaga N, Sano H, Oizumi N, Minami A.In vitro and finite element analysis of a novel rotator cufffixation technique. J Shoulder Elbow Surg 2008;17:986-992.

21. Burkhart SS, Athanasiou KA. The twist-lock concept oftissue transport and suture fixation without knots:Observations along the Hong Kong skyline. Arthroscopy2003;19:613-625.

22. Burkhart SS, Adams CR, Burkhart SS, Schoolfield JD.A biomechanical comparison of 2 techniques of footprintreconstruction for rotator cuff repair: The SwiveLock-FiberChain construct versus standard double-row repair.Arthroscopy 2009;25:274-281.

INTERCONNECTED KNOTLESS ANCHORS 2959