HAND/PERIPHERAL NERVE

A Knotless Flexor Tendon Repair TechniqueUsing a Bidirectional Barbed Suture: An ExVivo Comparison of Three MethodsW. Thomas McClellan, M.D.

Matthew J. Schessler, M.D.David S. Ruch, M.D.L. Scott Levin, M.D.

Richard D. Goldner, M.D.Morgantown, W.Va.; Durham, N.C.;

Philadelphia, Pa.

Background: Flexor tendon repairs using conventional suture require knotsthat enlarge the cross-sectional area at the repair site. This enlargement in-creases the force of finger flexion and jeopardizes the integrity of a nascenttendon repair during rehabilitation. The authors hypothesized that a knotlessflexor tendon repair using bidirectional barbed suture has similar strength andwith reduced cross-sectional area compared with traditional techniques.Methods: Sixty-six fresh porcine flexor digitorum profundus tendons were di-vided randomly into three groups. Tendons were transected and repaired withone of the following techniques: two-strand Kessler technique, four-strand Sav-age technique, or four-strand knotless technique. The cross-sectional area ofeach tendon was calculated at the repair site before and after repair. All tendonsunderwent mechanical testing to assess the 2-mm-gap formation force andultimate strength.Results: The 2-mm-gap formation force and ultimate strength of the Savage andknotless technique groups were not significantly different; however, both weresignificantly greater than those of the Kessler repair group (p ! 0.05). Therepair-site cross-sectional area of tendons repaired with the knotless techniquewas significantly smaller than that of tendons repaired with the Kessler or Savagetechnique (p ! 0.01). Tendons repaired with the knotless technique also hada significantly smaller change in repair-site cross-sectional area (p ! 0.01).Conclusions: The authors demonstrate that knotless flexor tendon repair withbarbed suture has equivalent strength and reduced repair-site cross-sectionalarea compared with traditional techniques. The smaller tendon profile maydecrease gliding resistance, thus reducing the risk for postsurgical tendonrupture during rehabilitation. (Plast. Reconstr. Surg. 128: 322e, 2011.)

Flexor tendon lacerations are devastating in-juries that hand surgeons all-too-commonlyencounter. These injuries require surgical

repair to restore the patient’s function. Despite aplethora of research, the basis of primary flexortendon repair has changed little since the 1970s.

Tendon repairs have traditionally been per-formed with permanent suture that requires knotseither inside or outside the tendon. Knots are theweak point of the tendon repair,1,2 cause de-creased tendon apposition,3 and are operator de-pendent. However, several studies show a positivecorrelation between suture caliber and number ofknots with strength of repair.4–6

Early active motion rehabilitation protocols re-quire strong tendon repairs. These protocols havegreatly improved patients’ function after surgical re-pair of flexor tendon lacerations by decreasing ad-hesions and increasing the repair’s strength.5,7 Sev-

From the Department of Surgery, Section of Plastic Surgery,West Virginia University School of Medicine, the Departmentof Orthopaedic Surgery, Duke University School of Medicine,and the Department of Orthopaedic Surgery, Hospital of theUniversity of Pennsylvania.Received for publication August 6, 2010; accepted March 8,2011.Presented at Plastic Surgery 2010: 79th Annual Meeting ofthe American Society of Plastic Surgeons, in Toronto, On-tario, Canada, October 1 through 5, 2010.Copyright ©2011 by the American Society of Plastic SurgeonsDOI: 10.1097/PRS.0b013e3182268c1f

Disclosure: The authors have no financial interestin the products or surgical techniques employed inthis study.

www.PRSJournal.com322e

eral studies indicate that friction created bytendons gliding against pulleys during active fingerflexion increases load.5,8,9 In addition, increased su-ture caliber and number of knots increases tendoncross-sectional area, causing increased glidingresistance.4,10 This increased load endangers the nas-cent repair during active rehabilitation. Clearly, onemust balance repair strength with the increased ten-don cross-sectional area within the tendon sheath.We hypothesized that a four-strand knotless tendonrepair using a bidirectional barbed suture has com-parable strength and reduced repair-site cross-sec-tional area when compared with traditional flexortendon repairs.

MATERIALS AND METHODSSixty-six fresh flexor digitorum profundus ten-

dons were obtained from adult pigs. These tendonshave been used frequently in prior studies becausethey are similar in structure and strength to a humanmiddle finger flexor tendon.11–15 The tendons wereexamined for abnormalities such as synovitis anddegeneration, and were rejected if an abnormalitywas present.

Cross-Sectional Area MeasurementsEach tendon’s height and width were measured

at the repair site and 1 cm proximal and distal to therepair site using a Brown & Sharpe IP67 digital cal-iper (part no. 00530300; Hexagon Metrology, Inc.,North Kingstown, R.I.). The measurements were an-alyzed to ensure all tendons were a similar size. Thecross-sectional area at each site was calculated usingthe formula for area of an ellipse (area " !ab, wherea equals one-half tendon height and b equals one-half tendon width). Measurements were taken at allthree sites before tendon transection and after re-pair to determine prerepair and postrepair cross-sectional area. Blinded intrarater analysis was con-ducted both before transection and after repair toensure measurement consistency.

RepairThe tendons were divided randomly into three

repair groups (A, B, and C), transected, and thenrepaired as described in Table 1.

All knots in groups A and B received six throwsto maximize effectiveness.1,16 The modified Kesslertechnique was chosen to represent a two-strandgrasping technique and the modified Savage tech-nique was chosen to represent a four-strand lockingtechnique. One surgeon (W.T.M.) performed allrepairs under 3.5# loupe magnification. The mod-ified Kessler and Savage repairs were performed with

nonabsorbable, 3-0 braided polyester suture (Ethi-bond Excel; Ethicon, Inc., Somerville, N.J.), whereasthe knotless repair was performed using barbed,nonabsorbable, 0-diameter monofilament polypro-pylene suture (Quill SRS; Angiotech, Inc., Vancou-ver, British Columbia, Canada). The 0-diameter su-ture was selected because of its similar strength to 3-0polyester.17,18 A core suture purchase of 1 cm wasused on all repairs.11,14

Knotless Repair TechniqueAlthough a novel repair method, the knotless

technique incorporates elements of both the mod-ified Kessler and the Savage methods. The knotlessmethod uses four strands and a locking grasp of theepitenon. A diagram of our technique is shown inFigure 1. The barbed repair first incorporates astraight pass through the tendon until the barbscatch the opposite side of the tendon. Then, thedouble-armed suture is passed back through the cen-tral core of the tendon to the transection site. Thecentral core limbs of the barbed repair are thenpassed diagonally across the tendon twice and anexternal bite of the epitenon is performed. The di-agonal passes serve two purposes. First, they increasethe number of barbs within the tendon substance.Second, the diagonal passes allow for multiple graspsof the epitenon. Finally, the external bite locks thesuture onto itself and then a mirror stitch is applied tothe initial tendon. A running epitendinous suture wasnotperformedsothatonly thecoresuturestrengthwasanalyzed.

Biomechanical TestingAfter repair and surface area measurement,

each tendon was secured into the clamps of atensiometer (model 4411; Instron Corp., Canton,Mass.) with a load cell of 500 N. The clamps havea broad surface that prevented tendon slippageduring testing. The upper clamp had a preload of1.5 N and was advanced at a rate of 20 mm/minuteThe preload and rate were selected because theybest simulate forces acting on an immobilized ten-

Table 1. Brief Description of Tendon Repair Methods

Group

Group A B C

No. oftendons 22 22 22

Repairtechnique

ModifiedKessler

ModifiedSavage Knotless

No. ofstrands 2 4 4

Suture 3-0 Ethibond 3-0 Ethibond 0 barbed

Volume 128, Number 4 • Knotless Flexor Tendon Repair

323e

don during active flexion.7,11,12,19,20 The linear dis-traction was monitored with a video camera andthe digital caliper (previously noted) was placednear the repaired tendon. The force and tendondisplacement were recorded by Instron Series 9software. The force that produced a 2-mm gapbetween tendon halves at the repair site was re-corded as the 2-mm-gap formation force. Lineardistraction continued until the sutures were pulledout or ruptured. In all cases, the greatest force oc-curring immediately before repair failure was re-corded as the ultimate strength. The mode of repairfailure was reported as pullout or rupture. An ob-server blinded to the tendon repair technique per-formed all mechanical strength testing.

RESULTSPower analysis was performed to ensure a large

enough sample size. For 0.80 power, seven tendonswere needed in each group. The 2-mm-gap forma-tion force, ultimate strength, and cross-sectionalarea data were analyzed with one-way analysis of vari-ance. A log transformation of the 2-mm-gap forceand ultimate strength data were taken before anal-ysis of variance. Intergroup reliability was checked.Values of p ! 0.05 were considered significant.

The 2-mm-gap formation force results, ulti-mate strength results, and mode of failure arelisted in Table 2, and changes in tendon dimen-sions are listed in Table 3. All values are reportedas mean $ SD. Results are depicted graphically inFigures 2 and 3. Mode of failure is reported aseither suture rupture or suture pullout. Rupturefailure means the strands or knots broke. Pulloutfailure means that the strands tore from the ten-don without breaking.

2-mm-Gap Formation ForceForces necessary to produce a 2-mm gap at the

repair site are reported in Figure 2 and Table 2.Tendons repaired by the knotless method pos-sessed a 2-mm-gap formation force of 62.84 N. The

2-mm-gap formation force for tendons repaired bythe Savage method was 59.22 N. Tendons repairedwith the modified Kessler method required 23.45N to form a 2-mm gap. The knotless and Savagemethods demonstrated a significantly greater2-mm-gap formation force than the Kesslermethod (p ! 0.05). However, no significant dif-ference in 2-mm-gap formation force existed be-tween the knotless technique and the Savagemethod.

Ultimate StrengthThe force causing ultimate failure is reported

in Figure 2 and Table 2. Tendons repaired by theknotless method withstood 72.39 N before failing.The ultimate failure force for tendons repaired bythe Savage method was 69.18 N. Tendons repairedby the modified Kessler method ultimately failed

Fig. 1. Diagram of the four-strand knotless flexor tendon repair technique.

Table 2. Data from Mechanical Strength Testing ofTendon Repairs Including 2-mm-Gap FormationForce, Ultimate Strength, and Mode of Failure

Tensile Strength (N) Failure Mode

RepairMethod

2-mm-GapFormation Ultimate Rupture Pullout

Knotless 62.84 $ 17.30 72.39 $ 15.16 18 4Savage 59.22 $ 15.12 69.18 $ 8.96 22 0Kessler 23.45 $ 5.32 32.03 $ 5.36 17 5

Table 3. Comparison of PostrepairCross-Sectional Area*

Repair SiteCross-Sectional Area (mm2)

RepairTechnique

AbsoluteSize

Change(vs. native tendon)

Knotless 24.4 %7.10 $ 4.58Savage 31.9 %13.6 $ 3.35Kessler 32.3 %14.3 $ 5.55*The knotless technique had a significantly smaller tendon size andchange in cross-sectional area compared with the Kessler and Savagetechniques.

Plastic and Reconstructive Surgery • October 2011

324e

at 32.03 N. The knotless and Savage methods dem-onstrated a significantly greater ultimate strengththan the Kessler method (p ! 0.05). However, nosignificant difference in ultimate strength was ob-served between the knotless technique and theSavage method.

Cross-Sectional AreaThe postrepair cross-sectional area for each

tendon repair technique is reported in Figure 3.The change in tendon size with the knotless tech-nique was significantly less than with the Savageand Kessler techniques (p ! 0.01). No significantdifference was observed in cross-sectional areaproximal or distal to the repair site among thetechniques (Fig. 4).

DISCUSSIONMcKenzie first reported using a unidirectional

barbed steel wire to repair flexor tendons in1967.21 His repair showed theoretical advantagescompared with traditional repair techniques, butthe use of a barbed suture repair was lost to theliterature until recently.22,23 Current barbed suturetechnology has advanced radically. Barbed suturesare bidirectional, with barbs spiraling around thecentral core suture. Barbed suture can now becreated using absorbable and nonabsorbable ma-terials, unlike the original steel wire description.Using these types of materials is advantageous fortendon repair.

According to Strickland, the ideal character-istics of a primary flexor tendon repair includesecure and easily placed sutures to allow for earlypostsurgical mobilization, smooth apposition ofthe tendon sections, minimum gapping forces,and minimal tendon vasculature disturbance.24,25

Trail et al. listed the ideal suture characteristics ashaving high tensile strength and being inexten-sible, absorbable, and, most importantly, easy touse.1,2 No current technique or suture meets all ofthe criteria of Strickland and Trail et al.

Traditional flexor tendon repair techniquesrely on knots either within the tendon or posi-tioned externally. When the knots lie within thetendon, they may also impede the tendon’s ulti-mate healing potential because of interposition ofthe knot between the tendon halves.3,26 When us-ing a nonabsorbable suture, a permanent obstruc-tion is placed between the tendon ends. With a

Fig. 2. Comparison of tensile strength among tendon repair techniques. Average2-mm-gap formation force (yellow bars) and ultimate strength (blue bars) areshown for each tendon repair technique. Knotless and Savage repairs were signif-icantly stronger than the Kessler repairs (p ! 0.05). Knotless and Savage methodswere not significantly different in strength.

Fig. 3. Comparison of postsurgical cross-sectional area amongtendon repair techniques. Average cross-sectional area at the re-pair site is shown for each repair technique. The knotless methodhad a significantly smaller cross-sectional area than the Savage orKessler method.

Volume 128, Number 4 • Knotless Flexor Tendon Repair

325e

knotless repair, there is no knot interposition be-tween the tendon ends, so the potential exists forbetter healing and increased long-term strength.

Knots within the repair site do not affect therepair’s overall strength unless they are greaterthan 26 percent of the tendon’s cross-sectionalsurface area.3 Although larger suture calibers im-part greater strength to the tendon repair, they areharder to manipulate and create larger knots.6,26

Some authors have even recommended at leastfive throws to create a secure knot.1,16 In addition,increased foreign material within a repair site hasbeen shown to decrease wound healing by stim-ulating an inflammatory response.26 Althoughlarge knots impart greater repair strength, theyare less than ideal because of their deleteriouseffect on healing and increased tendon profile.

Bulky knots increase the tendon’s cross-sec-tional area, thus increasing gliding resistance dur-ing active flexion.4,10,27–29 Furthermore, this in-creased load at the repair site can cause gapformation and failure.7,9,30

Furthermore, knots have been shown to im-pede vasculature.29,31–33 This deprives the ten-don of vital nourishment necessary to heal, caus-ing extrinsic neovascularization and adhesionformation.34

One way to prevent the problems caused byknots is to completely eliminate them. We reporta four-strand knotless flexor tendon repairmethod using a bidirectional barbed suture. Ourresults show no significant difference in repairstrength between the four-strand knotless tech-nique and the criterion standard four-strand Sav-age technique. Most importantly, the four-strandknotless method creates a lower tendon profile atthe repair site. It is plausible that decreasing thetendon profile will decrease gliding resistance,thus reducing the risk of gapping and failurein vivo.

The knotless method does not decrease thedifficulty of flexor tendon repair or change itsindications. Further study of this four-strandedknotless technique should include cyclical loadingand angular tensile strength. Cyclical loading stud-ies would be important to determine the risk ofbarbed sutures slicing through the freshly re-paired tendon during early, active rehabilitation.Angular tensile strength studies would moreclosely resemble forces acting on the tendon dur-ing rehabilitation.

Another weakness of our study is the lack ofclinical data and outcomes from application inpatients. In vivo studies should certainly be per-formed to assess repair healing and its environ-mental interactions in an animal model. Success-ful application and outcomes in patients wouldallow the knotless four-stranded technique to beincorporated into clinical practice.

CONCLUSIONSThis report shows that our four-stranded knot-

less technique yields a repair as strong as a con-temporary four-stranded method but with asmaller cross-sectional area. Because our repair isas strong as current techniques and has a lowertendon profile, further ex vivo and in vivo studiesare warranted. Our knotless technique may im-prove outcomes in patients with zone II flexortendon lacerations by allowing for more aggressiverehabilitation with reduced risk of repair failure.

W. Thomas McClellan, M.D.Morgantown Plastic Surgery Associates

1085 Van Voorhis Road, Suite 350Morgantown, W.Va. 26505

wtmcclellan@yahoo.com

REFERENCES1. Trail IA, Powell ES, Noble J. An evaluation of suture materials

used in tendon surgery. J Hand Surg Br. 1989;14:422–427.

Fig. 4. An unrepaired tendon (below) and a tendon repaired with theknotless technique (above).

Plastic and Reconstructive Surgery • October 2011

326e

2. Trail IA, Powell ES, Noble J. The mechanical strength ofvarious suture techniques. J Hand Surg Br. 1992;17:89–91.

3. Pruitt DL, Aoki M, Manske PR. Effect of suture knot locationon tensile strength after flexor tendon repair. J Hand SurgAm. 1996;21:969–973.

4. Aoki M, Manske PR, Pruitt DL, Larson BJ. Work of flexionafter tendon repair with various suture methods: A humancadaveric study. J Hand Surg Br. 1995;20:310–313.

5. Barrie KA, Tomak SL, Cholewicki J, Merrell GA, Wolfe SW.Effect of suture locking and suture caliber on fatiguestrength of flexor tendon repairs. J Hand Surg Am. 2001;26:340–346.

6. Taras JS, Raphael JS, Marczyk SC, Bauerle WB. Evaluation ofsuture caliber in flexor tendon repair. J Hand Surg Am. 2001;26:1100–1104.

7. Tang JB, Cao Y, Xie RG. Effects of tension direction onstrength of tendon repair. J Hand Surg Am. 2001;26:1105–1110.

8. Coert JH, Uchiyama S, Amadio PC, Berglund LJ, An KN.Flexor tendon-pulley interaction after tendon repair: A bio-mechanical study. J Hand Surg Br. 1995;20:573–577.

9. Thurman RT, Trumble TE, Hanel DP, Tencer AF, Kiser PK.Two-, four-, and six-strand zone II flexor tendon repairs: Anin situ biomechanical comparison using a cadaver model.J Hand Surg Am. 1998;23:261–265.

10. Momose T, Amadio PC, Zhao C, Zobitz ME, An KN. Theeffect of knot location, suture material, and suture size on thegliding resistance of flexor tendons. J Biomed Mater Res. 2000;53:806–811.

11. Cao Y, Zhu B, Xie RG, Tang JB. Influence of core suturepurchase length on strength of four-strand tendon repairs.J Hand Surg Am. 2006;31:107–112.

12. Mishra V, Kuiper JH, Kelly CP. Influence of core suturematerial and peripheral repair technique on the strength ofKessler flexor tendon repair. J Hand Surg Br. 2003;28:357–362.

13. Smith AM, Evans DM. Biomechanical assessment of a newtype of flexor tendon repair. J Hand Surg Br. 2001;26:217–219.

14. Tang JB, Zhang Y, Cao Y, Xie RG. Core suture purchaseaffects strength of tendon repairs. J Hand Surg Am. 2005;30:1262–1266.

15. Zatiti SC, Mazzer N, Barbieri CH. Mechanical strengths oftendon repair: An in vitro comparative study of six tech-niques. J Hand Surg Br. 1998;23:228–233.

16. Viinikainen A, Goransson H, Huovinen K, Kellomaki M,Tormala P, Rokkanen P. Material and knot properties ofbraided polyester (Ticron) and bioabsorbable poly-L/D-lac-tide (PLDLA) 96/4 sutures. J Mater Sci Mater Med. 2006;17:169–177.

17. Leung JC, Ruff GL, Megaro MA. Barbed bi-directional med-ical sutures: Biomechanical properties and wound closureefficacy study. Society for Biomaterials 28th Annual MeetingTransactions. No. 724. Mt. Laurel, NJ: Society for Biomateri-als; 2002.

18. Villa MT, White LE, Alam M, Yoo SS, Walton RL. Barbedsutures: A review of the literature. Plast Reconstr Surg. 2008;121:102e–108e.

19. Xie RG, Zhang S, Tang JB, Chen F. Biomechanical studies of3 different 6-strand flexor tendon repair techniques. J HandSurg Am. 2002;27:621–627.

20. Xie RG, Tang JB. Investigation of locking configurations fortendon repair. J Hand Surg Am. 2005;30:461–465.

21. McKenzie AR. An experimental multiple barbed suture forthe long flexor tendons of the palm and fingers: Preliminaryreport. J Bone Joint Surg Br. 1967;49:440–447.

22. Parikh PM, Davison SP, Higgins JP. Barbed suture tenor-rhaphy: An ex vivo biomechanical analysis. Plast Reconstr Surg.2009;124:1551–1558.

23. Trocchia AM, Aho HN, Sobol G. A re-exploration of the useof barbed sutures in flexor tendon repairs. Orthopedics 2009;32:731.

24. Strickland JW. Flexor tendon injuries: I. Foundations of treat-ment. J Am Acad Orthop Surg. 1995;3:44–54.

25. Strickland JW. Development of flexor tendon surgery: Twenty-five years of progress. J Hand Surg Am. 2000;25:214–235.

26. van Rijssel EJ, Brand R, Admiraal C, Smit I, Trimbos JB.Tissue reaction and surgical knots: The effect of suture size,knot configuration, and knot volume. Obstet Gynecol. 1989;74:64–68.

27. Momose T, Amadio PC, Zhao C, Zobitz ME, Couvreur PJ, AnKN. Suture techniques with high breaking strength and lowgliding resistance: Experiments in the dog flexor digitorumprofundus tendon. Acta Orthop Scand. 2001;72:635–641.

28. Strick MJ, Filan SL, Hile M, McKenzie C, Walsh WR, TonkinMA. Adhesion formation after flexor tendon repair: A his-tologic and biomechanical comparison of 2- and 4-strandrepairs in a chicken model. J Hand Surg Am. 2004;29:15–21.

29. Zhao C, Amadio PC, Momose T, Couvreur P, Zobitz ME, AnKN. The effect of suture technique on adhesion formationafter flexor tendon repair for partial lacerations in a caninemodel. J Trauma 2001;51:917–921.

30. Barrie KA, Wolfe SW, Shean C, Shenbagamurthi D, Slade JFIII, Panjabi MM. A biomechanical comparison of multistrandflexor tendon repairs using an in situ testing model. J HandSurg Am. 2000;25:499–506.

31. Datillo PP, King MW, Cassill NL, Leung JC. Medical textiles:Application of an absorbable barbed bi-directional surgicalsuture. J Text Apparel Technol Manage. 2002;2:1–5.

32. Manske PR, Gelberman RH, Vande Berg JS, Lesker PA. In-trinsic flexor-tendon repair: A morphological study in vitro.J Bone Joint Surg Am. 1984;66:385–396.

33. Amadio PC, Hunter JM, Jaeger SH, Wehbe MA, SchneiderLH. The effect of vincular injury on the results of flexortendon surgery in zone 2. J Hand Surg Am. 1985;10:626–632.

34. Seiler JG III, Chu CR, Amiel D, Woo SL, Gelberman RH. TheMarshall R. Urist Young Investigator Award. Autogenousflexor tendon grafts: Biologic mechanisms for incorporation.Clin Orthop Relat Res. 1997;345:239–247.

Volume 128, Number 4 • Knotless Flexor Tendon Repair

327e