t

SCIENTIFIC ARTICLE

Epitendinous Suture Techniques in Extensor Tendon

Repairs—An Experimental Evaluation

James Henderson, MD, Michael Sutcliffe, PhD, Patrick Gillespie, MD

Purpose The tension-band principle might be relevant to extensor tendon repairs, and adorsal-only Silfverskiöld epitendinous repair is stronger and stiffer than more conventionaltechniques in vitro. We aimed to evaluate the strength and stiffness of the strongestepitendinous sutures described, using an in vitro model that subjects the repair to angularforce over a pulley, thereby creating a tension-band model.

Methods Silfverskiöld dorsal-only epitendinous extensor tendon repairs in porcine foottendons (n � 8) were compared to reverse (buried) Silfverskiöld (n � 8), Halsted (n � 8),and interrupted horizontal mattress (IHM) repairs (n � 6) in vitro with a tensiometer arounda 45° pulley. Thirty tendons total were tested to assess the force required for 2-mm gappingand ultimate tensile strength.

Results The IHM repair had a significantly higher ultimate tensile strength (43 N; SD, 10 N) thanthe other repairs, which had strengths between 27 N (SD, 4 N) and 31 N (SD, 7 N). The IHMwas also significantly more resistant to gapping than the Silfverskiöld and Halsted repairs.

Conclusions Interlocking horizontal mattress, dorsal-only extensor tendon repairs were sig-nificantly stronger and more resistant to gapping than Silfverskiöld and Halsted repairs.Other repairs were still strong and resistant to gapping in comparison to previously publisheddata for conventional repairs.

Clinical relevance The IHM is a relatively difficult technique to perform, and it remains to beseen whether the additional strength translates to clinical benefits over the easier Silfver-skiöld technique. (J Hand Surg 2011;36A:1968–1973. Copyright © 2011 by the AmericanSociety for Surgery of the Hand. All rights reserved.)

Key words Extensor tendon repairs, Halsted, Silfverskiöld, tension-band principle.

tctawt

mmbpl

THE TENSION-BAND PRINCIPLE describes that whena structure is submitted to a deforming force,one side of the structure is under compression

and the other under tension.1 We previously suggestedhat this might apply to extensor tendon repairs.2 The

dorsal surface of the repair can be under greater tension

From the Department of Plastic and Reconstructive Surgery, Norfolk and Norwich University HospitalNHSTrust,Norwich,UK;DepartmentofEngineering,UniversityofCambridge,Cambridge,UK;Depart-ment of Plastic and Reconstructive Surgery, Addenbrooke’s University NHS Trust, Cambridge, UK.

Received for publication February 22, 2011; accepted in revised form August 30, 2011.

No benefits in any form have been received or will be received related directly or indirectly to thesubject of this article.

Corresponding author: James Henderson, MD, Department of Plastic Surgery, Norfolk and Nor-wich University Hospital NHS Trust, Norwich, NR4 7UY, UK; e-mail: jh@jameshenderson.net.

0363-5023/11/36A12-0011$36.00/0

edoi:10.1016/j.jhsa.2011.08.038

1968 � © ASSH � Published by Elsevier, Inc. All rights reserved.

han the deep surface because the tendon runs over theonvexity of the joints. It might be beneficial for repairso offer maximum strength dorsally. This has beenlluded to in the study of flexor repairs,3 where repairsith a stronger dorsal component were superior when

ested around a 90° pulley.4,5 Extensor tendons repairedwith dorsal-only Silfverskiöld epitendinous sutures arestronger than conventional extensor repair techniqueswhen tested on a curved apparatus.2

A suture loop tied within a tendon causes tenocyteigration,6 so sparing part of the tendon from suturesight benefit healing, although this might be a minor

enefit in a thin extensor tendon. There is little workublished comparing tendon repairs subjected to angu-ar tension in vitro, and this had generally focused

ntirely on models designed to simulate flexor tendon

DORSAL EXTENSOR EPITENDINOUS SUTURE TECHNIQUES 1969

repairs.3–5 A cadaver study comparing modified Bun-nell, modified Krackow-Thomas, and augmentedBecker 4-strand repairs of extensor tendons in zoneIV found the augmented Becker repair to be lessresistant to gapping than the other repairs,7 but ex-tensor tendons are often quite flat, and it is difficult toplace a 4-strand core suture. There are no previouscomparisons of epitendinous dorsal-only extensorrepairs.

Initially intended to prevent tendon fraying, epiten-dinous sutures have been demonstrated to contribute upto 50% of the repair strength in flexor tendons.8–10

Multiple epitendinous sutures are described. Graspingcircumferential sutures are more resistant to gappingand failure than running sutures.11,12 A running, grasp-ing, cross-stitch technique is reliable in clinical use.13,14

The Halsted epitendinous technique has been found tobe stronger than the Silfverskiöld cross-stitch in flexortendon repairs.15 In vitro testing suggested that aninterrupted horizontal mattress (IHM) epitendi-nous technique might be stronger than a simplerunning stitch, cross-stitch, or interlocking cross-stitch.16 The IHM technique is a continuous suturethat includes IHM components (Fig. 1D), a cir-cumferential form of which has been validated forflexor tendon repairs.17 A buried Silfverskiöld re-pair was stiffer and stronger than conventionalSilfverskiöld or Halsted epitendinous repairs in aflexor tendon model,18 although no significant dif-ference in strength was found between simple running,locking, Halsted, Lin, or Lembert techniques.19 Deeperepitendinous sutures are significantly stronger than su-perficially placed sutures.8 Although a peripheral cross-stitch epitendinous suture was found to be weaker thana core and epitendinous suture in a flexor tendon repairmodel,20 dorsal-only Silfverskiöld repairs have beenfound to be stronger than Kessler or mattress core-onlyrepairs.2

We aimed to compare the strength of Silfverskiölddorsal-only epitendinous extensor tendon repairs withburied Silfverskiöld, Halsted, and IHM sutures, all per-formed as dorsal-only repairs. These were chosen asbeing among the strongest described in the literaturethat would be practicable for extensor tendon repair. Anattempt to more accurately re-create the biomechanicalsituation in the fingers was made by testing the repairsover a pulley, thereby creating a tension-band model, aspreviously described.2

MATERIALS AND METHODSThirty fresh-frozen and thawed porcine rear-foot exten-

sor tendons (Fresh Tissue Supplies Horsham, West

JHS �Vol A, D

Sussex, UK) were harvested and sorted. Tendons werematched into 4 groups by diameter so that the meanmedial-lateral diameter within each group was 7 mm.All tendons used were trimmed, and a 6-cm length oftendon was marked out. Tendons were sharply dividedin the center of this 6-cm length, and repaired as follows(Fig. 1) with dorsal-only sutures:

(1) Silfverskiöld running cross-stitch with 5-0 nylonand 2 knots external to the repair (n � 8).

(2) Reverse Silfverskiöld running cross-stitch with 5-0nylon and 2 knots external to the repair (n � 8).

(3) Halsted running epitendinous repair with 5-0 ny-lon and knots external to the repair (n � 8).

(4) Interlocked running horizontal mattress with 5-0nylon (n � 6).

All repairs were performed so that the suture bites

FIGURE 1: A Silfverskiöld running cross stitch. B Reverse(buried) Silfverskiöld running cross-stitch. C Halsted runningepitendinous repair. D Interlocked running horizontal mattressepitendinous repair.

were 6 mm from the site of tendon division, and each

ecember

1970 DORSAL EXTENSOR EPITENDINOUS SUTURE TECHNIQUES

bite was 2 mm in width and depth. There were 6 passesperformed in each case. Repairs were performed un-der magnification after the tendons were marked toensure consistency. Ethilon (Ethicon, Kirkland,Scotland, UK) nylon sutures on multipass needleswere used for all repairs.

Testing of the repairs was performed as previouslydescribed,2 which was consistent with our previousfindings and those of others.12,20 Repairs were tested ina specially constructed rig (Fig. 2) in which the tendonpassed over a lubricated pulley so that an angle of 45°is maintained at the site of tendon repair. The workingdiameter of the pulley was 17.8 mm, and the surfacewas smooth plastic, lubricated with liquid paraffin.These parameters were chosen to approximate the dor-sal surfaces of the proximal and distal interphalangealjoints. The tendon was positioned so that the repair wasover the inferior convexity of the pulley, ensuring that itremained in contact throughout the test unless alarge gap developed in the repair. The tendon wasgrasped at either end in a sandpaper-lined frictionclamp so that the length of tendon between theclamps was the 6 cm previously marked out and

FIGURE 2: Photograph of the repair-testing rig. The pulley ispositioned so that the tendons pass through a 45° angle andthat the repair is in contact with the lubricated pulley.

the repair was at the center.

JHS �Vol A, D

Testing was performed using an Instron 5544 tensi-ometer with a 500-N load cell (Instron, Norwood, MA).A preload of 1 N was applied, and this was increased by20 N per minute. These values were chosen to ensureconsistency with our previous experiments and thoseperformed by Tang et al.12,20 A calibrated scale waspositioned adjacent to the pulley for measurement ofrepair site gapping on magnified video playback. Gap-ping was measured between the ends of the tendonsubstance, which was always the tension side of therepair. Video recording of the testing was made using avideo camera (JVC Everio camera; JVC UK, London,UK), and playback and data recording were performed(VLC Mediaplayer 1.0.1 Goldeneye; Videolan, Paris,France).

The data points recorded were (1) the force requiredto achieve a 2-mm gap in the tendon repair, (2) thefailure force of the repair, indicated by the peak forcerecorded by the tensiometer, (3) the maximum exten-sion (used to calculate stiffness) of each repair, and (4)the mechanism of failure. Data analysis was performedusing analysis of variance (ANOVA) techniques,namely Tukey for parametric and Kruskal-Wallis fornonparametric data, after testing for normality of dis-tribution. Dunn’s method for multiple pairwise compar-isons was also used for nonparametric data.

RESULTS

Gapping

The mean force required to cause a 2-mm gap (and SD)are shown in Figure 3. On ANOVA testing, the IHMwas significantly stronger than the Halsted (P � .004)and the Silfverskiöld (P � .001). It showed a trend forbeing stronger than the reverse Silfverskiöld, but thisdid not reach significance (P � .2). There were nosignificant differences between the gapping strength ofthe other repair types.

Failure

All the repairs failed by breaking of the suture materialrather than by pullout. The force required for totalfailure of the repair was recorded by the tensiometer asthe peak force achieved. These data are shown in Figure3. On ANOVA testing, the IHM was significantlystronger than the Halsted (P�.001), the Silverskiold(P � .004), and the reverse Silfverskiöld (P � .015).There were no significant differences between the fail-ure strengths of the other repair types.

Resistance to gapping and stiffness of the repair

In our study, the force applied to the tendon and the

total extension of the sample, as measured by the ten-

ecember

DORSAL EXTENSOR EPITENDINOUS SUTURE TECHNIQUES 1971

siometer, is used to deduce a stiffness of the repair. Thischaracteristic of the repaired tendon measures the resis-tance to repair extension in a more general way thantaking the force at an arbitrary gap of 2 mm, althoughthe 2 measures are related, given the reasonable linear-ity of the force-extension plots up to gaps of 2 mm. Nodeformation was observed at the specimen grips (ie, thetendon-tensiometer interface) and the elastic deforma-tion within the tendon can be assumed to be smallcompared to the extension at the repair site. Hence, thespecimen extension is a good estimate of the gap at therepair, and it avoids the need to estimate the gap fromthe video image. Stiffness was calculated from the slopeof the portion of the force/extension curve betweenextension values of 0 and 2 mm. The median calculatedstiffness (and SD) for each repair is shown in Figure 4.On ANOVA testing, there were no significant differ-ences in stiffness between any of the repair types,although the power of the study was 0.049 (desired is0.080 or greater), meaning that the negative findingsshould be interpreted cautiously.

DISCUSSIONThe theory that a tension-band concept might be applica-

FIGURE 3: Mean (and SD) force (N) required for 2-mm gap anin vitro.

ble to tendon repairs was raised as early as 1994.21 Tang et

JHS �Vol A, D

al showed that the 2-mm gap formation force and ultimatetensile strength of tendon repairs decreased as they weresubjected to tensile force around angles from 0° to 90°.3

This group tested a variety of flexor tendon repair tech-niques around a curved apparatus.3,4 The model used forrepairing and testing tendons was the same as previouslydescribed.3 Our experimental apparatus was designed toclosely replicate that of Tang, although our study questionsrelated entirely to techniques of potential use for extensortendon repair, whereas Tang et al looked exclusively atflexor tendon repairs.

Porcine tendons were chosen for their availabilityand similar characteristics to human tendons during invitro testing.22 The tendons are not totally flat, but ovalin cross-section. Tendons were matched into groups bydiameter so that the mean medial-lateral diameterwithin each group was 7 mm, although it has beenfound that tendon diameter does not affect outcomes ifsuture techniques are standardized.23 We recognize thatthis might not extrapolate to clinical tendon repairs, butit allows comparison between repair techniques in vitro.Extensor tendons were chosen because they are moresimilar than porcine flexors in size and morphology tohuman extensors, with a flatter cross-section and longi-

tal failure of repair given by the maximum recorded force

d to

tudinal orientation of fibers. Porcine tendons are thicker

ecember

erski

1972 DORSAL EXTENSOR EPITENDINOUS SUTURE TECHNIQUES

than those of humans, and those used in this experimentwere more rounded than the flatter tendons found inhuman fingers.

Cyclical testing might be more sensitive than statictesting by better simulating the behavior of repairsduring active motion rehabilitation regimes.24 In addi-tion to static testing in vitro, cyclical testing in vitro andcadaveric studies can be used for testing tendon repairs.Other types of extensor tendon repairs have previouslybeen evaluated in this setting,7 and it might be thatdorsal-only extensor tendon repairs could be tested inthis way before becoming adopted for clinical use.25

Although our results are a promising start, further val-idation of the epitendinous-only extensor tendon repairtechniques is warranted.

The Silfverskiöld superficial-only repair with 5-0nylon has previously been shown to be superior to thecommonly used Kessler and mattress sutures performedwith 4-0 nylon.2 The other techniques were selected forevaluation because they had been found to be superiorto Silfverskiöld for flexor tendon repairs.15–18 Our find-ings with the extensor tendon model are consistent withthose of these authors. Although not reaching signifi-cance, the buried Silfverskiöld was stronger than theconventional Silfverskiöld, and the IHM was signifi-cantly stronger than the conventional Halsted as well asthe other repairs. The IHM was also stiffer, although

FIGURE 4: Mean (and SD) stiffness (N/mm) of Silfv

this did not reach significance.

JHS �Vol A, D

Technically, the IHM is the most complex repair toperform, and due to the external suture passes, thistechnique leaves a large amount of suture material onthe surface of the tendon, potentially increasing adhe-sions. This problem might also be an issue with theconventional Silfverskiöld repair. Conversely, the bur-ied Silfverskiöld and Halsted techniques leave less su-ture material exposed, and these repairs might benefitfrom reduced adhesions.

Comparing these findings with our previous work,every type of dorsal-only epitendinous extensor repairwas stronger than Kessler or mattress techniques whentested in our angular jig. This gives further support to ourhypothesis that the tension-band principle can be appliedto extensor tendon repairs, possibly to an even greaterextent than in flexors.3 Cao et al found gap formation at 16N for modified Kessler 2-strand repairs and 34 N for theTang 6-strand flexor repairs around a pulley. In both cases,this was significantly weaker than when tested in a straightline.5 Because the flat extensor tendons are less amenableto the placement of core sutures, the finding that a dorsal-only epitendinous-type repair is strong is encouraging. Ourresults are favorable in comparison to those of Tang,although the testing conditions were not identical. Inter-locking horizontal mattress sutures are technically difficultto perform in flat tendons, and, although mattress suturesare technically easier, these data support dorsal sutures

öld, reverse Silfverskiöld, Halsted, and IHM sutures.

over these traditional techniques. The Silfverskiöld tech-

ecember

DORSAL EXTENSOR EPITENDINOUS SUTURE TECHNIQUES 1973

nique is familiar to most surgeons, but it is not quite asstrong as the IHM. It might be that the Silfverskiöld repairis strong enough to allow early active mobilization, which,as with flexors, can lead to a higher proportion of good orexcellent outcomes.26,27 Interlocking horizontal mattresssutures also leave a larger amount of external suture ma-terial. Accounting for strength and stiffness, as well asconsidering ease of repair and amount of external suture,the buried Silfverskiöld might be the optimum repair. Aswith flexor repairs, only clinical testing will determinewhether this translates to improved clinical results whenthe confounding factors of suture material exposure, adhe-sion formation, and the complex changes in tendon com-position and strength after repair are included.

REFERENCES1. Pauwels F. Biomechanics of the locomotor apparatus. Effects of

tension direction on strength of tendon repair. New York, Berlin,Heidelberg: Springer-Verlag, 1980.

2. Henderson J, Sutcliffe M, Gillespie P, The tension band principleand angular testing of extensor tendon repairs. J Hand Surg 2011;36E:297–302.

3. Tang JB, Wang B, Chen F, Pan CZ, Xie RG. Biomechanical eval-uation of flexor tendon repair techniques. Clin Orthop Relat Res2001;386:252–259.

4. Xie RG, Zhang S, Tang JB, Chen F. Biomechanical studies of 3different 6-strand flexor tendon repair techniques. J Hand Surg2002;27A:621–627.

5. Cao Y, Xie RG, Tang JB. Dorsal-enhanced sutures improve tensionresistance of tendon repair. J Hand Surg 2002;27B:161–164.

6. Wong JK, Cerovac S, Ferguson MW, McGrouther DA. The cellulareffect of a single interrupted suture on tendon. J Hand Surg 2006;31B:358–367.

7. Howard RF, Ondrovic L, Greenwald DP. Biomechanical analysis offour-strand extensor tendon repair techniques. J Hand Surg 1997;22A:838–842.

8. Diao E, Hariharan JS, Soejima O, Lotz JC. Effect of peripheralsuture depth on strength of tendon repairs. J Hand Surg 1996;21A:234–239.

9. Wade PJ, Muir IF, Hutcheon LL. Primary flexor tendon repair: themechanical limitations of the modified Kessler technique. J HandSurg 1986;11B:71–76.

10. Kim JB, de Wit T, Hovius SE, McGrouther DA, Walbeehm ET.What is the significance of tendon suture purchase? J Hand Surg

2009;34E:497–502.

JHS �Vol A, D

11. Mashadi ZB, Amis AA. Strength of the suture in the epitenon andwithin the tendon fibres: development of stronger peripheral suturetechnique. J Hand Surg 1992;17B:172–175.

12. Tan J, Tang JB. Locking repairs for obliquely cut tendons: Effects ofsuture purchase and directions of locking circles. J Hand Surg2004;29A:891–897.

13. Silfverskiöld KL, May EJ. Flexor tendon repair in zone II with a newsuture technique and an early mobilization program combining pas-sive and active flexion. J Hand Surg 1994;19A:53–60.

14. Sirotakova M, Elliot D. Early active mobilization of primary repairsof the flexor pollicis longus tendon. J Hand Surg 1999;24B:647–653.

15. Zatiti SC, Mazzer N, Barbieri CH. Mechanical strengths of tendonsutures. An in vitro comparative study of six techniques. J Hand Surg1998;23B:228–233.

16. Dona E, Turner AW, Gianoutsos MP, Walsh WR. Biomechanicalproperties of four circumferential flexor tendon suture techniques.J Hand Surg 2003;28A:824–831.

17. Vigler M, Palti R, Goldstein R, Patel VP, Nasser P, Lee SK.Biomechanical study of cross-locked cruciate versus Stricklandflexor tendon repair. J Hand Surg 2008;33A:1826–1833.

18. Wang B, Tang JB. Embedded cross-stitch suture: an alternative tocurrent cross-stitch peripheral suture. J Hand Surg 2003;28B:471–474.

19. Kubota H, Aoki M, Pruitt DL, Manske PR. Mechanical properties ofvarious circumferential tendon suture techniques. J Hand Surg 1996;21B:474–480.

20. Papaloïzos MY, Schärer N, Rehart S, Stüssi E, Meyer VE. Crossstitch peripheral tendon repair: a mechanical comparison with corestitch techniques. Chir Main 2000;19:128–133.

21. Tan J, Wang B, Tan B, Xu Y, Tang JB. Changes in tendon strengthafter partial cut and effects of running peripheral sutures. J HandSurg 2003;28B:478–482.

22. Tang JB, Shi D, Gu YQ, Chen JC, Zhou B. Double and multiplelooped suture tendon repair. J Hand Surg 1994;19B:699–703.

23. Hausmann JT, Vekszler G, Bijak M, Benesch T, Vécsei V, Gäbler C.Biomechanical comparison of modified Kessler and running suturerepair in 3 different animal tendons and in human flexor tendons.J Hand Surg 2009;34A:93–101.

24. Boyer MI, Meunier MJ, Lescheid J, Burns ME, Gelberman RH,Silva MJ. The influence of cross-sectional area on the tensile prop-erties of flexor tendons. J Hand Surg 2001;26A:828–832.

25. Pruitt DL, Manske PR, Fink B. Cyclic stress analysis of flexortendon repair. J Hand Surg 1991;16A:701–707.

26. Sylaidis P, Youatt M, Logan A. Early active mobilization for exten-sor tendon injuries. The Norwich regime. J Hand Surg 1997;22B:594–596.

27. Crosby CA, Wehbé MA. Early protected motion after extensor

tendon repair. J Hand Surg 1999;24A:1061–1070.

ecember