~

G CL-

:1

~ILately, many suture material\: have been introduced.Their physical characteristics in combination with knotsare not well known.

In this study, seven knots (square-1=1, 2=1, 2=1-Sand l=l=l-and sliding-SxSxS, S=S//S and l-S//S/ /S) made in seven suture material\: (plain catgut,Dexon@ [polYglycolic acid)J Maxon@ [polygiyconateJ, PDS@

oolydiaxoneJ, Vicrytf!l [p°lyglactine 910J, Mersiúmé1!:laolyester fiberJ, Prolen{;l!!> [polypropyleneJ were tested

aynamically to ascertain tensile strength. The knotswere classified as '~edominantly breaking" (PB) and"l1redominantly slipping" (PS). A new method for sta-tistical analysis, the Kaplan-Meier survival estimate,was introduced.

Square knots provided good mecha ni cal results butdid not prevent slippage completely. Most sliding knotswere weak. The 1=1=1 knot was superiO1: PS knots(1=1, 2=1, SxSxS and S=S//S) were unsuitable forsurgical practice in monofilament or coated multifila-ment suture material\:. The classification PB and PSknots gave an easy impression of the knot holdingcapacities. Application of the Kaplan-Meier estimateresulted in a more reallitic analysis than classical meth-ods.

SUTURES FAIL for various reasons. Suture materialscan break and knots can slip (1). Failures canresult in wound dehiscence, which is a seriouscomplication in surgery (2). Therefore, knowl-edge of mechanical properties or surgical knotsin different suture materials is important. Duringthe past ten years, a number of new suture ma-terials have been introduced but have not been

extensively investigated.In most publications, static loading (tear speed

5 centimeters per minute) is used to measuretensile strength of suture materials (3-6). How-ever, after wound closure, considerable forces athigh speed can be present, which are evokedby coughing or vomiting. Thus, in this study,fue mechanical properties of different materialsand both square and sliding knots are investigatedafter dynamic loading (tear speed 75 centimetersper minute) (7).

From the Departments of Surgical Research and Medical Statistics,Academic Hospital, University of Amsterdam, Amsterdam, The Neth-erlands.

Reprint requests: Dr. J. E. Brouwers, Department of Surgical Re-search, Academic Medical Center, Meibergdreef 9, 1105 AZ Am-sterdam, The Netherlands.

4'

Analysis of data was performed by using classicmethods. However, those methods cannot differ-entiate between slippage and breakage failures.Therefore, analysis was algo performed by usingfue Kaplan-Meier survival estimate, fue results ofwhich can differentiate between break and slipfailures (8).

MATERIAlS AND METHODS

In this study, plain catgut, three monofilamentmaterials-Prolene@ (polypropylene), PDS@ (po-lydiaxone) and Maxon@ (polyglyconate)-andthree multifilament materials-Mersilene@ (poly-ester fiber), Dexon@ (polyglycolic acid) and Vi-cryl@ (polyglactine 910) were tested. A micro-scope grid of 0.016 millimeter was used tomeasure thread diameter (3-0 USP) precisely. Itvaried between 0.24 and 0.36 millimeters (Table1). Four square knots (1=1,2=1, 2=1-S and 1=1=1)and three sliding knots (SxSxS, S=S/ /S and 1-S/ /S/ /S) were tested (Fig. 1) (6, 9).

Two different experiments were performed.First, the properties of single untied suture threadwere assessed. Second, knotted loops of suturematerial were tested. These loops were form'edby knotting fue material around a device thatconsisted of two metal cylinders with a diameterof 1 centimeter and fixed on a board at a distanceof 6.5 centimeters.

Single threads were toro apart six times andloops ten times each by a homemade testingmachine based on fue Hounsfield tensometer(Monsant9, United Kingdom). In total, 532 ex-perimentS were performed. Dynamic load (tearspeed of 75 centimeters per minute) was obtainedusing a pneumatically driven piston of a Marton-air cylinder. Elongation and force were simulta-neously registered with a magnetic displacementdevice (HBM W 100, Novotechnic, Germany) anda force transducer (PR 9310/03, Philips, TheNetherlands). The result, a stress-strain diagram,was recorded on a X-y potentiometric recorder(PR 9340, Philips, The Netherlands) after ap-

propriate amplification.Four parameters were used for evaluation of

data. Tensile strength of the untied fiber (TSuf)'1:3

RESULTSis fue maximal force in fue stress-strain diagram Most loops could withstand a force of 20 N.necessary to break an untied thread of suture The mean of all observations for slippage wasmaterial, expressed in Newton (N). Mean loop 19.4 N and for breakage, 39.9 N. The meanholding capacity (mLHC) is the maximal force break percentages for fue tested materials andin fue stress-strain diagram necessary to break knots are arranged and presented in Figure 3.from fue inside, a knotted loop of suture material, Values of Tsuf and mLHC are presented inexpressed in Newton (N). Breaking and slipping Tables I and 11. The values of fue kmLHC arepercentage is fue percentage break or slip seen in accordance with those of mLHC (48 of 49in a knot-thread combination. Slippage was de- kmLHC values are within one standard deviationfined as more than 2 millimeters extra elongation of the mLHC). Interpretation of fue data ison fue stress-strain diagram. A knot that slips mainly performed by comparing fue mLHC val-more often than 50 per cent (n=70) is called ues and by performing a statistical analysis witha "predominantly slipping" knot (PS knot), a fue generalized Savage test on fue survival graphs.knot that slips in less than 50 per cent a "pre- In the following, fue most interesting results aredominantly breaking" knot (PB knot). Statistical emphasized. Unless otherwise stated, all resultsanalysis methods that have be en used until now are significant.do not account for discrimination between slip- Sliding knots. Knots SxSxS and S=S/ /S are PSpage and breakage of knots. The Kaplan-Meier knots and a knot l-S/ /S/ /S is a PB knot. Thesurvival estimate, originally used in oncology, pro- SxSxS knot, which is easy to make, has fue lowestvides a measure of loop holding capacity (Kaplan- mLHC values. If fue last throw is made arourdMeier loop holding capacity [kmLHC]) including fue other thread (S=S/ /S) instead of three throwsslip and break events. around the same thread (SxSxS), the knot

The survival graph (Fig. 2) shows force on strength increases significantly (except Maxon,fue X-axis and knot survival percentage (survival which is nonsignificant [NS]). Compared withchance) on fue Y-axis. The curve begins at force S=S/ /S, fue mLHC of fue rather complex 1-zero and 100 per cent survival rateo Mter each S/ /S/ /S knot in creases dramatically, especiallybreak, fue total number of surviving knots and when monofilaments and Vicryl are used.force is displayed. A break results in a decrease Square knots. In Figure 3, knot 1=1 is a trueof fue curve. Slipping of a knot is not displayed PS knot and 2=1 is a PS knot in general. Bothas a direct decrease. It is visualized after fue 1=1=1 and 2=1-S are PB knots.next break by counting fue total of surviving According to fue rnLHC, knot 1=1 is a weakknots. Therefore, slipping is indirectly included. knot. The addition of a third square throw

Statistical analysis is performed by using fue (1=1=1) makes it fue strongest one. Tensile

TABLE ll.-MFAN LOOP HOLDING CAPACIn: STANDARD DEVlAnON, IN NEWTON, AS DETERMINED IN TEN TFSrS FOR EACH

COMBINAnON OF KNOT AND MATERIAL, 3-0 VSP

Catgut Maxon PDS Prolene Dexon Mersilene ViaylSquare knots .171 (81)1=1 23.7 (3.8) 32.3 (10.1) 12.5 (4.3) 7.4 (lA) 24.1 (4.1) 20.5 (4.0) .9.8)2=1 26.2 (2.8) 28.5 (9.2) 12A (4.9) 6.2 (1.7) 34.1 (9.0) 36.9 (2.3) 2~.g ~4'7)2=1-S 24.5 (2.9) 36.9 (9.2) 36.5 (8.3) 26.7 (1.6) 35.8 (5.1) 36.9 (5.4) 2.2 (27)1=1=1 29.2 (1.5) 49.1 (9.2) 31A (6A) 24.8 (1.7) 39A (3.8) 37.3 (1.9) 34. .

Sliding knots 4 141 (1.7)SxSxS 24.6 (4.7) 26.0 (8.8) 16.2 (3.9) 13.1 (3.7) 24A (8.1) 21.2 (64) 17:8 (4.7)S=S//S 25.5 (1.7) 22.1 (4.8) 22.3 (6.9) 20.7 (4.6) 38A (3.2) 34.8 (6.) 388 (79)1-S/ /S/ /S 29.6 (2.5) 46.1 (3.5) 30.5 (4.6) 25.2 (3.2) 37A (11.1) 37.8 (4.2) ..

TABLE I.-PH\SICAL PROPERTIES OF TESTEO MATERJALS

d TSufPlain catgut 0.36 25.5

Three monofilament

Maxon 0.31 34.5POS'!> (polydiaxone) '.. '" ...0.30 27.2

Prolene'!> (polypropylene) 0.26 16.7Three multifilament

Oexon'!> (polyglycolic acid) 0.24 29.1Mersilene 0.26 28.3Vicryl'!> (polygactin 910) 0.29 34.6

-d, Diameter in millimeten, and TSuf, tensile strength of the untied fiber(N).

Al! experimento were performed by the uses of suture materia!s of thesize category of ~ USP.

generalized Savage test. In this test, survivalgraphs are compared (p<0.05). Of course, slipcan be accentuated instead of break by displayingslippage directIy and breakage indirectIy. Direct-break analysis is mainly used for PB knots,whereas direct-slip anaIysis is used for PS knots.The kmLHC is defined as fue force at which50 per cent of fue knots survive (50th percentilein fue survivaI graph).

r

Brouwers el al.: DYNAMIC LOADING OF SURGICAL KNOTS 445

$x$x$

1=1=1

5 m 5115

strength generally improves when the surgeonsknot (2=1) is used instead ofthe 1=1 knot (exceptmonofilamepts and Vicryl, which are NS). Sup-plementation of a simple sliding throw to the2=1 knot yields a stronger knot: 2=I-S (exceptcatgut, Dexon and Mersilene, which are NS).

Sliding knots versus square knots. The three-throwsquare-knot (1=1=1) is stronger than the strongestthree-throw sliding knot (S=S/ /S) (except Dexonand Mersilene, which are NS). The four-throwsliding knot (I-S/ /S/ /S) is as strong as the three-throw square-knot.

Monofilament material. Maxon is the strongestmaterial examined and Prolene is the weakest.

Multifilament material. Vicryl is thoe weakest ma-terial despite its high TSuf (except I-S/ /S/ /S,which is NS). Dexon and Mersilene are compa-rable.

Catgut. The variation in mLHC is small whenvarious knots are used. The break percentage is

high.Catgut versus monofilament versus multifilament ma-

terials. In general, monofilament materials slipmore often than multifilament materials (exceptVicryl). Catgut slips less often than multifilaments.

When the strength of catgut is compared withthe strongest monofilament (Maxon) and mul-tifilament material (Dexon), catgut is weaker (ex-cept when catgut'is compared with Maxon 2=1NS). Comparison of an absorbable monofilament(Maxon) and an absorbable multifilament (Dex-on) results in a higher holding capacity for Max-on (except 2=1 and 5=5/ /S knots, which areN5). The two strongest nonabsorbable materialscompared, Prolene versus Mersilene, show an in-variably higher mLHC for the latter one.

DISCUSSION

2 = 1-51-51/51/5

'-"'-"

a b

FIG. l. Configuration and cacle of types of knots. a, Squarethrow, and b, sliding throw.

caused by stiffness of fue material, which doesnot allow easily asymmetric knotting.

The 1=1=1 square-knot and !:he 1-8/ /S/ /S slid-ing-knot have equal mechanical characteristics,but making fue 1=1=1 knot requires less materialand, therefore, diminishes inflammation of fuetissue. Thus, we generally recommend applicationof fue 1=1=1 knot. If sliding knots cannot beavoided, as in ligating in the depth of a wound,fue 1-5/ /S/ /S should be used (10, 11).

In surgery, fue choice of material depends pri-marily upon whether or not resorption is nec-essary. If nonabsorbable materials are needed,Mersilene is fue first choice. Prolene is quiteelastic and has unfavorable mechanical proper-tieso Choosing a resorbable material is more com-plexo Among knot holding capacities, many fac-tors should be accounted for, such as degradationspeed, tissue inflarnmation, friction coefficientand capillary suction of bacteria. When a mul-tifilament is needed, Vicryl should not be used.This material behaves like a weak monofilamentbecause of its smooth coating.

Naming knots "slipping" knots suggests fuetendency of all those knots to slip. The exper-iments contradict such a prejudice. Simple knots,such as SxSxS and S=S/ /S (PS knots), indeedslip often and cannot be recommended. However,fue l-S/ /S/ /S knot is a PB knot. This is probablya result of fue increased internal contact afeathat upgrades fue friction inside the knot.

Forming a square-knot does not always preventslippage, for example, 1=1 and 2=1 are PS knots.Adding a third throw is necessary to obtain astrong knot (1=1=1). Using the more complex2=1-S knot gives no additional advantage. Re-markably, in monofilaments, the mLHC does notimprove when fue more complex 2=1 knot isused instead of the 1=1 knot. This could be

446 SURGERY, Gynecology & Obstetrics'December 1991. Volume 173o/a *.

100 , -1

..L.: * = break

.= slip

50

~~

-*

*--,o

o 50

tensile strength (N)

FIG. 2. Kaplan-Meier survival graphs for the 1=1=1 knot made with Prolene@(polypropylene) (a) and Maxon@ (polyester fiber) (b). The graph displays forceversus knot survival percentage. The curve begins at force zero and 100 percent survival (n=10). Mter each break, the; total arnount of survived knots andforce is displayed. A break, *, results in a decrease of the curve. Slipping, closedcircle, is not displayed as a direct decrease, it is visualized after the next breakby counting the total of survived knots. The line is drawn at the 50th percentile,Kaplan-Meier loop holding capacity.

25

The loops of knotted suture material have beenpulled apart from fue inside of fue loop (12,13). This results in two thread parts: one withand one without a knot. It was stated that fuesetwo parts are loaded unequally; the unknottedpart can lengthen more, move around the sup-port and relieve fue knotted parto Therefore, itwas advised that one should pull at fue two looseends of fue thread (single strand method) (3,4). Because of a combination of high speed andfriction around fue supports, this argument seemsless important in our study.

The high TSuf of some modero materials doesnot imply that those materials are, therefore,more suitable in surgical practice (for example,modero Vicryl Tsuf with a relatively low mLHC,in conventional catgut, it is fue other way around.Additional parameters like mLHC should be re-viewed. Another useful parameter could be "knotefficiency" (KE). It is calculated by combiningTSuf and LHC through the formula:

chance of an early wound dehiscence is mini-mized. As stated, the LHC of PB knots is muchhigher compared with PS knots.

The results of statistical analysis using Studentt and chi-square tests do not discriminate betweenslippage and breakage. These methods attach anequal value to each evento This is a false pre-sumption, since the results of the study provedthat slippage occurs at lower tensile forces thanbreakage. The introduced Kaplan-Meier survivalestimate can express slippage and breakage be-cause it has a "censored" modality. Besides, thesurvival graph is an elegant graphic tool thatdisplays the differences between knots and ma-terials. Whether or not it is more advisable touse the 50th percentile of the Kaplan-Meier es-timate (kmLHC) instead of the mLHC is partof a further study. ,

Of course, the circurrlstances in laboratoriesdiffer considerably from those encountered dur-ing clinical intervention. The tensile strength canbe affected by various fluids in the wound. More-over, each surgeon ties a knot in a differentway. For example, the force applied to eachthrow, the speed of the knotting and the useof instruments determine the knot holding ca-pacity (14-16). Even research methods are notstandardized. Some investigators use dry threads,others soak threads in plasma or saline solution

KE=0.5+ (LHC/TSuf) + 100 per cent.

The KE reflects fue de crease in tensile strengthof the suture material caused by knotting.

An easy ~ethod to evaluate knots is to arrangethem in PS and PB knots. For PB knots, fuedegree of slippage is less than 50 per cent; the

-"'--

material knot

FIG. 3. Mean break-percentage for the tested materials (n=70) and knots (n=70)indicating the difference between "predominantly slipping," PS, and "predom-inantly breaking," PB, knots. 1, 1=1; 2, SxSxS; 3, 2=1; 4, S=S/ /S; 5, 2=1-8; 6,1=1=1, and 7, 1-S/ /S/ /S. PDS, polydiaxone.

(17-19). " Sometimes .knots are made with instru-

ments or tightened to a force of 80 per centof fue TSuf (20). Various tear speeds and dif-ferent knots have been applied (7,21-23). There-fore, fue results of most studies cannot be com-pared with each other.

The lack of unity calls for intemational stan-dardization of suture material research and itsclinical application. The use of fluids, knottingmethods, tear speed, kind of knot used and fuechoice of loop or single strand method shouldat least be standardized. Differentiation betweenslip and break in analyzing knots and suturematerials should be performed by means of theKaplan-Meier survival estimate.

SUMMARY

In this study, seven knots (1=1, 2=1, 2=1-S,1=1=1, SxSxS, S=S/ /S and l-S/ /S/ /S) made inseven materials (plain catgut, Dexon, Maxon,PDS, Vicryl, Mersilene and Prolene) were testedwith a dynamic loado The need for internationalstandardized testing of suture material is accen-tuated in view of fue increasing number of prod-ucts.

As a new method to judge statistically fue me-chanical properties of knots and materials, theKaplan-Meier estimate is advised. This enablesone to differentiate between slip and break.

Square knots provided adequate mechanical re-sults, but did not always prevent slippage. The1=1=1 knot provided optimal results in all ma-terials tested. The use of sliding and PS knotshas to be limited, in particular, to monofilamentor coated multifilament suture material.

REFERENCESl. Taylor, F. W. Surgical knots. Ann. Surg., 1938, 107:

458-468.2. van Geldere, D. De gebarsten buik. Amsterdam, Uni-

versity of Amsterdam, 1986.3. Holmlund, D. E. W. Knot properties of surgical suture

material. Eur. Surg. Res., 1974, 6: 65-71.4. Ide.m. Knot properties of surgical suture materials. Acta

Chlr. Scand., 1974, 140: 355-362.5. Meiss, L., and Holzrichter, D. Blockierende Gleitknoten

in der Chirurgie. ChiTo Praxis, 1982, 30: 573-581.6. Tera, H., and Aberg, C. Tensile stren~ of twelve

type ofknots employed in surgery, using dlfferent suturematerials. Acta Chir. Scand., 1976, 142: 1-7.

7. Visser, J. D. Dynamic strength of surgical suture ma-terials. In: Evaluation of Biomaterials. Edited by G. D.Winter, J. L. Leray and K. de Groot. Pp. 269-274. Lon-don: John Wiley & Sons, Ltd., 1980.

8. Matthews, D. E., and Farewell, V. T. Using and un-derstanding medical statistics. Basel: S. Karger, 1985,67-74.

9. Trimbos, J. B. Security of various knots commonly usedin surgical practice. Obstet. Gynecol., 1984,64: 274-280.

10. Trimbos, J. B., Rijssel, E.J. C van, and Klopper, P. J.Performance of sliding knots in monofilament and mul-tifilament suture material. Obstet. Gynecol., 1986, 68:425-430.

11. Meiss, L. Blockierende Gleitknoten, eine Erganzung chi-rurgische Knoten? Zentralbl. Chir., 1983, 108: 1051-1056.

12. Nilsson, T. Mechanical properties of Prolene, Ethilonand surgical steelloop~. Scand. J. Plast. Reconstr. Surg.Hand Surg., 1981, 15. 111-115.

13. Herrmann, J. B. Tensile strength and knot security ofsurgical suture materials. Am. Surg., 1971, 37: 209-217.

14. Cono, J., Ir., Oyasu, R, Welsh, M., and Beal, J. M.Vicryl (polyglactin 910) synthetic absorbable sutures.Am. J. Surg., 1974, 128: 19-23.

15. CTaig, P. H., Williams, J. A, Davis, K. W., and others.Biological comparison of synthetic. absorbable sutures.Surg. Gynecol. Obstet., 1975, 141. 1-10.

16. Herrmann, J. B. Changes in tensile strength and knotsecurity of surgical sutures in vivo. Arch. Surg., 1973,106: 707-710.

17. Howes, E. L. Strength studies of polyglycolic acid versuscatgut sutures of the same size. Surg. Gynecol. Obstet.,1973, 137: 15-20.

18. Thiede, A., Stuewe, W., and Luenstedt, B. Vergleich

¡ l

448 SURGERY, Gynecology & Obstetncs. December 1991 .Volume 173

van physikalischen Parametem und handhabungs Ei-genschaften kurzfristig und mittelfristig absorbierbarerNahtmaterialen. Chirurg, 1985, 56: 803-808.

19. Rodeheaver, G. T., Thacker, J. G., and Edlich, R. F.MechanicaI performance of polyglycolic acid and poly-glactin 910 synthetic absorbable sutures. Surg. Gynecol.Obstet., 1981, 153: 835-841.

20. Luenstedt, B., and Thiede A. Polydiaxanon (PDS), einneues monofiles synthetisches, absorbierbares Nahtma-terial. Chirurg, 1983, 54: 103-107.

21. Marchant, L. H., Knapp, S., and Apter, J. T. Effect ofelongation rate on tensile strength of surgical suturematerials. Surg. Gynecol. Obstet., 1974, 139: 231-233.

22. Marchant, L. H., Knapp, S., Braun, H., and Apter, J.T. Effect of elongation rate on the percentage elon-gation of surgical suture materials. Surg. Gynecol. Ob-stet., 1974, 139: 389-391.

23. Nilsson, T. Effect of strain rate on tensile strengthand strain of surgical suture materials. Scand. J. Recon-str. Surg. Hand Surg., 1982, 16: 97-100.