J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/jmbbm

Research paper

Failure analysis of the fractured wires in sternal perichronalloops

Jesús Chaoa,∗, Roberto Vocesb,1, Carmen Peñaa,2

aNational Center for Metallurgical Research (CENIM-CSIC), Avda. Gregorio del Amo, 8, 28040 Madrid, SpainbDivision of Cardiac Surgery, Hospital de Cruces, Plaza Cruces-Gurutzeta 12, 48902 Barakaldo, Spain

A R T I C L E I N F O

Article history:

Received 25 January 2011

Received in revised form

24 February 2011

Accepted 28 February 2011

Published online 6 March 2011

Keywords:

Sternal dehiscence

Perichronal loop

Fatigue

Biomechanics

Stainless steel

A B S T R A C T

We report failure analysis of sternal wires in two cases in which a perichronal fixation

technique was used to close the sternotomy. Various characteristics of the retrieved wires

were compared to those of unused wires of the same grade and same manufacturer

and with surgical wire specifications. In both cases, wire fracture was un-branched and

transgranular and proceeded by a high cycle fatigue process, apparently in the absence of

corrosion. However, stress anlysis indicates that the effective stress produced during strong

coughing is lower than the yield strength. Our findings suggest that in order to reduce the

risk for sternal dehiscence, the diameter of the wire used should be increased.c⃝ 2011 Elsevier Ltd. All rights reserved.

d

1. Introduction

Median sternotomy is often used in cardiothoracic surgery,and stainless steel figure-of-eight and/or single wiring isthe most common procedure for sternal closure. Recently, anew fixation technique, so-called perichronal closure (Fig. 1),has been proposed (Voces-Sánchez, 2009). The idea behindthis technique is that the closure should have the sufficientrigidity to allow primary osseous healing but also sufficientflexibility for uniform distribution of the stresses to preventwire cutting through the bone (Voces-Sánchez, 2009). Failureof postoperative sternal closure due to sternal dehiscence

∗ Corresponding author. Tel.: +34 91 5538900; fax: +34 91 534 74 25.E-mail addresses: jchao@cenim.csic.es (J. Chao), robertovoces@vo

1 Tel.: +34 94 6006339.2 Tel.: +34 91 5538900.

1751-6161/$ - see front matter c⃝ 2011 Elsevier Ltd. All rights reservedoi:10.1016/j.jmbbm.2011.02.015

dafone.es (R. Voces), c.pena@cenim.csic.es (C. Peña).

is a serious issue (Bruhin et al., 2005). Dehiscence occurs,typically, within the first two weeks after surgery resultingfrom the fracture of suture wire or as often occurs, whenit is being pulled through the bone (Brantigan et al., 1979;Wilkinson and Clarke, 1998). Comparative analysis of thedehiscence cases associated to 1188 sternotomies (669 singleclosures, 181 figure-of-eight closures, and 338 perichronalclosures) performed from 2003 to 2008 at the institution ofone of the authors (R.V) has revealed that perichronal closureallows a similar consolidation to those of single and figure-of-eight closures (Voces-Sánchez, 2009). Moreover, in practicallyall cases in which single or figure-of-eight loops were used,dehiscence via cutting through bone occurred whereas when

.

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0 1005

Fig. 1 – Schematic diagram of the use of perichronalclosure to close the chest after an open-heart surgicalprocedure.

perichronal loops were used dehiscence occurred by wirefracture (Voces-Sánchez, 2009).

Previous works have attributed the wire fracture in singleand figure-of-eight loops to the combination of two factors:the cyclic loading arising from the patient activity and thecorrosive in vivo medium (Brantigan et al., 1979; Shih et al.,2005). These studies concluded that in order to prevent wirefracture it is necessary to improve the quality of the sternalwire. However, it has been also reported that the corrodedpores developed in stainless steel wires remain shallow, evenafter decades of exposure and that they are unlike to producethe mechanical failure of the knot (Tamizawa et al., 2006).

The aim of this paper is twofold: firstly, to analyse thecauses of in vivo breakage of perichronal loops; and secondly,to relate the location and features of the wire fracture withthe way in which wire has worked prior to failure. To theknowledge of the authors the failure mechanism of theperichronal loop has not been studied yet. On the other hand,information is very scarce on the structural response of thesternal closure following median sternotomy in the openliterature (Bruhin et al., 2005).

2. Materials and methods

2.1. Patient characteristics and surgical outcome

The study was on sternal wires retrieved from two malepatients, aged 74 and 68 years (Patients A and B), each witha body mass index of 25. Both were drinkers, smokers, andhad respiratory difficulties. After open-heart surgery stainlesssteel of 0.78 mm diameter (Ethicon Ltd., Edinburgh, UK)was used to close the chest. The sternal closure of bothpatients consisted of five loops: one trans-sternal single loopplaced at the manubrium, follow two single perichronal loopsplaced on the third and fourth ribs and finally two doubleperichronal loops placed on the fifth and six ribs. Althoughthe convalescence of both patients was normal, they did notcomply with the recommendations of the medical staff. Thetimes to fracture of sternal wires in Patients A and B were 7and 5 days, respectively. An X-radiograph taken of the thoraxof the Patient A prior to retrieving the fractured wires isshown in Fig. 2.

2.2. Wire characteristics and mechanical properties

After removing the fibrous tissue from the retrieved wires,they were ultrasonically cleaned in ethanol for 30 min and

Fig. 2 – X-radiograph showing multiple wire fractures inPatient A. The arrows indicate the fracture location on theloops.

then, for each wire: (1) the diameter at a point 0.5 mm fromthe fracture site and the diameter at a point 15.0 mm fromthe fracture site were measured; (2) the type and amount ofinclusions were determined per ASTM E45-97 (ASTM, 2005);(3) the grain size was determined per ASTM E112-96 (ASTM,2004); and (4) the amount of each of the alloying elements inthe steel was determined per ASTM E 352-93 (ASTM, 2006).The procedures described in items (1)–(4) were also madeusing an un-used wire. For the microstructural work, the wirespecimen was cut normal to the longitudinal axis of the wire,polished to a finish of 1 µm, and then etched in a solutioncomprising 100 mL CH3–CH2OH, 10 mL ClH and 10 mL H NO3for 20 s. The microstructure of the material was obtainedusing an optical microscope (Model PME3 Olympus, Münster,Germany). The morphology of the fracture surface wasobtained using a scanning electron microscope (Model S2100,Hitachi Europe, Kretell, Germany) operated at an acceleratedvoltage of 15 kV, in the secondary electron imaging mode.

For the un-used wires, the yield strength (YS), ultimatetensile strength (UTS), and% elongation at fracture (El) weredetermined from 350 mm-long specimens using a universaltesting machine (Model EM2/100/FR-10 kN; Microtest, Madrid,Spain) operated according to ASTM 1350-91 (ASTM, 2008) at astrain rate of 0.02 s−1.

Alternating bending fatigue tests were carried out onthree 180 mm-long specimens of the un-used wire witha reciprocating cantilever bending fatigue machine (ModelUA246, Amsler, Schafhouse, Switzerland). The test consistsof alternatively bending the specimen on rounded grips of25 mm radius. The distance between the gripping zone andthe area in which the force was applied was 80 mm and theapplied bending angle was ±15◦. The tests were performed inlaboratory air, at room temperature and a frequency of 0.5 Hz.

2.3. Stress analysis

The maximum intra-thoracic pressure (P), due to inspirationand expiration, during normal breathing, is about 1.4 kPa

1006 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0

(Sharpey-Schafer, 1953; Ratan, 1993). A simple “ahem”,ordinary coughing, a strong coughing (i.e. one that leads tosyncope) produce P of 3.3 kPa, between 7 kPa and 13 kPaand between 20 kPa to 40 kPa, respectively (Sharpey-Schafer,1953). Idealizing the thoracic cage as a thin walled cylindricalvessel (Casha et al., 1999), leads to the following expressionfor the pressure-induced hoop stress in the sternal wire (Gere,2002a):

σHw =2 · P · r · L

N · π · d2w(1)

where r and L are the radius and length respectively of thethoracic cage, N is the closure number, and dw is the diameterof the sternal wire. On the other hand, the intrathoracicpressure leads to a concentrated radial force on each wire of:

FRw = P · Bs · dw (2)

where Bs is the breath of the sternum. Considering thetransverse wires of the loops as beams loaded at theirmidpoints by the FRw force and supported by the lateraledges of the sternum halves, the maximum stress on the wiresurface is (Gere, 2002b):

σBw =8 · FRw · Bs

π · d3w. (3)

Thus the maximum effective stress is:

σeff = σHw + σBw.

Using r = 0.12 m, L = 0.25 m, Bs = 0.03 m, dw = 0.00078 m andN = 7 we calculated values of σHw, σBw, and σeff for variouslevels of P. The value of the factor of safety against yield ofthe wire (FOS) was calculated by dividing the material yieldstrength by the effective stress.

3. Results

In the retrieved wires, the diameter of the wire at 0.5 mmfrom the fracture site was not significantly different fromthat at 15.0 mm from that site and of that of un-usedwire, which indicates that fracture occurred in the absenceof macroscopic plastic deformation. The total amount ofinclusions (sulfides, alumina, silicates, and globular oxides)in the retrieval wires was not significantly different fromthat of the un-used wire. The total amount of inclusions ineach of these types of wire was significantly lower than thatcorresponding to the severity level of 5 specified for surgicalwires (ASTM, 2008), which means that stock wire used tofabricate the sternal wires was very clean. The grain size ofthe retrieved wires from Patients A and B was 22 ± 8 µm and45 ± 6 µm, respectively, while that of the un-used wire was10±6 µm. These differences in grain size could be attributed todifferences in the thermo-mechanical treatment of the wire.In spite of these differences, the grain size of the retrievedwires is in agreement with the specification for surgical wires(ASTM, 2008). In terms of chemical composition of the steel,there was no significative difference in the amount of a givenalloying element in the retrieved wires and those of the un-used wire, Table 1. In both cases the chemical composition isin agreement with the specification for surgical wires (ASTM,

Fig. 3 – Features of the fracture surface at highmagnifications of wire retrieved from Patient A (a) and fromPatient B (b). The arrows indicate the fracture initiationpoints.

2008). The tensile properties of the un-used wire (YS = 379 ±

15 MPa;UTS = 738 ± 8 MPa;El = 48.6%) are also in agreementwith specifications (ASTM, 2008). In the alternating (±15◦)fatigue tests, fracture occurred after 92 000 ± 12 000 loadingcycles. Supposing the fatigue specimen as a cantilever beamof 80 mm length and 0.78 mm diameter, it can be calculatedaccording to Gere (2002b) that the stress on the wire surfacecorresponding to 15◦ bending angle is slightly higher than YS.

Un-used sternal wires contained some scratches andshallow crevices probably produced during the drawing andsterilization stages of the manufacturing process.

In both patients sternal dehiscence occurred due to thewire fracture at themidpoint of one of the transverse portionsof all perichronal loops, Fig. 2. In the wires retrieved fromPatient A, the main fracture initiated at a point on the wiresurface and propagated through the cross section (Fig. 3(a)),whereas in the wires retrieved from Patient B, the mainfracture started at two points located on opposite sides andpropagated towards the wire centre (Fig. 3(b)). In both Patientsthe propagation direction is near the anterior–posteriordirection. At high magnifications a finely spaced striationtypical of fatigue fracture was observed (Fig. 4). Secondarysurface microcracks parallel to the main fracture initiated onthe same side as the main fracture in Patient A, whereasin Patient B they initiated on the same side as initiated thelargest fatigue crack of the main fracture, Fig. 5. The fracturesurface of the analysed wires of both patients is normalto the wire axis and the crack morphology is un-branched

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0 1007

Tabl

e1

–A

mou

nt

ofal

loyi

ng

elem

ents

inth

est

ain

less

stee

l(in

wt%

,bal

ance

Fe).

Sam

ple

Cr

Ni

Mo

Mn

CSi

PS

Cu

N

Retriev

edwire

16.55

±0.15

13.94

±0.19

2.20

±0.01

1.48

±0.01

0.02

8±

0.00

50.36

±0.03

0.08

±0.00

50.00

4±

0.00

10.28

±0.05

0.06

±0.02

As-rece

ived

wire

17.10

±0.15

13.20

±0.23

2.51

±0.01

1.38

±0.01

0.01

3±

0.00

70.42

±0.05

0.07

±0.00

30.00

6±

0.00

10.32

±0.06

0.04

±0.02

AST

Msp

ecifica

tion

17–1

913

–15.5

2–3

2a0.03

a0.75

a0.02

5a0.01

0a0.50

a0.1a

aMax

imum

content.

1008 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0

Fig. 4 – Blocks of parallel fatigue striations on fracturesurface of wire retrieved from Patient A (a) and fromPatient B (b).

and transgranular, Fig. 5. The fatigue fracture occupies 95%and 80% of the cross section of the wires analysed fromPatients A and B, respectively (Fig. 3), indicating that theclosure of Patient A was subjected to a lower load than thatof the Patient B. No macroscopic damage, severe pitting, orgeneralised corrosion was appreciated on the wire surfaceand fracture surface in the analysed samples. Fig. 6 showsthe calculated σHw, σBw, σeff, and FOS values for the aboveconsidered physiological actions. It can be observed thatthe values of σBw and σHw stress components are similar.Moreover, regardless of the physiological action, σeff is lowerthan YS and thus it follows that cyclic plastic strain would notoccur during sternal consolidation.

4. Discussion

There is scarce information in the open literature regardingthe structural response of the sternal closure followingsternotomy and no accurate conclusions can be reached onthe analysis of sternal dehiscence cases via wire fractureor cutting through bone. In this study two cases of sternaldehiscence via wire fracture of perichronal loops wereanalysed. On the basis of this analysis the structural responseof the perichronal closure is proposed, as well as a possiblecause for the observation that in the case of single wireclosure dehiscence via cutting through bone occurs moreoften than via wire fracture.

The chemical composition, microstructure (grain size andinclusion content), and mechanical properties (UTS, El) of theretrieved wires were found to be in agreement with standardrequirements for stainless steel surgical fixation wires. Since

Fig. 5 – Features of the profile of the fracture surface in wire retrieved from Patient A. Magnification X200 (a) and X500 (b).

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0 1009

Fig. 6 – Calculated values of hoop (σHw), bending (σBw),effective (σeff), and FOS for different physiological actions.

no mechanical damage was caused in the course of thefabrication and/or during surgical operation, the failure of theretrieved wires cannot be attributed to the wire material.

Using the results of the fractographic study it is possibleto establish the fracture mechanism and the sequence ofevents that have occurred for the failure of the perichronalclosure. The fracture initiated at one point or at twodiametrally opposed points on the wire surface. The fracturepropagates across the cross section in the anterior–posteriordirection. The striation throughout most of the fracturesurface indicates a fatigue process, presumably, enhancedby corrosion due to the presence of the low pH chloridecontaining the body fluids surrounding a fresh wound(Magnin and Lardon, 1985). It seems that the wire fractureoccurred due to bending stresses or a combination of tensileand bending stresses. However, whereas in Patient A thebending component was of fluctuating type, in Patient B itwas of alternating type. Furthermore, the alternating loadingmode was un-symmetrical, because the main cracks weredifferent in size (Fig. 3(b)). The presence of secondary cracksparallel to the main fatigue crack also reveals the existenceof an important bending stress component in the effectivestress.

Taking into account that inspirations and expirationsduring normal breathing take place about 12 times a minuteand that the time to dehiscence was 7 days, the numberof cycles elapsed before the wire fracture would be about7 × 104. However, the wire failure at the loops is unexpected,because for normal breathing the effective stress level onthe wire is about 11 MPa and the fatigue endurance limitof the material for 107 cycles is slightly higher than the YS(NRIM, 1979). The difference between the elapsed cycles tofailure estimated above and those of fatigue tests could beattributed to the effect of corrosive environment (Magninand Lardon, 1985) just in the case of the magnitude of allstress cycles corresponds to that of the maximum stress(330 MPa) of strong coughing. It is also noteworthy that thegreat similitude in the fracture features of the retrieved wireof Patient B (Fig. 3(b)) in which presumably the effective stressis lower than YS and those of fatigue specimens (Fig. 7)

Fig. 7 – Features of the fracture surface at highmagnifications of unused wire after 92 000 cycles ofalternative bending stresses. The arrows indicate thefracture initiation points.

Fig. 8 – Schematic diagram of the effect of the intrathoracicpressure on the bending and shear stresses induced in thewire and sternum respectively of the perichronal loop (a)and single loop (b). The open arrows indicate the locationof maximum shear regions in the perichronal loop; thebold arrows indicate the intrathoracic pressure; and theshaded areas in the single loop indicate the shear regions.

in which the stress on the wire surface is slightly higherthan YS. Therefore some factors that have not been takeninto account could have contributed to the wire fracture.Because of the conical cylinder geometry of the thoracic cageit can be asserted the forces are not evenly distributed onthe N wires (McGregor et al., 1999). Effectively, the differentwire loops of the sternal closure were not at equal risk offracture. As could be deduced from Eq. (1), the wire loopson the lower half sternum are subjected to more stress thanthe upper half sternum, because of the greater transverseand anterior–posterior dimensions of the lower versus upperthoracic cage. On the other hand, ribs 7 through 10 allattack the seventh costal cartilage causing a concentration offorces at the wire loops on the lower sternum. In reality, themaximum separation and stress concentration occur at thewire loop placed on the xyphoid region (Bruhin et al., 2005).Further studies on the stress state of the sternal closure forthe different physiological actions are required to evaluate indepth the mechanics of the sternal dehiscence.

Fig. 8 schematises the effect of an intra-thoracic pressureon the perichronal (Fig. 8(a)) and peristernal single (Fig. 8(b))closures. From the observation of this figure the following canbe noted:

– In both closures, the maximum bending stress occurs atthe transverse section of the wire on the sternotomy line.

1010 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 4 ( 2 0 1 1 ) 1 0 0 4 – 1 0 1 0

The magnitude of the bending stress component is higherin perichronal closure than in single closure because inthe latter the closure hinders the anterior opening of thesternum halves. Therefore the dehiscence risk via wirefracture using perichronal closures is higher than whenusing single closures. Taking into account that σHw ∝ 1/d2w(Eq. (1)) and σBw ∝ 1/d3w (Eq. (3)) the dehiscence risk ofthe perichronal closure via wire fracture can be reducedby increasing the wire diameter. Since the surgeon needsa flexible wire to make the loop, a 0.9 mm wire diametercould result an appropriate selection. This increase (15%)in the wire diameter would produce a decrease of about25% in the effective stress.

– In the case of the perichronal closure the shear stress ismaximum on the faces of the ribs supporting the closure(Fig. 8(a)), whereas in the single closure it is situated at theanterior regions of the sternum adjacent to the sternotomyline and to a lesser extent on the posterior regions of thesternumnear the ribs (Fig. 8(b)). Comparatively, the highestshear stress occurs in the anterior region of single loop dueto the stress concentration at the corners. Therefore thedehiscence risk via cutting through the bone using singleloops is higher than using perichronal loops.

5. Conclusions

It was found that sternal wires of perichronal loops fracturedat the anterior posterior direction by a high-cycle fatiguemechanism.

The hoop and bending stress components induced by theintrathoracic pressure are similar in magnitude.

For the physiological activities considered the effectivestress on the wire is lower than the YS.

To reduce the risk of the wire fracture, an increase of0.12 mm in the standard diameter is proposed.

Acknowledgement

The authors would like to thank Dr. T. De Cock for help withmanuscript revision.

R E F E R E N C E S

ASTM E112-96, 2004. Standard Test Methods for DeterminingAverage Grain Size. American Society for Testing of Materials,West Conshohocken, PA.

ASTM E45-97, 2005. Standard Test Methods for Determining theInclusion Content of Steel. American Society for Testing ofMaterials, West Conshohocken, PA.

ASTM E352-93, 2006. Chemical Analysis of Tool Steels and OtherSimilar Medium and High-Alloy Steels. American Society forTesting of Material, West Conshohocken, PA.

ASTM F1350-91, 2008. Standard Specification for Wrought 18Chromium-14 Nickel-2.5 Molybdenum Stainless Steel SurgicalFixation Wire. American Society for Testing of Materials, WestConshohocken, PA.

Brantigan, C.O., Brown, R.K., Brantigan, O.C., 1979. The brokenwire suture. Am. Surgeon 38–41.

Bruhin, R., Stock, U.A., Drücker, J.-P., Azhari, T., Wippermann, J.,Albes, M., Hintze, D., Eckardt, S., Könke, C., Wahlers, T.,2005. Numerical simulation techniques to study the structuralresponse of the human chest following median sternotomy.Ann. Thorac. Surg. 80, 623–630.

Casha, A.R., Yang, L., Kay, P.H., Saleh, M., Cooper, G.J.,1999. A biomechanical study of median sternotomy closuretechniques. Eur. J. Cardiothorac. Surg. 15, 365–369.

Gere, J.M., 2002a. Thimoshenko Resistencia de Materiales, 5aed. Internacional Thomson Editores Spain, Paraninfo, S.A,pp. 564–567.

Gere, J.M., 2002b. Thimoshenko Resistencia de Materiales, 5aed. Internacional Thomson Editores Spain, Paraninfo, S.A,pp. 289–290.

Magnin, T., Lardon, J.M., 1985. The influence of a 3.5 NaCl solutionon the fatigue damage evolution in a planar slip fcc stainlesssteel. Mater. Sci. Eng. 76, L7–L10.

McGregor, W.E., Trumble, D.R., Magovern, J.A., 1999. Mechanicalanalysis of midline sternotomy wound closure. J. Thorac.Cardiovasc. Surg. 117, 1144–1150.

NRIM Fatigue Data Sheet, 1979. Data Sheets on elevated—temperature high-cycle and low-cycle fatigue properties ofSUS 316—HP (18Cr-12Ni-2Mo) hot rolled stainless steel plate.No. 15. National Research Institute for Metals, Tokyo, Japan.

Ratan, V., 1993. Handbook of Human Physiology, 7th ed.In: Respiratory System, Jaypee Brothers Medical PublishersLtd., New Delhi, India, 99-119.

Sharpey-Schafer, E.P., 1953. The mechanism of syncope aftercoughing. Br. Med. J. 860–863.

Shih, C.M., Su, Y.Y., Lin, S.J., Shih, C.C., 2005. Failure analysis ofexplanted sternal wires. Biomaterials 26, 2053–2059.

Tamizawa, Y., Hanawa, T., Kuroda, D., Nishida, H., Endo, M., 2006.Corrosion of stainless steel wire alter long-term implantation.J. Artif. Organs 9, 61–66.

Voces-Sánchez, R., 2009. Nueva Técnica de Cierre Posester-notomía. Estudio Clinico y Experimental. Ph.D. Thesis. Depar-tamento de Cirugía de la Universidad de Medicina de Sala-manca, Spain.

Wilkinson, G.A.L., Clarke, D.B., 1998. Median sternotomy dehis-cence: a modified wire suture closure. Eur. J. Cardiothorac.Surg. 2, 287–290.