外文翻译--对于梗死后心室间隔破裂的心内膜片修补大小刚度和接触条件的有限元分析 英文版

ORIGINAL ARTICLEFinite element analysis regarding patch size, stiffness, and contactcondition to the endocardium in surgery for post infarctionventricular septal ruptureToshiaki ItoHiroaki HagiwaraAtsuo MaekawaTakenori YamazakiReceived: 17 February 2013 / Accepted: 15 April 2013 / Published online: 23 April 2013C211 The Japanese Association for Thoracic Surgery 2013AbstractBackground The purpose of this study is to establish arational surgical design to minimize suture line stress of thepatch and thus prevent residual leakage in surgery for postinfarction ventricular septal rupture (VSR).Methods A computer model that simulates surgery forVSR was developed. Stress force on the suture line of thepatch was calculated at varying size, stiffness, and contactcondition of the endocardial patch to the inner surface ofthe heart using a finite element analysis. Clinical resultsand echo findings of 34 consecutive patients with a meanage of 72.6 9.5 (range 5589) who underwent emer-gency surgery for VSR from 1995 to 2012 were reviewed.Results Suture line stress decreased by two-thirds as thesize or stiffness of the patch increased in comparison withthe basic conditions that mimic a pericardial patch fitted tothe septal plane. On the other hand, suture line stressincreased sixfold when there was a dead space beneath thepatch. 30-day mortality was 12 %, and in-hospital mor-tality 18 %. On echocardiography, all three patients whohad dead space beneath the patch had residual leak.Another patient who had huge posterior defect also showedresidual leak. Clinical results were well matched to modelresults. 5-year survival rate of all patients who receivedoperation was 68.7 9.3 %.Conclusion In endocardial patch type surgery for VSR,proper sizing of the patch to sufficiently fit to endocardialsurface in a tension-free condition is the most important toavoid residual leak.Keywords Myocardial infarction C1 Ventricular septalrupture C1 Computer applicationsIntroductionPostoperative residual leak after patch closure for post-infarction ventricular septal rupture (VSR) is a strongfactor of poor outcome 1, 2. Dehiscence of the patch iscaused by fragility of myocardium acutely after myocardialinfarction. Therefore, some investigators have proposeddelaying operation until the tissues surrounding the septaldefect have become durable owing to scar formation 3.However, critically ill patients cannot always wait until thechronic stage due to ongoing multiple organ failure.Maltais et al. 2 recently reported that time delay betweenVSR diagnosis and surgery was a significant predictor ofpoor outcome.To avoid detachment of the patch in emergency surgeryfor VSR, several operative techniques have been proposed.David et al. 4 focused on enlargement of the patch thatexcludes all infarcted areas. Others concentrated on rein-forcement of the suture line using transmural interruptedsutures 1, 5, 6. Regarding patch material, xenopericardium,Teflon felt, a combination of both 7, and further rein-forcement using biological glues 7, 8 have been reported.The purpose of this study is to estimate the change insuture line stress, which may lead to patch dehiscence inT. Ito (&) C1 A. MaekawaDepartment of Cardiovascular Surgery,Japanese Red Cross Nagoya First Hospital, 3-35 Michishita-cho,Nakamura, Nagoya 453-8511, Japane-mail: cvs1nagoya-1st.jrc.or.jpH. HagiwaraDepartment of Cardiovascular Surgery, Kasugai MunicipalHospital, 1-1-1, Takaki-cho, Kasugai 486-8510, JapanT. YamazakiDepartment of Cardiovascular Surgery, Kainan Hospital,396 Minamihonda, Maekasu-cho, Yatomi 498-8502, Japan123Gen Thorac Cardiovasc Surg (2013) 61:632639DOI 10.1007/s11748-013-0255-zVSR surgery, according to various changes in patchproperties using a computer model, and analyzing theclinical results based on the theoretical model.Patients and methodsThis study was approved by Institutional Ethics Commit-tee, and written informed consent for surgery was obtainedfrom each patient. Consent for data collection and utiliza-tion was obtained in written form or by telephone fromeach patient or a member of his or her family.Theoretical model analysisTo simulate a left ventricle with septal perforation, a sim-plified computer model was developed. A sphere with a 6 cminternal diameter and 1 cm wall thickness was assumed forthe left ventricle. Septal defect was represented as a rounddefect, 1.4 cm in diameter, on the wall of the sphere. Thismodel was divided into tetrahedron meshes for finite elementanalysis (FEA). Solid Works Simulation (Solid Works,Tokyo, Japan) was used for FEA. As a material used for FEAcalculation, a sphere made of soft rubber is assumed (Youngsratio = 2 9 106/m2, Poissons ratio = 0.49).As basic conditions, a concave-shaped rubber (Youngsratio = 2 9 106/m2, Poissons ratio = 0.49), 0.25 mmthickness was attached inside the sphere at a circular lineaway from the center of the defect by a 40C176 central angle(Fig. 1). The membrane is fitted to the inner surface of thesphere before inner pressure was applied. This modelrepresents a xenopericardial endocardial patch that widelycovers the septal plain. The membrane and sphere wereconnected only at the edge of the membrane. Frictionbetween the membrane and the internal surface of thesphere was assumed to be zero because the contact con-dition in a beating heart is not static as in a model condi-tion, and friction seemed to be negligible. The width of theattachment line was defined as 0.5 mm. Suture line stressand deformity of the patch were calculated for each con-dition under 120 mmHg of internal pressure. Since sutureline disruption usually occurs on the endocardial surface,not on the patch in clinical settings, stress force on thesuture line was calculated on the inner surface of thesphere. Because wall stress usually has a direction and avalue, von Misses method was used to express average wallstress as an absolute value.Effect of patch size on suture line stressOnly the size of the membrane representing the pericardialpatch was changed from the above-mentioned basic con-ditions. For a model of a small patch that only coversaround the defect, the central angle between the center ofthe defect and the circular suture line was reduced to 30C176.Then, for a model of a large patch that covers wider area ofendocardium as infarct exclusion technique, the centralangle of the circular suture line was increased to 60C176.Effect of stiffness of the patchTo mimic secondary reinforcement of the pericardial patchwith Teflon felt to increase the stiffness of the patch, a sec-ondary patch made of solid rubber (Youngs ratio = 2 9107/m2, Poissons ratio = 0.49), 1 mm in thickness wasattached underneath the soft rubber patch under the basicconditions. A secondary patch was firmly attached to theprimary patch, but not to the ventricular sphere.Effect of contact condition between the patchand the internal surface of the left ventricleWhen the pericardial patch is tailored too small, and doesnot fit well to the concave endocardial surface of the leftventricle, stress on the suture line may increase. To mimicthis situation, a model in which a plain rubber membraneinstead of concavely shaped one was attached to a circularline 40C176 apart from the center of the VSR.Effect of defect size on suture line stressTo mimic a very large septal defect or surgically enlargeddefect by infarctectomy, a defect diameter of 2.5 cm wasFig. 1 Tetrahedron mesh image of the left ventricular model and theendocardial patch at static basic conditions. A shaped patch along theinner surface of the sphere is attached at a circular line with a 40C176central angle. The center of the patch is aligned to the center of thedefect. Diameter of the sphere is 6 cm, wall thickness is 1 cm, anddiameter of the defect is 14 mmGen Thorac Cardiovasc Surg (2013) 61:632639 633123set. Other conditions were the same as the basic conditions.Suture line stress was also calculated when a secondaryreinforcement patch (Youngs ratio = 8 9 107/m2, Pois-sons ratio = 0.49), 1 mm in thickness, was added to thissituation.Clinical evaluationBetween 1995 and 2012, 34 patients received an emer-gency operation for post infarction VSR at our facilities.Clinical profiles of the patients are listed in Table 1. Allpatients were referred and were operated on an emergencybasis on the day of admission or as soon as diagnosis wasconfirmed. Five different surgeons performed the opera-tions. We did not postpone operation waiting for fibrosis ofinfarcted tissue until chronic stage, and operated on allpatients referred to our facilities without any arbitraryselection criteria, even if the patient was in salvage status.Two patients had been accompanying left ventricular free-wall rupture. Three patients were already in chronic state(14 days after VSR) when referred from cardiologistsduring this study period and were excluded, becausechronic VSR differs greatly from the acute state, andusually is not difficult to handle.Data were collected retrospectively. Hospital morbidityand mortality were obtained from the hospital records.Current status and late complications of hospital survivorswere ascertained by telephone or inquiry form contact withthe patient or a member of his/her family. Survivalinformation was obtained between March and June 2012.Follow-up rate was 97 % with a mean period of 44 months.Findings of postoperative echocardiography were evalu-ated with special interest given to patch size, use of asecondary patch, contact conditions of the patch to theendocardium, and residual leakage. If significant residualleakage was suspected on echocardiography, shunt ratiowas calculated by oximetry.Surgical techniqueWe used three different types of surgical techniques duringthis period: infarct exclusion technique with bovine orequine pericardium (Edwards Lifesciences, Irvine, CA),septal patch with xeno-pericardium, and xeno-pericardialseptal patch plus a secondary reinforcement patch of Teflonfelt. Details of our septal patch technique (entire septalpatch technique) and its refinement have already beenreported 5, 6. In recent patients, after suturing a largepericardial patch that almost entirely covers the ventricularseptum, a Teflon felt patch tailored smaller than the peri-cardial patch was inserted beneath and fixed to the peri-cardial patch with fibrin glue and several interruptedsutures. This secondary felt patch was intended to preventdeformity and protrusion of the soft pericardial patch intothe septal defect that may lead to undue stress on the sutureline. This secondary patch is fixed to the primary patch, butnot to the ventricular septum.Statistical analysisAll statistical data analyses were performed using SPSSstatistics 17.0 (Japan IBM, Tokyo, Japan). Actuarial sur-vival rate was calculated using the KaplanMeier methodand comparison between groups was evaluated using log-rank test.ResultsTheoretical model analysisDeformity of each element in the basic conditions wascalculated, and shown in Fig. 2a. A thin rubber patch alongthe inner surface of the sphere protruded at the defect by4.6 mm when 120 mmHg of inner pressure was applied.The color bar shows the degree of deformity. Each elementof the sphere received stress force as shown in Fig. 2b. Thecolor bar shows stress force calculated using the vonMisses method. Only the sphere was depicted in Fig. 2b,and the membrane was not, because suture line stress wascalculated on the sphere side. Elements along the sutureline of the patch received high stress force as expressed byTable 1 Preoperative clinical profile of patients (n = 34)VariablesAge (years) 72.6 0.5 (5589)Male/female 11/23Diabetes mellitus 8Hypertension 14Previous MI 4Thrombolysis/direct PCI 1/131/2/3-Vessel disease 30/4/0Anterior/posterior VSR 25/9AMI to VSR (days) 2.9 2.5 (0.310)VSR to surgery (days) 1.6 2.2 (0.213)Preoperative IABP 32Cardiogenic shock 22Endotracheal intubation 8Inotropic support 32MOF 6Qp/Qs3.4 1.0 (1.56.1)AMI acute myocardial infarction, IABP intra aortic balloon pumping,MI myocardial infarction, MOF multiple organ failure, PCI percuta-neous coronary intervention, Qp/Qspulmonary to systemic flow ratio,VSR ventricular septal rupture634 Gen Thorac Cardiovasc Surg (2013) 61:632639123the color red. Average suture line stress in the basic settingwas calculated as 1.60 9 105N/m2.Effect of the size of the patch on suture line stressSuture-line stress for a small size patch was 1.99 9 105,and 1.05 9 105N/m2for a large size patch. Suture-linestress decreased as the size of the patch increased (Fig. 3).Effect of stiffness of the patchWhen a secondary buttressing patch was attached to thethin primary patch under the basic conditions, suture linestress decreased to 1.15 9 105N/m2. Deformity of thepatch was down-regulated to 1.4 mm when a secondarypatch was added compared to that without buttressing(Fig. 4).Effect of contact condition between the patchand the internal surface of the left ventricleWhen the patch was flat and did not contact the innersurface of the sphere, calculated average suture line stresswas 1.05 9 106N/m2, almost six times higher than thatunder basic conditions (Fig. 5).Effect of an enlarged defectWhen defect size was enlarged from 1.4 to 2.5 cm indiameter, suture line stress increased to 4.0 9 105N/m2.By adding a secondary reinforcement patch under the pri-mary patch in these conditions, suture line stress decreasedto 2.5 9 105N/m2, and deformity of the patch apparentlydecreased.Clinical resultsThirty-day and in-hospital mortality of all patients whoreceived an operation was 11.8 (4/34) and 17.6 % (6/34),respectively. Average cardiac ischemic time was 69 (range29120), cardio-pulmonary bypass time 163 (range82504), and operative time 270 (range 145600) min.Four patients were operated on without aortic crossclamping. In three patients, left ventricular closure-linebleeding which was difficult to control occurred and led toearly deaths. Of these, two underwent infarct exclusion,and one septal patch operation. Another patient died intra-operatively of severe heart failure as the family membersdeclined to place the patient on assisted perfusion. Residualshunt could not be evaluated in these four patients. Type ofsurgery, location of VSR, in-hospital deaths, and presenceof residual leak are summarized in Table 2.Significant residual leak occurred in four patients.Among these, two patients were operated on with infarctexclusion for anterior VSR. Pulmonary to systemic flowratio (Qp/Qs) using oximetry was 1.18 and 1.35, respec-tively. According to echocardiography, the pericardialpatch did not fit tightly to the left ventricular endocardialsurface in these patients. Residual shunt flow passedthrough the free space between the pericardial patch andthe endocardium, and into the septal defect. Reoperationwas not required because they were clinically stable, butFig. 2 a Deformity of elements when 120 mmHg of inner pressure isapplied to the basic conditions model. Color bar shows amount ofdeformity. Protrusion of the center of the patch facing the septaldefect is 4.6 mm. The sphere is also slightly deformed. b Stress toeach element when 120 mmHg of inner pressure was applied to thebasic conditions model. Stress force calculated using the von Missesmethod increases as the color changes from blue to red. Elementsalong the circular suture line of the patch receive high stress forceGen Thorac Cardiovasc Surg (2013) 61:632639 635123later two patients needed reoperation. A 78-year-oldwoman who underwent pericardial septal patch operationfor anterior VSR had residual leakage with a Qp/Qsratio1.8. Operation was performed on this patient with the heartbeating due to severe aortic calcification. According toechocardiography, there was a free space beneath thepatch, presumably because the patch was trimmed duringempty heart beating and was too small for a blood-filledheart. The patient died in the hospital as a result of braininfarction even though the residual shunt disappeared afterreoperation. A 56-year-old man had very large (2 9 5cmin size, Qp/Qs= 4) posterior VSR. He underwent septalpatch operation using a bovine pericardium. Postoperativeechocardiography showed protrusion of pericardial patchinto septal defect toward the right ventricle, and significantresidual leak with partial patch dehiscence. There was noapparent free space between the patch and the septal plane.At reoperation, the disrupted suture-line was reinforcedwith several felt buttressed sutures. Residual shunt disap-peared and the patient was discharged. However, 5 monthsafter initial operation, the patient died of congestive heartfailure. In recent consecutive 18 patients in whom sec-ondary Teflon felt patch was added to septal pericardialpatch, 17 patients survived the operation and were dis-charged from hospital without significant residual leak.There was no apparent free space between the patch andthe left ventricular inner surface in these cases. Actuarialsurvival rate at 5 and 10 years of all the patients (includingin-hospital deaths) was 68.7 9.3 and 61.9 10.6 %(mean SE) for all causes of death. Survival rates did notsignificantly differ depending on rupture site (anterior orposterior), or the presence or absence of additional CABGby log-rank test. Late recurrence of shunt was not observedin any patient.DiscussionSignificant residual shunt is a strong predictor of pooroutcome after emergency surgery for VSR 1, 2. Mu-rashita et al. 1 reported that 83 % of patients with residualshunt after surgery for VSR resulted in hospital mortality.To prevent patch dehiscence, two different approaches canbe considered. One is to reinforce the suture line, and theother is to devise a surgical design to decrease suture linestress. In the clinical setting, reinforcement of the suture-line has some limitations even using felt buttressed inter-rupted sutures 1, 5 because the non-infarcted myocardiumFig. 3 Images of suture line stress in a large and a small patch Suture line stress of a large patch was about half compared to that of a smallpatchFig. 4 Deformity of elements when a secondary buttressing patchwas added. Deformity of the patch decreased to 1.4 mm636 Gen Thorac Cardiovasc Surg (2013) 61:632639123is sometimes fragile in patients suffering from VSR.Therefore, in this study, we focused on a surgical design todecrease suture line stress.In the computer model with FEA, when a large patchwas sutured to the line well apart from the edge of septalrupture, stress on the patch suture line decreased about35 % compared to that of the basic conditions which rep-resented septal exclusion patch. On the other hand, sutureline stress increased by 25 % when a small patch wassutured near the edge of the defect. Suture line stress maysimply be decreased, being distributed to longer suture linewhen a larger patch is used. These results imply that aninfarct exclusion technique 4 where a large pericardialpatch is sutured onto healthy myocardium well apart fromVSR is superior. However, it is sometimes difficult to tailorone piece of plain xeno-pericardial patch into a three-dimensional shape that well fits to the concave left-ven-tricular inner surface including cardiac apex. If the patchdoes not fit well to the endocardial surface, and free spaceexists between the patch and the endocardium, left ven-tricular inner pressure is directly applied to the suture lineof the patch. The average suture line stress was six timeshigher when the patch lifted off the endocardial surfacethan when the patch fit as in the conditions we set in thecomputer model. Although increasing the size of the patchis theoretically beneficial, it may become more difficult toalign the patch to the three-dimensional inner surface of theleft ventricle. If a free space is left between the patch andendocardium, effect is nullified or may be even worse thanwhen a smaller patch that fits well to the endocardium isapplied. Shibata et al. 9 reported a modified infarctexclusion technique using two separated patches which arejoined to create a cup shape. Kobayashi et al. 10 reporteda technique which tailors a plain xenopericardial patch intoa three-dimensional shape. These are logical approaches toeliminate dead space underneath the patch. Several sur-geons 7, 11 have used a secondary patch around a septaldefect in addition to a pericardial patch sutured apart fromthe defect. In our theoretical models, this type of secondarypatch prevented undue deformity of the pericardial patch,and decreased suture-line stress by two-thirds: an almostidentical effect as when increasing the size of the patchfrom the basic condition to large. The secondary but-tressing patch may decrease the suture line stress like acommonly observed phenomenon as follows. A drain ofkitchen sink can be obstructed with a tiny paper when theorifice is covered with a mesh strainer. Without a strainer,the paper can easily be deformed and evacuated into thedrain. Likewise, additional Teflon buttressing patch maydecrease suture line stress by preventing undue deformityof the pericardial patch around the septal defect. ThisFig. 5 A model showing insufficient contact of the patch to theendocardial surface. a A plain patch was attached to the same sutureline under basic conditions. The model was divided into a tetrahedronmesh. b Very high stress force along the suture line of the patch isshown by the red circular bandTable 2 Surgical results according to the location of VSR and typeof surgeryVSR location Type of surgery In hospital death LeakAnterior Infarct exclusion 2/5 2/3Septal pericardial patch 3/9 1/8Buttressed septal patch 0/11 0/11Posterior Infarct exclusion 0 0Septal pericardial patch 0/2 1/2Buttressed septal patch 1/7 0/6Presence of residual leak was evaluated in 30 patients excluding 4patients with very early deaths* Numbers represent the number of patientsVSR ventricular septal ruptureGen Thorac Cardiovasc Surg (2013) 61:632639 637123mechanism should also work when the defect is coveredwith independent secondary patch 7, 11. Gluing a peri-cardial patch around the septal defect 8 may show asimilar effect as a secondary patch with the stiffening effectof a biological glue. Results of our computer model anal-ysis support the assertions of previous researchers.As size of VSR increases, suture line stress alsoincreased by 2.5 times compared to the basic condition.Therefore, a large septal defect may be a risk factor inpatch dehiscence. Similarly, resecting infarcted myocardialtissue around the septal defect may negatively affect sutureline stress. Stiffening the patch with a double layer patch orgluing may be effective when the defect is very large, orinfarctectomy is necessary in the procedures 12.Theoretical results were in good agreement with theclinical results. Four patients showed significant postoper-ative residual shunt. Of these, the existence of free spacebetween the patch and endocardium was seen in threepatients. The patch seemed to be trimmed too small inthese patients. Another patient with residual leak had alarge posterior septal defect (2 9 5 cm). Although thepericardial patch was well fixed to the left ventricular innersurface, the patch and the suture-line may have received avery large force when exposed to left ventricular pressure.Currently, our standard procedure is trans-infarct ven-triculotomy, no resection of infarcted myocardium, place-ment of a xenopericardial patch almost entirely coveringthe septal surface 5, and the addition of secondary but-tressing Teflon felt patch around the defect to increasestiffness of the patch 6. Instead of an apical half of the leftventricle as done in conventional infarct exclusion tech-niques, covering over the septal plane with a pericardialpatch without leaving dead space beneath the patch is notdifficult. In 18 recent operations of this type, 17(94 %) ofthe patients were discharged walking out of the hospitaland no major leaks were observed. It is noteworthy thatSkillington et al. 13 already have achieved a 30-daymortality of 11.1 % in 36 consecutive patients whoreceived an operation between 1987 and 1988 using a largepericardial patch that widely covered the septal wall.Study limitationsAs for the EFA model, the shape of the simulated ventriclewas a very simple and properties of the wall are assumed tobe uniform. This differs from a real left ventricle withinfarction that has a complex shape, and has both fragileand healthy portions. Of course the patch is not sutured in acompletely circular line in real surgery. Therefore, thedefinite values of suture-line stress do not have muchmeaning, but the ratios between different conditions do.Based on their own clinical experience, many surgeonsseemed to be aware of importance of enlarging patch size,prevention of dead space beneath the patch, and the addi-tion of secondary stiffening patch or biological glues.However, the numerical data about suture line stress hasnot been shown. Our models are very simple, but addedtheoretical background to these previous reports.In the computer model, we did not evaluate left-ven-tricular closure-line stress and possibility of bleeding to theoutside. Bleeding from the left ventricle is another seriousproblem in clinical settings, and needs further study.Finally, we respect the various efforts by other surgeonsto overcome this difficult-to-treat disease, and do not insistthat our current surgical method is superior to any other.However, we sincerely hope that our theoretical analysiswill be useful to surgeons seeking to refine surgery for VSR.ConclusionIn endocardial patch type surgery for post infarct