Percutaneous Occlusion Of The Left Atrial Appendage:Instruments Development And Experimental Study

Background： Percutaneous left atrial appendage(LAA) occlusion for strokeprevention in patients with nonvalvular atrial fibrillation (NVAF) has emerged as aninnovative valid procedure which issimple, shorter recovery time and minimallyinvasive. With increased operator experience, there is a significant improvement inthe safety of percutaneous LAA closure. Accumulating randomized trials andregistries with longer term follow up continue to support a role for LAA occlusion asan alternative to anticoagulation in carefully selected patient populations. With anaging population of our society, the prevalence of AF is likely to rapid increase.AF-related stroke is particularly catastrophic resulting in the worst prognosis amongthe various causes of stroke which would impose a heavy burden on healthcare. In ourcountry, the development of LAA occluder had an early start and made slow progress.In order to promote the application of this technique and overcome the limitationsincluding need for relatively larger delivery sheaths and limited recapture andrepositioning capabilities. We cooperate with the Shanghai Push Medical DeviceTechnology Co. LTD. to develop a novel self-expanding LAA occluder and itsdelivery system. The feasibility, safety, and healing characteristics of the newocclusion system was evaluated in canine.Objective：In this study, we aimed at evaluating the feasibility and safety ofpercutaneous implantation of the new homemade occluder in a healthy caninemodel.Methods：1、Applied anatomy ofTranscatheter Occlusion of the Left AtrialAppendage：（1）To study various morphologic parameters of the left atrialappendage (LAA) by muti-slice computed tomography（MSCT）:A workstation(TOSHIBA Vitrea Core) was used for analysis. The diameters of LAA ostium weremeasured by2D oblique method using multiplanar reformations（MPR）, then LAAostial plane was marked. Three-dimension structures of the left atrium(LA), andLAA were constructed using the volume rendered(VR) postprocessing technique.The angle of the first LAA bend, the distance from the first bend to the LAA orifice,and the approximate distance from the assumed puncture site at the septum to theLAA orifice were studied in3D images. LAA morphology was classified accordingto the shape of LAA. The angle of the first LAA bend formed by the primary lobewas measured along the central axis of the primary lobe. The distance from the left superior pulmonary vein (LSPV) orifice and from the mitral valve to the LAA orificewas measured in endocardiac view of3D images. The assumed puncture site waslocatedin axial plane, horizontal plane and sagittal plane.The distance from thepuncture site to the left atrial appendage ostium was measured, and the angle ofconnection from the puncture site to the center of the marked LAA ostial plane withthe horizontal was measured. LAA was taken off from the LA and was rotated toobtain the shape of LAA orifice.Fourteen fresh healthy canine hearts were dissected.The long diameter, short diameter and the depth of LAA were measured.The distancefrom the LAA orifice to the mitral valve and LSPV were also measured.2.Establishment of a method for percutaneous LAA occlusion in a canine model.Transseptal puncture was performed via femoral vein approach under fluoroscopicand angiographic guidance in12dogs. The diameters of the neck of LAA weremeasured from LAA angiogram in right anterior oblique (RAO) cranial(CRA), rightanterior oblique (RAO) and right anterior oblique (RAO) caudal (CAU) projection.Amethod of placing the delivery sheath was established to implanting the novel LAAoccluder. Five dogs were sacrificed immediately after procedure. The hearts weregrossly examined.3.Development of a novel LAA occlusion system: Wecooperated with the Shanghai Push Medical Device Technology Co. LTD to developa new double disk bowl-like LAA occluder and a delivery cable with a flexible tip.The angle and length of delivery sheath of transcatheter closure of ventricular septaldefect(VSD) were modified to suit the morphology of animal LAA. The bench testand test of implanting the occluder into10LAAs of dog in vitro were developed toevaluate the feasibility of new device.4. Experimental study of percutaneousLAA occlusion：Twelve healthy dogs received novel LAA occluders. The devicewas delivered by the modified sheath and delivery cable with a flexible tip.Afterprocedure all dogs received clopidogrel (75mg/day)4-week and aspirin (75mgmg/day) till the end of the experiment. Three days after procedureelectrocardiogram(ECG) and fluoroscopy were performed before animals weresacrificed. The other dogs were followed up by ECG, transesophagealechocardiography(TEE) and fluoroscopy30days or60days after procedure. Theywere sacrificed in groups at Days3(n=2),30(n=4) and90(n=4) and hearts wereharvested for gross examination or histological evaluation.Results:1. From November2012to March2013,30consecutive patients（age60±8.9years，73%were male） with AF undergoing MSCT angiography were selected for the study. Of30patients,6patients had paroxysmal AF,8had persistentAF and16had permanent AF. The general morphology of LAA was classified into4types including:Windsock type (47%), Cauliflower type(30%),Chicken wing(17%)and Cactus(6%).The angle of the first bend was110±12.8°. The distance from thefirst bend to the LAA orifice was15.2±3.1mm. The width of LAA ridge from LSPVto LAA was4.4（4.2，5.0）mm, the distance from LAA to MV was18.3±1.3mm.Diameters of the LAA ostium in the long axis and the short axis were24.6±3.1mm,15.9±2.4mm respectively. The distance from the puncture site to the left atrialappendage ostium was53.4（48.9，58.9）mm, and the angle of connection from thepuncture site to the center of LAA ostial plane with the horizontal was119.2±8.9°.The shape of the LAA orifice was classified into4typesincluding:oval(66.7%),triangular (16.7%), water drop-like (13.3%), and foot-like(6.7%). The long diameter, short diameter and the depth of LAA of dog were14.6±2.3mm,6.8±1.1mm and10±1.2mm respectively. The distance from the LAAorifice to the mitral valve and LSPV of dog were6.4±0.9mm and7.1±0.8mm.2.Experimental animals supine fixed before procedure, puncture site at the septum waslocated under fluoroscopic guidance from fixed bony landmark. Success rate oftransseptal puncture was67%. Pericardial tamponade occurred in one dog because ofinjure of LA. The neck of LAA was visualized in RAO cranial, and the diameter ofLAA measured in RAO cranial was16.4±2.7mm. The delivery sheath was safelyadvanced into the LAA using pigtail catheter. The fluoroscopy time was10±1.4minute and the procedure time was63minute. Gross pathological examinationsshowed puncture site was almost at the center of fossa ovalis.3. we developed a newdouble disk bowl-like LAA occluder with6bars and a delivery system with a joint atthe tip of the delivery cable to allow optimal assessment of final device position. Thelength and angle of delivery sheath were90cm and110-120°to accommodate thevariation of LAA anatomy. The device was complete collapse and repositioning.However a few bars would change direction when the occluder was completecollapse more than3times which made it unable to be used again. Feasibility of thenew occluder was confirmed in vitro.4. Percutaneous occlusion of LAA wassuccessful in ten of12dogs. One dog died during the procedure due to acutemyocardial infarction(AMI), other dog died due to device dislodgment the day afterprocedure. No other postoperative events were noted following the initial implantprocedure in other animals. The fluoroscopy time was22.9±2.3minute and the procedure time was112.8±16.6minute. ECG was normal and fluoroscopy showedadequate position before sacrifice. TEE showed successful occlusion of theappendages without obvious peri-device leak and trivial mitral regurgitationoccurred in2dog. Gross pathological examinations showed that the disk of theimplant occluded LAA ostium. There was complete neointimal coverage and pannusformation in dogs that been followed up day≥1month.The organized neointimaspread across the device-left atrial interfaces to seal the LAA openings completely.Both gross and light microscopic examinations of the major organs showed noevidence of infarct.Conclusion: The neck of LAA with less morphologic variation is an ideal positionto implant the occluder. The morphology of LAA in canine is similarity to that of thehuman LAA which makes the canine is chosen as a canine model. The new doubledisk bowl-like LAA occluder is feasible in vitro. The new homemade LAA occlusiondevice is feasible and safe in a canine model in vivo. The delivery cable withdirection-adjustable tip allows optimal assessment of final device position. At30or90days follow up, all occluders display complete healing and occlusion of the LAAwith biocompatibility by transesophageal echocardiography. There is no adverseevent.