OR

IGIN

AL

AR

TIC

LE

Original Article

Percutaneous Closure of Post-MyocardialInfarction Ventricular Septal Defects:

A Single Centre ExperienceJamil Ahmed, MBBS a,∗, Peter N. Ruygrok, MD, FRACP a,

Nigel J. Wilson, MBChB, FRACP b, Mark. W.I. Webster, MBChB, FRACP a,Sally Greaves, MBChB, FRACP a and Ivor Gerber, MD, FRACP a

a Green Lane Cardiovascular Service, Auckland City Hospital, Auckland, New Zealandb Paediatric and Congenital Cardiac Service, Auckland City Hospital, Auckland, New Zealand

Background: Ventricular septal defect (VSD) is a serious complication of myocardial infarction (MI), occurring inabout 0.2% of cases. Untreated, mortality in high and early surgical repair is difficult because of friable necrotic tissue.Percutaneous closure may be an alternative treatment option in selected patients.

Methods: We report our complete single centre experience of percutaneous post-MI VSD closure using the Amplatzerdevice. The VSD was closed under general anaesthesia with fluoroscopic and transoesophageal echocardiographicguidance. Clinical characteristics and outcomes are reported.

wswuc

o

K

I

DitrvGcs9dot

R22∗AZE

©A

Results: The five patients were aged from 66 to 76 years. Closure was attempted from 1 to 64 days post-MI. VSD closureas performed in two patients after surgical patch dehiscence (infero-apical and infero-basal VSDs). The procedure was

uccessful in four patients, with failure to cross the interventricular septum in one. Thirty-day and 12-month survivalas 3/5 (60%); no patient undergoing VSD closure early post-MI survived, whereas 3/3 in whom the procedure wasndertaken >14 days post-MI survived, suggesting that it may be appropriate to delay attempts at percutaneous VSDlosure after MI.

Conclusion: Percutaneous VSD closure post-MI appears to be a viable alternative to surgical repair, with favourableutcomes if it can be undertaken >14 days post-infarct.

(Heart, Lung and Circulation 2008;17:119–123)© 2007 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society of Australia and

New Zealand. Published by Elsevier Inc. All rights reserved.

eywords. Post-infarct ventricular septal defect; Percutaneous closure; VSD closure device

ntroduction

evelopment of a ventricular septal defect (VSD) isa serious complication of transmural myocardial

nfarction (MI). It occurs in about 0.2% of MI casesreated with thrombolysis1; with an incidence that may beeducing now that primary percutaneous coronary inter-ention has replaced thrombolysis in many centres.2 In theUSTO trial 30-day mortality was 74% in those with VSDs,

ompared with 7% in those without, and those with VSDselected for surgery had a mortality of 47%, compared with4% in patients treated medically.1 Most cases occur onay 2 or 3 post-MI, but may occur as early as 24 hoursr be delayed up to two weeks. Perforation of the sep-um is usually single with a diameter that can be greater

eceived 13 August 2007; received in revised form 23 August007; accepted 1 September 2007; available online 3 December007

Corresponding author at: Green Lane Cardiovascular Service,uckland City Hospital, Private Bag 92024, Auckland 1030, Newealand. Tel.: +64 9 307 4949; fax: +64 9 307 4950.-mail address: JamilA@adhb.govt.nz (J. Ahmed).

than a centimetre. The size of the defect determines themagnitude of the left to right shunt, and subsequentdevelopment of right ventricular dysfunction, pulmonaryhypertension, congestive heart failure, low cardiac outputand subsequent impact on other organ systems. Surgi-cal repair is difficult due to friable tissue surrounding theinfarction, leading to poor suture support. Posterior VSDsare much more difficult to repair than apical VSDs. Dehis-cence of a patch is common after surgical repair3,4 andearly mortality in patients post-surgical repair remains ashigh as 45% in some series1,5–8 despite improved surgicaltechniques.

Percutaneous catheter closure using VSD occluderdevices is a promising alternative to surgical closure ofpost-infarct VSD in selected patients. We report our expe-rience of five cases.

Methods

This study reviews the first five consecutive cases ofpercutaneous post-MI VSD closure performed at Auck-

2007 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society ofustralia and New Zealand. Published by Elsevier Inc. All rights reserved.

1443-9506/04/$30.00doi:10.1016/j.hlc.2007.09.001

OR

IGIN

AL

AR

TIC

LE

120 Ahmed et al. Heart, Lung and CirculationPercutaneous Post-MI VSD closure 2008;17:119–123

Figure 1. LV angiography before closure showing left to right shunt (left hand panel) and angiographic image (right hand panel) showing theAmplatzer device in position (black arrow), before release from delivery cable (white arrow).

land City Hospital since the procedure was introducedin August 2003. We describe the patients, procedure,immediate outcome and long-term results to Decem-ber 2006. Demographic, clinical, echocardiographic andangiographic data were collected and recorded into a ded-icated database.

Procedure

The percutaneous VSD closure procedures describedbelow were all performed under general anaesthesia (GA)with transoesophageal echocardiographic (TOE) and fluo-roscopic guidance (Figs. 1 and 2). An initial left ventricular(LV) angiogram was performed to establish landmarksand guide the procedure, via the right femoral arteryapproach (Fig. 1). An 8.5 French sheath was insertedinto right internal jugular vein (RIJ) and 5000 IU intra-venous heparin and 1 g cephazolin was administered. AJR4 catheter was then advanced to the LV and manipulatedinto the mouth of the VSD, through which a Glidewirewas passed across the VSD and the catheter advancedinto the right ventricle (RV). The Glidewire was thenreplaced with a 260 cm or 400 cm Wholey wire, whichwas advanced into the pulmonary artery where it was

captured using a snare advanced from the RIJ vein. Thewire was then exteriorised resulting in a complete arterio-venous loop. The RIJ sheath was then exchanged for a12 French AGA delivery sheath, which was then care-fully advanced from venous side across the VSD andthe dilator removed. The chosen device was advancedthrough the delivery sheath and LV disk deployed. Thesheath and device were then pulled back until the LVdisk firmly effaced the LV septal wall under TOE guid-ance and the sheath retracted to deploy the waist of thedevice and then the RV disk under fluoroscopic guidance.The device position was confirmed by TOE (Fig. 2) andangiography before release (Fig. 1), and a transthoracicechocardiogram (TTE) performed the following day. Endo-carditis prophylaxis was recommended for all patientsfor six months and indefinitely for those with a residualshunt.

The Amplatzer post-infarction muscular VSD closuredevice (AGA Medical Corporation, Minnesota, MN) is aself-expandable, double-disc made from a nitinol wiremesh filled with polyester patches. The connecting waistis 10 mm long and connecting disks are 5 mm larger inradius, than the waist of the device. It is available in sizesfrom 16 mm to 24 mm with 2 mm increments.

st-infnferio

Figure 2. Transoesophageal echocardiographic images showing the pohand panel) and post-procedure showing device placement in the mid i

arct ventricular septal defect location in the mid inferior septum (leftr septum (right hand panel).

OR

IGIN

AL

AR

TIC

LEHeart, Lung and Circulation Ahmed et al. 121

2008;17:119–123 Percutaneous Post-MI VSD closure

Description of Cases

Case 1A 76-year-old male was thrombolysed with streptokinasefollowing presentation to hospital with an anterior MI.Two days later he developed a new murmur due to aninfero-apical VSD. Coronary angiography, showed severemid left anterior descending (LAD) artery and distal cir-cumflex lesions. An intra-aortic balloon pump (IABP) wasinserted and he underwent urgent coronary artery bypassgrafting (CABG) with a mammary artery graft to the LAD,a saphenous vein graft to obtuse marginal and pericardialpatch VSD repair. He was discharged home on postop-erative day 7 but was readmitted on postoperative day 49with increasing shortness of breath, hypotension and signsof right heart failure. A transoesophageal echocardiogram(TOE) showed dehiscence of the VSD repair patch, signif-icant left to right shunt and a pulmonary artery systolicpressure of 60 mmHg. The diameter of the VSD measured10–15 mm. He proceeded to percutaneous VSD closure,which was performed employing an 18 mm Amplatzermuscular VSD occluder device with a fall in pulmonarypressure to 40 mmHg. He made a steady recovery andwas discharged 13 days post-procedure with a mild resid-ual left to right shunt. Forty-one weeks post-procedure heremains asymptomatic with a small residual leak.

Case 2AtTtpoeua1bdsasa5f

CAwpwtbfiad2g

a small residual shunt. A large groin haematoma requireda 2 unit blood transfusion. She was discharged on day 3post-procedure. At 18 months follow-up, she remains wellapart from haemolysis requiring monthly blood transfu-sions, thought due to two tiny persistent shunt jets.

Case 4A 66-year-old male was admitted with chest pain andtroponin T rise to 0.9 �g/l without significant electrocar-diographic changes. Coronary angiography the followingday showed a 65% left main stem lesion and an occludedmid right coronary artery. A TTE showed severe hypoki-nesis of the basal to mid inferior wall and inferior septumand mild mitral regurgitation (MR). On day 9 he developedpersistent rest pain and inpatient CABG was planned. Amurmur thought to be due to mild mitral regurgitationwas detected on auscultation. At surgery a post-inductionTOE showed only trivial MR and an inferior septal VSD.Following considerable discussion a decision was made toproceed with the grafting only and delaying percutaneousVSD closure for approximately four weeks, provided heremained stable. Three vessel CABG was performed, fromwhich he made an uncomplicated recovery. He was dis-charged from hospital on post-operative day 8 followinga TTE that showed preserved biventricular function withno increase in VSD size or shunt.

The VSD, with a measured diameter of 18 mm, wascAhrpVp

CAtSaondistbwirha

R

Tsi

V

69-year-old woman presented with a late presenta-ion infero-posterior MI. A murmur was noted and aTE confirmed a basal inferior VSD, a dilated right ven-

ricle with impaired function and a pulmonary arteryressure of 50 mmHg. Coronary angiography revealed anccluded distal right coronary artery and mild to mod-rate disease elsewhere. An IABP was inserted and shenderwent emergency surgical repair of the VSD withpericardial and Gortex patch, without CABG. On the

0th post-operative day she became increasingly short ofreath redeveloped a systolic murmur. A TTE showedehiscence of the VSD patch with a large left to righthunt. She proceeded to percutaneous VSD closure with12 mm Amplatzer muscular VSD occluder device, with

ignificant reduction in the shunt. She failed to improvend subsequently developed multi-organ failure. On daypost-procedure, active support was withdrawn at the

amily’s request and she died soon after.

ase 375-year-old woman presented with acute inferior MI andas thrombolysed with streptokinase. Coronary angiogra-hy showed an occluded mid right coronary artery whichas stented. On day 9 post-MI she developed a new sys-

olic murmur and a TTE revealed a VSD in the inferiorasal septum with a measured diameter of 15 mm, suitable

or percutaneous closure. As she remained haemodynam-cally stable, the procedure was deferred for six weeks tollow time for the tissue to become firm and fibrotic. Onay 50 percutaneous VSD closure was performed using a0 mm Amplatzer post-infarct VSD occluder device, with aood angiographic and echocardiographic result, and only

losed percutaneously on day 53 post-MI with a 20 mmmplatzer post-infarct VSD occluder device. A large groinaematoma required a blood transfusion and vascularepair. He recovered well and was discharged on day 6ost-procedure. A pre-discharge TTE confirmed a stableSD closure device with a trivial residual shunt, which hasersisted at 9-month follow-up.

ase 575-year-old woman admitted to a peripheral hospi-

al with an inferior MI was thrombolysed with reteplase.he subsequently developed hypotension, hypoxia, anuriand runs of ventricular tachycardia. She transferred tour institution in cardiogenic shock, and underwent coro-ary angiography and insertion of an IABP. An occludedominant circumflex artery was stented, but she remained

n shock requiring noradrenaline and dopamine infu-ions. A TOE showed severe hypokinesis in the circumflexerritory and revealed a moderate-sized VSD in the infero-asal septum. On day 2 post-MI percutaneous closureas attempted but was unsuccessful due to an inabil-

ty to cross the defect with the delivery system, likelyelated to the angulation and tortuosity of the defect, andaemodynamic and rhythm instability. The procedure wasbandoned and the patient died soon after.

esults

he demographic and clinical findings of our patients areummarised in Table 1 and the outcomes of the proceduresn Table 2.

The median time between MI and the diagnosis ofSD was two days (range 1–10 days), and the median

OR

IGIN

AL

AR

TIC

LE

122 Ahmed et al. Heart, Lung and CirculationPercutaneous Post-MI VSD closure 2008;17:119–123

Table 1. Demographics and Clinical Data of Patients withPost-Infarct VSD

Demographic and Clinical Data VSD (n = 5)

Age (years) 72 (66–76)Weight (kg) 84 (68–121)Height 166 (152–182)Male 2 (40%)Female 3 (60%)Anterior MI 1 (20%)Inferior MI 4 (80%)Thrombolysis 3 (60%)Shock prior to procedure 3 (60%)IABP prior to procedure 2 (40%)Coronary angiography 5 (100%)Single vessel disease 2 (40%)Two vessel disease 2 (40%)Three vessel disease 1 (20%)Occluded RCA 3 (60%)Occluded dominant circumflex 1 (20%)Coronary intervention 2 (40%)Surgical VSD closure 2 (40%)CABG 2 (40%)Follow-up, patients alive (days) 665 (263–1248)

time between MI and percutaneous closure or attemptedclosure was 50 days (range 1–64 days). The Amplatzer post-infarct VSD closure device was used in all cases. The devicesize ranged from 12 to 20 mm and was deployed success-fully in four of five patients. The median procedure timewas 96 min (range 65–191 min) and the median fluoroscopytime 27 min (range 17–79 min). A small residual shunt per-sisted in all patients after successful device deploymentbut in no patients has this caused haemodynamic com-promise, although one has developed haemolysis.

Both patients in whom we attempted to close the VSDwithin first two weeks died. In contrast, the three who

underwent late closure have all experienced a good techni-cal and clinical outcome. Two patients in the series (Cases 1and 2) underwent percutaneous closure because of dehis-cence of a surgical patch repair, both of whom had thesurgical repair two days post-MI. The patient with earlypatch dehiscence died but the patient with late patchdehiscence is alive and well over three years later. Thesefindings indicate a better outcome if closure is delayed,either primarily or following surgical patch dehiscence.

Discussion

The first reported attempts at percutaneous closure ofpost-infarct VSDs were from Lock et al. in the 1980s.9

Unfortunately these sick patients had a persisting shuntand all died post-procedure. Since then there have beenimprovements in device technology, in particular the ded-icated Amplatzer device, and many isolated case reportsof procedural success.

We have described our first five cases of attemptedpercutaneous closure of post-MI VSDs, performed atAuckland City Hospital since August 2003. All threepatients treated more than two weeks post-MI had aprocedural success and have survived for at least sixmonths. Despite some residual shunting, there has beenno late increase in the size of the shunt or progressiveventricular dysfunction secondary to the shunt; the only

Table 2. Immediate and Long-Term Outcome of the Percutaneous P

Patient DeviceReleased

Successfully

Device Typeand Size

Residual Shunt:Immediate

Residual Shunt:Late

1 Yes Amplatzer 18 mm Small Small N

2 Yes Amplatzer 12 mm Moderate Moderate N

3 Yes Amplatzer 20 mm Very small Very small Lr

4 Yes Amplatzer 20 mm Trivial Trivial AchtrsC

5 No – – –

residual problem being haemolysis in one case. We hadno instances of device embolisation or dislodgement. Vas-cular complications were of concern in two patients, whorequired blood transfusion and vascular surgery, respec-tively, prolonging hospital stay.

There are only a small number of published case series.In the largest, a multicentre Amplatzer registry report-

rocedure

Complications:Procedure Related

Complications:Unrelated/Unknown

Outcome

il Nil Alive and well at41 months (1248days).

il Persisting shock, renalfailure.

Died at day 5post-VSD closuredue to multi-organfailure.

arge groin haematomaequiring transfusion.

Haemolytic anaemiarequiring monthlyblood transfusion.

Alive at 18 months(484 days).

trial flutter requiringardioversion. Largeaematoma requiring

ransfusion, arterial damageequiring surgical repair ofuperficial femoral artery.

Nil Alive at 9 months(263 days).

ontinued shock, runs of VT. Procedureabandonedbecause ofhaemodynamicinstability. Patientdied 2 hpost-attemptedclosure.

OR

IGIN

AL

AR

TIC

LEHeart, Lung and Circulation Ahmed et al. 123

2008;17:119–123 Percutaneous Post-MI VSD closure

ing 18 attempted procedures between 2000 and 2003, fivepatients were treated within one week of the infarct andthe remainder more than 14 days post-MI, with a mediantime from MI presentation to VSD closure of 25 days.10 Theauthors suggest that this delay may have been responsiblefor improved outcomes, through careful patient selection.Procedural success was achieved in 16/18, and 11/18 werealive at 30 days post-procedure. Of the 11 survivors twowere treated within six days of MI (40% survival) andnine were of patients treated more than 14 days afterMI (69% survival). Two patients required a second proce-dure, six out of ten patients had a small residual shuntand two out of ten patients were left with a moderateshunt. In a single centre Polish series of seven patients,procedural success was achieved in five, mostly using theAmplatzer atrial septal occluder.11 There are a numberof other publications that describe fewer patients treatedwith the Amplatzer VSD device.3,12

The outcomes are difficult to compare with those ofpublished surgical series given the small patient num-bers and the patient selection bias present in all studies.Predictors of increased surgical mortality include haemo-dynamic instability, ventricular dysfunction, inferior VSDlocation and surgery early post-MI.13–15

One of the cases in our series was an elective hybridprocedure; inpatient CABG was performed, employingcardio-pulmonary bypass, on a semi-urgent basis fol-lltnaidyc

smpamb

CPcgbcmwupc

AWR

References

1. Crenshaw BS, Granger CB, Birnbaum Y, Pieper KS, MorrisDC, Kleiman NS, Vahanian A, Califf RM, Topol EJ. Risk factors,angiographic patterns, and outcomes in patients with ventric-ular septal defect complicating acute myocardial infarction.GUSTO-I (Global Utilization of Streptokinase and TPA forOccluded Coronary Arteries) Trial Investigators. Circulation2000;101:27–32.

2. Kinn JW, O’Neill WW, Benjuly KH, Jones DE, Grines CL.Primary angioplasty reduce the risk of myocardial rupturecompared to thrombolysis for acute myocardial infarction.Cathet Cardiovasc Diag 1997;42:151–7.

3. Goldstein JA, Casserly IP, Balzer DT, Lee R, Lasala JM.Transcatheter closure of recurrent postmyocardial infarctionventricular septal defects utilizing the Amplatzer postin-farction VSD device: a case series. Cathet Cardiovasc Interv2003;59:238–43.

4. Cutfield NJ, Ruygrok PN, Wilson NJ, Raudkivi PJ, Greaves SC,Gentles TL, Kerr AJ. Transcatheter closure of a complex postmyocardial infraction ventricular septal defect after surgicalpatch dehiscence. Intern Med J 2005;35:128–30.

5. Labrousse L, Choukroun E, Chevalier JM, Madonna F, Rober-tie F, Merlico F, Coste P, Deville C. Surgery for post infarctionventricular septal defect (VSD): risk factors for hospitaldeath and long term results. Eur J Cardiothorac Surg 2002;21:725–31.

6. Deja MA, Szostek J, Widenka K, Szafron B, Spyt TJ, Hickey MS,Sosnowski AW. Post infarction ventricular septal defect—canwe do better? Eur J Cardiothorac Surg 2000;18:194–201.

7. Jeppsson A, Liden H, Johnsson P, Hartford M, Radegran

owed by deferred percutaneous VSD closure, 44 daysater. Stability of the patient, good biventricular func-ion, a well-defined margin of the VSD on the LV side,o pulmonary hypertension, availability of device closure,nd the high mortality associated with early repair of annferior VSD were taken into account when making theecision. The patient remains alive and well almost oneear later. Hybrid procedures should be considered forarefully selected patients.

The major limitation to our study, and that of all otheringle centre series is the small number of patients. Treat-ent was not randomised and the decision to perform

ercutaneous closure was made on the clinical situationt the time. The improved outcomes of those treated lateray be a result of selection bias, a factor that will need to

e accounted for in future studies.

onclusionercutaneous device closure of post-infarct VSD may beonsidered in selected patients as an alternative to sur-ical repair or conservative management. Patients shoulde carefully selected and counselled and, if the clinicalondition permits, we believe that delaying the procedureay improve outcomes. It can be undertaken in patientsith surgical patch dehiscence. The frequent small resid-al shunt remains a concern; there remains considerableotential for future development of percutaneous VSDlosure devices.

cknowledgemente thank Barbara Semb, Research Secretary, Green Lane

esearch and Education Fund, for secretarial assistance.

K. Surgical repair of post infarction ventricular septaldefects: a national experience. Eur J Cardiothorac Surg 2005;27:216–21.

8. Ellis CJ, Parkinson GF, Jaffe WM, Campbell MJ, KerrAR. Good long-term outcome following surgical repair ofpost-infarction ventricular septal defect. Aust NZ Med J1995;25:330–6.

9. Lock JE, Block PC, McKay RG, Baim DS, Keane JF. Tran-scatheter closure of ventricular septal defects. Circulation1988;78:361–8.

10. Holzer R, Balzer D, Amin Z, Ruiz CE, Feinstein J, Bass J, VanceM, Cao QL, Hijazi ZM. Transcatheter closure of postinfarctionventricular septal defects using the new Amplatzer muscularVSD occluder: results of an U.S. Registry. Catheter CardiovascInterv 2004;61:196–201.

11. Szkutnik M, Bialkowski J, Kusa J, Banaszak P, Baranowski J,Gasior M, Chodor P, Zembala M. Postinfarction ventricularseptal defect closure with Amplatzer occluders. Euro J ThoracicSurg 2003;23:323–7.

12. Pesonen E, Thilen U, Sandstrom S, Arheden H, Koul B, OlssonSE, Wilson RF, Toher C, Bank A, Bass J. Transcatheter closureof post-infarction ventricular septal defect with the AmplatzerSeptal Occluder device. Scand Cardiovasc J 2000;34:446–8.

13. Moore CA, Nygaard TW, Kaiser DL, Cooper AA, Gibson RS.Postinfarction ventricular septal rupture: the importance oflocation of infarction and right ventricular function in deter-mining survival. Circulation 1986;74:45–55.

14. Cummings RG, Califf R, Jones RN, Reimer KA, Kong Y-H,Lowe JE. Correlates of survival in patients with postinfarctionventricular septal defect. Ann Thorac Surg 1989;47:824–30.

15. Held AC, Cole PL, Lipton B, Gore JM, Antman EM, HochmanJS, Corrao J, Goldberg RJ, Alpert JS. Rupture of the inter-ventricular septum complicating acute myocardial infarction:a multicenter analysis of clinical findings and outcome. AmHeart J 1988;116:1330–6.