- Email: [email protected]
Risks and Benefits of Combined Maze Procedure for Atrial Fibrillation Associated With Organic Heart Disease1
985
J A C Vol.C 28, No. 4 October 1996:985-90
ELECTROPHYSIOLOGY
Risks and Benefits of Combined Maze Procedure for Atrial Fibrillation Associated With Organic Heart Disease T. KAWAGUCHI, AKIRA
MD, YOSHIO KOSAKAI, MD, YOSHIKADO SASAKO, MD, KIYOYUKI EISHI, MD, KIYOHARU NAKANO, MD, YASUNARU KAWASHIMA, MD
Suitaj Osaka,Japan
Objectives.Thisstudysoughtto identifythe risksandbenefits of addingthe mazeprocedurein patientswithatria]fibrillation (AF)undergoingoperationforunderlyingorganiccardiacdisorders. Background. PersistentAF oftenleavespatientssymptomatic evenafterotherwisesuccessfulcardiacsurgery. Methods. Fifty-onepatientsundergoingvalvularoperationand the mazeprocedure(n = 43) or repairof congenitalanomalies (n= 8) combinedwiththemazeprocedurewerecomparedwith51 patients (controlgroup)matchedfor underlyingdiseases and proceduresexceptforthe mazeoperation.Eachgroup,including 31 patientstith a concomitanttricuspidannuloplastyand 12 undergoingreoperation,weresimilarin age,durationof arrhythmia, degreeof cardiomegalyand NewYorkHeartAssociation functionalclass. Results.Patientsundergoingthe mazeprocedurehad longer cardiopulmonary bypasstime (213 vs. 144 rein, p < 0.0001),
longercardiacarrest (134vs. 93 rein,p < 0.0001)and greater bloodlosswithlongerrespiratorycare(39vs. 18h,p = 0.021)and intensivecareunitstaybutnomortality,Nosignificantdifferences werefoundin catecholamineor transfusionrequirementsimme? diatelyafteroperation.SustainedAFwas muchless frequentin themazegroup(12%at 1 year)thanthecontrolgroup(86%,p < 0.0001.),withan averagefollow-upperiodof 32 months(range25 to 42). Atria]contractionwas documentedin 41 (80%0)and 40 (78%)patientsfor rightand left ventricularfilling,respectively, after the maze procedure,resultingin a significantlysmaller cardiacsize and improvedfunctionalcapacity.Medicationwas discontinuedin sevenpatientsin the mazegroupcomparedwith MO in the controlgroup. Conclusions. Improvedrestorationof atrial rhythmand contractionwith combinedmazeoperationappearedto justifi the increasedoperativetimeandcomplexityandpostoperativecare. (JAm CoilCardiol1996;28:985-90)
Cox and colleaguesdeveloped(1) and applied(2) the maze (7) so as to shortenthe operatingtimeand preservethe sinus proceduresuccessfullyin patientswith lone atrialfibrillation node artery (8). Moreover,myocardialchangesand fibrosis derivedfromunderlyingdiseasesmayrenderthe mazeproce(AF).AlthoughisolatedAF itselfhasbeenreportedto carrya durelesseffective thanin loneAF (2).To identi@therisksand low risk of thromboembolism(3), AF increasesthe risk benefitsof combining the mazeprocedurein patientsundersignificantly when associatedwith organicdisease(4). Once goingcardiacsurge~,weretrospectively comparedthe first51 AF becomessustainedin thesepatients,it usuallypersistseven suchpatientswitha case-controlled g roupof 51 patientswith afterotherwisesuccessfuloperationfor the underlyinglesions AF matchedforunderlying diseases andprocedures exceptfor (5,6).BecausedevelopingAF signifiespathologicdegradation the mazeoperation. and symptomaticdeteriorationin thesepatients,simultaneous treatmentof the rhythmand organiclesionshas long been desired(5,6)and expectedto improveprognosis.Nonetheless, Methods addingthe mazeproceduremayincreasethe riskbecauseof Studypatients. We modifiedthe originalmazeprocedure extensiveatrial incisionand reanastomosisrequiringlonger (1) and beganto combineit withotheropen heart operations cardiac arrest and cardiopulmonarybypasstime. For this in March1992; the initial14patientswereselectedmainlyon reason,weusedcryoablationandmodifiedthemazeatriotomy the basisof simplicityof the combinedprocedurefor safety. After a furthermodificationin atriotomy(Fig.1) (7), contraFromthe NationalCardiovascular Center,Suita,Osaka,Japan.Thisstudy wassupportedinpartbythe SpecialCoordination FundsforPromotingScience indicationsfor the combinedapproachwereabandonedin the next37 patientswithAF undergoingcardiacoperation,with and Technology(EncouragementSystemof COE) from the Scienceand TechnologyAgency,Tokyo;and ResearchGrant 6C-4for Cardiovascular the exceptionoftwopatientsduringthe sameperiodwhowere Diseases,from the Ministryof Health and Welfare,Tokyo,Japan. It was presentedat the 43rdAnnualScientificSession,AmericanCollegeof Cardiol- judged unable to tolerate the combinedoperations.Up to August1993,51patientswithchronicAF (Table1)underwent Ogy,Atlanta,Georgia,March1994, ManuscriptreceivedMay11,1995;revisedmanuscriptreceivedFebruary16, the mazeoperationsimultaneously withmitralvalveoperation 1996,acceptedMay14,1996. (n = 41),isolatedaorticvalvesurgery(n = 2) and closureof Addressforcorrespondence: Dr.AkiraT.Kawaguchi, Nationalcardiovascular atrial septaldefect([ASD]ostiumsecundum, n = 7; ostium Center,Fujishirodai 5-7-1,Suita565,Osaka,[email protected]. 0735-1097/96/$15.00 PII SO735-1097(96)OO275-6
01996 by the Amelican College of Cardiology Published by Elsevier Science Inc.
—-
986
JACC Vol. 28, No. 4 October 1996:985–90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
Table1. UnderlyingLesions and Procedures in the Maze Group Abbreviationsand Acronyms AF
No.(%) of Operations
= atrialfibrillation
ASD = CTR = Icu = LAD = NYHA=
Total Reoperation GM
atrialseptaldefect cardiothoracic ratio intensivecareunit leftatrialdimension NewYorkHeartAssociation
Atriafseptaldefect Isolated +Tricuspid valve
+Tncuspidvalve+ mitralvalve Aorticvalvedisease Mitralvalvedisease
Isolated
primum, n = 1). Concomitanttricuspidannuloplastywas carriedout in 31patients(60.8%)andleftatrialplicationin 3. Twelvepatients (23.5%)had a previousvalvularoperation, withreplacementin 9 and repairin 3. Control patients. For each patient undergoingthe maze procedure,a controlpatientwasselectedfrompatientsundergoingthe sameproceduresfor the sameunderlyingdiseases, with the samehistoryof previousoperation,but withoutthe mazeprocedure.Controlpatientswereselectedretrospectively accordingto best matchwithregardto age and preoperative NewYorkHeart Associationfunctionalclass. Surgical modifications. Our initial modificationof the originalmazeprocedure(1) includeduse of cryoablationand changesin atriotomy(Fig.1) to shortenthe atrialsutureline. Thisprocedurewasfurthermodified(7) to avoidtransecting the sinusnodeartery(8)(Fig.1).Othermodifications included transectionof the superiorvenacavafor better exposureand easiermanipulationofthemitralvalvein allexceptonepatient who had ASD closureand tricuspidannuloplasty.Circumferentialincisionaroundthe pulmonaryveinsmobilizedthe left ventricleand improvedexposureof the mitralvalve. Postoperative management.Atrial fibrillationoccurring early after operationassociatedwith hemodynamiccompromisewasfirsttreated with overdrivepacingusingtemporary wiresand intravenousmedicationand then by electricalcardioversionin bothgroupsof patients.Atrialfibrillationoccur-
+Tricuspidvalve +Aorticvalve +Tricuspidvalve+ aortic valve Total
Repair
19 10 6
0 (o) 0 (o) o(o) o(o) 1(50) 11(27) 0 (o) 8 (42) 2 (20) 1(17)
o (o) 8 (1OQ) o (o) 2(100) o (o) 4 (loo) o (o) 2 (loo)* o (o) o (o) 4 (lo) 14(33) 2(33) 4 (67) 1(5) 7(37) 1(lo) 3 (30) o (o) o(o)
51
12(24)
4(8)
8 2 4
2* 2 41
6
22(43)
*Patient withostiumprimum defect.GLA= giantleftatrium.
ringlaterwithoutobviouscompromisewastreatedin the same way except for electricalcardioversion,which required an informedconsent.Directcurrentcardioversion wascarriedout under intravenousanesthesia,delivering100 to 300 J with intravenousantiarrhythmicagentsin case of failure.Afterward, patientswere usuallystarted on oral antiarrhythmic agents,mainlyprocainamideor quinidine,untiltherhythmwas consideredstable,unlessintolerancedeveloped.Anticoagulation withwarfarinwas institutedin patientswith mechanical valvesand those with persistentAF. Patientswho regained atrialrhythmand contractionafter reparativeoperationwere maintainedwithanticoagulation fortheinitial3 to 6 months,at whichtimeanticoagulation wasdiscontinued,and antiarrhythmicagents,if any,weretapered. Data collection. Informationcollectedfor comparisonincludeddurationof cardiacarrestand cardiopulmonary bypass time in the operatingroom. Early after operation,patients were evaluatedfor hemodynamicvariables,requirementfor
Figure1. Illustrationof operativetechniques:Atriotomy (solidline withcrossbars)and cryoablation(dottedarea) withvariationsin the sinusnode arteryfor the initial (left panel)andcurrentmodifications (rightpanel).Upperpanels illustrateendocardialviewsof the atria, and lowerpanels showposteriorviewsof the cardiacbase.FO = fossaovalis; IVC = inferiorvena cava;LAA = left atrial appendage; LSA = left sinusnode artery;MV = mitralvalve;PSA = posterior sinus node artery; RAA = right atrial appendage; RSA = right sinus node artery; SN = sinusnode; SVC=
superiorvenacava;TV = tricuspidvalve.
.
.
JACC Vol. 28, No. 4 October 1996:985-90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
987
Table2, ClinicalCharacteristicsof the Mazeand ControlGroups catecholamines, cardiacindex,pulmonaryarterypressuresand pulmonaryarterywedgepressure12h after admissionto the MazeGroup Control Group intensivecare unit (ICU).Timeuntilextubation(extubation) (n= 51) (n= 51) p Value and before dischargefrom the ICU (ICU stay) were also Preoperative recorded.Chest tube drainage(bloodloss) and transfusion Age(yr) 58.4? 8.3 58.8~ 8.1 0.797 requirementduringthe ICU staywerecalculated.Forpatients AF (J@ 7.6? 6.6 8.0* 6.7 0.760 f wave(mV) with the maze procedure,cardiacrhythmwas closelymoni0.16* 0.09 0.14t 0.08 0.209 LVDD(mm) 51.3t 10.1 51.3t 11.2 0.988 tored after cardioversionduringcardiopulmonary bypassimLVDS(mm) 35.3&8.7 35.228.7 0.965 mediatelybeforeand after operationand continuouslythere%FS 32.2f 8.6 31.9t 7.1 0.840 after.Afterchestclosure,atrialrhythmwascheckedwithatrial Perioperative pacemakerwires dailyin the ICU and with decreasingfre93+ 36 Arrest(rein) 134&25 <0.0001 quencyin thewarduntilremovalofthewiresbeforedischarge CPB(rein) <0.0001 213*40 144*51 from the hospital.Postoperatively, earlydiastolicventricular Operation(rein) 343? 94 <0.0001 496? 93 Blood10SS (mf) 1,014* 513 fillingand the presenceof atrialA waveduringtranstricuspid 727? 728 0.030 709* 731 Transfusion(ml) 452? 917 0.112 and transmittalflowweredocumentedbyDopplerechocardiPostoperative ographyscheduled1,3,6 and 12monthsafteroperation.Other CI (liters/reinper m’)” 3.67? 0.94 3.74? 0.65 0.443 measurementsincludedleft atrialdimension(LAD) and left 9.9~ 29 CVP(mmHg) 7.9? 3.3 0.001 ventriculardimensions.ChestX-rayfilmswere taken on the MPAP(mmHg) 24.8? 6.3 21.8* 5.9 0.028 samescheduleforcardiothoracicratio(CTR).Cardiacrhythm Extubation(h) 18.1t 11.8 38.9* 60.1 0.021 and functionalclasswere assessedat dischargeand at every ICUstay(h) 92.5? 86.9 59.8? 43.6 0.024 LVDD(mm) 47.8? 6.0 45.6? 7.5 ambulatoryvisitthereafter. 0.202 LVDS(mm) 34.3? 7.3 32.1* 8.0 0.262 Statistics. Continuousvariableswith equalvariancewere %FS 29.3t 9.2 30.1t 9.8 0.750 comparedby two-tailedttest. Whenvariancewasconsidered unequal,twosamplettestswithWelch’scorrectionwasused. Rhythm Sinusrhythm <0.0001 43(84%) 3 (6%) Discretevariableswereanalyzedbycontingencytableanalysis. Junctionalrhythm 2 (4%) 3 (6%) Freedom from postoperativeAF was analyzedby KaplanPersistentAF 45(88%) 6 (12%) Meieractuarialcurves.Changesin CTR,LADand functional *Cardiacindex(CI),centralvenouspressure(CVP)andmeanpulmonary class between the groups were analyzedwith analysisof artery pressure(MPAP)were measured12 h after ICU admission.Data variancecombinedwith a multiplecomparisonprocedure. presentedare meanvalue2 SDor number(%) of patients.AF = durationof Resultsare presentedas averagevalue53D. Differenceswere atrialfibrillation;Arrest= durationof cardiacarrest;Bloodloss= chesttube drainageintheICU;CPB= durationofcardiopulmonary bypass;Extubation= consideredstatisticallysignificantat p <0.05.
Results Patientgroups. Becauseno earlyor latedeathsoccurredin the mazegroup,controlpatientswereselectedfromsurviving patients with the same diseasesand undergoingthe same procedures.Thus, the case-matchednonmazecontrolgroup includedthe samenumberof patientswiththe sameunderlyingdiseases,historyof previousoperation(24’%)and surgical proceduresas the mazegroup(Table1).Among43 patients with valvulardisease,both the maze and control groups included36patientswithrheumaticdiseaseand 7 withdegenerativelesions(puremitralregurgitationin 6,including2 with repeatplasty,and bioprostheticvalvefailurein 1).Among36 patientswith rheumaticetiology,15 had predominantlystenoticlesions,5 had predominantlyregurgitanthemodynamic variables,andthe other 16had combineddiseases,including9 withreoperation.Preoperatively, the controlgroupwassimilar in averageage, f wavevoltage,left ventriculardimensions, cardiomegaly andfunctionalclassto themazegroup(Table2), exceptfor an averagedate of operation29 monthsearlier. Althoughaveragedurationof AF wascomparable,the maze groupincludedfivepatients(9.8%)with a historyof AF <1 year comparedwith nine (17.6%)in the controlgroup.Althoughthe mazegroupincludedno patientwith a historyof
—-
timebeforeextubation; fwave= maximalfwavevoltageinVIlead;LAD= left atrialdimension; LVDD= leftventricularend-diastolic dimension; LVDS= left ventricularend-systolicdimension;FS = fractionalshortening;Operation= timein the operatingroom.
AF <6 months,the controlgroup had three such patients, includingonewithAF for only2 monthsbeforeoperation. Operativedata. Intraoperatively, the mazegrouprequired significantly longercardiacarrest (41minlongeron average), cardiopulmonary bypass(69rein)andoperativetime(153rein) than the controlgroup (Table2). Althoughthe maze group had significantly greaterbloodloss,theydid not requiremore transfusions.Intraaorticballoonpumpingwasrequiredin four patient in the maze groupversusone in the controlgroup. Although no differenceswere found in cardiac index or catecholaminerequirements,pulmonaryartery and central venouspressureswere higherin the maze group 12 h after admissionto the ICU (Table 2). The maze group required significantly longerrespiratorycare, resultingin a prolonged ICU stay(Table2). Postoperative course. Electrical cardioversionwas performedin 11controlgrouppatientsmainlyto controltachyarrhythmiaearlyafteroperationandin 17mazegrouppatients mainlyto reverseAF beforedischarge,with11responding.Ail sixpatientswithrefractoryAF had mitralvalvediseaseas the
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
JACC Vol. 28, No. 4 October 1996:985-90
Cardiacsize, contractionand functional class. Left ven-
“%
100 90 80 70 60 50 40 30 2010-
o~ POSTOP YEARS
Postoperative(POSTOP)rhythmwasevaluatedat hospital di~chargeand it everyambulato~vkit thereafter.To analyze-the AF-freerate usingKaplan-Meieractuarialcurves,patientswho~ventuallyregainedsinusrhythmwere consideredAF free regardlessof temporaryrhythmdisturbances, whereasthoseinwhomAFeventually returnedwereconsideredto havedevelopedAFat thelastrecurrence. TheactuarialcurvesrepresentfreedomfromAFafterdischargeinthe mazegroup (solidcircles)versusthe controlgroup (opencircles), exceptfor one maze group patient who regainedsinus rhythm3 monthspostoperatively, showinga highlysignificantdifference(p < 0.0001).Verticalbars= patientswitha designatedperiodoftimeafter operation;all patientswerefollowedup >2 years.
underlyingpathology(rheumaticin four,degenerativein two). Cardiacrhythmwasstabilizedby the timeof discharge,when postoperativerhythmwasdefined,exceptfor twomazegroup patients(one experiencedreturn of AF; the other had spontaneousablationof AF); three controIpatientsexperienced return of AF after discharge.SustainedAF was much less frequentin the mazegroup(Fig.2,Table2).Threepatientsin the controlgroupwhoregainedsinusrhythmwithoutthe maze procedurehad had AF for 7, 7 and 17 months,respectively, beforemitralvalvuloplasty for rupturedchordaetendineaein two and ASD closurein the other. Althoughthree control patientsresumedand toleratedjunctionalrhythm,sinusnode dysfunctionrequiredatrial pacemakerimplantationin three patientsundergoingthe initialmodification(21.4%).In twoof them, preoperativesinusnode functioncould not be determined,and one regainedsinusrhythmlater, overdrivingthe pacemaker.No need for permanentpacingoccurredin the controlgroupandin the last37patientsundergoingthe current modification. Atrial contraction.Au A wave was documentedin 41 (80%) of 51 patients during transtricuspidflow and in 40 (78%) of 51 during transmittal flow after the combined operations.Whereasearlydiastolicventricularfillingremained high,the transtricuspidA wavesignificantlyincreasedfrom 30 ~ 9 cm/~at 1 monthto 39 ~ 14 cm/sat 3 monthsafter operation (p = 0.024).Similarly,the transmittalA wave tendedto increasefrom46 & 22 cm/sat 1 monthto 54 f 25 cm/sat 3 monthsafter operation,comparableto the normal valuefor age (9),and leveledoffthereafter.
triculardimensionsand fractionalshorteningdid not change beforeand afteroperationor differbetweenthe groups(Table 2). For CTR, LAD and functionalclass,observationtime pointswerecombinedas early(1 to 3 months)and late (1 to 2 years) after operation (Fig. 3). Although cardiac size as assessedby CTR remainedunchangedin the controlgroup after an initialdecrease,it continuedto decreaseonlyin the mazegroup(p = 0.011).Both groupsshowedreducedLAD afteroperation;however,in the controlgroup,LADincreased significantly(p = 0.03),returned to preoperativelevelsand becamesignificantly largerthan that in the mazegroup(p < 0.0015),in whomLAD was essentiallyunchanged(Fig. 3). Both atria appearedto contractmore efficientlyovertime in sinusrhythm,withdecreasingLAD.Increasedphysicalactivity andloadafterdischargeresultedin lessfatigueanddyspneain the mazegroup,indicatinga significantly improvedfunctional capacityin thesepatients(p <0.009, Fig.3). Medications. Among22patientsafterreparativeoperation (8 afterASDclosure,14after mitralvalvuloplasty) (Table1), anticoagulation couldbe discontinuedin 16patients(73$%) in the maze group.Medicationwas totallydiscontinuedin two control group patients and in seven maze group patients whoregainedatrialrhythmand contraction(fouraftertricuspid annuloplastywithASDclosure,three aftermitralvalvuloplasty). Late complications. Despite warfarinanticoagulation,one transientnecrologicischemicattackoccurredin a maze group patient with normal sinus rhythmand contraction4 months aftermechanicalmitralvalvereplacement,andcerebralinfarction occurred in one control patient with persistent AF 6 months after aortic valve replacement.Intracranialbleeding requiringadmission to the hospital occurred in one maze grouppatientand two controlgrouppatients;all were receiving warfarinfor anticoagulation.
Discussion Studydesignand limitations. This retrospectivestudywas carriedout to evaluatethe risksandbenefitsof combiningthe maze procedure with surgicalinterventionfor underlying disorderscausingAF; however,a prospective,randomized studyshouldhavebeen performed.After beinginformedof the initialresults,everypatient requestedthe combinedapproach despitepotentiallyincreasedrisksand undetermined efficacy.Although case-matchedcontrol patients were included, obviousdifferencesin the date of operation and treatmentof postoperativeAF couldnot be eliminated,underminingthe inferencesderivedfrom the results.The decisionto addthe mazeprocedurerestonwhetherthepotentially greater risksof increasedcomplexityare outweighedby the benefitsfromregainedatrialrhythmandcontraction.Thus,the risk-benefitbalancemay help to determineindicationsfor the combinedapproach,whichshouldbe differentfromindicationsfor the isolatedmazeprocedurefor loneAF (2).
JACC Vol. 28, No. 4 October 1996:985-90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
989
implantation. Althoughtheincidenceofsinusdysfunction requiringatrialpacing waslowerthanin the seriesof Coxet al.(2,10), 66frequentcomplications promptedusto furthermodifytheproce64dure(7),resultinginno needforartificialpacingamongthenext 37patientsundergoingthe currentmodification. Timesrequired 62forthecombinedoperationswere-40,70 and150minlongerfor 60cardiacarrest,cardiopuhnonary bypassandthe completeoperation,respectively, than for the controloperationsalone.Never58theless,thesetimesweresignificantly shorterthanthoserequired 56for the mazeIII procedureof Coxet al. (10)combinedwith 54 J operationsfororganiclesionsin a similarcohort(11),suggesting LATE EARLY PRE the simplicityof the current modikation. Preoperativeand 60postoperativeleftventriculardimensionsand contractionwere LAD(m) T 58simifarbetweenthe groups,suggesting that longercardiacisch56emia does not result in persistentor significantventricular 54dysfunction.Despiteinitiafcomplicationsattributableto the 52combinedmaze procedure,subsequentmodifications and in50creasedexperienceimprovedresultsand alloweduneventful recoveryin the restof the mazegroupdespiteextendedindica48tions. 46Benefits. Evenwhen the combinedapproachwas consid44. ered to havefailedto restoresinusrhythmor to defibrillatein 42 J LATE &lRLY PRE allpatientswhorequiredelectricalcardioversion(n = 17),AF 2.61 wassurgicallyablatedin the remainingtwo thirdsof patients NYHA (34of 51)afterthe combinedapproach.Thisnumberis much 24 Ii higherthan that in currentcontrolgroupsincludingtwiceas 1 2.2manypatientswitha briefhistoryofAF,whoweremorelikely 2to undergodefibrillation(12).Satoet al. (6)reportedthatAF 1.6wasabolishedin only2870ofpatientslateafteroperationwith aggressivetreatmentusingrepeatedcardioversionafterrepair 1.6of underlyinglesionsalone. Even lower long-termmainte1.4nance of sinus rhythmwas reported by Hansen (5), who 1.2identifiedpreoperativeAF <12 monthsin durationasthe only IJ variableaffectingthe resultsof electricalcardioversionafter IATE PRE EARLY mitral valve surgery.These reports (5,6) discouragedthe Figure3. Changesincardiothoracic ratio(CTR),leftatrialdimension controlpatientsfromundergoingcardioversion, resultingin a (LAD)andNewYorkHeartAssociation(NYHA)functionalclassin significant d ifferencein treatment of postoperative AF and the maze(circles)versusthe controlgroup(squares)before(PRE),1 earlier operation.However,these factorsalone may not acto 3 months(EARLY)and 1 to 2 years after operation(LATE). Afthoughtherewereno significantdifferences(p = 0.824)in periop- countforthefactthat34patientshadno arrhythmiasrequiring erativeCTRreductionbetweenthe maze(n = 50)andcontrolgroups earlycardioversion, and 88%remainedAF free late after the (n= 39),CTRcontinuedto decreaselate afteroperationin the maze combinedapproach;althoughadditionof the mazeprocedure grouponly(p = 0.011),resultingina significant differencebetweenthe alone may not be solelyresponsiblefor these advantages groups(p = 0.029).Therewere alsono significantdifferences(p = either. In contrast,the rate of regainingatrial rhythmand 0.509)betweenthe maze (n = 32) and controlgroups(n = 8) in perioperativereductionin LAD; however,LAD returned to the contractionafterthe combinedoperationswasstilllowerthan preoperativelevellate afteroperationonlyin the controlgroup(p = afterloneAFwiththe isolatedmazeprocedure(2,10).Results 0.0015),resultingin a significantdifferencebetweengroups(p = of our recentreview(11)and analysis(13)indicatedthat the 0,012).A significantdifference(p = 0.036)in perioperativeimprovedifferenceappearedto resultfromthe durationof the arrhythment in functionalclassremainedlate after operationbetweenthe maze(n = 40)and controlgroups(n = 17),resultingin a significant mia and LAD rather than the technicalmodificationitself. Despite concern about recoveryof atrial contractility,the differencebetweengroups(p = 0.009).Verticalbars= 1 SE. majorityof patientswerefoundto havesignificantright(n = Risks.Risksassociatedwiththe combinedapproachinclude 41 [80%])and left atrialcontraction(n = 40 [78%])for late thecomplexity ofthemazeprocedure,whichrequiresprolonged diastolicventricularfilling.Moreover,atrial contraction(A cardiacarrestand cardiopulmonary bypasswithsubsequentde- wave)appearedto improveover time in sinusrhythm,with layedpostoperative recovery.Amongthefirst14patientsunder- decreasingLAD,approachingthe normalvaluefor age (9)by goingthe initialmodification, 3 (21%)requiredintraaorticbal- 3 months after operation.Thus, not only atrioventricular loon pumping, and 3 (21%) required atrial pacemaker synchrony,but also active diastolicventricularfillingwas 68-
——
990
JACC Vol. 28, No. 4 October1996:985–90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
improved.Thisfindingmayaccountfor continuedreductionof cardiacsizein the mazegroup,as reportedby Gosselinket al. (14),whoobservedreductionin leftatrkdsizeonlyin patients remainingin .c$inus rhythmaftercardioversion. Improvedventricular Ming and reducedatrialsizemay accountfor improved functionalcapacityafter the combinedprocedure.Left atrial isolationhas been reported (15) to be less complicatedyet comparablyeffectivein restoringregularventricularcontraction or the RR intervalin 7190of patientslate after operation. Nonetheless, persistentAF inbothatriaaftertheproceduremay limithemodynamic improvement, whichwasnot demonstrated, andrequirecontinuedanticoagulation becausetheincreasedrisks inherentto AF persist(4).Of22patientsundergoingrepairwith ourcombinedapproach,anticoagulation wasnolongernecessary in 16(73%),andmedicationwastotallydiscontinued in7 (32%). Althoughtherewereno signitlcant differences intheincidenceof thromboembolic or bleedingcomplications in thecurrentreview withlimitednumbersandfollow-up, restorationof atrialrhythm and contractionby the mazeproceduremaymakea long-term differencein this cohortwithunderlyingorganiclesionsand a greaterriskof stroke(4). Conclusions. Becauseof the lackof prospectiverandomization, the present results are suggestiveand inferential. Nonetheless,resultsof the currentstudyconvincedus that the substantialbenefitsfromregainedatrialrhythmand contraction outweighthe potentialrisksof addingthe maze procedure,justifyingthe combinedapproachin patientswithestablishedAF andorganicdisorders,exceptfor a fewpatientswho wereconsideredunableto toleratethe procedure.Prospective randomizationseems warranted to examinethe value of addingthe mazeprocedurein patientswithAF ofrecentonset (12) and in those with presumedmaze-refractoryAF (13) undergoingoperationfor organiclesions. Weappreciatethe editorialassistanceof LeonardM.Linde,MD,Professorof Pediatrics(Cardiology), Universityof SouthernCalifornia,LosAngeles.
References L CoxJL.Thesurgicaltreatmentof atrialfibrillation.IV.Surgicaltechnique. J ThoracCardiovascSurg1991; 101:584-92. 2. CoxJL, BoineauJP, SchuesslerRB, Rater KM,LappasDG. Five-year experiencewiththe mazeprocedureforatrialfibrillation. AnnThoracSurg 1993;56:814-24. 3. KopeckySL,GershBJ,McGoonMD,et al.‘1’henaturalhistoryof loneatrial fibrillation:a population-based studyoverthree decades.N EnglJ Med 1987; 317:669-74. 4. ChesebroJH, FusterV, HalperinJL. Atrialfibrillation-riskmarkerfor stroke.N EnglJ Med1990;323:1556-8. 5. HansenJF, AndersonED, OlesenKH, et al. DC-conversion of atrial fibrillationaftermitralvalveoperation:an analysisofthe long-termresults. ScandJ ThoracCardiovascSurg1979; 13:267-70. 6. Sato S, KawashimaY, HiroseH, NakanoS, MatsudaH, ShimkuraR. Long-termresultsof direct-currentcardioversion afteropencommissurotomyfor mitralstenosis.AmJ Cardiol1986;57:629-33. 7. KosakaiY,Kawaguchi AT,IsobeF, et al.Coxmazeprocedureforpatients undergoing simultaneous surgery.J ThoracCardiovasc Surg1994; 108:1049– 55. 8. McAlpineWA.HeartandCoronaryArteries.Heidelberg:Springer-Verlag, 1975:151-9. 9. Miyatake& OkamotoM, KinoshitaN, et al. Augmentationof atriaf contributionto leftventricularinflowwithagingas assessedbyintracardiac Dopplerflowmetry. AmJ Cardiol1984;53:586-9. 10.CoxJL,BoineauJP, SchuesslerRB,JaquissRD,LappasDG.Modification of the mazeprocedurefor atrial flutter and atrial fibrillation.J Thorac Cardiovasc Surg1995; 110:473-84. 11.KoskaiY, KawaguchiAT, IsobeF, et al. Modifiedmazeprocedurefor patientswithatrialfibrillationundergoing simultaneous openheartsurgery. Circulation1995;92 Suppl11:11-359-64. 12.ChuaYL, SchaffI-IV,OrszulakTA, MorrisJJ. Outcomeof mitralvafve repair in patientswith preoperativeatrial fibrillation.Shouldthe maze procedurebecombinedwithmitralvalvuloplasty? J ThoracCardiovasc Surg 1994; 107:408-15. 13.Kawaguchi AT,KosakaiY,IsobeF, et al.Surgicalstratificationforpatients with atria] fibrillationsecondaryto organiccardiaclesionsundergoing surgery[abstract].Circulation1995;92 Suppl1:1-441. 14.Gosselink ATM,CrijnsHJGM,HamerHPM,HillegeH,LieRI.Changesin leftandrightatrialsizeaftercardioversion ofatrialfibrillation: roleofmitral valvedisease.J AmCoilCardiol1993;22:1666-72. 15.GraflignaA,PaganiF,Mirrzioni G,SalernoJ, ViganoM.Leftatrialisolation associatedwithmitralvalveoperation.AnnThoracSurg1992;54:1093-8.
J A C Vol.C 28, No. 4 October 1996:985-90
ELECTROPHYSIOLOGY
Risks and Benefits of Combined Maze Procedure for Atrial Fibrillation Associated With Organic Heart Disease T. KAWAGUCHI, AKIRA
MD, YOSHIO KOSAKAI, MD, YOSHIKADO SASAKO, MD, KIYOYUKI EISHI, MD, KIYOHARU NAKANO, MD, YASUNARU KAWASHIMA, MD
Suitaj Osaka,Japan
Objectives.Thisstudysoughtto identifythe risksandbenefits of addingthe mazeprocedurein patientswithatria]fibrillation (AF)undergoingoperationforunderlyingorganiccardiacdisorders. Background. PersistentAF oftenleavespatientssymptomatic evenafterotherwisesuccessfulcardiacsurgery. Methods. Fifty-onepatientsundergoingvalvularoperationand the mazeprocedure(n = 43) or repairof congenitalanomalies (n= 8) combinedwiththemazeprocedurewerecomparedwith51 patients (controlgroup)matchedfor underlyingdiseases and proceduresexceptforthe mazeoperation.Eachgroup,including 31 patientstith a concomitanttricuspidannuloplastyand 12 undergoingreoperation,weresimilarin age,durationof arrhythmia, degreeof cardiomegalyand NewYorkHeartAssociation functionalclass. Results.Patientsundergoingthe mazeprocedurehad longer cardiopulmonary bypasstime (213 vs. 144 rein, p < 0.0001),
longercardiacarrest (134vs. 93 rein,p < 0.0001)and greater bloodlosswithlongerrespiratorycare(39vs. 18h,p = 0.021)and intensivecareunitstaybutnomortality,Nosignificantdifferences werefoundin catecholamineor transfusionrequirementsimme? diatelyafteroperation.SustainedAFwas muchless frequentin themazegroup(12%at 1 year)thanthecontrolgroup(86%,p < 0.0001.),withan averagefollow-upperiodof 32 months(range25 to 42). Atria]contractionwas documentedin 41 (80%0)and 40 (78%)patientsfor rightand left ventricularfilling,respectively, after the maze procedure,resultingin a significantlysmaller cardiacsize and improvedfunctionalcapacity.Medicationwas discontinuedin sevenpatientsin the mazegroupcomparedwith MO in the controlgroup. Conclusions. Improvedrestorationof atrial rhythmand contractionwith combinedmazeoperationappearedto justifi the increasedoperativetimeandcomplexityandpostoperativecare. (JAm CoilCardiol1996;28:985-90)
Cox and colleaguesdeveloped(1) and applied(2) the maze (7) so as to shortenthe operatingtimeand preservethe sinus proceduresuccessfullyin patientswith lone atrialfibrillation node artery (8). Moreover,myocardialchangesand fibrosis derivedfromunderlyingdiseasesmayrenderthe mazeproce(AF).AlthoughisolatedAF itselfhasbeenreportedto carrya durelesseffective thanin loneAF (2).To identi@therisksand low risk of thromboembolism(3), AF increasesthe risk benefitsof combining the mazeprocedurein patientsundersignificantly when associatedwith organicdisease(4). Once goingcardiacsurge~,weretrospectively comparedthe first51 AF becomessustainedin thesepatients,it usuallypersistseven suchpatientswitha case-controlled g roupof 51 patientswith afterotherwisesuccessfuloperationfor the underlyinglesions AF matchedforunderlying diseases andprocedures exceptfor (5,6).BecausedevelopingAF signifiespathologicdegradation the mazeoperation. and symptomaticdeteriorationin thesepatients,simultaneous treatmentof the rhythmand organiclesionshas long been desired(5,6)and expectedto improveprognosis.Nonetheless, Methods addingthe mazeproceduremayincreasethe riskbecauseof Studypatients. We modifiedthe originalmazeprocedure extensiveatrial incisionand reanastomosisrequiringlonger (1) and beganto combineit withotheropen heart operations cardiac arrest and cardiopulmonarybypasstime. For this in March1992; the initial14patientswereselectedmainlyon reason,weusedcryoablationandmodifiedthemazeatriotomy the basisof simplicityof the combinedprocedurefor safety. After a furthermodificationin atriotomy(Fig.1) (7), contraFromthe NationalCardiovascular Center,Suita,Osaka,Japan.Thisstudy wassupportedinpartbythe SpecialCoordination FundsforPromotingScience indicationsfor the combinedapproachwereabandonedin the next37 patientswithAF undergoingcardiacoperation,with and Technology(EncouragementSystemof COE) from the Scienceand TechnologyAgency,Tokyo;and ResearchGrant 6C-4for Cardiovascular the exceptionoftwopatientsduringthe sameperiodwhowere Diseases,from the Ministryof Health and Welfare,Tokyo,Japan. It was presentedat the 43rdAnnualScientificSession,AmericanCollegeof Cardiol- judged unable to tolerate the combinedoperations.Up to August1993,51patientswithchronicAF (Table1)underwent Ogy,Atlanta,Georgia,March1994, ManuscriptreceivedMay11,1995;revisedmanuscriptreceivedFebruary16, the mazeoperationsimultaneously withmitralvalveoperation 1996,acceptedMay14,1996. (n = 41),isolatedaorticvalvesurgery(n = 2) and closureof Addressforcorrespondence: Dr.AkiraT.Kawaguchi, Nationalcardiovascular atrial septaldefect([ASD]ostiumsecundum, n = 7; ostium Center,Fujishirodai 5-7-1,Suita565,Osaka,[email protected]. 0735-1097/96/$15.00 PII SO735-1097(96)OO275-6
01996 by the Amelican College of Cardiology Published by Elsevier Science Inc.
—-
986
JACC Vol. 28, No. 4 October 1996:985–90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
Table1. UnderlyingLesions and Procedures in the Maze Group Abbreviationsand Acronyms AF
No.(%) of Operations
= atrialfibrillation
ASD = CTR = Icu = LAD = NYHA=
Total Reoperation GM
atrialseptaldefect cardiothoracic ratio intensivecareunit leftatrialdimension NewYorkHeartAssociation
Atriafseptaldefect Isolated +Tricuspid valve
+Tncuspidvalve+ mitralvalve Aorticvalvedisease Mitralvalvedisease
Isolated
primum, n = 1). Concomitanttricuspidannuloplastywas carriedout in 31patients(60.8%)andleftatrialplicationin 3. Twelvepatients (23.5%)had a previousvalvularoperation, withreplacementin 9 and repairin 3. Control patients. For each patient undergoingthe maze procedure,a controlpatientwasselectedfrompatientsundergoingthe sameproceduresfor the sameunderlyingdiseases, with the samehistoryof previousoperation,but withoutthe mazeprocedure.Controlpatientswereselectedretrospectively accordingto best matchwithregardto age and preoperative NewYorkHeart Associationfunctionalclass. Surgical modifications. Our initial modificationof the originalmazeprocedure(1) includeduse of cryoablationand changesin atriotomy(Fig.1) to shortenthe atrialsutureline. Thisprocedurewasfurthermodified(7) to avoidtransecting the sinusnodeartery(8)(Fig.1).Othermodifications included transectionof the superiorvenacavafor better exposureand easiermanipulationofthemitralvalvein allexceptonepatient who had ASD closureand tricuspidannuloplasty.Circumferentialincisionaroundthe pulmonaryveinsmobilizedthe left ventricleand improvedexposureof the mitralvalve. Postoperative management.Atrial fibrillationoccurring early after operationassociatedwith hemodynamiccompromisewasfirsttreated with overdrivepacingusingtemporary wiresand intravenousmedicationand then by electricalcardioversionin bothgroupsof patients.Atrialfibrillationoccur-
+Tricuspidvalve +Aorticvalve +Tricuspidvalve+ aortic valve Total
Repair
19 10 6
0 (o) 0 (o) o(o) o(o) 1(50) 11(27) 0 (o) 8 (42) 2 (20) 1(17)
o (o) 8 (1OQ) o (o) 2(100) o (o) 4 (loo) o (o) 2 (loo)* o (o) o (o) 4 (lo) 14(33) 2(33) 4 (67) 1(5) 7(37) 1(lo) 3 (30) o (o) o(o)
51
12(24)
4(8)
8 2 4
2* 2 41
6
22(43)
*Patient withostiumprimum defect.GLA= giantleftatrium.
ringlaterwithoutobviouscompromisewastreatedin the same way except for electricalcardioversion,which required an informedconsent.Directcurrentcardioversion wascarriedout under intravenousanesthesia,delivering100 to 300 J with intravenousantiarrhythmicagentsin case of failure.Afterward, patientswere usuallystarted on oral antiarrhythmic agents,mainlyprocainamideor quinidine,untiltherhythmwas consideredstable,unlessintolerancedeveloped.Anticoagulation withwarfarinwas institutedin patientswith mechanical valvesand those with persistentAF. Patientswho regained atrialrhythmand contractionafter reparativeoperationwere maintainedwithanticoagulation fortheinitial3 to 6 months,at whichtimeanticoagulation wasdiscontinued,and antiarrhythmicagents,if any,weretapered. Data collection. Informationcollectedfor comparisonincludeddurationof cardiacarrestand cardiopulmonary bypass time in the operatingroom. Early after operation,patients were evaluatedfor hemodynamicvariables,requirementfor
Figure1. Illustrationof operativetechniques:Atriotomy (solidline withcrossbars)and cryoablation(dottedarea) withvariationsin the sinusnode arteryfor the initial (left panel)andcurrentmodifications (rightpanel).Upperpanels illustrateendocardialviewsof the atria, and lowerpanels showposteriorviewsof the cardiacbase.FO = fossaovalis; IVC = inferiorvena cava;LAA = left atrial appendage; LSA = left sinusnode artery;MV = mitralvalve;PSA = posterior sinus node artery; RAA = right atrial appendage; RSA = right sinus node artery; SN = sinusnode; SVC=
superiorvenacava;TV = tricuspidvalve.
.
.
JACC Vol. 28, No. 4 October 1996:985-90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
987
Table2, ClinicalCharacteristicsof the Mazeand ControlGroups catecholamines, cardiacindex,pulmonaryarterypressuresand pulmonaryarterywedgepressure12h after admissionto the MazeGroup Control Group intensivecare unit (ICU).Timeuntilextubation(extubation) (n= 51) (n= 51) p Value and before dischargefrom the ICU (ICU stay) were also Preoperative recorded.Chest tube drainage(bloodloss) and transfusion Age(yr) 58.4? 8.3 58.8~ 8.1 0.797 requirementduringthe ICU staywerecalculated.Forpatients AF (J@ 7.6? 6.6 8.0* 6.7 0.760 f wave(mV) with the maze procedure,cardiacrhythmwas closelymoni0.16* 0.09 0.14t 0.08 0.209 LVDD(mm) 51.3t 10.1 51.3t 11.2 0.988 tored after cardioversionduringcardiopulmonary bypassimLVDS(mm) 35.3&8.7 35.228.7 0.965 mediatelybeforeand after operationand continuouslythere%FS 32.2f 8.6 31.9t 7.1 0.840 after.Afterchestclosure,atrialrhythmwascheckedwithatrial Perioperative pacemakerwires dailyin the ICU and with decreasingfre93+ 36 Arrest(rein) 134&25 <0.0001 quencyin thewarduntilremovalofthewiresbeforedischarge CPB(rein) <0.0001 213*40 144*51 from the hospital.Postoperatively, earlydiastolicventricular Operation(rein) 343? 94 <0.0001 496? 93 Blood10SS (mf) 1,014* 513 fillingand the presenceof atrialA waveduringtranstricuspid 727? 728 0.030 709* 731 Transfusion(ml) 452? 917 0.112 and transmittalflowweredocumentedbyDopplerechocardiPostoperative ographyscheduled1,3,6 and 12monthsafteroperation.Other CI (liters/reinper m’)” 3.67? 0.94 3.74? 0.65 0.443 measurementsincludedleft atrialdimension(LAD) and left 9.9~ 29 CVP(mmHg) 7.9? 3.3 0.001 ventriculardimensions.ChestX-rayfilmswere taken on the MPAP(mmHg) 24.8? 6.3 21.8* 5.9 0.028 samescheduleforcardiothoracicratio(CTR).Cardiacrhythm Extubation(h) 18.1t 11.8 38.9* 60.1 0.021 and functionalclasswere assessedat dischargeand at every ICUstay(h) 92.5? 86.9 59.8? 43.6 0.024 LVDD(mm) 47.8? 6.0 45.6? 7.5 ambulatoryvisitthereafter. 0.202 LVDS(mm) 34.3? 7.3 32.1* 8.0 0.262 Statistics. Continuousvariableswith equalvariancewere %FS 29.3t 9.2 30.1t 9.8 0.750 comparedby two-tailedttest. Whenvariancewasconsidered unequal,twosamplettestswithWelch’scorrectionwasused. Rhythm Sinusrhythm <0.0001 43(84%) 3 (6%) Discretevariableswereanalyzedbycontingencytableanalysis. Junctionalrhythm 2 (4%) 3 (6%) Freedom from postoperativeAF was analyzedby KaplanPersistentAF 45(88%) 6 (12%) Meieractuarialcurves.Changesin CTR,LADand functional *Cardiacindex(CI),centralvenouspressure(CVP)andmeanpulmonary class between the groups were analyzedwith analysisof artery pressure(MPAP)were measured12 h after ICU admission.Data variancecombinedwith a multiplecomparisonprocedure. presentedare meanvalue2 SDor number(%) of patients.AF = durationof Resultsare presentedas averagevalue53D. Differenceswere atrialfibrillation;Arrest= durationof cardiacarrest;Bloodloss= chesttube drainageintheICU;CPB= durationofcardiopulmonary bypass;Extubation= consideredstatisticallysignificantat p <0.05.
Results Patientgroups. Becauseno earlyor latedeathsoccurredin the mazegroup,controlpatientswereselectedfromsurviving patients with the same diseasesand undergoingthe same procedures.Thus, the case-matchednonmazecontrolgroup includedthe samenumberof patientswiththe sameunderlyingdiseases,historyof previousoperation(24’%)and surgical proceduresas the mazegroup(Table1).Among43 patients with valvulardisease,both the maze and control groups included36patientswithrheumaticdiseaseand 7 withdegenerativelesions(puremitralregurgitationin 6,including2 with repeatplasty,and bioprostheticvalvefailurein 1).Among36 patientswith rheumaticetiology,15 had predominantlystenoticlesions,5 had predominantlyregurgitanthemodynamic variables,andthe other 16had combineddiseases,including9 withreoperation.Preoperatively, the controlgroupwassimilar in averageage, f wavevoltage,left ventriculardimensions, cardiomegaly andfunctionalclassto themazegroup(Table2), exceptfor an averagedate of operation29 monthsearlier. Althoughaveragedurationof AF wascomparable,the maze groupincludedfivepatients(9.8%)with a historyof AF <1 year comparedwith nine (17.6%)in the controlgroup.Althoughthe mazegroupincludedno patientwith a historyof
—-
timebeforeextubation; fwave= maximalfwavevoltageinVIlead;LAD= left atrialdimension; LVDD= leftventricularend-diastolic dimension; LVDS= left ventricularend-systolicdimension;FS = fractionalshortening;Operation= timein the operatingroom.
AF <6 months,the controlgroup had three such patients, includingonewithAF for only2 monthsbeforeoperation. Operativedata. Intraoperatively, the mazegrouprequired significantly longercardiacarrest (41minlongeron average), cardiopulmonary bypass(69rein)andoperativetime(153rein) than the controlgroup (Table2). Althoughthe maze group had significantly greaterbloodloss,theydid not requiremore transfusions.Intraaorticballoonpumpingwasrequiredin four patient in the maze groupversusone in the controlgroup. Although no differenceswere found in cardiac index or catecholaminerequirements,pulmonaryartery and central venouspressureswere higherin the maze group 12 h after admissionto the ICU (Table 2). The maze group required significantly longerrespiratorycare, resultingin a prolonged ICU stay(Table2). Postoperative course. Electrical cardioversionwas performedin 11controlgrouppatientsmainlyto controltachyarrhythmiaearlyafteroperationandin 17mazegrouppatients mainlyto reverseAF beforedischarge,with11responding.Ail sixpatientswithrefractoryAF had mitralvalvediseaseas the
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
JACC Vol. 28, No. 4 October 1996:985-90
Cardiacsize, contractionand functional class. Left ven-
“%
100 90 80 70 60 50 40 30 2010-
o~ POSTOP YEARS
Postoperative(POSTOP)rhythmwasevaluatedat hospital di~chargeand it everyambulato~vkit thereafter.To analyze-the AF-freerate usingKaplan-Meieractuarialcurves,patientswho~ventuallyregainedsinusrhythmwere consideredAF free regardlessof temporaryrhythmdisturbances, whereasthoseinwhomAFeventually returnedwereconsideredto havedevelopedAFat thelastrecurrence. TheactuarialcurvesrepresentfreedomfromAFafterdischargeinthe mazegroup (solidcircles)versusthe controlgroup (opencircles), exceptfor one maze group patient who regainedsinus rhythm3 monthspostoperatively, showinga highlysignificantdifference(p < 0.0001).Verticalbars= patientswitha designatedperiodoftimeafter operation;all patientswerefollowedup >2 years.
underlyingpathology(rheumaticin four,degenerativein two). Cardiacrhythmwasstabilizedby the timeof discharge,when postoperativerhythmwasdefined,exceptfor twomazegroup patients(one experiencedreturn of AF; the other had spontaneousablationof AF); three controIpatientsexperienced return of AF after discharge.SustainedAF was much less frequentin the mazegroup(Fig.2,Table2).Threepatientsin the controlgroupwhoregainedsinusrhythmwithoutthe maze procedurehad had AF for 7, 7 and 17 months,respectively, beforemitralvalvuloplasty for rupturedchordaetendineaein two and ASD closurein the other. Althoughthree control patientsresumedand toleratedjunctionalrhythm,sinusnode dysfunctionrequiredatrial pacemakerimplantationin three patientsundergoingthe initialmodification(21.4%).In twoof them, preoperativesinusnode functioncould not be determined,and one regainedsinusrhythmlater, overdrivingthe pacemaker.No need for permanentpacingoccurredin the controlgroupandin the last37patientsundergoingthe current modification. Atrial contraction.Au A wave was documentedin 41 (80%) of 51 patients during transtricuspidflow and in 40 (78%) of 51 during transmittal flow after the combined operations.Whereasearlydiastolicventricularfillingremained high,the transtricuspidA wavesignificantlyincreasedfrom 30 ~ 9 cm/~at 1 monthto 39 ~ 14 cm/sat 3 monthsafter operation (p = 0.024).Similarly,the transmittalA wave tendedto increasefrom46 & 22 cm/sat 1 monthto 54 f 25 cm/sat 3 monthsafter operation,comparableto the normal valuefor age (9),and leveledoffthereafter.
triculardimensionsand fractionalshorteningdid not change beforeand afteroperationor differbetweenthe groups(Table 2). For CTR, LAD and functionalclass,observationtime pointswerecombinedas early(1 to 3 months)and late (1 to 2 years) after operation (Fig. 3). Although cardiac size as assessedby CTR remainedunchangedin the controlgroup after an initialdecrease,it continuedto decreaseonlyin the mazegroup(p = 0.011).Both groupsshowedreducedLAD afteroperation;however,in the controlgroup,LADincreased significantly(p = 0.03),returned to preoperativelevelsand becamesignificantly largerthan that in the mazegroup(p < 0.0015),in whomLAD was essentiallyunchanged(Fig. 3). Both atria appearedto contractmore efficientlyovertime in sinusrhythm,withdecreasingLAD.Increasedphysicalactivity andloadafterdischargeresultedin lessfatigueanddyspneain the mazegroup,indicatinga significantly improvedfunctional capacityin thesepatients(p <0.009, Fig.3). Medications. Among22patientsafterreparativeoperation (8 afterASDclosure,14after mitralvalvuloplasty) (Table1), anticoagulation couldbe discontinuedin 16patients(73$%) in the maze group.Medicationwas totallydiscontinuedin two control group patients and in seven maze group patients whoregainedatrialrhythmand contraction(fouraftertricuspid annuloplastywithASDclosure,three aftermitralvalvuloplasty). Late complications. Despite warfarinanticoagulation,one transientnecrologicischemicattackoccurredin a maze group patient with normal sinus rhythmand contraction4 months aftermechanicalmitralvalvereplacement,andcerebralinfarction occurred in one control patient with persistent AF 6 months after aortic valve replacement.Intracranialbleeding requiringadmission to the hospital occurred in one maze grouppatientand two controlgrouppatients;all were receiving warfarinfor anticoagulation.
Discussion Studydesignand limitations. This retrospectivestudywas carriedout to evaluatethe risksandbenefitsof combiningthe maze procedure with surgicalinterventionfor underlying disorderscausingAF; however,a prospective,randomized studyshouldhavebeen performed.After beinginformedof the initialresults,everypatient requestedthe combinedapproach despitepotentiallyincreasedrisksand undetermined efficacy.Although case-matchedcontrol patients were included, obviousdifferencesin the date of operation and treatmentof postoperativeAF couldnot be eliminated,underminingthe inferencesderivedfrom the results.The decisionto addthe mazeprocedurerestonwhetherthepotentially greater risksof increasedcomplexityare outweighedby the benefitsfromregainedatrialrhythmandcontraction.Thus,the risk-benefitbalancemay help to determineindicationsfor the combinedapproach,whichshouldbe differentfromindicationsfor the isolatedmazeprocedurefor loneAF (2).
JACC Vol. 28, No. 4 October 1996:985-90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
989
implantation. Althoughtheincidenceofsinusdysfunction requiringatrialpacing waslowerthanin the seriesof Coxet al.(2,10), 66frequentcomplications promptedusto furthermodifytheproce64dure(7),resultinginno needforartificialpacingamongthenext 37patientsundergoingthe currentmodification. Timesrequired 62forthecombinedoperationswere-40,70 and150minlongerfor 60cardiacarrest,cardiopuhnonary bypassandthe completeoperation,respectively, than for the controloperationsalone.Never58theless,thesetimesweresignificantly shorterthanthoserequired 56for the mazeIII procedureof Coxet al. (10)combinedwith 54 J operationsfororganiclesionsin a similarcohort(11),suggesting LATE EARLY PRE the simplicityof the current modikation. Preoperativeand 60postoperativeleftventriculardimensionsand contractionwere LAD(m) T 58simifarbetweenthe groups,suggesting that longercardiacisch56emia does not result in persistentor significantventricular 54dysfunction.Despiteinitiafcomplicationsattributableto the 52combinedmaze procedure,subsequentmodifications and in50creasedexperienceimprovedresultsand alloweduneventful recoveryin the restof the mazegroupdespiteextendedindica48tions. 46Benefits. Evenwhen the combinedapproachwas consid44. ered to havefailedto restoresinusrhythmor to defibrillatein 42 J LATE &lRLY PRE allpatientswhorequiredelectricalcardioversion(n = 17),AF 2.61 wassurgicallyablatedin the remainingtwo thirdsof patients NYHA (34of 51)afterthe combinedapproach.Thisnumberis much 24 Ii higherthan that in currentcontrolgroupsincludingtwiceas 1 2.2manypatientswitha briefhistoryofAF,whoweremorelikely 2to undergodefibrillation(12).Satoet al. (6)reportedthatAF 1.6wasabolishedin only2870ofpatientslateafteroperationwith aggressivetreatmentusingrepeatedcardioversionafterrepair 1.6of underlyinglesionsalone. Even lower long-termmainte1.4nance of sinus rhythmwas reported by Hansen (5), who 1.2identifiedpreoperativeAF <12 monthsin durationasthe only IJ variableaffectingthe resultsof electricalcardioversionafter IATE PRE EARLY mitral valve surgery.These reports (5,6) discouragedthe Figure3. Changesincardiothoracic ratio(CTR),leftatrialdimension controlpatientsfromundergoingcardioversion, resultingin a (LAD)andNewYorkHeartAssociation(NYHA)functionalclassin significant d ifferencein treatment of postoperative AF and the maze(circles)versusthe controlgroup(squares)before(PRE),1 earlier operation.However,these factorsalone may not acto 3 months(EARLY)and 1 to 2 years after operation(LATE). Afthoughtherewereno significantdifferences(p = 0.824)in periop- countforthefactthat34patientshadno arrhythmiasrequiring erativeCTRreductionbetweenthe maze(n = 50)andcontrolgroups earlycardioversion, and 88%remainedAF free late after the (n= 39),CTRcontinuedto decreaselate afteroperationin the maze combinedapproach;althoughadditionof the mazeprocedure grouponly(p = 0.011),resultingina significant differencebetweenthe alone may not be solelyresponsiblefor these advantages groups(p = 0.029).Therewere alsono significantdifferences(p = either. In contrast,the rate of regainingatrial rhythmand 0.509)betweenthe maze (n = 32) and controlgroups(n = 8) in perioperativereductionin LAD; however,LAD returned to the contractionafterthe combinedoperationswasstilllowerthan preoperativelevellate afteroperationonlyin the controlgroup(p = afterloneAFwiththe isolatedmazeprocedure(2,10).Results 0.0015),resultingin a significantdifferencebetweengroups(p = of our recentreview(11)and analysis(13)indicatedthat the 0,012).A significantdifference(p = 0.036)in perioperativeimprovedifferenceappearedto resultfromthe durationof the arrhythment in functionalclassremainedlate after operationbetweenthe maze(n = 40)and controlgroups(n = 17),resultingin a significant mia and LAD rather than the technicalmodificationitself. Despite concern about recoveryof atrial contractility,the differencebetweengroups(p = 0.009).Verticalbars= 1 SE. majorityof patientswerefoundto havesignificantright(n = Risks.Risksassociatedwiththe combinedapproachinclude 41 [80%])and left atrialcontraction(n = 40 [78%])for late thecomplexity ofthemazeprocedure,whichrequiresprolonged diastolicventricularfilling.Moreover,atrial contraction(A cardiacarrestand cardiopulmonary bypasswithsubsequentde- wave)appearedto improveover time in sinusrhythm,with layedpostoperative recovery.Amongthefirst14patientsunder- decreasingLAD,approachingthe normalvaluefor age (9)by goingthe initialmodification, 3 (21%)requiredintraaorticbal- 3 months after operation.Thus, not only atrioventricular loon pumping, and 3 (21%) required atrial pacemaker synchrony,but also active diastolicventricularfillingwas 68-
——
990
JACC Vol. 28, No. 4 October1996:985–90
KAWAGUCHI ET AL. RISKS AND BENEFITS OF MAZE PROCEDURE
improved.Thisfindingmayaccountfor continuedreductionof cardiacsizein the mazegroup,as reportedby Gosselinket al. (14),whoobservedreductionin leftatrkdsizeonlyin patients remainingin .c$inus rhythmaftercardioversion. Improvedventricular Ming and reducedatrialsizemay accountfor improved functionalcapacityafter the combinedprocedure.Left atrial isolationhas been reported (15) to be less complicatedyet comparablyeffectivein restoringregularventricularcontraction or the RR intervalin 7190of patientslate after operation. Nonetheless, persistentAF inbothatriaaftertheproceduremay limithemodynamic improvement, whichwasnot demonstrated, andrequirecontinuedanticoagulation becausetheincreasedrisks inherentto AF persist(4).Of22patientsundergoingrepairwith ourcombinedapproach,anticoagulation wasnolongernecessary in 16(73%),andmedicationwastotallydiscontinued in7 (32%). Althoughtherewereno signitlcant differences intheincidenceof thromboembolic or bleedingcomplications in thecurrentreview withlimitednumbersandfollow-up, restorationof atrialrhythm and contractionby the mazeproceduremaymakea long-term differencein this cohortwithunderlyingorganiclesionsand a greaterriskof stroke(4). Conclusions. Becauseof the lackof prospectiverandomization, the present results are suggestiveand inferential. Nonetheless,resultsof the currentstudyconvincedus that the substantialbenefitsfromregainedatrialrhythmand contraction outweighthe potentialrisksof addingthe maze procedure,justifyingthe combinedapproachin patientswithestablishedAF andorganicdisorders,exceptfor a fewpatientswho wereconsideredunableto toleratethe procedure.Prospective randomizationseems warranted to examinethe value of addingthe mazeprocedurein patientswithAF ofrecentonset (12) and in those with presumedmaze-refractoryAF (13) undergoingoperationfor organiclesions. Weappreciatethe editorialassistanceof LeonardM.Linde,MD,Professorof Pediatrics(Cardiology), Universityof SouthernCalifornia,LosAngeles.
References L CoxJL.Thesurgicaltreatmentof atrialfibrillation.IV.Surgicaltechnique. J ThoracCardiovascSurg1991; 101:584-92. 2. CoxJL, BoineauJP, SchuesslerRB, Rater KM,LappasDG. Five-year experiencewiththe mazeprocedureforatrialfibrillation. AnnThoracSurg 1993;56:814-24. 3. KopeckySL,GershBJ,McGoonMD,et al.‘1’henaturalhistoryof loneatrial fibrillation:a population-based studyoverthree decades.N EnglJ Med 1987; 317:669-74. 4. ChesebroJH, FusterV, HalperinJL. Atrialfibrillation-riskmarkerfor stroke.N EnglJ Med1990;323:1556-8. 5. HansenJF, AndersonED, OlesenKH, et al. DC-conversion of atrial fibrillationaftermitralvalveoperation:an analysisofthe long-termresults. ScandJ ThoracCardiovascSurg1979; 13:267-70. 6. Sato S, KawashimaY, HiroseH, NakanoS, MatsudaH, ShimkuraR. Long-termresultsof direct-currentcardioversion afteropencommissurotomyfor mitralstenosis.AmJ Cardiol1986;57:629-33. 7. KosakaiY,Kawaguchi AT,IsobeF, et al.Coxmazeprocedureforpatients undergoing simultaneous surgery.J ThoracCardiovasc Surg1994; 108:1049– 55. 8. McAlpineWA.HeartandCoronaryArteries.Heidelberg:Springer-Verlag, 1975:151-9. 9. Miyatake& OkamotoM, KinoshitaN, et al. Augmentationof atriaf contributionto leftventricularinflowwithagingas assessedbyintracardiac Dopplerflowmetry. AmJ Cardiol1984;53:586-9. 10.CoxJL,BoineauJP, SchuesslerRB,JaquissRD,LappasDG.Modification of the mazeprocedurefor atrial flutter and atrial fibrillation.J Thorac Cardiovasc Surg1995; 110:473-84. 11.KoskaiY, KawaguchiAT, IsobeF, et al. Modifiedmazeprocedurefor patientswithatrialfibrillationundergoing simultaneous openheartsurgery. Circulation1995;92 Suppl11:11-359-64. 12.ChuaYL, SchaffI-IV,OrszulakTA, MorrisJJ. Outcomeof mitralvafve repair in patientswith preoperativeatrial fibrillation.Shouldthe maze procedurebecombinedwithmitralvalvuloplasty? J ThoracCardiovasc Surg 1994; 107:408-15. 13.Kawaguchi AT,KosakaiY,IsobeF, et al.Surgicalstratificationforpatients with atria] fibrillationsecondaryto organiccardiaclesionsundergoing surgery[abstract].Circulation1995;92 Suppl1:1-441. 14.Gosselink ATM,CrijnsHJGM,HamerHPM,HillegeH,LieRI.Changesin leftandrightatrialsizeaftercardioversion ofatrialfibrillation: roleofmitral valvedisease.J AmCoilCardiol1993;22:1666-72. 15.GraflignaA,PaganiF,Mirrzioni G,SalernoJ, ViganoM.Leftatrialisolation associatedwithmitralvalveoperation.AnnThoracSurg1992;54:1093-8.