Atrial Flutter: A Review of Its History, Mechanisms, Clinical Features, and Current Therapy Ken W. Lee, MD, MS, Yanfei Yang, MD, and Melvin M. Scheinman, MD Introduction W hile atrial flutter (AFL) was first recognized shortly after the birth of electrocardiography, its mechanism and therapy were hotly debateduntilrecently.Drugtherapyprovedtobenotoriouslypoor for those with recurrent AFL in that drugs failed to prevent recurrences and large doses of AV nodal blockers were often needed for rate control. Over the past decade monumental shifts have occurred with respect to betterdefinitionofthearrhythmiamechanismandtheremarkableefficacy of catheter ablative therapy. It is especially appropriate for this essay to review and take stock of where we have been, where we are, and where we hope to be in the future. Historical Perspectives ThestoryofAFLbeginsmorethanacenturyago.1,2Thefirstpublished description of AFL dates back to 1886 when McWilliam described observing regular, rapid excitations of the atrium in an animal.3 In 1906 Einthoven made an electrocardiographic recording of AFL.4 Character- istic sawtooth waves in the inferior ECG leads were described by Jolly andRitchie5in1911.TheseauthorswerethefirsttodistinguishAFLfrom atrial fibrillation (AF). In 1913 Lewis and coworkers6 also described the distinctive sawtooth waves. Lewis and his colleagues were the first to investigate the mechanism of this arrhythmia.7,8 Using a combination of epicardial maps and ECG recordings from a canine model of AFL induced by rapid atrial pacing, they showed that constant activation of at least some part of the atrium resulted in the flutter waves seen in the surfaceECG.Theyalsoshowedthattheactivationsequencewasorderly, ie, the wavefront circulated in either a cranial-caudo or a caudo-cranial CurrProblCardiol2005;30:121–168. 0146-2806/$–seefrontmatter doi:10.1016/j.cpcardiol.2004.07.001 CurrProblCardiol,March2005 121 direction in the right atrium.7 From this groundbreaking experimental work, Lewis and his colleagues concluded that AFL was due to intra- atrial circus movement around the vena cavae.8 Subsequent works that supported the notion that flutter was due to intra-atrialreentryincludedthoseofRosenbleuthandGarcia-Ramoswho constructed a crush injury model of this arrhythmia by creating a lesion between the vena cavae.9 Based on the epicardial maps, the authors deduced that the reentry loop circled around the atrial crush lesion. Interestingly, they also noted that when the crush lesion was extended from the inferior vena cava (IVC) to the AV groove, the arrhythmia disappeared and could not be induced. This important finding suggests that the true circuit may have included the cavotricuspid isthmus. Intra-atrialmacro-reentryasthemechanismofAFLwasnotuniversally accepted.Gotoetal10andAzumaetal11hadshownthataconitinecaused abnormalautomaticityatrapidratesintherabbitatria.Itwasthoughtthat iftheatrialaconitinesitefiredfastenough,eitherflutter(1:1conduction) or fibrillation (fibrillatory conduction because the atrial rate was too fast and 1:1 conduction could not be supported) occurred.12 Based on these and other works with aconitine,10,11,13-16 Scherf felt that flutter was due to abnormal automaticity. Building on the work of Rosenbleuth and Garcia-Ramos, Frame et al17-19 showed that the flutter reentry loop could exist outside of an atrial crush lesion. They created a “Y” lesion in the canine right atrium by extendingtheintercavalcrushlesiontotherightatrialfreewall.The“Y” lesionproducedacircuitthatrotatedaroundthetricuspidannulus.Similar fluttercircuitsmayexistinpatientswhohaveundergonerightatriotomies during repair of congenital heart defects.20,21 Over a span of nearly two decades, detailed experiments in various animal models and clinical studies have not only confirmed that the mechanismofflutterwasduetointra-atrialmacro-reentrybutalsosetthe stage for the development of curative catheter ablation therapy.22-30 Of particular importance were the elegant works of Waldo et al,27 Inoue et al,28 and Stevenson et al,29,30 all of described techniques of manifest and concealed entrainment. The latter allowed for identification of a site for catheter ablation. AFL as an arrhythmia that could be successfully ablated with radiofre- quency (RF) energy depended on the identification of a vulnerable, critical zone in the reentrant circuit. In 1986 Klein et al31 reported their findingsonintra-operativemappingstudiesoftwopatientswithpersistent flutter. They found that the narrowest part of the circuit had relatively slow conduction and localized to the low right atrium, between the IVC 122 CurrProblCardiol,March2005 TABLE1.Atrialflutteroccurrencecomparedwithatrialfibrillation Atrial Atrial Authors Patients fibrillation(%) flutter(%) Doliopoulosetal36 3780 24.6 0.7 Makinsonetal37 9458 7.6 0.2 KatzandPick38 50,000 11.7 0.5 and tricuspid ring. Furthermore, cryosurgical ablation of this critical region and its surrounding tissue prevented short-term recurrences of the arrhythmia. Subsequent studies by Chauvin and Brechenmacher32 and Saoudietal33usingdirectcurrent(DC)shockstodisruptthecriticalzone and eliminate the tachycardia supported the prospect that flutter could be permanently abolished by disruption of the isthmus. However, one drawbackofusingDCshockwasthattheshockitselfcouldconvertAFL. In the early 1990’s, groups led by Feld and coworkers34 and Cosio and coworkers35 found that disruption of the isthmus of the flutter circuit could be carried out safely with RF catheter ablation. B. Gersh: This is an interesting account of the history of atrial flutter and provides a perspective ranging from the initial recognition of this distinctive electrophysiologicentitytoanunderstandingoftheanatomiccharacteristics of the circuit, which in turn have led to the therapeutic applications of radiofrequency ablation. It is said that a historical perspective helps to understandboththepresentandthefuture,andinthisrespect,readersmight enjoy a scholarly and thoroughly entertaining review of a “close relative” of atrial flutter, namely, atrial fibrillation (Silvermann ME: From Rebellious PalpitationstotheDiscoveryofAuricularFibrillation:ContributionsofMack- enzie, Lewis, and Einthoben. Am J Cardiol 1994;73:384-9). Epidemiology, Risk Factors While detailed epidemiologic studies of AF have been available for almost two decades, similar studies of AFL have only been available for the past five years. AFL occurs less than one-tenth as often as AF (Table 1).36-38 Based on the 1990 Commission on Profession and Hospital Activity (CPHA) database, of the 517,699 discharges nationwide with arrhythmia as the principal diagnosis, AF made up 179,018 (34.6%), whileAFLmadeup23,420(4.5%).39RecentstudiesfromtheMarshfield Epidemiologic Study Area (MESA) database have reported that the overall incidence of AFL is about 88 per 100,000 person-years40,41 and haveestimatedthatthereare200,000newAFLcasesintheUnitedStates annuallywith80,000ofthesecasespresentingas“atrialflutteronly.”40,41 CurrProblCardiol,March2005 123 FIG1.Incidenceofatrialflutterbyageandgender(100,000person-years).(FromGranadaJ, etal.JAmCollCardiol2000;36:2242-6.) The incidence of AFL is about two to five times higher in men than in women, and like AF, AFL increases dramatically with age (Fig 1)41-43: the incidence of AFL in those younger than 50 years old is about 5/100,000 but rises sharply to 587/100,000 in those older than 80 years old.41 Besides advanced age and male gender, risk factors for AFL include heartfailure,chronicpulmonarydisease,previousstroke,andmyocardial infarction.41,44 Conditions associated with flutter include thyrotoxicosis, valvular heart disease, pericardial disease, congenital heart disease, post-open heart surgery, post-major noncardiac surgery, and especially postsurgical repair of congenital heart defects (eg, Mustard, Senning, Fontan) (Table 2).42,45,46 The possibility of a genetic predisposition for developing AFL is unclear. Preliminary studies suggest that a genetic cause may exist, although flutter presentation is more likely a manifes- tation of an important cardiac genetic abnormality that results in dilated cardiomyopathy and conduction system disease.47 While the morbidity and mortality associated with AF have been well recognized,39,48-51itisnotuntilrecentlythatthemorbidityandmortality specifically associated with AFL have been examined. This may be that AFL was previously grouped under the AF/AFL category. Studies have reported that AFL is associated with increased mortality,44,52,53 though not as high as AF or a combination of AF and AFL (Fig 2).44 124 CurrProblCardiol,March2005 TABLE2.Conditionsassociatedwithatrialflutter Valvularheartdisease(ie,rheumatic,mitral,tricuspid) Myocardialinfarction Pericardialdisease Cardiactumors Hypertrophiccardiomyopathy Congenitalheartdisease Postsurgicalrepairofcongenitalheartdefects Postcardiothoracicsurgery Postmajornoncardiacsurgery Severepulmonarydisease Pulmonaryembolus Thyrotoxicosis Acutealcoholintoxication FIG 2. Kaplan–Meier survival plots of subjects with atrial fibrillation, atrial flutter, or both arrhythmias,comparedwithcontrols.(FromVidailletH,etal.AmJMed2002;113:365-70.) Interestingly,intheircohortofpatientsVidailletetal44reportedthatearly mortality (first 6 months) was higher in patients with flutter than in controls but the number did not reach statistical significance, while late mortality was both increased and statistically significant. An intriguing question is whether early curative treatment of AFL will alter late mortality. CurrProblCardiol,March2005 125 B. Gersh: In a population-based study of “lone” atrial flutter from Olmstead County,Minnesota,Halliganetal(ref.53)demonstratedoveralongperiodof follow-up that atrial fibrillation developed in 56% of patients an initial diagnosis of lone atrial fibrillation. This emphasizes the close relationship between the two arrhythmias (ref. 53). Clinical Presentation, Diagnosis AFL usually occurs in paroxysms, lasting seconds to hours. Less commonly it exists as a stable, persistent rhythm. AFL is frequently associated with AF. Symptoms are most prevalent when flutter is paroxysmal and when the ventricular rate response is rapid. Palpitations are the most common symptom54; others include dyspnea, chest discom- fort, presyncope, and weakness. Syncope, in the absence of significant cardiac disease, is rare.55 AFL is not infrequently a precipitant of congestive heart failure in patients with significant cardiac disease. Patients with both AFL and AF may be more symptomatic than those with just AF because the heart rate tends to be more rapid during AFL, while AF is usually associated with increased AV nodal penetration and slower ventricular responses. Notable physical examination findings include a rapid peripheral pulse that is more often regular than irregular. Cannon “a” waves due to atrial contraction against a closed tricuspid valve may be observed. Cardiac auscultation may reveal a first heart sound of variable intensity: constant if the association of the atrial and ventricular contractions is maintained, and variable if it is not.56 In most cases, one can make the diagnosis of AFL with a 12-lead surfaceECG,lookingfordistinctivesawtoothwavesinleadsII,III,aVF, and V (Figs 3A and B). In cases when the flutter waves are not readily 1 discernible, interventions that transiently increase AV block to remove the QRS complexes can be helpful. They include vagal maneuvers (eg, carotid sinus massage, Valsalva maneuver) or administration of rapid- acting AV nodal blocking agents (eg, diltiazem, adenosine). An electro- gramobtainedwithanesophagealelectrodecanalsobeusedtomakethe diagnosis. AV conduction in AFL is usually 2:1 (Fig 3A), resulting in a regular rhythm,butconductionmaybevariable(Fig3B),resultinginanirregular rhythm. Rarely, 1:1 AV conduction can occur and may be lethal.57-59 Situations when 1:1 AV conduction can occur include use of drugs that slow the flutter rate and paradoxically increase the ventricular response, in patients with the Wolff–Parkinson–White syndrome and a short 126 CurrProblCardiol,March2005 FIG3.12-leadECG’softhesamepatientwithcounterclockwise(typical)atrialflutterwith2:1 AVblock(A)andvariableblock(B).Notethenegativeflutterwavesintheinferiorleadsand positivewavesinV . 1 antegrade refractory period of the accessory pathway, in those with acceleratedAVnodalconduction,orduringintensecatecholaminesurges (eg, exercise). ThesurfaceECGflutterwavemorphologycanprovideinsightsintothe specificmechanismofthecircuit.Readilydiscernibleflutterwavesinthe inferior and V leads are specific (up to 90%) for cavotricuspid-isthmus- 1 dependent (CTI-dependent) flutter.60,61 For example, negative flutter waves in the inferior leads and positive in V are suggestive of a 1 CurrProblCardiol,March2005 127 FIG4.SimultaneousECGandintracardiacelectrogramsofcounterclockwiseflutter.Notethe positiveflutterwavesinV andnegativewavesintheinferiorleads.DisplayedbelowtheECG, 1 the intracardiac recordings show activation from the lateral right atrium (TA5) to the septum (HBE).Theisthmusisactivatedfromlowlateraltricuspidannulus(TA1)tocoronarysinus(CS ). P HBE(cid:1)recordingfromtheregionofanteriorseptumaroundHisbundleregion;CS (cid:1)proximal P coronary sinus; CS (cid:1) middle of the coronary sinus; CS (cid:1) distal coronary sinus; TA (cid:1) M D recordings from the 20-pole, “halo” electrode catheter (see Fig. 7) positioned along the tricuspidannuluswithitsdistalpole(TA1)at7o’clockintheleftanteriorobliqueprojection,and proximalathighrightatrium(TA5). counterclockwise right atrial circuit with a lateral-to-medial activation over the CTI (typical) (Figs 3A, 4, and 8A), while positive flutter waves intheinferiorleadsandnegativeinV aresuggestiveofaclockwiseright 1 atrial circuit with medial-to-lateral activation over the CTI (Figs 5 and 8B). Attempts have been made to correlate flutter wave morphologies with non-CTI-dependent flutter mechanisms (see below). In most non- CTI-dependent flutter circuits, the surface ECG findings are non-specific andarenotpredictiveofthemechanismofthecircuit.Milliezetal62have examined the association of flutter wave morphologies of CTI-dependent AFL with echocardiographic findings and clinical patient characteristics. Theyhavefoundthataterminalpositivecomponentoftheflutterwavein the inferior leads in counterclockwise (CCW) right atrial AFL is associ- ated with heart disease, AF, and left atrial enlargement (left atrial 128 CurrProblCardiol,March2005 FIG5.SimultaneousECGandintracardiacelectrogramsofclockwiseflutter.Notethenegative flutterwavesinV andpositivewavesintheinferiorleads.Theintracardiacrecordingsshow 1 thattheisthmusisactivatedfromproximalcoronarysinus(CS )tolowlateralrightatrium(TA1). P The tricuspid annulus is activated in a clockwise fashion from isthmus to lateral right atrium (TA6). dimensiongreaterthan4cminthelongaxisviewand/orgreaterthan5.2 cm in the four-chamber view with transthoracic echocardiography). B.Gersh:Theauthorsmaketheusefulpointthatthesurfaceelectrocardiogram in this condition is really useful from a clinical standpoint, particularly when radiofrequency ablation is under consideration. The highest success rates are obtained in “typical” atrial flutter. This is well clarified by the authors in the subsequentdiscussionofmechanismsandtheclassificationutilizingTable3. Mechanisms, Nomenclature Atrial flutter is a condition in which, as has recently been shown, the contraction wave follows a circular and never ending path in the auricle, the circuits being completed a rate of from 240 to 350 per minute in different subjects. Sir Thomas Lewis 19207 CurrProblCardiol,March2005 129 TABLE3.Classificationofatrialflutter RightatrialCTI-dependentflutter Counterclockwise(CCW)flutter Clockwise(CW)flutter Double-wavereentry Lowerloopreentry Intra-isthmusreentry Rightatrialnon-CTI-dependentflutter Scar-relatedflutter Upperloopflutter Leftatrialflutter Mitralannularflutter Scar-andpulmonaryvein-relatedflutter Coronarysinusflutter Leftseptalflutter CTI(cid:1)Cavotricuspidisthmus. Counterclockwisevs.clockwisedirectionofwavefrontrotation,whenvisualizedfromtheleft anteriorobliquefluoroscopicview. AFL is a macro-reentrant arrhythmia. The atrial rhythm is regular (usually with a rate of 250-350/min) with little or no isoelectric interval on the ECG. Its activation pattern can usually be determined by detailed, intracardiac mapping studies. Over the years, terms used to describe various types of AFL have been ambiguousandhavecreatedagreatdealofconfusion.Theyincluderare, common, uncommon, typical, atypical, fast, slow, type I, type II, and left atrial. As we have learned more about the various mechanisms of this arrhythmia, the terminology of AFL has evolved. In 2001, the European Society of Cardiology and the North American Society of Pacing and Electrophysiology have published a treatise on AFL nomenclature.61 Morerecently,Scheinmanandhiscolleagues63haveprovidedanupdated classification and nomenclature. Since an understanding of AFL mecha- nismsisneededtorememberAFLterminology,wedescribethecurrently known AFL circuits below, modeled after the classification proposed by Scheinman et al.63 Table 3 provides a classification of AFL. Right Atrial Cavotricuspid-Isthmus-Dependent Flutter Fig6showstheanatomicstructuresofinterestinCTI-dependentflutter. Fig 7 shows a diagram of catheter arrangement in the right atrium frequently used in our electrophysiology laboratory to study flutter. CCW Atrial Flutter (Fig 8A) Classically referred to as typical, this is the most common type of flutter and makes up about 90% of clinical cases.61 It is the most 130 CurrProblCardiol,March2005