our study, we found a positive correlation, althoughmodest, in patients who had sinus node dysfunction butnot in the control subjects who did not have sinus nodedysfunction. The correlation was even better after auto-nomic blockade. The reason for the differential effects ofadenosine in subjects who had sinus node dysfunctionand those who did not is not clear.

The sensitivity and specificity of the adenosine testwere quite close to those of the standard atrial over-drive pacing test in our study. We also found a posi-tive correlation of CMSNRT between the 2 methods,with an improved correlation after autonomic block-ade. Due to its simplicity and safety, the bolus aden-osine test may be a useful and quick bedside “confir-mation” of sinus node dysfunction. However, thesensitivity of adenosine may be lower in patients who

have “suspected” sinus node dysfunction. Our datashould not be used to address the utility of the bolusadenosine test to “diagnose” sinus node dysfunction inpatients who have syncope of unknown origin.

1. Lerman BB, Belardinelli L. Cardiac electrophysiology of adenosine. Circula-tion 1991;83:1499–1509.2. Belardinelli L, Isenberg G. Isolated atrial myocytes: adenosine and acetylcho-line increase potassium conductance. Am J Physiol 1983;244:H734–H737.3. Benedini G, Cuccia C, Bolognesi R, Affatato A, Gallo G, Renaldini E, VisioliO. Value of purinic compounds in assessing sinus node dysfunction in man: anew diagnostic method. Eur Heart J 1984;5:394–403.4. Resh W, Feuer J Wesley RC. Intravenous adenosine: a noninvasive test for sicksinus syndrome. PACE 1992;15:2068–2073.5. Burnett D, Abi-Samra F, Vacek JL. Use of intravenous adenosine as anoninvasive diagnostic test for sick sinus syndrome. Am Heart J1999;137:435–438.

Efficacy of Transthoracic Cardioversion of AtrialFibrillation Using a Biphasic, Truncated ExponentialShock Waveform at Variable Initial Shock Energies

Eric J. Rashba, MD, Michael R. Gold, MD, PhD, Fred A. Crawford, MD,Robert B. Leman, MD, Robert W. Peters, MD, and Stephen R. Shorofsky, MD, PhD

Biphasic shocks are more effective than damped sinewave monophasic shocks for transthoracic cardiover-sion (CV) of atrial fibrillation (AF), but the optimalprotocol for CV with biphasic shocks has not beendefined. We conducted a prospective, randomizedstudy of 120 consecutive patients with persistent AFto delineate the dose-response curve for CV of AFwith a biphasic truncated exponential shock wave-form and to identify clinical predictors of shock effi-cacy. Our data suggest that the initial shock energyfor CV with this waveform should be 200 J if thepatient weighs <90 kg and 360 J if the patientweighs >90 kg. �2004 by Excerpta Medica Inc.

(Am J Cardiol 2004;94:1572–1574)

This study enrolled 120 consecutive patients whounderwent elective transthoracic cardioversion

(CV) at the University of Maryland Medical Center,the Baltimore VA Medical Center, and the MedicalUniversity of South Carolina. Written informed con-sent was obtained from all patients, and the protocolwas approved by the institutional review board at allparticipating institutions. The inclusion criteria werepersistent atrial fibrillation (AF) (�48 hours) and age�21 years. Exclusion criteria were a contraindication

to CV (e.g., digoxin toxicity or left atrial thrombus bytransesophageal echocardiography), unstable angina,or myocardial infarction during the preceding 2weeks, or unwillingness or inability to give informedconsent. All patients underwent therapeutic anticoag-ulation (international normalized ratio �2.0) for �3weeks before CV or had left atrial thrombus excludedby transesophageal echocardiography.

• • •Standard self-adhesive defibrillation pads (Quan-

tum Edge Quik-Combo, Physio-Control Corp., Red-mond, Washington) were placed in the anteroposteriorposition. The anterior pad was placed in the rightparasternal area and the posterior pad was placed justbelow the left scapula.1 All patients were sedated with1% propofol under the supervision of an anesthesiol-ogist. Cardioversion was performed by delivering R-wave synchronized biphasic shocks using a Life-Pak12 external defibrillator (Physio-Control Corp.). Suc-cessful CV was defined as the presence of �1 clearlyvisible P wave within 30 seconds after administrationof the shock. Early recurrence of AF was defined asthe resumption of AF within 1 minute after a shockthat resulted in sinus rhythm for �1 beat.2 The pa-tients were randomly assigned to 1 of 4 initial biphasicshock energies. To fit a dose-response curve, the suc-cess rates of the 4 shock energies should be approxi-mately 10%, 30%, 70%, and 100%. Based on previousdata,3 we chose shock energies of 20, 50, 100, and 200J. If the first shock was unsuccessful, a 200-J shockwas delivered, followed by a maximum output 360-Jbiphasic shock if the second shock failed. If the 360-Jshock was unsuccessful, then additional 360-J shocks

From the Division of Cardiology, University of Maryland School ofMedicine, Baltimore, Maryland; and Division of Cardiology, MedicalUniversity of South Carolina, Charleston, South Carolina. Dr. Rashba’saddress is: Division of Cardiology, University of Maryland School ofMedicine, 22 South Greene Street, Room N3W77, Baltimore, Mary-land 21201-1595. E-mail: erashba@medicine.umaryland.edu.Manuscript received July 7, 2004; revised manuscript received andaccepted August 11, 2004.

1572 ©2004 by Excerpta Medica Inc. All rights reserved. 0002-9149/04/$–see front matterThe American Journal of Cardiology Vol. 94 December 15, 2004 doi:10.1016/j.amjcard.2004.08.044

could be administered with anterior chest pressure,ibutilide pretreatment,2 or after repositioning of theelectrodes at the discretion of the attending electro-physiologist. All patients underwent continuous te-lemetry monitoring throughout the procedure.

The efficacy of the 4 initial shock energies was com-

pared using the chi-square test. A pvalue of �0.05 was required for statis-tical significance. The sample size of30 subjects per group was chosen toprovide �80% power to detect a 25%difference in success rate between thestudy groups with a 2-tailed � of 0.05.A dose-response curve for CV with thefirst biphasic shock was fitted using asigmoidal function. Demographic datawere collected on each patient, includ-ing age, gender, duration of AF, his-tory of AF, coronary artery disease,hypertension, diabetes, height, weight,and cardiovascular medications. Echo-cardiographic data included left atrialsize and left ventricular ejection frac-tion. The clinical characteristics of thepatient groups were compared usingthe chi-square test for categorical vari-ables and analysis of variance for con-tinuous variables. Univariate and mul-tivariate logistic regression analyses

were performed to identify clinical predictors of requir-ing a 360-J shock for successful CV (STATA Version7.0; StataCorp LP, College Station, Texas). A p value�0.10 was required to enter a variable into the multivar-iate analysis.

The clinical characteristics of the study populationwere typical of patients with persistent AF who arereferred for elective CV (Table 1). The patients werepredominantly men and hypertensive, with moderateleft atrial dilation and mild depression of left ventric-ular function. One-third of the patients had a history ofsignificant coronary artery disease. The mean durationof AF was 4 months; in addition, 41% of the patientshad AF of indeterminate duration, and 50% had ahistory of AF. At the time of the CV procedure, 37%of the patients were treated with Vaughan-Williamsclass III antiarrhythmic drugs, and 6% were treatedwith class IC agents. The 4 randomization groupswere well matched with respect to clinical character-istics and cardiovascular medication use, with theexception of minor differences in left ventricular func-tion and �-blocker utilization (Table 1).

All patients completed the protocol without com-plications. The CV success rates for the first shock aredepicted in Figure 1. The observed CV success rateswere 7% at 20 J, 23% at 50 J, 63% at 100 J, and 83%at 200 J. The dose-response curve predicted CV suc-cess rates of 50% at 75 J, 80% at 175 J, and �95% at360 J. None of the 5 patients in whom the first 200-Jshock failed were successfully cardioverted with asecond 200-J shock. In contrast, 45 of 60 patients(75%) in whom the first shock failed with energies�200 J were successfully cardioverted with a second200-J shock. Only 1 subject in whom the first 2 shocksfailed could not be cardioverted with a 360-J shock.This patient was treated with ibutilide and CV wasreattempted with 360 J, without success. In 2 othersubjects, 360-J shocks were unsuccessful with theself-adhesive electrodes, even when manual pressure

FIGURE 1. Dose-response curve for CV of AF with the initialshock. The dose-response curve predicts CV success rates of 50%at 75 J, 80% at 175 J, and �95% at 360 J.

TABLE 1 Clinical Characteristics of the Study Population, Stratified byRandomization Group

20 J 50 J 100 J 200 J

Age (yrs) 65 � 12 69 � 13 65 � 12 63 � 10Men 20 (67%) 24 (80%) 21 (70%) 19 (63%)Atrial fibrillation duration (days) 71 � 80 86 � 100 136 � 177 176 � 371Hypertension 19 (63%) 17 (57%) 22 (73%) 19 (63%)Coronary artery disease 10 (33%) 9 (30%) 10 (33%) 8 (27%)Diabetes mellitus 4 (13%) 8 (27%) 8 (27%) 7 (23%)Previous AF 17 (53%) 14 (47%) 17 (53%) 14 (47%)Height (cm) 175 � 13 176 � 9 170 � 34 174 � 13Weight (kg) 93 � 21 96 � 20 94 � 28 100 � 31Left atrial size (mm) 44 � 13 50 � 7 45 � 10 47 � 7Left ventricular ejection fraction (%)* 50 � 16 41 � 16 50 � 13 50 � 15Cardiovascular medications

� blockers* 17 (53%) 26 (87%) 15 (50%) 24 (80%)Calcium channel blockers 5 (17%) 6 (20%) 8 (27%) 8 (27%)Digoxin 12 (40%) 9 (30%) 10 (33%) 9 (30%)Amiodarone 7 (23%) 6 (20%) 8 (27%) 9 (30%)Other class III agents 3 (10%) 3 (10%) 2 (7%) 7 (23%)Propafenone 1 (3%) 2 (7%) 3 (10%) 1 (3%)

*p �0.05.

TABLE 2 Univariate Predictors of Requiring �200 J forCardioversion

OddsRatio 95% CI p Value

Age (10 yrs) 0.54 0.36–0.83 0.004AF duration (wks) 1.02 1.00–1.04 0.037Height (10 cm) 2.27 1.30–3.97 0.004Weight (10 kg) 1.64 1.29–2.08 �0.001

BRIEF REPORTS 1573

was applied to the anterior patch. Both of these pa-tients were successfully cardioverted with another360-J shock delivered with paddles in the anterior–lateral position with vigorous manual pressure. Tenpatients (8.3%) had early recurrence of AF.

Of the 120 patients enrolled in this study, 20 re-quired a CV attempt with 360 J. The duration of AF,age, height, and weight were all univariate predictorsof a high-energy requirement (Table 2; p �0.05).Increased age was associated with lower odds of re-quiring 360 J, whereas increases in the duration of AF,height, and weight were associated with a greaterlikelihood of requiring 360 J. There was a moderatelystrong negative correlation between age and weight (r� �0.42, p �0.0001). Multivariate logistic regressionanalysis identified body weight as the only indepen-dent predictor of requiring �200 J for CV (p �0.001).The odds of requiring �200 J for CV increased 64%for each 10-kg increment in body weight. Shocks�200 J were required in 25% of patients who weighed�90 kg (the median weight in our study population)but in only 8% who weighed �90 kg (p � 0.01).

• • •The principal finding of this study is that maximum

output 360-J shocks are required for CV of AF with abiphasic truncated exponential waveform in a substan-tial minority of patients (�15%). The success rate for200-J biphasic shocks with this waveform is similar tothat of 360-J shocks with a monophasic damped sinewave waveform (80% to 85%). Furthermore, bodyweight is the best clinical predictor of high-shockenergy requirements for CV with both biphasic trun-cated exponential and monophasic waveforms.

Our results are consistent with previous studies, inthat the duration of AF and measures of body habitus(height and weight) were all significant univariatepredictors of shock efficacy.3–6 The association ofincreased age with lower odds of requiring 360 J islikely explained by the negative correlation betweenage and body weight that we observed. Our studydesign differs from previous reports of transthoracicCV with biphasic shock waveforms in several impor-tant respects.3–9 The first shock energy was random-ized (20, 50, 100, or 200 J) so that a dose-responsecurve for CV with biphasic truncated exponentialshocks could be constructed. This method is prefera-ble to drawing inferences regarding the appropriatechoice of initial shock energy from the results of astep-up protocol, because the probability of successfor higher energy shocks with a step-up protocol isinfluenced by the patient having already failed toconvert with lower energy shocks. Importantly, ourstudy population exclusively comprised patients withpersistent AF (�48 hours). Nearly 1/3 of the patientsin a previous study had AF durations of �48 hours,

which likely explains the better success rates that wereobserved with the same biphasic truncated exponentialwaveform.6 Our results represent the most commonclinical scenario, because patients with brief episodesof AF are typically treated conservatively in the ab-sence of uncontrolled heart rates or hemodynamicinstability.10

Based on our data, we propose the following pro-tocol for CV of persistent AF with the biphasic trun-cated exponential device that was employed in ourstudy: if the patient weighs �90 kg, an initial shockenergy of 200 J should be used, as it is effective in�90% of patients. Because the results of our studyand others indicate that a second 200-J shock is rarelyeffective if the first 200-J shock fails,3,6 a 360-J shockis recommended if the patient remains in AF. Forpatients who weigh �90 kg, the initial shock energyshould be 360 J. If the first 360-J shock is ineffective,adjunctive measures, such as anterior chest pressure,ibutilide pretreatment, or repositioning of the elec-trodes, should be employed.1,2 We do not advocate theuse of initial 360-J shocks in all patients becausehigher energy shocks have been associated with anincreased incidence of skin injury even when biphasicwaveforms are employed.6 Because other defibrillatormodels employ biphasic rectilinear3 or different bi-phasic truncated exponential4 waveforms, additionalstudies are required to determine if this protocol isapplicable to CV with these devices.

1. Ewy GA. The optimal technique for electrical cardioversion of atrial fibrilla-tion. Clin Cardiol 1994;17:79–84.2. Oral H, Souza JJ, Michaud GF, Knight BP, Goyal R, Strickberger SA, MoradyF. Facilitating transthoracic cardioversion of atrial fibrillation with ibutilidepretreatment. N Engl J Med 1999;340:1849–1854.3. Mittal S, Ayati S, Stein KM, Schwartzman D, Cavlovich D, Tchou PJ,Markowitz SM, Slotwiner DJ, Scheiner MA, Lerman BB. Transthoracic cardio-version of atrial fibrillation. Comparison of rectlinear biphasic versus dampedsine wave monophasic shocks. Circulation 2000;101:1282–1287.4. Page RL, Kerber RE, Russell JK, Trouton T, Waktare J, Gallik D, Olgin JE,Ricard P, Dalzell GW, Reddy R, et al, for the Bicard Investigators. Biphasicversus monophasic shock waveform for conversion of atrial fibrillation. J AmColl Cardiol 2002;39:1956–1963.5. Marinsek M, Larkin GL, Zohar P, Bervar M, Pekolj-Bicanic M, Mocnik FS,Noc M, Podbregar M. Efficacy and impact of monophasic versus biphasiccountershocks for transthoracic cardioversion of persistent atrial fibrillation. Am JCardiol 2003;92:988–991.6. Neal S, Ngarmukos T, Lessard D, Rosenthal L. Comparison of the efficacy andsafety of two biphasic defibrillator waveforms for the conversion of atrial fibril-lation to sinus rhythm. Am J Cardiol 2003;92:810–814.7. Ricard P, Levy S, Boccara G, Lakhal E, Bardy GH. External cardioversion ofatrial fibrillation: comparison of biphasic versus monophasic waveform shocks.Europace 2001;3:96–99.8. Khaykin Y, Newman D, Kowalewski M, Korley V, Dorian P. Biphasic versusmonophasic cardioversion in shock-resistant atrial fibrillation. J CardiovascElectrophysiol 2003;14:868–872.9. Scholten M, Szili-Torok T, Klootwijk P, Jordaens L. Comparison of monopha-sic and biphasic shocks for transthoracic cardioversion of atrial fibrillation. Heart2003;89:1032–1034.10. Fuster V, Ryden LE. ACC/AHA/ESC guidelines for the management ofpatients with atrial fibrillation: executive summary. J Am Coll Cardiol 2001;38:1231–1265.

1574 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 94 DECEMBER 15, 2004