Author + information
- Received November 13, 2001
- Revision received March 12, 2002
- Accepted April 1, 2002
- Published online June 19, 2002.
- Richard L Page, MD, FACC*,* (, )
- Richard E Kerber, MD, FACC†,
- James K Russell, PhD‡,
- Tom Trouton, MD§,
- Johan Waktare, MD∥,
- Donna Gallik, MD, FACC¶,
- Jeff E Olgin, MD, FACC#,
- Philippe Ricard, MD**,
- Gavin W Dalzell, MD††,
- Ramakota Reddy, MD, FACC‡‡,
- Ralph Lazzara, MD, FACC§§,
- Kerry Lee, PhD∥∥,
- Mark Carlson, MD, FACC¶¶,
- Blair Halperin, MD, FACC##,
- Gust H Bardy, MD, FACC***,
- BiCard Investigators
- ↵*Reprint requests and correspondence:
Dr. Richard L. Page, Department of Internal Medicine (Cardiology, Clinical Cardiac Electrophysiology), University of Texas Southwestern Medical Center, Room CS7.102, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9047, USA.
Objectives This study compared a biphasic waveform with a conventional monophasic waveform for cardioversion of atrial fibrillation (AF).
Background Biphasic shock waveforms have been demonstrated to be superior to monophasic shocks for termination of ventricular fibrillation, but data regarding biphasic shocks for conversion of AF are still emerging.
Methods In an international, multicenter, randomized, double-blind clinical trial, we compared the effectiveness of damped sine wave monophasic versus impedance-compensated truncated exponential biphasic shocks for the cardioversion of AF. Patients received up to five shocks, as necessary for conversion: 100 J, 150 J, 200 J, a fourth shock at maximum output for the initial waveform (200 J biphasic, 360 J monophasic) and a final cross-over shock at maximum output of the alternate waveform.
Results Analysis included 107 monophasic and 96 biphasic patients. The success rate was higher for biphasic than for monophasic shocks at each of the three shared energy levels (100 J: 60% vs. 22%, p < 0.0001; 150 J: 77% vs. 44%, p < 0.0001; 200 J: 90% vs. 53%, p < 0.0001). Through four shocks, at a maximum of 200 J, biphasic performance was similar to monophasic performance at 360 J (91% vs. 85%, p = 0.29). Biphasic patients required fewer shocks (1.7 ± 1.0 vs. 2.8 ± 1.2, p < 0.0001) and lower total energy delivered (217 ± 176 J vs. 548 ± 331 J, p < 0.0001). The biphasic shock waveform was also associated with a lower frequency of dermal injury (17% vs. 41%, p < 0.0001).
Conclusions For the cardioversion of AF, a biphasic shock waveform has greater efficacy, requires fewer shocks and lower delivered energy, and results in less dermal injury than a monophasic shock waveform.
Atrial fibrillation (AF) is the most common arrhythmia requiring treatment and is expected to increase in frequency as the population ages (1–4). Transthoracic cardioversion is an important component of AF treatment. Traditionally, cardioversion used monophasic shocks, primarily with a damped sine waveform. Monophasic shocks are limited by failure to convert, typically occurring around 10% of the time (5–7)but ranging up to 20% (8)or 30% (9), depending on technique and patient population characteristics. Techniques to increase cardioversion success have included maneuvers such as alternate electrode placement (10), compression on the electrodes (11)and simultaneous shocks from two defibrillators (12). Ibutilide infusion before shock has shown promise in raising the success rate to 100% (13,14), although it carries a proarrhythmic risk. The limitations of monophasic shocks, and the success of the biphasic waveform for internal and external cardioversion of ventricular fibrillation, led to the development of biphasic cardioversion of AF. Ricard et al. (7)and Mittal et al. (8)have published reports of improved cardioversion of AF with biphasic waveform. We present the results of a multicenter, randomized, double-blind, international trial comparing monophasic and biphasic waveform shocks for the conversion of AF.
Patients enrolled were 18 years or older, hemodynamically stable and scheduled for elective cardioversion of AF. Exclusion criteria were epicardial defibrillator electrodes, pacemaker dependence, participation in a double-blind antiarrhythmic trial, dependence on vasopressors or inability to place defibrillation electrodes in the positions defined by the study.
Proper anticoagulation for AF of >48 h duration was required, as defined by either three weeks with international normalized ratio ≥2.0 or intravenous heparin with transesophageal echocardiogram (15)negative for left atrial (LA) thrombus; anticoagulation was required for the four weeks after cardioversion. For AF known to be of <48 h duration, no specific prior anticoagulation protocol was required.
Atrial fibrillation was defined by absence of organized atrial activity, presence of AF waves and irregularly irregular ventricular activity.
Identical cardioversion electrodes were used for all procedures, measuring 10.6 cm × 15.8 cm. The posterior electrode was placed to the right of the patient’s spine at the midscapular level, while the anterior electrode was placed to the left of the sternum with its upper edge at the level of the fourth anterior intercostal space.
Anesthesia method was at the discretion of the individual institution, although propofol was recommended.
Electrocardiographic (ECG) data were obtained from a 12-lead ECG before the procedure and a multichannel rhythm strip, a three-lead Holter monitor (model 20/20D, Phillips Medical Systems, Zymed, Oxnard, California) and the defibrillator recorder during the procedure.
The defibrillators were outwardly identical other than for the serial number. The monophasic device was a standard unit (CodeMaster XL M1722B, Hewlett Packard, Palo Alto, California), with modified labeling so the investigator could not determine from the energy labeling which shock waveform was being administered. The biphasic device contained circuitry to deliver an impedance-compensated biphasic shock waveform (same as the Heartstream XL M4735A [Philips Medical Systems, Seattle, Washington], then in development). The waveforms are shown in Figure 1. The devices were labeled at four energy levels: 100 J, 150 J, 200 J and 200/360 J. The final level either delivered a repeat shock of 200 J biphasic or provided 360 J monophasic. Both devices provided an impedance measurement for each shock.
The study was approved by the respective institutional review boards, and informed consent was obtained from each patient. The two defibrillators were placed side by side on a cart. Randomization by sealed envelope was used to determine which device would be placed, out of the view of the investigator, on the left of the cart—the location used for all cardioversion attempts before crossover.
All shocks were synchronized to the QRS complex on the surface ECG. Up to five shocks were delivered per protocol. The first four shocks were delivered by the assigned device, with progression from 100 J, 150 J, 200 J and then either 200 J biphasic or 360 J monophasic. If these four shocks were ineffective, the patient crossed over to the alternative device at maximum shock output (200 J biphasic or 360 J monophasic).
The protocol was terminated by conversion to another rhythm or by completion of the five-shock sequence. Subsequent therapy was at the investigator’s discretion, using standard equipment. Successful conversion was defined by two consecutive P waves uninterrupted by AF occurring any time within 30 s of the shock. If the shock resulted in atrial flutter or atrial tachycardia that required further intervention, the protocol was terminated and the shock was not counted as a success.
Skin burns, common after monophasic waveform shocks (16), were assessed 24 h to 48 h after the procedure, by either direct inspection or telephone interview. A prospectively defined scale was employed, defining burn as: none; mild (erythema, no tenderness); moderate (erythema, tenderness) or severe (blistering or necrosis, tenderness). The investigators remained blinded to the waveform used throughout the study.
The study was powered to detect a difference of 20% in first shock (100 J) efficacy, in a two-tailed fashion, with a confidence level of 95% and power of 80%. This called for 94 patients in each arm. To allow for errors in enrollment and patient drop-out, a target of 105 patients per arm was chosen.
A data and safety monitoring board reviewed the progress monthly, blinded to treatment until six months, when they assessed for justification of early termination in an unblinded fashion. There were no safety concerns nor were criteria for early termination met, so the trial was completed.
All ECG rhythms were reviewed by an event committee of two cardiac electrophysiologists who resolved discrepancies by consensus. ST segments after successful shocks were measured by one cardiac electrophysiologist from digital ambulatory ECG recorders (20/20D recorders, Philips Medical Systems, Zymed, Oxnard, California). All ECG review was blinded as to treatment.
Binary counts were evaluated with Fisher exact test, multiple alternative counts with the Fisher-Freeman-Halton test, using StatXact, version 4.0.1 (Cytel Software, Cambridge, Massachusetts), which also calculated the adjusted relative risks and their exact 95% confidence intervals. Other statistical tests were performed with Statistica, version 5.5 (StatSoft, Tulsa, Oklahoma). Ordinal data and non-Gaussian continuous data (checked with the Shapiro-Wilks W test) were evaluated with nonparametric Kruskal-Wallis analysis of variance (ANOVA). Other continuous data were evaluated by ANOVA. Multivariate analysis of binomial outcomes, under logistic transformation, was performed on mixed categorical and continuous variables using a generalized linear/nonlinear model with stepwise backward removal on the basis of the Wald statistic with a retention criterion of p < 0.05. Significance criteria in multiple sets of analyses (post-shock rhythms, ST levels) were adjusted by the Bonferroni method.
The study enrolled 210 patients for 212 procedures at 11 sites between March 3, 1999, and March 29, 2000. In two cases a patient was enrolled twice, and the second procedure was excluded. One patient was excluded due to incorrect electrode placement, and six patients were excluded due to later assessment that the original rhythm was not AF. The remaining 203 cases are analyzed here.
Patient characteristics and demographic data were similar between the treatment groups (Table 1). There was no difference in age, gender, weight, New York Heart Association (NYHA) congestive heart failure class, ejection fraction or LA size. Causes of heart disease and use of cardiac medications were similarly distributed. The duration of AF was similar, with just 13% of each group having AF of <48 h.
Results of conversion
There was a significantly greater probability of conversion with the biphasic waveform at 100 J (60% biphasic vs. 22% monophasic, p < 0.0001), at 150 J (77% vs. 44%, p < 0.0001) and at 200 J (through three shocks: 90% vs. 53%, p < 0.0001). After four shocks, including repeat 200 J biphasic or escalation to 360 J monophasic, the cumulative success was similar (91% biphasic vs. 85% monophasic, p = 0.29) (Table 2, Fig. 2).
Crossover occurred in 14% of the monophasic patients versus 6% of the patients assigned to the biphasic defibrillator (p = 0.10). The resultant sample is small, but it is of interest that 9 of 15 patients who crossed-over to biphasic were converted, while only 1 of the 6 who failed to convert with biphasic shocks was converted when crossed-over to monophasic cardioversion (p = 0.15).
A significantly greater number of shocks and greater total energy were delivered to the patients randomized to monophasic shocks (Fig. 3). With the monophasic waveform, 2.8 ± 1.2 shocks and 548 ± 331 J were required, compared with 1.7 ± 1.0 shocks and 217 ± 176 J for biphasic shocks (p < 0.0001 for both comparisons). Peak currents were higher for monophasic than for biphasic shocks at each energy setting (100 J: monophasic 22 ± 3 A, n = 107, biphasic 19 ± 4 A, n = 96, p < 0.0001; 150 J: monophasic 28 ± 4 A, n = 83, biphasic 22 ± 4 A, n = 36, p < 0.0001; 200 J: monophasic 32 ± 4 A, n = 60, biphasic 26 ± 2 A, n = 43, p < 0.0001) and were particularly high for monophasic 360 J shocks (44 ± 4 A, n = 54).
Assessment of the degree of skin burn was performed by telephone interview (76%) or direct inspection (24%); no significant difference in the degree of severity was observed between the two follow-up methods. Biphasic shocks were associated with less dermal injury (Fig. 4, p < 0.0001). Tenderness with erythema was present in 17% with biphasic, compared with 41% after monophasic shocks. Blistering or necrosis occurred in two patients, both of whom received exactly four monophasic shocks. Further evidence for the deleterious effect of high energy shocks in the monophasic arm is apparent in the progressive tendency for skin tenderness (moderate or severe skin burn) with increasing energy in that study arm (p = 0.0006), an effect not evident in the biphasic arm (p = 0.77) (Table 3).
Multivariate analysis of significant predictors identified by univariate analysis identified shock waveform (p < 0.0001), patient weight (p = 0.0007), duration of current AF episode (p = 0.0008) and transthoracic impedance (p = 0.005) as independent predictors of first shock success. Of these, only duration of current AF episode (p = 0.0001) and transthoracic impedance (p = 0.02) remained significant predictors of success for patients treated with four or fewer shocks from the initial device.
Overall, regardless of shock waveform, patients failing first shock cardioversion weighed more (93 ± 23 kg vs. 80 ± 17 kg, p = 0.0007) and had higher transthoracic impedance (83 ± 16 ohms vs. 71 ± 19 ohms) than patients successfully converted on the first shock and had longer durations of AF (Table 4). The latter was true regardless of treatment (monophasic, p = 0.04, biphasic p = 0.02, Table 4). When conversion within the initial four shocks was considered, this tendency was no longer statistically significant.
Patients who failed cardioversion with the initially randomized device (failing to convert with all four shocks) had higher transthoracic impedances than those who were successfully converted (86 ± 13 vs. 77 ± 19, p = 0.04). This influence was confined to the monophasic population; among monophasic patients, transthoracic impedances were higher for those who failed cardioversion in four or fewer shocks (91 ± 13 ohms vs. 79 ± 19 ohms, p = 0.02), while among biphasic patients there was no difference (76 ± 6 ohms vs. 76 ± 18 ohms, p = 0.94).
There were no significant differences between the treatment populations in ECG rhythms at 5 s or 10 s, at 1 min or 5 min or 1 h after the shock, providing successful cardioversion had occurred. After 1 h 68/70 (97%) of the biphasic and 77/85 (91%) of the monophasic patients were in sinus rhythm (p = 0.11).
Median ST levels and ST deviations were <200 μV at all time points measured (5 s, 10 s, 1 min, 5 min, 1 h postshock) for both treatments. In 20 comparisons (2 leads × 5 time points × [levels or deviations]), no statistically significant difference was seen.
This study demonstrates the advantages of biphasic shocks over monophasic shocks in conversion of AF. In a double-blind, randomized trial in well-matched populations, a step-up protocol with biphasic shocks resulted in greater cumulative success at energies of 100 J, 150 J and 200 J, although success with the four shock protocol (through biphasic repeat 200 J or monophasic 360 J) was similar. The biphasic waveform also required fewer shocks and less energy and produced less dermal injury.
The biphasic shock waveform was the strongest independent predictor of initial shock success. Other independent predictors were lower weight, shorter duration of AF episode and lower transthoracic impedance, consistent with findings in other studies with both monophasic (5,17)and biphasic defibrillators (8). The only independent predictor of success within the first four shocks was lower transthoracic impedance. The influence of transthoracic impedance was restricted to the monophasic arm; it had no residual effect in the four-shock protocol with the biphasic waveform.
Previous studies with biphasic cardioversion
Ricard et al. (7)and Mittal et al. (8)previously compared biphasic and monophasic shock waveforms for transthoracic conversion of AF. The present study agrees with those studies in demonstrating the superiority of biphasic shocks, but important methodological differences must be recognized. The Ricard et al. study (7)was relatively small (57 patients), involved only a single-center, employed anterolateral pad placement and had only a two-step protocol (either 150 J to 360 J monophasic or fixed 150 J biphasic). The Mittal et al. study (8)was multicenter and used pad placement similar to that used here but with different shock protocols (100 J, 200 J, 300 J, 360 J monophasic vs. 70 J, 120 J, 150 J, 170 J biphasic) and different biphasic waveform shape (“rectilinear” first pulse). There are differences among the studies in patient population as well, particularly in duration of AF, which emerged in our study as an important predictor of first shock success. While very few (7%) of the patients in the Ricard et al. study (7)had brief AF, (defined as <7 days), it is unclear how many had longer-term AF because all durations >7 days are considered “chronic.” The classification in the Mittal et al. study (8)is similar to that used here except that it does not separate patients with AF of greater than one year. The Mittal et al. study (8)included a higher proportion of patients with short (<48 h) (19%, compared with 13% here) and a lower proportion of patients with long (>6 months) (13%, compared to 25% here) duration AF (p = 0.007). This may, in part, account for the small difference in four-shock efficacy between their biphasic arm (94%) and that in the present study (91%). Finally, neither of the previous studies was blinded, and neither apparently used an independent event committee, both potential sources of investigator bias.
Although the findings are consistent with one report of experience with rectilinear biphasic shock (Mittal et al. ), they may not apply to other devices that employ alternate biphasic waveforms.
What protocol should be used clinically for biphasic cardioversion of AF?
Based on the present study and other data (7,8), it is reasonable to consider what protocol is appropriate for elective cardioversion of AF using a biphasic shock waveform. Because there are no data demonstrating that cardiac injury results from shock energies of up to 200 J biphasic, one could consider using 200 J (the maximum output on some biphasic devices) as the appropriate first choice; in fact, this may be appropriate for patients with AF >1 year duration. On the other hand, 150 J appears to have approximately 80% likelihood of success, so it represents a reasonable compromise between the desire to limit any possible deleterious energy effects on the myocardium and the efficient accomplishment of conversion. Failure with this first shock would warrant advancing to 200 J, however. Finally, for AF of <48 h duration, a first shock of 100 J could be justified, as it results in 80% conversion.
Independent of the initial energy selected, if 200 J is unsuccessful in conversion, one may consider alternate electrode placement. Although studies of the effect of electrode placement have yielded mixed results (6,10), in some cases changing the electrode position results in successful conversion when previously unsuccessful (at least for monophasic shocks) (11). Thus, after one or two failed cardioversion attempts with 200 J biphasic waveform shocks, one might consider changing the placement of electrodes to an anterior location (high right parasternal/left apical), administration of an antiarrhythmic agent before repeat shock delivery, double shocks or even catheter cardioversion.
Investigators and Participating Institutions.University of Texas Southwestern Medical Center, Dallas, TX: Richard Page, MD (principal investigator [PI]), Lauren Nelson, RN, Jose Joglar, MD, Karthik Ramaswamy, MD; Antrim Area Hospital, Antrim, Northern Ireland, U.K.: Thomas Trouton, MD (PI), Siobhan O’Mullan, RN, Mark Harbinson, MD, Simon Baird, MD; University of Iowa Hospital, Iowa City, IA: Richard Kerber, MD (PI), Joyce Vavra, RN, Carol Erenberger, RN, James B. Martins, MD; St. George’s Hospital, London, England, U.K.: Johan Waktare, MD (PI), Ann-Marie Murtagh, RN, A. John Camm, MD; Cedars Sinai Medical Center, Los Angeles, CA: Donna Gallik, MD (PI), Carol Vaughan, RN; Indiana University School of Medicine, Indianapolis, IN: Jeff Olgin, MD (PI), Julie Haschel, RN, William Groh, MD, John Miller, MD; Hospital Nord, Marseille, France: Phillippe Ricard, MD (PI); Royal Victoria Hospital, Belfast, Northern Ireland, U.K.: Gavin Dalzell, MD (PI), Maggie McIntyre, RN, Denise Shearer, RN, Thomas Matthew, MD, Mark Harbinson, MD; Wright Patterson Air Force Base, OH: Ramakota Reddy, MD (PI), Mary Caldwell, RN; Harborview Medical Center, Seattle, WA: Keith Comess, MD (PI), Margaret Russell, RN, Frances DeRook, MD; University of Western Ontario, London, Ontario, Canada: Allan Skanes, MD (PI), Bonnie Spindler, RN.
Data and Safety Monitoring Board.Ralph Lazzara, MD (Chair), University of Oklahoma Health Sciences, Oklahoma City, OK; Mark Carlson, MD, University Hospitals of Cleveland, Division of Cardiology, Cleveland, OH; Kerry Lee, PhD, Duke Clinical Research Institute, Durham, NC.
Event Committee.Blair Halperin, MD, Oregon Health Sciences University, Portland, OR; Mark Carlson, MD, University Hospitals of Cleveland, Division of Cardiology, Cleveland, OH.
☆ Supported by a grant from Heartstream, Philips Medical Systems, Seattle, Washington.
- atrial fibrillation
- analysis of variance
- left atrial
- New York Heart Association
- Received November 13, 2001.
- Revision received March 12, 2002.
- Accepted April 1, 2002.
- American College of Cardiology Foundation
- American Heart Association
- Dalzell G.W.N.,
- Anderson J.,
- Adgey A.A.J.
- Ricard P.,
- Levy S.,
- Boccara G.,
- Lakhal E.,
- Bardy G.
- Mittal S.,
- Ayati S.,
- Stein K.M.,
- et al.
- Levy S.,
- Lauribe P.,
- Dolla E.,
- et al.
- Botto G.L.,
- Politi A.,
- Bonini W.,
- Broffoni T.,
- Bonatti R.
- Saliba W.,
- Juratli N.,
- Chung M.K.,
- et al.
- Kerber R.E.,
- Martins J.B.,
- Kienzle M.G.,
- et al.