Adenosine-Induced Atrial Arrhythmia: A Prospective Analysis

Methods

Patients

The study participants were 200 consecutive patients (74 men and 126 women) with PSVT who were referred for an electrophysiology procedure and possible ablation. The mean (±SD) patient age was 43 ± 16 years. No patient had structural heart disease. These patients had had symptoms from PSVT for a mean of 13.7 ± 12.1 years before the electrophysiology procedure. No patient had a clinical history of atrial fibrillation or atrial flutter. The mechanism responsible for PSVT during electrophysiologic evaluation was typical (n = 114) or atypical (n = 10) atrioventricular nodal reentrant tachycardia in 124 patients, atrioventricular reciprocating tachycardia with a manifest accessory pathway in 40 patients, a concealed accessory pathway in 28 patients, and atrial tachycardia in 8 patients.

Electrophysiologic Testing

The investigational protocol was approved by the Committee for Human Research at the University of Michigan. Patients gave informed consent for all procedures. Electrophysiology procedures were performed in the fasting state after therapy with all antiarrhythmic medications had been discontinued for at least 5 half-lives. Three 7-French quadripolar electrode catheters were inserted into the right femoral vein and positioned in the high right atrium, across the tricuspid valve to record the His-bundle electrogram, and in the right ventricle. Leads V₁, II, and III and the intracardiac electrograms were displayed on an oscilloscope and recorded on a Mingograph 7 recorder (Siemens-Elema, Solna, Sweden) at a paper speed of 100 mm/s. Pacing was performed with a programmable stimulator (Bloom Associates, Reading, Pennsylvania).

The goal of electrophysiologic testing was to induce and determine the mechanism of PSVT and to measure the conduction properties and the refractory periods of the atrioventricular node [15]. If tachycardia could not be provoked in the baseline state, isoproterenol (2 µg/min) was given intravenously and programmed stimulation of the right atrium and right ventricle was repeated. Isoproterenol was required for tachycardia induction in 71 patients. The likelihood of requiring isoproterenol for the induction of tachycardia was similar in patients with atrioventricular nodal reentrant tachycardia (n = 44 [57%]), those with atrioventricular reentrant tachycardia (n = 23 [54%]), and those with atrial tachycardia (n = 4 [50%]) (P > 0.2).

Study Protocol

Eligible patients were presumed to have PSVT and were referred for an electrophysiology procedure and possible ablation. Only patients with a history of reactive airway disease were excluded. After the mechanism of PSVT was determined, the investigational protocol was performed. During tachycardia, 12 mg of adenosine (Fujisawa USA, Inc., Deerfield, Illinois) was administered as a bolus as quickly as possible through a sheath positioned in the femoral vein. The adenosine bolus was immediately followed by a 20-mL bolus of normal saline. Termination of tachycardia and development of atrial fibrillation or atrial flutter were noted. The cycle length of atrial fibrillation or atrial flutter and the type of atrial fibrillation were determined [16]. The types of atrial fibrillation were defined as follows: type I, discrete atrial electrograms of variable structure with an isoelectric baseline; type II, discrete atrial electrograms of variable structure with a nonisoelectric baseline; type III, neither discrete atrial electrograms nor an isoelectric baseline; and type IV, a combination of type III and either type I or type II atrial electrograms [16].

The occurrence of atrial or ventricular premature complexes was also noted. Upon termination of PSVT by adenosine and upon induction of atrial fibrillation, the time from the preceding atrial complex to the premature atrial complex that induced atrial fibrillation was measured. The ratio of the coupling interval of the premature complex to the preceding cycle length was determined, as was the total duration of atrial fibrillation or atrial flutter. If atrial fibrillation persisted for more than 10 minutes, sinus rhythm was restored with cardioversion.

Statistical Analysis

Continuous variables are expressed as the mean ±SD and were compared by using an unpaired t-test. The frequency of induction of atrial fibrillation or atrial flutter with adenosine, based on the mechanism of PSVT, was compared by using analysis of variance. Nominal variables were compared by using the Fisher exact test. A P value less than 0.05 was considered statistically significant.

Results

Main Findings

Paroxysmal supraventricular tachycardia terminated after adenosine administration in 198 patients (99% [95% CI, 96% to 100%]) (Figure 1). Adenosine administration resulted in atrial fibrillation (n = 22) or atrial fibrillation and atrial flutter (n = 2) in 24 patients (12% [CI, 7.5% to 16.5%]) a mean of 6.3 ± 6.2 seconds (range, 0.4 to 20.7 seconds) after PSVT termination (Figure 2). The rhythm preceding the atrial premature complex that provoked the atrial arrhythmia was sinus rhythm in 21 patients and an atrial ectopic complex in 3 patients. When atrial fibrillation or atrial flutter occurred, 20 patients had atrioventricular block and 4 patients had a prolonged P-R interval. In 1 patient, atrial fibrillation lasting 22 seconds precipitated atrial flutter that lasted 3.4 minutes; in another patient, atrial flutter lasting 1.5 minutes provoked atrial fibrillation that lasted 5.3 minutes.

View larger version:

• In this window
• In a new window

Figure 1. The mechanism responsible for PSVT was atrioventricular reentrant tachycardia that used a manifest left free-wall accessory pathway. From top to bottom, the figure shows leads V1, I, and II; the high right atrial (HRA) bipolar recording; a low septal right atrial (LRSA) recording; the right ventricular (RVA) bipolar recording; and lead III. A representative example of electrocardiographic recordings of paroxysmal supraventricular tachycardia (PSVT) termination after adenosine administration.

View larger version:

• In this window
• In a new window

Figure 2. The mechanism responsible for PSVT was atrioventricular reentrant tachycardia that used a left free-wall accessory pathway. The atrial premature complex that induced atrial fibrillation ( ) occurred 180 milliseconds after the preceding sinus beat (arrowhead). The ratio of the atrial premature complex cycle length to the preceding sinus cycle length was 0.29. The mean preexcited R-R rate during atrial fibrillation was 162 beats/min, and the shortest preexcited R-R interval was consistent with a heart rate of 200 beats/min. This figure and show recordings from the same patient. From top to bottom, the figure shows leads V1, I, and II; the high right atrial (HRA) bipolar recording; a low septal right atrial (LRSA) recording; the right ventricular (RVA) bipolar recording; and lead III. Electrocardiographic recordings showing atrial fibrillation that developed 8.8 seconds after paroxysmal supraventricular tachycardia (PSVT) was terminated by adenosine administration (shown in Figure 1).arrowFigure 1

Type I atrial fibrillation occurred in 1 patient, type III occurred in 10 patients, type IV occurred in 13 patients, and atrial flutter occurred in 2 patients. No patient had type II atrial fibrillation. The mean ventricular rate during atrial fibrillation was 107 ± 43 beats/min. The mean duration of the atrial arrhythmias was 5.6 ± 6.7 minutes (range, 8 seconds to 20.7 minutes). Sinus rhythm was restored spontaneously in 16 patients (67%), and cardioversion was required to restore sinus rhythm in 8 patients (33%).

Risk Factors for Adenosine-Induced Atrial Arrhythmias

Among the 24 patients who developed atrial fibrillation or both atrial fibrillation and atrial flutter, 13 had atrioventricular nodal reentrant tachycardia (11% of all patients with atrioventricular nodal reentrant tachycardia), 9 had atrioventricular reentrant tachycardia (13% of all patients with atrioventricular reentrant tachycardia), and 2 had atrial tachycardia (25% of all patients with atrial tachycardia) (P > 0.2). The frequency of atrial arrhythmia induction was similar in patients with typical (11%) and atypical (10%) atrioventricular nodal reentrant tachycardia (P = 1.0) and in patients with manifest (15%) and concealed (11%) accessory pathways (P > 0.2). During the electrophysiology procedure, atrial fibrillation or atrial flutter only developed immediately after the administration of 12 mg of adenosine. No patient had a history of atrial fibrillation or atrial flutter.

After PSVT termination, one or more ventricular or atrial premature complexes occurred in 17% and 63% of patients, respectively. Occurrence of a ventricular premature complex was not associated with the development of atrial fibrillation (P = 0.2). However, atrial premature complexes occurred in all 24 patients who developed atrial fibrillation or both atrial fibrillation and atrial flutter and in 102 of the 176 patients (58%) who did not develop an atrial arrhythmia (P < 0.001) (Figure 2). The time from the preceding beat to premature atrial complex was 265 ± 124 milliseconds (range, 80 to 530 milliseconds) when atrial fibrillation or atrial flutter developed (Figure 2) and 436 ± 141 milliseconds (range, 200 to 920 milliseconds) when atrial fibrillation or atrial flutter did not develop (P < 0.001). Furthermore, the ratio of the coupling interval of the atrial premature complex to the preceding atrial cycle length was 0.37 ± 0.16 when atrial fibrillation was induced after adenosine administration (Figure 2) and 0.49 ± 0.16 when atrial fibrillation was not induced (P = 0.002).

The development of an atrial arrhythmia after the administration of adenosine was not associated with age, sex, PSVT cycle length, the requirement of isoproterenol infusion for PSVT induction, the atrioventricular block cycle length, the ventriculoatrial block cycle length, or the R-R interval immediately preceding the development of atrial fibrillation (P > 0.2 for the first six variables; P = 0.2 for the last variable). In addition, atrial overdrive pacing was used to terminate PSVT 6.2 ± 3.1 times per patient (range, 5 to 20 times per patient) and never led to atrial fibrillation or atrial flutter.

Ventricular Response during Adenosine-Induced Atrial Fibrillation

The mean ventricular rate during adenosine-induced atrial arrhythmias was similar in patients with atrioventricular nodal reentrant tachycardia (105 ± 46 beats/min), those with atrioventricular reentrant tachycardia (114 ± 41 beats/min), and those with atrial tachycardia (91 ± 24 beats/min) (P > 0.2). Among the six patients with accessory pathways that were capable of anterograde conduction, four developed preexcited atrial arrhythmias. The mean preexcited R-R rate during atrial fibrillation was 150 ± 36 beats/min (range, 102 to 188 beats/min), and the mean maximum preexcited R-R rate was 176 ± 51 beats/min (range, 102 to 214 beats/min). Preexcited atrial flutter occurred in one patient who also developed atrial fibrillation. In this patient, the mean preexcited R-R rate during atrial flutter was 188 beats/min, and the maximum preexcited heart rate was 214 beats/min.

Discussion

Major Findings

Intravenous administration of a 12-mg dose of adenosine resulted in PSVT termination in 99% of patients, and atrial fibrillation or both atrial fibrillation and atrial flutter subsequently developed in 12% of patients. Regardless of the mechanism responsible for PSVT (that is, atrioventricular nodal reentrant tachycardia, atrioventricular reentrant tachycardia, or atrial tachycardia), 12% of patients developed atrial arrhythmias an average of 6 seconds after intravenous amiodarone terminated PSVT.

The induction of atrial arrhythmia usually depends on a premature atrial complex, which causes a “long-short” sequence. The atrial electrograms obtained during atrial fibrillation are usually very rapid and disorganized (reflecting type III or type IV atrial fibrillation [16]), and the arrhythmia usually lasts for several minutes and then terminates spontaneously. In one third of our patients, however, atrial fibrillation lasted longer than 10 minutes. When an accessory pathway is present and capable of anterograde conduction, preexcitation of the ventricles by the accessory pathway often results if an atrial arrhythmia develops secondary to adenosine administration. In our study, 12 mg of adenosine administered centrally resulted in atrial fibrillation in 13% of patients with an accessory pathway capable of anterograde conduction.

Mechanism of Induction of Atrial Fibrillation

We could not determine the mechanism by which adenosine induces atrial fibrillation, but at least two mechanisms may be solely or partially responsible. First, the induction and maintenance of atrial fibrillation are believed to require a critical number of wavelets and a critical wavelength [17-19]. The wavelength is the product of the atrial refractory period and the conduction velocity [17-19]. Shorter wavelengths are thought to promote the induction and maintenance of atrial fibrillation [18, 19]. Therefore, factors that shorten refractory periods or slow conduction velocity favor the induction and maintenance of atrial fibrillation. Adenosine shortens the duration of atrial action potential and thus is believed to directly shorten atrial refractory periods [20, 21]. In our study, types III and IV atrial fibrillation were noted in all but one patient. Atrial cycle lengths are usually shorter with types III and IV atrial fibrillation and are correlated with shorter atrial refractory periods [22]. Adenosine also results in a reflex increase in circulating catecholamine levels and sympathetic nerve traffic [23, 24]. Catecholamines improve conduction velocity and shorten refractory periods [25-27] and therefore have competing effects on the wavelength. Catecholamines may shorten the wavelength if the shortening of the refractory period is greater than the improvement in conduction velocity. Together, these direct and indirect effects of adenosine administration result in the potential for shorter wavelengths, which in turn promote the induction of atrial fibrillation [18, 19]. These same concepts theoretically apply to adenosine-induced atrial flutter, except that this arrhythmia involves only one circuit [28].

A second potential mechanism is that adenosine administration results in frequent atrial premature complexes and therefore a long-short atrial sequence that induces atrial fibrillation. In our study, long-short atrial sequences were strongly associated with the development of atrial fibrillation, and the responsible mechanism could be similar to the mechanism that causes the development of ventricular fibrillation when a ventricular premature complex falls on the vulnerable portion of the T wave. Atrial premature complexes were usually seen after PSVT termination, and the occurrence of atrial premature complexes was significantly associated with the development of atrial fibrillation. Although we did not compare the frequency of atrial premature complexes after overdrive pacing to terminate PSVT with the frequency of these complexes after adenosine administration, atrial premature complexes in the latter situation occurred in approximately two thirds of patients. If adenosine causes atrial premature complexes, the mechanism would be difficult to hypothesize. In fact, one would not expect adenosine to cause triggered atrial premature complexes because adenosine activates an outward potassium current and hyperpolarizes atrial tissue [29-31]. Adenosine can also decrease intracellular calcium by blocking catecholamine-induced calcium uptake [32]. These two effects are thought to explain adenosine's ability to terminate triggered and catecholamine-dependent ventricular arrhythmias, and one might expect that the same mechanisms prevent atrial premature complexes [32]. One possible explanation is that an increase in circulating catecholamine levels and sympathetic nerve traffic [23, 24] promotes atrial premature complexes, which in turn create long-short atrial sequences and result in atrial fibrillation.

Previous Studies

In previous studies, adenosine terminated PSVT in approximately 90% of patients [33]. For this reason, the authors of the American Heart Association's guideline on advanced cardiac life support recommended adenosine for the short-term management of PSVT [6]. The higher frequency of successful termination of PSVT noted in our study probably relates to the central administration of adenosine. Because adenosine is safe and effective therapy for PSVT, many investigators recommend it for this purpose with few caveats [1-6]. Adenosine administration has also been suggested as a way to identify latent anterograde accessory pathway conduction either before or after an ablation procedure [34, 35]. Adenosine, however, is believed to shorten anterograde accessory pathway refractory periods [36], perhaps through reflex sympathetic stimulation [22, 23, 36]. In addition, anecdotal evidence suggests that adenosine is associated with the development of atrial fibrillation with a rapid ventricular response due to conduction over an accessory pathway [7-12]. Our study is the first to determine that the frequency of atrial fibrillation and atrial flutter after termination of PSVT with 12 mg of adenosine is 12% and that atrial fibrillation occurs with a similar frequency regardless of the mechanism of PSVT. A previous study [33] found an incidence of atrial fibrillation after PSVT termination with adenosine of only 1%. In that study, however, 80% of 163 patients treated with adenosine received less than 12 mg of the drug. The results of a recent small study [13] suggested that atrial fibrillation is provoked only by large doses of adenosine (≥12 mg). This finding indicates that the frequency of atrial fibrillation precipitated by adenosine may be dose related and may occur more often with larger doses of adenosine, such as those used in our study.

Previous studies have shown that patients with the Wolff-Parkinson-White syndrome who experience an episode of cardiac arrest have a minimum preexcited ventricular response during atrial fibrillation of more than 240 beats/min [37]. No patient from our study who developed atrial fibrillation had a minimum preexcited R-R interval in this range. However, this finding is probably due to the small number of patients with preexcitation who developed atrial fibrillation. Our results suggest that approximately 10% of patients with the Wolff-Parkinson-White syndrome and PSVT who are treated with adenosine may develop atrial fibrillation. According to the data of Klein and coworkers [37], one would expect that half of these patients would develop a rapid preexcited ventricular response during atrial fibrillation.

Finally, an anecdotal report suggests that adenosine may provoke ventricular arrhythmias in addition to atrial arrhythmias [38]. However, we did not observe ventricular proarrhythmia with the administration of adenosine in any patient.

Limitations

Our study has several potential limitations. First, the manufacturers of adenosine recommend that this agent be initially administered as a 6-mg bolus and that 12 mg be administered only if the initial bolus does not terminate PSVT [39]. We administered one 12-mg bolus through the femoral vein. Our findings, therefore, may not be applicable to patients receiving other doses or to patients who receive this agent through a peripheral vein. Second, our study was done while a catheter was placed in the right atrium. Intracardiac electrode catheters may precipitate atrial premature complexes and atrial fibrillation; thus, the incidence of adenosine-induced atrial fibrillation may be lower when intracardiac catheters are not present. Third, the development of atrial fibrillation may have been only coincidentally related to adenosine administration. This seems unlikely, however, because atrial fibrillation developed only after the effects of adenosine were observed. Another technique to terminate PSVT, overdrive pacing, never resulted in atrial fibrillation, and otherwise these patients never developed atrial fibrillation during the ablation procedure.

Fourth, all of our patients were referred for an electrophysiology procedure and possible ablation for the treatment of PSVT. The clinical characteristics of these patients with PSVT is typical of patients referred for an electrophysiologic procedure and possible ablation. Patients not referred for invasive evaluation of PSVT may represent a different clinical population. Finally, although our study did not address the reproducibility of adenosine-induced atrial fibrillation, the results of a previous small study suggest that adenosine-induced atrial fibrillation is reproducible [13].

Clinical Implications

In conclusion, the incidence of adenosine-induced atrial fibrillation after the central administration of 12 mg of intravenous adenosine to terminate PSVT is 12%. Because our study was performed during an electrophysiologic procedure while a right atrial catheter was in place, the results may not be applicable to patients with PSVT treated with adenosine in other clinical settings. However, if the mechanism of PSVT is not known and the Wolff-Parkinson-White syndrome is possible, administration of adenosine should be limited to medical facilities with emergency resuscitation equipment because of the potential for a rapid, preexcited ventricular response during atrial fibrillation.