Adenosine-Induced Atrial Arrhythmia


Synopsis: Adenosine should only be used in situations where the cardiac rhythm can be continuously monitored, and any deterioration of the rhythm after termination of the re-entrant tachycardia can be diagnosed and treated properly.

Source: Strickberger SA, et al. Ann Intern Med 1997; 127:417-422.

Strickberger and associates administered adenosine to 200 consecutive patients who were undergoing electrophysiologic study for paroxysmal supraventricular tachycardia (PSVT). The patients had either atrioventricular nodal re-entrant tachycardia (124 patients); atrioventricular re-entrant tachycardia with a manifest (40 patients); or concealed (28 patients) accessory pathway or atrial tachycardia (8 patients). During the electrophysiologic study, the patient’s tachycardia was induced. Once the patient was in stable tachycardia, 12 mg of intravenous adenosine followed by a 20 mL normal saline flush was infused through a sheath positioned in a femoral vein. The investigators note whether the tachycardia was terminated and whether atrial fibrillation or atrial flutter developed after termination of the tachycardia. If atrial fibrillation developed, the type of electrogram noted during fibrillation was classified as either class I, II, III, or IV. The occurrence of atrial or ventricular premature beats was also recorded.

Adenosine terminated the tachycardia in 198 out of 200 patients (99%). In 22 patients, atrial fibrillation occurred, and, in two patients, atrial flutter developed a mean of 6.3 ± 6.2 seconds after PSVT termination. In 21 patients, at least one sinus beat was noted before the onset of the atrial arrhythmia while in the remaining three patients, an atrial ectopic complex was the preceding beat. Twenty-three of the 24 patients had either type III or type IV atrial fibrillation characterized by disorganized atrial electrograms that had neither a discrete morphology nor an isoelectric baseline. The mean duration of the atrial fibrillation was 5.6 ± 6.7 minutes. The range was from 8 seconds to 20.7 minutes. The median duration was not specified. Sixteen of the 24 patients spontaneously converted to sinus rhythm, whereas eight patients underwent cardioversion.

Six of the 40 (15%) patients with accessory pathways capable of anterograde conduction developed atrial fibrillation. Four of these developed pre-excited atrial arrhythmias. The mean pre-excited R-R rate during atrial fibrillation was 150 ± 36 bpm, and the mean maximum pre-excited R-R rate was 176 ± 51 bpm. One patient also developed pre-excited atrial flutter with a mean pre-excited R-R rate of 188 bpm. No patient developed syncope or ventricular fibrillation. Strickberger et al could not identify any predictors for the occurrence of atrial fibrillation in response to adenosine and conclude that adenosine should not be given unless emergency resuscitation equipment and personnel are available.


The electrophysiologic effects of adenosine on supraventricular tissues have been well characterized. Adenosine produces conduction block in the AV node, and this is the mechanism by which it terminates most forms of PSVT. The conduction block in the AV node usually occurs at a concentration lower than that which produces major changes in atrial electrophysiology. However, adenosine does have dramatic effects on the atrial action potential. It shortens action potential duration, shortens the atrial effective refractory period, and slightly hyperpolarizes the membrane. It has been shown in several prior studies that atrial premature beats with a short coupling interval during the period of action potential shortening will result in atrial fibrillation in virtually all patients (O’Nunain S. Cardiovasc Res 1992;26:939-943; Kabell G. Am J Cardiol 1996;78:1443-1446).

The frequency with which atrial fibrillation will develop during routine clinical treatment of PSVT appears to be greatly overestimated in this paper. The effects of adenosine are extremely short-lived. Once injected into the circulation, adenosine is cleared from the circulation by cellular uptake and metabolism with a half-life that is measured in only seconds. The dosage guidelines for adenosine are for injection into a peripheral vein. With this type of infusion, the onset of drug action is usually 20-30 seconds after injection. The recommended dosages are an initial infusion of 6 mg followed, if necessary, by a second infusion of 12 mg. The two-dose sequence is specified specifically to avoid producing adverse effects from high adenosine concentrations while still allowing for interindividual variability in circulation time, clearance, or sensitivity. If one injects adenosine through a central vein or through a large femoral vein sheath, the onset of action will be within 5-8 seconds after injection and the effect of the recommended doses will be markedly greater (McIntosh-Yellin NL. J Am Coll Cardiol 1993;22:741-745). The infusion technique used by the investigators used the recommended second dose of adenosine without testing a lower dose first, and it used a large sheath in a femoral vein as the injection site. In the multi-center trial that led to the recommended dosage guidelines, only three patients who received adenosine developed atrial fibrillation, and all of those episodes were of short duration (DiMarco JP. Ann Intern Med 1990;13:104-110). It should also be noted that those three patients and all of the patients in this study developed atrial fibrillation when they received adenosine in the setting of an electrophysiologic study. During the electrophysiologic study, many patients are receiving isoproterenol and also have multiple catheters in the heart. The sudden change in heart rate associated with tachycardia termination when catheters are in the heart may lead to an increased frequency of atrial premature beats that can precipitate atrial fibrillation. Atrial and ventricular ectopy may also be noted in patients without intracardiac instrumentation after adenosine, but the frequency is decreased.

Despite these reservations, the final conclusion of Strickberger et al still seems justifiable. Although the frequency of inducing atrial fibrillation or atrial flutter during routine clinical use with adenosine should be much lower than they report here, physicians should recognize that this can happen. Therefore, I fully agree with Strickberger et al that adenosine should only be used in situations where the cardiac rhythm can be continuously monitored, and any deterioration of the rhythm after termination of the re-entrant tachycardia can be diagnosed and treated properly.