OBJECTIVE Although transesophageal pacing has been used successfully for the interruption of cardiac arrhythmias, the efficacy of this technique for the interruption of spontaneous atrial flutter remains poorly defined. The utility of transesophageal pacing to interrupt atrial flutter that was persistent despite standard antiarrhythmic drug therapy (mean duration: 70.3 days; range: one day to more than 365 days) was studied prospectively in 39 consecutive patients. METHODS After written informed consent was obtained from each patient, transesophageal pacing was performed with a programmable stimulator, using the distal electrode as the cathode and the proximal electrode as the anode. All patients continued to receive a type 1 antiarrhythmic drug or amiodarone throughout the period of transesophageal pacing. The response to transesophageal pacing was classified as follows: (1) direct conversion; (2) indirect conversion; or (3) failure to interrupt atrial flutter. RESULTS The mean stimulus amplitude and pulse duration required for atrial capture were 19.8 +/- 7.5 mA and 18.4 +/- 7.9 msec. Atrial flutter was successfully converted to sinus rhythm by transesophageal pacing in 82% of patients. In 38% of patients, atrial flutter was converted directly to sinus rhythm without another intervening arrhythmia (direct conversion). The mean pacing rate required for direct conversion was 341 +/- 27 beats/minute. In 44% of patients, the cycle length of atrial flutter was accelerated to less than 180 msec or was converted to atrial fibrillation with spontaneous conversion to sinus rhythm within 24 hours (mean 8.4 +/- 9.3 hours, indirect conversion). The mean pacing rate inducing accelerated atrial flutter or transient atrial fibrillation was 372 +/- 61 beats/minute (p = NS compared to direct conversion). Atrial flutter was not interrupted or atrial fibrillation was induced that did not spontaneously convert to sinus rhythm within 24 hours in an additional seven patients (18%). The underlying cardiac disease, age, previous drug therapy, atrial size, atrial flutter cycle length, history of prior atrial fibrillation, left ventricular function, and concomitant medical illnesses did not predict the efficacy of transesophageal pacing. CONCLUSIONS The present study suggests that transesophageal pacing is highly effective for interrupting spontaneous atrial flutter that does not terminate with standard antiarrhythmic drug therapy.