After informed consent, all patients underwent an electrophysiologic study in the fasting state and after stopping all antiarrhythmic drugs except for amiodarone at least 48 h in advance. Two catheters were introduced through the right femoral vein into the right atrium. A 6-F quadripolar catheter with an interelectrode distance of 5 mm (Bard Electrophysiology, Tewksbury, Massachusetts) was placed for pacing at the low lateral right atrium with electrograms from the distal bipole coinciding with the beginning of the surface electrocardiogram (ECG) plateau in inferior leads during counterclockwise AF or with the nadir of the flutter wave during clockwise AF. A 7-F quadripolar deflectable catheter with a 4-mm tip electrode (Cordis-Webster, Baldwin Park, California or Medtronic, Minneapolis, Minnesota) was used for cavotricuspid isthmus mapping and ablation. Surface ECG (leads I, II, III, V1) filtered through a bandpass of 1 to 500 Hz, bipolar intracardiac electrograms filtered at 30 to 500 Hz and amplified at high gain (0.1 mV/cm) were simultaneously recorded with a polygraph (model Midas, PPG Biomedical Systems, Overland Park, Kansas) at paper speeds of 100 mm/s. A programmable stimulator (Cardiostimulateur Orthorhythmique, Savita, Paris, France) with a 2-ms output pulse width and four times threshold amplitude was used. Descending lateral, ascending septal right atrial activation and lateral to medial isthmus activation during counterclockwise AF and ascending lateral, descending septal and medial to lateral isthmus activation during clockwise AF were confirmed by sequential mapping. This meant that electrograms coinciding with the end of the surface ECG plateau in inferior leads (during counterclockwise AF) and coinciding with the peak of the positive flutter wave (during clockwise AF) were recorded at the coronary sinus ostium. Withdrawal mapping was then performed in the cavotricuspid isthmus from the margin of the TV to the IVC in the line of electrograms coinciding with the center of the surface ECG plateau in inferior leads during counterclockwise AF and coinciding with the initial downslope of the positive flutter wave during clockwise AF (Figure 1). This ensures a catheter tip position perpendicular to the advancing wave front, and moreover catheter displacement to either side can be nonfluoroscopically recognized by the altered timing of the site electrogram. Recognition of seemingly minor changes in position is facilitated by the naturally lower conduction velocities in this region during AF. Atrial electrograms on this line were categorized as DPs (if separated by an isoelectric interval of more than 30 ms), single potentials (SPs) and fractionated potentials (FPs). The extent of contiguous DPs on the line (uninterrupted by SPs or FPs) was fluoroscopically estimated in the right anterior oblique view as large (more than or equal to half the isthmus from the TV to the IVC edge), intermediate or small (less than one third of the isthmus width). Low amplitude DPs were recorded after ablation at sites of RF delivery (during AF and low lateral atrial pacing) and used to recognize the position of the incomplete ablation line (18- 19). The maximal amplitudes of DPs (at sites with both potentials of equal amplitude) mapped before ablation and on sites after RF delivery just before AF termination were compared.