Typical atrial flutter ablation has become anatomically guided to 2 separate sites within the isthmus at the inferior right atrium: (1) between the inferior vena cava and the tricuspid annulus (anterior side of the isthmus [A]), (2) between the eustachian crest, the coronary sinus ostium and tricuspid annulus (posterior side of the isthmus [P]). We prospectively compared ablation results at these sites in 72 consecutive patients. Patients were randomized in group P or A according to the initial target site. If ablation failed at 1 site after 15 radiofrequency (RF) pulses, the other side of the isthmus was targeted. Before 15 RF pulses, complete bidirectional isthmus block was achieved in 30 of 36 group A patients and in 25 of 36 group P patients, with similar mean RF pulses number, procedure time, and fluoroscopy time. After shifting to the other target, success was finally obtained at P in 2 of 6 group A patients, and at A in 8 of 11 group P patients before a maximum of 30 RF pulses. Among successful patients, number of RF pulses, procedure time, and fluoroscopy time were significantly lower in group A (7.2 +/- 5.4 vs 11.0 +/- 8.1 pulses, p = 0.03; 131 +/- 44 vs 163 +/- 66 minutes, p = 0.03; 31 +/- 19 vs 46 +/- 24 minutes, p = 0.01, respectively). Impairment of atrioventricular (AV) nodal conduction occurred in 5 patients only during ablation at P. AV block was transient in 4 patients and permanent in 1. Although atrial flutter ablation is equally effective at P and A, success seems easier to obtain when A is first targeted. Ablation at P is associated with a significant risk of AV block.