We assessed the feasibility of low energy endocardial defibrillation i
n patients with atrial fibrillation or atrial flutter who had failed a
trial of pharmacological reversion with amiodarone. Low energy endoca
rdial defibrillation under general anesthesia was attempted in 9 patie
nts, 5 with atrial flutter and 4 with atrial fibrillation (median dura
tion of arrhythmia 3.75 months). Two large surface area endocardial le
ads were introduced percutaneously and sited in the right atrial appen
dage and at the right ventricular apex. A cutaneous patch electrode wa
s placed on the left thorax. Biphasic shocks synchronized to the ventr
icular electrogram were used to terminate atrial arrhythmias. Three el
ectrode configurations were evaluated in the following sequence at eac
h energy level: atrial cathode to ventricular anode; ventricular catho
de to atrial anode; atrial cathode to a combined ventricular and cutan
eous anode. If endocardial defibrillation failed (0.5-10 J), transthor
acic defibrillation using 200 joules followed by 360 joules, if requir
ed, was performed. Endocardial defibrillation was successful in all fi
ve patients with atrial flutter (0.5 J, 1.0 J, 1.0 J, 4.0 J, and 10.0
J) but in only one patient with atrial fibrillation (10 J). On no occa
sion did successful defibrillation occur with one configuration when i
t had failed with an alternate configuration at that particular energy
level. Ventricular fibrillation did not occur, and there were no othe
r significant complications. Low energy endocardial defibrillation is
feasible in patients with atrial flutter using large surface area elec
trodes. Although the success rate of atrial defibrillation was low, fu
rther work is required, particularly in patients with more recent onse
t of the arrhythmia and using a right to left electrode configuration.