Cardioversion of atrial fibrillation (AF) using traditional monophasic shock waveform is unsuccessful in up to 20% of cases, and often requires several shocks of up to 360 J. Based on the success with biphasic shock waveform in converting ventricular fibrillation, it was postulated that biphasic shocks would allow cardioversion with lower energy. In a international multicenter, double-blind, randomized trial of 203 patients, damped sine wave monophasic shocks were compared with impedance-compensated truncated exponential biphasic waveform shocks. Patients received up to five shocks: 100 J, 150 J, 200 J, a fourth shock at maximum output for the initial waveform (200 J biphasic, 360 J monophasic) and a final cross-over shock at maximum output of the alternate waveform. For each energy level, the biphasic waveform compared favorably to the monophasic waveform in successful cardioversion (100 J: 60% versus 22%, P < 0.0001; 150 J: 77% versus 44%, p < 0.0001; 200 J: 90% versus 53%, p < 0.0001). Success with 200 J biphasic was equivalent to 360 J monophasic shock (91% versus 85%, p = 0.29). Patients randomized to biphasic waveform required fewer shocks and lower total energy delivered; in addition, this waveform was associated with less dermal injury and no blistering. Biphasic shocks converted AF present for less than 48 hours with 80% efficacy, but conversion of AF present for more than 48 hours and more than 1 year the success rate was only 63 and 20%, respectively. The results of this study is similar to other investigations comparing biphasic and monophasic shock waveforms for conversion of atrial fibrillation. We recommend starting with biphasic energy of 100 J for atrial fibrillation of less than 48 hours duration, but using higher energies (150 J, 200 J or greater) when AF has been present for longer periods.