Analyses of randomised clinical trials have suggested that only in selective populations may antiarrhythmic drugs be effective in improving prognosis: therapy of cardiac arrhythmias, in contrast to other cardiovascular pathological conditions, has not been fully successful. The ideal treatment of arrhythmias should be guided by a sound understanding of the relative arrhythmogenic mechanisms and vulnerable parameters of the different arrhythmias. New model agents are pure class III agents, developed to fulfil these ideal characteristics and are now under active investigation (dofetilide, ibutilide, azimilide, ambasilide, E 4031, almokalant, sematilide, RP 58866 and tedisamil).