We developed a new protocol for diagnosis and treatment of patients with sustained tachycardias (heart rate > 150 beats.min-1). The patients first underwent vagal manoeuvres; if those remained unsuccessful, i.v. adenosine in increasing doses of 6, 12, and 18 mg was administered until sinus rhythm (SR) or transient atrioventricular (AV) block, unmasking the underlying rhythm, was recorded. In the latter and in the non-responding cases other antiarrhythmics were applied. Ninety-three episodes of tachycardia in 46 patients were treated according to this protocol. Six episodes (6%) were terminated by carotid massage, 64 of the remaining 87 episodes (74%) responded to adenosine with return to SR. Conversion to SR occurred more often in episodes with narrow- than in wide-complex tachycardia (81 vs. 59%, P < 0.05). To achieve SR, the mean adenosine dose was lower in narrow- than in wide-complex tachycardia (13 +/- 8 vs 21 +/- 10 mg; P < 0.01). The duration of asystole after adenosine did not differ between these two groups, whereas the duration of arrhythmia after adenosine differed significantly (8.5 +/- 5.8 vs 18.6 +/- 22.9 s; P < 0.05). Side effects of adenosine such as flush, dyspnoea, and chest pain did not seem to be dose dependent and occurred in about 20%. According to our protocol, in more than 75% SR was achieved in patients with sustained tachycardias after vagal manoeuvres and adenosine.