The influence of the clinical presentation on the long-term outcome in 213 consecutive patients with ICDs, ECG storage capability, and nonthoracotomy leads, was analyzed. Sixty-six patients presented with cardiac arrest (CA), 81 patients with hemodynamically stable VT, and 66 patients with syncope (SY). Patient characteristics were: mean age CA 62, VT 61, SY 61 years; mean ejection fraction CA 31%, VT 29%, SY 30%; coronary artery disease CA 71%, VT 71%, SY 64% (all P > 0.05 Fisher's exact test); female gender CA 40%, VT 14%, SY 19% (CA vs VT and SY, P < 0.005); inducibility by programmed stimulation CA 50%, VT 84%, SY 61% (VT vs CA and SY, P < 0.001, CA vs SY, P > 0.05). During a mean follow-up of 14.5 months, 29 patients died: CA 12%, VT 14%, SY 9% (P > 0.05). Comparing Kaplan-Meier curves, no difference in the time course of overall mortality was found (log-rank P > 0.05). In the CA, VT, and SY groups, 543, 1,630, and 189 ICD therapies (including antitachycardia pacing, low energy cardioversion, and defibrillation) were observed, respectively. Actuarial analysis showed a shorter interval between implantation and first ICD therapy for VT versus CA and SY (log-rank P < 0.005). Patients presenting with VT experienced earlier and more frequent ICD therapies than patients with CA or SY independent of age, ejection fraction, and heart disease. No difference in overall mortality and time course of fatal events was observed among the three groups.