Electrophysiological testing and left ventricular endocardial mapping in sinus rhythm were performed in 61 patients with coronary artery disease who presented with cardiac arrest in an attempt to relate the results of these studies to clinical outcome. Forty-one patients (67%) had inducible sustained arrhythmias (18 uniform ventricular tachycardia, 23 polymorphic ventricular tachycardia/ventricular fibrillation) and 20 had no inducible arrhythmia. Patients with inducible arrhythmia had 45% abnormal and 6% fractionated electrograms versus 31% and 0% for those without inducible arrhythmia (P greater than 0.05 for both comparisons). Sixteen of 59 patients (27%) with adequate follow-up had arrhythmia recurrence (11/39 [31%] with inducible arrhythmia and 5/20 [25%] without inducible arrhythmia) over a mean follow-up period of 27 months. Of five patients without inducible arrhythmia who experienced recurrence, two did so despite the anti-ischemic therapy. In the 20 patients without inducible arrhythmia, the 15 who remained arrhythmia-free had a mean of 78 +/- 22% normal sites versus 46 +/- 24% normal sites in those with recurrence (P greater than 0.05). We conclude that in patients with coronary artery disease and cardiac arrest: 1) patients without inducible arrhythmia have less marked endocardial electrical abnormality than those with inducible arrhythmia, 2) those patients who have marked endocardial abnormality despite the lack of inducible arrhythmia are at risk for clinical recurrence which suggests that these abnormalities may represent an anatomic substrate for arrhythmia which cannot be identified by programmed stimulation. These patients are candidates for AICD implantation and 3) patients with relatively normal endocardial electrograms do well with anti-ischemic therapy alone.