One hundred consecutive men with a normal ECG at rest had a maximal treadmill test using 14 leads during and postexercise. Coronary arteriography performed the following day revealed coronary stenoses k70% in 66 patients. Test results obtained from a V6 lead were compared to different lead combinations and were correlated with arteriographic findings. A positive exercise test occurred in 37 men using an isolated V, lead compared to 50 men (P < 0.05) using 11 leads, 52 men (P < 0.05) using a combined CM6, CC5, Cl (inferior) lead system and 58 (P < 0.001) men using all 14 leads. The predictive value of a positive test varied between 89-95% and was not changed significantly by the addition of multiple leads. The 14 lead ECG was positive in 43/45 (96%) patients with multivessel disease.Parameters which helped to predict multivessel disease using 14 leads were 1) the time that ischemia first appeared 2) the pressurerate product at the time ischemia first appeared, and 3) the maximum workload that could be attained. In general, the magnitude of STsegment depression and the time required for a positive ECG to return to normal postexercise were not useful predictors of multivessel disease.We conclude that the use of multiple leads improves the sensitivity and efficiency of the maximal treadmill exercise test. The usefulness of exercise test results can be further improved if multiple leads are combined with physiologic data collected during exercise.