To assess the clinical usefulness of continuous on-line vectorcardiography (VCG), we studied 61 patients admitted to the coronary care unit (CCU) with chest pain, supposedly ischemic. Continuous VCG was performed for 24 h, monitoring QRS vector difference (QRS-VD), ST-change vector magnitude (STC-VM) and ST vector magnitude (ST-VM) measured 20 and 60 ms after the termination of the QRS complex. The patients were divided into four groups based on the final diagnosis; group A, 15 patients with normal exercise tests and extracardiac causes of chest pain; group B, 15 patients with unstable angina; group C, 15 patients with non-Q-wave myocardial infarction (MI); group D, 16 patients with Q-wave MI. Treatment was given according to a normal routine. Of 31 patients with MI, 16 received treatment with streptokinase. Groups A and B showed no significant permanent changes in QRS-VD, STC-VM or ST-VM. However, group B showed a higher occurrence of transient episodes (duration: 2 min-6 h) of a significant change of QRS-VD by > 15 µVs and of STC-VM, ST-VM 20 and ST-VM 60 by > 0.1 mV. Groups C and D showed both permanent changes and transient episodes for the studied vector parameters. Transient episodes were significantly fewer in group D than in group B. In patients with MI, the permanent change of vector parameters evolved more rapidly and reached a plateau earlier in those treated with streptokinase (QRS-VD: 178 ± 82 vs. 293 ± 100 min, p < 0.001; ST-VM 20: 142 ± 75 vs. 293 ± 89 min, p < 0.005). The magnitude of the end value for QRS-VD correlated with infarct size estimated by the maximal value of creatine kinase (r = 0.89; p < 0.001). We conclude that in patients admitted to the CCU with chest pain, continuous VCG monitoring early differentiates patients suffering from ischemic heart disease (IHD) from patients without IHD. It also differentiates patients with unstable angina from patients with MI.