To evaluate whether the extent of left ventricular (LV) asynchrony plays a role in the impairment of LV rapid filling in patients with coronary artery disease (CAD), 48 patients underwent both radionuclide angiography and cardiac catheterization. Patients were divided into group I (n = 33), with normal LV kinesis or only mild hypokinesia, and group II (n = 15), with LV dyskinesia or akinesia. Radionuclide ejection fraction was higher in group I than in group II (62 +/- 12 vs 44 +/- 20%; p less than 0.001). Peak filling rate was significantly lower in group II (1.9 +/- 0.8 vs 2.6 +/- 0.9 end-diastolic counts/s; p less than 0.01). Time to end-systole coefficient of variation, an index of the extent of LV asynchrony, was significantly higher in group II than in group I (43 +/- 10 vs 35 +/- 6; p less than 0.0002). In group I, a highly significant inverse relation was found between this index of asynchrony and peak filling rate (r = 0.71; p less than 0.0001). This correlation was found even when time to end-systole coefficient of variation was normalized to the RR interval (r = 0.49; p less than 0.01) and when peak filling rate was expressed in stroke counts (r = 0.57; p less than 0.001). The correlation between peak filling rate and index of asynchrony was maintained up to an end-systole coefficient of variation value of approximately 35. In group II patients (most with an asynchrony value greater than or equal to 35) no relation was found between time to end-systole coefficient of variation and peak filling rate.