The ventricular rate and percent of pre-excited QRS complexes during atrial fibrillation were compared in two groups of patients with the Wolff-Parkinson-White syndrome. Group A consisted of 22 patients whose anterograde effective refractory period of the accessory pathway was longer than that of the atrioventricular (AV) node. Group B consisted of 23 patients in whom this relation was reversed. No patient had organic heart disease. Both groups had a similar effective refractory period of the accessory pathway (288 +/- 37 vs. 280 +/- 26 ms), whereas that of the AV node was shorter in group A than group B (242 +/- 25 vs. 285 +/- 27 ms, p = 0.0001). Patients in group A had a lower percent of pre-excited QRS complexes during atrial fibrillation (39 +/- 43% vs. 93 +/- 20%, p = 0.0001). In the 21 patients whose refractory period was measured, the difference was plotted against the percent of pre-excited QRS complexes; there was a significant correlation between the two (r = -0.83, p < 0.001). In patients in whom pre-excited RR intervals were present, the pre-excited RR intervals were compared between the two groups. Both groups had similar effective refractory periods of the accessory pathway (265 +/- 22 vs. 280 +/- 27 ms) and ventricle (200 +/- 17 vs. 211 +/- 26 ms). The effective refractory period of the AV node was shorter in group A (248 +/- 22 vs. 285 +/- 28 ms, p = 0.0005). The shortest pre-excited RR interval did not show any difference (244 +/- 37 vs. 265 +/- 41 ms). However, both the average (328 +/- 39 vs. 397 +/- 56 ms, p = 0.001) and longest (495 +/- 109 vs. 666 +/- 205 ms, p = 0.02) pre-excited RR intervals were shorter in group A. These data suggest that interaction between the refractory periods of the AV node and accessory pathway contributes to the percent of pre-excited QRS complexes. The effective refractory period of the AV node also indirectly contributes to the duration of pre-excited RR intervals. This contribution is greatest when RR intervals are long.