Atrial synchronous pacing with short, nonphysiological atrioventricular (AV) intervals has been reported to increase cardiac output in selected patients with severe dilated heart failure. The aim of this study was to determine the acute effect of atrial synchronous pacing with short AV intervals in a consecutive series of patients with dilated heart failure. Twelve patients with a mean ejection fraction of 21% +/- standard error 2.5% were studied. Pacing catheters were placed in the high right atrium and right ventricular apex and a balloon flotation catheter in the pulmonary artery for measurement of cardiac output. Simultaneous transthoracic echocardiography was performed for measurement of left ventricular filling time and mitral regurgitation. In a randomized crossover design, measurements were made during VDD pacing at programmed AV intervals of 100 and 60 msec and during a control period in sinus rhythm. Left ventricular filling time increased at AV intervals of 100 and 60 msec (mean difference 37 +/- 9 and 34 +/- 11 msec, respectively, both P < 0.01 compared to control). Despite increases in ventricular filling time, stroke, and cardiac index declined with short atrioventricular intervals (at an AV interval of 60 msec, stroke index fell by 2.1 +/- 0.5 mL/m(2), P < 0.05 and cardiac index by 125 +/- 45 mL/m(2); P = NS). Heart rate was unchanged at both AV intervals (78 +/- 4.9 at control, 78 +/- 5.2 at 100 msec and 79 +/- 4.9 beats/min at 60 msec; P = NS). The decrease in stroke index at an AV interval of so msec was inversely related to control left ventricular filling time (r = 0.74; P = 0.01) and ejection fraction (r = 0.69; P = 0.02) and directly related to heart rate (r = 0.77; P ( 0.02). The change in stroke index at an A V delay of 60 msec wets also inversely related to the change in mitral regurgitation induced by pacing (r = 0.84; P = 0.01). Thus, in a group of patients with stable dilated heart failure, atrial synchronous pacing with short AV intervals did not improve cardiac output. The change in cardiac output with pacing was inversely related to baseline left ventricular function and to the change in mitral regurgitation induced by pacing.