To assess the potential effects of coronary artery bypass surgery on left ventricular diastolic filling, 12 patients, aged 65 +/- 11 years, were studied by serial transesophageal Doppler echocardiograms. Doppler measures of mitral inflow velocity were made before, immediately after, 4 h after, and 20 h after cardiopulmonary bypass (CPB). Left atrial pressure was directly measured and controlled at 10 +/- 2 mm Hg for each study period. Mitral maximal early inflow velocity (E)/maximal atrial velocity (A) ratios and atrial filling fractions were calculated as indexes of diastolic function from maximal E and A velocities and their time velocity integrals, respectively. Data sets were available for serial comparison in 11 patients and were also compared with an age-matched control group of normal values. The results of E/A ratios were as follows: control group--1.4 +/- 0.2; before CPB--1.7 +/- 0.6; immediately after CPB--1.0 +/- 0.2 (p < 0.05 vs control group, before CPB, and 20 h after CPB values); 4 h after CPB--0.8 +/- 0.2 (p < 0.05 vs control group, before CPB, and 20 h after CPB values); and 20 h after CPB--1.3 +/- 0.4. Atrial filling fractions were as follows: control group--0.29 +/- 0.05; before CPB--0.25 +/- 0.06; immediately after CPB--0.43 +/- 0.07 (p < 0.05 vs control group, before CPB, and 20 h after CPB values); 4 h after CPB, 0.46 +/- 0.07 (p < 0.05 vs control group, before CPB, and 20 h after CPB values); and 20 h after CPB--0.35 +/- 0.06. Alterations in Doppler indexes of left ventricular filling occurred immediately after CPB and persisted 4 h after CPB. These indexes returned to baseline values by 20 h after CPB. This suggests reversible diastolic dysfunction in patients after coronary artery bypass surgery.