Atrial natriuretic peptide and brain natriuretic peptide are known to be indices for heart failure. Atrial natriuretic peptide and brain natriuretic peptide changes in off-pump and on-pump coronary artery bypass grafting is hypothesized to be correlated to clinical implications of coronary artery bypass grafting. 20 consecutive off-pump and 20 consecutive on-pump coronary artery bypass grafting patients were studied. Perioperative atrial natriuretic peptide and brain natriuretic peptide values were measured and statistically analyzed in terms of 14 factors related to myocardial damage and recovery. Postoperative atrial natriuretic peptide plateaued on the third postoperative day and it decreased gradually down to the preoperative level by one month in the off-pump group. Postoperative brain natriuretic peptide plateaued, showed very slow decrease and it never reached down to the preoperative level. The peak brain natriuretic peptide level was correlated with aortic cross-clamp time, postoperative pleural effusion, and postoperative atrial fibrillation (p < 0.01). The atrial natriuretic peptide change reflected surgical prevention of ventricular remodeling. Brain natriuretic peptide > 450 microg.mL(-1) had strong predictive power for atrial fibrillation and pleural effusion and is a useful marker for management of coronary surgery patients.