We undertook a study to determine (1) the frequency and prognostic significance of preexisting MRI brain abnormalities in patients undergoing coronary artery bypass grafts (CABG) and (2) whether MRI can detect surgery-related brain damage in 31 neurologically asymptomatic CABG patients (mean age, 61.0 +/- 6.6 years). MRIs were performed within 7 days before and 8 to 17 days after surgery. When we compared the preoperative images with those of 31 age- and risk factor-matched neurologically asymptomatic controls free of cardiac disease (mean age, 60.3 +/- 6.1 years), higher rates of thromboembolic infarcts (16% versus 0%), lacunes (58.1% versus 32.3%), and brainstem lesions (22.6% versus 3.8%) were noted. Subjective rating demonstrated significantly larger ventricles in patients than in controls (p = 0.002). CABG candidates also had significantly increased ventricular-to-intracranial cavity ratios (VICR) as determined by semiquantitative volumetric measurements (6.9 +/- 2.5% versus 4.9 +/- 1.6%; p = 0.0004). Eleven patients had postsurgical complications, with eight having symptoms consistent with diffuse encephalopathy. The only MRI finding that separated encephalopathic from complication-free patients was ventricular size (VICR 9.0 +/- 2.5% versus 4.9 +/- 1.6%; p = 0.006). This difference remained statistically significant after adjustment for the effects of age (p = 0.04). Postoperative MRI consistently failed to demonstrate surgery-related brain damage responsible for the encephalopathy.