To determine the predictive value of adenosine thallium-201 myocardial imaging for perioperative cardiac events, 60 consecutive patients referred for preoperative cardiac evaluation were studied before vascular (n = 25), orthopedic (n = 14), or general (n = 21) surgery. Tomographic (n = 52) and planar (n = 8) thallium-201 imaging was performed after adenosine infusion at a rate of 140 micrograms/kg/min for 6 minutes. Two blinded expert observers graded results of adenosine thallium-201 studies as normal (33%), fixed defect only (2%), reversible defect only (48%), and combined (fixed and reversible) defects (17%). After 6 +/- 3 months of follow-up, 81% proceeded to surgery and 43% underwent preoperative coronary angiography. Clinical variables that correlated with perioperative cardiac events were a history of diabetes mellitus (p = 0.05), left bundle branch block (p = 0.02), and left ventricular hypertrophy (p = 0.06) on the resting ECG. This clinically "high-risk" group had an event rate of 22% as compared with no cardiac events in patients in the "low-risk" group without these clinical characteristics (p = 0.005). Stepwise logistic regression analysis revealed that the presence of a combined (fixed and reversible) adenosine thallium-201 defect (p = 0.0007), three-vessel coronary artery disease (p = 0.001), and left bundle branch block (p = 0.02) was predictive of subsequent cardiac events with relative risk ratios of 4.9, 2.9, and 2.2, respectively. Therefore the presence of an adenosine thallium-201 perfusion defect is correlated with and predictive of an increased risk of perioperative cardiac events in patients referred for preoperative risk evaluation.