Twenty-two studies enrolling a total of 3096 patients were identified from 1984 to date. Selection bias, blinding of the results, different cut-off limits, and several retrospective studies were some of the problems preventing a comprehensive analysis. The resting LVEF was not found to be a consistent predictor of perioperative ischaemic cardiac events. In the perioperative phase, poor LV function was, mainly, predictive of congestive heart failure, and, in the long-term, of cardiac outcome. The presence of myocardial wall motion abnormalities was also associated with both a higher chance of postoperative cardiac complications and a worse long-term cardiac outcome. Although measurements of LV function seem to play a key role in defining a patient's long-term prognosis, the value of routinely measuring LVEF preoperatively is limited and, therefore, MUGA scanning cannot be recommended as a general screening test. Despite this, it has been widely used for cardiac risk assessment in vascular surgery, and only recently its popularity has started declining. Other tests, such as stress-echocardiography and myocardial perfusion imaging, used selectively in moderate-risk patients can refine prediction of cardiac risk. In the future, gated stress myocardial perfusion scintigraphy, perhaps combined with ANP/BNP plasma level determination, may become a first choice test in preoperative cardiac risk assessment.