To determine the clinical usefulness of the thromboelastogram in the prediction of postoperative hemorrhage in cardiac patients, we related the results of routine coagulation tests (RCTs) and thromboelastography with the amount of chest tube drainage postoperatively in 101 patients requiring cardiopulmonary bypass. Our data indicated that there was no correlation between RCT results and thromboelastographic variables. No single variable of RCTs and thromboelastography correlated well with the amount of chest tube drainage postoperatively. Before the onset of cardiopulmonary bypass, the most frequent abnormalities detected by thromboelastograms were fibrinolysis and hypocoagulability resulting from factor deficiency. Hypercoagulability detected by thromboelastograms occurred in 13% of patients after cardiopulmonary bypass and usually was not detected by RCTs. The incidence of false-negative thromboelastograms and RCT results in patients who had excessive hemorrhage of unknown cause was 46% and 52%, respectively. The incidence of fibrinolysis as detected by thromboelastograms was similar before and after bypass, but only 2 of the 18 patients with fibrinolysis had excessive hemorrhage postoperatively. Our results indicate that neither RCTs nor thromboelastography predicts the likelihood of excessive hemorrhage in patients after cardiopulmonary bypass. The thromboelastographic results should be interpreted cautiously because of the high rate of unreliable results.