Objective: To determine the accuracy of coagulation profile laboratory tests, thromboelastography, and Sonoclot (SCT) values for predicting microvascular bleeding after cardiopulmonary bypass (CPB). Design: A prospective, blinded trial. Setting: A large academic medical center. Participants: Eighty-two adult patients undergoing elective cardiac surgery. Interventions: Ten minutes after CPB, thromboelastography, SCT, and coagulation profile tests (bleeding time, prothrombin time, activated partial thromboplastin time, fibrinogen, fibrin split products, platelet count, mean platelet volume, and platelet hematocrit) were determined from a whole blood sample taken from an existing arterial catheter. Patients were subjectively defined as ''bleeders'' or ''non-bleeders'' by blinded clinical observers. Preoperative baseline tests were also obtained. Measurements and Main Results: Thirty of the 82 patients (36.6%) were characterized as bleeders. Coagulation profile tests had the best correlation with intraoperative and postoperative blood loss, The specificity, sensitivity, and negative and positive predictive values were determined by receiver operating characteristic analysis, and the test values that differentiated normal from abnormal (bleeding) patients were determined. The coagulation profile laboratory tests had the greatest maximal sensitivity and specificity for predicting bleeding. These predictive values were outside the normal range for these laboratory tests. The thromboelastography values that produced maximal sensitivity and specificity were in the normal range for that test. Conclusion: Contrary to previous studies, coagulation profile tests had the greatest sensitivity and specificity to differentiate patients with excessive bleeding (abnormal) from those without excessive bleeding (normal) after CPB. Therefore, these tests should be used to guide transfusion therapy in patients who have excessive bleeding after CPB. Copyright (C) 1997 by W.B. Saunders Company.
