Abstract
Excessive perioperative bleeding is associated with increased morbidity and mortality as well as increased economic costs. A range of whole blood laboratory tests for hemostatic monitoring has emerged, but their ability to predict perioperative bleeding is still debated. We conducted a systematic review of the existing literature assessing the ability of whole blood coagulation (thromboelastography [TEG]/thromboelastometry [ROTEM]/Sonoclot), platelet function tests, and standard plasma-based coagulation tests to predict bleeding in the perioperative setting. We searched PubMed and Embase, covering the period from 1966 to November 2016. In total, 99 original studies were included. The included studies assessed TEG/ROTEM/Sonoclot (n = 29), platelet function tests (n = 27), both test types (n = 8), and standard coagulation tests only (n = 18), and some (n = 17) investigated the predictive value of testing in patients receiving antithrombotic medication. In general, studies reported low positive predictive values for perioperative testing, whereas negative predictive values were high. The studies yielded moderate areas under receiver operator characteristics (ROC) curve (for the majority, 0.60–0.80). In conclusion, while useful in the diagnosis and management of patients with overt bleeding, whole blood coagulation and platelet function tests as well as standard coagulation tests demonstrated limited ability to predict perioperative bleeding in unselected patients. Therefore, we recommend that both whole blood and plasma-based coagulation tests are primarily used in case of bleeding and not for screening in unselected patients prior to surgery.
Keywords
thromboelastography - platelet function analysis - blood coagulation tests - perioperative bleeding - surgical blood loss - predictive value
