Abstract
We recently demonstrated that patients with intraoperative coagulopathy and increased blood loss show a persistent pre-, intra- and postoperative increase in fibrin monomer concentration as well as an excessive intraoperative consumption of fibrinogen and F. XIII. We therefore wanted to test the hypothesis that preoperative fibrin monomer concentrations can be used as a risk indicator for intraoperative blood loss.
In 168 patients admitted to the surgical service of our hospital, median intraoperative blood loss increased significantly with preoperative fibrin monomer (FM) quartiles (50, 100, 200 and 400 ml in preoperative FM quartiles 1 to 4; p<0.001, ANOVA on ranks and p<0.05 for group wise comparison, Rank Sum test). Stratification in FM quartile groups was unrelated to diagnoses with the exception of laparoscopic cholecystectomy (found significantly less frequent in quartile 4 than 1). Most importantly, accuracy evaluation showed that preoperative fibrin monomer concentration < 3 μg/l excluded intraoperative blood loss > 500 ml with 92% sensitivity, 95% negative predictive value and 41% exclusion rate. This compares well to other exclusion strategies such as the exclusion of deep venous thrombosis with the help of D-dimer and probability scores. In contrast, blood loss was unrelated to preoperative values of prothrombin time, platelet count and actived partial thromboplastin time. The FM ROC curve compared to the PT ROC curve is shown in the figure; the area under the curve for the FM ROC curve was significantly greater than the one for the PT ROC curve (0.743, 95% CI 0.655 – 0.811 vs. 0.555, 95% CI 0.473 – 0.635).
These results - obtained in a second, independent and prospective study - confirm the hypothesis generated from our first study that preoperative fibrin monomer concentrations allow prediction on excessive intraoperative blood loss. We suggest that preoperative fibrin monomer concentrations be further studied for identification of patients that might benefit from intensified intraoperative monitoring.
Figure
