Evaluating risk of esophageal variceal bleed in children with cirrhosis and waitlisted for liver transplantation

Background: Children with cirrhosis and waitlisted for liver transplant are prone to bleeding from ruptured esophageal varices; for which presence of grade 2, 3 varices and red signs are known risk factors. The involvement of hemostasis remains controversial at the moment because of the rebalanced state of coagulation during cirrhosis. Patients and methods: Prospective observational study was designed, including children with portal hypertension and decompensated cirrhosis. Portal hypertension assessment was done by ultrasound, endoscopy and hemostasis evaluation with dynamic parameters of thromboelastometry (ROTEM®), platelet function with Multiplate®, in addition to conventional tests. The clinical end point was occurrence of upper gastrointestinal bleeding. Above mentioned parameters were compared between children with and without bleeding using univariate statistical methods. Additionally an earlier developed predictive model for risk of variceal bleeding was applied and validated on the prospective cohort, which comprises of presence of grade 2-3 varices and / or red spots on the upper endoscopy and fibrinogen <150 mg / dl;. Results: Twenty children were enrolled (18 - biliary atresia, median age 9 months (4-129)), six were in the bleeding group as they had upper gastrointestinal bleed during the pre-transplant period. Waiting time before transplantation, presence and severity of oesophageal varices, levels of factor V (49% vs 70), INR (1.8 vs. 1.5) or platelets count (104 10exp3 /µl vs 219) were not statistically different between the two groups. Significant differences were observed in fibrinogen levels (109 mg / dl vs 257 mg / dl , p <0.05), ADP dependent platelet aggregation (103 AU / min versus 368, p <0.05 ), the thrombin dependent platelet aggregation (265 AU / min vs 558 , p <0.05) and clotting time (64 sec vs 52, p <0.05) of EXTEM analysis. The bleeding risk model was tested in this prospective cohort and predictive performance of bleeding risk (accuracy) was 85.18% (sensibility 90.5%, specificity 66.7%, NPV 90.5%, PPV 66.7%). Conclusion: We demonstrate involvement of hemostasis in bleeding risk of esophageal varices. A low fibrinogen levels seems to be risk factor of bleeding tendency in patients with decompensated cirrhosis, which suggests the potential benefit of prophylactic treatment with fibrinogen in high risk cases. Multiplate® analysis should also help us in the future to determine the risk of bleeding esophageal varices in children with decompensated cirrhosis and would be further integrated into our predictive model.