Esophageal variceal bleeding risk assessment in children with decompensated cirrhosis waitlisted for liver transplantation

Background

Cirrhotic children who are waitlisted for liver transplantation are prone to bleeding from ruptured esophageal varices, with Grade 2-3 varices and red signs as known risk factors. The involvement of hemostasis remains controversial due to the rebalanced coagulation during cirrhosis.

Material and methods

This was a prospective observational study involving children with portal hypertension and decompensated cirrhosis. Portal hypertension was assessed by ultrasound and endoscopy, whereas hemostasis was evaluated using dynamic parameters of thromboelastometry (ROTEM®) and platelet function analysis (Multiplate®), in addition to conventional laboratory tests. The clinical endpoint was the occurrence of upper gastrointestinal bleeding. The above-mentioned parameters were compared between children with and without bleeding using univariate statistical methods. Furthermore, an earlier developed predictive model for estimating the risk of variceal bleeding, comprising Grade 2-3 varices, red spots on upper endoscopy, and fibrinogen levels <150mg/dL, was applied to our prospective cohort.

Results

In total, 20 children were included in the study, of which 18 exhibited biliary atresia, with a median age of 9 months (range:4-129). Six of them were classified in the bleeding group since they had upper gastrointestinal bleeding during the pre-transplant period. No statistically significant difference was noted between the two groups regarding waiting time prior to transplantation, presence and severity of esophageal varices, Factor V levels (49% vs. 70%), INR (1.8 vs. 1.5), and platelet counts (104 10e3/µL vs. 219 10e3/µL). However, significant differences were observed with respect to fibrinogen levels (109mg/dL vs. 257mg/dL, p<0.05), platelet adenosine diphosphate (ADP)-dependent platelet aggregation (103 AU/min vs. 368AU/min, p<0.05), thrombin-dependent platelet aggregation (265 AU/min vs. 558 AU/min, p<0.05), and clotting time (64 sec vs. 52 sec, p<0.05) in EXTEM analysis. The predictive model of bleeding risk was tested in our prospective cohort, yielding a predictive performance (accuracy) of 85.18% (sensitivity 90.5%, specificity 66.7%, negative predictive value [NPV] 90.5%, and positive predictive value [PPV] 66.7%).

Conclusion

Our study demonstrated that hemostasis is involved in the risk of bleeding from esophageal varices. Low fibrinogen levels were identified as a risk factor for bleeding in children with decompensated cirrhosis, thus suggesting the potential benefit of prophylactic fibrinogen administration in high-risk cases. In the future, the Multiplate® analysis should aid practitioners determining the risk of esophageal variceal bleeding in children with decompensated cirrhosis, and it may possibly be integrated in our predictive model.