Novel insights into the assessment of risk of upper gastrointestinal bleeding in decompensated cirrhotic children.

OBJECTIVES

Cirrhotic children wait-listed for liver transplant are prone to bleeding from gastrointestinal varices. Grade 2-3 esophageal varices, red signs, and gastric varices are well-known risk factors. However, the involvement of hemostatic factors remains controversial because of the rebalanced state of coagulation during cirrhosis.

METHODS

Children suffering from decompensated cirrhosis were prospectively included while being on waitlist. Portal hypertension was assessed by ultrasound and endoscopy. Coagulopathy was evaluated through conventional tests, thromboelastometry, and platelet function testing. The included children were followed up until liver transplantation, and all bleeding episodes were recorded. Children with or without bleeding were compared according to clinical, radiological, endoscopic, and biological parameters. In addition, validation of a predictive model for risk of variceal bleeding comprising of grade 2-3 esophageal varices, red spots, and fibrinogen level <150 mg/dL was applied on this cohort.

RESULTS

Of 20 enrolled children, 6 had upper gastrointestinal bleeding. Significant differences were observed in fibrinogen level, adenosine diphosphate, and thrombin-dependent platelet aggregation. The model used to compute the upper gastrointestinal bleeding risk had an estimated predictive performance of 81.0%. Platelet aggregation analysis addition improved the estimated predictive performance up to 89.0%.

CONCLUSIONS

We demonstrated an association between hemostatic factors and the upper gastrointestinal bleeding risk. A low fibrinogen level and platelet aggregation dysfunction may predict the risk of bleeding in children with decompensated cirrhosis. A predictive model is available to assess the upper gastrointestinal bleeding risk but needs further investigations. Clinicaltrials.gov number: NCT03244332.