Platelet count in predicting bleeding complication after elective endoscopy in children with portal hypertension and thrombocytopenia

ABSTRACT

Background: Thrombocytopenia is a frequent and challenging clinical disorder in patients with portal hypertension. It can increase the risk of bleeding associated with invasive procedures. Standard treatment for thrombocytopenia usually consists of platelet transfusions, which may cause transfusion-related complications including viral or bacterial infection, allo-immunization and febrile non-hemolytic reactions. Uncertainty still exists as to the platelet level at which transfusion is indicated in conjunction with invasive procedures such as endoscopy.Objective: To determine the predictive value of platelet level for post elective endoscopy bleeding in children with portal hypertension and thrombocytopenia.Patients and methods: Children with portal hypertension and thrombocytopenia (platelet count \< 100,000 per μL) were enrolled. Those who had active gastrointestinal (GI) bleeding at the time of admission were excluded. All patients underwent elective upper GI endoscopy for esophagogastric varices surveillance and tested for complete blood count.Results: There were 41 children (malefemale = 2021) with portal hypertension enrolled in this study. Age ranged from 2-16 years (mean±SD = 8.42±4.32 years). The etiology of portal hypertension was biliary atresia in 19 and extra-hepatic portal vein obstruction in 22. Thirty of 41 experienced previous upper GI bleeding. Endoscopic finding was 32 esophageal varices (EVs), two EVs plus gastric varices, and seven without varices. Twenty-one patients underwent endoscopy without therapeutic procedure and 20 went through endoscopic variceal ligation. Two patients developed GI bleeding post endoscopy. Platelet count of children with and without post endoscopy bleeding was not significantly different (68,500±40,305 vs. 73,025±24,030/μL, respectively; p=0.8). By applying receiver operating characteristic (ROC) curves, a platelet count cut-off value of 41,500/μL was obtained, which gave positive and negative predictive values of 12.5% and 96.9%, respectively. The accuracy of this cut-off value as evaluated by applying ROC curves was 80.5%.Conclusion: Patients with platelet count of more than 40,000/μL were almost certainly safe to undergo elective endoscopy without prophylactic platelet transfusions and hence minimize the cost and complications of transfusion.