Copyright 2013 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited. Ultrasonographic Predictors of Esophageal Varices  Roberta V. de Alcantara, y Roberto M. Yamada,  Sı´lvia R. Cardoso, Maria de Fa ´tima,  C.P. Servidoni, and  Gabriel Hessel ABSTRACT Objective: The aim of this study was to identify ultrasonographic predictors of esophageal varices (EVs) in children and adolescents with chronic liver disease (CLD) and extrahepatic portal venous obstruction (EHPVO). Methods: This study evaluates 53 patients younger than 20 years with CLD or EHPVO and no history of bleeding or prophylactic EVs treatment. They were divided into 2 groups: group I (35 with CLD) and group II (18 with EHPVO). Splenorenal shunt (SS), gallbladder wall varices, gallbladder wall thickening (GT), and lesser omental thickness (LOT) were compared with the presence of EVs, gastric varices, and portal hypertensive gastropathy (PHG). Univariate (x 2 test, Fisher exact test, and Wilcoxon signed rank test) and multivariate (logistic regression) analyses were performed. The area under the receiver operating curve was calculated. Results: EVs were observed in 48.5% of patients with CLD and in 83.3% of patients with EHPVO. SS (P ¼ 0.0329) and LOT (P ¼ 0.0151) predicted EV among patients with CLD. A median of 5.3 mm of LOT was considered a predictor of EVs among these patients. Multivariate analysis showed SS as an independent predictor of EVs in patients with EHPVO (odds ratio 15). Gallbladder varices (P ¼ 0.0245) and GT (P ¼ 0.0289) predicted EVs among patients with EHPVO. PHG occurred more often among patients with CLD who had SS (P ¼ 0.0384) and greater LOT (P ¼ 0.0226). Conclusions: SS and a greater LOT were indicative of EV among children and adolescents with CLD. Gallbladder varices and GT were indicative of EVs among patients with EHPVO. SS and a greater LOT were indicative of PHG among patients with CLD. Key Words: children, cirrhosis, extrahepatic portal venous obstruction, lesser omental thickness (JPGN 2013;57: 700–703) L iver cirrhosis is considered the main cause of portal hyper- tension (PH) in the pediatric age group in most countries (1–6); however, extrahepatic portal vein obstruction (EHPVO) is the most frequent cause of PH in Indian children (7). Findings of PH detected by abdominal ultrasound (US) include the presence of splenomegaly, ascites, and portosystemic collateral vessels, such as dilatation of the left gastric vein, paraumbilical vein, and splenorenal shunt (SS) (8,9). In 1 study, the lesser omental thickness (LOT) was found to be greater in 42 children and adolescents with PH when compared with a control group (2). In addition, a greater LOT/aorta diameter ratio was associated with the prediction of esophageal varices (EVs) in the pediatric age group (10). Esophageal variceal bleeding is the major complication of PH in adults experiencing liver cirrhosis (LC), being responsible for 14% to 33% of deaths until 6 weeks after the first bleeding episode (11). Risk factors for esophageal variceal bleeding in these patients include hepatic venous pressure gradient >12 mmHg, large-caliber varices, the presence of ‘‘red wale markings’’ on varices at endo- scopy, and evidence of severe liver disease (ie, Child-Pugh class C patients) (12,13). The prevalence of EVs in adults with LC ranges from 30% to 70% (13 – 18). The risk of developing EV after diagnosis is 9%/year (13). After the first bleeding episode, the mean incidence of bleeding recurrence in untreated patients is 63% within 2 years (19). In children with LC, the prevalence of EV ranges from 50% to 70% (5,20,21). Studies involving adults with LC showed a decreased bleeding risk in patients undergoing primary prophylaxis. Therefore, primary prophylaxis of EV bleeding in cirrhotic adults must be performed in selected patients (22). In children with PH, there are few studies on variceal bleeding prophylaxis (3,4,23–27). Only 1 randomized controlled study is available involving the pediatric age group (3). In that study, 50 of 100 children underwent prophylactic sclerotherapy for EV. This group exhibited 6% of upper gastrointestinal hemorrhage (UGIH), whereas it was observed in 42% of the control group. The presence of EV in children with EHPVO may be observed in 85% to 100% of patients (5,21,28,29). The prevalence of UGIH in these patients ranges from 50% to 77% (28–30). Once EHPVO is diagnosed in children, upper gastrointestinal endoscopy (UGIE) is recommended; however, there are insufficient data in the literature to support whether primary prophylaxis should be performed using a b-blocker or endoscopic therapy (22,31). Considering the high prevalence of EV in cirrhotic adults and the elevated morbidity and mortality related to UGIH, the consensus is that UGIE should be performed at diagnosis and every 1 to 2 years, depending on the severity of liver disease (32,33). Nevertheless, endoscopic screening of all of the patients diagnosed as having LC applies to 50% of normal endoscopies (5,21). Not only are such patients subjected to an unnecessary invasive procedure but also most children will require general anesthesia under such a procedure. In addition, medical services are often overcrowded, and the cost of medical treatment is escalating in the health care system. Studies have attempted to solve this problem by correlating the presence of EV with diversified clinical, laboratory, or ultrasound parameters (5,6,14–16,21,34–39). Restricting UGIE to those patients who have indicators capable of predicting the presence of varices would result in a significant decrease in the Received March 6, 2013; accepted July 29, 2013. From the  Department of Pediatrics, Hospital de Clı ´nicas, UNICAMP, Campinas, and the y Department of Medicine, Universidade Federal de Sa ˜o Carlos, Sa ˜o Carlos, Sa ˜o Paulo, Brazil. Address correspondence and reprint requests to Roberta V. de Alcantara, Departamento de Pediatria/FCM/UNICAMP, Rua Tessa ´lia Vieira de Camargo, 126 Bara ˜o Geraldo, Campinas, Sao Paulo 13083-887, Brazil (e-mail: rva@terra.com.br). The authors report no conflict of interest. Copyright # 2013 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0b013e3182a7bc2e ORIGINAL ARTICLE:HEPATOLOGY AND NUTRITION 700 JPGN  Volume 57, Number 6, December 2013