A Clinical Prediction Rule and Platelet Count Predict Esophageal Varices in Children

Transcription

1 GASTROENTEROLOGY 2011;141: A Clinical Prediction Rule and Platelet Count Predict Esophageal Varices in Children JUAN CRISTÓBAL GANA,* DAN TURNER, GIORGINA MIELI VERGANI, MARK DAVENPORT, TAMIR MILOH, YARON AVITZUR, #, ** JASON YAP, VERONIQUE MORINVILLE, HERBERT BRILL, and SIMON C. LING** *Division of Pediatrics, Gastroenterology, Hepatology and Nutrition Unit, Pontificia Universidad Católica de Chile, Santiago, Chile; Pediatric Gastroenterology and Nutrition Unit, Shaare Zedek Medical Center, The Hebrew University of Jerusalem, Jerusalem, Israel; Paediatric Liver, Gastroenterology and Nutrition Centre, King s College London School of Medicine at King s College Hospital, London, United Kingdom; Department of Paediatric Surgery, Kings College Hospital, London, United Kingdom; Mount Sinai School of Medicine, Mount Sinai, New York; Phoenix Children s Hospital, Phoenix, Arizona, United States; # Institute of Gastroenterology, Nutrition and Liver Diseases, Schneider Children s Medical Centre of Israel, Petah-Tikva, Israel; University of Alberta, Edmonton, Alberta, Canada; Division of Pediatric Gastroenterology and Nutrition, Montreal Children s Hospital, Montreal, Canada; McMaster Children s Hospital, Hamilton, Canada; and **Division of Gastroenterology, Hepatology & Nutrition, Hospital for Sick Children, and Department of Paediatrics, University of Toronto, Toronto, Canada See Covering the Cover synopsis on page BACKGROUND & AIMS: The validation of noninvasive tests to diagnose esophageal varices is a priority in children because repeated endoscopic evaluations are too invasive. We measured the ability of a previously developed noninvasive clinical prediction rule (CPR) to predict the presence of esophageal varices in children. METHODS: We analyzed data from 108 children, younger than age 18, who received endoscopies at 8 centers, to assess portal hypertension from chronic liver disease or portal vein obstruction. Blood test and abdominal ultrasound scan results were obtained within 4 months of endoscopy. Grading of varices identified by endoscopy was confirmed by independent blinded review. Spleen size, based on data from the ultrasound scan, was expressed as a standard deviation score relative to normal values for age. RE- SULTS: Of the children studied, 74 had esophageal varices (69%), including 35 with large varices (32%). The best noninvasive predictors of esophageal varices of any size were as follows: platelet:spleen size z-score ratio (area under the receiver operating characteristic curve [AUROC], 0.84; 95% confidence interval [CI] ), CPR (AUROC, 0.80; 95% CI, ), and platelet count (AUROC, 0.79; 95% CI, ). The positive predictive values for the CPR and platelet count were 0.87 and 0.86, the negative predictive values were 0.64 and 0.63, the positive likelihood ratios were 3.06 and 2.76, and the negative likelihood ratios were 0.64 and 0.63, respectively. Based on positive and negative predictive values, the most accurate noninvasive tests were the CPR and platelet counts. CONCLUSIONS: Noninvasive tests such as CPR and platelet count can assist in triaging children for endoscopy to identify esophageal varices. Keywords: Portal Hypertension; Pediatric Chronic Liver Disease; Diagnostic Tests. Many children with chronic liver disease or portal vein obstruction are at risk of variceal bleeding, which is associated with significant morbidity and mortality. 1 8 Guidelines for adults with portal hypertension recommend performing esophagogastroduodenoscopy (EGD) to identify those with varices who may benefit from prophylactic therapy. Treatment with nonselective -blockade or endoscopic variceal ligation is effective for the prevention of variceal bleeding in cirrhotic adults Currently, there are inadequate data to support a similar approach in children. EGD is the reference standard test for the diagnosis of varices, but it is considered especially invasive in children, time consuming, expensive, and associated with risk. There is, therefore, a pressing need for a noninvasive test that is reliable enough to target EGD only to those children with the highest risk of varices. The American Association for the Study of Liver Disease and the American College of Gastroenterology recognize this as one of the most important areas for research in portal hypertension. 12 In a previous retrospective study, we showed that esophageal varices in children could be predicted accurately by a clinical prediction rule (CPR, calculated from platelet count, spleen size z-score, and albumin concentration) and by the ratio of the platelet count and spleen size z-score (P/SSAZ). 13 A lower CPR score means a higher likelihood for the presence of varices. We now present the results of a multicenter prospective study in which we aimed to validate the ability of the CPR and P/SSAZ ratio to predict the presence of esophageal varices in children with portal hypertension. This study was conducted according to the requirements set by the Quality Assessment of Diagnostic Accuracy Studies checklist. 14,15 Materials and Methods Settings and Eligibility This was a prospective, multicenter study in which 8 centers recruited consecutive children younger than age 18 with chronic liver disease or portal vein thrombosis between 2007 and Abbreviations used in this paper: CI, confidence interval; CPR, clinical predictor rule; P/SSAZ, ratio of platelet count and spleen size z-score; EGD, esophagogastroduodenoscopy; ROC, receiver operator characteristic; AUROC, area under receiver operator characteristic by the AGA Institute /$36.00 doi: /j.gastro

2 2010 GANA ET AL GASTROENTEROLOGY Vol. 141, No Inclusion criteria included those undergoing EGD, either to screen for esophageal varices or for investigating gastrointestinal symptoms. Exclusion criteria were previous portal-systemic shunt surgery or transjugular intrahepatic portal-systemic shunt, previous ligation or sclerotherapy of varices, therapy with -blocker during the previous 6 months, organ transplantation, previous upper gastrointestinal bleeding, and malignancy. The study was approved by the Institutional Review Boards or Research Ethics Committees of all participating centers. End Points The primary end point was the presence of esophageal varices of any size. Secondary outcomes included the presence of large varices and changes in the noninvasive test variables in a subset of patients who underwent repeat endoscopy for clinical indications. Data Collection Data were recorded on a standardized case report form, including demographic information, primary diagnoses, comorbidities, medications, and details of physical examination. Hepatic encephalopathy was determined using the classification of the Pediatric Acute Liver Failure Study Group for children younger than age For older children, the West Haven classification system was used. 17 Blood test results and abdominal ultrasound scan data were obtained from tests performed within 3 weeks and 4 months of the endoscopy, respectively. Spleen size on ultrasound scan was expressed as a standard deviation score relative to previously established normal values for age. 18 Derivation of the CPR has been described previously. 13 The formula for calculation of the CPR is as follows: 0.75 Platelets 2.5 albumin (1) SAZ 5 The severity of liver disease was measured by the Child Pugh score and the Model for End-stage Liver Disease (for ages 12 and older) or Pediatric End-stage Liver Disease (for ages less than 12) scores Physicians who performed the EGD were asked to describe the esophageal varices appearance using two classification schemes: the Paquet 22 and the Cales et al 23,24 classifications. In the Paquet 22 classification, varix size is graded on a 4-point Likert scale: grade 1 varices are small and flattened by insufflation of air; grade 2 varices are slightly larger and do not flatten; grade 3 varices are larger but do not touch in the middle of the lumen; and grade 4 varices are large and touch each other in the middle of the lumen. Physicians were provided with a pictorial description of the 4 grades derived from the original report by Paquet. 22 The Cales et al 23,24 criteria are based on a 3-size graduation: small varices flatten with air insufflation and are not confluent around the esophageal wall, medium varices do not flatten with air insufflation and are not confluent around the esophageal wall, and large varices do not flatten with air insufflation and are confluent around the esophageal wall. This semiquantitative approach was chosen because it is the best validated tool of all variceal sizing systems and it provides good interobserver agreement. 25 Further endoscopic evaluation included the stomach and duodenum. Gastric or duodenal varices were described alongside edema, submucosal petechial areas, and a snake-skin appearance of the stomach was described to be consistent with portal hypertensive gastropathy. Because EGD was performed as part of routine clinical care, the endoscopists were not blinded to the results of the noninvasive blood tests and ultrasound scans. To minimize this potential bias, EGD findings were recorded by videos and/or pictures, which subsequently were reviewed in random order by two additional experienced pediatric gastroenterologists with at least 5 years of experience, blinded to all clinical data. Each assessor independently graded the variceal size and the presence of red wales or red spots. The final variceal grade was assigned to each patient by majority (2 of 3 assessments). If all 3 assigned a different grade, then agreement was reached by discussion and consensus. As a post hoc analysis conducted at the time of data analysis, participating children were identified who had undergone a repeat EGD for clinical indications. Data were collected from chart review including the presence and grading of varices, measurement of spleen size on ultrasound scan (performed within 4 months of the repeat EGD), and bloodwork variables including platelet count and albumin (performed within 3 weeks of the EGD). Children with repeat EGD performed after banding ligation (n 4), liver transplantation (n 2), surgical shunt (n 1), or variceal hemorrhage (n 1) were excluded from this post hoc analysis. Statistical Analysis Data are presented using means standard deviations, medians (interquartile range), and proportions 95% confidence interval (95% CI), as appropriate. Patients were divided into 2 groups: those with varices on EGD and those without. First, data were explored to search for variables that differed between these 2 groups. Continuous variables (such as age, disease duration, laboratory values, and spleen size) were compared using the Student t test or the Wilcoxon rank sum test, as appropriate for the data normality. Categoric variables (such as sex, presence of cirrhosis, and comorbidity) were compared using the chi-square test or the Fisher exact test as appropriate. We emphasize that these analyses were exploratory in nature and, thus, no correction was made for multiple comparisons and a P value of less than.05 was considered statistically significant for all analyses. Interobserver variability for the grading of varices was calculated using kappa statistics with quadratic weights. Primary analysis. The primary analysis assessed the ability of CPR and P/SSAZ to predict the presence of any esophageal varices using diagnostic utility methods. A receiver operator characteristic (ROC) curve was constructed and the area under the ROC (AUROC) was calculated with the corresponding 95% CI. An AUROC of greater than 0.7 was considered indicative of a fair test, 0.8 was considered good, and more than 0.9 was considered an excellent test. The optimal cut-off value of the CPR and P/SSAZ score to predict esophageal varices was determined as the point at which the second diagonal crosses the ROC curve (ie, the point where the shoulder of the curve is closest to the left upper corner of the graph). Sensitivity, specificity, predictive values, and likelihood ratios were calculated for this optimal cut-off score and for 2 other cut-off values, one aimed to optimize sensitivity and the other to optimize specificity. Secondary analyses. Similar diagnostic utility statistics were used for each of the following test variables, as defined in the secondary aims: spleen size, platelet count, and the ratio between aspartate aminotransferase and alanine aminotransferase levels. A logistic regression model then was constructed with the presence of varices as the dependent variables and, governed

3 December 2011 PREDICTING ESOPHAGEAL VARICES IN CHILDREN 2011 Table 1. Baseline Characteristics of 108 Patients Included Variable EV ( ) (n 74) EV ( ) (n 34) P value Diagnosis Biliary atresia 19 (26%) 11 (32%) Autoimmune hepatitis 10 (14%) 5 (15%) Portal vein thrombosis 10 (14%) 0 (0%) Primary sclerosing cholangitis 5 (7%) 4 (13%) Congenital hepatic fibrosis 7 (9%) 0 (0%) Progressive familial intrahepatic cholestasis 6 (8%) 1 (3%) Others a 17 (23%) 13 (38%) Male sex 16 (47%) 37 (50%) NS Mean age, (y) NS Age range 9 mo to 17 y 2 mo to 18 y Age, 2 y 7 (9%) 3 (8%) Age, 2 10 y 22 (30%) 9 (26%) Age, 10 y 45 (61%) 22 (65%) Child Pugh, A/B/C 59/13/2 25/7/2 Mean Child Pugh NS Splenomegaly, n b 70 (95%) 19 (56%).001 a Others includes choledochal cyst (4), hepatitis C (3), Wilson s disease (3), Alagille syndrome (3), overlap syndrome (2), Caroli s disease (2), neonatal sclerosing cholangitis (2), drug-induced liver failure (2), 1-antitrypsin deficiency (2), glycogen storage disease (1), focal nodular hyperplasia (1), lipid storage disease (1), schistosomiasis (1), cryptogenic cirrhosis (1), parenteral nutrition cholestasis (1), and hepatoportal sclerosis (1). b At physical examination. by sample size, the CPR, platelet count, and albumin as the explanatory variables, to determine whether the CPR explains the prediction contribution of the others. This study was not powered for statistical comparison of the performance of competing prediction rules and thus formal analysis was not used. Analyses then were repeated after dichotomizing the outcome by large varices (at least grade 2 by the Cales classification) 23 vs small or no varices. For the post hoc analysis of data from the subset of children who had undergone repeat EGD, the change in CPR was calculated for children grouped according to stable appearance of varices and compared with the change in CPR among children with progression of varices. Because of the small number of children in this subset, no statistical testing was undertaken. Statistical analyses were performed using SPSS (IBM, Armonk, NY) version Results Of 111 children who consented to participate in the study, 3 patients were excluded because of previous gastrointestinal bleeding. The main characteristics of the remaining 108 children are depicted in Table 1. The mean age ( standard deviation) of the entire cohort was 11 years ( 5.4 y) (48% males). Only 1 patient had grade 2 encephalopathy associated with large esophageal varices. All patients underwent EGD, blood tests, and an ultrasound scan within the expected time frame (mean interval between the EGD and blood test was 9 10 days and between the EGD and ultrasound scan was days). The indications for EGD were screening for esophageal varices in 99 children (92%), evaluation of inflammatory bowel disease in 4 children, pretransplant assessment of portal hypertension in 1 child, and for other clinical symptoms in 4 children. Esophageal varices were found in 74 of 108 patients (69%), of whom 35 had large varices (Table 1). Forty-eight patients (44%) had portal hypertensive gastropathy, 14 (13%) had both gastric and esophageal varices, and only 2 (2%) had isolated gastric varices. The interobserver variability for the diagnosis of esophageal varices had a weighted kappa value of 0.67 for the Paquet 22 classification and 0.65 for the Cales classification, which implies good agreement. Most variables did not differ between patients with or without esophageal varices, including age, sex, Child Pugh Score, and Pediatric End-stage Liver Disease or Model for End-stage Liver Disease score (Table 1). Variables found to differ significantly between children with and without varices included splenomegaly on physical examination, spleen length Z-score measured by ultrasound scan, platelet count, P/SSAZ, white blood cell count, presence of collaterals on ultrasound scan, aspartate aminotransferase/platelet ratio, and our previously derived CPR (Table 2). In the multivariate model, only CPR (odds ratio, 0.62; 95% CI, ; P.002) and albumin (odds ratio, 3.1; 95% CI, ; P.004) were independent predictors of varices, whereas platelet count was not (odds ratio, 1.02; 95% CI, ; P.12). By using a univariate model we determined that for every 10-point decrease in the CPR, the likelihood for the presence of varices increases by 95%. In ROC curve analysis, the best predictors of esophageal varices of any size were P/SSAZ (AUROC, 0.84; 95% CI, ), CPR (0.80; 95% CI, ), platelet count (0.79; 95% CI, ), and spleen size z-score (0.76; 95% CI, ) (Figure 1). Cut-off values for the best predicting variables for esophageal varices of any size that maximized sensitivity and negative predictive value are

4 2012 GANA ET AL GASTROENTEROLOGY Vol. 141, No. 6 Table 2. Variables to Differentiate Those With any Esophageal Varices Versus Those Without No varices (n 34) Any varices (n 74) P value Spleen length and markers of hypersplenism SAZ Platelets, 10^9/L P/SSAZ 15 (10 37) 8 (4 10).001 White blood count, 10^9/L 5.4 ( ) 3.8 (3-5.1).001 Hemoglobin level, g/l NS Other USS variables Collaterals 4 (12%) 29 (39%).006 Ascites 3 (9%) 8 (11%) NS Liver synthetic function Albumin level, g/l NS INR NS Other tests and ratios Conjugated bilirubin level, umol/l 5.1 (0 27) 1.7 (0 86) NS AST level 64 ( ) 62 (36 148) NS ALT level 59 (32 159) 56 (30 107) NS AST/ALT ratio NS AST/PLT ratio 0.45 ( ) 0.7 ( ).016 GGT level 73 (33 121) 76 (26 204) NS Alkaline phosphatase level 265 ( ) 311 ( ) NS Creatinine level, umol/l NS CPR NOTE. Data are presented as mean standard deviation or medians (interquartile range), as appropriate for the distribution normality. ALT, alanine aminotransferase; AST, aspartate aminotransferase. presented in Table 3. The best predictors of large esophageal varices were platelet count (AUROC, 0.73; 95% CI, ), P/SSAZ (0.73; 95% CI, ), followed by the CPR (0.68; 95% CI, ). Excluding patients with portal vein thrombosis had no significant effect on the diagnostic performance of the CPR or of the other noninvasive test variables. For instance, in a subgroup analysis of the 98 children with parenchymal disease (ie, excluding the 10 patients with portal vein thrombosis), the AUROC for the CPR to differentiate patients with varices of any size was similar to the full cohort (0.81; 95% CI, ) and to differentiate those with large varices from others was 0.68 (95% CI, ). Sixteen children underwent repeat endoscopy for clinical indications at a median time of 15 months (range, 1 28 mo) after their initial endoscopy, and had noninvasive test results available within the appropriate time window of their repeat endoscopy. The change in the CPR correlated with the observed change in varices. Children in whom varices developed (n 1) or enlarged (n 5) showed a change in CPR of 5.2% ( vs ). Children with no change in variceal size (n 10) showed a change in CPR of 1.2% ( vs ). Figure 1. ROC curve to differentiate those with any esophageal varices vs those without.

5 December 2011 PREDICTING ESOPHAGEAL VARICES IN CHILDREN 2013 Table 3. Performance Characteristics of the Best Variables for Diagnosis of Esophageal Varices Cut-off value Sensitivity, % Specificity, % PPV NPV LR LR AUROC P value P/SSAZ CPR Platelet count, ( 10^9/L) LR, positive likelihood ratio; LR, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value. Discussion In this prospective multicenter study, we aimed to validate the findings of our retrospective study in which we derived a CPR for the noninvasive diagnosis of esophageal varices in children, and also tested the predictive ability of other simple, commonly available, and reproducible variables. 13 When compared with EGD, the AUROC obtained here indicates that the CPR is a good noninvasive diagnostic test for esophageal varices. The CPR was responsive to changes in varices in a subset of children who underwent repeat endoscopy. If applied to a similar population of children with 69% prevalence of varices, then 87% of children with a positive CPR will have varices on EGD and 21% of children will be classified incorrectly by the CPR (including 18% of those with varices). Alternatively, 86% and 79% of children with a positive platelet count and P/SSAZ, respectively, will have varices on EGD, with incorrect classification of 22% by platelet count and 27% by P/SSAZ. The preferred noninvasive tests are therefore the CPR or platelet count, despite the slightly higher AUROC for P/SSAZ. The CPR or platelet count therefore may be used as noninvasive tests for esophageal varices in a clinical or research setting to triage children to undergo EGD for confirmation and grading of varices. In clinical practice, we previously showed that up to 70% of pediatric hepatologists and gastroenterologists perform screening endoscopies in selected children at risk of varices and that the majority of children with portal hypertension favor undergoing EGD to quantify the risk of bleeding, even if they know that no prophylactic treatment will be made available thereafter. 26 Knowledge of the CPR or platelet count will enable more appropriate selection of patients for EGD. Instead of following adult guidelines that recommend a screening endoscopy for all patients with cirrhosis, pediatric hepatologists can now calculate the CPR or interpret the platelet count to provide a useful measure of the likelihood of varices being present in their patients before undertaking EGD (Figure 2). Such practice will reduce the number of children who otherwise would be required to undergo invasive testing. Future studies should investigate the ability to further improve the sensitivity of this test and thus enhance its clinical utility. For research, the results of our study will enhance the ability to conduct pediatric studies of primary prophylaxis against variceal bleeding. Institutional Review Boards may be concerned by protocols that require children to undergo invasive testing such as EGD to determine their eligibility for a research study. Use of the CPR or platelet count to triage pediatric research subjects for screening EGD will reduce the number of children who must otherwise undergo EGD to determine that they have no varices, and thus reassure institutional review boards that unnecessary testing is minimized. Although 50% of those clinicians who perform screening EGD also will provide prophylactic therapy against bleeding if varices are seen, 26 there is currently no evidence for the efficacy of primary prophylaxis in children. The feasibility of conducting a study of primary prophylaxis in children has been questioned and the development of techniques to enable the various challenges to be overcome is therefore a research priority in pediatric hepatology. 27 Platelet count may be preferred over the CPR because of its simplicity when compared with the need to calculate the CPR. However, platelet count did not reach significance as an independent predictor of varices in the multivariate regression analysis and the CPR provides a slight advantage in diagnostic accuracy (eg, the positive likelihood ratio of CPR was slightly better than that of platelet count). Nevertheless, because the performance characteristics of the CPR are not substantially better than those of the platelet count, we suggest that based on our data either of these variables can be used. We expect that development of readily available electronic tools will enable the CPR calculation to be quickly undertaken by clinicians and thereby optimize the accuracy of clinical decisions in the future. Platelet count and P/SSAZ are independent of liver function tests and therefore might be used in patients with portal hypertension resulting from portal vein Figure 2. Flowchart describing the use of the CPR or platelet count in clinical practice.

6 2014 GANA ET AL GASTROENTEROLOGY Vol. 141, No. 6 thrombosis. In support of this, the removal of these patients from the overall analysis had no significant effect on the diagnostic accuracy among the remaining patients with liver disease. The expression of spleen length as a standard deviation score (Z-score) enables appropriate comparison between children of different ages. The ratio between spleen diameter and platelet count repeatedly has shown high diagnostic accuracy for esophageal varices in studies of cirrhotic adults, with the AUROC ranging from 0.86 to The optimal cut-off value for this ratio has yet to be determined and depends in part on the severity of disease in the patient population under consideration. The platelet count/spleen diameter ratio has a logical pathophysiological basis in children and adults with portal hypertension. The increase in spleen length in patients with chronic liver disease almost always reflects the increased portal pressure 33,34 and thrombocytopenia may be the result of splenic pooling of platelets owing to portal hypertension, immune-mediated mechanisms, or lower thrombopoietin synthesis Data to support the noninvasive identification of varices in children are sparse. In a recent study of children with portal hypertension, cirrhotic children with splenomegaly were 14.6-fold more likely to have esophageal varices compared with cirrhotic children without splenomegaly. Hypoalbuminemia increased the likelihood of varices (odds ratio, 4.17; 95% CI, ), although the significance of thrombocytopenia in the univariate analysis did not hold in the multivariable modeling. 38 Ultrasound transient elastography (Fibroscan, Echosens Paris, France) has shown promise for the diagnosis of esophageal varices in children with biliary atresia in 2 studies, although the expense and availability of the equipment is a challenge for many centers, and the reproducibility in children (especially young children and infants) has not yet been established. 39,40 In addition, we are concerned that some of these previous pediatric studies have inadequate reporting of their methodology such that the risk of bias cannot be evaluated adequately (eg, blinding of investigators to the results of the index or reference tests, clarity of the inclusion and exclusion criteria, presence of previous history of esophageal bleeding, and other components of the Quality Assessment of Diagnostic Accuracy Studies guidelines). 14,15 Fifteen children with portal vein cavernoma and esophageal varices were evaluated in another pediatric study that analyzed ultrasound markers for esophageal varices. Abdominal ultrasound scan revealed an increased lesser omentum/aorta diameter ratio in children with portal hypertension, compared with controls (P.001). 41 In contrast, many studies in cirrhotic adults have shown the potential for noninvasive tests to identify esophageal varices, including the ratio of platelet count to spleen size, blood test scores of liver fibrosis, capsule endoscopy, imaging with computerized tomography or magnetic resonance imaging scans, and so forth. 28,30 32,42 56 Cochrane systematic reviews currently are being undertaken by us to help establish which of these different approaches offers the greatest noninvasive diagnostic accuracy for esophageal varices Ultimately, future studies will explore the ideal use of these various approaches alone or in combination to predict the presence of varices and, more importantly, future episodes of variceal bleeding. In the design and conduct of this study, we closely followed the Quality Assessment of Diagnostic Accuracy Studies checklist, which ensured appropriate methodology. 14,15 This list consists of 14 items and allows for inclusion of additional items as appropriate for a specific study. According to these checklist items, we included an appropriately representative spectrum of patients (ie, only children with cirrhosis or portal hypertension at risk of having varices), clearly described our selection criteria, used the appropriate reference standard, minimized the delay between the noninvasive and reference standard tests, and avoided partial and differential verification bias. Reporting of blood test and ultrasound results was blinded to the results of the EGD. Blinded reporting of the EGD was achieved by later review of video and/or pictures by additional assessors. In summary, our results show that noninvasive tests have good diagnostic accuracy in the identification of children with esophageal varices. We have confirmed that the CPR and platelet count have the best overall performance characteristics and that either can be used to triage children with portal hypertension to undergo EGD. References 1. Lykavieris P, Gauthier F, Hadchouel P, et al. Risk of gastrointestinal bleeding during adolescence and early adulthood in children with portal vein obstruction. J Pediatr 2000;136: Goncalves ME, Cardoso SR, Maksoud JG. Prophylactic sclerotherapy in children with esophageal varices: long-term results of a controlled prospective randomized trial. J Pediatr Surg 2000;35: Miga D, Sokol RJ, Mackenzie T, et al. Survival after first esophageal variceal hemorrhage in patients with biliary atresia. J Pediatr 2001;139: van Heurn LW, Saing H, Tam PK. Portoenterostomy for biliary atresia: long-term survival and prognosis after esophageal variceal bleeding. J Pediatr Surg 2004;39: Kobayashi A, Itabashi F, Ohbe Y. Long-term prognosis in biliary atresia after hepatic portoenterostomy: analysis of 35 patients who survived beyond 5 years of age. J Pediatr 1984;105: Toyosaka A, Okamoto E, Okasora T, et al. Outcome of 21 patients with biliary atresia living more than 10 years. J Pediatr Surg 1993;28: Mitra SK, Kumar V, Datta DV, et al. Extrahepatic portal hypertension: a review of 70 cases. J Pediatr Surg 1978;13: Webb LJ, Sherlock S. The aetiology, presentation and natural history of extra-hepatic portal venous obstruction. QJM 1979;48: Khuroo MS, Khuroo NS, Farahat KL, et al. Meta-analysis: endoscopic variceal ligation for primary prophylaxis of oesophageal variceal bleeding. Aliment Pharmacol Ther 2005;21: Garcia-Pagan JC, Bosch J. Endoscopic band ligation in the treatment of portal hypertension. Nat Clin Pract Gastroenterol Hepatol 2005;2:

7 December 2011 PREDICTING ESOPHAGEAL VARICES IN CHILDREN Triantos CK, Burroughs AK. Prevention of the development of varices and first portal hypertensive bleeding episode. Best Pract Res Clin Gastroenterol 2007;21: Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Hepatology 2007;46: Gana JC, Turner D, Roberts EA, et al. Derivation of a clinical prediction rule for the noninvasive diagnosis of varices in children. J Pediatr Gastroenterol Nutr 2010;50: Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3: Whiting PF, Weswood ME, Rutjes AW, et al. Evaluation of QUADAS, a tool for the quality assessment of diagnostic accuracy studies. BMC Med Res Methodol 2006;6: Bucuvalas J, Yazigi N, Squires RH Jr. Acute liver failure in children. Clin Liver Dis 2006;10: , vii. 17. Ferenci P, Lockwood A, Mullen K, et al. Hepatic encephalopathy definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, Hepatology 2002;35: Megremis SD, Vlachonikolis IG, Tsilimigaki AM. Spleen length in childhood with US: normal values based on age, sex, and somatometric parameters. Radiology 2004;231: Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001;33: McDiarmid SV, Anand R, Lindblad AS. Development of a pediatric end-stage liver disease score to predict poor outcome in children awaiting liver transplantation. Transplantation 2002;74: Pugh RN, Murray-Lyon IM, Dawson JL, et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: Paquet KJ. Prophylactic endoscopic sclerosing treatment of the esophageal wall in varices a prospective controlled randomized trial. Endoscopy 1982;14: Cales P, Zabotto B, Meskens C, et al. Gastroesophageal endoscopic features in cirrhosis. Observer variability, interassociations, and relationship to hepatic dysfunction. Gastroenterology 1990;98: Cales P, Buscail L, Bretagne JF, et al. [Interobserver and intercenter agreement of gastro-esophageal endoscopic signs in cirrhosis. Results of a prospective multicenter study]. Gastroenterol Clin Biol 1989;13: Cales P, Oberti F, Bernard-Chabert B, et al. Evaluation of Baveno recommendations for grading esophageal varices. J Hepatol 2003;39: Gana JC, Valentino PL, Morinville V, et al. Variation in care for children with esophageal varices: a study of physicians, patients, and families approaches and attitudes. J Pediatr Gastroenterol Nutr 2011;52: Ling SC, Walters T, McKiernan PJ, et al. Primary prophylaxis of variceal hemorrhage in children with portal hypertension: a framework for future research. J Pediatr Gastroenterol Nutr 2011;52: Giannini E, Botta F, Borro P, et al. Platelet count/spleen diameter ratio: proposal and validation of a non-invasive parameter to predict the presence of oesophageal varices in patients with liver cirrhosis. Gut 2003;52: Giannini EG, Botta F, Borro P, et al. Application of the platelet count/spleen diameter ratio to rule out the presence of oesophageal varices in patients with cirrhosis: a validation study based on follow-up. Dig Liver Dis 2005;37: Giannini EG, Zaman A, Kreil A, et al. Platelet count/spleen diameter ratio for the noninvasive diagnosis of esophageal varices: results of a multicenter, prospective, validation study. Am J Gastroenterol 2006;101: Sethar GH, Ahmed R, Rathi SK, et al. Platelet count/splenic size ratio: a parameter to predict the presence of esophageal varices in cirrhotics. J Coll Physicians Surg Pak 2006;16: Zimbwa TA, Blanshard C, Subramaniam A. Platelet count/spleen diameter ratio as a predictor of oesophageal varices in alcoholic cirrhosis. Gut 2004;53: Liangpunsakul S, Ulmer BJ, Chalasani N. Predictors and implications of severe hypersplenism in patients with cirrhosis. Am J Med Sci 2003;326: Pockros PJ, Duchini A, McMillan R, et al. Immune thrombocytopenic purpura in patients with chronic hepatitis C virus infection. Am J Gastroenterol 2002;97: Peck-Radosavljevic M. Thrombocytopenia in liver disease. Can J Gastroenterol 2000;14(Suppl D):60D 66D. 36. Peck-Radosavljevic M. Review article: coagulation disorders in chronic liver disease. Aliment Pharmacol Ther 2007;26(Suppl 1): Giannini E, Botta F, Borro P, et al. Relationship between thrombopoietin serum levels and liver function in patients with chronic liver disease related to hepatitis C virus infection. Am J Gastroenterol 2003;98: Fagundes ED, Ferreira AR, Roquete ML, et al. Clinical and laboratory predictors of esophageal varices in children and adolescents with portal hypertension syndrome. J Pediatr Gastroenterol Nutr 2008;46: Chang HK, Park YJ, Koh H, et al. Hepatic fibrosis scan for liver stiffness score measurement: a useful preendoscopic screening test for the detection of varices in postoperative patients with biliary atresia. J Pediatr Gastroenterol Nutr 2009;49: Chongsrisawat V, Vejapipat P, Siripon N, et al. Transient elastography for predicting esophageal/gastric varices in children with biliary atresia. BMC Gastroenterol 2011;11: Moreno E, Torres P, Trejo C, et al. [Predictive value of ultrasonography in portal hypertension]. G E N 1991;45: Ng FH, Wong SY, Loo CK, et al. Prediction of oesophagogastric varices in patients with liver cirrhosis. J Gastroenterol Hepatol 1999;14: Schepis F, Camma C, Niceforo D, et al. Which patients with cirrhosis should undergo endoscopic screening for esophageal varices detection? Hepatology 2001;33: Thomopoulos KC, Labropoulou-Karatza C, Mimidis KP, et al. Noninvasive predictors of the presence of large oesophageal varices in patients with cirrhosis. Dig Liver Dis 2003;35: Zein CO, Lindor KD, Angulo P. Prevalence and predictors of esophageal varices in patients with primary sclerosing cholangitis. Hepatology 2004;39: Thabut D, Trabut JB, Massard J, et al. Non-invasive diagnosis of large oesophageal varices with FibroTest in patients with cirrhosis: a preliminary retrospective study. Liver Int 2006;26: Kazemi F, Kettaneh A, N kontchou G, et al. Liver stiffness measurement selects patients with cirrhosis at risk of bearing large oesophageal varices. J Hepatol 2006;45: Cottone M, D Amico G, Maringhini A, et al. Predictive value of ultrasonography in the screening of non-ascitic cirrhotic patients with large varices. J Ultrasound Med 1986;5: Chalasani N, Imperiale TF, Ismail A, et al. Predictors of large esophageal varices in patients with cirrhosis. Am J Gastroenterol 1999;94: Pilette C, Oberti F, Aube C, et al. Non-invasive diagnosis of esophageal varices in chronic liver diseases. J Hepatol 1999;31: Zaman A, Hapke R, Flora K, et al. Factors predicting the presence of esophageal or gastric varices in patients with advanced liver disease. Am J Gastroenterol 1999;94: Madhotra R, Mulcahy HE, Willner I, et al. Prediction of esophageal varices in patients with cirrhosis. J Clin Gastroenterol 2002;34:

8 2016 GANA ET AL GASTROENTEROLOGY Vol. 141, No Zaman A, Becker T, Lapidus J, et al. Risk factors for the presence of varices in cirrhotic patients without a history of variceal hemorrhage. Arch Intern Med 2001;161: Vanbiervliet G, Barjoan-Marine E, Anty R, et al. Serum fibrosis markers can detect large oesophageal varices with a high accuracy. Eur J Gastroenterol Hepatol 2005;17: Sethar GH, Ahmed R, Rathi SK, et al. Platelet count/splenic size ratio: a parameter to predict the presence of esophageal varices in cirrhotics. J Coll Physicians Surg Pak 2006;16: Hong WD, Dong LM, Jiang ZC, et al. Prediction of large esophageal varices in cirrhotic patients using classification and regression tree analysis. Clinics (Sao Paulo) 2011;66: Gana JC, Turner D, Yap J, et al. Platelet count, spleen length, and platelet count/spleen length ratio for the diagnosis of oesophageal varices in patients with chronic liver disease or portal vein thrombosis (Protocol). Cochrane Database Syst Rev 2010;10: CD Gana JC, Turner D, Yap J, et al. Transient ultrasound elastography and magnetic resonance elastography for the diagnosis of oesophageal varices in patients with chronic liver disease or portal vein thrombosis (Protocol). Cochrane Database Syst Rev 2010; 10:CD Gana JC, Turner D, Yap J, et al. Capsule endoscopy for the diagnosis of oesophageal varices in patients with chronic liver disease or portal vein thrombosis (Protocol). Cochrane Database Syst Rev 2010;10:CD Gana JC, Turner D, Yap J, et al. Non-invasive test of liver fibrosis for the diagnosis of oesophageal varices in patients with chronic liver disease or portal vein thrombosis (Protocol). Cochrane Database Syst Rev 2010;10:CD Gana JC, Turner D, Yap J, et al. Magnetic resonance imaging, computer tomography scan, and oesophagography for the diagnosis of oesophageal varices in patients with chronic liver disease or portal vein thrombosis (Protocol). Cochrane Database Syst Rev 2010;10:CD Received April 12, Accepted August 29, Reprint requests Address requests for reprints to: Juan Cristóbal Gana, MD, Division of Pediatrics, Gastroenterology, Hepatology and Nutrition Unit, Pontificia Universidad Católica de Chile, Lira 85, 5th Floor, Santiago, Región Metropolitana, Chile jcgana@gmail.com; fax: Acknowledgments The authors would like to acknowledge The Canadian Association for Study of the Liver Schering Victor Feinman Research Training Fellowship; Isabel Ansaldo and Margarita Mena for their support and graphics. Conflicts of interest The authors disclose no conflicts.