Biliary atresia (BA) is a rare cholestatic liver disease of infancy, but still the most common indication for liver transplantation in childhood (1). The first-line surgical treatment of BA is portoenterostomy aiming to establish bile flow. Advancing fibrosis of the liver is associated with development of portal hypertension, esophageal varices, and variceal bleeding in a significant number of patients also after successful portoenterostomy (2,3). For adult cirrhotic patients, screening for varices is recommended, and for patients with large varices, a primary prophylaxis of bleeding with either β-blockers or band ligation (4,5).
Evidence-based recommendations of surveillance endoscopies for esophageal varices in children with liver cirrhosis are scarce (4,6,7). Platelet count with or without spleen size assessment showed predictive value as a noninvasive predictor of varices in children with chronic liver disease or portal vein obstruction (8). In patients with BA specifically, endoscopic injection sclerotherapy has been described with favorable results as both primary (9–11) and secondary prophylaxis of variceal hemorrhage (11–13). Endoscopic variceal ligation has been successfully used as primary prophylaxis in patients with BA older than 3 years (14,15). These studies have included selected patient materials based on survival following portoenterostomy (11), size of esophageal varices (9), or signs of portal hypertension (16). A prospective study demonstrated that patients with BA with gastric varices and red markings on esophageal varices were at the highest risk for variceal bleeding and should receive primary prophylaxis of variceal bleeding (16); however, which patients with BA should undergo screening endoscopy for varices remains unclear.
At our institution, patients with BA have unselectively undergone routine upper gastrointestinal endoscopic surveillance and prophylactic injection sclerotherapy of esophageal varices. We aimed to assess the efficiency and complications of primary prophylaxis sclerotherapy and risk factors of esophageal varices and associated upper gastrointestinal bleeding. To this end, we related disease characteristics, clearance of jaundice, laboratory tests reflecting liver function and hypersplenism, and sonographic signs of portal hypertension recorded at set time intervals after portoenterostomy and at the time of each endoscopy to endoscopic findings and bleeding episodes.
A total of 59 consecutive patients with BA born between 1987 and 2008 were included. Eleven patients had not attended surveillance endoscopies: 6 died as infants (2 of severe heart defect and 4 of liver failure before receiving a liver transplant in the early years of the liver transplantation program), 3 underwent liver transplantation before beginning of surveillance at 4 to 13 months, and 2 had not yet started surveillance. One patient's charts were unavailable for review. Of the 47 eligible patients 15 were referred to our hospital after having undergone portoenterostomy elsewhere. All of the patients had undergone at least 1 upper gastrointestinal endoscopy (UGE). The median follow-up was 1.7 years (range 0.5–18.9), and actual overall survival at 2 years was 71%. In 2005, BA treatment in Finland was centralized to Helsinki, and thus the patients in the current series are slightly concentrated to recent years (34 born 1987–2004, mean 2/year, 13 born 2005–2008, mean 3/year).
The patients were followed up in the Children's Hospital, Helsinki University and Helsinki University Central Hospital, a tertiary referral center for severe pediatric liver disease in Finland. Here a nationwide liver transplantation program was started in 1987 and since then patients with BA have undergone a program of endoscopic surveillance and prophylactic injection sclerotherapy of esophageal varices. The surveillance was started at 12 months, or earlier, if signs of advanced cirrhosis or bleeding occurred. Grade 2 and 3 (17) varices were treated with injection sclerotherapy when encountered. In sclerotherapy 0.5 mL of sodium tetradecyl sulfate (Fibro-Vein, Std Pharmaceutical Product Ltd, Hereford, UK, or Sotradecol, AngioDynamics, New York) was injected per site. A median of 2 (range 1–7) sites were injected per session. Sclerotherapies were repeated at 2- to 4-week intervals until varices were considered eradicated. If no or grade 1 varices were encountered UGE was repeated at yearly intervals. An experienced endoscopist performed all endoscopies under general anesthesia with a Pentax (Hoya Corporation, Tokyo, Japan) flexible endoscope.
The patient charts were reviewed and the clinical data on diagnostic modalities, operative findings, disease characteristics, as well as technique and outcome of portoenterostomy were recorded. Also, the endoscopic findings, bleeding episodes, red blood cell transfusions, abdominal ultrasound reports for signs of portal hypertension (enlarged spleen, ascites), management of sclerotherapies, and complications were evaluated. Sonographic signs of portal hypertension and laboratory tests were recorded at 3 and 6 months after portoenterostomy and at the time of each endoscopy. The following blood values were recorded: prothrombin ratio, hemoglobin, platelets, albumin, prealbumin, alanine aminotransferase, and total and conjugated bilirubin. The follow-up ended to liver transplantation, death or end of study period on December 31, 2009. Patients were listed for liver transplantation based on the following parameters: jaundice (bilirubin and bile acid levels), liver function (clotting factors, prealbumin), portal hypertension (portal flow, spleen size, varices, ascites), radiological findings of the liver (cirrhotic change, size, biliary lakes), and patient growth and quality of life (itching, nutrition, cholangitis episodes, bone health). All liver grafts were from deceased donors.
Diagnosis of BA was confirmed by cholangiographic findings and/or histopathology of the biliary remnant and the liver. Clearance of jaundice after portoenterostomy was defined as a decrease in serum bilirubin <20 μmol/L (18–20). The degree of liver fibrosis in biopsies obtained at portoenterostomy was classified as no fibrosis, mild fibrosis, moderate (bridging) fibrosis, or cirrhosis. Esophageal varices were graded from 0 to 3 according to Baveno I Consensus Workshop (17). For analytical purposes patients with no or grade 1 varices were pooled and considered as having no varices. Presence or absence of gastric varices and portal hypertensive gastropathy were also recorded. Upper gastrointestinal bleeding was defined as hematemesis with or without associated melena treated with red blood cell transfusion. We considered esophageal varices eradicated if no or grade 1 varices were encountered in subsequent surveillance endoscopies following sclerotherapy. Variceal bleeding was considered recurrent if episodes of upper gastrointestinal bleeding reoccurred after commencing sclerotherapies. Cholangitis was defined as a pyrexial illness not attributed to other source, treated with intravenous antibiotics.
The study conforms to the principles of the 1975 Declaration of Helsinki. The ethics committee of the Hospital District of Helsinki and Uusimaa approved this study a priori.
Unless otherwise stated, data are expressed as median and range. For comparisons between groups, we used the chi square or the Fisher exact test for dichotomous variables, and the Mann-Whitney U test for continuous variables. For paired comparisons within a group we used Wilcoxon signed ranks test. Actuarial transplant-free survival and cumulative incidence of esophageal varices and bleeding were calculated by the Kaplan-Meier method. To assess risk factors of varices and occurrence of upper gastrointestinal bleeding, we used binary logistic regression analyses. For statistical tests we used SPSS version 17.0 (SPSS, Chicago, IL).
Patient characteristics are shown in Table 1. The median age at portoenterostomy was 69 days and a total of 40% of patients cleared their jaundice. Eight patients (17%) had associated splenic malformation syndrome. In 4 (9%), spleen size assessment was impossible: 2 had polysplenia, 1 asplenia, and 1 had a multilobular spleen.
Endoscopic Appearance of Varices and Sclerotherapy
In total, 230 upper gastrointestinal endoscopies with a median of 3 (range 1–30) per patient were performed and esophageal varices were encountered in 28 (60%) patients at a median age of 11 (range 4–165) months. The first endoscopy was carried out at a median age of 8 (4–87) months. When first encountered, varices were classified as grade 2 in 10 patients and grade 3 in 18 patients. At the first endoscopy, 31 patients had no or grade 1 varices. Of them, 12 later developed grade 2 or 3 varices at the median age of 1.5 (0.5–13.8) years.
A total of 27 patients received sclerotherapy in 115 sessions (median 2; range 1–19). In 9 patients up to 2 variceal columns were treated in 1 or 2 sessions. In 18 patients sclerosant was injected at least into 3 sites in 3 or more sclerotherapy sessions. Minor bleeding occurred in 12 (10%) sessions. Bleeding necessitating red blood cell transfusions occurred after 2 sessions; superficial esophageal ulceration was detected in 3 patients. No strictures were encountered. One patient had an esophageal perforation, which was followed by mediastinitis and death—this patient had terminal liver failure and repeated variceal hemorrhages.
More than 1 endoscopy (median 4; range 2–30) was performed in 34 patients, of whom 15 had failed and 18 successful portoenterostomy. The median interval between the first and the last examination was 10 (range 0–200) months (Table 2). Among failed portoenterostomy patients variceal size diminished in 5 (33%), remained unchanged in 5 (33%), and progressed in 5 (33%); among successful portoenterostomy patients varices diminished in 3 (17%), remained unchanged in 11 (61%), and progressed in 4 (22%) (Table 2).
Risk Factors of Esophageal Varices
The median bilirubin did not significantly differ between the first and the last endoscopy among patients whose varices diminished (76 [10–445] vs 69 [17–922] μmol/L, P = 0.219), or remained unchanged (33 [3–556] vs 16 [3–693] μmol/L, P = 0.232), but was significantly higher at the last endoscopy among patients whose varices progressed (211 [5–766] vs 612 [5–1252] μmol/L, P = 0.031). At 6 months after portoenterostomy the patients who developed varices had significantly higher median serum bilirubin levels (205 [5–1209] μmol/L) compared with those without varices (22 [4–468] μmol/L, P = 0.041). A logistic regression analysis including bilirubin values, platelet levels, and prothrombin ratio at 6 months after portoenterostomy as variables was performed to assess risk factors for development of esophageal varices. Serum bilirubin >40 μmol/L was the only significant risk factor of varices with an odds ratio (OR) of 7.9 (95% confidence interval [CI] 1.2–53, P = 0.046). At the time of endoscopy when varices were first discovered, patients with varices had significantly higher bilirubin, lower prothrombin ratio, and more often an enlarged spleen in relation to those with no varices at the time of their last surveillance endoscopy (Table 3). At the last endoscopy or at variceal detection 6 of 31 patients had a normal-sized spleen (Table 3). Five (16%) of these had both a normal platelet count (>200/E9/L) and a normal serum bilirubin level (<20 μmol/L), and none of these 5 had varices (P = 0.005).
Varices did not modify the overall transplant-free survival (Table 3); however, patients with a normal serum bilirubin level (<20 μmol/L) at the time of discovery of varices had a significantly longer median transplant-free survival (4.2 [0.2–9.5] years) after variceal detection compared with those with increased bilirubin (0.2 [0.0–1.0] years, P = 0.001) (Fig. 1).
Of the 19 patients with a successful portoenterostomy, 10 had grade 2 or 3 varices. The patients with varices had cleared their jaundice later (4 [1–26] vs 2 [1–4] months, P = 0.017) and already at the first endoscopy had significantly more often an enlarged spleen (8/10 vs 0/7, P = 0.002). At the time when varices were first detected, patients with varices had higher bilirubin (18 [5–37] vs 8 [2–16] μmol/L, P = 0.014], higher alanine aminotransferase (72 [42–1703] vs 40 [16–77] U/L, P = 0.033], lower platelets (127 [40–141] vs 224 [44–405] E9/L, P = 0.022), lower prealbumin (95 [52–172] vs 163 [114–207] mg/L, P = 0.038), and more often an enlarged spleen (9/9 vs 4/8, P = 0.029) when compared with those without varices at the time of their last surveillance endoscopy. Of the 10 patients who developed esophageal varices after successful portoenterostomy the liver function had deteriorated before development of varices in 2, and 8 patients developed varices with stable liver function. When these 8 patients were compared with patients with no varices at their last endoscopy, the patients with varices had significantly lower platelets (127 [27–187] vs 224 [44–405] E9/L, P = 0.036), higher bilirubin (17 [5–19] vs 8 [2–16] μmol/L, P = 0.045), and higher conjugated bilirubin (7 [2–10] vs 2 [1–8] μmol/L, P = 0.050).
Occurrence and Risk Factors of Upper Gastrointestinal Bleeding
Upper gastrointestinal bleeding occurred in 16 (34%) patients; they underwent endoscopic examination during the same admission (Table 4). Bleeding originated from esophageal varices in 13 patients, whereas in 3 patients neither bleeding site nor esophageal or gastric varices were identified. The first bleeding episode occurred at a median age of 8.5 (5–20) months. Six patients experienced their first bleeding episode before the first surveillance endoscopy, at a median age of 7 (5–10) months. Six patients bled median 3 (1–10) months after first endoscopy, but no varices had been present in endoscopies before the bleeding. Overall, of the 28 patients with varices, 16 received sclerotherapy before the first bleeding episode, and 4 of them (25%) bled. Recurrent bleeding after commencing sclerotherapies occurred in 12 patients.
No patient bled from varices after a successful portoenterostomy. Already at 3 months after portoenterostomy, the bleeders had significantly higher serum bilirubin levels than the rest (283 [28–815] μmol/L vs 21 [5–558] μmol/L, P < 0.001), and at 6 months after portoenterostomy, higher bilirubin (385 [26–1209] μmol/L vs 18 [4–385] μmol/L, P < 0.000) and lower prothrombin ratio (48 [16–112]% vs 73 [35–170]%, P = 0.007). A logistic regression analysis including bilirubin >40 μmol/L, platelet levels <200 E9/L, and prothrombin ratio <70% at 3 months after portoenterostomy as variables was performed to assess risk factors of bleeding. The only significant risk factor of upper gastrointestinal bleeding was bilirubin >40 μmol/L with an OR of 17 (95% confidence interval, 1.7–175, P = 0.017). At the time of the first endoscopy, the grade of varices was similar between the nonbleeders and the bleeders, but the bleeders had significantly more often ascites present (Table 4). At the time of the first bleeding episode, the bleeders had significantly higher bilirubin and alanine aminotransferase levels, lower prothrombin ratio and hemoglobin, and more often ascites present, when compared with patients without bleeding (Table 4). Before first bleeding 9 of 12 patients had an enlarged spleen, whereas 16 of 22 patients with no bleeding had an enlarged spleen (Table 4). Three patients with a normal-sized spleen before bleeding also had normal platelets (>200 E9/L), but these 3 were failed PE patients with marked hyperbilirubinemia (bilirubin 144–314 μmol/L). Deterioration of hepatic function by the first bleeding episode translated into decreased transplant-free survival among the bleeders (Table 4).
Effect of Clearance of Jaundice on Development of Esophageal Varices and Efficiency of Sclerotherapy
Overall occurrence of esophageal varices was equally common after failed and successful portoenterostomy (18/28 vs 10/19, P = 0.547). Following failed portoenterostomy, esophageal varices were encountered significantly earlier (8.3 [4–23] vs 19 [4–165] months, P = 0.004) (Fig. 2A). Upper gastrointestinal bleeding occurred only after failed portoenterostomy (16/28 vs 0/19, P < 0.001) (Fig. 2B). Eradication of varices succeeded significantly less often after failed portoenterostomy (0/16 vs 6/10, P = 0.001). After variceal detection, transplant-free survival was significantly shorter among patients with failed portoenterostomy (0.2 [0.0–0.9] vs 2.4 [0.2–9.5] years, P = 0.001).
In this population-based study, 47 unselected patients with BA underwent surveillance endoscopies and primary prophylactic sclerotherapy of grade 2 and 3 varices when encountered. Previous data on unselective routine endoscopic surveillance and prophylactic sclerotherapy of esophageal varices in BA are scarce (7). Esophageal varices occurred with similar frequency after failed and successful portoenterostomy in 28 (60%) patients. All 16 patients with upper gastrointestinal bleeding had undergone a failed portoenterostomy, whereas no bleeding occurred after successful portoenterostomy. Following failed portoenterostomy, esophageal varices were encountered significantly earlier. Increased serum bilirubin concentration >40 μmol/L at 3 months after portoenterostomy was a significant risk factor of upper gastrointestinal bleeding. After successful portoenterostomy, patients who developed varices cleared their jaundice later and had more often an enlarged spleen at the first endoscopy.
Here, the patients with grade 2 or 3 varices received sclerotherapy every 2 to 4 weeks until eradication. The efficiency of this sclerotherapy protocol among the failed portoenterostomy patients was poor—despite initially successful eradication varices eventually reappeared in all. Upper gastrointestinal bleeding occurred in 72% of failed portoenterostomy patients with varices, and in 25% of failed portoenterostomy patients who received primary prophylaxis sclerotherapy. These findings underscore the importance of continuing surveillance after variceal eradication until liver transplantation following failed portoenterostomy. Duché et al (9) reported 13 patients with BA with a mean age of 13 months of whom all but 1 had serum bilirubin >30 μmol/L, and all had large varices. Their patients received sclerotherapy weekly for 2 to 5 weeks, and eradication was successful in 8 of 13 patients, and only 1 child bled. Accordingly, an aggressive approach for prophylactic sclerotherapy seems indicated after failed portoenterostomy.
Earlier studies have reported 17% to 41% frequency of variceal hemorrhage in the subgroup of patients who developed varices after an initially successful portoenterostomy (11,12). In our material no upper gastrointestinal bleeding occurred in patients who underwent a successful portoenterostomy, suggesting a benefit of prophylactic sclerotherapy in this subset of patients with BA. Whether bleeding would have occurred without primary prophylactic sclerotherapy remains unknown.
Some earlier studies suggest that sclerotherapy treated patients are at risk for bleeding from treatment-induced portal hypertensive gastropathy (10,21). In our study, no significant hypertensive gastropathy was encountered, and no patient experienced bleeding from gastric varices; however, in 3 patients the upper gastrointestinal bleeding site remained unidentified, leaving the possibility that they had developed varices in the Roux loop of the jejunum, which is not reached by a gastroscope. Despite similar sclerotherapy protocol, variceal hemorrhage was significantly more frequent after a failed portoenterostomy than in patients who developed varices several years after an initially successful portoenterostomy. This is not unexpected in the light that after failed portoenterostomy children display rapidly progressive cholestatic liver failure and cirrhosis, whereas gradually ensuing live fibrosis portrays patients who eventually develop cirrhosis after successful portoenterostomy (22). Here, transplant-free survival was similar among patients with normal serum bilirubin level with or without varices. In accordance with previous studies, these data suggest that appearance of esophageal varices in patients with BA with stable liver function does not predict soon approaching need for liver transplantation (3,9,13,23–25).
This study had several limitations. In Finland, BA treatment was centralized in 2005 from 5 centers to Helsinki. The centralization resulted in improvement in portoenterostomy success rate, being now 75%, and thus the patients with a successful portoenterostomy are concentrated in the recent years (26). Owing to the retrospective nature of the data drawn from endoscopy reports, we were unable to reliably analyze other important endoscopic risk factors of bleeding such as red spots associated with varices (16).
During this surveillance and sclerotherapy protocol, no patient with a successful portoenterostomy bled from varices. Variceal hemorrhage only occurred after failed portoenterostomy, the first bleeding episode occurred at a median age of 8.5 months, and 6 patients experienced their first bleeding episode before the first surveillance endoscopy, the earliest bleeding occurring at the age of 5 months. In a study by Gana et al (8) platelet count and spleen size assessment predicted varices in children. Our study included 3 patients with normal platelets and no spleen enlargement observed (which may be doubted owing to the retrospective nature of our data) before a bleeding from varices—these 3 had a failed portoenterostomy. On the contrary, 5 patients with a successful portoenterostomy had a normal-sized spleen, normal platelets, and normal serum bilirubin level, and presented no varices during the surveillance. A proportion of patients with BA, in our material 4 (9%), have splenic anomalies such as poly- or asplenia that make spleen size assessment impossible.
In the future, surveillance should be planned separately for patients with failed or successful portoenterostomy. After failed portoenterostomy, endoscopic surveillance and prophylactic sclerotherapy of esophageal varices may be started by the age of 6 months. The encountered large varices should receive frequently repeated sclerotherapy sessions until eradication (16). After successful portoenterostomy, the surveillance may be initiated when serum bilirubin concentration exceeds 40 μmol/L, or clinical signs of advanced portal hypertension, such as splenomegaly or thrombocytopenia develop, before gastrointestinal bleeding occurs. The progression and prognosis of varices in patients with BA after failed and successful portoenterostomy are divergent, which should be considered also when planning future studies.
1. Hartley JL, Davenport M, Kelly DA. Biliary atresia. Lancet
2. Houwen RH, Zwierstra RP, Severijnen RS, et al. Prognosis of extrahepatic biliary atresia. Arch Dis Child
3. Lykavieris P, Chardot C, Sokhn M, et al. Outcome in adulthood of biliary atresia: a study of 63 patients who survived for over 20 years with their native liver. Hepatology
4. Ling SC, Shneider BL. Portal Hypertension
: current practice and the need for evidence. In: de Franchis R. Portal hypertension
. Proceedings of the Fifth Baveno International Consensus Workshop.
London: Wiley & Blackwell; 2011.
5. Triantos CK, Burroughs AK. Prevention of the development of varices and first portal hypertensive bleeding episode. Best Pract Res Clin Gastroenterol
6. Molleston JP. Variceal bleeding in children
. J Pediatr Gastroenterol Nutr
7. 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
8. Gana JC, Turner D, Mieli-Vergani G, et al. A clinical prediction rule and platelet count predict esophageal varices in children
9. Duche M, Habes D, Roulleau P, et al. Prophylactic endoscopic sclerotherapy
of large esophagogastric varices in infants with biliary atresia. Gastrointest Endosc
10. Goncalves ME, Cardoso SR, Maksoud JG. Prophylactic sclerotherapy
with esophageal varices: long-term results of a controlled prospective randomized trial. J Pediatr Surg
11. Stringer MD, Howard ER, Mowat AP. Endoscopic sclerotherapy
in the management of esophageal varices in 61 children
with biliary atresia. J Pediatr Surg
12. Lilly JR, Stellin G. Variceal hemorrhage in biliary atresia. J Pediatr Surg
13. Sarin SK, Misra SP, Singal AK, et al. Endoscopic sclerotherapy
for varices in children
. J Pediatr Gastroenterol Nutr
14. Celinska-Cedro D, Teisseyre M, Woynarowski M, et al. Endoscopic ligation of esophageal varices for prophylaxis of first bleeding in children
and adolescents with portal hypertension
: preliminary results of a prospective study. J Pediatr Surg
15. Sasaki T, Hasegawa T, Nakajima K, et al. Endoscopic variceal ligation in the management of gastroesophageal varices in postoperative biliary atresia. J Pediatr Surg
16. Duché M, Ducot B, Tournay E, et al. Prognostic value of endoscopy in children
with biliary atresia at risk for early development of varices and bleeding. Gastroenterology
17. de Franchis R, Pascal JP, Ancona E, et al. Definitions, methodology and therapeutic strategies in portal hypertension
: a consensus development workshop, Baveno, Lake Maggiore, Italy, April 5 and 6, 1990. J Hepatol
18. Davenport M, Puricelli V, Farrant P, et al. The outcome of the older (≥100 days) infant with biliary atresia. J Pediatr Surg
19. Serinet MO, Broue P, Jacquemin E, et al. Management of patients with biliary atresia in France: results of a decentralized policy 1986–2002. Hepatology
20. Wildhaber BE, Majno P, Mayr J, et al. Biliary atresia: Swiss national study, 1994–2004. J Pediatr Gastroenterol Nutr
21. Sasaki T, Hasegawa T, Shimizu Y, et al. Portal hypertensive gastropathy after surgery for biliary atresia. Surg Today
22. Haafiz AB. Liver fibrosis in biliary atresia. Expert Rev Gastroenterol Hepatol
23. Miga D, Sokol RJ, MacKenzie T, et al. Survival after first esophageal variceal hemorrhage in patients with biliary atresia. J Pediatr
24. van Heurn LWE, Saing H, Tam PKH. Portoenterostomy for biliary atresia: long-term survival and prognosis after esophageal variceal bleeding. J Pediatr Surg
25. Ohi R, Mochizuki I, Komatsu K, et al. Portal hypertension
after successful hepatic portoenterostomy in biliary atresia. J Pediatr Surg
26. Lampela H, Ritvanen A, Kosola S, et al. National centralization of biliary atresia care to an assigned multidisciplinary team provides high-quality outcomes. Scand J Gastroenterol