Extrahepatic portal vein obstruction (EHPVO) is the most common cause of long-standing portal hypertension (PH) in children. In paediatric patients, this condition occurs shortly after birth; therefore, in noncirrhotic subjects, EHPVO is complicated by bleeding episodes and hypersplenism, but does not progress to end-stage liver disease (1,2). These subjects have no indication to liver transplantation and thus they represent the main paediatric population requiring long-term care of PH (3–6).
Unlike in adults, there is little evidence on efficacy of nonselective β-blockers, endoscopic variceal obliteration (EVO), realization of portosystemic shunts in children with PH, and even less in noncirrhotic PH (7,8). Despite the relatively recent introduction of a new surgical tool to cure EHPVO, the mesoportal bypass (MPB), there is no wide agreement on feasibility, indications, timing, and success of such procedure in children with portal vein thrombosis (9–14).
Our aim is to present the results of a stepwise approach we have used in this setting, based on severity of PH, response to medical and endoscopic management, indications to surgery, favouring the surgical choice of MPB when feasible.
We reviewed the notes of all consecutive children with a diagnosis of chronic noncirrhotic EHPVO managed at our institution and with a minimum follow-up period of 1 year. Our institutional review board approved data collection and notes evaluation that have been anonymised for this report.
EHPVO was diagnosed in patients with clinical features of noncirrhotic, prehepatic PH with the evidence of obstruction of the extrahepatic portal vein demonstrated at Doppler ultrasound and/or computerised tomographic angiogram (15). Malignancies or inflammation were ruled out. Liver biopsy was done in selected patients to confirm the absence of cirrhosis.
Large varices are varices scored as grade 3, or with red wales (irrespective of the size), or presence of gastric varices. Uncontrolled varices are as all the events in which there was an episode of oesophageal bleeding or an endoscopy demonstrating large varices.
Prophylaxis with NSBBs is primary when started upon endoscopic detection of large varices, and secondary if started after endoscopic treatment of varices or after a bleeding episode (16).
Successful medical management is when there were no bleeding episodes, no large varices at follow-up endoscopy, and no severe hypersplenism for at least 12 months.
Successful surgical management is when resolution of PH was achieved and the bypasses or shunts were maintained patent at the last Doppler ultrasound evaluation and for a minimum of 6 months after the intervention.
All referred children were approached according to a step-by-step program (Fig. 1):
- Workup at diagnosis: blood tests (full blood cell count, liver and renal function, coagulation, blood ammonia), imaging (Doppler ultrasound, contrast-enhanced computerised tomography if indicated), oesophagogastroduodenoscopy.
- Follow-up visits every 4, 6, or 12 months according to the clinical status, the severity of PH, and the course of the disease: clinical, laboratory, and endoscopic evaluations.
- Patients who developed large varices were either started on NSBBs or received EVO. Those who bled underwent EVO before starting NSBBs.
- Patients taking NSBBs who continued to have large varices at follow-up endoscopy or experienced gastrointestinal bleeding underwent EVO.
- A retrograde portogram and a contrast-enhanced computerised tomography angiogram were carried out in patients with relapsing bleeding or large varices following endoscopic eradication.
- Among these, patients with a patent Rex recessus underwent an MPB, whereas the others underwent shunt surgery or transjugular intrahepatic portosystemic shunt (TIPS) if feasible. Shunting was considered also in children with severe hypersplenism (platelet count <30 × 10–9/L, white blood cell count <2.0 × 10−9 cells/L, or transfusion-dependent anaemia).
The endpoint of this study was to test the applicability and the efficacy of such protocol to control major complications of PH.
Endoscopic procedures were carried out using a forward-viewing flexible fiberoptic endoscope (Olympus GIF-160, GIF-XP160, Olympus, Tokyo, Japan) under general anaesthesia with endotracheal intubation in case of endoscopic treatment, and under deep sedation in case of diagnostic procedures.
Eradication involved sequential sessions of EVO every 15 to 20 days until varices disappearance was achieved; sclerotherapy was carried out in younger children (<2 years or <10 kg) (17–19), whereas banding ligation was the preferred method adopted in all other patients (20). Sclerotherapy and ligation at our centre were performed by 3 experienced endoscopists.
We exclusively used propranolol that was started at a dose of 0.5 mg · kg−1 · day−1 and increased to a maximum of 1.5 mg · kg−1 · day−1 if tolerated to achieve a 20% to 30% drop in heart rate from baseline evaluation (21).
The workup and the operative technique to perform the MPB have been described previously (9). Briefly, at laparotomy, the umbilical vein was recanalised and a catheter was inserted for direct portography. When necessary, a band of liver parenchyma bridging the umbilical scissura was resected; the intrahepatic left portal vein was then approached and its ventral and lateral aspects were dissected. A venous autograft (usually the left internal jugular vein) was anastomosed end-to-side to the ventral portion of the intrahepatic left portal vein. The graft was then positioned behind the pylorus and across the mesocolon to reach the superior mesenteric vein to which it was anastomosed end-to-side. The techniques used to perform surgical portosystemic shunts and TIPS have been described elsewhere (14,15,22–26).
Data are given as medians and ranges. The values of individual continuous variables were compared for significance using Student t test for paired samples. A P value of ≤0.05 was assigned significance. Multivariate analysis was performed to disclose the relation between the variables and the studied effect. The statistical analysis was carried out with SPSS 13.0 for Windows (SPSS Inc, Chicago, IL).
In the last 15 years, 65 patients (40 boys, 25 girls) with EHPVO were managed at our institution. Thirty-three (50%) were born preterm, median gestational age 35 weeks (range 24–41), median birth weight 2100 g (870–4100); in 36 (55%) patients, an umbilical venous catheter had been inserted. Eight patients had an uneventful postnatal period and apparently no risk factors for the development of EHPVO.
At the time of this review, 6 patients have been lost to follow-up (data are updated at the last visit), 1 received a liver transplant following the development of hepatopulmonary syndrome, and 2 patients died from causes unrelated to PH (interstitial pneumonia in immunodeficiency and sepsis respectively). The clinical features of the studied group are reported in Table 1.
Among the 65 patients studied, 32 (49%) presented with bleeding, whereas 33 (51%) were diagnosed because of clinical signs of PH (Fig. 2). Patients presenting with bleeding had a median age at onset of 3.8 years (0.5–15.5), and 25 of 32 received immediate endoscopic treatment, whereas 5 had first EVO shortly after the onset. Overall, 30 of 32 patients presenting with bleeding had endoscopic treatment as primary management of uncontrolled varices; 1 patient was started on NSBBs without endoscopic treatment, and 1 went directly to surgery because of mandibular ankylosis, making endoscopy unfeasible. During the follow-up, 24 of 30 patients who had endoscopic treatment were started on secondary prophylaxis with NSBBs, whereas 1 of 30 was considered under control with no further treatment and 5 of 30 were sent directly to surgery (because of refractory varices in 3, portal hypertensive biliopathy in 1, electively in 1). Among the 25 started on NSBBs, 13 (52%) had a new episode of bleeding or developed large varices and therefore had endoscopic treatment, 2 went to surgery (because of portal hypertensive biliopathy and hepatopulmonary syndrome), whereas 10 (40%) were controlled. Among the 13 patients having a second round of endoscopic treatment, 1 was controlled, whereas 12 failed and required a surgical procedure because of relapsing bleeding or large varices (MPB in 5, proximal splenorenal shunt in 3, distal splenorenal shunt in 2, mesocaval shunt in 1; 1 was managed by TIPS) at a median age of 7.8 years (3.8–13.5). Overall, 12 of 32 (37%) patients presenting with bleeding were controlled without surgery or TIPS.
Thirty-three patients (51%) were diagnosed before an episode of bleeding at a median age of 3 years (0.2–17.5); 8 had large varices at first endoscopy and all received primary endoscopic obliteration as the first treatment, of whom 1 went straight to surgery because the condition failed to be eradicated. Seven of 8 patients were subsequently started on secondary prophylaxis with NSBBs. Of these 7, 3 were controlled, whereas 4 went to surgery (because of relapsing varices in 3 and severe hypersplenism in 1) after a second endoscopic treatment. In the group diagnosed before a bleeding episode, 25 of 33 received initially no treatment. Among these, 5 of 25 remained well controlled (but they have a median follow-up of only 1.7 years), 13 were eventually started on primary NSBBs, whereas 7 underwent primary endoscopic treatment (3 after a bleed) before starting secondary NSBBs. Of the 20 who went on NSBBs (13 as primary and 7 as secondary prophylaxis), 9 were well controlled, whereas 11 required further endoscopic treatment; 5 of these 11 were well controlled, whereas 6 required surgery (because of relapsing varices in 2, refractory ascites in 2, electively in 2). Overall, 22 of 33 (67%) patients with no bleeding at presentation were controlled without surgery; however, 3 of these were sent to surgery electively.
During the follow-up, 43 patients (66%) had at least 1 episode of gastrointestinal bleeding, of which 32 occurred at presentation (median age 3.8, range 0.5–15.5 years) and 11 at a median age of 7 years (2.7–17.5) and at a follow-up time of 2.5 years (1–11); 23 patients (35%) had >1 bleeding episode (up to 6 episodes). On the contrary, none died from bleeding.
A multivariate analysis showed that bleeding at diagnosis was not associated with a history of prematurity, low birth weight, placement of umbilical catheter, neonatal illness, and age at diagnosis. Patients bleeding at diagnosis, although, were more likely to require surgery or TIPS (P = 0.02, 2-tailed Fisher exact test) during a similar follow-up time (bleeders vs nonbleeders medians of 8.85 vs 7.0 years, t test; P = 0.5).
Fifty-five of 65 (85%) patients had uncontrolled varices and required treatment at a median age of 4 years (0.3–21), and 40 of 55 at presentation. The number of patients who developed uncontrolled varices was 53 of 65 (81%) and 54 of 65 (83%) by 5 and 10 years of follow-up, respectively (Fig. 3A). Ten patients had controlled varices either with no treatment or only with NSBBs, but they had a much shorter follow-up period as compared with those with uncontrolled varices (median of 2.3 vs 9.2 years, t test; P = 0.01), leading one to think that varices would progress eventually.
Patients developing uncontrolled varices after the presenting episode, and namely after 6 months from presentation (conditional survival), were 25 of 65 (38%) and 31 of 65 (48%) by 5 and 10 years of follow-up, respectively (Fig. 3B).
Use of NSBBs
Prophylaxis with NSBBs was prescribed to 52 patients (80%) at a median age of 5 years (0.8–17.8) and at a median time from presentation of 0.5 years (0.3–11.5). Among those who received NSBBs, 30 (58%) failed to control varices, 8 of 13 in the group of primary prophylaxis and 22 of 39 in the group of secondary prophylaxis; however, the dose was adjusted according to weight gain only in 21 of 52 (40%) patients, whereas the remainders outgrew the dose of the drug. Evidence of reached target heart rate decrease was found in the notes of only 8 of 52 patients (15%). Therefore, NSBBs failed if they were used subtherapeutically.
Overall, a total of 254 procedures (172 sclerotherapies, 82 banding ligations) were carried out, and all of the patients underwent at least 1 endoscopy at a median age of 3.9 years (0.3–17.7). Of 62 patients from whom data of first endoscopy are available, 42 (68%) already had large varices at diagnosis and 38 of 42 (90%) received a first endoscopic eradication at a median age of 4.5 years (1.1–21). Overall, 55 patients (85%) required EVO, of whom 22 (40%) were well controlled, whereas 33 (60%) eventually required surgery (Fig. 2).
Evaluation of the Rex Recessus
Retrograde portogram was performed in 54 of 65 patients (83%) at a median age of 8 years (0.6–19); the Rex recessus was patent in 24 (44%), whereas in 30 (56%), it was not suitable for MPB (obstructed in 25, hypoplastic in 5). The median age of patients with a patent Rex (7.5 years, range 0.6–19) did not differ from that of patients with an obstructed recessus (8.5, 2–13.5) (t test; P = 0.9). There was an association between obstructed Rex recessus and history of umbilical catheter placement (Pearson correlation; P = 0.01). Among 42 patients who had an umbilical vein cannulation at birth, 35 underwent retrograde portogram and 15 (42%) had a patent Rex recessus.
Thirty-four patients (53%) were considered for interventions because of relapsing uncontrolled varices (n = 21, 62%), huge splenomegaly with severe hypersplenism (n = 8, 32%), portal hypertensive biliopathy (n = 2, 6%); 3 patients were sent to MPB electively. Thirteen of 34 (38%) patients underwent a proximal splenorenal shunt, 13 of 34 (38%) a MPB, 3 a mesocaval shunt, 2 a TIPS, 2 a distal splenorenal shunt, 1 a liver transplantation because of hepatopulmonary syndrome (Fig. 4). Eleven of 13 patients (85%) who underwent an MPB have a patent conduit after a follow-up of 2.4 years (0.6–9). Ten of 13 proximal splenorenal shunts (77%) are patent after a follow-up of 2.5 years (0.6–15.8). The 2 patients who underwent a distal splenorenal shunt and the 2 who had TIPS have a patent conduit at their last follow-up visit. Among the 3 patients who had a mesocaval shunt, 2 have a patent conduit at the last follow-up visit. The patient who received a transplant because of hepatopulmonary syndrome is well with a functional graft.
Overall, the management of patients requiring surgical or radiological interventions was successful in 28 of 34 (82%), with a high rate of patency of shunts and bypass (27/33, 82%). Patients who had a TIPS placed presented with a mild increase in blood ammonia without clinical encephalopathy. Among those who underwent a successful surgical shunting procedure, 1 developed episodes of hepatic encephalopathy that were well controlled by medical management.
Adherence to the Protocol
Comparing the algorithm adopted in our centre (Fig. 1) with the results studied retrospectively (Fig. 2), we found that the adherence to the protocol was good. Among 32 patients who bled at presentation, 25 were started on NSBBs after EVO. One patient was controlled by endoscopic treatment; 6 patients were sent directly to surgery, 3 because of severe refractory bleeding, 1 because of mandibular ankylosis, making endoscopy difficult, whereas 2 were sent electively to MPB despite well-controlled varices.
In the 33 patients who were diagnosed before an episode of bleeding, NSBBs or endoscopic treatment were adopted without preference. Eight patients had large varices at presentation and underwent endoscopic eradication in the first instance; among the 20 who developed uncontrolled varices late in the study, 13 were prescribed NSBBs and 7 were sent to endoscopic eradication (Fig. 5).
Three patients with a patent Rex recessus were sent to MPB electively, despite good control of varices. Retrograde portogram was carried out in many more patients than expected, allowing us to define the prevalence of patent Rex recessus in this population.
After a median follow-up from diagnosis of 8.4 years (1–16), 2 patients died from causes unrelated to PH, and 6 were lost to follow-up. Of the remaining 57 patients, after a median follow-up after the last intervention of 3.1 years (0.6–15.8), 53 (93%) achieved or maintained bleeding control. On the contrary, 13 (25%) still carry some complications of PH: hypersplenism (4), requirement of further endoscopic treatment (3), high ammonia (2), portal hypertensive biliopathy (2), and nodular regenerative hyperplasia (2).
In children, prehepatic PH has become more prevalent than PH of other aetiologies. In fact, the severity of disease progression of many paediatric cirrhotic disorders (eg, biliary atresia) along with the relatively larger availability of liver segments for paediatric transplantation have led to a rapid decrease in time spent by these patients on the care for end-stage liver disease (2). Conversely, most children with prehepatic PH require lifelong care. The prompt availability of liver transplantation for children with cirrhotic PH has limited the possibilities of testing the efficacy of NSBBs and EVO to prevent bleeding. These reasons, along with the large sample size required and the difficulty in recruiting children in a multicentre study, render paediatric trials unfeasible and robust data unavailable (27,28).
Our study showed that EHPVO is strictly related to neonatal illnesses and occurs at birth in the majority of cases. In more than half of these patients, an umbilical venous catheter had been placed (29). In our cohort, there was a male preponderance, which, we believe, is coincidental. Patients presenting with bleeding are more likely to require surgery or TIPS by 10 years of follow-up. Overall, 85% of our patients had uncontrolled varices, either at presentation or developed thereafter, of whom 81% within 5 years, suggesting a rapid progression of PH.
Variceal bleeding in adults with cirrhosis is a life-threatening condition with high mortality rate; this study confirms that children are more resistant to bleeding episodes, especially in patients with noncirrhotic PH. Indeed in this series, bleeding episodes were not associated with risk of death; therefore, PH in this setting carries a high morbidity but a low mortality rate.
In our patients, treatment with NSBBs failed in approximately 60% of subjects, but many outgrew the starting dosage and in the majority of them, the target heart rate reduction was not met. Hence, from our data, we cannot gather reliable information on the efficacy of propranolol in preventing bleeding or rebleeding in this setting.
A few studies on endoscopic treatment as primary prophylaxis of gastrointestinal bleeding in children have been published, showing good tolerance and low rate of bleeding recurrence (30,31). In our cohort, nearly 68% of the patients had varices at diagnosis. During the follow-up, endoscopic obliteration was required in 55 patients (85%) and allowed to control varices in 40% of them.
In our centre, a retrograde portogram was carried out in almost all of the patients, even before nonsurgical management failed, giving us the chance to know the rate of its patency. Interestingly, only 54% of those tested had a patent Rex recessus and were therefore suitable for MPB; given the anatomy of the portal circulation in the newborn, it is not surprising that such obstruction was associated with the insertion of an umbilical vein catheter after birth (32–34). Although one may hypothesise that a portogram carried out at a younger age would show a higher rate of patency, in our patients, this was not found. Indeed, it has been suggested that portal vein thrombosis does not progress to the intrahepatic portal branches (35). Whether the chronic reduction of the portal flow causing a relative liver atrophy may lead to hypoplasia of the Rex recessus has not been clearly demonstrated.
In approximately half of our patients, the complications of PH could not be controlled by NSBBs and variceal obliteration. There is no concordance among the reports on the natural history of EHPVO managed conservatively, with some studies suggesting a tendency to a reduced incidence of bleeding episodes across adolescence and others reporting the development of severe complications of PH even after 20 years of follow-up (36–38). Therefore, one could wonder whether we should be aggressive with these patients and carry out an operation in those unresponsive to variceal obliteration or simply manage the episodes of bleeding. The answer is obvious in children having a patent Rex recessus who, after a successful MPB, definitely gain a large advantage through the resolution of PH and complete restoration of the physiological portal blood flow to the liver. A different solution is offered by shunting procedures, carrying probably higher risks of surgical complications as well as the risk of subtle hepatic encephalopathy. The observation that the frequency of variceal bleeding is reduced after puberty because opening spontaneous shunts has not been confirmed in long-term studies. In children, the bleeding risk factor (number of bleeds per year) is approximately 1.3 and remains unchanged after puberty (39). Continuous episodes of bleeding set relevant limitations to the patient and the family, and the perspective of awaiting such frightening events make quality of life obviously poor. Hospitalisation and management of bleeding carry remarkable procedural costs. A study carried out in patients of a median of 23 years of age affected by EHPVO showed that portal hypertensive biliopathy can be demonstrated in >30% of cases (37), but as many as 80% have biliary abnormalities at ERCP (40). A further reason in favour of bypass surgery is the fact that EHPVO occurring in the prepubertal period often results in growth retardation, possibly because of the reduced portal blood supply to the liver with resistance to growth hormone function (14). A spurt in growth has been observed in children with EHPVO both after MPB and after shunt surgery (38,41). In our view, all of these reasons justify the choice of a surgical decompression of the portal system in children unresponsive to medical and endoscopic treatment. This conclusion is also supported by the high rate of success of such procedures in our cohort of patients, one of the highest reported so far and by the success of this stepwise approach that warranted complete bleeding control in 93% of the patients.
In reviewing the data of our patients, we saw that the adherence to the initial protocol, as shown in Figure 1, was good. This increases the reliability of our study that can be considered the retrospective review of a prospective evaluation. In our protocol, NSBBs and endoscopic obliteration were alternative first-line options to treat large varices in patients who did not bleed at presentation; however, most of the times endoscopic obliteration was preferred and NSBBs added later. This is probably because of the fact that we always perform endoscopies in a setting allowing variceal treatment, which is therefore carried out as soon as large varices are detected, skipping the step of a diagnostic procedure.
A violation to the protocol is represented by the fact that 3 patients were sent directly to the MPB despite good control of bleeding and no further complications. These operations have been carried out recently, reflecting a slight change of the use of this procedure in our centre, which is becoming a good option also as a preemptive technique, when feasible.
In conclusion, we have reported our successful experience with a stepwise approach to paediatric EHPVO comprising medical, endoscopic, and surgical options, providing optimal survival and low morbidity to a large cohort of children affected by this condition. Most of these children had an umbilical venous catheter inserted because of a neonatal illness, and this was associated with Rex recessus obliteration, making MBP unfeasible. Almost all of the patients had severe PH and several episodes of variceal bleeding early in life that were only partially controlled by non-surgical means. Half of all patients eventually required decompression of the portal system by surgery or TIPS, especially those presenting with bleeding episode. Our stepwise approach warranted bleeding control in >90% of cases and we therefore consider it a successful protocol for children with EHPVO.
1. Alonso EM, Hackworth C, Whitington PF. Portal hypertension
. Clin Liver Dis
2. D’Antiga L. Medical management of esophageal varices and portal hypertension
. Semin Pediatr Surg
3. Sarin SK, Agarwal SR. Extrahepatic portal vein obstruction. Semin Liver Dis
4. Alvarez F, Bernard O, Brunelle F, et al. Portal obstruction in children
. I. Clinical investigation and hemorrhage risk. J Pediatr
5. Thompson EN, Williams R, Sherlock S. Liver function in extrahepatic portal hypertension
6. Webb LJ, Sherlock S. The aetiology, presentation and natural history of extra-hepatic portal venous obstruction. Q J Med
7. Shneider BL, Bosch J, de Franchis R, et al. Portal hypertension
: expert pediatric opinion on the Report of the Baveno V Consensus Workshop on Methodology of Diagnosis and Therapy in Portal Hypertension
. Pediatr Transplant
8. de Franchis R. Revising consensus in portal hypertension
: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension
. J Hepatol
9. de Ville de Goyet J, Alberti D, Clapuyt P, et al. Direct bypassing of extrahepatic portal venous obstruction in children
: a new technique for combined hepatic portal revascularization and treatment of extrahepatic portal hypertension
. J Pediatr Surg
10. Bambini DA, Superina R, Almond PS, et al. Experience with the Rex shunt (mesenterico-left portal bypass) in children
with extrahepatic portal hypertension
. J Pediatr Surg
11. Dasgupta R, Roberts E, Superina RA, et al. Effectiveness of Rex shunt in the treatment of portal hypertension
. J Pediatr Surg
12. Stringer MD. Improved body mass index after mesenterico-portal bypass. Pediatr Surg Int
13. Mack CL, Superina RA, Whitington PF. Surgical restoration of portal flow corrects procoagulant and anticoagulant deficiencies associated with extrahepatic portal vein thrombosis. J Pediatr
14. Wolff M, Hirner A. Current state of portosystemic shunt surgery
. Langenbecks Arch Surg
15. Belli L, Puttini M, Marni A. Extrahepatic portal obstruction: clinical experience and surgical treatment in 105 patients. J Cardiovasc Surg (Torino)
16. Bosch J, García-Pagán JC. Prevention of variceal rebleeding. Lancet
17. Dilawari JB, Chawla YK, Ramesh GN, et al. Endoscopic sclerotherapy in children
. J Gastroenterol Hepatol
18. Howard ER, Stringer MD, Mowat AP. Assessment of injection sclerotherapy in the management of 152 children
with oesophageal varices. Br J Surg
19. Stringer MD, Howard ER. Longterm outcome after injection sclerotherapy for oesophageal varices in children
with extrahepatic portal hypertension
20. McKiernan PJ, Beath SV, Davison SM. A prospective study of endoscopic esophageal variceal ligation using a multiband ligator. J Pediatr Gastroenterol Nutr
21. Shashidhar H, Langhans N, Grand RJ. Propranolol in prevention of portal hypertensive hemorrhage in children
: a pilot study. J Pediatr Gastroenterol Nutr
22. Botha JF, Campos BD, Grant WJ, et al. Portosystemic shunts in children
: a 15-year experience. J Am Coll Surg
23. Orloff MJ, Orloff MS, Girard B, et al. Bleeding esophagogastric varices from extrahepatic portal hypertension
: 40 years’ experience with portal-systemic shunt. J Am Coll Surg
24. Orozco H, Takahashi T, Mercado MA, et al. Surgical management of extrahepatic portal hypertension
and variceal bleeding
. World J Surg
25. Senzolo M, Tibbals J, Cholongitas E, et al. Transjugular intrahepatic portosystemic shunt
for portal vein thrombosis with and without cavernous transformation. Aliment Pharmacol Ther
26. Di Giorgio A, Agazzi R, Alberti D, et al. Feasibility and efficacy of transjugular intrahepatic portosystemic shunt
(TIPS) in children
. J Pediatr Gastroenterol Nutr
27. 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
28. Ozsoylu S, Kocak N, Demir H, et al. Propranolol for primary and secondary prophylaxis of variceal bleeding
with cirrhosis. Turk J Pediatr
29. Morag I, Epelman M, Daneman A, et al. Portal vein thrombosis in the neonate: risk factors, course, and outcome. J Pediatr
30. 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
31. Zargar SA, Javid G, Khan BA, et al. Endoscopic ligation compared with sclerotherapy for bleeding esophageal varices in children
with extrahepatic portal venous obstruction. Hepatology
32. Lawson AJ, Rischbieter P, Numanoglu A, et al. Imaging the Rex vein preoperatively using wedged hepatic venous portography. Pediatr Radiol
33. Kwan SW, Fidelman N, Durack JC, et al. Rex shunt preoperative imaging: diagnostic capability of imaging modalities. PLoS One
34. Gibelli NE, Tannuri AC, Pinho-Apezzato ML, et al. Extrahepatic portal vein thrombosis after umbilical catheterization: is it a good choice for Rex shunt? J Pediatr Surg
35. Ogawa A, Takamori S, Kojima K, et al. Hemodynamics in extrahepatic portal vein obstruction and its changes during long-term follow-ups. Hepatol Res
36. Rangari M, Gupta R, Jain M, et al. Hepatic dysfunction in patients with extrahepatic portal venous obstruction. Liver Int
37. Chattopadhyay S, Govindasamy M, Singla P, et al. Portal biliopathy in patients with non-cirrhotic portal hypertension
: does the type of surgery affect outcome? HPB (Oxford)
38. Sarin SK, Bansal A, Sasan S, et al. Portal-vein obstruction in children
leads to growth retardation. Hepatology
39. Sarin SK, Sollano JD, Chawla YK, et al. Consensus on extra-hepatic portal vein obstruction. Liver Int
40. Dilawari JB, Chawla YK. Pseudosclerosing cholangitis in extrahepatic portal venous obstruction. Gut
41. Lautz TB, Sundaram SS, Whitington PF. et al RA. Growth impairment in children
with extrahepatic portal vein obstruction is improved by mesenterico-left portal vein bypass. J Pediatr Surg