• •

  It is defined as an extrahepatic occlusion of the portal vein, with or without involvement of the intrahepatic portal veins, splenic vein or superior mesenteric vein.

• •

  It is frequently characterized by the presence of a portal cavernoma.

• •

  Isolated occlusion of the splenic vein or superior mesenteric vein do not constitute EHPVO and may require distinct management approaches.

• •

  Portal vein obstruction associated with chronic liver disease or neoplasia does not constitute EHPVO.7,8

Level of evidence 1, grade of recommendation A.

According to WHO, EHPVO fulfils the criteria for rare diseases, with a prevalence of < 5 per 10,000 inhabitants. It is the cause of PHT in more than 30% of cases of children with esophageal variceal bleeding.5,9

The frequency of EHPVO is variable and depends on the location of the obstruction to portal flow, being involved in 9% to 76% of all cases of portal hypertension in children (Table 2).10–17

Level of evidence 4, grade of recommendation C

Several causes have been associated with EHPVO.2

• •

  Direct injury to the portal vein due to omphalitis and umbilical vein catheterization, the latter is considered to be one of the major risk factors and is associated with: a delay in placement, catheterization for more than 3 days, misplacement, trauma and type of solution used.

• •

  Portal vein abnormalities such as stenosis, atresia, and agenesis often are associated with congenital shunting and do not fit into the typical paradigm of EHPVO.

• •

  Indirect factors: neonatal sepsis, dehydration, multiple exchange transfusions and prothrombotic states.

• •

  Idiopathic, with no identifiable etiology.

Causes of EHPVO, according to reports in the literature, are shown in table 3. Neonatal events such as omphalitis, umbilical vein catheterization or sepsis are present in 6% to 49% of cases; in those in which abdominal trauma or sepsis are present, the frequency varies between 3% and 22%.18–26

Level of evidence 4, grade of recommendation C

Prothrombotic states associated with EHPVO have been little studied and the results vary (Table 4).

There is evidence that deficiencies of coagulation inhibitors (protein C, protein S, antithrombin III), prothrombin (PTHR) gene G20210A mutation, methylene tetrahydrofolate reductase (MTHFR) gene C677T mutation and factor V Leiden (FVL) mutation may predispose to portal vein thrombosis. In 5 studies of children and adolescents thrombophilia was found in a total of 42 of 134 (31%) cases.23,25,27–29 Protein C (PC) deficiency appears to be the most frequent followed by deficiencies in protein S (PS) and antithrombin III (ATIII); approximately 20% of cases had combined deficiencies. A genetic origin has not yet been proven; the studies reported involved a small number of patients or were isolated case reports. In a prospective randomized trial in 17 children with EHPVO, seven had low levels of PC, PS and ATIII; none of the 24 controls nor any of the 25 parents of the children studied were found to have deficiencies in these factors and as for genetic mutations, there was no significant difference in the frequency of MTHFR between cases and controls; only a few isolated cases of mutant PTHR were seen.29

Level of evidence 3, grade of recommendation C

EHPVO is between 35% to 78% of cases idiopathic18–25 so that a greater number of studies seeking to identify the etiology are required, especially those aimed at detecting prothrombotic states.

Level of evidence 4, Grade of recommendation D

• 1.

  Liver ultrasound: allows assessment of liver size and echogenicity, spleen size and presence of cavernous degeneration of the portal vein (vessels or collateral veins around the thrombosed portal vein appear to become recanalized). In children with history of PVT in the neonatal period, it has been shown that abdominal ultrasounds is the most sensitive method for detecting progression of portal hypertension caused by PVT. It is therefore advisable to perform periodic abdominal ultrasound examinations in patients with a history of neonatal PVT.31,34,38,41 Ultrasound is also recommended during follow up in order to detect portal biliopathy (irregular diltation of the biliary tree) and associated colelithiasis and cholecystitis.2

• 2.

  Doppler ultrasound: used to assess patency of the portal and the splenic vein, the hemodynamic status of the portal system, distinguish the portal vein from the inferior vena cava and the hepatic artery and determine the direction of blood flow: hepatopetal or hepatofugal.42

• 3.

  Splenoportography: allows evaluation of the portal venous system, extent and location of the obstruction and presence of collateral vessels; its invasive nature limits its use.5

• 4.

  Other techniques: computed tomography and MRI angiography should be the first option to assess the anatomy of the portal system and it have been recommended in cases where surgery is required.5,7

Level of evidence 3, grade of recommendation C.

• 5.

  Measurement of portal pressure/hepatic venous pressure gradient. Assesses type and severity of portal hypertension; it is an invasive technique and published clinical experience reports discuss its use in children with intrahepatic portal hypertension.43 It is of no value in EHPVO as hepatic vein pressure gradient is normal in this condition and will not detect the portal hypertension otherwise it is associated with intrahepatic disease.

Level of evidence 4, grade of recommendation D.

Abdominal Doppler ultrasound, because it evaluates portal blood flow velocity and arterio-portal ratio, is a non-invasive technique that has been reported in the evaluation of EHPVO in children, with a sensitivity of 91% and a specificity of 100%.39,40,42,43

Level of evidence 3, grade of recommendation C.

Soehendra’s classification of esophageal varices is the one most widely used in endoscopic practice45 (Table 5) although at the most recent workshops on methodology of diagnosis in Baveno in 2005 and 20108,46 it was recommended that varices be defined according to size (Table 6) as:

• a)

  Small varices. Minimally elevated veins above the esophageal mucosa surface.

• b)

  Medium varices. Tortuous veins occupying less than one third of the esophageal surface, and

• c)

  Large varices. Those occupying more than one third of the esophageal surface.

Recommendations for medium-sized varices are the same as for large varices because this is how they were grouped in prophylactic trials47 as well as in a meta-analysis of randomized controlled clinical trials.44,45,48,49

Other classification used in children to evaluate the severity of the esophageal varices by endoscopic findings had been reported as:

• •

  Grade I. When varices are flattened by insufflations.

• •

  Grade II. When varices are not flattened by insufflations but are separated by healthy mucosa.

• •

  Grade III. When varices are not flattened by insufflations and had mucosa red signs.50

Level of evidence 5, grade of recommendation D.

Risk factors for gastric variceal hemorrhage include the size of fundal varices which are classified by size into large (> 10 mm), médium-sized (5–10 mm) and small (< 5 mm), respectively.45,51

According to Sarin, gastric varices are commonly classified based on the presence or absence of esophageal varices and their location in the stomach (fundus or lesser curvature)52 (Table 7). They are divided into GOV-1, with esophageal varices extending along the lesser curvature; GOV-2, with esophageal varices located in the gastric fundus; IGV-1 located in the fundus, without presence of esophageal varices, and, IGV-2, with varices in the lesser curvature but no esophageal varices. The rate of bleeding depends on location, varices located in the gastric fundus cause bleeding more often. Sarin’s classification continues in use up to the present and has been useful in helping to determine the type of treatment in adults, according to the findings.

Level of evidence 5, grade of recommendation D.

Hypertensive gastropathy is a mucosal lesion characterized by ectatic gastric mucosal vessels, mainly in the fundus and body of the stomach. It is classified as mild or severe (Table 8).53 In mild gastritis there is fine pink speckling, erythema on the surface of the folds, presence of a reticular white mosaic pattern separating areas of erythematous and edematous mucosa that resemble snakeskin. Severe gastritis presents with cherry red smudges or spots with diffuse hemorrhagic gastritis. The presence of gastro-esophageal varices is a predictive factor for hypertensive gastropathy and the presence of gastropathy is high in patients with esophageal varices who have undergone endoscopic treatment.53,54

Level of evidence 4, grade of recommendation C.

Propranolol and nadolol are the most widely studied non-selective beta-blockers in adult patients with cirrhosis; their mechanism of action involves blocking the β1 receptors that causes decreased cardiac output and splanchnic vasoconstriction by blocking β2 receptors. Reduction of the hepatic venous pressure gradient (HVPG) to less than 12 mmHg or a reduction greater than 20% from baseline is the only appropriate parameter for evaluating reduction in portal pressure; however performing this invasive procedure in children raises ethical issues, so it is not routinely performed.1,44 To assess the effective dose, a 25% reduction in heart rate (HR) has been recommended however in adults not even a 40% reduction in the HR has produced a significant decrease in HVPG.55 The decrease in HR only seems to be useful for determining tolerance to the drugs and not response to treatment. The dose used in children was between 1 to 2 mg/kg/day and in some cases up to 8 mg/kg/day, to achieve a 25% reduction in HR from baseline.56 Reported adverse effects in adults are: bradycardia, heart failure, behavioral disturbances, bronchospasm and hypoglycemia, which have been reported in 10 to 20% of the cases treated.57 There is not enough information to recommend their use in children with EHPVO.

Level of evidence 4, grade of recommendation D.

There are no reports available in the reviewed literature, in either children or adults, to justify use of non-selective beta-blockers (NSBB) in preventing the development of varices.

Level of evidence 4, grade of recommendation D.

There are few studies in children that prove BBs are effective in reducing the risk of bleeding. Retrospective and prospective case series in children treated with propranolol with or without cirrhosis, first bleeding occur in 16% and 35% of cases in a follow-up time between 3 and 5 years respectively,58–60 these percentage is similar for the first bleeding that occurs in 30% of the cases in the natural course of the disease over a follow-up period of 24 months.60,61 In adult patients with cirrhosis controlled studies showed that propranolol reduces the risk of bleeding between 30% and 50%; this effect is limited to those with medium or large varices.62,63

Based on the limited evidence, we cannot recommend propranolol for primary prophylaxis in children with EHPVO.64

Level of evidence 4, grade of recommendation D.

Endoscopic therapy (sclerotherapy or variceal ligation) in adult patient with medium and large varices had been reported to reduce the risk of bleeding by 30% without altering mortality.65 In children sclerotherapy for primary prophylaxis has been reported in 26 cases, 42% had variceal bleeding within an average of 2.4 years after the endoscopic therapy;66 in a prospective study of 100 children with portal hypertension with different etiologies, variceal hemorrhage appeared in 6% with sclerotherapy vs. 42% (p < 0.05) of those who did not undergo the procedure during a mean follow-up of 36 months. However the rate of variceal gastric hemorrhage and hypertensive gastropathy (HTG) seems to be higher in those who had sclerotherapy 38% vs. 6% who did not have sclerotherapy (p < 0.05),14 this is an important aspect to consider before initiate an esclerotherpay as primary prophylaxis treatment. In most of these cases the patients had cirrhosis, so that on the basis of these studies, prophylactic sclerotherapy cannot be recommended in children with EHPVO.

Variceal ligation for primary prophylaxis has been reported in only two series of children with and without cirrhosis and showed a bleeding rate of 0%–10% during a follow-up of 16 and 23 months, respectively, this procedure was well tolerated without major complications.15,67 There are still few studies in children with EHPVO to allow recommendation of sclerotherapy and ligation for primary prophylaxis. However, in children with large varices who are at a high risk of bleeding or living far from the hospital, endoscopy therapy is justified, otherwise it must be done in a protocolized study.

Level of evidence 4, grade of recommendation D.

There is no information regarding the combined use of propranolol and sclerotherapy/ligation in children, so therefore it cannot be recommended.

Level of evidence 4, grade of recommendation D.

It is recommended that endoscopy be performed every 2–3 years in adult patients with cirrhosis without varices. Patients with small varices should have one every 1–2 years and patients with big varices should have one every year.46

Nothing has been published to support endoscopic screening in children with EHPVO or to define appropriate follow-up times so that studies are necessary to do so.56 Extrapolating from adults to children and taking the natural course of the disease into consideration,61 we believe that an endoscopy should be performed initially in every child with EHPVO and follow up must be under a study protocol every 1–2 years in children with esophageal varices with no history of bleeding.

Level of evidence 4, grade of recommendation D.

• 1.

  Stabilization and restitution of blood volume to be done cautiously to maintain hemoglobin between 8–10 mg/dL.1

• 2.

  Vasoactive drugs. Terlipressin, somatostatin and octreotide, they should be maintained between 48 h and 5 days.46

• 3.

  Endoscopic treatment. Ligation is the treatment of choice for bleeding from esophageal varices (in older children) within the first 24 hours after admission and after the child is hemodynamically stable.2 For bleeding gastric varices the procedure is obturation with cyanoacrylate.1,2

• 4.

  Sengstaken-Blakemore tube. Stops bleeding in almost 90% of cases, to be used as a temporary measure and should be inserted by trained staff because of the high risk of complications.68

Level of evidence 1, grade of recommendation A.

Vasoactive drugs cause splachnic vasoconstriction, thereby reducing portal pressure and controlling variceal bleeding in 75% to 80% of patients. They should be started as soon as possible, prior to endoscopy in patients with suspected variceal hemorrhage.46,67

Level of evidence 1, grade of recommendation A.

a) Octreotide. It is a somatostatin analogue that is equally effective; control of bleeding was reported in 70% of 3 pediatric case series. The recommended dose is a 1 to 2 mcg/kg bolus followed by infusion of 1 to 2 µcg/kg/h. The infusion is adjusted according to response. When the bleeding is under control, the dose is reduced by 50% every 12 to 24 h.69,70

Level of evidence 1, grade of recommendation A.

b) Terlipressin is a synthetic analog of vasopressin, with a longer shelf life and fewer side effects. Control of bleeding within the first 48 h is achieved in 75% to 80% of adults and of 67% within 5 days. It is administered as an initial IV dose of 1 mg for < 50 kg, 1.5 mg for 50–70 kg, 2 mg for > 70 kg followed by 1–2 mg every 4 h for 5 days. In pediatrics, use of terlipressin has been reported as rescue treatment in patients with refractory septic shock and bleeding in the digestive system in a case series in which control of bleeding was achieved in 8/15 (53%) of cases.71,72 More information is required before recommendations for its routine use in the treatment of esophageal varices in children can be made.56

Level of evidence 5, grade of recommendation D.

Depending on the age of the patient and skill of the endoscopist, both sclerotherapy and ligation have been proven effective in controlling acute variceal bleeding. Endoscopic therapy should be performed within the first 12 to 24 h of the onset of bleeding, after the child is stable and hemodynamically compensated.1,2,46

Level of evidence 1, grade of recommendation A.

• •

  Sclerotherapy. The obliteration of varices can be achieved in approximately 92% of children with extrahepatic portal hypertension and this technique can be used in infants, children and adolescents.11,73–75

• •

  Ligation of esophageal varices. Has been reported to be effective for controlling acute bleeding in 96% of cases, similar to sclerotherapy, which has a 92% rate.73,75,76

Level of evidence 1, grade of recommendation A.

• •

  Combination therapy. A meta-analysis of adults treated with a combination of vasoactive drugs and endoscopic therapy showed better control of initial bleeding and at day 5.77 A case report in children showed control of bleeding control in 90%.78

Level of evidence 4, grade of recommendation D.

• •

  Adults. Persistence of hematemesis > 2 h after endoscopic therapy is considered a failure to control acute variceal bleeding, Hb decrease of > 3 g/dL.8

• •

  Children. Fresh hematemesis or nasogastric aspiration of > 2 mL/kg/h or 100 mL of fresh blood > 2 h after the therapeutic endoscopy.2,46

Level of evidence 1, grade of recommendation A.

The device consists of two balloons, one for the esophageal varices and one for gastric varices; it stops bleeding in nearly 90% of cases. However, it is only temporary measure and should not be left in place for more than 24 h; it has a high rate of complications such as mucosal ischemia, perforation, aspiration and airway obstruction. It is recommended for use in cases when drug and/or endoscopic treatment fail, or as a temporizing measure until surgery.68,79

Level of evidence 4, grade of recommendation D.

It is unknown in children with EHPVO, it has been reported to be between 0% to 8% in cases with intra- and extrahepatic portal hipertensión.58,74,80–82

Level of evidence 4, grade of recommendation D.

The criteria have not been clearly defined for children with EHPVO; extrapolating from adults to children, the following criteria should be taken into account:1,2

The time frame for the acute bleeding episode is 120 h (5 days).

Treatment failure signifies the need to change therapy of at least one of the following criteria:

• 1.

  Fresh hematemesis is ≥ 2 h after the start of a specific drug or endoscopic treatment.

• 2.

  For patients with a nasogastric tube, aspiration of more than 100 mL of fresh blood or more than 2 mL/kg.

• 3.

  Three gram drop in Hb and 9% drop in Hct if no blood transfusion is administered.

• 4.

  Death.

• 5.

  Some people consider a score ≥ 0.75 on the ABRI index at any time point denotes failure, there is no data related to the use of ABRI in children.

ABRI = blood units transfused/(final hematocritinitial hematocrit)+ 0.01)

Level of evidence 5, grade of recommendation D.

The most widely used non-selective beta-blocker is propranolol. The evidence is insufficient to recommend its use in adults with prehepatic portal hiypertension;1 most studies are in cirrhotic patients and it has not been shown to be superior to endoscopic treatment. In children, most of who presented with intrahepatic portal hypertension, variceal rebleeding has been reported in between 25% to 53% of cases, in a follow-up time between 3 and 5 years.58,59,83

To date, there are insufficient data to allow recommendation for use of propranolol as standard therapy for the secondary prophylaxis of bleeding esophageal varices in children with EHPVO.

Level of evidence 5, grade of recommendation D.

Non-selective beta-blockers (especially propranolol) have been used in combination with ligation for secondary prophylaxis in adult patients with prehepatic portal hypertension.1

A meta-analysis of adults with both prehepatic and intrahepatic portal hypertension showed that a combination of propranolol and ligation or sclerotherapy is superior to endoscopic treatment alone for the control of bleeding, which is why it is recommended as first line management.84

To date, combination therapy for secondary prophylaxis cannot be recommended, as there are no data regarding its use in children. Studies in this regard are needed and it could be used within a protocol.

Level of evidence 5, grade of recommendation D.

There are no data about mortality following re-bleeding from esophageal varices in pediatric patients with portal hypertension due to EHPVO.

Level of evidence 5, grade of recommendation D.

a) Sclerotherapy.

• •

  Description. Endoscopic procedure that involves intra- or paravariceal injection of pharmacological agents in order to obliterate esophageal varices and/or control active bleeding.

• •

  Types of sclerosing solutions. The most common are: 1% and 3% polidocanol, 5% ethanolamine oleate, 100% ethanol, sodium morrhuate. There are no studies comparing the different sclerosing agents for efficacy, adverse effects or with recommendations for their use in children, so we cannot recommend any drug in particular.85

• •

  Indications. During acute bleeding episodes and for the eradication of esophageal varices. 3 to 7 sessions are required to eradicate the varices.11,73,74,86

• •

  Response to treatment. In pediatric patients with pre-hepatic PHT, variceal eradication by sclerotherapy is achieved in 80 to 100% of cases,16,66,73,80,87–90 and the rate of recurrence of esophageal varices is 10% to 40% after eradication. Gastrointestinal rebleeding is frequently caused by gastric varices and occurs in 0 to 42% of cases. Bypass surgery is necessary in 4% to 13% of cases (Table 9).

  Table 9.

  Secondary prophylaxis in children with variceal bleeding due to extrahepatic portal hypertension.

  Autor Type of study	Operation	Cases (n)	Follow-up time (years)	Obliteration (%)	Duration endoscopic surveillance after obliteration (months)	Recurrent EV (%)	Re-bleeding in digestive tract (%)	Bypass surgery (%)
  Stringer, 1994 Case series	Sclerosis	32	8.7	80	-	12.5	31	13
  Yachha, 1996 Case series	Sclerosis	50	1.6	100	-	10	0	-
  Sokoku, 2003 Case series	Sclerosis	21	2	80	6	28	42	4.7
  Zargar, 2004 Case series	Sclerosis	69	15	91	06–12	5.7	12	14
  Itha, 2006 Case series	Sclerosis	183	10	89	03–6	40	7	6.7
  Maksoud, 2009 Case series	Sclerosis	82	17	95	6	4.6	36	4.8
  Poddar, 2011 Prospective series	Sclerosis bands + sclerosis	101 57	3.5 2.7	95 100	03–6	31 25	10.5 3.9	-
  Poddar, 2005 Retrospective series	Sclerosis bands + sclerosis	106 30	8 8	96 100	27 27	10 6.6	-	-
  Zargar, 2002 Controlled study	Sclerosis Bands	24 25	1.8 1.7	91 96	-	17.4 10**	25 4*	4 0

  *

  p = 0.049.

  **

  NS: not significant.

Level of evidence 3, grade of recommendation C.

b) Ligation of esophageal varices.

• •

  Description. Ligation is an endoscopic procedure that consists in the use of ties or bands to obliterate the varix and/or control active bleeding. The main limitation in pediatric patients is the size of the overtube for band placement, as it can measure as much as 11 mm in diameter. There are no reports of its use based on the age and weight of children, in our setting it is performed after the age of one year or weight > 10 kg. Nevertheless, it must be individualized for each patient.91

• •

  Indications. For controlling acute bleeding and for the obliteration of varices already receiving sclerotherapy.

• •

  Response to treatment. Ligation controls acute EV bleeding in 96% of cases.73,90 Obliteration of the varices is achieved in 16 to 100% of cases treated with 2 to 4 sessions being necessary to achieve obliteration.91–94 Recurrence of EV in pediatric patients has been reported to be from 5% to as much as 75%.90–95

Level of evidence 4, grade of recommendation D.

By comparison with sclerotherapy, ligation requires fewer sessions (3–9 vs. 6.1) and as such, fewer anesthetic procedures; there are fewer major complications (25% vs. 4%), whereas there is no difference in control of bleeding or in the obliteration of esophageal varices, which is achieved in 91.7% and 96% respectively (p < 0.61).70,73

Level of evidence 1, grade of recommendation A.

• a)

  Failure to control active bleeding or a re-bleeding episode after two separate attempts of the same endoscopic treatment.

• b)

  Three or more re-bleeding episodes requiring endoscopic treatment and blood transfusion.73

Level of evidence 5, grade of recommendation D.

• •

  Sclerosis. The most common complications of sclerotherapy are bleeding, ulceration, odynophagia and dysphagia. Other possible complications are perforation, pneumonia, sepsis, esophagopleural fistula, gastroesophageal reflux and esophageal motility disorders.87–90

• •

  Ligation. The most common complications are chest pain, dysphagia, odynophagia and ulceration. Esophageal perforation, fever and post-ligation bleeding can occur.90–95

Level of evidence 1, grade of recommendation A.

In the adult population, the use of tissue adhesives such as N-butyl-cyanoacrylate or isobutyl-2-cyanoacrylate has proven more effective than ligation and sclerotherapy for the control of bleeding, in preventing rebleeding and also in reducing mortality.4,98

In case series of pediatric patients there have been reports of good response in controlling acute bleeding.99,100

It is recommended for active bleeding from gastric varices.

Level of evidence 4, grade of recommendation D.

Several complications have been reported from the use of tissue adhesives (cyanoacrylate) such as cerebral and pulmonary embolism, damage to endoscopic equipment and adhesion of the endoscopic needle.93–95

Level of evidence 4, grade of recommendation D.

To date, the use of drug therapy for secondary prophylaxis of gastric or duodenal varices cannot be recommended, as there are no data regarding its use.

Level of evidence 5, grade of recommendation D.

The use of cyanoacrylate is useful for secondary prophylaxis of gastric variceal bleeding; although its use in children has been limited, results appear to have been good.98–101

Level of evidence 4, grade of recommendation D.

Secondary hypersplenism refers to the sequestration of blood cells that is associated with splenomegaly and causes a decrease in all blood cell lines (thrombocytopenia, anemia, leukopenia). It can cause gastrointestinal bleeding and episodes of epistaxis in patients with portal hypertension.102–104

Level of evidence 1, grade of recommendation A.

• •

  Splenomegaly.

• •

  Decrease in one or more cell lines.

• •

  Bone marrow hyperplasia.

• •

  Clinical evidence of bleeding.

• •

  Improvement in cytopenias following splenectomy.

• •

  Imaging: abdominal Doppler ultrasound and computed axial tomography scan.

Level of evidence 1, grade of recommendation A.

The aim of surgery is to divert portal blood flow into the systemic circulation. It has been reported that between 4% and 24% of children with EHPVO require bypass surgery after failure of medical and endoscopic therapy, with recurrence of bleeding being the main indication (Table 10).16,23,24,73,85,88,89,130,131

Different surgical techniques have been used in the surgical treatment of EHPVO (Table 11); Mesocaval and distal splenorenal or “Warren” shunts are the types most often used in children, with no difference as far as postoperative complications shunt patency and mortality are concerned.134–138

Level of evidences 1, grade of recommendation A.

It has recently been suggested that the meso-Rex procedure is the best alternative; it involves recanalization of the regular portal system, creating a bypass between the superior mesenteric vein and the intrahepatic left portal vein using an autograft; experience in Mexico is limited. Although there are few reports in the literature, results are encouraging. Restoration of venous flow is achieved in most cases, but this shunt can only be performed when a patent left branch is available and the thrombosis is higher.132,139,140

This technique can be recommended for those who have experience in performing this type of bypass.

Level of evidence 3, grade of recommendation C.

• •

  Immediate. Bleeding, encephalopathy, thrombosis, infection.

• •

  Delayed. Variceal rebleeding.

• •

  Recent studies have reported 0% to 11% recurrence of bleeding, 2% to 16% anastomotic thrombosis and 0% to 4% mortality134–138 (Table 12).

  Table 12.

  Results of bypass surgery in children with EHPVO.

  Author	Cases operated	Follow-up (years)	Rebleeding after surgery (%)	Anastomotic thrombosis (%)	Mortality (%)
  Alvarez, 1983	76	2	4	8	4
  Mitra, 1993	81	4.6	11	16	0
  Orloff, 1994	162	10	2	2	0
  Prasad, 1994	140	15	11	-	1.9
  Zargar, 2009	69	12.5	0	-	0

Level of evidence 1, grade of recommendation A.

• •

  Clinical. Disappearance of bleeding, gradual reduction of splenomegaly and hypersplenism.

• •

  Imaging. Doppler ultrasound or angiography to assess shunt patency.

• •

  Endoscopy. Follow-up endoscopy to evaluate the reduced size or disappearance of the varices (3 to 6 months after surgery).133

Level of evidence 1, grade of recommendation A.

This is a measure of last resort when everything else has failed to control acute bleeding. It is not without serious complications and experienced staff must carry out placement; it may be used as a bridge to stabilize the patient until surgical bypass can be performed. There is very little experience in children.135

Level of evidence 4, grade of recommendation D.