Management of Gastropathy and Gastric Vascular Ectasia in Portal Hypertension

Synopsis

Portal hypertensive gastropathy and gastric antral vascular ectasia may cause gastrointestinal hemorrhage in patients with portal hypertension. While the former presents exclusively in patients with portal hypertension; gastric antral vascular ectasia can also be observed in patients with other conditions. Diagnosis is established with upper gastrointestinal endoscopy, although some cases may require a biopsy to confirm the diagnosis. The most frequent manifestation is ferropenic anemia that may become transfusion dependent. Treatment in portal hypertensive gastropathy is focused on portal pressure reducing drugs, mainly non selective beta-blockers while in gastric antral vascular ectasia it is based on endoscopic ablation. More invasive options can be utilized in case of failure of first line therapies, although this should be evaluated on a case-by-case basis.

Patients with cirrhosis are at an increased risk of gastrointestinal hemorrhage, with the most common source being gastroesophageal varices. However, there are gastrointestinal mucosal lesions typical of cirrhosis that may also bleed in these patients, namely portal hypertensive gastropathy (PHG) and gastric vascular ectasia (GAVE). These are two clearly distinct mucosal lesions with different pathophysiology, endoscopic appearance, histopathology and treatment (Table 1). PHG, as its name indicates, is associated with the presence of portal hypertension and therefore is only observed in patients with this condition, while GAVE is observed in patients without portal hypertension or liver disease. Nevertheless and despite the differences between both entities, they may lead to similar clinical manifestations, most frequently to symptoms due to chronic ferropenic anemia associated to chronic occult blood loss and only rarely lead to acute, overt gastrointestinal hemorrhage.

Table 1.

Differential characteristics between Portal Hypertensive Gastropathy and Gastric Antral Vascular Ectasia

Portal Hypertensive Gastropathy		Gastric Antral Vascular Ectasia
Causal	Relationship with portal hypertension	Coincidental
Mainly proximal	Distribution in stomach	Mainly distal
+	Presence in other territories of GI tract	-
Mosaic pattern + red spots	Endoscopic findings	Linear pattern + red spots
Severe portal hypertensive gastropathy	Difficult differential diagnosis	Diffuse gastric vascular ectasia
Dilated capillaries and venules No inflamation	Pathology	Thrombi Spindle cell proliferation Fibrohyalinosis
Portal pressure reducing	Treatment	Endoscopic
TIPS/Shunt surgery	Salvage therapy *	Antrectomy and Billroth I

^*to be evaluated on an individual basis

SMT: somatostatin; TIPS: transjugular intrahepatic shunt

Portal Hypertensive Gastropathy

Portal hypertensive gastropathy (PHG) is characterized by more or less typical gastric mucosal lesions presenting in patients with portal hypertension (either pre-hepatic or hepatic). Its typical location is in the gastric fundus and upper body of the stomach although it can affect the whole stomach and even other areas of the gastrointestinal tract, such as the small bowel or the colon ^1-8.

Epidemiology

The prevalence of PHG in patients with portal hypertension has been reported to vary between 20% and 80% ^{1, 9-12}. The wide variation in the reported prevalence is most likely due to differences in the study population specifically, the proportion of patients with non-cirrhotic portal hypertension, the severity of the underlying liver disease, and the proportion of patients with previous endoscopic treatment. A higher rate of PHG is observed in patients with more severe liver disease ^{10, 11} and in patients who have had previous endoscopic treatment with sclerotherapy or endoscopic variceal ligation ^{1, 9, 12} There is controversy regarding the specific endoscopic technique used for variceal eradication that leads to a higher prevalence of PHG, with some studies showing a higher incidence after sclerotherapy than after ligation and other studies showing a similar incidence with both techniques ^13-15. Although many studies have concluded that the presence of PHG is associated with indirect signs of portal hypertension such as the presence of larger varices ^{1, 11}, splenomegaly ^{10, 11} and low platelet count ¹¹, the studies that evaluated the association between the hepatic venous pressure gradient, a well established method to measure portal pressure ¹⁶, and PHG have lead to controversial results. Two studies found no clear relationship between HVPG and PHG ^{17, 18}, while another study found that patients with more severe gastropathy had a higher HVPG than patients with mild or no gastropathy¹⁹. These controversial results may be due to selection bias, as all patients who were included in these studies had clinically significant portal hypertension ^{17, 18}. The issue would be clearer if it could be demonstrated that PHG presents only in patients with cirrhosis and portal hypertension (HVPG ≥ 6mmHg) or only in those with clinically significant portal hypertension (HVPG ≥ 10mmHg). However, such studies are still lacking.

Pathophysiology

The pathophysiology of PHG is unclear although portal hypertension plays a major role. Studies that have evaluated the prevalence PHG have revealed that it is almost exclusively observed in patients with portal hypertension, with or without liver disease ^{20, 21}. PHG was not observed in 100 patients with chronic alcoholism without signs of liver disease or portal hypertension (on ultrasonographic examination) or in 10 patients with cirrhosis without portal hypertension ²¹. It is not a peptic process, as mucosal changes do not respond to antisecretory drugs and histological changes, predominantly dilated capillaries and venules in the mucosa and submucosa without significant inflammation, are clearly distinct from those typically observed in peptic related disease ²². Nevertheless, it seems that the gastric mucosa has an increased susceptibility to injury by noxious factors, as well as impaired healing ^23-26.

Diagnosis

The diagnosis of PHG, and by extension portal hypertensive enteropathy (similar mucosal changes at other sites of the gastrointestinal tract) is established when the characteristic endoscopic findings are observed in patients with portal hypertension. PHG is classified as mild when only the snake-skin mosaic pattern is present or severe when, in addition to the mosaic pattern, flat or bulging red marks or black-brown spots are observed ²⁷. The clinical relevance of this classification has been established since patients with severe PHG are more likely to have acute bleeding or chronic anemia than patients with mild PHG ^{11, 28}. Furthermore, this classification is reproducible as a fairly good concordance between observers has been shown, especially regarding the mosaic pattern and red marks ²⁸. Similar lesions as the ones observed in the stomach in patients with PHG have been observed in the small bowel ^{3, 5, 7, 29, 30} and the colon ^{2, 7}.

Some studies have evaluated alternative non-endoscopic methods for the diagnosis of PHG ^{31, 32} such as MRI or CT although until further evaluation in larger populations is available, endoscopy still remains the chief diagnostic method. In a study evaluating the efficacy of capsule endoscopy in the evaluation of the presence and size of varices, capsule endoscopy was shown to have only moderate sensitivity and specificity for the detection of PHG ³³. Future studies should specifically evaluate its efficacy in evaluating not only the presence but also the severity of PHG as capsule endoscopy will be particularly important in the evaluation of lesions in the small bowel.

Natural history

PHG may change with time in an individual patient. Studies that have evaluated the natural history of PHG have described controversial results, possibly due to differences in study population. If only patients with cirrhosis who do not require primary or secondary prophylaxis are included, approximately 30% of patients with mild PHG develop severe PHG during a follow up period ranging from 12 to 103 months¹¹. Only a few cases of improvement in PHG are described. Most of the cases that bleed from PHG occur in patients with de novo PHG or in those with worsening of previous PHG ^{9, 11}. Patients with diffuse lesions are more likely to bleed ^{9, 12}. Patients who have PHG associated to cirrhosis-related portal hypertension have more frequently persistent and progressive PHG (which is more likely to bleed) than patients with PHG related to non-cirrhotic portal hypertension ⁹. Although, as mentioned above, patients with previous endoscopic therapy (sclerotherapy or endoscopic variceal ligation) have a higher prevalence of PHG ^{1, 9, 12}, the clinical course of the PHG in this context, particularly in non cirrhotic portal hypertension, may be milder and transient ^{9, 14}. Studies that have focused on patients with cirrhosis have not observed differences in severity or course of PHG in patients who develop it spontaneously or after endoscopic therapy ¹.

Clinical picture

PHG is mostly asymptomatic but, when symptomatic, it most frequently causes chronic gastrointestinal blood loss and ferropenic anemia. No study has evaluated the prevalence of PHG in cirrhotic patients with chronic anemia. The rate of chronic hemorrhage in patients with PHG ranges between 6-60% in different studies ^{1, 9, 11, 34}. This wide variation is due to a great heterogeneity in the patient population included in the different studies, from patients without cirrhosis ⁹ to patients with cirrhosis and severe PHG ³⁴. Chronic bleeding from PHG is suspected in a patient with portal hypertension and chronic ferropenic anemia and its presence is confirmed upon thorough examination of the whole gastrointestinal tract including upper endoscopy, colonoscopy and evaluation of the small bowel which is most easily done with capsule endoscopy.

Although PHG may bleed acutely leading to hematemesis and/or melena, this accounts for only a minority of cases of acute gastrointestinal hemorrhage in patients with cirrhosis. In a study that included 250 patients with cirrhosis presenting with GI hemorrhage in whom a source could be identified, PHG accounted for only 5% of the cases, with varices (57%) and peptic disease (17%) being the most frequent sources ³⁵. Studies that evaluate the probability of bleeding from PHG report an incidence between 2.5-30% ^{1, 9, 11, 34}, with the greatest incidence observed in patients with severe PHG ^{9, 11, 34}. Diagnosis of acute hemorrhage from PHG is established when active bleeding from gastropathy lesions or non-removable clots overlying these lesions is observed or when there is PHG and no other cause of acute bleeding can be demonstrated after thorough evaluation of the gastrointestinal tract.

Management

Treatment of PHG relies on two main pillars. Firstly, general measures used in gastrointestinal bleeding independent of cause should be applied and secondly a specific approach to treat the cause of the gastrointestinal bleed should be undertaken. The most effective specific treatments in patients with PHG are those aimed at reducing portal pressure. The main pharmacological agent that has been investigated in this setting is the nonselective beta-blocker propranolol.

a) Chronic hemorrhage

In the most frequent context of chronic blood loss, iron supplementation should be provided in order to counteract the continuous depletion of iron deposits. The most effective specific therapy is that aimed at reducing portal pressure. Non-selective beta-blockers have been shown to decrease bleeding from both acute and chronic forms of hemorrhage from PHG. The first trial ³⁶ was a randomized controlled trial that included only 24 patients. Patients with portal hypertension evidenced by the presence of varices and PHG diagnosed at least 6 weeks before inclusion were randomized to the administration of long acting propranolol (160 mg/day) or placebo initially for 6 weeks and then crossed-over to the other arm for another 6 weeks. While patients were on propranolol, there was a lower rate of hemorrhage, an increase in hemoglobin level and an apparent improvement in the endoscopic appearance of the lesions, compared to the placebo period. However, the study was underpowered to yield any significant statistical or clinical conclusions.

More solid evidence supporting the use of propranolol in the prevention of recurrent bleeding in patients with cirrhosis and severe PHG was obtained from a randomized controlled trial in which patients who had previously bled acutely or chronically from PHG were randomized to receive propranolol (26 patients) or no therapy other than iron administration if needed (28 patients) ³⁴. The end-point of the trial was recurrent hemorrhage from PHG. This was defined as a) the presence of acute upper gastrointestinal bleeding with a decrease in hematocrit and the finding on emergency endoscopy (performed within the first 12 hours) of gastric red spots as the source of bleeding, or b) the presence of chronic bleeding defined by occult blood loss with transfusion requirements of 3 or more units of red packed blood cells in 3 months or continuous iron repacement therapy for more than 50% of the time of follow-up. The actuarial probability of remaining free of hemorrhage was greater in patients randomized to propranolol both in the acute (85% vs. 20%) and in the chronic setting (63% vs. 40% at 30 months), although in the latter this difference did not achieve statistical significance ³⁴. On multivariate analysis, the only independent predictor of recurrent hemorrhage was the absence of propranolol ³⁴. Therefore, non-selective beta-blockers should be used both in the chronic setting and in the acute setting (once the acute episode is controlled). Propranolol is given at an initial dose of 20 mg BID, gradually escalated to 160 mg BID or the maximum tolerated dose according to heart rate (50-55 bpm) or secondary effects (light-headedness, asthenia, etc). This therapy should be maintained as long as the patient continues to have portal hypertension.

Use of other pharmacologic agents in PHG such as losartan ³⁷, thalidomide ³⁸ and corticosteroids³⁹ has been described. However the evidence supporting the use of these agents is weak with small open-label studies and case reports.

Non-response to beta-blockers should be considered when the patient continues to bleed and is transfusion-dependent despite betablockers and iron replacement therapy. Patients who only require occasional transfusions will not benefit from more invasive therapeutic options, although this should be evaluated on an individual basis.

Since portal hypertension is the main underlying factor that leads to the development of PHG, shunt therapies have been evaluated as salvage therapies. Data has been reported regarding the use of both the transjugular intrahepatic portosystemic shunt (TIPS) ^40-43 and shunt surgery ^{44, 45}. TIPS placement consists in the creation of a shunt between a hepatic vein and a portal vein. From a hemodynamic point of view it is similar to a side to side surgical portacaval shunt and is very effective in lowering portal pressure. Almost all patients who receive TIPS demonstrate an improvement of PHG lesions on endoscopic evaluation and a decrease in transfusion requirements ^40-43. In the largest study (40 patients with PHG) endoscopic improvement was detected as early as 6 weeks after TIPS placement, although in patients with severe PHG it could take at least 3 months ⁴⁰. Similar results have been shown with shunt surgery, that is, improvement in endoscopic appearance and lower transfusion requirements ^{44, 45}. Due to the morbidity associated with shunt surgery, particularly in patients with cirrhosis, its role may be more relevant in patients with non-cirrhotic portal hypertension ⁴⁵.

Only one study has evaluated the use of endoscopic treatment of PHG with argon plasma coagulation (APC)⁴⁶. This study included 29 patients, of which 11 were considered to be bleeding from PHG, defined as diffuse vascular ectasia with a background mosaic pattern distributed throughout the stomach in patients with portal hypertension. The APC was set at 30 to 40 W and 1.5 to 2 l/min of argon plasma flow. The aim of each session was to ablate as much of the surface as possible, at least 80% of the mucosa in diffuse lesions. Sessions were repeated as needed every 2 to 4 weeks. The treatment success, defined by the absence of further episodes of upper gastrointestinal bleeding or a reduction in blood transfusion requirements, was between 81 and 90% without any significant differences between groups. The data is limited and this endoscopic approach could be considered in patients in whom severe recurrent bleeding occurs despite beta-blockers and who are not candidates for TIPS.

b) Acute hemorrhage

Although PHG may present as acute gastrointestinal hemorrhage, this presentation is less frequent than chronic blood loss. As in all patients with acute hemorrhage, general measures should be undertaken. Adequate volume resuscitation should be provided. Even though data specific for acute hemorrhage from PHG is not available given its low frequency, it would appear reasonable to apply general measures proven to be efficacious in improving outcomes in patients with acute variceal hemorrhage such as cautious blood transfusion aimed at maintaining a hemogobin of around 8 g/dL and the use of prophylactic antibiotics ⁴⁷. Oral or intravenous quinolones are the recommended antibiotics such as norfloxacin (400 mg BID) or ciprofloxacin (500 mg BID PO, 200 mg BID IV). Intravenous ceftriaxone (1 g/day) may be preferred in patients with Child B or C cirrhosis, or in those undergoing prophylaxis with quinolones⁴⁸.

Besides the general measures, a specific portal hypotensive approach should be undertaken. The use of betablockers in the acute setting was evaluated in an open trial in acute severe bleeding from endoscopically-proven PHG (14 patients) ³⁶. In this context, all but one patient ceased to bleed within 3 days of the initiation of propranolol ³⁶. However, the gradual dose titration of the drug necessary to achieve adequate beta-blockade is not ideal in this setting, particularly when blunting the physiologic compensatory increase in heart rate that results from acute bleeding would appear counterintuitive. In this setting, vasoactive drugs used routinely in the setting of variceal hemorrhage such as somatostatin and its analogues (octreotide and vapreotide) and vasopressin and its analogue, terlipressin, which are effective by means of producing a decrease in portal pressure ^{49, 50}, could be useful from a theoretical point of view. In fact, three trials have evaluated the use of these drugs in acute hemorrhage from PHG ^51-53. Hemorrhage was satisfactorily controlled with somatostatin ⁵³, octreotide ^{52, 53} or terlipressin ⁵¹ although vasopressin ⁵² did not show any benefit in comparison to omeprazole. Furthermore, in the trial that evaluated the effect of terlipressin, patients who received a larger (1 mg/4 hours) dose of terlipressin had a higher proportion of control of hemorrhage and a lower recurrence rate than those who received a lower dose (0.2mg/4 hours) ⁵¹.

Therefore, following current recommendations ⁴⁷, resuscitation, a safe vasoactive drug (somatostatin or terlipressin) and antibiotic prophylaxis should be started as soon as gastrointestinal bleeding is suspected in a patient with cirrhosis, preferably before upper endoscopy. It seems reasonable that once the diagnosis of bleeding due to PHG is established, these drugs should be maintained.

In cases of acute hemorrhage that does not respond to medical treatment, alternative approaches should be contemplated. Non response to medical treatment in the acute setting may be defined according to the standards of non response to variceal bleeding ²⁷, as both are associated to portal hypertension and require a change of therapy. Treatment failure is considered when there is a new hematemesis after at least 2 hours of treatment initiation, or a 3 gr drop in hemoglobin in the absence of transfusion of packed red blood cells or an inadequate hemoglobin increase in response to blood transfusions, or death ²⁷.

Rescue therapies in patients in whom standard therapy fails are the same as those recommended in patients who fail standard therapy for chronic bleeding.

Gastric Antral Vascular Ectasia

Gastric antral vascular ectasia (GAVE) is characterized by the presence of red spots without a background mosaic pattern that are typically located in the gastric antrum. GAVE is most frequently observed in patients with cirrhosis and portal hypertension, although they have also been observed in patients without cirrhosis such as autoinmune connective tissue disorders, bone marrow transplantation or chronic renal failure ^54-57. Seemingly, patients with cirrhosis have more often diffuse disease ⁵⁸ while in non-cirrhotic patients the disease is most frequently limited to the antrum ^{59, 60}. In contrast to PHG, GAVE is only observed in the stomach and not in other parts of the gastrointestinal tract. The prevalence of GAVE in cirrhosis is very low. In a recent study performed in patients awaiting liver transplantation, GAVE was observed in only 8/345 (2%) of the patients ⁶¹. A similar prevalence has been described in patients with HCV and advanced fibrosis ¹⁰.

Pathophysiology

The pathophysiology of GAVE is not fully understood. In patients with cirrhosis, portal hypertension appears not to be essential in its development as patients do not respond to portal pressure-reducing therapies, such as TIPS or surgical shunt ^{40, 62}. Liver insufficiency seems to play a significant role in the development of GAVE because it develops in patients with more severe liver dysfunction ⁵⁹ and it has been shown to resolve after liver transplantation ^{61, 63}. Speculation regarding an accumulation of substances not metabolized by the liver which may induce vasodilatation and/or angiogenesis has been suggested as a posible mechanism ⁶². The association between GAVE and hormones with vasodilating properties such as gastrin ^{18, 54, 59} and prostaglandin E₂⁶⁴ has also been suggested. Finally abnormal antral motility ⁶⁵ and mechanical stress ¹⁸ have also been associated to the pathogenesis of GAVE which is further supported by the antral distribution of the lesions.

Diagnosis

The diagnosis of GAVE is established when the characteristic red spots (without a background mosaic pattern) are seen in the stomach at endoscopy. These lesions are typically found in the antrum, and in occasions are aggregated in a linear distribution, so that one may observe red longitudinal stripes which may be flat or raised, with visible blood vessels and pale normal mucosa in between, similar to a watermelon, for which the term “watermelon-stomach” is used ⁵⁴. In occasions, the lesions are distributed diffusely throughout the proximal and distal stomach. In these cases the term “diffuse gastric vascular ectasia” is preferred ⁵⁴. The appearance of diffuse GAVE may be very similar to severe PHG and it is important to differentiate them as treatment will be different ⁶⁶.

Histologically, GAVE lesions are completely distinct from PHG so biopsy can be used to distinguish between these two entities particularly in those cases in which the endoscopic appearance of the lesions may lead to some confusion ⁵⁹. Findings in full-thickness mucosal biopsies that are highly suggestive of GAVE are mucosal vascular ectasia, fibrin thombi, fibrohyalinosis and spindle cell proliferation without signs of inflammation. However, the absence of these characteristics on biopsy does not preclude its diagnosis as biopsies are normally not deep enough ⁶⁷.

Clinical presentation

From a clinical perspective, GAVE presents in a similar manner as PHG, that is, while most patients are completely asymptomatic, others present with acute or chronic hemorrhage and iron deficiency anemia. Acute hemorrhage appears to be more frequent in patients with cirrhosis, while non-cirrhotic patients will generally present with anemia ⁶⁰. In a study in which patients (with or without cirrhosis) underwent endoscopic evaluation of nonvariceal upper gastrointestinal hemorrhage, GAVE was the cause of hemorrhage in 26/744 patients (4%). Approximately a third of these patients with GAVE had underlying portal hypertension ⁵⁸.

Management

In the setting of acute hemorrhage, the general therapeutic measures recommended for patients with cirrhosis and acute hemorrhage from varices or PHG (see above) will apply to patients with cirrhosis who bleed acutely from GAVE.

Specific measures to treat patients with GAVE with acute or chronic bleeding are substantially different from those used in PHG. The mainstay of therapy in GAVE is the endoscopic ablation of the lesions. Some other studies have evaluated the use of different drugs and other more invasive options, but these should be used once an endoscopic therapy has failed.

Many different endoscopic approaches have been used in the setting of GAVE although most studies have evaluated the use of thermoablative techniques. Typically several sessions are needed to achieve a control of acute bleeding and to correct anemia due to chronic blood loss.

The use of neodymium:yttrium-aluminium-garnet (Nd:YAG) laser coagulation has shown to be effective in reducing rebleeding and transfusion requirements ^{54, 56, 68-71}. Its main advantage is that it can be applied to a fairly large suface area of the mucosa in a single session and that its hemostatic response may be observed earlier. On the other hand the risk of perforation is greater than with other thermablative approaches such as argon plasma coagulation as the thermal effect penetrates deeper. The Nd:YAG laser is placed 1 cm from mucosa surface and is used with a power setting between 40-90 W with short pulse durations (0.5-1 sec). Sessions should be repeated every 2-4 weeks until the therapeutic goal is achieved.

The most numerous studies evaluating the use of thermoablative methods in the treatment of GAVE are with argon-plasma coagulation which produces thermal coagulation by applying high frequency electric current that is passed through with argon gas without direct contact with the mucosa. Several studies have demonstrated the beneficial effect of this method in patients with GAVE ^{46, 58, 60, 72-78}. The main advantage of argon-plasma coagulation is that it is easy to use and the risk of perforation is lower than with Nd:YAG laser. Large areas of mucosa may be treated in a single session although this is very time consuming. Complications associated to this method are the development of hyperplasic polyps ^78-80 and gastric outlet obstruction ^{77, 81}. The settings for the electrical power (20 to 80W) and gas flow (0.5-2 L/min) vary throughout the studies. The technique combines both focal pulse and “paint brush”. The sessions should be repeated every 2-6 weeks as needed.

Other endoscopic therapeutic options such as sclerotherapy and heater probe ablation ⁸² do not appear to have significant advantages over the previously described methods. However there are other endoscopic approaches with more limited experience that may offer some additional advantages. Cryotherapy ^{83, 84} consists of rapid expansion in the stomach of compressed nitrous oxide resulting in a local decrease of the temperature with consequent freezing of the mucosa. The cryospray is applied with a specific catheter until ice is formed. Endoscopy should be repeated until all lesions are fully treated. The main advantage of this approach is that it allows for a faster and more extensive treatment of diffuse lesions. Recently, the use of banding in the stomach antrum with the same system that is used for variceal ligation has been evaluated for the treatment of GAVE ^{85, 86}. This method offers the advantage that a large area of mucosa can be treated at once and it is a technique that is easily accesible to many centers. In a controlled trial comparing banding to thermoablative treatment, patients who received banding had a significantly greater increase in hemoglobin, decrease in blood transfusion requirements and hospital admissions ⁸⁵.

Limited experience has been reported regarding the use of different pharmacological treatments for GAVE. Pilot studies have suggested that the use of estrogen-progesterone may be useful, so that this could be tried if endoscopic treatment fails ^87-89. Some success has been described in case reports with the use of octreotide⁹⁰, corticosteroids ^{91, 92}, tranexamic acid ⁹³, thalidomide ⁹⁴, and a serotonin antagonist ⁹⁵. Some cases have resolved with treatment of the underlying disease such as liver transplantation in patients with cirrhosis ^{61, 63}, although there has been a report of worsening GAVE despite adequate control of the underlying disease, systemic sclerosis and interstitial pneumonitis ⁹⁶

In extreme cases in which the combination of endoscopic and pharmacological treatment is unsuccessful, surgery with antrectomy can be considered on an individual basis ^97-99. The preferred procedure is antrectomy with Billroth I anastomosis. A thorough endoscopic evaluation of the gastrointestinal tract should be performed in order to exclude other causes of bleeding before turning to surgery. This is the most definitive therapy with low rates of rebleeding and anemia. However, morbidity and mortality are high, particularly in patients with decompensated cirrhosis in who GAVE usually presents.

Conclusion

PHG and GAVE are two different clinical entities that share a common clinical manifestation, gastrointestinal bleeding. Most cases will present as chronic bleeding, although there may be some cases of acute life threatening bleed. Besides the common clinical manifestation, these two entities are clearly distinct with different pathophysiology, endoscopic appearance and therapeutic approach. The treatment of PHG is based on measures that reduce portal pressure, namely the administration of betablockers. On the other hand, gastric vascular ectasia responds to endoscopic treatment. Most studies in this context have used thermoablative techniques, mainly argon-plasma coagulation. Refractory cases can respond to more invasive therapeutic options, although this has to be evaluated individually.

Figure 1.

Figure 2.

Figure 3.