At endoscopy, the lesions from PHG are found in the fundus of the stomach, but similar lesions can be seen throughout the gastrointestinal tract. Mild PHG has the appearance of a snake-skin mosaic pattern with severe PHG appearing flat or as bulging red spots. There are multiple different classification systems based on endoscopic appearance which are outlined in Table 12.3 with the North Italian Endoscopic Club (NIEC) classification being most commonly used. On histology, there is mild-to-moderate dilation of the veins and capillaries of the gastric mucosa and submucosa and no changes in the blood vessel wall [4].

GAVE is typically noted in the antrum of the stomach on endoscopy and is limited to the stomach. The lesions of GAVE appear as flat red spots without the background mosaic pattern that is seen in PHG. The red spots can merge causing stripes into the pylorus, which has led to the term “watermelon stomach.” The histology is characterized by marked dilation of capillaries and venules in the gastric mucosa, submucosa with areas of intimal thickening, spindle cell proliferation, fibrohyalinosis, and thrombi [4, 10]. The features of the histologic appearance of GAVE are used in the GAVE score (Table 12.4) which has an 80 % diagnostic accuracy in separating GAVE from PHG with a cutoff of GAVE score ≥ 3 [10].

In patients with chronic bleeding, the first-line therapy is iron replacement. If iron therapy is not sufficient to maintain the hemoglobin, further treatments are necessary with the initial goal of lowering the portal pressure in patients with chronic bleeding. Nonselective beta-blockers have been used for medical therapy with studies finding the most benefit in patients with mild PHG [14] with only modest effects noted in patients with severe PHG [15]. In patients who do not respond to medical therapy, TIPS procedure may be considered. In a study by Kamath et al., 75 % of patients with severe PHG showed improvement after TIPS as measured by improvement on endoscopic appearance and a decrease in transfusion requirement [1]. Improvement post TIPS procedure was noted as early as 2 weeks based on stabilization of hemoglobin and decreased transfusion requirements [1].

Acute bleeding can rarely occur from PHG; one large study found an incidence of 2.5 % for acute bleeding from PHG compared to 10.8 % for chronic bleeding [16]. Acute bleeding from PHG is managed similarly to variceal bleeding, with initial management consisting of appropriate resuscitation and antibiotic coverage. The main goal of medical therapy in acute bleeding is the reduction of the portal pressure. Agents that have been described in the literature include somatostatin [17], vasopressin [18], octreotide [19], and terlipressin [20]. In severe cases of acute bleeding, TIPS or surgical shunt may be necessary for management. For patients who fail medical therapy and who are not surgical candidates, there is evidence that endoscopic therapy with argon plasma coagulation (APC) could be an alternative therapy [21]. In a prospective trial by Herrera et al., APC therapy was examined in 29 patients admitted for upper gastrointestinal bleeding, with 11 patients having an underlying diagnosis of PHG. Herrera et al. found that APC was a successful therapy in 81 % patients with APC with success defined as no further episode of upper gastrointestinal bleeding and an increase of hemoglobin by 30 % or an increase in hematocrit of 10 %. In this study, the average number of sessions of APC required was 2.2 for patients with PHG and the average follow-up was 23.1 months [21]. One small case series of four patients examined the use of hemospray, a hemostatic agent licensed for endoscopic hemostasis in non-variceal upper gastrointestinal bleeding in Europe and Canada [22]. The rationale for the use of hemospray was that it would allow for the treatment of a large area and in this study hemostasis was achieved in all four patients; However, one patient passed away from a perforated viscous and subsequent sepsis indicating the need for caution [22].