Although the risk of bleeding from gastric varices is relatively low (10–36%) and less than that of esophageal varices, once bleeding from gastric varices does occur, the mortality is high.1,2 The treatment of bleeding gastric varices is still a challenge to clinicians, and prophylactic obliteration of risky gastric varices remains controversial. Gastric varices generally develop as a part of gastrorenal shunt communicating to systemic circulation, resulting in a higher blood flow through the gastric mucosa.3 Standard endoscopic sclerotherapy is not effective for the treatment of bleeding gastric varices, and endoscopic gastric variceal obturation with cyanoacrylate is necessary for an initial hemostasis. Endoscopic treatment with cyanoacrylate results in good initial hemostasis rates (≥ 90%), but is limited by a higher rebleeding rate (20–60%).4–6 Therefore, a subsequent definitive treatment is required to prevent rebleeding of gastric varices.
In Japan, balloon-occluded retrograde transvenous obliteration (B-RTO) has been developed and established as a reliable treatment for the prevention of primary bleeding from risky gastric varices and secondary bleeding of gastric varices after an initial hemostasis and in portosystemic encephalopathy.7–9 B-RTO involves inserting a balloon catheter into a major outflow vessel, such as the gastrorenal shunt, and embolization with 5% ethanolamine oleate under balloon occlusion. The 5-year cumulative rebleeding rate is 0–5.5% in patients with gastric variceal bleeding.3,8,10 This rebleeding rate after B-RTO is extremely low compared with that after endoscopic gastric variceal obturation with cyanoacrylate. In primary prophylaxis for risky gastric varices, several studies have reported no bleeding over long-term follow up.3,7,8 Although these studies from Japan have a potential limitation related to the lack of randomized, controlled trials, we strongly suggest that primary and secondary prophylaxes of gastric varices with B-RTO confers a significant advantage over the absence of specific therapy.
Clinical long-term studies have consistently noted an association between portal pressure and the risk of esophageal variceal bleeding, generally suggesting that reducing the hepatic venous pressure gradient (HVPG) below the threshold of 12 mmHg, or at least by 20%, considerably lowers the risk of variceal bleeding.11,12 This goal is the current therapeutic standard, and decompressive procedures, such as transjugular intrahepatic portosystemic shunt (TIPS) and β-blockers, are recommended for the prevention of esophageal variceal hemorrhage. However, HVPG in patients with gastric variceal bleeding are lower than that with esophageal variceal bleeding, and gastric variceal bleeding can occur even at a HVPG < 12 mmHg, because gastric varices are associated with a well-developed, high-flow, low-pressure portosystemic shunt.13,14
Patients in whom pharmacological interventions achieve the threshold of HVPG below 12 mmHg (or at least a 20% reduction) have been shown to have a better overall prognosis than patients who do not respond.15 However, one study has shown that in a group with HVPG > 12 mmHg, the prognosis of patients with gastric variceal bleeding was better than that in patients with esophageal variceal bleeding.14 Decompressing procedures, such as TIPS and β-blockers, are less effective in overall outcomes, including rebleeding, in patients with gastric variceal bleeding than in those with esophageal variceal bleeding.3,14 It is therefore reasoned that the therapeutic goal of gastric varices should not be to reduce HVPG to below 12 mmHg, but to obliterate gastric varices with B-RTO or to devascularize the upper stomach by Hassab's operation.3,16
In this issue of the Journal of Gastroenterology and Hepatology, Uehara and colleagues report that B-RTO caused a mean elevation of HVPG from 11.7 mmHg to 16.4 mmHg, 44% above the baseline;17 this result is consistent with a previous study.18 However, B-RTO not only increases portal venous pressure by occlusion of a large collateral vessel, such as a gastrorenal shunt, but also augments portal venous blood flow and improves liver function tests.7–9,17,18 Interestingly, the authors' previous study reported that there was no significant difference in the survival rate after B-RTO between Child–Pugh classes A and B or class C.8 Another study also reported that the prognosis of patients without hepatocellular carcinoma who underwent B-RTO was not affected by the severity of liver function.7 Several studies, including those by Jiao et al., using a rabbit fibrosis model,19 and by Cardoso et al., using isolated perfused cirrhotic rat and human livers,20 have demonstrated that an increase in portal venous blood flow produced by mechanically pumping not only increases portal inflow pressure, but also decreases intrahepatic portal resistance (IHPR) and dilates sinusoidal spaces in cirrhotic liver, changes that resulted in improving liver function.
In liver cirrhosis, portal hypertension is characterized by increased intrahepatic vascular resistance and elevated splanchnic blood flow. Hepatic stellate cells play a crucial role in regulating sinusoidal vascular tone by their contraction. In turn, such contractility is regulated by a counterbalance between vasoactive agents, such as endothelin-1, and vasorelaxing agents, such as nitric oxide (NO). Recent studies have shown that NO production in hepatic sinusoidal endothelial cells is decreased in the cirrhotic liver, leading to increased intrahepatic resistance.21,22 Generally, the increased shear stress induced by blood flow augments NO production in the vascular endothelium and mediates vasodilatation.23 This decreased IHPR might result from sinusoidal dilatation by NO overproduction following the augmented portal blood flow.19 One clinical study in 14 cirrhotic patients who underwent B-RTO showed that hepatic blood flow significantly increased 4 weeks after the procedure and was associated with reduced IHPR.18 B-RTO is likely to enhance portal blood flow, and subsequently to reduce IHPR through shear stress-induced vasodilatation, finally leading to improve liver function. Mechanical portal pumping might be a useful therapeutic modality in cirrhotic portal hypertension, but it would be a difficult procedure to apply in the clinical setting. Therefore, we suggest that B-RTO could be a procedure potentially to enhance portal blood flow with benefits in intrahepatic hemodynamics that are similar to those elicited by mechanical portal pumping.
In the present study, the authors demonstrated that patients with an increase in HVPG ≥ 20% showed a significant improvement of liver function 6 months after B-RTO, whereas those with an increase in HVPG < 20% showed no significant change. Shear stress is determined mainly by three factors: vessel radius, flow rate, and viscosity.23 It is calculated from the flow rate, pressure change, and vessel length. If viscosity and vessel length are considered constant in the intrahepatic portal venous system, shear stress in the portal venous system can be estimated as an index calculated from the changes in portal pressure and portal blood flow. In the present study, shear stress in the intrahepatic portal venous system could be much more pronounced in patients with an increase in HVPG ≥ 20% than in those with an increase in HVPG < 20%, leading to reduced IHPR and sinusoidal dilatation, and then resulting in improved hepatic microcirculation.
In conclusion, we recommend that B-RTO is a reliable and safe procedure as a radical treatment after initial hemostasis of gastric variceal bleeding, as well as for prophylactic treatment for risky gastric varices. Beyond obliteration of gastric varices, B-RTO is likely to be a potential procedure to enhance portal blood flow and improve liver function in liver cirrhosis. To clarify the overall efficacy of B-RTO, randomized, controlled trials to compare B-RTO and gastric variceal obturation with cyanoacrylate or TIPS are necessary.