An 8-year-old female with no medical history, was referred to the Pediatric Intensive Care Unit due to severe hemorrhagic shock from esophageal varices bleeding. The child underwent blood transfusions and emergent gastroscopy with successful variceal ligation and hemodynamic stabilization.
On physical examination, she only presented an evident venous reticulum on her chest, without hepatosplenomegaly or other signs of portal hypertension (such as ascites, encephalopathy or confusion). Blood tests, including platelet and white blood cells count, liver function and coagulation, were within normal limits both at the admission and discharge.
Based on these findings and no liver disease history, abdominal Computed Tomography Angiography (CTA) and Doppler Ultrasound were performed to assess the liver parenchyma and vascularization. CTA showed normal liver morphology, volume and density; absence of splenomegaly and ascites. The extrahepatic portal vein was not detected at the expected site, and a single large mesenteric retroduodenal venous collateral (Fig. 1a) connected with the intrahepatic bifurcation of the right portal branch was observed. The anatomy of the intrahepatic portal venous system was substantially normal except for a dilated right portal branch. The splenic and superior mesenteric veins were patent with normal caliber, merging into a slightly dilated left gastric vein: a normal portal vein origin was absent. The enlarged gastric vein fed the esophageal varices. Doppler Ultrasound showed the presence of a stenosis at the confluence between the mesenteric venous collateral and the right intrahepatic portal branch, with a peak velocity of 150 cm/s and a pre-stenotic hepatopetal flow of 13 cm/s. No other vascular abnormalities such as arterio-venous fistulas or Abernethy shunts were noted. The biliary tree was not dilated; gall bladder and pancreas were normal. Based on the non-invasive imaging findings, a percutaneous transhepatic portography was indicated to better assess the morphology and hemodynamics of the splanchnic venous system. Under general anesthesia ultrasound-guided transhepatic catheterization of the right portal branch was performed with a 4 Fr coaxial introducer system (Neff Percutaneous Access Set, Cook Incorpo-155 rated, Bloomington, IN, USA). Percutaneous portography revealed a normal intrahepatic portal anatomy except for a dilated right portal branch. Under fluoroscopic guidance (Allura Xper FD20; Philips Healthcare, Best, the Netherlands) the mesenteric venous collateral was catheterized with a 4 Fr Simmons-shaped catheter (Cordis Corporation, Miami Lakes, FL, USA). Percutaneous venography of the splanchnic venous system confirmed the stenosis at the confluence between the mesenteric venous collateral and the right portal branch, with retrograde flow to the gastric vein and esophageal varices (Fig. 2a). Venous pressure in the mesenteric vein was above normal range (25 mmHg), suggestive of severe portal hypertension, with a trans-stenotic gradient of 18 mmHg. The 4 Fr introducer was exchanged with a 6 Fr vascular sheath (Merit Medical System, South Jordan, UT, USA) on a 0.035″ Amplatz guidewire (Cook Incorpo-155 rated, Bloomington, IN, USA). A 10-mm balloon (Mustang, Boston Scientific, Marlborough, MA, USA) angioplasty of the stenotic tract was performed and control venography showed an improved portal flow to the liver with restoration of the normal anterograde flow in the gastric and splenic veins (Fig. 2b). Final venous pressure in the mesenteric vein decreased to 14 mmHg; the same pressure was measured in the intrahepatic branches with no trans-stenotic gradient. The transhepatic tract was embolized using cyanoacrylate (Glubran 2, GEM, Viareggio, Italy) without intraprocedural complications and the child was discharged in good conditions after 1 week. Doppler Ultrasound performed 72 h, 1 month and 6 months after the procedure showed an adequate hepatopetal flow (50 cm/s) within the mesenteric venous collateral. The confluence between the collateral and the right portal branch presented a 5-mm caliber with a maximum velocity of 80 cm/s (Fig. 2c), with no residual hemodynamically significant stenosis. Intrahepatic portal venous flow was present, turbulent, with a mean velocity of 50 cm/s. The clinical benefit of the treatment was proven by the absence of varices at the 1-month followup gastroscopy and no bleeding episodes, with normal blood counts, during the 6-month follow-up.