A 70-year-old woman who was under regular treatment at our hospital for cirrhosis and hepatocellular carcinoma due to nonalcoholic steatohepatitis had developed IGV. Laboratory tests on admission revealed blood levels of albumin of 3.9 g/dL, bilirubin of 2.0 g/dL, creatinine of 0.74 mg/dL, sodium of 140 mmol/L, prothrombin time of 69.6%, hemoglobin of 9.0 g/dL, and platelet count of 69,000 /µL. Her Child-Pugh score was 7 (class B) and her Model for End-stage Liver Disease (MELD) score was 11. Upper gastrointestinal endoscopy showed IGV in the fundus and thin esophageal varices. The IGV were beaded and moderately enlarged, with no red spots. Computed tomography (CT) showed a thin PCV via the inferior phrenic vein (IPV) as the main efferent vein. The varices were mainly fed by the posterior gastric vein, and the dilated left gastric vein joined in the drainage route of the varices (Fig. 1). There was no development of a gastrorenal shunt in continuity with the varices. BRTO via the PCV was selected for treatment of the varices.
The left internal jugular vein was punctured with a 20-gauge needle under ultrasound guidance, and a 5F sheath (SuperSheath; Medikit Co. Ltd., Tokyo, Japan) and a 0.035" wire (Radifocus Guide Wire M; Terumo, Tokyo, Japan) were inserted into the selected PCV. A 5.2F balloon catheter (Selecon MP catheter II; Terumo, Tokyo, Japan) was inserted up to the level of the diaphragm. Balloon-occluded retrograde transvenous venography did not show the varices, although it showed drainage via the gastrocaval shunt and intercostal veins (Hirota Grade 4, Fig. 2) [3]. At the subdiaphragmatic level, the route to the varices was quite tortuous. A 1.9F micro balloon catheter (LOGOS GrandMaster; PIOLAX Medical Devices, Yokohama, Japan) could not reach the efferent vein near the varices. Only a 1.7F microcatheter (Progreat λ17; Terumo, Tokyo, Japan) was able to advance the catheter closer to the varices. Therefore, CARTO with the inverted catheter tip technique was performed, as described below (Fig. 3a). Venography from the IPV did not show the afferent veins or varices (Fig. 3b). First, the microwire was inverted using the arcuate edge of the IPV and followed by the microcatheter (Fig. 3c). Next, a fibered coil (Tornado18; Cook Medical, Bloomington, IN, USA) was placed from the inverted microcatheter in the IPV (Fig. 3a-i). Then, the loop of the microcatheter was released (Fig. 3d). Venography showed that the left gastric vein was the afferent vein joining the drainage route of the varices, although it did not show varices (Fig. 3e). Then, the left gastric vein was embolized with a coil (Target XL; Stryker Neurovascular, Kalamazoo, MI, USA) (Fig. 3a-ii). And 5% ethanolamine oleate iopamidol (EOI) was injected near the varices (Fig. 3a-iii and f). Cone-beam CT showed stagnation of the sclerosing agent in the IGV. On the third postoperative day, abdominal contrast-enhanced CT showed thrombosis of the IGV. At 2 months after BRTO, endoscopy showed shrinkage of the GV. The esophageal varices remained unchanged and will continue to be monitored.