Endoscopic ultrasound guided transgastric shunt occlusion or ETSO procedure for recurrent hepatic encephalopathy. None of the authors have any financial disclosures and we have not received any specific funding for this procedure. Spontaneous portosystemic shunts are common in patients with cirrhosis. Sometimes they may be large and in relation to the bowel wall which may lead to a risk of bleeding. They might also cause extensive shunting of the portal blood into the systemic circulation contributing to recurrent hepatic encephalopathy. In these situations, these shunts may need to be occluded. Usually these procedures are performed by interventional radiology going from systemic circulation into the portal circulation and blocking of the access. This involves occluding the portosystemic junction using a balloon, plug or coils and sclerosing of the proximal or the portal segment of this collateral channel. Once the portosystemic communication is blocked, it is very difficult to go back in in case the procedure needs to be repeated or there is incomplete occlusion of the collateral channel. We here describe this procedure being performed using endoscopic ultrasound guidance where we can visualize the collateral channel and directly block it off using coil and glue. This 74-year-old gentleman was diagnosed to have cirrhosis of liver almost two years before presentation to us. His initial presentation was with ascites and hepatic encephalopathy. The ascites responded very well to diuretics and he had not required any diuretics for over a year and a half. However, he had recurrent episodes of overt hepatic encephalopathy over the last 6-8 months. The episodes had increased in frequency and he had required at least 5-6 emergency room visits over the last 3-4 months. This was despite him being on optimal medical management. A contrast CT of the abdomen showed a large splenorenal shunt. In the absence of any other clear precipitating factors, we concluded that this shunt was likely the culprit of his recurrent episodes of overt hepatic encephalopathy. A careful study of the cross-sectional imaging was done and volume-rendered sequences were obtained in order to better understand the anatomy of the collaterals. A linear echoendoscope was then placed in the cardia of the stomach and the splenic vein was traced from its origin in the renal hilum all the way to its portal venous insertion. A dilated tortuous collateral arising from the splenic vein was visualized. It could be traced taking a convoluted course from the splenic vein to the left renal vein. A 19-gauge bevel-tipped needle was used to puncture this collateral just beyond its point of origin from the splenic vein. A 20-millimeter 0.035-inch embolization coil was placed within this area. This was followed by injection of 2 milliliters of undiluted N-butyl sinoacrylate glue just at the point of the coiling injection. We can see the gluonitis forming here with incomplete occlusion of the varix. Instead of placing another coil and putting more glue in, we decided to do this as a staged procedure as this patient had a history of ascites and we wanted to avoid precipitating a decompensation. We also thought it possible that this partial occlusion of the shunt might gradually lead to complete thrombosis of the whole splenorenal shunt due to sluggish flow. Over the next six weeks, the patient did not develop any episodes of overt encephalopathy and did not need to visit the ER. He also reported some subjective improvement in his cognition, however, this was not quantifiable. We did a transabdominal ultrasound which revealed a partially occluded splenorenal shunt, however, the flow in the shunt was maintained. As the patient was clinically doing well, we decided to wait before doing a completion procedure. Around 7th week from the procedure, the patient developed another episode of overt encephalopathy. This improved after a lactulose enema, however, this is when we decided to go for a completion procedure to block off the shunt as the patient had otherwise maintained liver functions. With the echoendoscope in the cardiac, a 19-gauge needle was used to puncture the ectatic varix just in the region of the previous gluonitis. A 20 mm 0.035 embolization coil was deployed right next to the previous coil. Another 2 ml of undiluted N-butyl sinoacrylate glue was injected in the region of the coil. The gluonitis with the post-acoustic shadowing can be clearly seen here. Here we can see the tortuous splenorenal shunt which has clearly thrombosed. These are coronal cuts of contrast enhanced CT showing the splenorenal shunt. The one on the left is the pre-procedure scan, the one on the right is the post-procedure scan where the splenorenal shunt is completely thrombosed and there is some spillover of thrombus into the left renal vein. We can see that ascites has developed post-procedure. Occlusion of photosystemic shunts can lead to hepatic decompensation even in carefully selected patients. For example, this patient had a relatively preserved liver function with a child score of 7 and male score of 10. However, he did develop ascites as a consequence of acute portal hypertension after closure of the splenorenal shunt. The advantage with doing an EUS guided procedure is that it is quite repeatable and if required can be done in a staged manner. It is important to take some time to study the cross-sectional imaging before the procedure is taken on the table. However, the procedure itself is quite quick and can be completed in 15 to 20 minutes. The cost of the procedure is also modest as compared to alternative procedures as the only accessories used are a 19-gauge needle, two coils and some cyanoacrylate glue. Moreover, the procedure time is also short. To conclude, we suggest that endoscopic ultrasound guided transgastric shunt occlusion is a good potential alternative to transvenous procedures in order to occlude large portosystemic shunts.

endoscopic ultrasound guided transgastric shunt occlusion

ETSO procedure

recurrent hepatic encephalopathy

large splenorenal shunt

collateral channel blocking