Background
Recurrent or refractory hepatic encephalopathy in patients with cirrhosis can be difficult to manage. Liver transplant is the most effective modality, but may be precluded by financial, logistical, or organ availbility limitations. Many of these patients have demonstrable large portosystemic shunts, occluding which could lead to clinical improvement. However, this may lead to new onset/worsening hepatic decompensation due to acute portal hypertension in patients with poor liver reserves

Case Presentation
A 45 year old male with ethanol related cirrhosis was referred to us in view of recurrent hepatic encephalopathy(HE) for over 1 year. He had multiple breakthrough episodes of overt HE despite optimally dosed lactulose, rifaximin, and branched chain amino acids. He had persistent asterixis, severe slurring of speech. He also had facial dyskinesias, which after extensive neurological workup, were attributed to atypical extrapyrimidal symptoms of HE.

Bloodwork showed pancytopenia, bilirubin 3.5mg/dL, albumin 2.5g/dL, INR 2.1, and MELD score was 23. He was advised liver transplantation, however, this was not an option due to financial constraints. A contrast enhanced tomography of the abdomen showed a large splenorenal shunt arising from a short gastric vein. Shunt occlusion via balloon retrograde transvenous obliteration was considered, but considering the poor liver parameters, there was high risk of worsening hepatic decompensation.

We realized that occluding the splenic vein near its origin from the portal vein would create a porto-systemic split, prevent shunting of mesenteric venous blood (which contained gut derived toxins-ammonia, lipopolysaccharides) to the systemic circulation, and force it through liver parenchyma. Moreover, the splenic outflow would be maintained by the splenorenal shunt and not cause additional congestion of the liver.

Endoscopic Methods
With a linear echoendoscope in the gastric cardia, the splenic vein was targeted behind the neck of pancreas. Three 20mm 0.035” embolization coils were injected followed by 4ml of undiluted n-butyl-cyanoacrylate glue in 2 aliquots. This led to reduction of flow in the splenic vein from 25cm/s to <10cm/s. A large glue cast was documented covering the splenic vein, and maintained splenic outflow via the splenorenal shunt was documented.
The patient tolerated the procedure well, with no immediate post procedure adverse events. At 1 and 3 month followups, he showed good clinical improvement- no further episodes of overt encephalopathy, imporved cognition, minimal asterixis, and marked improvement in extrapyramidal symptoms

Conclusions
EUS guided portosplenic split can be a potential therapeutic strategy in some patients with recurrent hepatic encephalopathy with poor liver reserve and no transplant options. Prospective, long term data is needed to establish its place in clinical scenarios

Ritu Shree

Appearances