Successful reversal of chronic incapacitating post-- TIPS encephalopathy by balloon occlusion of the stent

Transjugular intrahepatic portosystemic shunt (TIPS) placement is a new technique allowing decompression of the portal system without the need for abdominal surgery or general anesthetic. This promising procedure appears safe, and is being evaluated in the context of life threatening uncontrollable variceal hemorrhage as well as ascites refractory to medical treatment. Following TIPS, portal flow diversion is associated with hepatic encephalopathy in up to 25% of patients. This is most often mild and treatable but may become uncontrollable, incapacitating and even life threatening in up to 3 to 5% of cases. The authors present two patients in whom such life threatening encephalopathy and stupor was reversed by transjugular balloon occlusion of the TIPS.

T HE MAJOR COMPLICATION OF POR- ta l hyperten sion and primary cause of death in cirrhotics remains variceal rupture.It carries a h igh morbidity an d mortali ty ( 1 ), is often associa ted with deteriora ting liver function, and may thus make t he need fo r eventual organ replacement even more imminent in patients with chron ic li ver disease.
Despite available h emostatic procedures, includ ing balloon tamponade and endoscopic sclerotherapy, the hemorrhage may be uncontrollable.A more definitive intervention is then requ ired.Unfortunately in cirrhotics with more advanced underlying liver disease, the morbid ity and mortality rates associated with shunt surgery remain prohibitive.T hese poten tial candidates need a nothe r option to prevent th em fro m succumbing to hemorrhage and to provide a 'bridge' until organ replacement (2).
Such a life-saving procedure is the transjugu lar intrahepatic portosystemic shun t (TIPS) .T he technique consists of transjugu lar placeme nt of an intravascular Stent into the liver parenchyma to con nect the portal and hepatic veins (3,4 ).T his c reates a shunt, thus decompressing the portal system without the t ion importante du debit portal qui peut etre compliquee d'une encephalopathie hepa t ique chez plus de 25 % des patients.Celle-ci est la pluparr du te mps moderee et facilement traitable, mais peut devenir incontrolable, incapacitan te, et meme mettre la vie du patient en danger clans 3 a 5 % des cas.Les auteurs presentent deux cas ch ez qui une te lle encephalopathie grav iss ime a pu etre traitee de fa~on efficace par l'occlusion du shunt a u moyen d'un ballonnet mis en p lace par voie tran sj ugu la ire.

TABLE l
Hepatic manometry pre-and post -transjugular intrahepatic shunt (TIPS) Infe rior vena c a va pressure (mmHg) Portal veno us pressure (mmHg ) Po rtohepatic gra die nt (mmHg) need fo r a gene ral anesthetic or intraabdom inal surgery.The procedure, being evaluated m several centres, shows promise.lt seems effective in a rresting he morrhage from variceal rupture a nd, in experienced hands, appears relative ly safe (5-10).In light of good initial results, TIPS is also being evaluated for use in intracta ble ascites (11 -14).
Successful portal decompression by the c reation of a portosystemic shunt is accompanied, as is its surgical counte rpart, by the development of c h ron ic e ncephalopathy in up to a quarter of patients (1 5-17).This can often be controlled with med ications; however, severe in tracta ble hepatic encephalopathy a nd/or progressive liver fa ilure occur and present a serious problem.
We descri be two patien ts in whom placemen t of TIPS for intractable ascites resulted in the developmen t of severe, intractable hepat ic encephalopathy a nd coma, a nd in whom complete reversal was achi eved by permane ntly occluding t he intrahepa tic stent shunt.

CASE 1 PRESENTATION
T he patien t was an 80-year-old male suffering fro m myelo id metaplasia, which has slowly evolved over the past 15 years.T he d isease was accompanied by portal hypertension and increasing ascites, with an associated umbil ical hern ia and important periphe ral ede ma (treated with d iuretics since December 199 1).
From March 1992, fo llowing variceal rupture, the patie nt underwent re- O n October 21, 1992, the patient was readmitted with massive ascites accompanied by a tense protuberant, excoriated um bilical hernia and massive leg edema.He was on a 'no added salt' d ie t, and his medication included aldactone 150 mg and furosemide 60 mg daily.
The patient was alert and oriented, pale, cachectic, with absen ce of icterus or stigmata of chro nic liver disease and, in particular, no en cephalopathy.He was perfectly lucid with no flapping.In itial laboratory investigations revealed: hemoglobin 90 g/L; white blood cell count 12,100 cells~mm 3 ; plate le t count 238,000 cells/mm ; prothrombin t ime/partial thromboplastin t ime n ormal; bilirubin 24 µmol/L; aspartate aminotransferase 22 IU/L; alanine aminotransferase 11 IU/L; alkaline phosphatase 546 IU/L; albumin 34 g/L; urea 25.8 mmol/L; c reatinine 254 µmol/L; sodium 134; potassium 5.3; c hloride 105; 24 h urin a ry sodium excretion 6 mmol; and creatinine clearance 0. 17 mL/s (n ormal 1.2 to 2.4).After in itial stabilization including fur ther paracentesis, it was decided to provide more defini tive treatment due to fear of potent ial umbilical hernia rupture.
O n October 29, under a nt ibiotic coverage and after mformed consent was obtained, a n intrahepatic portosyste mic stent shunt was placed via the transjugula r approach as previously described and without complications (6).Prior abdominal Doppler examination had shown a patent porto-splcno-mcsence ric ax is with heparopedal flow in all except the inferio r mesenteric vein (whe re flow was reversed ).T here was a sma ll patent para-umbil ical vein wi th hepa tofugal flow.T he coronary vei n was not visua lized.
The baseline a nd post-TIPS hepatic manometric values arc summarized in T able l.Li ver biopsy done during the procedure revealed a large nu mber of megakaryocytes with in the sinusoids as well as discre te sinusoidal fibrosis, all compa tible with the underlying diagnosis of myelo id metaplasia.
A bdominal Dopple r exami nation one week fo llowing TIPS placemen t confirmed a patent shunt between the right portal and hepatic veins.T he para-umbilical vei n was no longer patent and the inferior mesenteric vein h ad a reversed but slower flow.Bloodflow in the right portal radicals was h epatofuga l, and that in the left intrahepatic po rtal radicals was back a nd forth .Ascites were still present.
The patie nt was well for the first week post-TIPS but then developed enceph alopathy.It was ini tially m ild and treatable (grade I to II) but became progressively more severe and more diffic ult to control (grade III ) despite persistent absence of any precipitating fac tors.The patient kept fl ucruating between a grade Il and lil encephalopathy over a period of two weeks despite progressive and addit ive use of lactulose up to 15 ml qid, metronidazole 250 mg bid a nd finall y sodium benzoate 3 g tid.O n November 25, t he patie nt's level of consciousness deteriorated further and he became comatose within the ensuing three days.It was decided to sec if occlusion of the intrahepatic shunt was feasible and could achieve restoration of portal perfusion a nd reversal of the hepatic coma.
On Novemher 28, the prostheses and portal vein were cannulated via the transjugular approach, demonstrating all portal venous flow to be <liverted through the shunt.The inferior vena cava pressure was 11 mmHg and Lhe portal venous pressure was 20 mmHg, resulting in a porrohepatic gradient of 9 mmHg.The stent was blocked with an angioplasL y balloon ( 10 mm X 4 cm, Cook Cana<la) dilated to its maximum diameter using weak pressure; this resulted in an increase in portal venous pressure to 40 mmHg.Injection of contrast material into the portal vein after shunt occlusion revealed the immediate re-opacification of the right and left intrahepatic vein branches.The balloon catheter was secured at the neck for evaluation of the effect of this temporary manoeuvre on the patient's level of consciousness before proceeding to a more permanent solution.
Within 24 h the patient was fully awake.The TIPS was then occluded in a definitive fashion under antibiotic cover on November 30.Using a standard neuroangiographic technique routinely used for carotid occlusion (25 ), a latex balloon (Debrun #9, Nycomed lngenor, Paris, France), hand-tied with an elastic ligawre to a microcatheter (Tracker 10, Target Therapeutic, California), was placed within the shunt just above the portal segment of the prosthesis.It was inflated with contrast material until it was secured aga inst the wall of the stent ( 10 mm), then released in that position by traction on the microcatherer.It was shown to occlude the shunt completely on angiographic control.The procedure was we ll tolerated.The patient remaine<l completely alert and lucid despite cessation of metronidazole, so<lium benzoate and lacrulose.
The ascites re-accumulated rapidly, necessitating paracentesis one week following complete blockage of the shunt.An abdominal Doppler examination 10 days after the intervention showed the presence of a patent paraumbilical vein and hepatope<lal flow in the intrahepatic portal veins.Two and a half months later, the patient was well but require<l paracentesis (8 L) every two weeks.He had no encephalopathy an<l was nor on any medication.

CASE 2 PRESENTATION
A 65-year-ol<l male with alcoholic cirrhosis who had been abstinent since March 1991 presented with cachexia, spider angioma, increased abdominal girth and bilateral leg edema.Investigations ruled out other etiologies for the underlying liver disease.A transjugular manometry and biopsy at that time confirmed the cirrhosis with associated portal hypertension.Endoscopy showed the presence of esophageal varices grade III/IV.
Over the ensuing seven months, the patient received successively triamterene, furosemide, metolazone and amiloride for massive ascites.He required constant adjustments in diuretic therapy due to insufficient effect and increasing creatinine levels.From October 1991, the patient required paracentesis (8 to 10 Lat a time) every two weeks; the diuretics were discontinued due to development of the hepatorenal syndrome.
The patient was admitted in April 1992 for consideration of intrahepatic shunt placement.He had a moderately advanced hepatic insufficiency (Pugh score 9).Baseline values were: 24 h urinary sodium excretion 15 mEq; urea 14.2 mmol/L; creatinine 186 µmol/L; and creatinine clearance 0.44 mL/s (normal 1.2 to 2.4).An abdominal ultrasound and Doppler examination revealed the presence of a patent porto-spleno-mesenteric axis; portal flow was hepatopedal.
On April 22, a TIPS was performed.Baseline an<l po~t-T!PS manometry are shown in Table l.The procedure was well tolerated with no immediate complications.A control abdominal Doppler examination one week fo llowing the procedure confirmed a patent stent shunt between the right portal and hepatic veins.The left intrahepatic portal bloodflow was still hepatopedal whereas in the right lobe, bloodflow exhibited back an<l forth motion.
The patient required rwo subsequent shunt revisions in July and October 1992, the last one 5.5 months following the original procedure, for partial obstruction at the portal and hepatic vein extremities, respectively.These revisions resulted in a porto- Balloon occlusion of TIPS hepatic gradient reduction from 25 to 1 7 an<l finally 9 mmHg.
Three weeks after the secon<l shunt revision, the patient was admitted with an episode of confusion thought to have been precipitated by dehydration.He was rehydrated, but also required treatment with lactulose 30 mL qid and merronidazole 250 mg bid.After three weeks he was <lischarged, alert and oriented on these medications as well as on triamterene 50 mg and furosemide 40 mg daily for a moderate residual ascites.
On November 29, seven weeks after the last shunt revision, the patient was adm itted with increasing encephalopathy (grade Ill), again with no obvious precipitating cause.His level of consciousness progressively declined and he became comatose within the ensuing 72 h.Because of the catastrophic deterioration, on December 3, two months after the last shunt revision, the patient underwent catheterization of the intrahepatic shunt with a view to evaluate the effect of temporary scent occlusion on his level of consciousness.Baseline manometry revealed: inferior vena cava pressure 8 mmHg; portal venous pressure 18 mmHg; and portohepatic gradient 10 mmHg.Portography showed passage of the dye via a patent portohepatic shunt, with absence of opacification of the intrahepatic portal vein branches (Figure I).An angioplasric balloon (10 mm X 4 cm, Cook Canada) was dilated and left in p lace within the stent shunt, resulting in an immediate increase of the portal venous pressure to 39 mmHg and opacification of the previously absent intrahepatic portal veins (Figure 2).The balloon catheter was secure<l at the neck for evaluation of this temporary manoeuvre on rhe patient's level of consciousness.
Within 24 h of shunt blockage, the patient was awake, although still slow and not fully oriente<l.Because of this dramatic though partial recovery, the TIPS was occluded permanently on December 4 using a detachable latex balloon ( Debrun #9) inflated to l O mm as described in the previous case.Angiographic control confirmed complete occlusion of the shunt with assn-Figure 1) lntrahepatic stent shunt in position between the right portal and hepatic veins.Injection of contrast material in the main portal vein trunk is followed by immediate wash-out of the dye via the shunt, with total absence of opacification of any intrahepatic portal vein branches Figure 2) Upper right lntrahepatic shunt is occluded using a balloon with immediate re-opacification of intrahepatic portal vein branches Figure 3) Lower right Permanent occlusion of the intrahepatic shunt with a dilated balloon left in place within the stent ciated refilling of the intrahepatic portal branches (Figure 3).The procedure was well tolerated.Follow-up Doppler examination revealed a return of hepatopedal flow in both right and left intrahepatic portal branches.In addition, there was re-accumulation of ascites and a left pleural effusion.
The patient's level of consciousness continued to improve; he was fully oriented and lucid five days later.He was discharged on December 12 (nine days after permanent shunt occlusion) without any specific treatment for encephalopathy.
The ascites is being treated on an out-patient basis, with repeated paracentesis every two weeks.On March 10, 1993, three months after shunt occlusion, the patient presented with an esophageal variceal rupture.He was successfully treated with sclerotherapy and has embarked on monthly injections for full eradication of the varices.Encephalopathy never recurred despite cessation of all specific treatment.

DISCUSSION
We presented two patients in whom severe disabling chronic recurrent encephalopathy, occurring following creation of an intrahepatic shunt, was reversed by transjugular balloon occlusion of the stent.
TIPS placement is a promising new technique that allows portosystemic decompression by creating an intrahepatic shunt without the need for a general anesthetic or laparotomy.It can be a life-saving procedure in the face of otherwise uncontrollable variceal hemorrhage, an<l may become a useful tool in the management of intractable ascites.Surgical portal decompression with shunting of portal blood away from the liver and directly into the systemic circulation may result in chronic hepatic encephalopathy in approximately 25% of cirrhotics.This usually responds to medical therapy, but may become uncontrollable.Selective decompression with a distal splenorenal shunt does not appear to le~n the incidence of hepatic encephalopathy due to reversal of portal bloodflow (18).In addition, we are unable to predict who will develop encephalopathy following portal flow diversion (19).
As with its surgical counterpart, published reports on TIPS placement describe an incidence of hepatic encephalopathy ranging from l O to 24% (15-l7,20), with symptoms being uncontrollable, incapacitating and even potentially life threatening in 3 to 5% of cases.This is of particular concern as we gain confidence with the technique and expand its indications, offering TIPS to patients with a relatively well-preserved liver function or as an elective therapeutic alternative to those with intractable ascites who require repeated paracentesis, some before eventual liver transplantation but others mostly for a better quality of life.
Our first presented patient did not have any underlying chronic liver disease.His portal hypertension was related to a long-standing myeloproliferative disorder (21 ).Significant portal flow diversion, even in the absence of underlying liver dysfunction, can lead to severe hepatic encephalopathy as our patient clearly demonstrated.
For patients with chronic liver disease, it has been suggested that pre-TIPS Pugh score is not predictive of postprocedure encephalopathy, and that advanced age, larger shunt diameter and greater shunt flow may all contribute REFERENCES It is possible to reverse encephalopathy following surgical portal flow diversion by operative ligation of the shunt (except cn<l-to-side portacaval anastomosis) and restoration of portal perfusion (22,23).Unfortunately, this is accompanied by a high operative mortality, especially in those with associated marked liver dysfunction.Recently, Pott:.et a l (24) demonstrated the restoration of hepatorenal flow with reversal of disabling encephalopathy and improved liver function by balloon occlusion of a portorenal graft shunt without need for a laparotomy.
Faced with a life threatening situation, we attempted transjugular occlusio n of the intrahepatic shunts.The balloon occlusion method used in these two patients is a well-recognized technique for permanent occlusion of carotid arteries and carotid cavernous fistulae (25,26).When latex balloons are filled with contrast material, they slowly deflate over several months, by which time the formation of an organized thrombus has already long ensured permanent obliteration (2 7).Inadvertent detachment of a latex balloon used for occlusion of intracranial aneurysms has been described (28).The risk of such an event occurring in the present context, however, is minimal since the latex balloon, hand-tied with elastic ligatures, is difficult to detach, and must be adequately secured against a rigid tubular structure to allow it to be pulled away and released from its microcatheter.We prefer this to the comvanceal bleeding.Can J Surg 1987;30:45-50.Balloon occlusion of TIPS mercially available balloon systems that arc equipped with a self-sealing valve and deliberately easier to detach (ITC, California).In addition, should such a contrast-fi lled balloon migrate and cause occlusion of an important vessel, it can he perforated under fluoroscopic guidance using a 22 gauge needle.
Onset of chronic disabling encephalopathy after TIPS is a serious prohlem.Unfortunately, once in place and dilated, the stent cannot be reduced in size in an attempt to decrease the amount of portal flow diversion while preservmg some advantage of the shunung procedure with respect co management of intractable a:.c1tes or variceal rupture.Furthermore, in our experience, initial 'down' calibration of the shunt diameter results in insufficient flow diversion and necessity of further dilation and shu nt revisions to decompress the portal system adequately and to render the shunt efficacious.

CONCLUSIONS
The authors demonstrated the feasibility, safety and ease of shunt obliteration using t ransjugular balloon occlusion of the scents, with consequent reve rsal of the hepatic coma.In addition, such shunt occlusion <lid not induce portal vein thrombosis and thus did not endanger any eventual liver transplantation.le is paramount, however, as these cases clearly demonstrate, that very strict cncena be used to select patients for TIPS.