| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Vascular and Interventional Radiology |
1 From the Departments of Radiological Sciences (P.R., F.M.S., F.F., M.B., M.R., G.M., D.P., R.P.) and Gastroenterology (O.R.), Policlinico Umberto I, Viale Regina Elena 324, 00161 Rome, Italy. Received March 4, 2003; revision requested May 23; final revision received September 29; accepted October 28. Address correspondence to P.R. (e-mail: plinio.rossi@uniroma1.it).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Fifty-three consecutive patients underwent TIPS procedures between January 2000 and February 2002. Minimum patient follow-up was 9 months (mean, 16.3 months). Fifty-six stent-grafts were implanted in 53 patients; eight of the devices were 8 mm in diameter and 48 were 10 mm in diameter. The stent length varied from 4 to 7 cm. Indications for the procedure included recurrence of bleeding after sclerotherapy (28 patients with cirrhosis, one patient without), refractory ascites or hydrothorax (21 patients with cirrhosis, one patient without), and Budd-Chiari syndrome (two patients).
RESULTS: A technical success rate of 100% was obtained, with an early clinical success rate of 96.2%. During the follow-up period, the recurrence rate was 3.4% (one of 29 patients) for bleeding and 9.0% (two of 22 patients) for ascites. Shunt malfunction occurred in nine of 53 patients (16.9%); in one of these nine patients, shunt occlusion was evident after revision, and a parallel shunt was created. The 1-year primary and secondary patency rates were 83.8% and 98.1%, respectively. In this series, the incidence of encephalopathy (included even as a single short-lived episode) was 47.1% (25 of 53 patients). The 30-day mortality rate was 3.8% (two of 53), and the late mortality rate was 17.3% (eight of 46), excluding seven patients who underwent transplantation.
CONCLUSION: The new PTFE-covered nitinol stent-graft used appears to be excellent in preventing the need for repeated interventions. A primary patency rate of 83.8% and a secondary patency rate of 98.1% were achieved.
© RSNA, 2004
Index terms: Hypertension, portal, 95.711 • Liver, interventional procedures, 761.1269, 95.1268 • Shunts, portosystemic, 95.453 • Stents and prostheses, 95.1268 • Venography, 95.124
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
To overcome the great difficulty represented by shunt obstruction and a high bleeding recurrence rate, many experimental and clinical studies have been performed with the use of covered stent-grafts to improve the long-term patency of TIPS. The best results with the use of stent-grafts covered with polytetrafluoroethylene (PTFE) were reported by Nishimine et al (5), and these positive results were confirmed by Saxon et al (6), Haskal (7), and Andrews et al (8), as well as by our own experience.
The purpose of our study was to prospectively evaluate the use of the recently developed Viatorr stent-graft (W.L. Gore & Associates, Flagstaff, Ariz) in preventing the need for repeated interventions after TIPS.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
In this study, however, we included only those patients who underwent stent-graft placement before February 28, 2002, so that all surviving patients completed a minimum follow-up period of 9 months. Our study group, therefore, included 53 consecutive patients treated with stent-graft placement, with a mean age (±SD) of 56.5 years ± 13.8 and an age range of 16–75 years; there were 23 female patients (mean age, 62.47 years ± 7.43; age range, 39–73 years) and 30 male patients (mean age, 51.93 years ± 15.82; age range, 16–75 years).
The inclusion criteria were as follows: recurrent variceal bleeding after a session of variceal sclerotherapy (as an emergent or elective procedure), refractory ascites, or Budd-Chiari syndrome.
The exclusion criteria were as follows: severe right-sided heart failure, severe liver failure (bilirubin above 4 mg/dL [68.4 µmol/L]), polycystic liver disease, or dilated biliary ducts. Portal vein thrombosis was not considered an absolute contraindication.
The review board at our institution approved the protocol for TIPS creation with this type of stent-graft, and written informed consent was obtained from all patients.
Indications
Indications for TIPS creation included bleeding or refractory ascites in 51 patients and Budd-Chiari syndrome in two patients. Of the 51 patients with bleeding or ascites, 49 had cirrhosis and two did not have cirrhosis.
Of the 49 patients with cirrhosis, 28 underwent TIPS procedures for recurrent variceal bleeding that did not respond to or was not amenable to endoscopic therapy. Nineteen of these 28 patients were not actively bleeding and underwent the procedure electively; the remaining nine patients had an acute hemorrhage and underwent an emergency TIPS procedure. All 28 patients had already been treated with a complete cycle of sclerotherapy or banding of varices for the first episode of bleeding.
The remaining 21 patients with cirrhosis underwent TIPS procedures for refractory ascites or hepatic hydrothorax. Conditions in all of these patients had failed to respond to high doses of diuretics, bed rest, low-sodium diets, or repeated paracentesis.
Among the four patients without liver cirrhosis, two had Budd-Chiari syndrome. The third patient had portal hypertension, ascites, and bleeding secondary to resection of the right lobe of the liver for hepatoblastoma and thrombosis of the left portal branch. The fourth patient developed graft versus host disease after liver transplantation, and TIPS creation was performed for refractory ascites while the patient awaited a second liver transplantation.
Cirrhosis was the main cause of portal hypertension and was present in 49 of 53 patients; cirrhosis was secondary to viral hepatitis (type B, type C, or both) in 34 patients, excessive chronic ethanol consumption in 13 patients, and cryptogenic hepatitis in two patients. According to the Child-Pugh classification, 16 patients had class A disease (A5 in six patients, A6 in 10 patients), 24 had class B disease (B7 in four patients, B8 in 11 patients, B9 in nine patients), and nine had class C disease (C10 in five patients, C11 in four patients).
None of the 53 patients was febrile or had any laboratory evidence of a systemic infection. In patients with refractory ascites, paracentesis was performed before TIPS creation, and no bacterial peritonitis was observed.
Fifty-six of the evaluated stent-grafts were implanted in 53 patients. Eight stent-grafts were 8 mm and 48 were 10 mm in diameter; the length of the covered portion varied from 4 to 6 cm in the 8-mm stents and from 5 to 7 cm in the 10-mm stents. The mean follow-up period was 16.3 months ± 8.3 (range, 9–35 months).
Definitions
Technical success.—Technical success was defined as the creation of a shunt between the hepatic vein and the portal vein, with successful deployment of the stent-graft (9).
Clinical success.—We divided clinical success into subcategories of early and late success. We considered early clinical success to be the prompt arrest of acute variceal hemorrhage, as indicated by cessation of demonstrable gastrointestinal bleeding, decreased transfusion requirements, pharmacologic support, balloon tamponade, and return of hemodynamic stability with decompression of varices and improvement in hydrothorax and ascites (ie, progressive clearing of hydrothorax or diminution of ascites, based on clinical definitions), as estimated on chest radiographs and ultrasonographic (US) images.
For late clinical success, we evaluated the interval after discharge during which the patient remained free of the symptoms alleviated by the TIPS. For patients treated for variceal hemorrhage, this was the period after TIPS and before recurrence of bleeding.
For patients with ascites and hydrothorax, we considered late clinical success to be the complete disappearance of ascites and hydrothorax or a considerable improvement in the patient’s condition (ie, achievement of a consistent clinical status, without any worsening of symptoms). Ascites were classified at US and clinical evaluation as minimal (fluid restricted to the abdominal gutter and the area around the liver, detectable only at US), moderate (fluid in the flanks, clinically apparent), or severe (fluid in the entire abdominal cavity that results in a tense abdomen).
The follow-up period was measured in months from the date of the procedure until death, liver transplantation, or the most recent clinical examination. Patients were considered lost to follow-up if they missed two consecutive 6-month clinical examinations.
Shunt malfunction.—Shunt malfunction was suggested when a color Doppler US image showed midshunt flow velocities of less than 50 cm/sec (10) and/or in the presence of recurrent clinical signs of portal hypertension, such as ascites, varices seen at endoscopy, or bleeding (11).
All patients suspected of having shunt malfunction were followed up with portal venography. The criteria for the identification of stenoses were a more than 50% reduction of the stent lumen and/or a portosystemic gradient (PSG) greater than 12 mm Hg in symptomatic patients.
Shunt occlusion was defined as the complete absence of blood flow within the lumen at that segment. Reintervention was indicated in the presence of stent malfunction (stenosed or occluded shunts) in patients with symptoms that indicated recurrence of gastrointestinal bleeding or ascites and that did not respond to medical therapy. The treatment consisted of (a) revision, which involved mechanical dilation or repeated stent placement in symptomatic patients with stenoses greater than 50%, (b) recanalization in patients with completely occluded stents, or (c) creation of a parallel shunt.
Patency.—We adhered to the classification used by Zhuang et al (12), in which primary patency was defined as uninterrupted patency without reintervention; thus, primary patency ended at the first reintervention. Primary assisted patency was the duration of continuous patency of the shunt, with or without intervention, until complete occlusion. Secondary patency was the duration of patency after reintervention, with or without prior occlusion, until the next complete occlusion or complete loss of the shunt function.
Encephalopathy.—Hepatic encephalopathy was diagnosed in the presence of any clinically evident alterations in the patient’s mental status. It was diagnosed and graded according to the criteria defined by Parson-Smith et al (13); hepatic encephalopathy was graded as moderate (grade I or II) or severe (grade III or IV) by a hepatologist (O.R.) with 20 years of experience.
Mortality.—Early mortality included any death that occurred within 30 days after a TIPS procedure. Late mortality included the total number of deaths during the study period, with the exclusion of those patients referred for liver transplantation.
Preprocedural Evaluation
Before a TIPS procedure was performed, cardiac status, hepatic reserve, renal function, and coagulation ability were evaluated in all patients. Portal vein patency was documented at US. In selected cases, spiral multi–detector row computed tomographic (CT) angiography (Volume Zoom; Siemens Medical Systems, Erlangen, Germany) was also performed, to provide a better evaluation of liver anatomy and associated diseases.
TIPS Procedure
All TIPS procedures were performed by interventional radiologists with eight (P.R. and F.M.S.) or five (F.F., M.B., and M.R.) years of experience.
TIPS procedures were performed after deep sedation was induced in patients by using a laryngeal mask with a mixture of oxygen, nitrous oxide, and isofluorane (1%–2%) (14).
After puncture and catheterization of the jugular vein, a 12-F introducer sheath (35–40 cm long; Cook, Bloomington, Ind; Daig, Minnetonka, Minn) was advanced into the right atrium for pressure measurement and then into the inferior vena cava (IVC). After cannulation (preferably of the right hepatic vein) with a TIPS set (Angiodynamics, Queensberry, NY), the portal vein was punctured without radiologic guidance in 50 (94.3%) of 53 patients; in three patients, US guidance was necessary to reach the portal vein, which either was very small or had anatomic variations. We preferred to puncture the right portal vein 1–2 cm from its origin; on rare occasion, the left portal vein was punctured instead.
After cannulation of the portal vein and pressure measurement, the intrahepatic tract was dilated by using a low-profile Wanda balloon (Boston Scientific); the balloon was inflated to a size that corresponded with the stent diameter or, in sicker patients (older patients with more advanced disease), to only 8 mm in diameter.
Types of Stents and Stent-Grafts
De novo shunt creation.—The device that was evaluated in this study (Viatorr) is a self-expanding nitinol stent that has a high radial strength and is covered with an ultrathin expanded PTFE tube. The outer surface of the tube is a modified expanded PTFE film that minimizes transmural permeation of bile and mucin, and thus inhibits hyperplastic tissue ingrowth, and that provides radial reinforcement to the inner surface. The intraluminal surface is made of expanded PTFE from the same manufacturer.
This stent-graft has a characteristic structural design. It is composed of a 2-cm-long uncovered portion for the portal region and an expanded PTFE-covered portion for the intrahepatic tract. A gold marker band indicates the interface between the covered and the uncovered areas. Moreover, a smaller gold marker is incorporated into the proximal end of the covered portion to facilitate fluoroscopic imaging during stent-graft deployment.
The stent-graft is mounted on a 10-F device; the covered portion is secured by a constraining expanded PTFE sleeve, while the self-expanding uncovered portion is secured by an introducer sheath.
This stent-graft is available commercially in three different diameters (8, 10, and 12 mm), and the length of the covered portion ranges from 4 to 7 mm.
Revisions.—In some patients with shunt malfunction, it was necessary to place an additional stent to extend the PTFE stent-graft to the IVC. In those patients, we used a Jostent device (Jomed, Beringen, Belgium) instead of the Viatorr, because the Jostent is a balloon-expandable stent-graft that has a very high radial strength and no bare portion, is also covered with expanded PTFE, and is easier to deploy in the correct position. This stent has a "sandwich" design, in which a thin layer of expanded PTFE is placed between two stainless steel (316 L) stents; it is commercially available in different lengths, which range from 12 to 58 mm, and in two different diameters (4–9 mm or 6–12 mm).
In some cases in which there was stenosis at the level of the portal region, revisions were performed by using available self-expanding bare stents (Wallstent [Boston Scientific] or Dynalink [Guidant, Menlo Park, Calif]). In one case, a second Viatorr stent-graft was deployed to completely cover the stenotic intrahepatic tract that was not covered by the Viatorr device positioned 9 months earlier.
Stent diameter reduction.—In patients with refractory hepatic encephalopathy assessed by a trained hepatologist (O.R.), the stent diameter could be reduced to permit an increase in portal blood flow to the liver (15–17). The reduction of the TIPS diameter was achieved by deployment of a Jostent device within the shunt and by dilation of the device to an hourglass shape. To obtain such a shape, the stent is mounted on a balloon with a ligature in its center so that, at inflation, the proximal and distal portions of the balloon will be dilated to the nominal size of the balloon, while the central portion will not, because of the tight ligature. The second stent will therefore have an hourglass appearance. The central portion of the stent can then be dilated to the desired size with a new balloon (5-, 6-, or 7-mm balloon).
Stent-graft length selection.—The Viatorr device provides the best results when its length is correctly selected. The entire length of the intrahepatic tract should be covered by the stent-graft, from the portal vein to the junction of the hepatic vein and IVC (7).
The length of the intrahepatic tract can be evaluated at angiography performed with a marked catheter. In our experience, the easiest way to measure the distance between the portal vein and the junction of the hepatic vein and IVC is to visualize both the hepatic and the portal veins with a simultaneous double injection of contrast material administered through an angiographic catheter in the portal vein and through an introducer sheath in the hepatic vein. An 8-cm-long balloon is sometimes used for measurement.
It is, however, very difficult to measure the exact length of the hepatic parenchyma and vein to be covered by the stent-graft, since a deployed stent has a different position within the tract compared with the catheter. For these reasons, after the initial difficulties, the graft length in our study was selected to be 1 cm longer than the measured length of the intrahepatic tract.
Stent-Graft Deployment
The deployment of the Viatorr device is performed in two steps. The first step is the release of the uncovered portion. The 12-F introducer sheath should be advanced at least 3 cm into the portal system and then gently retracted to permit the release of the uncovered portion of the stent-graft within the portal vein. It should be emphasized that an incorrect release of the bare portion within the hepatic tract cannot be modified. Once the stent is fully released, the whole system is gently retracted until resistance is felt, which indicates that the proximal portion of the bare stent has reached the junction of the portal vein within the intrahepatic tract.
The second step of deployment is the release of the covered portion. After the correct position of the uncovered portion in the portal vein has been confirmed, the introducer is removed far above the beginning of the stent and, while the device is held in place, the covered portion of the device is released by pulling the string. The stent-graft is then dilated with the balloon catheter used for the dilation of the tract; dilation is 8 mm for the 8-mm stents and 10 mm for the 10-mm stents. After deployment of the stent-graft, the mean atrial and portal pressures are measured again to confirm the hemodynamic success of the procedure.
Medications
An intravenous antibiotic (ceftriaxone, Rocefin; Hoffmann-La Roche, Kaiseraugst, Switzerland) treatment regimen was started the day before the procedure and was continued for 48 hours. After the procedure, in those patients with normal coagulation parameters, prophylactic heparin anticoagulation therapy was prescribed to reduce the incidence of shunt occlusion. This is a widespread practice in Europe, especially among gastroenterologists, who usually oversee patient care after TIPS procedures. This therapy, however, has not been scientifically proved. During the procedure, 5,000 IU of heparin was injected after stent-graft deployment. All patients underwent postimplantation therapy with 12,500 IU of heparin diluted in 500 mL of saline solution for the first 24 hours; thereafter, 0.5 mg of low-molecular-weight heparin was administered twice a day for at least 1 week.
Follow-up
After TIPS, all 53 patients were followed up for a mean period of 16.3 months (range, 9–26 months) according to a protocol established in cooperation with the gastroenterologist (O.R.).
Clinical and laboratory tests were performed before patient discharge, at 1-month follow up, and every 3 months thereafter.
Color Doppler US for evaluation of flow direction in the portal branches, patency of the hepatic veins, stent flow velocity, and diameters of both the stent and the portal vein was performed 6–7 days after the TIPS procedure (before patient discharge), at 3 months and 6 months after discharge, and every 6 months thereafter. All US examinations were performed by two different operators with more than 7 years of experience each.
It should be noted that color Doppler US performed in patients less than 48–72 hours after the TIPS procedure may provide false results. In fact, in the first few days after the procedure, there is an acoustic barrier that may not permit exploration of the shunt lumen because of the presence of microbullae embedded inside the expandable PTFE (18).
We consider stent malfunction to be present when the midshunt velocity is less than 50 cm/sec (10); therefore, a portal venogram with measurement of the PSG should be obtained.
In those patients treated for bleeding, upper gastrointestinal endoscopy was performed 1 week after the procedure and then every 6 months thereafter, to evaluate the presence of varices (11). (All endoscopy was performed by O.R. and members of the gastroenterology department with more than 15 years of experience.)
In our study we were able to obtain all the data described here in all patients.
Statistical Analysis
Results are expressed as mean ± SD. One-year primary patency and survival rates were calculated by using the Kaplan-Meier method, and the resultant curves were compared by means of the log-rank test. The association between the incidence of encephalopathy and PSG was calculated with the paired t test, while the association between incidence of encephalopathy and stent diameter was studied with the 
2 test. A logistic regression analysis was then performed for the three variables. Moreover, the Kaplan-Meier analysis and 2 test were performed to evaluate whether patient age and sex affect the patency rate, survival rate, and encephalopathy incidence. Statistical significance was established with a P value of less than .05, and calculations were performed by using a statistical software program (Number Cruncher Statistical System, version NCSS 2001; NCSS, Kaysville, Utah).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
In one patient, rupture of the releasing system was caused by overly forceful insertion of the stent-graft via the introducer. The uncovered portion of the stent-graft was released within the liver parenchyma, but the device was successfully retrieved, with some difficulty, by snaring and extracting it. The partially released stent, once snared, was pulled from the hepatic tract into the superior vena cava; however, removal through the jugular vein was considered too traumatic to be attempted because of the large diameter of the released bare portion, which would not collapse. We therefore used a femoral approach, and, with use of a 16-F introducer, the stent-graft was snared in its most distal portion, thus permitting removal. Despite this initial difficulty, the procedure was completed with the insertion of a new stent-graft, and a favorable outcome followed.
In four patients, extrahepatic portal vein puncture with intraabdominal extravasation of contrast media was treated with the rapid deployment of a stent-graft, which immediately arrested the intraabdominal bleeding. In one patient treated for variceal bleeding, the stent-graft was implanted by means of a left portal vein puncture. One week after implantation, we observed an obstruction of the left biliary duct caused by compression from the stent-graft, with an increased bilirubin level. This was treated with implantation of a plastic biliary endoprosthesis, and symptoms resolved completely until the patient’s death 19 months later.
Throughout the follow-up period, no hepatic infarction due to a large stent occluding the hepatic vein was detected at clinical or US follow-up examinations in any of our patients.
Evaluation of PSG
After TIPS placement, the mean PSG value decreased from 23.3 mm Hg ± 6.5 (range, 13–42 mm Hg) to 7.13 mm Hg ± 3.1 (range, 2–16 mm Hg). In two patients (3.8%) treated for refractory ascites, the PSG remained slightly higher (16 mm Hg in one patient and 15 mm Hg in the other) despite the use of 10-mm stents. At 1-month follow up, however, ascites had completely disappeared in both patients, which confirmed the successful decompression of the portal system.
Clinical Success
Early.—Early clinical success was achieved in 51 (96.2%) of 53 patients. In two patients, despite immediate technical success, no corresponding improvement was evident in their clinical conditions (persistent ascites and esophageal varices) before discharge. Therefore, after endoscopy, venography and PSG measurement were performed. In these two patients, the stent-graft, which had been positioned a few days before, was too short to cover the intrahepatic portion. A stenosis was present, which resulted in an increase in the PSG. An additional covered stent was implanted in each patient to complete the procedure and extend the covered portion to the IVC junction. Both patients recovered very well, and the clinical signs of portal hypertension disappeared.
Late.—Of the 29 patients treated for bleeding (28 with cirrhosis, one without cirrhosis), only one patient (3.4%) had recurrence of bleeding. This patient returned after 3 months with minimal recurrence of bleeding, not from varices, but because of congestive gastropathy. This patient had a PSG of 16 mm Hg. After balloon dilation of a short stenosed segment of the uncovered hepatic vein, the PSG returned to 8 mm Hg and the patient stopped bleeding. Among 22 patients whose indications for TIPS were ascites and hydrothorax (21 patients with cirrhosis and one without), 20 were treated for ascites only and two were treated for a combination of ascites and hydrothorax; 19 (86.3%) of 22 had an excellent response, with complete resolution or improvement of ascites, while in three patients we observed a minimal persistence of ascites. No increase in ascites was observed in any patient. The recurrence rate was 9.1% (two of 22 patients) at 11 and 13 months after treatment. In both patients, portal venography was performed, and the PSG was 21 mm Hg in one patient and 19 mm Hg in the other.
Follow-up
Shunt malfunction requiring reintervention.—After a mean follow-up period of 16.3 months ± 8.3, nine (16.9%) of 53 patients needed reintervention for a shunt malfunction detected at color Doppler US, endoscopy, or clinical examination (presence of varices or ascites). Stenoses or intraluminal defects of the covered portion of the stent-graft were not identified as the cause of malfunction in any patient.
Thirteen reinterventions were necessary in nine patients, as follows: one revision in each of six patients, two revisions each in two patients, and three reinterventions in one patient, including one recanalization and one creation of an additional shunt.
In an attempt to classify shunt malfunctions according to their cause, we selected three main groups, as follows:
1. The first cause is a stent of insufficient length, which does not completely cover the intraparenchymal tract. This was observed in three patients (three revisions). In two of these patients, the stent-graft did not reach the hepatic vein wall because of foreshortening, and a stenosis was evident during hospitalization 6 and 8 days after TIPS placement. Both procedures were completed with the placement of an additional covered stent (Jostent), and symptoms resolved. In the third patient, the short (10-mm x 5-cm) stent-graft (Viatorr) was originally released with the bare portion inside the intraparenchymal tract at the portal level, to allow the covered portion to reach the hepatic vein (Fig 1a). After 9 months, the patient returned with a stenosis that involved the bare portion of the stent within the intraparenchymal tract (Fig 1b). This was successfully treated by positioning a new Viatorr device (10 mm x 7 cm), correctly deployed, inside the previous one and thus covering the full length of the intrahepatic tract (stent-within-a-stent technique) (Fig 1c). The PSG in these patients decreased from a mean of 16.3 mm Hg ± 6.0 to a mean of 6.3 mm Hg ± 0.5.
|
|
|
Four of the five patients with hepatic vein stenosis underwent placement of an additional covered stent (Jostent) (Fig 2). In the fifth patient, who had repeated episodes of bleeding from congestive gastropathy, the increase in the PSG (to 16 mm Hg) was secondary to a kinking of the hepatic vein before reaching the IVC. A 10-mm balloon dilation was sufficient to reduce the PSG from 16 to 8 mm Hg, with clinical improvement and resolution of bleeding. The PSG in these five patients decreased from a mean of 16.4 mm Hg ± 4.7 to a mean of 8.6 mm Hg ± 3.7.
|
|
|
|
|
|
|
|
2 test, P = .88; identical values for logistic regression analysis). In 13 of 25 patients, the PSG values after shunt creation ranged from 8 to 12 mm Hg, and in the remaining 12 patients, they were below 8 mm Hg (below 4 mm Hg in five patients, below 6 mm Hg in three patients, and below 8 mm Hg in four patients).
Stent reduction.—In 22 of 25 patients, encephalopathy was successfully treated with lactulose and dietary restriction. Shunt reduction was required in the remaining three cases (5.6%), in which the patients had severe encephalopathy that was refractory to the medical treatment and in one case was associated with cardiac failure. Shunt reduction was performed after 3, 4, and 28 months by using a covered stent (Jostent) partially dilated inside the shunt, with an immediate increase in PSG from 6 to 21 mm Hg, from 9 to 16 mm Hg, and from 10 to 17 mm Hg in the three patients, respectively. Despite the increased pressure, all the patients did very well, and there were no signs of recurrent encephalopathy, ascites, or bleeding prior to the conclusion of the study (November 2002).
Orthotopic liver transplantation.—In our series, seven patients underwent orthotopic liver transplantation without procedural difficulties. An initial belief that longer stents would cause difficulty during surgery was later discarded because the covered portion of the stent-graft was not attached to the liver parenchyma and was easily displaced for clamping.
Histopathologic analysis of explanted Viatorr stent-grafts showed that they were patent and lined internally with a thin layer of fibrin. No signs of endothelialization or intimal hyperplasia were observed in the covered portion of these devices, and they were easily removed from the liver parenchyma.
The uncovered portion of the stent was well attached to the portal vein wall and, in a few cases, was considered too long because it crossed the line of portal venous section, which rendered the procedure more difficult (Fig 6).
|
Late mortality.—Excluding the seven patients who underwent transplantation, the survival rate was 82.6% (38 of 46 patients) (Fig 7) and the total mortality rate was 17.3% (eight of 46 patients) with a mean 16.3 months of follow-up. Two patients who were treated electively died within 30 days of the procedure; the other six patients died of worsening liver function and multiorgan failure (five patients [83%] within the first year). Among these eight patients, five underwent TIPS procedures for refractory ascites and three underwent TIPS procedures for recurrence of bleeding.
|
2 test. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
To obtain the best results with expanded PTFE-covered stent-grafts, and to avoid hepatic vein stenosis, it has been suggested that the device should extend from the portal vein to the IVC junction (7,18,19). In an experimental study, Haskal et al (20) reported that two of eight pigs treated with a flexible PTFE stent-graft developed mild hepatic vein stenosis. A third pig developed a stent occlusion from cellular overgrowth of the hepatic vein. Occlusion was also partly related to the rapid axial growth of the pig, which resulted in foreshortening of the stent. In 2001, Cejna et al (21) reported their initial experience with 16 cases of Viatorr stent implantation, with a 62% (10 of 16 patients) incidence of hepatic vein stenosis above the grafted portion, which required the placement of a new stent. In an article published in 2002, Otal et al (18) reported an experience with 15 de novo TIPS procedures and five revisions in which Viatorr stents were used. In one patient, a hepatic vein stenosis was treated with the addition of an uncovered stent.
The correct determination of stent length is key to achieving a longer patency; when the stent is too short, stenosis can be expected in the hepatic vein. This was confirmed in our experience. When we compared the incidence of malfunctions in those 23 patients in whom the hepatocaval junction was reached by the stent-graft, we found that only two patients needed a revision at the portal vein level, while seven of 30 patients in the group with too short a stent-graft required a revision. Although this difference may not be statistically significant, we believe that stent length is an important factor to be considered in efforts to reduce the incidence of reintervention, the length of hospital stay, and the total cost of the procedure.
In our first experiences with covered stents in 2000, we were unable to consider the length of the stent-graft such a vital factor in the outcome of the procedure for the following reasons: (a) our lack of full belief in the suggestions made by Haskal et al (20) and (b) the initial lack of availability of the Viatorr stent in the right diameters and lengths to fulfill the needs in every patient at the time of the procedure.
It is difficult to measure the exact length of a curved path from the portal vein to the IVC junction. In small hepatic veins, the use of a long covered stent-graft may cause hepatic infarction by blocking the hepatic venous return from the segment of the hepatic vein territory. Although this never occurred in our patients, in a study by Otal et al (18), two cases of segmental liver ischemia, detected at 24 hours and 1 week, respectively, after the TIPS procedure, were found at CT but were not clinically suspected.
Shunt Malfunction
The most important finding in our study was that, despite a 16.9% prevalence of stent malfunction, in no cases were there stenoses or intraluminal defects in the covered portion of the stent-graft.
The major cause of malfunction in the evaluated stent-grafts in our series was hepatic vein stenosis (five [9.4%] of 53 patients), which was a more frequent finding here than in patients treated with bare stents (three [3.5%] of 87 patients, based on laboratory data from our previous experience). The hemodynamic changes in the blood flow in hepatic veins may be responsible for hepatic vein intimal hyperplasia, since the blood flow within a covered stent is higher because of the absence of intrastent stenosis, which is usually present in bare stents. The extension of the covered portion of the stent-graft along the whole tract from the portal vein to the junction between the hepatic vein and the IVC represents a key point to be considered in order to avoid shunt dysfunction.
All reinterventions performed because of insufficient stent length should be considered technical errors caused by the operator and not the stent design. In the future, with the placement of stents of the correct length, the reintervention rate should decrease drastically.
Those difficulties that occurred at the portal vein level, however, were not secondary to the stent length selection. In the four cases in our study, one was due to intimal dissection of the portal vein, two were probably due to a previous partial thrombosis in the portal vein, and in the remaining case there was no apparent reason. The stent should probably be modified, since its bare portion appears too stiff and is probably too long, which may lead to intimal injury or stent overextension, especially in patients with a short portal vein.
Recurrence of Bleeding
In all randomized trials of TIPS versus sclerotherapy, one of the end points has been the rate of bleeding recurrence as a measure of the ability of those procedures to prevent a recurrence of bleeding (22–26). In our experience, TIPS creation with the Viatorr stent-graft appears to be the most efficient therapeutic procedure for preventing bleeding recurrence. In the 29 patients treated for bleeding in our study, only one (3.4%) returned with recurrent bleeding after the procedure.
In reviewing the literature, we found one case of bleeding recurrence reported by Cejna et al (21) among 16 TIPS procedures performed with the Viatorr device, and in a series of 20 patients by Otal et al (18), no bleeding recurrence was reported.
In comparison, TIPS creation with bare stents is significantly more effective than sclerotherapy (P = .026) in preventing a recurrence of bleeding (22), but TIPS creation with bare stents (3.5% [based on data from our previous experience]) is not as effective as it is with the Viatorr stent-graft. In a recent publication, Zhuang et al (12) reported a 32.26% incidence of bleeding recurrence among 92 patients treated with bare stents. Domagk et al (27) reported on a recent multicenter study, in which use of a self-expanding wire mesh stent in 139 patients with repeated episodes of variceal bleeding resulted in a bleeding recurrence rate of 14.4%.
Encephalopathy
The incidence of encephalopathy reported in the literature varies widely, from 3% to 75% (3,28,29). We can, however, consider that an average of 32% of patients undergoing a TIPS procedure will experience an episode of acute encephalopathy (30).
In our experience, hepatic encephalopathy occurred during the follow-up period in 25 (47.1%) of 53 patients. To reduce the frequency, we generally performed TIPS procedures by using only stent-grafts with a diameter of 8 or 10 mm. The 10-mm stent represents the best compromise between PSG reduction and good blood flow; moreover, if the stent proves to be too large, its lumen can be reduced with an additional stent. In no case was the 10-mm stent-graft inadequate to decompress the portal vein.
In evaluating our results, we found an equal prevalence of encephalopathy in patients treated with an 8-mm stent and in patients treated with a 10-mm stent. We recommend the use of 8-mm stent-grafts in older patients with poor liver function and cardiovascular problems.
In assessing the decrease in PSG in our patients with encephalopathy, we found that the PSG was below 12 mm Hg in all 25 patients, regardless of whether an 8- or a 10-mm stent was used. However, in 13 of 25 patients, the PSG ranged between 2 and 8 mm Hg, while in the remaining 12 patients, the PSG ranged between 8 and 12 mm Hg. The decrease in the PSG, therefore, is not predictable.
The majority of these patients will be free of encephalopathy within 1 month after TIPS creation, if they adhere to the medical therapy and dietary regimen. With the use of bare stents, shunt narrowing due to intimal hyperplasia is frequent, and encephalopathy represents only a minor problem. With the new covered stents, however, the shunts will not easily accommodate a narrowing of the lumen; the incidence of encephalopathy increases, with a higher number of cases refractory to the medical treatment, especially after 2 or 3 years, when hepatic function may decrease.
The rationale in treating portal hypertension with TIPS creation is to have the PSG decrease to less than 12 mm Hg after the TIPS procedure, since 12 mm Hg is considered a threshold for the risk of bleeding recurrence (31–33). However, in a study by Casado et al (34), as in our experience, all cases of encephalopathy developed with a PSG 12 mm Hg. Therefore, for a patient to be completely free of the risk of bleeding recurrence, we have to accept a high frequency of encephalopathy.
In a TIPS review published in 2002, Bilbao et al (35) referred to encephalopathy as the "price to pay" for a functioning shunt. It is our opinion that TIPS should bring about an improvement in the patient’s physical condition, but not at such a high price.
In a recent article, Rössle et al (36) questioned whether it is necessary to decrease the PSG to less than 12 mm Hg after TIPS, especially in sicker patients in whom, with a high shunt volume, there is decreased liver perfusion and increased encephalopathy and liver dysfunction. According to these authors, this threshold should be based on individual needs and not on a cohort observation (31–33). In their article, Rössle et al (36) reported a study of 225 patients hospitalized after TIPS procedures for recurrence of bleeding or for shunt revision motivated by the endoscopic finding of varices. In these patients, the authors found that the PSG at the time of revision was almost identical to the PSG measured during the TIPS procedure. On the basis of their calculations of the probability of bleeding recurrence after revision, they suggested that with a reduction in the initial PSG to 50%, the risk for bleeding recurrence should be no more than 1%; if the reduction is only 25%, the probability of bleeding recurrence will be around 7%.
This should not apply to those patients with a first variceal bleeding episode and a PSG of 16 mm Hg, a finding that suggests a fragility of the variceal walls (36). In these cases, therefore, the pressure should be lowered to less than 12 mm Hg.
Moreover, if we consider that a good percentage of patients (21 [39.6%] of 53 in our series) were referred for TIPS only for correction of refractory ascites, then—if we want to avoid encephalopathy—we have to contemplate the possibility in these patients of a postprocedural PSG considerably higher than 12 mm/Hg without any risk of bleeding.
This possibility is also confirmed by the outcome in three cases in which we needed to reduce the shunt diameter because of encephalopathy after TIPS creation; this reduction was achieved by deploying an additional stent, partially dilated, within the previous stent (16,17). In these three cases, the pressure rose from below 12 mm Hg to 17, 21, and 16 mm Hg, respectively, with complete relief of encephalopathy and no recurrent ascites or bleeding.
These observations should encourage us to evaluate each patient on an individual basis and, probably, to initially dilate the stent to a minimal diameter and then dilate it to a larger size if there is no clinical improvement. We have to make every effort to prevent encephalopathy refractory to medical treatment, since patients with this condition cannot have a normal lifestyle and may become a burden for their family and society.
For this reason we believe that, in sicker patients, we should initially attempt to dilate the intraparenchymal tract to a diameter of 6 or 7 mm and only later increase it to 8 or 10 mm if a full portal decompression does not occur.
The problem arises as to which stent should be used; balloon-expandable stents can be progressively dilated as needed but are stiff and do not follow the curved path from the hepatic to the portal vein, while self-expanding covered stents (like the Viatorr device) have a fixed diameter. Although there is a remote possibility of overdilating both types of stents, there is no way to reduce their caliber, except by placing another stent within the first stent or by initially dilating the intrahepatic tract to a smaller diameter, thus hoping that the radial force of the stent-graft does not produce a progressive dilation within a few days after placement.
A third option is represented by a newly designed stent that, to our knowledge, is not yet commercially available but has the dual characteristics of being self-expanding to 6–7 mm and then balloon-expandable to 8–10 mm.
Another consideration in evaluating patients for TIPS procedures is the presence of those factors described as predictors of postprocedural encephalopathy. These factors are as follows: advanced age and female sex; nonalcoholic liver cirrhosis with hypoalbuminemia, Child-Pugh class B or C; previous episodes of encephalopathy; or recent gastrointestinal bleeding (37).
Study Limitations
One limitation of this study is that the findings are not compared with those from a study of bare stents performed by us before the advent of covered stent-grafts, although comparative information could be obtained from the literature. In addition, the follow-up period in our study, although fairly long, should be prolonged to 3–4 years to obtain confirmation of our early results.
The expanded PTFE-covered stent-graft represents an important advance in prolonging TIPS patency and reducing the bleeding recurrence rate, which previously was the critical limiting factor of this technique. The primary patency rate of 83.8% and the secondary patency of rate 98.1% confirm the importance of the Viatorr device. This stent can also be easily removed by the surgeon during orthotopic liver transplantation. However, some modifications should be made to make the bare portion of the device shorter and less stiff.
A technical success of 100% and a clinical success of 96% indicate that the stent is easily implanted and, when the correct size is selected, leads to excellent clinical results. In seven of nine patients who required reintervention, the difficulty was caused by a stent that was too short, which should be considered a human error and not a stent failure. Therefore, we recommend careful selection of the stent-graft length. Encephalopathy occurs at a higher rate with stent-grafts; this may be due to greater shunt efficiency, which could be limited with a smaller shunt diameter. Further studies are needed to determine methods of limiting encephalopathy.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, P.R.; study concepts, F.F., P.R., M.B., M.R.; study design, P.R., F.F.; literature research, D.P., G.M.; clinical studies, F.M.S., F.F., D.P., O.R., M.R.; data acquisition, P.R., F.F., M.B.; data analysis/interpretation, F.F., O.R.; statistical analysis, G.M.; manuscript preparation, F.F., G.M., P.R.; manuscript definition of intellectual content, M.R., P.R., F.M.S., F.F., M.B.; manuscript editing, G.M., D.P.; manuscript revision/review, P.R., R.P.; manuscript final version approval, G.M., P.R.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
