Efficacy of covered and bare stent in TIPS for cirrhotic portal hypertension: A single-center randomized trial

We conducted a single-center randomized trial to compare the efficacy of 8 mm Fluency covered stent and bare stent in transjugular intrahepatic portosystemic shunt (TIPS) for cirrhotic portal hypertension. From January 2006 to December 2010, the covered (experimental group) or bare stent (control group) was used in 131 and 127 patients, respectively. The recurrence rates of gastrointestinal bleeding (18.3% vs. 33.9%, P = 0.004) and refractory hydrothorax/ascites (6.9% vs. 16.5%, P = 0.019) in the experimental group were significantly lower than those in the control group. The cumulative restenosis rates in 1, 2, 3, 4, and 5-years in the experimental group (6.9%, 11.5%, 19.1%, 26.0%, and 35.9%, respectively) were significantly lower (P < 0.001) than those in the control group (27.6%, 37.0%, 49.6%, 59.8%, 74.8%, respectively). Importantly, the 4 and 5-year survival rates in the experimental group (83.2% and 76.3%, respectively) were significantly higher (P = 0.001 and 0.02) than those in the control group (71.7% and 62.2%, respectively). The rate of secondary interventional therapy in the experimental group was significantly lower than that in the control group (20.6% vs. 49.6%; P < 0.001). Therefore, Fluency covered stent has advantages over the bare stent in terms of reducing the restenosis, recurrence, and secondary interventional therapy, whereas improving the long-term survival for post-TIPS patients.

As shown in Fig. 4, covered stent (experimental) in TIPS exerted a significant protective effect on patients' long-term overall survival (with 5-years follow up) compared to the bare stent (control) (log-rank test, p = 0.009). The median survival time of the experimental group was prolonged than the control group.

Discussion
The short-term efficacy of TIPS in treating the complications caused by portal hypertension has been widely acknowledged [15][16][17][18][19][20][21][22][23] . However, the long-term efficacy of TIPS is restricted due to the high incidence of shunt channel restenosis and hepatic encephalopathy 24 . We compared our results with other reported studies as shown in Table 4. Shunt channel restenosis could induce the recurrence of gastrointestinal bleeding or refractory ascites, and thus further affect the survival time. Application of covered stents could significantly reduce the incidence of shunt channel restenosis. A multicenter prospective study on 114 patients receiving Viatorr covered stents found that restenosis rate at 6, 12, and 24 months was 8.1%, 20.1%, and 24.1%, respectively 12 . Another study showed that the restenosis rate in 1, 2, or 3-years after applying polytetrafluoroethylene (PTFE) covered stents was 10%, 16%, and 26%, respectively 25 . In a retrospective study 26 Table 2. Recurrence rate of gastrointestinal bleeding and refractory hydrothorax/ascites, incidence rate of hepatic encephalopathy and secondary interventional therapy.
reported the primary patency rate of 76% and 36%, respectively (p = 0.001); clinical relapse recurrence rate of 10% and 29%, respectively (p < 0.05); free of encephalopathy rate of 67% and 51% (p < 0.05), respectively; and 2-year survival rate of 58% and 45%, respectively, for bare stents and Fluency covered stents. The 2-year patency rate of 76% and 36% (p < 0.001), clinical symptom recurrence rate of 10% and 29% (p < 0.05), and mortality rate of 58% and 45% (p < 0.05) for bare stents and Fluency covered stents, respectively; suggesting that Fluency covered stents could increase the long-term patency rate without affecting the mortality or incidence of hepatic encephalopathy and liver failure, and thus increase the middle-and long-term efficacy of TIPS.
In the present study, the cumulative restenosis rate in 1, 2, 3, 4, or 5-years in the experimental group was significantly lower than that in the control group, suggesting that covered stent can reduce both the short-and long-term restenosis rate compared to the bare stent. The recurrence rate of repeated gastrointestinal bleeding and refractory hydrothorax and ascites was significantly lower in the covered stent group than in the bare stent group. Notably, the 4-and 5-year survival rate was also higher in the covered stent group than in the bare stent group. These results were in accordance with other reports, especially the one using Viatorr stent specific for TIPS. Compared to the previous findings, the longer follow up time and the randomized controlled trial design of the present study made the results more convincing and with better comparability.
Viatorr stent applied 3 layers of PTFE with different pore-diameters of specifically processed degradations, which could completely block the leakage of bile into the stent lumen, prevent liver tissues from growing into the stent mesh, effectively cover the hepatic vein at the puncture site, and thus reduce the postoperative restenosis. Viatorr stent consists of two parts: a self-expandable metallic stent (a 2 cm bare stent for the implantation in the portal vein) and a PTFE-covered stent. A metal ring, which can be observed in imaging, is used to separate the covered and uncovered parts to help intraoperative positioning. Fluency covered stent, like the Viatorr stent also used PTFE. Two layers of PTEF cover are used in Fluency stent; the inner layer consists carbon to prevent the aggregation of platelet. Although the cover is not exactly as the same as that in Viatorr stent, the function of preventing shunt channel restenosis is similar; furthermore, the short-term efficacy of Fluency covered stent are similar to Viatorr stent, according to large-scaled international studies 13,14 . The findings of   the present study found that the long-term survival rate in the Fluency covered stent group was significantly higher than that in the bare stent group. Another important factor that could affect the long-term efficacy of TIPS is the development of postoperative hepatic encephalopathy. Current opinions indicate that the mechanism involved in this complication is central nervous system dysfunction caused by shunting induced or aggravated increase of nitrogen substances (e.g., blood NH3/N H4 + ). The development of hepatic encephalopathy is closely associated with the diameter of the shunt; in another word, the diameter of the shunt may determine the susceptibility to hepatic encephalopathy. A large diameter may lead to increased shunt volume, and the nitrogen substances may reduce the tolerance capacity of the central nervous system and thus induce hepatic encephalopathy [29][30][31][32][33][34][35][36][37][38][39] . Sarfeh 40,41 and Rypinh 42 also demonstrated that higher shunt volume in H-shaped shunt could increase the risk of hepatic encephalopathy. However, the incidence rate of hepatic encephalopathy after TIPS using covered stents ranged greatly (from 14.1% to 47.1%) among different studies [43][44][45][46] . Several studies showed that the incidence rate of hepatic encephalopathy in bare stents and 10 mm Fluency covered stents was similar (both of 20-30%); while the incidence rate of hepatic encephalopathy in 8 mm Fluency covered stent was only about 5-10%, suggesting that 8 mm Fluency covered stents could achieve effective shunt and reduce the incidence of hepatic encephalopathy 47 . Many studies have suggested that applying covered stents did not significantly reduce the incidence of hepatic encephalopathy, which is still one of the most important causes affecting the life of quality and overall survival of the patients after TIPS, which also brings heavy economic burden for the families. In the present study, we found that the incidence of hepatic encephalopathy was similar between the two groups, suggesting that using covered stents did not reduce, but slightly increased the incidence of hepatic encephalopathy as comparing with the bare stent group. We speculated that several factors including liver function, underlying diseases, and the size of the liver could be associated with this observation. A low protein diet, lactulose oral solution and dietary fiber can accelerate intestinal peristalsis and the excretion of the stool and toxic substances. All these strategies may reduce the incidence of hepatic encephalopathy.
In summary, Fluency covered stents could significantly reduce the short-and long-term rate of shunt channel restenosis and the recurrence rate of clinical symptoms, and increase the long-term survival of the patients; however, no evidence of reducing the incidence of hepatic encephalopathy was found in the present study.   Inclusion and exclusion criteria were carefully designed to exclude the confounding factors. The core item in the inclusion criteria was to minimize the diameter of the shunt channel (8 mm). The inclusion criteria included: 1) portal hypertension patients with defined indications for TIPS treatment; 2) scheduled for elective TIPS; and 3) aged between 18-70 years.
The patients with one or more of the following characteristics were excluded: 1) combined with hepatic encephalopathy before the treatment; 2) combined with portal vein thrombosis; 3) combined with malignant liver tumor or malignancies at the other sites; or 4) combined with hemorrhage of gastrointestinal ulcer.
The general principle of the exclusion criteria was to exclude all the potential factors that could affect the results; for instance, for cases underwent emergent TIPS, the preoperative examination might not reflect the actual conditions of the patients. If the adjustments of the severely abnormal parameters could not be performed, then the stents could not be randomly selected, so these patients were excluded from the present study. Age (< 18 or > 70 years), the existence of preoperative hepatic encephalopathy, portal vein thrombosis, malignancies, as well as gastrointestinal bleeding could affect the results. In addition, several patients who met the inclusion criteria, but did not accept the randomly assigned stents, or did not establish the shunt channel as expected due to specific conditions in the operation were also excluded. For instance, the patients who were scheduled to use bare stents but received covered stents to prevent major abdominal bleeding, received both covered and bare stents, or a stent nested in the 8 mm stent were excluded. The patients who were lost to follow up or not followed up as scheduled, underwent liver transplantation, developed liver cancer or other malignancies after the operation, or died of unrelated diseases were excluded as censored cases to ensure the accuracy, randomness, and controllability of the present study.
Sample size. Based on an expected incidence of the primary endpoint (survival rate) of about 60% at 5 years in the control group, we calculated that we would need at least 100 primary endpoint events (deaths) and a sample size of at least 250 patients to obtain 85% power to detect a significant difference between experimental and control, corresponding to a 10% reduction of relative risk (with a 2-sided type 1 error of 5%). Randomization. Each patient was assigned with an ID from 1 to 288 in order on admission, then a random number generator was used to randomly divide the 288 patients into group 0 or 1, with 144 patients in each. The random numbers were ordered from small to large to generate a random code. Patients with odd random codes were assigned to receive the treatments with 8 mm covered stents (experimental group), while the ones with even random codes receive the treatments with 8 mm bare stents (control group). The results of randomization were sealed in envelops, which were allocated to the clinicians in charge when the trial began. This was a double-blinded trial. Patients and physicians allocated to the intervention group were unaware of the allocated arm, outcome assessors and data analysts were kept blinded to the allocation.

Clinical characteristics. From
Operation procedures. Preoperative preparation. Liver function, coagulation, blood routine, blood type, electrocardiogram, computed tomography (CT), magnetic resonance imaging (MRI), color Doppler ultrasonography, gastroscopy, esophagography and other etiology-specific examinations were performed before the operation to exclude the bleeding caused by ulcer or other diseases. Patients' coagulation and platelet counts were regulated to meet the requirements of TIPS. The patients and the families were informed the risks and potential consequences of the operation, and operation agreements were signed.
Major operation processes. Jugular vein puncture and catheterization were performed. RUPS-100 (COOK Company) sheath was delivered into the angiographic catheter to display the hepatic vein and inferior vena cava. The appropriate site was selected on the hepatic vein or inferior vena cava for the puncture of the portal vein at appropriate angle. Contrast media was injected to ensure the puncture of the portal vein was successful. The sheath was then delivered into the portal vein. When the deliver was difficult, balloon dilation was performed after portography and then the stent was implanted. Pigtail catheter was used for portography, portal venous pressure was measured, and then the varicosed vein was embolized before (or after) shunting. The portal venous pressure was measured for the second time before an 8 mm balloon was used to dilate the shunt channel and an 8 mm stent was implanted. The portal venous pressure was measured for the third time along with portography. Finally, 131 patients underwent stenting with covered stents (Bard, Fluency) and 127 patients received bare stents (EV3, protégé; Cordis, Smart).
Postoperative observation and treatments. All the patients were asked to rest in bed for 24 hours after the operation; pressure dressing or sandbag pressing was used, and the vital signs were monitored. In addition, antibiotics were used prophylactically. Low molecular heparin (5000 IU, twice per day) was subcutaneously injected for 5 days, and then switched to warfarin for at least 1 year. The coagulation of each patient was examined every half month to ensure the international normalized ratio (INR) was between 2 and 3. Intravenous injection of branched chain amino acid and oral administration of lactulose were also applied to prevent hepatic encephalopathy. Liver protective strategy was taken for all the patients. For the patients with sepsis, treatments based on the results of blood culture and drug sensitive test were performed in time.
Follow up. For each patient, systemic examinations were performed at 3-and 6-months after the operation, and then re-examinations at every 6 months. Detailed medical history and symptoms were recorded. Examinations included liver function, coagulation, blood ammonia, blood routine, color ultrasonography, esophagography, CT, and gastroscopy. When color ultrasonography suggested stenosis of the shunt channel, aggravation of varicosity, or accompanied with gastrointestinal bleeding, refractory hydrothorax or ascites, imaging of the shunt channel would be repeated, and the portal venous pressure measured. If the blood flow in the shunt channel was normal, whereas the portal venous pressure increased or stenosis/occlusion of the shunt channel was identified, balloon dilation of the shunt channel and re-stenting was performed. When the shunt volume was insufficient or the patency of the shunt channel was difficult to resume, secondary TIPS was performed to establish a second shunt channel.
Statistical analyses. Statistical analysis was performed using SPSS software (version 17.0). Quantitative data were described with means and standard divisions, and compared by paired t-test. Qualitative data were compared by χ 2 test or Fisher's exact test. The difference in overall survival time between the control and experimental group was assessed using the log-rank test. The Kaplan-Meier method was used for calculating cumulative survival rate and survival curves. A P < 0.05 was considered statistically significant.