Abstract
Patients with Child-Pugh class C cirrhosis or Child-Pugh class B cirrhosis with persistent bleeding confirmed by endoscopy are at high risk of failing medical therapy and have a poor prognosis, and even remedial TIPS for patients who have failed medical therapy does not significantly reduce the risk of treatment failure or improve prognosis.
Methods
We enrolled 63 patients with cirrhosis combined with acute esophagogastric fundic variceal bleeding within 24 hours of admission, all of whom received vasoactive drugs + endoscopic therapy on admission, and randomly divided them into two groups, namely, the early TIPS group (32 patients) and the drug therapy-endoscopic ligation group (31 patients), with the early TIPS group undergoing TIPS within 72 hours of admission. The early TIPS group was treated with TIPS (using a PTFE-coated stent) within 72 hours of admission, while the pharmacotherapy-endoscopic ligation group continued to receive vasoactive medication, followed by propranolol or nadolol and long-term endoscopic ligation (EBL) after 3-5 days, and then emergency TIPS was performed as a remedial measure after the above treatment measures failed.
Results
During a mean follow-up period of 16 months, 14 patients in the control group experienced rebleeding or uncontrollable bleeding, corresponding to only 1 patient in the trial group (P = 0.001). The precise probability of maintaining this absence of a composite endpoint at one year was 50% in the control group and 97% in the trial group (P<0.001). A total of 16 patients died in both groups (4 in the trial group and 12 in the control group, P = 0.01). The one-year exact survival rate was 61% in the control group and 86% in the experimental group (P<0.001). Seven patients in the control group underwent TIPS during the observation period Seven patients in the control group underwent TIPS as a remedial treatment measure, and four of them died. The number of days the patients were treated in the intensive care unit during the follow-up period and the proportion of hospitalization dates during the follow-up period and the proportion of hospitalization dates to the overall follow-up dates were significantly higher in the control group than in the experimental group, and no significant differences in serious adverse events were observed between the two groups.
Conclusion
For those patients hospitalized for acute esophagogastric fundic variceal bleeding in cirrhosis and at high risk of treatment failure for acute gastrointestinal bleeding, early TIPS treatment significantly reduces the probability of treatment failure and death for those patients hospitalized for acute esophagogastric fundic variceal bleeding in cirrhosis and at high risk of treatment failure.
Ruptured esophagogastric variceal bleeding is a serious complication of portal hypertension and the leading cause of death in patients with cirrhosis combined with portal hypertension. Continued deterioration of liver function, uncontrollable gastrointestinal bleeding, early rebleeding, and excessive portal pressure directly increase mortality in these patients. In patients with acute esophagogastric fundic variceal bleeding, the currently recommended treatment regimen is a combination of vasoactive drugs, prophylactic antibiotics, and endoscopic hemostasis. However, in approximately 10-15% of patients, treatment fails and multiple endoscopic hemostasis and repeated transfusions are required. In these patients, TIPS is a more effective measure to control bleeding, but the mortality rate is also high, probably due to liver failure as a result of continued deterioration of liver function.
In a study by Monescillo et al. whose subjects included patients at high risk of treatment failure, early TIPS intervention had a better prognosis than drug therapy alone in patients with hepatic venous pressure gradients above 20 mmHg, as described in the literature. However, patients in its drug treatment group were not treated with the currently recommended treatment regimen and therefore, it may have gotten a worse outcome than expected.
This study wanted to demonstrate whether early TIPS with a polytetrafluoroethylene coated overmolded stent could improve the outcome of treatment in patients with cirrhotic esophagogastric fundic variceal bleeding who have failed treatment or are at risk of death.
Methods
Patient selection
From May 2004 to March 2007, we collected a total of 63 patients from 9 European centers (see Supplementary Appendix for details, available in full from the NEJM homepage).
Inclusion criteria: patients with acute esophagogastric fundic variceal bleeding in cirrhosis who had been treated with vasoactive drugs, endoscopic hemostasis, and prophylactic antibiotics; patients with Child-Pugh class C (10-13 points) or Child-Pugh class B (7-9 points) liver function along with active bleeding on endoscopy. those with Child-Pugh scores above 13 could not be enrolled (ChildCPugh classification of liver disease, grade A indicates mild disease; grade B indicates moderate disease; grade C indicates severe disease and poor prognosis). Endoscopic skin loop ligation (EBL) or endoscopic sclerotherapy (EIS) were performed at the same time as endoscopy within 12 hours of patient admission; patients were also required to have received vasoactive medication (i.e., terlipressin 2 mg/4H; growth inhibitor 250-500 μg/H; and sennin 50-100 μg/Hg). Diagnostic criteria for endoscopic active bleeding were in accordance with the Baveno Criteria.
Exclusion criteria: age over 75 years; pregnancy; hepatocellular carcinoma but not meeting the Milan Criteria for liver transplantation [i.e., single lesion <5 cm or multiple lesions (up to 3) but the largest ≤3 cm in diameter]; creatinine >265 μmol/L; Child-Pugh score >13; previous use of drugs in combination with endoscopic therapy to prevent bleeding; previous portal venous shunts or TIPS; bleeding from isolated varicose fundic veins or ectopic varicose veins; portal trunk thrombosis; and cardiac insufficiency.
All patients were required to sign an informed consent for treatment. The study protocol was approved by the ethics committee of each participating hospital and was performed in full compliance with the relevant regulations in the Code of Practice for the Administration of Pharmaceutical Clinical Trials.
Patients were randomized within 24 hours of admission and randomization codes were generated by computer application of a hidden block size of four. The resulting assignment codes for the treatment regimens were kept by the coordinating center in sealed, opaque envelopes and numbered consecutively. The randomization groups and the assignment of the different treatment protocols were obtained by telephone or fax from the coordinating center of the project (telephone and fax were available 24 hours a day in the coordinating center)
Drug + endoscopic skin collar ligation
Patients were given continuous vasoactive medication until the bleeding had stopped for more than 24 hours, preferably reaching more than 5 days, and then a non-selective B-blocker (takayasu or nadolol). Gradually increase the drug dose (every 2 to 3 days) until the maximum tolerated dose or the maximum allowed dose is reached, i.e., 160 mg/BID for take-home safety and 240 mg/QD for nadolol. after reaching these doses, give 10 mg of isosorbide mononitrate at bedtime and gradually increase to the maximum tolerated dose dose or the maximum allowed dose, 20 mg/BID. in addition, within the first endoscopic A second elective endoscopic skin loop ligation was performed within 7 to 14 days after the first treatment and every 10 to 14 days thereafter until the varicose vein disappeared completely (i.e., the varicose vein disappeared completely or the ligation operation could no longer be performed). Endoscopic lap ligation was performed using a multi-band ligator (6-Shooter Saeed Multi-Band Ligator, Cook, or Speedband SuperView Super 7, Boston Scientific), and the ligature was first applied at the esophage-fundus junction. Proton pump inhibitors were given to all patients during treatment until the varices disappeared. Endoscopy was repeated at months 1, 6, and 12 after the disappearance of the varices, and annually thereafter, and further endoscopic skin ligation was performed if varices were found. Criteria for treatment failure: one severe rebleed (requiring 2 or more units of red blood cell transfusion) or two less severe rebleeds but requiring TIPS as a remedial treatment measure.
Early TIPS treatment
TIPS was performed within 72 hours of diagnostic endoscopy (within 24 hours if possible), and all patients were treated with vasoactive drugs prior to this time. The intraoperatively placed polytetrafluoroethylene coated stents (Viatorr, Gore) were initially 8 mm in diameter and were expanded to 10 mm in diameter if the portal venous pressure gradient failed to fall below 12 mm Hg after insertion.
When complications related to portal hypertension reappear or abnormalities of the TIPS tract are detected by Doppler examination (i.e., portal vein flow velocity <28 cm/s; change in the direction of intrahepatic portal branch flow; decrease in portal flow velocity of more than 50%), further TIPS shunt angiography is required, and if TIPS shunt dysfunction is confirmed, shunt balloon dilation or placement of A second Teflon-coated stent was placed.
Follow-up
Clinical follow-up was performed at 1 month, 3 months, and every 3 months thereafter. Color Doppler ultrasound was performed regularly at 1 month, 6 months, and every 6 months thereafter. Endpoints for follow-up were patient death, liver transplantation, a maximum follow-up period of 2 years, or up to the end of the study (September 2007).
Trial endpoints
The first observed endpoint was failure to control acute bleeding or severe GI rebleeding within the first year of enrollment. The second endpoint was patient mortality within 6 weeks and 1 year, uncontrolled acute bleeding, early rebleeding (rate of rebleeding within 5 days and 6 weeks), rate of rebleeding from 6 weeks to 1 year, and other complications due to portal hypertension at follow-up, number of days in the intensive care unit and proportion of days in the hospital throughout the follow-up period, alternative treatment options, etc. The following statistics were used
Statistical treatment
In the study by Villanueva et al, in patients with active variceal rupture bleeding confirmed by endoscopy, the rate of failure to control bleeding within 5 days was 27% despite combined drug and endoscopic therapy, similarly, in another study, in patients with Child-Pugh grade B or C variceal rupture bleeding, the rate of uncontrolled bleeding despite drug and endoscopic therapy was 23%. The rate of uncontrolled bleeding was 23%. Assuming a 25% rate of uncontrolled bleeding within 5 days in our study, and considering a 20% incidence of rebleeding from day 5 to 1 year, the cumulative rate of uncontrolled bleeding and rebleeding within 1 year would be estimated to be as high as 45%. We hypothesized that early use of an e-PTFE overmolded stent would reduce this risk to 10%. Because early use of the TIPS overlay stent can provide a theoretical basis for early use of TIPS only if it is shown to be superior to current treatment options, we used a one-sided test to calculate the sample size in this study. We calculated a sample size of 31 cases per group to effectively account for this difference, with an alpha level of 0.05 and a beta level of 0.20.
All data were analyzed according to a pre-designed statistical scheme in accordance with the basic principles of intention-to-treat. Dichotomous variables were compared using Fisher’s exact test, and nonparametric continuous variables were tested using the Mann-Whitney rank sum test. The probability of reaching the first observed endpoint and the estimation of survival were then estimated using the Kaplan Meier method with log-rank test. Absolute risk reduction and number needed to treat were estimated using 95% confidence periods for precision and effect. p less than 0.05 was considered a statistically significant difference, and all tests were bilaterally symmetric. The statistical software was SPSS 16.0 and CIA 2.1.2 (University of outhampton,United Kingdom).
Results
A total of 296 patients were excluded from the study. 63 patients who were eligible for enrollment were randomly assigned to the drug+endoscopic lap banding group (control group, 31 patients) and the experimental group (experimental group, 32 patients). (test group, 32 patients). There was no significant difference in the basic characteristics of the patients in the two groups at the time of enrollment. A total of 7 patients (3 in the control group and 4 in the trial group) were lost to follow-up (mean 8 months, range 0.5 to 12 months) and failed to reach the composite endpoint. The mean follow-up time was 10.6±9.9 months in the control group and 14.6±8.4 months in the trial group. Six patients (2 in the control group and 4 in the trial group) underwent liver transplantation during the follow-up period.
In the control group, 22 patients were given propranolol (mean dose 55 mg, 10 to 160 mg) and 3 patients were given nadolol, leaving 6 patients who did not receive non-selective beta-blockers from the beginning because of uncontrollable bleeding, early recurrent bleeding, or death. Among patients treated with nonselective beta-blockers, isosorbide 5~nitrate was added in 12 patients (mean dose 25 mg, 10-40 mg), and isosorbide 5~nitrate was not given in another 13 patients because of hypotension, physician’s habit, or early death. 12 patients had their varices eradicated after a mean of 2 (1-7) endoscopic skin collar ligations, and None of the patients had rebleeding; four patients had rebleeding followed by endoscopic laparoscopy, which also led to the eradication of varicose veins. In the remaining 15 patients, the varicose veins were not eradicated, including 12 patients who reached the initial endpoint (7 had TIPS remediation and 5 died), 2 patients who were lost to follow-up, and 1 patient whose varicose veins were not eradicated despite 8 endoscopic ligations.
In the trial group, TIPS was performed early in all but 1 patient due to voluntary withdrawal of consent. One patient developed sudden supraventricular tachycardia, which improved with pharmacological treatment. The mean portal venous pressure gradient decreased from 20.2±7 mmHg to 6.2±3 mmHg in 27 patients with one stent and 4 patients with two stents (P<0.001). Two patients underwent variceal vein embolization (including the aforementioned one patient with a portal vein pressure gradient greater than 12 mmHg after TIPS).
Re-bleeding
The primary complex endpoint was reached in 14 patients in the control group. At the time of reaching the endpoint, the Model for End-Stage Liver Disease (MELD) score (6 to 40 points, with higher scores indicating more severe disease) increased from a mean of 18.8 ± 6.4 to 22.6 ± 11 points in these patients. Seven of these 14 patients opted for TIPS (using an e~PTFE overlay stent) as a palliative treatment measure, and although bleeding was controlled, four of them died within 36 days (1 to 36 days) after the procedure. Five of these 14 patients failed to undergo further treatment and eventually died because of severe liver failure. When the primary endpoint of the study was reached, nine patients who died had a mean Model for End-Stage Liver Disease (MELD) score of 28.2±9, and two other patients who were treated with endoscopic ligation at the time of the primary endpoint were alive at the end of follow-up.
In the trial group, the primary endpoint was achieved in only 1 patient (significantly different compared with the control group, P<0.001), and that patient still had a portal venous pressure gradient above 12 mm Hg after TIPS. the one-year actuarial probability of bleeding control and absence of recurrent bleeding was higher in the trial group than in the control group (97% vs 50%; 47 percentage point absolute risk reduction; 95% confidence interval 25 to 69; number of cases requiring treatment 2.1; 95% confidence interval 1.4 to 4.0). Four additional patients (three in the control group and one in the trial group) had recurrent bleeding, but it was not clinically significant (i.e., no hospitalization or transfusion was required).
Survival rate
There were 12 deaths in the control group, which was significantly different from the 4 deaths in the trial group (P=0.001). 6-week survival was higher in the trial group than in the control group (97%: 67%; absolute risk reduction of 30 percentage points, 95% confidence interval 12-48; number of cases requiring treatment 3.3; 95% confidence interval 2.1-8.3). 1-year survival was also higher in the trial group than in the control group (86% The 1-year survival rate was also higher in the test group than in the control group (86%: 61%; absolute risk reduction of 25 percentage points; 95% confidence interval 2 to 48; number of cases requiring treatment 4.0; 95% confidence interval 2.1 to 50.0).
Other complications of portal hypertension
The one-year actuarial probability of developing hepatic encephalopathy was 28% in the trial group compared with 40% in the control group (12 percentage point absolute difference; 95% confidence interval: 18 to 40; P=0.13). A total of 17 hepatic encephalopathies occurred in 12 patients in the control group, whereas a total of 10 hepatic encephalopathy events occurred in 8 patients in the trial group (Table 3). Most of the hepatic encephalopathy occurred when there was a possibility of hemorrhage. The probability of a hepatic encephalopathy event within one year after discharge was similar in both groups (10% in the control group and 19% in the trial group, P=0.80). Grade III hepatic encephalopathy occurred in three patients in the control group and in two patients in the trial group, and one patient in each group had mild, recurrent hepatic encephalopathy.
The one-year actuarial probability of new ascites or worsening of existing ascites was 33% in the control group and 13% in the trial group, with an absolute difference of 20 percentage points between the two groups (95% confidence interval: 8 to 47; P=0.11). Two patients in the control group developed and died of spontaneous bacterial peritonitis in the prehemorrhagic phase. A total of 7 patients in both groups developed hepatorenal syndrome in the prehemorrhagic period; 5 patients in the control group died in 4 cases, and 2 patients in the test group survived. The number of hospital days during follow-up was 15% (interquartile range, 5 to 100) of the entire follow-up period in the control group compared with 4% (interquartile range, 2 to 13) in the trial group (P=0.01).
Other adverse events
As shown in Table 3, there was no significant difference in the number of cases with adverse events between the two groups.
Discussion
The study by Monescillo et al. found that early TIPS treatment had a better prognosis compared with drug therapy in patients with uncontrolled or high risk of recurrent bleeding from esophagogastric fundic variceal bleeding due to a hepatic venous pressure gradient equal to or greater than 20 mmHg11. However, the study did not include continuous pharmacotherapy and endoscopic collar ligation in the medical treatment group, and the stent used for the TIPS procedure in the trial group was a bare stent. Related studies have shown that the use of e-PTFE-coated stents reduces the incidence of shunt dysfunction and recurrence of complications related to portal hypertension with TIPS. More importantly, the decision to perform TIPS was based on the measured hepatic venous pressure gradient, which has not been widely measured, especially in emergency patients. Therefore, it is difficult to generalize the results of this study to clinical practice.
Our study was deliberately designed to explore whether the early application of TIPS for varices at high risk of bleeding would improve the prognosis of patients treated with TIPS using an e-PTFE overmolded stent according to the current recommended protocol. The results found that early TIPS treatment resulted in a significant reduction in the risk of uncontrolled or recurrent esophagogastric fundic variceal bleeding.
More importantly, early TIPS treatment reduced the risk of death in patients. We even observed a survival benefit of TIPS as a remedy after failure of pharmacological treatment for esophageal fundic variceal bleeding in cirrhosis, which is also consistent with previous studies. Notably, five patients were not considered for TIPS because of severe deterioration in liver function (confirmed by MELD scores), and these five patients had higher MELD scores than those with excess mortality after TIPS. These data clearly indicate that in patients with active bleeding with ChildCPugh class C or B liver function, initial failure to control bleeding or early recurrence of bleeding can lead to further deterioration of liver function, and that further deterioration of liver function adds to the poor prognosis and may prevent patients from receiving remedial TIPS therapy.
Several previous studies evaluating the role of TIPS in the prevention of esophageal fundic variceal bleeding in cirrhosis have demonstrated that TIPS reduces the rate of recurrent bleeding from cirrhotic esophageal varices but increases their incidence of hepatic encephalopathy and, therefore, is not effective in improving patient survival. It is based on these findings that TIPS is now generally recommended only as a remedial treatment for esophagogastric fundic variceal bleeding in cirrhosis. Clearly, the results of our study challenge the recommendation that TIPS be used only as a remedial treatment measure. It is important to note that my study differs from previous related studies that used bare stents or did not exclude patients who were at high risk of treatment failure. In a study by Escorsell et al, more than 15% of patients with cirrhotic esophagogastric fundic variceal bleeding were not included in the study because they required emergency TIPS treatment or died within 5 days from uncontrollable bleeding. Therefore, the possibility that patients at risk could also benefit from TIPS treatment could not be confirmed in this study. Another study showed that patients who failed to control bleeding effectively with initial treatment had a mortality rate of up to 66%. For potentially high-risk patients, TIPS treatment (with e-PTFE-coated stents) is highly effective in controlling bleeding to counteract the potential adverse effects and prevent further deterioration of the patient’s condition for the overall benefit of the patient. Conversely, we do not recommend TIPS as a treatment for primary liver function ChildCPugh class A cases because of the low failure and mortality rates of drug therapy in such patients.
Although the risk of treatment failure is higher in patients with hepatic function ChildCPugh class C than in patients with hepatic function ChildCPugh class B, an appropriate subgroup analysis was not performed in our trial. Therefore, whether early TIPS treatment can benefit patients in both subgroups equally needs further evaluation.
Early TIPS treatment did not increase the incidence of hepatic encephalopathy or worsen pre-existing hepatic encephalopathy.
In conclusion, in patients with active bleeding esophagogastric varices in cirrhosis with Child-Pugh class C or B liver function, early TIPS treatment (with e-PTFE overlay stent) significantly reduced the risk of bleeding control failure, recurrent bleeding, and death without increasing the risk of developing or worsening pre-existing hepatic encephalopathy.