- Email: [email protected]
TIPSS trials: Design determines outcome
HEPATOLOGY Elsewhere JACQUELYN MAHER, EDITOR
ADVISORY COMMITTEE RAYMOND BURK, Nashville, TN DAVID CRABB, Indianapolis, IN ADRIAN DIBISCEGLIE, St. Louis, MO EMMET KEEFFE, Palo Alto, CA JOEL LAVINE, San Diego, CA MICHAEL NATHANSON, New Haven, CT JAMES SITZMANN, Washington, D.C. DWAIN THIELE, Dallas, TX
San Francisco General Hospital Building 40, Room 4102 1001 Potrero Avenue San Francisco, CA 94110
TIPSS TRIALS: DESIGN DETERMINES OUTCOME
(1) Cabrera J, Maynar M, Granados R, Gorriz E, Reyes R, Pulido Duque JM, San Roman JLR, et al. Transjugular intrahepatic portosystemic shunt versus sclerotherapy in the elective treatment of variceal hemorrhage. Gastroenterology 1996;110:832-839. (Reprinted with permission.) ABSTRACT
Background and Aims: Uncontrolled studies suggest that placement of a transjugular intrahepatic portosystemic shunt (TIPS) could be useful in the treatment of variceal bleeding. The aim of this study was to evaluate the efficacy and safety of TIPS in the elective treatment of hemorrhage from esophageal varices in a randomized controlled study that compared the effects of TIPS with those of endoscopic sclerotherapy (ES). Methods: Sixty-three consecutive cirrhotic patients with a hemorrhage from esophageal varices were included. Thirtytwo patients were randomly allocated to ES and 31 to TIPS groups. Results: One patient in each group died before the therapeutic procedure could be performed. During a mean follow-up period of 15 months, variceal rebleeding occurred in 51.6% of the patients in the ES group and 23% of those in the TIPS group. Uncontrolled rebleeding occurred in 10 of 31 patients in the ES group, whereas rebleeding did not occur in any patient of the TIPS group. Hepatic encephalopathy was more frequent in TIPS patients (33.3%) than in those treated by ES (13%). However, mortality was similar in both treatment groups. Conclusions: These preliminary results suggest that TIPS is more effective than ES in the prevention of variceal rebleeding in cirrhotic patients, even though no difference in survival was observed. (2) Rossle M, Deibert P, Haag K, Ochs A, Olschewski M, Siegerstetter V, Hauenstein KH, et al. Randomised trial of transjugular-intrahepatic-portosystemic shunt versus endoscopy plus propranolol for prevention of variceal rebleeding. Lancet 1997;349:1043-1049. (Reprinted with permission.) ABSTRACT
Background: The transjugular-intrahepatic-portosystemic shunt is a new interventional treatment for portal hypertension. The aim of our study was to compare the transjugular shunt with endoscopic treatment for the prophylaxis of recurrent variceal bleeding. Methods: Between March, 1993, and March, 1996, 126 patients with variceal bleeding were randomly assigned either transjugular shunt (n Å 61) or endoscopic treatment (n Å 65). Patients were followed up for
a median of 14 (IQR 8-25) months and 13 (8-25) months, respectively. In 31 (51%) of the shunted patients, simultaneous transjugular-variceal embolisation was done at the time of shunt placement. Endoscopic treatment consisted of sclerotherapy and/or banding ligation and was combined with propranolol medication. Findings: Technical success was achieved in all patients assigned to the shunt group. During follow-up, the cumulative 1-year variceal rebleeding rates in the shunted and endoscopically treated patients were 15% and 41% and the 2-year rates were 21% and 52% (P Å 0.001), respectively. In nine (12%) patients from the endoscopic group treatment failed and the patients received the transjugular-shunt treatment. A total of 19 bleeding episodes from any source occurred in 15 patients in the shunt group compared with 100 episodes in 33 patients in the endoscopic group. There was no difference in survival with estimated 1- year survival rates for shunted and endoscopically treated patients of 90% and 89%, and 2-year survival rates of 79% and 82%, respectively. The incidence of clinically significant hepatic encephalopathy after 1 year was higher in the shunt group (36% vs 18%, P Å 0.011). Interpretation: These results suggest that the transjugular shunt is more effective than endoscopic treatment in prevention of variceal rebleeding but has a considerable risk of hepatic encephalopathy. Survival is similar in the two groups. (3) Cello JP, Ring EJ, Olcott EW, Koch J, Gordon R, Sandhu J, Morgan DR, et al. Endoscopic sclerotherapy compared with percutaneous transjugular intrahepatic portosystemic shunt after initial sclerotherapy in patients with acute variceal hemorrhage: a randomized, controlled trial. Ann Intern Med 1997;126:858-865. (Reprinted with permission.) ABSTRACT
Background: Hemorrhage from esophageal varices remains a substantial management problem. Endoscopic sclerotherapy was preferred for more than a decade, but fluoroscopically placed intrahepatic portosystemic stents have recently been used with increasing frequency. Objective: To compare sclerotherapy with transjugular intrahepatic portosystemic shunt (TIPS) in patients with bleeding from esophageal varices. Design: Randomized, controlled clinical trial. Setting: Three teaching hospitals. Patients: 49 adults hospitalized with acute variceal hemorrhage from November 1991 to December 1995: 25 assigned to sclerotherapy and 24 assigned to TIPS. Intervention: Patients assigned to repeated sclerotherapy had the procedure weekly. In those assigned to TIPS, an expandable mesh stent was fluoroscopically placed between an intrahe-
1361
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
hepa
WBS: Hepatology
1362 HEPATOLOGY Elsewhere
HEPATOLOGY November 1997
patic portal vein and an adjacent hepatic vein. Measurements: Pretreatment measures included demographic and laboratory data. Postrandomization data included index hospitalization survival, duration of follow up, successful obliteration of varices, rebleeding from varices, number of variceal rebleeding events, total days of hospitalization for variceal bleeding, blood transfusion requirements after randomization, prevalence of encephalopathy, and total health care costs. Results: Mean follow-up ({ SE) was 567 { 104 days in the sclerotherapy group and 575 { 109 days in the TIPS group. Varices were obliterated more reliably by TIPS than by sclerotherapy (P õ 0.001). Patients having TIPS were significantly less likely to rebleed from esophageal varices than patients receiving sclerotherapy (3 of 24 compared with 12 of 25; P Å 0.012). No other follow-up measures differed significantly between groups. A trend toward improved survival, which was not statistically significant, was noted in the TIPS group (hazard ratio, 0.53 [95% CI, 0.18 to 1.5]). Conclusions: In obliterating varices and reducing rebleeding events from esophageal varices, TIPS was more effective than sclerotherapy. However, TIPS did not decrease morbidity after randomization or improve health care costs. It seemed to produce better survival, but the increase in survival was not statistically significant. (4) Sanyal AJ, Freedman AM, Luketic VA, Purdum III PP, Shiffman ML, Cole PE, Tisnado J, et al. Transjugular intrahepatic portosystemic shunts compared with endoscopic sclerotherapy for the prevention of recurrent variceal hemorrhage: a randomized, controlled trial. Ann Intern Med 1997;126:849-857. (Reprinted with permission.) ABSTRACT
Background: Transjugular intrahepatic portosystemic shunts (TIPS) have widened the use of portal decompression as therapy for variceal hemorrhage. However, no controlled studies have examined the efficacy of TIPS compared with that of other treatments. Objective: To compare the efficacy and safety of TIPS with those of endoscopic sclerotherapy for the prevention of recurrent variceal hemorrhage. Design: Randomized, controlled trial. Setting: Tertiary-care academic medical center. Patients: 100 patients with cirrhosis were evaluated a mean of approximately 10 days after an episode of acute variceal bleeding; 20 patients were excluded because of portal venous thrombosis (n Å 6), hepatoma (n Å 3), florid alcoholic hepatitis (n Å 6), and refusal to give consent (n Å 5). Interventions: TIPS (n Å 41) or sclerotherapy (n Å 39). The latter was performed by freehand injections of 5% Na morrhuate at 2- to 3-week intervals. Recurrent variceal hemorrhage was managed by sclerotherapy followed by angiographic assessment of TIPS and dilatation of the stents (TIPS group) or crossover to TIPS (sclerotherapy group). Measurements: Rebleeding and survival were the primary end points. Complications and rates of rehospitalization were secondary end points. Results: During a mean follow-up of approximately 1000 days, recurrent gastrointestinal bleeding resulted from variceal hemorrhage (9 patients in the TIPS group and 8 in the sclerotherapy group), portal gastropathy (1 patient in each group), and gastric lipoma (0 and 1 patients, respectively). A higher mortality rate was seen with TIPS (P Å 0.03). Death resulted from variceal bleeding (5 patients in the TIPS group and 3 in the
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
sclerotherapy group), sepsis (3 and 2 patients, respectively), liver failure (2 patients in each group), hepatoma (1 and 0 patients, respectively), and hemoperitoneum (1 and 0 patients, respectively). Encephalopathy was the most common complication in the TIPS group (n Å 12), and pain developing after sclerotherapy was the most common in the sclerotherapy group (n Å 10). The two groups had similar rates of rehospitalization. Conclusions: Endoscopic sclerotherapy and TIPS are equivalent with respect to rebleeding developing over the long term. However, sclerotherapy may be superior to TIPS with respect to survival. COMMENTS
Fine words and insinuating appearance are seldom associated with true virtue. Confucius, 550-478 B.C. The introduction of the transjugular intrahepatic portosystemic stent-shunt (TIPSS) into clinical practice was greeted with much enthusiasm. Several publications containing uncontrolled data have followed. The articles cited above represent the first four fully published, randomized controlled clinical trials comparing TIPSS with endoscopic treatment in the management of variceal hemorrhage. Interpretation of trials in variceal hemorrhage is difficult because of differences in the management of patients among various parts of the world. Moreover, TIPSS and endoscopic treatment can themselves be difficult to compare, because the success of a TIPSS procedure can be judged immediately by a reduction in portal pressure, whereas no such immediate criterion exists for endoscopic sclerotherapy (ES). Strict separation of patients into TIPSS or endoscopic treatment groups can also be problematic. For example, most patients entering randomized trials for variceal bleeding have already undergone endoscopic therapy to treat the index hemorrhage. Therefore, most patients in the TIPSS group will have had at least one session of endoscopic treatment. Cross-over between groups because of treatment failure also complicates the ‘‘intention-to-treat’’ analysis. All four of these trials have a potential confounding factor in the study design, which, although ethically responsible, makes interpretation of the mortality data difficult. This is the use of ‘‘TIPSS rescue’’ in patients randomized to ES. Rebleeding was significantly more frequent after ES than after TIPSS in three of the four studies (Cabrera: P õ .02; Rossle: P Å .001; Cello: P õ .01) (Table 1), and, in all of the studies, when ES patients experienced significant rebleeding from varices, TIPSS was performed. This measure was required in approximately one fifth of patients randomized initially to ES (Table 1). In Cabrera’s study, 51% of patients (16 of 31) in the ES group experienced rebleeding. Of these, 10 patients had uncontrolled variceal hemorrhage, and 9 were rescued with TIPSS. In Rossle’s study, 45% of patients (29 of 65) rebled in the ES group, and 9 patients underwent TIPSS rescue. In Cello’s study, 48% of patients (12 of 25) rebled in the ES group, of whom 6 patients required TIPSS rescue. In Sanyal’s study, only 21% of patients (8 of 39) rebled in the ES group, but 6 of them required TIPSS rescue. Although it is considered valid to analyze mortality data on an intentionto-treat basis, this does not reflect the true result. Mortality from uncontrolled variceal hemorrhage in patients treated
hepa
WBS: Hepatology
HEPATOLOGY Vol. 26, No. 5, 1997
HEPATOLOGY Elsewhere
1363
TABLE 1. Relationship Between Study Parameters and Results
Time of Randomization/ Primary Therapy
Study
Cabrera et al. (TIPPS 0 31; ES 0 32) Rossle et al. (TIPSS 0 61; ES 0 65) Cello et al. (TIPSS 0 24; ES 0 25) Sanyal et al. (TIPSS 0 41; ES 0 39)
Patients With Child C Cirrhosis
Patients Randomized With 1st Bleed
ES
TIPSS
ES
TIPSS
ES
TIPSS
Patients Needing ‘‘TIPSS Rescue’’
õ3 days
68%
06%
13%
51%
23%
16%
19%
28% (9 patients)
õ24 hours
39%
18%
18%
45%
15%
13%
12%
14% (9 patients)
õ24 hours
NA
48%
12%
16%
12%
24% (6 patients)
9-11 days?*
44%
21%
24%
29%
18%
15% (6 patients)
NA† 51%
46%
Rebleeding
Mortality
Abbreviation: NA, not available. * See text. † Child-Pugh score: TIPSS: 9 { 0.4; ES: 7.8 { 0.5; P õ .05.
with conventional measures approaches 80%. The use of TIPSS in this situation improves survival.1 If these studies were not designed to rescue patients in the ES arm, each of them might have shown a survival benefit for the patients in the TIPSS group. In fact, Garcia-Villareal2 presented a similar randomized trial in abstract form in which ‘‘TIPSS rescue’’ was not used. Although they randomized only a small number of patients, they showed a significant survival benefit for patients randomized to TIPSS. This tends to support the suggestion that the lack of significant differences in mortality in the first three studies may be due to rescue-TIPSS intervention for uncontrolled variceal hemorrhage. It could therefore be argued that these trials are not really TIPSS versus ES, but rather TIPSS versus [ES plus rescue TIPSS] trials. Of the above trials, the study by Sanyal et al. stands out in not reporting a difference in rebleeding (Table 1). This is probably because of the design and the methodology used in the study. One of the most important factors that determines outcome in clinical trials of variceal bleeding is the interval from the first bleed to randomization. Graham and Smith (1981)3 and Burroughs et al. (1989)4 confirmed in separate studies that shifting the starting point for analysis (time of randomization) of survival and rebleeding from the episode of bleeding by as little as 5 days significantly alters the rates of rebleeding and mortality. They also showed that, if treatments to prevent rebleeding were used long after control of a first episode of hemorrhage, then survival was not likely to be improved because many rebleeding deaths would have already occurred and the risk of death in the survivors is much lower. The time of randomization in the first three trials was within 3 days of the bleed (Table 1). In Sanyal’s study, the exact time of randomization is not clear. The authors state that ‘‘survivors of an episode of active esophageal variceal hemorrhage were considered for inclusion if they were clinically stable and were not actively hemorrhaging (absence of hemorrhage was indicated by a stable hemoglobin level and no need for transfusions) for at least 72 hours.’’ They then state that that ‘‘the primary intervention was started within 72 hours of randomisation.’’ This suggests that randomization was performed within 3 days of the initial hemorrhage, but a table in the manuscript indicates that 9 { 2 days elapsed between cessation of bleeding and TIPSS insertion, and that 11 { 3 days elapsed between cessation of bleeding and ES. Why patients received their respective treatments at a mean of 9 and 11 days, respectively, if the
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
criteria for randomization were met at 72 hours, is unclear. Furthermore, in Sanyal’s study, the rebleeding rate in the sclerotherapy group was much lower (21%) than that reported in the other three trials (Table 1). This may be due to a selection bias whereby the authors inadvertently selected survivors that were likely to do well with sclerotherapy. No significant differences were observed in mortality in the first three studies (Table 1), but an essential confounding factor is the ‘‘TIPSS rescue.’’ Sanyal’s study showed a significantly higher rate of mortality in the TIPSS group. When a proportional hazards model was constructed using variables such as age, Child class, encephalopathy, alcohol use, and primary intervention (TIPSS or ES), none of these factors were found to be independent predictors of mortality. It is also interesting to look at the causes of mortality. In the TIPSS group, of the 12 deaths, 5 were caused by variceal rebleeding. Seven patients died in the ES group, but only 3 deaths were caused by variceal rebleeding. Variceal rebleeding in the TIPSS group accounted for zero deaths in Cabrera’s study and zero deaths in Rossle’s study. Data for Cello’s study are not available. Three deaths occurred due to rebleeding in the ES group in Cabrera’s study, and 2 deaths were related to variceal rebleeding in Rossle’s study. The possible reason for the higher rate of variceal hemorrhage in the TIPSS group in Sanyal’s study may be related to the TIPSS procedure itself. TIPSS was either unsuccessful or inadequate in 5 patients (12%). Furthermore, most authorities would agree that the portal pressure gradient (PPG) should be reduced to less than 12 mm Hg after TIPSS. If this is not achieved, then the procedure should be considered unsuccessful. Although Sanyal et al. stated that some of their patients had a PPG ú 12 mm Hg, they did not indicate how many. They did not use a parallel stent to try to achieve this aim. The mean portal pressure they recorded at 6 months was approximately 18 mm Hg, suggesting that their method of surveillance (i.e., endoscopy and Doppler ultrasonography) did not adequately detect an increase in PPG. These factors may have contributed to the higher rate of rebleeding and mortality that they observed in the TIPSS group. Three of the four studies reported consistent differences in the incidence of encephalopathy between patients receiving TIPSS or ES. New-onset encephalopathy, or worsening of pre-existing encephalopathy, occurred in approximately 30% patients in the TIPSS group (Cabrera: 33%; Rossle: 36%; Sanyal: 31%), and approximately 15% in the sclerotherapy
hepa
WBS: Hepatology
1364 HEPATOLOGY Elsewhere
HEPATOLOGY November 1997
group (Cabrera: 13%; Rossle: 18%; Sanyal: 13%). Encephalopathy in Cello’s study was not significantly different between the two groups (TIPSS: 50%; ES: 44%). This observation is at variance with the other studies and may be due to the patients studied and the method of diagnosis of encephalopathy. In Cello’s study, only patients classified as having either massive or submassive hemorrhage were enrolled. The authors did not mention any strict criteria for the diagnosis of encephalopathy, but defined it as ‘‘the presence of asterixis, gross disorientation or agitation, or frank somnolence or coma in the absence of any other identifiable cause.’’ This would identify only patients with severe encephalopathy and overlook patients with milder encephalopathy. Only in Rossle’s study are clear diagnostic criteria set out for grading the severity of encephalopathy. It is notable that only a handful of patients in this study did not respond to medical treatment (3%), but responded quickly to reduction of the shunt. What is even more interesting is that there were 2 patients in the ES arm who developed encephalopathy that was refractory to treatment. These 2 patients had undergone TIPSS rescue. These observations are important because they highlight that the introduction of TIPSS is associated with increased incidence of encephalopathy, which is seldom troublesome, but also illustrate the shortcomings of an intention-to-treat analysis. Cello’s article provides the first evaluation of the relative costs incurred in the management of patients with variceal hemorrhage using TIPSS or ES. They found no significant differences in either the total health care costs or in the mean health care cost per day of survival. There were no significant differences between the two groups in the mean duration of hospitalization for encephalopathy. Sanyal et al. provided some data regarding the reasons for hospitalization. Patients in the ES group were hospitalized more frequently for decompensated ascites; patients in the TIPSS group had more hospitalizations for encephalopathy and sepsis. Rossle et al. found that patients in the TIPSS group spent less time in the hospital, but this was not statistically significant. Cabrera et al. found that, although there was no significant difference in the number of days spent in the hospital by patients in either group for the index admission, patients in the ES group spent significantly more time in the hospital for rebleeding. Overall, the total number of days spent per patient in the hospital was roughly 27 days in the ES group and 25 days in the TIPSS group (mean from Cabrera’s, Cello’s, and Rossle’s study; data from Sanyal are not available). It would have been interesting to see the extent to which patients were readmitted to intensive care, high-dependency care, or ward care, because these are important variables that have an enormous impact on overall cost. We recently compared TIPSS with variceal band ligation and found that patients in the banding arm required significantly more admissions for intensive and high-dependency care.5 Indeed, overall costs per patient were significantly lower in the TIPSS group. We accept, however, that measuring costs of various procedures and extrapolating them to different health systems is fraught with inaccuracies. Despite differences in the data, it is clear that TIPSS is not more expensive than ES in the management of variceal hemorrhage in patients with cirrhosis. Another confounding factor that makes interpretation difficult is that these studies do not strictly look at secondary prophylaxis of variceal hemorrhage, but at the role of the two treatments in the management of any episode of variceal bleeding. By definition, a study of secondary prophylaxis
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
should include only patients who are presenting with a first episode of variceal hemorrhage, and should build into the study an exclusion for patients who have bled previously. Patients with second or subsequent variceal bleeding are the survivors of a first bleed and may intrinsically have a different natural history. Cabrera randomized approximately 30% of patients in either group with previous variceal bleeding. In Rossle’s study, approximately 60% of the patients had previous bleeds, and, in Sanyal’s study, approximately 50% of the patients had previous bleeds. Cello et al. did not mention previous bleeds as an exclusion criteria and did not state how many of the patients in either group had previous bleeds. The fundamental principle of power calculations for clinical trials is to estimate the number of patients that would need to be randomized to show a difference in the primary end-point of the study. The primary end-point of the studies by Cabrera et al., Rossle et al., and Sanyal et al. was rebleeding, and the power calculations were made to this end. Cello et al. stated that the primary end-point of their study was survival, but their power calculations for estimation of sample size were for rebleeding. If mortality is considered a secondary end-point, as in all these studies, then it is inadvisable to overinterpret these data and state that ‘‘no differences in mortality were detected.’’ A study specifically designed to answer this question must be performed before such a conclusion can be made. One method for overcoming the problem of sample size is meta-analysis, but this is only reliable if the individual studies included in the analysis are homogenous. According to Friedman and Goldberg,6 ‘‘heterogeneity is the Achilles heel of a meta-analysis.’’ As shown above, studies examining therapy for variceal hemorrhage are often heterogeneous, not only because of variables such as interval to randomization, use of TIPSS rescue, or diagnosis of encephalopathy, but also because of confounding factors such as the use of banding, propranolol, vasoactive drugs, and nonstandardized methods for shunt surveillance. Bailar commented recently that, ‘‘in my own review of selected meta-analyses, problems were so frequent and so serious, including bias on the part of the meta-analyst, that it was difficult to trust the overall ‘best estimates’ that the method often produces.’’7 Even without formal meta-analysis, the studies highlighted above are extremely valuable and show five important conclusions. They leave little doubt that TIPSS substantially reduces rebleeding, its severity, and hospital admissions from this complication. One of the major problems following TIPSS is shunt insufficiency, which can be prevented and does not translate into increased rebleeding if shunt surveillance is performed regularly. TIPSS is associated with increased rates of encephalopathy, which is seldom troublesome. There is no difference in survival between the two treatment modalities, but this must take into account the effect of ‘‘TIPSS rescue’’ in patients with uncontrolled variceal hemorrhage in the patients in the endoscopy arm. Finally, TIPSS certainly does not cost more to the health service, but may, in some centers, be associated with reduced costs. There are now seven fully published trials in the literature comparing ES with variceal band ligation (VBL) that were combined in a meta-analysis.8 This included 547 patients and concluded that VBL carried a significantly lower rate of rebleeding, mortality, complications such as esophageal stricture, and the number of treatments required for variceal eradication. VBL is now considered by most as the endoscopic treatment of choice for patients with esophageal vari-
hepa
WBS: Hepatology
HEPATOLOGY Vol. 26, No. 5, 1997
HEPATOLOGY Elsewhere
ceal bleeding. In our recently completed trial comparing TIPSS with VBL (with a primary end-point of rebleeding),5 we showed that patients in the TIPSS group had significantly less rebleeding, no difference in encephalopathy or mortality, and reduced overall cost. Randomized, clinical trials comparing VBL with TIPSS must be performed on large numbers of patients with a first episode of variceal hemorrhage. Because very large numbers of patients will be required to assess a primary end-point of mortality, such a study should be international and involve multiple centers. The time of randomization is crucial, because any potential benefit of TIPSS will be progressively smaller with increasing time lag between the bleed and definitive procedure. We recognize the enormous problems facing the organizers of such a study. RAJIV JALAN, M.R.C.P. NIALL D. C. FINLAYSON, PH.D. PETER C. HAYES, PH.D. Department of Medicine and Centre for Liver and Digestive Diseases Royal Infirmary of Edinburgh Edinburgh, Scotland
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
1365
REFERENCES 1. Jalan R, John TG, Redhead DN, Finlayson NDC, Garden OJ, Hayes PC. A comparative study of the transjugular intrahepatic portosystemic stentshunt (TIPSS) and oesophageal transection in uncontrolled variceal hemorrhage. Am J Gastroenterol 1995;11:1932-1936. 2. Garcia-Villareal L, Martinez-Lagares F, Sierra A, Guevara C, HernandezCabrero T, Marrero JM, Jimenez E, et al. TIPS vs sclerotherapy for the prevention of variceal rebleeding. Preliminary results of a randomised study [Abstract]. HEPATOLOGY 1996;24:208A. 3. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-809. 4. Burroughs AK, Mezzanote G, Phillips A, Aiden McCormick P, McIntyre N. Cirrhotics with variceal hemorrhage: the importance of the time interval between admission and the start of analysis for survival and rebleeding rates. HEPATOLOGY 1989;9:810-807. 5. Jalan R, Forrest EH, Stanley AJ, Redhead DN, Forbes J, Dillon JD, MacGilchrist AJ, et al. A randomised trial comparing transjugular intrahepatic portosystemic stent-shunt with variceal band ligation in the prevention of rebleeding from oesophageal varices. HEPATOLOGY 1997;26:1115-1122. 6. Friedman HP, Goldberg JD. Meta-analysis: an introduction and point of view. HEPATOLOGY 1996;23:917-928. 7. Bailar III JC. The promise and problems of meta-analysis. N Engl J Med 1997;337:559-561. 8. Laine L, El-Newihi HM, Migikovsky B, Sloane R, Garcia F. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 1993;119:1-7.
hepa
WBS: Hepatology
ADVISORY COMMITTEE RAYMOND BURK, Nashville, TN DAVID CRABB, Indianapolis, IN ADRIAN DIBISCEGLIE, St. Louis, MO EMMET KEEFFE, Palo Alto, CA JOEL LAVINE, San Diego, CA MICHAEL NATHANSON, New Haven, CT JAMES SITZMANN, Washington, D.C. DWAIN THIELE, Dallas, TX
San Francisco General Hospital Building 40, Room 4102 1001 Potrero Avenue San Francisco, CA 94110
TIPSS TRIALS: DESIGN DETERMINES OUTCOME
(1) Cabrera J, Maynar M, Granados R, Gorriz E, Reyes R, Pulido Duque JM, San Roman JLR, et al. Transjugular intrahepatic portosystemic shunt versus sclerotherapy in the elective treatment of variceal hemorrhage. Gastroenterology 1996;110:832-839. (Reprinted with permission.) ABSTRACT
Background and Aims: Uncontrolled studies suggest that placement of a transjugular intrahepatic portosystemic shunt (TIPS) could be useful in the treatment of variceal bleeding. The aim of this study was to evaluate the efficacy and safety of TIPS in the elective treatment of hemorrhage from esophageal varices in a randomized controlled study that compared the effects of TIPS with those of endoscopic sclerotherapy (ES). Methods: Sixty-three consecutive cirrhotic patients with a hemorrhage from esophageal varices were included. Thirtytwo patients were randomly allocated to ES and 31 to TIPS groups. Results: One patient in each group died before the therapeutic procedure could be performed. During a mean follow-up period of 15 months, variceal rebleeding occurred in 51.6% of the patients in the ES group and 23% of those in the TIPS group. Uncontrolled rebleeding occurred in 10 of 31 patients in the ES group, whereas rebleeding did not occur in any patient of the TIPS group. Hepatic encephalopathy was more frequent in TIPS patients (33.3%) than in those treated by ES (13%). However, mortality was similar in both treatment groups. Conclusions: These preliminary results suggest that TIPS is more effective than ES in the prevention of variceal rebleeding in cirrhotic patients, even though no difference in survival was observed. (2) Rossle M, Deibert P, Haag K, Ochs A, Olschewski M, Siegerstetter V, Hauenstein KH, et al. Randomised trial of transjugular-intrahepatic-portosystemic shunt versus endoscopy plus propranolol for prevention of variceal rebleeding. Lancet 1997;349:1043-1049. (Reprinted with permission.) ABSTRACT
Background: The transjugular-intrahepatic-portosystemic shunt is a new interventional treatment for portal hypertension. The aim of our study was to compare the transjugular shunt with endoscopic treatment for the prophylaxis of recurrent variceal bleeding. Methods: Between March, 1993, and March, 1996, 126 patients with variceal bleeding were randomly assigned either transjugular shunt (n Å 61) or endoscopic treatment (n Å 65). Patients were followed up for
a median of 14 (IQR 8-25) months and 13 (8-25) months, respectively. In 31 (51%) of the shunted patients, simultaneous transjugular-variceal embolisation was done at the time of shunt placement. Endoscopic treatment consisted of sclerotherapy and/or banding ligation and was combined with propranolol medication. Findings: Technical success was achieved in all patients assigned to the shunt group. During follow-up, the cumulative 1-year variceal rebleeding rates in the shunted and endoscopically treated patients were 15% and 41% and the 2-year rates were 21% and 52% (P Å 0.001), respectively. In nine (12%) patients from the endoscopic group treatment failed and the patients received the transjugular-shunt treatment. A total of 19 bleeding episodes from any source occurred in 15 patients in the shunt group compared with 100 episodes in 33 patients in the endoscopic group. There was no difference in survival with estimated 1- year survival rates for shunted and endoscopically treated patients of 90% and 89%, and 2-year survival rates of 79% and 82%, respectively. The incidence of clinically significant hepatic encephalopathy after 1 year was higher in the shunt group (36% vs 18%, P Å 0.011). Interpretation: These results suggest that the transjugular shunt is more effective than endoscopic treatment in prevention of variceal rebleeding but has a considerable risk of hepatic encephalopathy. Survival is similar in the two groups. (3) Cello JP, Ring EJ, Olcott EW, Koch J, Gordon R, Sandhu J, Morgan DR, et al. Endoscopic sclerotherapy compared with percutaneous transjugular intrahepatic portosystemic shunt after initial sclerotherapy in patients with acute variceal hemorrhage: a randomized, controlled trial. Ann Intern Med 1997;126:858-865. (Reprinted with permission.) ABSTRACT
Background: Hemorrhage from esophageal varices remains a substantial management problem. Endoscopic sclerotherapy was preferred for more than a decade, but fluoroscopically placed intrahepatic portosystemic stents have recently been used with increasing frequency. Objective: To compare sclerotherapy with transjugular intrahepatic portosystemic shunt (TIPS) in patients with bleeding from esophageal varices. Design: Randomized, controlled clinical trial. Setting: Three teaching hospitals. Patients: 49 adults hospitalized with acute variceal hemorrhage from November 1991 to December 1995: 25 assigned to sclerotherapy and 24 assigned to TIPS. Intervention: Patients assigned to repeated sclerotherapy had the procedure weekly. In those assigned to TIPS, an expandable mesh stent was fluoroscopically placed between an intrahe-
1361
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
hepa
WBS: Hepatology
1362 HEPATOLOGY Elsewhere
HEPATOLOGY November 1997
patic portal vein and an adjacent hepatic vein. Measurements: Pretreatment measures included demographic and laboratory data. Postrandomization data included index hospitalization survival, duration of follow up, successful obliteration of varices, rebleeding from varices, number of variceal rebleeding events, total days of hospitalization for variceal bleeding, blood transfusion requirements after randomization, prevalence of encephalopathy, and total health care costs. Results: Mean follow-up ({ SE) was 567 { 104 days in the sclerotherapy group and 575 { 109 days in the TIPS group. Varices were obliterated more reliably by TIPS than by sclerotherapy (P õ 0.001). Patients having TIPS were significantly less likely to rebleed from esophageal varices than patients receiving sclerotherapy (3 of 24 compared with 12 of 25; P Å 0.012). No other follow-up measures differed significantly between groups. A trend toward improved survival, which was not statistically significant, was noted in the TIPS group (hazard ratio, 0.53 [95% CI, 0.18 to 1.5]). Conclusions: In obliterating varices and reducing rebleeding events from esophageal varices, TIPS was more effective than sclerotherapy. However, TIPS did not decrease morbidity after randomization or improve health care costs. It seemed to produce better survival, but the increase in survival was not statistically significant. (4) Sanyal AJ, Freedman AM, Luketic VA, Purdum III PP, Shiffman ML, Cole PE, Tisnado J, et al. Transjugular intrahepatic portosystemic shunts compared with endoscopic sclerotherapy for the prevention of recurrent variceal hemorrhage: a randomized, controlled trial. Ann Intern Med 1997;126:849-857. (Reprinted with permission.) ABSTRACT
Background: Transjugular intrahepatic portosystemic shunts (TIPS) have widened the use of portal decompression as therapy for variceal hemorrhage. However, no controlled studies have examined the efficacy of TIPS compared with that of other treatments. Objective: To compare the efficacy and safety of TIPS with those of endoscopic sclerotherapy for the prevention of recurrent variceal hemorrhage. Design: Randomized, controlled trial. Setting: Tertiary-care academic medical center. Patients: 100 patients with cirrhosis were evaluated a mean of approximately 10 days after an episode of acute variceal bleeding; 20 patients were excluded because of portal venous thrombosis (n Å 6), hepatoma (n Å 3), florid alcoholic hepatitis (n Å 6), and refusal to give consent (n Å 5). Interventions: TIPS (n Å 41) or sclerotherapy (n Å 39). The latter was performed by freehand injections of 5% Na morrhuate at 2- to 3-week intervals. Recurrent variceal hemorrhage was managed by sclerotherapy followed by angiographic assessment of TIPS and dilatation of the stents (TIPS group) or crossover to TIPS (sclerotherapy group). Measurements: Rebleeding and survival were the primary end points. Complications and rates of rehospitalization were secondary end points. Results: During a mean follow-up of approximately 1000 days, recurrent gastrointestinal bleeding resulted from variceal hemorrhage (9 patients in the TIPS group and 8 in the sclerotherapy group), portal gastropathy (1 patient in each group), and gastric lipoma (0 and 1 patients, respectively). A higher mortality rate was seen with TIPS (P Å 0.03). Death resulted from variceal bleeding (5 patients in the TIPS group and 3 in the
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
sclerotherapy group), sepsis (3 and 2 patients, respectively), liver failure (2 patients in each group), hepatoma (1 and 0 patients, respectively), and hemoperitoneum (1 and 0 patients, respectively). Encephalopathy was the most common complication in the TIPS group (n Å 12), and pain developing after sclerotherapy was the most common in the sclerotherapy group (n Å 10). The two groups had similar rates of rehospitalization. Conclusions: Endoscopic sclerotherapy and TIPS are equivalent with respect to rebleeding developing over the long term. However, sclerotherapy may be superior to TIPS with respect to survival. COMMENTS
Fine words and insinuating appearance are seldom associated with true virtue. Confucius, 550-478 B.C. The introduction of the transjugular intrahepatic portosystemic stent-shunt (TIPSS) into clinical practice was greeted with much enthusiasm. Several publications containing uncontrolled data have followed. The articles cited above represent the first four fully published, randomized controlled clinical trials comparing TIPSS with endoscopic treatment in the management of variceal hemorrhage. Interpretation of trials in variceal hemorrhage is difficult because of differences in the management of patients among various parts of the world. Moreover, TIPSS and endoscopic treatment can themselves be difficult to compare, because the success of a TIPSS procedure can be judged immediately by a reduction in portal pressure, whereas no such immediate criterion exists for endoscopic sclerotherapy (ES). Strict separation of patients into TIPSS or endoscopic treatment groups can also be problematic. For example, most patients entering randomized trials for variceal bleeding have already undergone endoscopic therapy to treat the index hemorrhage. Therefore, most patients in the TIPSS group will have had at least one session of endoscopic treatment. Cross-over between groups because of treatment failure also complicates the ‘‘intention-to-treat’’ analysis. All four of these trials have a potential confounding factor in the study design, which, although ethically responsible, makes interpretation of the mortality data difficult. This is the use of ‘‘TIPSS rescue’’ in patients randomized to ES. Rebleeding was significantly more frequent after ES than after TIPSS in three of the four studies (Cabrera: P õ .02; Rossle: P Å .001; Cello: P õ .01) (Table 1), and, in all of the studies, when ES patients experienced significant rebleeding from varices, TIPSS was performed. This measure was required in approximately one fifth of patients randomized initially to ES (Table 1). In Cabrera’s study, 51% of patients (16 of 31) in the ES group experienced rebleeding. Of these, 10 patients had uncontrolled variceal hemorrhage, and 9 were rescued with TIPSS. In Rossle’s study, 45% of patients (29 of 65) rebled in the ES group, and 9 patients underwent TIPSS rescue. In Cello’s study, 48% of patients (12 of 25) rebled in the ES group, of whom 6 patients required TIPSS rescue. In Sanyal’s study, only 21% of patients (8 of 39) rebled in the ES group, but 6 of them required TIPSS rescue. Although it is considered valid to analyze mortality data on an intentionto-treat basis, this does not reflect the true result. Mortality from uncontrolled variceal hemorrhage in patients treated
hepa
WBS: Hepatology
HEPATOLOGY Vol. 26, No. 5, 1997
HEPATOLOGY Elsewhere
1363
TABLE 1. Relationship Between Study Parameters and Results
Time of Randomization/ Primary Therapy
Study
Cabrera et al. (TIPPS 0 31; ES 0 32) Rossle et al. (TIPSS 0 61; ES 0 65) Cello et al. (TIPSS 0 24; ES 0 25) Sanyal et al. (TIPSS 0 41; ES 0 39)
Patients With Child C Cirrhosis
Patients Randomized With 1st Bleed
ES
TIPSS
ES
TIPSS
ES
TIPSS
Patients Needing ‘‘TIPSS Rescue’’
õ3 days
68%
06%
13%
51%
23%
16%
19%
28% (9 patients)
õ24 hours
39%
18%
18%
45%
15%
13%
12%
14% (9 patients)
õ24 hours
NA
48%
12%
16%
12%
24% (6 patients)
9-11 days?*
44%
21%
24%
29%
18%
15% (6 patients)
NA† 51%
46%
Rebleeding
Mortality
Abbreviation: NA, not available. * See text. † Child-Pugh score: TIPSS: 9 { 0.4; ES: 7.8 { 0.5; P õ .05.
with conventional measures approaches 80%. The use of TIPSS in this situation improves survival.1 If these studies were not designed to rescue patients in the ES arm, each of them might have shown a survival benefit for the patients in the TIPSS group. In fact, Garcia-Villareal2 presented a similar randomized trial in abstract form in which ‘‘TIPSS rescue’’ was not used. Although they randomized only a small number of patients, they showed a significant survival benefit for patients randomized to TIPSS. This tends to support the suggestion that the lack of significant differences in mortality in the first three studies may be due to rescue-TIPSS intervention for uncontrolled variceal hemorrhage. It could therefore be argued that these trials are not really TIPSS versus ES, but rather TIPSS versus [ES plus rescue TIPSS] trials. Of the above trials, the study by Sanyal et al. stands out in not reporting a difference in rebleeding (Table 1). This is probably because of the design and the methodology used in the study. One of the most important factors that determines outcome in clinical trials of variceal bleeding is the interval from the first bleed to randomization. Graham and Smith (1981)3 and Burroughs et al. (1989)4 confirmed in separate studies that shifting the starting point for analysis (time of randomization) of survival and rebleeding from the episode of bleeding by as little as 5 days significantly alters the rates of rebleeding and mortality. They also showed that, if treatments to prevent rebleeding were used long after control of a first episode of hemorrhage, then survival was not likely to be improved because many rebleeding deaths would have already occurred and the risk of death in the survivors is much lower. The time of randomization in the first three trials was within 3 days of the bleed (Table 1). In Sanyal’s study, the exact time of randomization is not clear. The authors state that ‘‘survivors of an episode of active esophageal variceal hemorrhage were considered for inclusion if they were clinically stable and were not actively hemorrhaging (absence of hemorrhage was indicated by a stable hemoglobin level and no need for transfusions) for at least 72 hours.’’ They then state that that ‘‘the primary intervention was started within 72 hours of randomisation.’’ This suggests that randomization was performed within 3 days of the initial hemorrhage, but a table in the manuscript indicates that 9 { 2 days elapsed between cessation of bleeding and TIPSS insertion, and that 11 { 3 days elapsed between cessation of bleeding and ES. Why patients received their respective treatments at a mean of 9 and 11 days, respectively, if the
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
criteria for randomization were met at 72 hours, is unclear. Furthermore, in Sanyal’s study, the rebleeding rate in the sclerotherapy group was much lower (21%) than that reported in the other three trials (Table 1). This may be due to a selection bias whereby the authors inadvertently selected survivors that were likely to do well with sclerotherapy. No significant differences were observed in mortality in the first three studies (Table 1), but an essential confounding factor is the ‘‘TIPSS rescue.’’ Sanyal’s study showed a significantly higher rate of mortality in the TIPSS group. When a proportional hazards model was constructed using variables such as age, Child class, encephalopathy, alcohol use, and primary intervention (TIPSS or ES), none of these factors were found to be independent predictors of mortality. It is also interesting to look at the causes of mortality. In the TIPSS group, of the 12 deaths, 5 were caused by variceal rebleeding. Seven patients died in the ES group, but only 3 deaths were caused by variceal rebleeding. Variceal rebleeding in the TIPSS group accounted for zero deaths in Cabrera’s study and zero deaths in Rossle’s study. Data for Cello’s study are not available. Three deaths occurred due to rebleeding in the ES group in Cabrera’s study, and 2 deaths were related to variceal rebleeding in Rossle’s study. The possible reason for the higher rate of variceal hemorrhage in the TIPSS group in Sanyal’s study may be related to the TIPSS procedure itself. TIPSS was either unsuccessful or inadequate in 5 patients (12%). Furthermore, most authorities would agree that the portal pressure gradient (PPG) should be reduced to less than 12 mm Hg after TIPSS. If this is not achieved, then the procedure should be considered unsuccessful. Although Sanyal et al. stated that some of their patients had a PPG ú 12 mm Hg, they did not indicate how many. They did not use a parallel stent to try to achieve this aim. The mean portal pressure they recorded at 6 months was approximately 18 mm Hg, suggesting that their method of surveillance (i.e., endoscopy and Doppler ultrasonography) did not adequately detect an increase in PPG. These factors may have contributed to the higher rate of rebleeding and mortality that they observed in the TIPSS group. Three of the four studies reported consistent differences in the incidence of encephalopathy between patients receiving TIPSS or ES. New-onset encephalopathy, or worsening of pre-existing encephalopathy, occurred in approximately 30% patients in the TIPSS group (Cabrera: 33%; Rossle: 36%; Sanyal: 31%), and approximately 15% in the sclerotherapy
hepa
WBS: Hepatology
1364 HEPATOLOGY Elsewhere
HEPATOLOGY November 1997
group (Cabrera: 13%; Rossle: 18%; Sanyal: 13%). Encephalopathy in Cello’s study was not significantly different between the two groups (TIPSS: 50%; ES: 44%). This observation is at variance with the other studies and may be due to the patients studied and the method of diagnosis of encephalopathy. In Cello’s study, only patients classified as having either massive or submassive hemorrhage were enrolled. The authors did not mention any strict criteria for the diagnosis of encephalopathy, but defined it as ‘‘the presence of asterixis, gross disorientation or agitation, or frank somnolence or coma in the absence of any other identifiable cause.’’ This would identify only patients with severe encephalopathy and overlook patients with milder encephalopathy. Only in Rossle’s study are clear diagnostic criteria set out for grading the severity of encephalopathy. It is notable that only a handful of patients in this study did not respond to medical treatment (3%), but responded quickly to reduction of the shunt. What is even more interesting is that there were 2 patients in the ES arm who developed encephalopathy that was refractory to treatment. These 2 patients had undergone TIPSS rescue. These observations are important because they highlight that the introduction of TIPSS is associated with increased incidence of encephalopathy, which is seldom troublesome, but also illustrate the shortcomings of an intention-to-treat analysis. Cello’s article provides the first evaluation of the relative costs incurred in the management of patients with variceal hemorrhage using TIPSS or ES. They found no significant differences in either the total health care costs or in the mean health care cost per day of survival. There were no significant differences between the two groups in the mean duration of hospitalization for encephalopathy. Sanyal et al. provided some data regarding the reasons for hospitalization. Patients in the ES group were hospitalized more frequently for decompensated ascites; patients in the TIPSS group had more hospitalizations for encephalopathy and sepsis. Rossle et al. found that patients in the TIPSS group spent less time in the hospital, but this was not statistically significant. Cabrera et al. found that, although there was no significant difference in the number of days spent in the hospital by patients in either group for the index admission, patients in the ES group spent significantly more time in the hospital for rebleeding. Overall, the total number of days spent per patient in the hospital was roughly 27 days in the ES group and 25 days in the TIPSS group (mean from Cabrera’s, Cello’s, and Rossle’s study; data from Sanyal are not available). It would have been interesting to see the extent to which patients were readmitted to intensive care, high-dependency care, or ward care, because these are important variables that have an enormous impact on overall cost. We recently compared TIPSS with variceal band ligation and found that patients in the banding arm required significantly more admissions for intensive and high-dependency care.5 Indeed, overall costs per patient were significantly lower in the TIPSS group. We accept, however, that measuring costs of various procedures and extrapolating them to different health systems is fraught with inaccuracies. Despite differences in the data, it is clear that TIPSS is not more expensive than ES in the management of variceal hemorrhage in patients with cirrhosis. Another confounding factor that makes interpretation difficult is that these studies do not strictly look at secondary prophylaxis of variceal hemorrhage, but at the role of the two treatments in the management of any episode of variceal bleeding. By definition, a study of secondary prophylaxis
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
should include only patients who are presenting with a first episode of variceal hemorrhage, and should build into the study an exclusion for patients who have bled previously. Patients with second or subsequent variceal bleeding are the survivors of a first bleed and may intrinsically have a different natural history. Cabrera randomized approximately 30% of patients in either group with previous variceal bleeding. In Rossle’s study, approximately 60% of the patients had previous bleeds, and, in Sanyal’s study, approximately 50% of the patients had previous bleeds. Cello et al. did not mention previous bleeds as an exclusion criteria and did not state how many of the patients in either group had previous bleeds. The fundamental principle of power calculations for clinical trials is to estimate the number of patients that would need to be randomized to show a difference in the primary end-point of the study. The primary end-point of the studies by Cabrera et al., Rossle et al., and Sanyal et al. was rebleeding, and the power calculations were made to this end. Cello et al. stated that the primary end-point of their study was survival, but their power calculations for estimation of sample size were for rebleeding. If mortality is considered a secondary end-point, as in all these studies, then it is inadvisable to overinterpret these data and state that ‘‘no differences in mortality were detected.’’ A study specifically designed to answer this question must be performed before such a conclusion can be made. One method for overcoming the problem of sample size is meta-analysis, but this is only reliable if the individual studies included in the analysis are homogenous. According to Friedman and Goldberg,6 ‘‘heterogeneity is the Achilles heel of a meta-analysis.’’ As shown above, studies examining therapy for variceal hemorrhage are often heterogeneous, not only because of variables such as interval to randomization, use of TIPSS rescue, or diagnosis of encephalopathy, but also because of confounding factors such as the use of banding, propranolol, vasoactive drugs, and nonstandardized methods for shunt surveillance. Bailar commented recently that, ‘‘in my own review of selected meta-analyses, problems were so frequent and so serious, including bias on the part of the meta-analyst, that it was difficult to trust the overall ‘best estimates’ that the method often produces.’’7 Even without formal meta-analysis, the studies highlighted above are extremely valuable and show five important conclusions. They leave little doubt that TIPSS substantially reduces rebleeding, its severity, and hospital admissions from this complication. One of the major problems following TIPSS is shunt insufficiency, which can be prevented and does not translate into increased rebleeding if shunt surveillance is performed regularly. TIPSS is associated with increased rates of encephalopathy, which is seldom troublesome. There is no difference in survival between the two treatment modalities, but this must take into account the effect of ‘‘TIPSS rescue’’ in patients with uncontrolled variceal hemorrhage in the patients in the endoscopy arm. Finally, TIPSS certainly does not cost more to the health service, but may, in some centers, be associated with reduced costs. There are now seven fully published trials in the literature comparing ES with variceal band ligation (VBL) that were combined in a meta-analysis.8 This included 547 patients and concluded that VBL carried a significantly lower rate of rebleeding, mortality, complications such as esophageal stricture, and the number of treatments required for variceal eradication. VBL is now considered by most as the endoscopic treatment of choice for patients with esophageal vari-
hepa
WBS: Hepatology
HEPATOLOGY Vol. 26, No. 5, 1997
HEPATOLOGY Elsewhere
ceal bleeding. In our recently completed trial comparing TIPSS with VBL (with a primary end-point of rebleeding),5 we showed that patients in the TIPSS group had significantly less rebleeding, no difference in encephalopathy or mortality, and reduced overall cost. Randomized, clinical trials comparing VBL with TIPSS must be performed on large numbers of patients with a first episode of variceal hemorrhage. Because very large numbers of patients will be required to assess a primary end-point of mortality, such a study should be international and involve multiple centers. The time of randomization is crucial, because any potential benefit of TIPSS will be progressively smaller with increasing time lag between the bleed and definitive procedure. We recognize the enormous problems facing the organizers of such a study. RAJIV JALAN, M.R.C.P. NIALL D. C. FINLAYSON, PH.D. PETER C. HAYES, PH.D. Department of Medicine and Centre for Liver and Digestive Diseases Royal Infirmary of Edinburgh Edinburgh, Scotland
AID
Hepa 0046
/
5p28$$$901
10-13-97 13:43:38
1365
REFERENCES 1. Jalan R, John TG, Redhead DN, Finlayson NDC, Garden OJ, Hayes PC. A comparative study of the transjugular intrahepatic portosystemic stentshunt (TIPSS) and oesophageal transection in uncontrolled variceal hemorrhage. Am J Gastroenterol 1995;11:1932-1936. 2. Garcia-Villareal L, Martinez-Lagares F, Sierra A, Guevara C, HernandezCabrero T, Marrero JM, Jimenez E, et al. TIPS vs sclerotherapy for the prevention of variceal rebleeding. Preliminary results of a randomised study [Abstract]. HEPATOLOGY 1996;24:208A. 3. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-809. 4. Burroughs AK, Mezzanote G, Phillips A, Aiden McCormick P, McIntyre N. Cirrhotics with variceal hemorrhage: the importance of the time interval between admission and the start of analysis for survival and rebleeding rates. HEPATOLOGY 1989;9:810-807. 5. Jalan R, Forrest EH, Stanley AJ, Redhead DN, Forbes J, Dillon JD, MacGilchrist AJ, et al. A randomised trial comparing transjugular intrahepatic portosystemic stent-shunt with variceal band ligation in the prevention of rebleeding from oesophageal varices. HEPATOLOGY 1997;26:1115-1122. 6. Friedman HP, Goldberg JD. Meta-analysis: an introduction and point of view. HEPATOLOGY 1996;23:917-928. 7. Bailar III JC. The promise and problems of meta-analysis. N Engl J Med 1997;337:559-561. 8. Laine L, El-Newihi HM, Migikovsky B, Sloane R, Garcia F. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 1993;119:1-7.
hepa
WBS: Hepatology