Beta‐blockers for prevention of oesophageal variceal rebleeding in cirrhotic patients
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
This systematic review will evaluate the beneficial and harmful effects of beta‐blockers on preventing oesophageal variceal rebleeding in cirrhotic patients who survived a variceal bleeding.
Background
Oesophageal varices develop in 50 to 60 per cent of patients with histologic evidence of cirrhosis (Jensen 2002), and congestive gastropathy is detected in 7 to 98 per cent of cirrhotic patients (Primignani 2001). Varices have been identified as the bleeding source in 65 to 80 per cent of patients with liver cirrhosis and upper intestinal bleeding (Sauerbruch 1988; Poynard 1991). The risk of recurrent bleeding is 30 per cent at six weeks and nearly 70 per cent at one year for those patients who survive their initial episode of variceal bleeding (Graham 1981). Mortality that follows a variceal bleeding is 30 to 50 per cent (Sauerbruch 1988; Poynard 1991;D'Amico 1995).
In order to decrease the portal vein pressure and prevent the occurrence of the second variceal bleeding, various therapies, including variceal sclerotherapy and/or banding, shunt surgery, pharmacotherapy, the transjugular intrahepatic portosystemic shunt, and liver transplantation, have been developed (Rossle 2001; Chung 2002). Endoscopic sclerotherapy and variceal banding are widely used to prevent rebleeding from oesophageal varices. However, sclerotherapy is still associated with a rebleeding rate as high as 50 per cent and with complications such as fever and oesophageal ulceration or stricture in up to 40 per cent of patients (Schuman 1987; Rossle 2001; Chung 2002). Variceal banding seems superior to sclerotherapy in decreasing rebleeding, but not in reducing mortality (Heresbach 1995).
Among drugs used to prevent oesophageal variceal rebleeding, beta‐blockers are regarded as the best candidate because of their ability to reduce the portal vein pressure. In cirrhotic patients, serum catecholamine concentrations are increased and contribute to the systemic and splanchnic hyperdynamic state (Gaudin 1991). By reducing the sympathetic nervous activity, beta1‐blockers are able to reduce the cardiac output and thereby the splanchnic blood flow; and beta2‐blockers are expected to lead to unopposed alpha‐adrenergic tone in the splanchnic vascular territory and thus to splanchnic arterial vasoconstriction. Reduction of splanchnic blood flow is, therefore, achieved by dual mechanisms and leads to a reduction of portal pressure (Groszmann 1990). In most trials, the dose of the beta‐blocker is titrated until a 20 to 25 per cent reduction of resting heart rate is reached. However, the efficacy of pharmacological treatments ought also to be evaluated by measurement of portal pressure (de Franchis 2001). Propranolol, a non‐selective beta‐blocker, is the only pharmacological agent that has been sufficiently evaluated. Propranolol has been shown to decrease portal venous pressure (Lebrec 1982), decrease the blood flow in the splanchnic region (Cales 1985(a)), and decrease blood flow in the superior portosystemic collateral circulation (Cales 1985(b)). Meta‐analyses of randomised clinical trials have shown that beta‐blockade significantly decrease variceal rebleeding (Pagliaro 1989; Ricca Rosellini 1991) and late meta‐analyses also showed that beta‐blockers significantly reduce death due to bleeding and all‐cause mortality (Hayes 1990; D'Amico 1995; Bernard 1997; D'Amico 1999). None of these meta‐analyses considered the potential influence of methodological quality of the included randomised trials on trial results (Kjaergard 2001).
Beta‐blockers may cause adverse events, which include bronchospasm, heart failure, prolonged hypoglycaemia, bradycardia, heart block, intermittent claudication, and Raynaud's phenomenon, and neurological reactions such as depression, fatigue, and nightmares (Frishman 1988). In cirrhotic patients, the liver function is damaged and drug toxicity may be more severe during the treatment. Systematically assessing the harmful effects of beta‐blockers for secondary prevention of oesophageal variceal bleeding in cirrhotic patients is needed for safe use of beta‐blockers.
Objectives
This systematic review will evaluate the beneficial and harmful effects of beta‐blockers on preventing oesophageal variceal rebleeding in cirrhotic patients who survived a variceal bleeding.
Methods
Criteria for considering studies for this review
Types of studies
We will include randomised clinical trials irrespective of blinding, publication status, or language. For cross‐over trials, we will only include data from the first period. We will exclude quasi‐randomised studies and observational studies. However, for the evaluation of rare, serious adverse events we will consider also such study designs as adverse events are rarely captured in randomised clinical trials.
Types of participants
Patients with cirrhosis of different etiologies (diagnosed by liver biopsy specimens or clinical, biochemical, and ultrasonographic findings) who had survived an oesophageal variceal bleeding. We will not include patients in whom control of bleeding failed. Not all trials have included patients only with cirrhosis and therefore we will include also trials in which at least 80 per cent of the patients have cirrhosis. However, such trials will be analysed separately (see Methods).
Types of interventions
-
Beta‐blocker of any dose or duration versus placebo/no intervention.
-
Beta‐blockers of any dose or duration versus another drug intervention.
-
Beta‐blockers of any dose or duration versus endoscopic interventions (sclerotherapy or band ligation).
We will also include randomised clinical trials with co‐interventions if received by both intervention arms.
Types of outcome measures
The primary outcome measures at maximal follow‐up will be:
1. All cause mortality: number of deaths irrespective of cause.
2. Mortality due to gastrointestinal bleeding: any death occurring within six weeks of time zero, that is, the time of admission of the patient to hospital with gastrointestinal bleeding, will be considered as related to gastrointestinal bleeding regardless of the mode (de Franchis 2001).
3. Mortality due to variceal rebleeding: any death occurring within six weeks of time zero, that is, the time of admission of the patient to hospital due to variceal rebleeding, will be considered as related to variceal rebleeding, regardless of the mode (de Franchis 2001). Variceal bleeding is diagnosed as endoscopy proven active bleeding; or signs of recent bleeding ('white nipple' or 'clot that could not be washed off'); or varices without other potential bleeding source (de Franchis 2001).
4. Number of patients experiencing gastrointestinal bleeding, irrespective of the site of bleeding (oesophageal variceal bleeding; gastric variceal bleeding; portal hypertensive gastropathy; duodenal variceal bleeding; colonic variceal bleeding).
5. Number of patients experiencing variceal rebleeding, that is, endoscopy proven active bleeding; or signs of recent bleeding ('white nipple' or 'clot that could not be washed off'); or varices without other potential bleeding source (de Franchis 2001).
The secondary outcome measures at maximum follow‐up will be:
6. Portal vein pressure.
7. Adverse events. Adverse events will be defined as any untoward medical occurrence not necessarily having a causal relationship with the treatment, but resulting in a dose reduction or discontinuation of the treatment. Severe adverse events will be defined according to the ICH guidelines (ICH‐GCP 1997) as any event that would increase mortality; is life‐threatening; requires inpatient hospitalisation; results in a persistent or significant disability; or any important medical event, which may jeopardise the patient or requires intervention to prevent it. We also plan to compare the frequency of a specific adverse event during the intervention with beta‐blockers.
8. Quality of life.
9. Cost‐effectiveness.
Search methods for identification of studies
We will search The Cochrane Hepato‐Biliary Group Controlled Trials Register, The Cochrane Central Register of Controlled Trials in The Cochrane Library, MEDLINE, EMBASE, and The Chinese Biomedical Database using the preliminary search strategies proposed in Appendix 1. We also plan to search LILACS with the help of a Spanish‐speaking member of The Cochrane Collaboration.
We will identify further trials by reading the reference lists of the identified studies. We will write to the principal authors of the identified randomised clinical trials and enquire about additional trials of which they might be aware. We will also write to the main pharmaceutical companies producing beta‐blockers to obtain any unpublished randomised clinical trials they may know.
Data collection and analysis
We will conduct the review according to the present protocol and the recommendations by The Cochrane Reviewers' Handbook (Clarke 2003).
We will list identified trials and will make an evaluation whether the trials fulfil the inclusion criteria. We will also list the excluded trials with the reasons for exclusion. WC, SLF, and DN will extract the data, and CG will arbitrate in case of disagreements.
Assessment of methodological quality
We will define the methodological quality as the confidence that the design and report restrict bias in the intervention comparison (Moher 1998; Kjaergard 2001). We will assess the methodological quality by four separate components (Kjaergard 2001) due to the risk of overestimation of intervention effects in randomised trials with inadequate components (Schulz 1995; Moher 1998; Kjaergard 2001). We will analyse the influence of methodological quality by four separate components:
Generation of the allocation sequence
-
Adequate, if the allocation sequence was generated by a computer or random number table. Drawing of lots, tossing of a coin, shuffling of cards, or throwing dice will be considered as adequate if a person who was not otherwise involved in the recruitment of participants performed the procedure.
-
Unclear, if the trial was described as randomised, but the method used for the allocation sequence generation was not described.
-
Inadequate, if a system involving dates, names, or admittance numbers were used for the allocation of patients. These studies are known as quasi‐randomised and will be excluded from the present review when assessing beneficial effects.
Allocation concealment
-
Adequate, if the allocation of patients involved a central independent unit, on‐site locked computer, identically appearing numbered drug bottles or containers prepared by an independent pharmacist or investigator, or sealed envelopes.
-
Unclear, if the trial was described as randomised, but the method used to conceal the allocation was not described.
-
Inadequate, if the allocation sequence was known to the investigators who assigned participants or if the study was quasi‐randomised.
Blinding (or masking)
-
Adequate, if the trial was described as double blind and the method of blinding involved identical placebo or active drugs. Due to beta‐blockers' effect on the heart rate we are well aware that it may be very difficult to properly blind such trials.
-
Unclear, if the trial was described as double blind, but the method of blinding was not described.
-
Not performed, if the trial was not double blind.
Follow‐up
-
Adequate, if the numbers and reasons for dropouts and withdrawals in all intervention groups were described or if it was specified that there were no dropouts or withdrawals.
-
Unclear, if the report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated.
-
Inadequate, if the number or reasons for dropouts and withdrawals were not described.
Furthermore, we will register whether or not the randomised clinical trials have used an 'intention‐to‐treat' analysis (Gluud 2001) and a sample size calculation.
Data extraction
We will extract the following characteristics from each randomised clinical trial: primary author, number of patients randomised, patient inclusion and exclusion criteria, methodological quality, sample size estimation, 'intention‐to‐treat' analysis, intervention regimens, mean age, proportion of males, etiology of cirrhosis, Child‐Pugh grade of cirrhosis (Pugh 1973), number of ascites, serum albumin, prothrombin time, serum bilirubin, serum creatinine, size of oesophageal varices, previous treatment for portal hypertension, time to follow up, number of outcomes and the number and type of adverse events. We will write to the principal investigator of the trial for further information in case we can not find the relevant data.
Statistical methods
We will perform the analyses in RevMan Analyses. The analyses will include all patients irrespective of compliance or follow‐up, according to the 'intention‐to‐treat' principle, and using the last reported observed response ('carry forward'). Regarding all cause mortality, we will perform both a 'worst‐best‐case scenario' analysis, which considers all dropout patients in the beta‐blocker group as dead and the dropout patients in the control group as alive, and a 'best‐worst‐case scenario' analysis, which considers all dropout patients in the beta‐blocker group as alive and the dropout patients in the control group as dead. In the assessment of the portal vein pressure response and quality of life, we will perform per protocol analyses. We plan to perform the meta‐analysis by both random effects (DerSimonian 1986) and fixed effect model (DeMets 1987). We will only report the results of fixed effect model if there are no significant differences between the two methods, otherwise we will report the results produced by both models. We will explore evidence of publication bias and other biases by funnel plot analyses (Egger 1997). We will present the binary outcome measures as relative risks (RR) with 95% confidence intervals (CI), and the continuous outcome measures as weighted mean differences (WMD) with 95% CI.
As stated under 'Type of participants', trials in which all patients have cirrhosis and trials in which more than 80 per cent have cirrhosis will be analysed separately. However, if the treatment effect does not show significant heterogeneity (P < 0.1) in the two groups of trials and the estimate of intervention effect does not differ significantly (P < 0.05), we will pool the trials and analyse them as one group in the sensitivity analyses described below.
We will compare the trials with beta‐blockers versus placebo/no intervention to the trials with beta‐blockers versus endoscopic interventions (sclerotherapy or band ligation). In addition, variceal rebleeding will be analysed according to the site of haemorrhage (not reported, reported as oesophageal variceal bleeding, and reported as gastric bleeding).
Subgroup and sensitivity analyses
We will perform the following subgroup/sensitivity analyses based on:
1. Methodological quality of the randomised clinical trials ‐ we will compare the trials with adequate methodological quality to the trials with inadequate methodological quality.
2. Reduction of heart rate by beta‐blocker ‐ we will compare the intervention effect of trials in which beta‐blocker reduced the heart rate 20 per cent or more to the trials in which the reduction of heart rate by beta‐blocker was less than 20 per cent or not reported.
3. Selectivity of beta‐blocker ‐ we will compare the trials assessing non‐selective beta‐blockers to the trials assessing beta1‐selective beta‐blockers.
4. We will compare the trials with co‐interventions to the trials without co‐interventions.
5. Etiology of cirrhosis ‐ we will compare the effects of beta‐blockers in different etiologies of liver cirrhosis.
6. Child‐Pugh grade (Pugh 1973) ‐ we will compare the effects of beta‐blockers in different Child‐Pugh grades (A, B, and C).
7. We will compare the trials in which the site of haemorrhage is defined as variceal bleeding to those in which it is not reported.
8. Size of oesophageal varices ‐ we will compare the effects of beta‐blockers in patients with different size of oesophageal varices (small, medium, large) at entry into the trials.
9. Publication status ‐ we will compare full manuscript trials to all other identified trials.
Due to the large number of comparisons planned we will interpret any significant findings conservatively. The main analysis will focus on the primary outcome measures in trials with adequate methodology.
