SEARCH

SEARCH BY CITATION

Keywords:

  • β-blockers;
  • combined therapy;
  • meta-analysis;
  • variceal ligation;
  • variceal rebleeding

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Background & Aims

Combined therapy with endoscopic variceal ligation (EVL) and β-blockers ± isosorbide mononitrate (ISMN) is currently recommended to prevent variceal rebleeding. However, the role of this combined therapy has been challenged by some studies. We performed a systematic review to assess the value of combined therapy with EVL and β-blockers ± ISMN as compared with each treatment alone to prevent rebleeding.

Methods

Databases, references and meeting abstracts were searched to retrieve randomized trials comparing combined therapy with EVL and β-blockers ± ISMN vs either treatment alone, to prevent variceal rebleeding in cirrhosis. Random-effects model was used for meta-analysis.

Results

We identified five studies comparing EVL alone or combined with drugs, including a total of 476 patients. Combination therapy reduced overall rebleeding [risk ratios (RR) = 0.44, 95% confidence interval (CI) = 0.28–0.69], and showed a trend towards lower mortality (RR = 0.58, 95% CI = 0.33–1.03), without increasing complications. We identified four trials comparing drugs alone or associated with EVL, including 409 patients. All used β-blockers plus ISMN. Variceal rebleeding decreased with combined therapy (P < 0.01) but rebleeding from oesophageal ulcers increased (P = 0.01). Overall, there was a trend towards lower rebleeding (RR = 0.76, 95% CI = 0.58–1.00) without effect on mortality (RR = 1.24, 95% CI = 0.90–1.70).

Conclusions

The addition of drug therapy to EVL improves the efficacy of EVL alone. However, the addition of EVL to β-blockers and ISMN achieves a non-significant decrease of rebleeding with no effect on mortality. Although combination therapy with EVL plus β-blockers ± ISMN is adequate to prevent rebleeding, β-blockers + ISMN alone may be a valid alternative.

Abbreviations
CI

confidence interval

EST

endoscopic sclerotherapy

EVL

endoscopic variceal ligation

ISMN

isosorbide mononitrate

RCTs

randomized controlled trials

RR

risk ratios

Patients who have recovered from acute variceal haemorrhage have a median rebleeding rate of 63% within 1–2 years and mortality of 33% [1]. For many years, nonselective β-blockers and endoscopic sclerotherapy (EST) were the first-line therapy to prevent rebleeding [1, 2]. However, endoscopic variceal ligation (EVL) has been shown to improve the safety and the efficacy of EST and is the current endoscopic treatment of choice [2, 3]. Adding EST to EVL to obliterate perforating veins has shown no additional beneficial effects as compared with EVL alone [4, 5]. Pharmacological therapy also has been improved in recent years. The combination of β-blockers and isosorbide mononitrate (ISMN) enhances the reduction of portal pressure induced by β-blockers [6]. This combined drug therapy has been shown to be superior to β-blockers alone and to EST [7, 8]. Meta-analyses of trials comparing β-blockers + ISMN with EVL have shown no significant differences in rebleeding or survival [9-11]. A recent study suggests an improvement in survival favouring combined drug therapy over EVL in the long-term follow-up [12].

Combining endoscopic and pharmacological therapies seems a rational approach because β-blockers may protect against rebleeding before variceal obliteration and would prevent variceal recurrence [13, 14]. Several randomized controlled trials (RCTs) have shown that this combination therapy is more effective in preventing rebleeding than EVL alone [15, 16], and this was confirmed in a subsequent meta-analysis which included both EVL and EST in the endoscopic therapy arm [17]. However, the effect of combination therapy on survival is unclear [15]. Furthermore, recent studies and subsequent meta-analysis have challenged the superiority of combination therapy over EVL [10, 11, 18]. It is also unclear whether EVL adds to the effect of β-blockers plus nitrates [19-21]. The present up-dated systematic review and meta-analysis of RCTs was performed to assess whether combined treatment with EVL and drug therapy with β-blockers ± ISMN improve the efficacy of either EVL alone or drug therapy alone for the prevention of rebleeding.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Retrieval of pertinent studies was based on electronic searches performed on MEDLINE, EMBASE, the Cochrane Central Register of Controled Trials, the Cochrane Database of Systematic Reviews, Pubmed and ClinicalTrials.gov, until March 2012. Eligible studies were RCTs evaluating combined therapy with β-blocker ± nitrates and EVL, compared with either drugs or EVL alone on the prevention of variceal rebleeding. The electronic search strategy used the terms: band ligation, ligation, β-blockers, propranolol, nadolol, oesophageal varices and rebleeding. The studies were eligible irrespective of publication status (in full or abstracts) or language. We performed a hand search of reference lists of the recovered articles, editorials and reviews, and of abstracts presented at the meetings of the Digestive Disease Week, American College of Gastroenteroloy, American Association for the study of the Liver, European Association for the Study of Liver Diseases and the British Society of Gastroenterology over the past 10 years.

Inclusion criteria

Studies were included when the following criteria were fulfilled: (i) randomized, controlled trials where one arm received β-blockers ± ISMN with EVL, and another arm received either EVL alone or β-blockers ± ISMN alone; (ii) previous variceal bleeding, (iii) cirrhosis, (iv) age over 18 years. Studies with one or more of the following exclusion criteria were rejected: those dealing with primary prevention of bleeding or with treatment of acute variceal bleeding, those dealing with the treatment of gastric varices or portal hypertensive gastropathy, those not detailing rebleeding or mortality and those involving patients with non-cirrhotic portal hypertension.

Trial selection, data extraction and quality assessment

Two reviewers independently selected the studies by screening the titles and abstracts of studies to identify the inclusion criteria. Disagreement was resolved through consensus decision or by consulting a third party in case of no agreement. If necessary, contact with the original investigator for further data was attempted. Data concerning studies and patients' characteristics, interventions, outcomes, follow-up and loss to follow-up were extracted. We assessed the risk of bias by using a modified established standard based on allocation concealment, adherence of analysis to the intention-to-treat principle and Jadad scale.

The following data were extracted: study population (age, gender, cause of liver disease, Child-Pugh score); intervention, outcomes, study risk of bias, follow-up and loss to follow-up, withdrawal, intolerance or adverse effects and non-compliance.

Statistical analysis

Primary end points were overall rebleeding and all cause mortality. Rebleeding was defined as bleeding from any upper gastrointestinal source after control of the first haemorrhage. We defined mortality as death from any cause. Secondary end-points were variceal rebleeding and rebleeding related to oesophageal ulcers. We also explored adverse effects.

The effects of treatment were expressed as risk ratios (RR) comparing the outcome rate in the group treated with combined therapy with EVL plus drugs and, respectively, the groups treated with single therapy either EVL or drugs. Therefore, RR < 1 indicated benefit from combined therapy. Pooled RR and their 95% confidence intervals (CI) were estimated using a random effect model with the Mantel–Haenszel method. Heterogeneity was assessed with the I2 statistic that indicates the percentage of total variation across studies not attributable to random error. An I2 of >60% was considered significant heterogeneity. A 2-sided α error <0.05 was considered to be statistically significant.

A priopri subgroup analysis was performed to explore subgroups related to stage of chronic liver disease and length of follow. Studies were arbitrarily divided according to whether or not they had ≥20% patients with Child–Pugh class C and to whether they had a length of follow-up ≥15 months. All analyses were performed using RevMan software (Review Manager v5.1.4, Copenhagen, Denmark).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The search strategy identified 760 potential references (Fig. 1). After reviewing the titles and abstracts, 685 studies were excluded. A further 66 studies that did not fulfil our inclusion criteria were excluded: eight of them were RCTs comparing EVL with combined drug therapy using non-selective β-blockers and ISMN [22-29]. Finally, 10 RCTs comparing combined therapy with EVL plus pharmacological treatment with β-blockers ± ISMN vs either EVL or drug therapy alone, for the prevention of variceal rebleeding, were identified [15, 16, 18-20, 29-32]. We excluded two studies published as an abstract because one was not an RCT [32], and non-cirrhotic portal hypertension could not be excluded in the other [33]. One study randomized patients into four arms treated with propranolol, propranolol plus ISMN, EVL and EVL combined with propranolol plus ISMN [29]. The arm treated with EVL was included as a control group to analyse EVL vs EVL plus drugs, and the arm treated with propranolol plus ISMN was included as a control group to analyse drugs vs drugs plus EVL. The arm treated with EVL plus propranolol and ISMN was included as the combined therapy group in both analyses. One study included 15% of patients (n = 26) with non-cirrhotic portal hypertension [18]. This subset of patients was excluded because the information provided allowed to include only patients with cirrhosis.

image

Figure 1. Flow diagram of studies through the process or retrieval. EVL, endoscopic variceal ligation; ISMN, isosorbide mononitrate.

Download figure to PowerPoint

Finally, five studies were included for analysis of EVL vs EVL plus drugs. They included a total of 450 patients [15, 16, 18, 29, 30]. Of these, 228 were randomized to EVL alone and 222 to EVL plus drugs. Four studies were included to analyse drugs vs drugs plus EVL. They included a total of 409 patients [19, 20, 29, 31], 206 of whom were randomized to drugs alone and 203 to drugs plus EVL.

Study characteristics

Most studies were published as full-text articles [15, 16, 18-20, 29, 31] and only one as an abstract [30]. Only two trials were multicentre [16, 19]. Treatments are summarized in Table 1. The EVL technique varied among the RCTs. EVL was performed with multiband devices in all RCTs, although a pneumatic-active ligating device with overtube was used in one study for patients with active bleeding [15]. EVL sessions were repeated at intervals that varied between 10 and 12 days [16, 30] and 4 weeks [20, 31], until varices were obliterated. The number of elastic bands placed in each session was not specified in most RCTs [15, 16, 19, 20, 29]. Obliteration of varices was defined only in some studies [15, 18, 19, 31]. In those reporting this data, it was achieved in a proportion of patients that varied from 40 to 100% [15, 18, 19, 31]. Surveillance endoscopy after variceal obliteration varied among trials. Whether proton pump inhibitors or sucralfate were used with EVL was not reported in most RCTs [16, 20, 29, 30]. The use of conscious sedation for EVL was reported only in some studies [15, 16, 19]. The index bleeding was treated with EVL in most RCTs, although EVL or EIS were used in several studies [16, 19, 29].

Table 1. Characteristics of trials
Study, ReferenceTrial designPatientsEndoscopic sessions to variceal obliterationTime to start of β-blocker therapyDrug therapy Mean dose (range or SD)Follow-up Median or mean (range or SD)
ScreenedRandomized
  1. EST, endoscopic injection sclerotherapy; EVL, endoscopic variceal ligation; HCC, hepatocelullar carcinoma; ISMN, isosorbide mononitrate; NA, not available; SD, standard deviation; TIPS, transjugular intrahepatic portosystemic shunt.

EVL and β-blockers ± ISMN (Group A) vs EVL alone (Group B)
Lo et al. [15]

Single centre

No blinding

183

122 (67%)

Group A: 60

Group B: 62

Group A: 3.2 ± 1.0

Group B: 3.8 ± 1.2

2 daysNadolol: 60 ± 18 mg/day plus sucralfate

21.5 months

Group A: 22 m

Group B: 21 m

De la Peña et al. [16]

Multicentre

No blinding

240

84 (35%)

Group A: 43

Group B: 37

Group A:

3 (1–7)

Group B:

3 (1–7)

5 daysNadolol: 58 (10–120) mg/day

16.4 months

Group A: 17.5 ± 7.8

Group B: 15 ± 8

Kumar et al. [18]

Single centre

No blinding

653

177 (27%)

Group A: 88 (76 cirrhosis)

Group B: 89 (75 cirrhosis)

Group A: 4.6 ± 1.9

Group B: 4.6 ± 1.6

5 days

Propranolol: 120 (40–320) mg/day

ISMN: 40 (20–40) mg/day

15 ± 12 months

Group A: 15 ± 12

Group B: 15 ± 11

Ahmad et al. [29]

Single centre

No blinding

NA

160

Group A: 37

Group B: 39

Group A: 3.0 ± 1.3

Group B: 3.5 ± 1.4

5 days

Propranolol: 53 ± 21 mg/day

ISMN: 35 ± 9 mg/day

274 days

Group A: 292 (49–609) days

Group B: 256 (32–164) days

Sollano et al. [30]

Single centre

No blinding

NA

31

Group A: 16

Group B: 15

NAAfter variceal erradicationPropranolol: 10 mg tid

Until a rebleeding episode

Group A: 282 ± 40 days

Group B: 292 ± 300 days

β-blockers ± ISMN and EVL (group A) vs β-blockers ± ISMN alone (group B)
Lo et al. [20]

Single centre

No blinding

286

120 (42%)

Group A: 60

Group B: 60

Group A: 3.8 ± 0.648 h

Groups A & B:

Nadolol: 40 (20–120) mg/day

ISMN: 20 (0–40) mg/day

23 months

Group A: 22.7 months

Group B: 23.2 months

Villanueva et al. [31]

Single centre

No blinding

83

59 (71%)

Group A: 29

Group B:30

Group A:

4 (1–7)

5 days

Group A:

Nadolol: 98 ± 44 mg/day

ISMN: 45 ± 23 mg/day

Group B:

Nadolol: 83 ± 30 mg/day

34 months Group A: 35 ± 19 months

Group B: 34 ± 12 months

García Pagan et al. [19]

Multicentre

No blinding

342

160 (47%)

Group A: 80

Group B: 78

Group A: 2 ± 15 days

Group A:

Nadolol: 102 ± 52 mg/day

ISMN: 36 ± 9 mg/day

Group B:

Nadolol: 90 ± 48 mg/day

ISMN: 36 ± 10 mg/day

15 months

Group A: 14.4 ± 7.8 months

Group B: 15.3 ± 8.3 months

 Ahmad et al.

Single centre

No blinding

NA

160

Group A: 37

Group B: 35

Group A: 3.0 ± 1.35 days

Group A:

Propranolol: 50 ± 21 mg/day

ISMN: 33 ± 10 mg/day

Group B:

Propranolol: 53 ± 21 mg/day

ISMN: 35 ± 9 mg/day

289 days

Group A: 287 (45–619) days

Group B: 292 (49–609) days

β-blockers were started at 5 days from index bleeding in most trials (Table 1) and after variceal obliteration in one trial [30]. Nadolol was used in five trials and Propranolol was used in the others (Table 1). The dose of β-blocker was titrated to reduce resting heart rate by 25% or to 55 beats/min. [15, 16, 20, 29, 30], or up to the maximum tolerated dose without reducing heart rate bellow 55 beats/min [18, 19, 31]. No trial documented the proportion of patients reaching these targets and only a minority of trials reported heart rates achieved at follow-up [16, 31]. Baseline HVPG measurements were performed in three trials [18, 19, 31] and were repeated during the follow-up in 2 [19, 31]. Patients received ISMN in addition to β-blocker in all the trials for analysis of drugs vs drugs + EVL [19, 20, 29, 31] and only in two trials for the analysis of EVL vs EVL + drugs [18, 29]. The dose of ISMN was titrated up to 20 mg twice daily in all trials except one in which it was titrated up to 40 mg twice daily [31]. In one study prazosin was used in addition to nadolol in patients without haemodynamic response to nadolol plus ISMN [31]. One study used sucralfate in addition to β-blocker [15].

The mean follow-up of the trials varied from 14 to 35 months (Table 1). Demographic data and data on the cause and severity of liver disease are reported in Table 2. Alcohol was the cause of liver disease in 0–70% of cases. Patients with Child's class C comprised a median of 22.5% of patients (range 13–40%). Study risk of bias is summarized in Table 3. No trial used double-blinding. Concealment of allocation was unclear only in one trial published as an abstract [30]. Intention-to-treat analysis was adequate in most trials.

Table 2. Clinical characteristics of patients included in the studies
StudyGroup A/BGroup AGroup B
N Age (range or SD), yearsMales (%)Alcohol (%)ABCABC
  1. EVL, endoscopic variceal ligation; ISMN, isosorbide mononitrate; NA, not available; SD, standard deviation.

EVL and β-blockers ± ISMN (Group A) vs EVL alone (Group B)
Lo et al. [15]60/6253 ± 12/51 ± 1175/7932/2811 (18)30 (50)19 (31)12 (19)28 (45)22 (35)
De la Peña et al. [16]43/3760 (36–75)/60 (18–75)77/7370/636 (14)25 (58)12 (28)6 (16)20 (54)11 (30)
Kumar et al. [18]76/7542 ± 14/41 ± 1488/8543/4035 (46)31 (41)10 (13)26 (35)34 (45)15 (20)
Ahmad et al. [29]37/3950 ± 11/52 ± 1064/810/34 (11)27 (73)6 (16)7 (18)23 (59)9 (23)
Sollano et al. [30]16/15NANANANANANANANANA
β-blockers ± ISMN alone (Group A) vs β-blockers ± ISMN alone (Group B)
Lo et al. [20]60/6054 ± 10/52 ± 1175/7035/2520 (33)29 (48)11 (19)21 (35)31 (52)8 (13)
Ahmad et al. [29]37/3550 ± 11/52 ± 1060/810/04 (11)27 (73)6 (16)2 (6)19 (54)14 (40)
Villanueva et al. [31]29/3064 ± 12/62 ± 1265/6334/431 (3)24 (83)4 (14)4 (11)20 (67)6 (20)
García Pagan et al. [19]80/7857 ± 12/56 ± 1168/8150/5516 (20)46 (58)18 (22)18 (23)42 (54)18 (24)
Table 3. Risk of bias of included trials
StudyRCTPublication typeAllocation concealmentaIntention-to-treatbWithdrawals and dropoutsJadad scorec
  1. a

    Adequate, when the method used prevented the investigators from knowing which treatment was next before allocation (central randomization, sealed envelopes).

  2. b

    Adequate, when explicit mention that all randomized patients were analyzed in the group to which they were originally allocated, regardless of the treatment received.

  3. c

    One point for randomization, adequate method of randomization, double blinding, adequate method of double blinding, and description of withdrawals/dropouts.

  4. EVL, endoscopic variceal ligation; ISMN, isosorbide mononitrate; RCT, randomized controlled trial.

EVL and β-blockers ± ISMN vs EVL alone
Lo et al. [15]YesArticleAdequateAdequateYes3
De la Peña et al. [16]YesArticleAdequateAdequateYes3
Kumar et al. [18]YesArticleAdequateAdequateYes3
Ahmad et al. [29]YesArticleAdequateUnclearYes3
Sollano et al. [30]YesAbstractUnclearUnclearNo1
β-blockers ± ISMN and EVL vs β-blockers ± ISMN alone
Lo et al. [20]YesArticleAdequateAdequateYes3
Ahmad et al. [29]YesArticleAdequateUnclearYes3
Villanueva et al. [31]YesArticleAdequateAdequateYes3
García-Pagan et al. [19]YesArticleAdequateAdequateYes3

Results for the comparison of combined therapy vs EVL alone

Pooling all five studies comparing combined therapy with EVL and β-blockers ± ISMN vs EVL alone resulted in a significant reduction in overall rebleeding rate favouring combined therapy (17% vs 32%), without heterogeneity (Fig. 2). Combination therapy was significantly more effective than EVL in two trials [15, 16], and had a no-significant trend towards more efficacy in the remaining trials [18, 29, 30]. The pooled rate of oesophageal variceal rebleeding was also lower with combined therapy than with EVL alone (14% vs 25%; RR = 0.51, 95% CI = 0.32–0.82, I2 = 4%). Subgroup analysis showed a favourable effect of combination therapy on rebleeding, mainly in trials with a greater proportion of Child's class C and in trials with a longer follow-up (Table 4).

Table 4. Subgroup analyses by follow-up length (<15 months or ≥15 months) and Child class (<20% or ≥20% of patients Child C)
 Studies, nPatients, nRisk ratio (95% CI)P-valueHeterogeneity (I²)
  1. EVL, endoscopic variceal ligation; ISMN, isosorbide mononitrate; NA, not applicable.

EVL and β-blockers ± ISMN vs EVL alone
Rebleeding54500.44 (0.28–0.69)0.00030
Follow-up <15 months21070.69 (0.46–1.05)0.080
Follow-up ≥15 months33430.53 (0.35–0.80)0.00315
Child C >20%22020.45 (0.29–0.70)0.00050
Child C ≤20%22170.76 (0.45–1.29)0.310
Mortality54500.58 (0.33–1.03)0.070
Follow-up <15 months21070.85 (0.35–2.03)0.710
Follow-up ≥15 months33430.59 (0.33–1.04)0.070
Child C >20%22020.61 (0.33–1.13)0.124
Child C ≤20%22170.79 (0.34–1.84)0.590
β-blockers ± ISMN and EVL vs β-blockers ± ISMN alone
Rebleeding44090.76 (0.58–1.00)0.050
Follow-up <15 months1720.84 (0.37–1.93)0.68NA
Follow-up ≥15 months33370.76 (0.57–1.01)0.060
Child C >20%22300.81 (0.54–1.21)0.300
Child C ≤20%21790.73 (0.51–1.06)0.100
Mortality44091.24 (0.90–1.70)0.180
Follow-up <15 months1721.10 (0.41–2.96)0.84NA
Follow-up ≥15 months33371.26 (0.90–1.75)0.180
Child C >20%22301.06 (0.62–1.80)0.840
Child C ≤20%21791.35 (0.91–2.01)0.180
image

Figure 2. Pooled meta-analysis of trials comparing EVL vs EVL plus drugs. (A) Forest plots for overall rebleeding. (B) Forest plots for mortality; EVL, endoscopic variceal ligation. (*) From the study by Kumar et al. [18], only cirrhotic patients were included.

Download figure to PowerPoint

Pooling mortality rates resulted in a trend favouring the combination of β-blockers ± ISMN with EVL over EVL alone, although the difference was not significant (10% vs 16%) and there was no heterogeneity (Fig. 2). Subgroup analysis showed no significant differences on mortality (Table 4).

The rate of adverse events increased with the association of β-blockers ± ISMN to EVL (18% vs 7%), although the difference was not significant (RR = 2.30, 95% CI = 0.84–6.28) and there was some heterogeneity (I2 = 51%) [15, 16, 18]. There was no intervention-related mortality. Bleeding from oesophageal ulcer was the most severe adverse event observed and it was more frequent in patients treated only with EVL, although the difference was not significant (5% vs 2%; RR = 0.46, 95% CI = 0.16–1.31, I2 = 0%) [15, 16, 18, 29].

Results for the comparison of combined therapy vs drug therapy alone

Pooling all four studies comparing β-blockers + ISMN combined with EVL vs β-blockers + ISMN alone resulted in a clear, although not significant, trend towards benefit with combination therapy for overall rebleeding (29% vs 37%), without heterogeneity (Fig. 3). The pooled risk of oesophageal variceal rebleeding was significantly lower with combined therapy than with drugs alone (21% vs 34%; RR = 0.51, 95% CI = 0.33–0.80, I2 = 0%).

image

Figure 3. Pooled meta-analysis of trials comparing drugs vs drugs plus EVL. (A) Forest plots for overall rebleeding. (B) Forest plots for mortality. EVL, endoscopic variceal ligation.

Download figure to PowerPoint

Pooled rates of mortality showed no statistically significant differences between groups, although higher rates were observed with combined therapy than with β-blockers + ISMN alone (28% vs 23%), without heterogeneity (Fig. 3). Subgroup analysis showed no significant differences between groups on rebleeding or mortality (Table 4).

Adverse events were available for analysis in three trials [19, 20, 31]. The pooled rate of adverse events was higher with combined therapy than with drugs alone (56% vs 45%; RR = 1.53, 95% CI = 1.03–2.10), but with significant heterogeneity (I2 = 82%). There was no intervention-related mortality. Bleeding from oesophageal ulcer was available for analysis in the four trials and was significantly more frequent with EVL plus drugs than with drugs alone (0.5% vs 6%; RR = 5.51, 95% CI = 1.45–20.97, I2 = 0%).

Mortality

Our results suggest that mortality may be improved by adding drug therapy to treatment with EVL only, while it is not improved with the addition of EVL to drug therapy alone. This suggests that drug therapy with β-blockers ± ISMN may have a greater effect on survival than EVL. To investigate this possibility, we performed an exploratory analysis pooling the RCTs that compared EVL alone vs β-blockers ± ISMN either alone or combined with EVL. Eight trials comparing EVL alone vs β-blockers ± ISMN alone were identified [22-29]. Of these, one did not provide data on mortality [25], and another was a 4-arm study that had been already included [29]. We pooled the six remaining RCTs with the five trials comparing EVL alone vs β-blockers + ISMN with EVL. Overall, these 11 RCTs allocated 1219 patients, 589 treated with EVL alone and 630 treated with drugs ± EVL (Fig. 4). The pooled risk of mortality was significantly lower in the group treated with β-blockers + ISMN, alone or combined with EVL, than in the group treated only with EVL (16% vs 21%; RR = 0.79, 95% CI = 0.62–0.99), without heterogeneity (Fig. 4).

image

Figure 4. Pooled meta-analysis for mortality of trials comparing EVL alone vs β-blockers ± ISMN either alone or combined with EVL, for secondary prophylaxis of variceal bleeding. EVL, endoscopic variceal ligation; ISMN, isosorbide mononitrate.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The current systematic review shows that, in the high-risk setting of preventing variceal rebleeding, the addition of drug therapy with β-blockers ± ISMN to EVL markedly improves the efficacy of EVL alone, achieving a relevant reduction of rebleeding risk. The beneficial effect of combination therapy was observed considering both variceal rebleeding and overall rebleeding. As compared with previous systematic reviews assessing EVL alone or in combination with drugs [10, 17, 34], the current analysis incorporates trials previously unavailable [29, 31] and also some full-text articles that previously were only available as abstract [18-20]. Consequently, our study has a larger sample size than the largest previously reported [10], with 450 patients finally included for the analysis. This has allowed us to demonstrate that the previously observer trend towards a lower rebleeding with combination therapy [10], is actually a significant difference favouring such combined therapy. The addition of drug therapy to EVL decreased rebleeding rate in all trials. However, the difference was significant in only two RCTs [15, 16]. The RR was >0.75 only in one study that had the lowest proportion of patients included from among those eligible (only 27% of cases) [18]. Such a high rate of exclusions may have induced unnoticed selection bias in this study. The present review also shows that combination therapy reduced mortality. However, although the death risk was markedly decreased with the association of drugs to EVL, the difference failed to reach statistical significance. The present results were robust and not affected by subgroup analysis by severity of liver dysfunction or the length of follow-up. The effect on mortality did not reach statistical significance probably because of a type II error, suggested by the relatively low rate of mortality achieved with EVL (only 16%).

The beneficial effect achieved with the addition of drug therapy to EVL is clinically sound and supported by a large amount of previous data. The efficacy of drug therapy is related to its effect attenuating portal hypertension [35, 36]. This provides an additional benefit to local therapies, such as EVL, which are directed at eradicating varices but do not decrease portal pressure and, consequently, have no effect on other complications of portal hypertension [13, 14]. In the setting of secondary prophylaxis, a number of studies have demonstrated that reducing portal pressure with drugs to <12 mmHg or ≥20% from baseline value decreases rebleeding risk to below 10% of cases [35, 36]. Furthermore, reaching these targets is associated with a significant decrease in the risk to develop other complications related to portal hypertension such as ascites, SBP or hepatorenal syndrome and significantly improves survival [37, 38]. Moreover, the protective effect of nonselective β-blockers may be due not only to a reduction in portal pressure but also to other haemodynamic and non-haemodynamic mechanisms, such as a decrease in portocollateral blood flow and variceal pressure (increasing collateral resistance) and a reduction in bacterial translocation (increasing bowel motility) [39]. Indeed it has been shown that β-blockers can improve survival even when there is no effect on rebleeding [12].

The current review also analysed RCTs investigating whether the addition of EVL to drug therapy may improve the efficacy of drugs alone to prevent variceal rebleeding. Four RCTs, including 409 patients, were available for the analysis. All trials used β-blockers plus ISMN for drug therapy. The beneficial effect of combined therapy in this case was less clear. There was a significant reduction on risk of variceal rebleeding favouring the addition of EVL to drug therapy. This effect was consistent across all trials and without statistical heterogeneity. However, the reduction on overall rebleeding failed to achieve statistical significance. This could be because of a type II error, although other factors may account such as the higher incidence of rebleeding from oesophageal ulcers with the addition of EVL to drug therapy. Similar results were observed in the largest trial included [19]. This trial, showed a lower rate of variceal rebleeding with combined therapy but no significant differences in rebleeding from all sources because of bleeding from oesophageal ulcers induced by EVL [19]. In our study, the overall complications rate was also higher in patients treated with drugs plus EVL. Furthermore, mortality was slightly higher with combined therapy than with the use of drugs alone although the difference was not significant. These results suggest that the addition of EVL to drug therapy with β-blockers and ISMN achieves only a slight improvement of rebleeding with no effect on mortality. The association of β-blockers plus ISMN achieves a greater reduction of portal pressure than β-blockers alone and probably also achieves lower rebleeding rates [6-8]. Given the low residual risk of rebleeding in haemodynamic responders, no further treatment will be required in these patients [35, 36]. Furthermore, from available data it is not clear that the addition of EVL can improve efficacy in haemodynamic non-responders [35]. In fact, the addition of EVL to drug therapy in nonresponders did not improve the efficacy of β-blockers plus MNIS in the largest RCT included [19]. Further studies are required to clarify whether EVL can improve the efficacy of drug therapy with β-blockers plus MNIS with different approaches, such as the use of different techniques for EVL to avoid bleeding related to oesophageal ulcers, or the addition of EVL only in patients at high risk of failure with drugs alone.

The current meta-analysis shows that adding drug therapy with β-blockers ± ISMN to treatment only with EVL results in a trend towards an improvement in mortality, while the addition of EVL to drug therapy with β-blockers plus ISMN does not improve the effect of such drug treatment on mortality. To investigate whether drug therapy with β-blockers ± ISMN improves the effect of EVL alone on survival, in an exploratory analysis we pooled trials comparing EVL alone vs trials using β-blockers ± ISMN either alone or combined with EVL [19, 20, 22-29, 31]. This analysis showed that the probability of survival was significantly better in patients treated with drugs, whether administered alone or in combination with EVL, than in those treated only with EVL. Several factors may help to explain this effect on mortality. The current review shows that combined therapy with EVL and drugs is associated with a higher risk of complications than any of these treatments alone. However, severe complications are more commonly associated with EVL than with drugs and eventually can contribute to mortality [40]. It should also be taken into account that a large proportion of deaths are not related to rebleeding but to other complications of portal hypertension or liver dysfunction [41]. Both liver function and portal hypertension are not affected by therapies that act locally on varices but are not pathophysiologically oriented, such as EVL [9-11]. Accordingly, an improvement in mortality may occur more likely with therapies that improve portal hypertension [35, 36]. Future studies should investigate whether further improvements in survival can be achieved using drugs that induce a more pronounced reduction of portal pressure, such as carvedilol [42] or β-blockers plus prazosin [31], or by adding drugs that have a beneficial effect on liver function such as simvastatin [43].

Our meta-analysis has limitations intrinsic to this type of analysis. Relatively few studies are available for analysis, while the amount of the pooled data is important because results become more reliable and the margin for error decreases as the amount of data increases. Trials with negative findings are less likely to be published and none of the studies included was blinded. These factors induce a risk for bias. However, blinding is difficult for trials involving endoscopic procedures. Moreover, the heterogeneity of patient populations and different treatment protocols may also be a concern in our meta-analysis. However, summary estimates did not change when we stratified patients according to severity of cirrhosis and the duration of follow-up. Another point to take into account is that the definition for rebleeding differs among studies and can thus confound conclusions from this data set. The interpretation of adverse event rates can also be problematic and should cautious given that there were no standard definitions of adverse events [14].

In conclusion, the present systematic review shows that adding drug therapy with β-blockers ± ISMN to EVL significantly reduces the probability of rebleeding achieved with EVL alone, with a trend towards lower mortality. The addition of EVL to drug therapy with β-blockers plus ISMN decreases the probability of variceal rebleeding but increases bleeding from oesophageal ulcers, with a clear trend towards lower overall rebleeding and without effect on mortality. These results suggest that β-blockers ± ISMN should be added to patients treated with EVL. In patients treated with β-blockers plus ISMN future studies should define the ideal method of EVL, to avoid ulcer-related bleeding and to improve efficacy particularly in patients at high risk of failure with drug therapy alone.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Financial support: This study has been supported in part by a grant from the Fondo de Investigaciones Sanitarias (PI10/01552) and by the Fundació Investigació Sant Pau (CAIBER).

Conflict of interest: The authors do not have any disclosures to report.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    D'Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995; 22: 33254.
  • 2
    De Franchis R, Primignani M. Endoscopic treatment for portal hypertension. Semin Liver Dis 1999; 19: 43955.
  • 3
    Laine L, Cook D. Endoscopic ligation compared with sclerothepary for treatment of esophageal variceal bleeding. A meta-analysis. Ann Intern Med 1995; 123: 2807.
  • 4
    Singh P, Pooran N, Indaram A, Bank S. Combined ligation and sclerotherapy versus ligation alone for secondary prophylaxis of esophageal variceal bleeding: a meta-analysis. Am J Gastroenterol 2002; 97: 6239.
    Direct Link:
  • 5
    Karsan HA, Morton SC, Shekelle PG, et al. Combination endoscopic band ligation and sclerotherapy compared with endoscopic band ligation alone for the secondary prophylaxis of esophageal variceal hemorrhage: a meta-analysis. Dig Dis Sci 2005; 50: 399406.
  • 6
    Garcia-Pagán JC, Feu F, Bosch J, Rodés J. Propranolol compared with propranolol plus isosorbide-5-mononitrate for portal hypertension in cirrhosis. A randomized controlled study. Ann Intern Med 1991; 114: 86973.
  • 7
    Gournay J, Masliah C, Martin T, Perrin D, Galmiche JP. Isosorbide mononitrate and propranolol compared with propranolol alone for the prevention of variceal rebleeding. Hepatology 2000; 31: 123945.
  • 8
    Villanueva C, Balanzó J, Novella M, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996; 334: 16249.
  • 9
    Ding SH, Liu J, Wang JP. Efficacy of β-adrenergic blocker plus 5-isosorbide mononitrate and endoscopic band ligation for prophylaxis of esophageal variceal rebleeding: a meta-analysis. World J Gastroenterol 2009; 15: 21515.
  • 10
    Cheung J, Zeman M, Van Zanten SV, Tandon P. Systematic review: secondary prevention with band ligation, pharmacotherapy or combination therapy after bleeding from oesophageal varices. Aliment Pharamcol Ther 2009; 30: 57788.
  • 11
    Li L, Yu C, Li Y. Endoscopic band ligation versus pharmacological therapy for variceal bleeding in cirrhosis: a meta-analysis. Can J Gastroenterol 2011; 25: 14755.
  • 12
    Lo GH, Chen WC, Lin CK, et al. Improved survival in patients receiving medical therapy as compared with banding ligation for the prevention of esophageal variceal rebleeding. Hepatology 2008; 48: 5807.
  • 13
    Garcia-Tsao G, Sanyal A, Grace ND, Carey W; Practice Guidelines Committee of the American Association for the Study of Liver Diseases, the Practice Parameters Committee of the American College of Gastroenterology. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Hepatology 2007; 46: 92238.
  • 14
    De Franchis R. Revising consensus in portal hypertension: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2010; 53: 7628.
  • 15
    Lo GH, Lai KH, Cheng JS, et al. Endoscopic variceal ligation plus nadolol and sucralfate compared with ligation alone for the prevention of variceal rebleeding: a prospective, randomized trial. Hepatology 2000; 32: 461565.
  • 16
    De la Peña J, Brullet E, Sanchez-Hernández E, et al. Variceal ligation plus nadolol compared with ligation for prophylaxis of variceal rebleeding: a multicenter trial. Hepatology 2005; 41: 5728.
  • 17
    Gonzalez R, Zamora J, Gomez-Camarero J, et al. Meta-analysis: combination endoscopic and drug therapy to prevent variceal rebleeding in cirrhosis. Ann Intern Med 2008; 149: 10922.
  • 18
    Kumar A, Jha SK, Sharma P, et al. Addition of propranolol and isosorbide mononitrate to endoscopic variceal ligation does not reduce variceal rebleeding incidence. Gastroenterology 2009; 137: 892901.
  • 19
    Garcia-Pagan JC, Villanueva C, Albillos A, et al. Nadolol plus isosorbide mononitrate alone or associated with band ligation in the prevention of recurrent bleeding: a multicenter randomized controlled trial. Gut 2009; 58: 114450.
  • 20
    Lo GH, Chen WC, Chan HH, et al. A randomized, controlled trial of banding ligation plus drug therapy versus drug therapy alone in the prevention of esophageal variceal rebleeding. J Gastroenterol Hepatol 2009; 24: 9827.
  • 21
    Lo GH, Hernández-Gea V, Aracil C, Villanueva C. Management of varices in cirrhosis.. N Engl J Med 2010; 362: 23313.
  • 22
    Villanueva C, Miñana J, Ortiz J, et al. Endoscopic ligation compared with combined treatment with nadolol plus isosorbide mononitrate to prevent recurrent variceal bleeding. N Engl J Med 2001; 345: 64755.
  • 23
    Patch D, Sabin CA, Goulis J, et al. A randomized controlled trial of medical therapy versus endoscopic ligation for the prevention of variceal rebleeding in patients with cirrhosis. Gatroenterology 2002; 123: 10139.
  • 24
    Lo GH, Chen WC, Chen MH, et al. Banding ligation versus nadolol and isosorbide mononitrate for the prevention of esophageal variceal bleeding. Gastroenterology 2002; 123: 72834.
  • 25
    El-Tahawy M, Mahmoud S, Waked I, Selim K, Hassanein T. Band ligation versus medical therapy in prevention of variceal rebleeding: a randomized controlled study in schistosomal hepatic fibrosis and HCV coinfection. Gastyroenterol 2003; 124: A6634.
  • 26
    Sarin SK, Wadhawan M, Gupta R, et al. Evaluation of endoscopic variceal ligation (EVL) versus propanolol plus isosorbide mononitrate/nadolol (ISMN) in the prevention of variceal rebleeding: comparison of cirrhotic and noncirrhotic patients. Dig Dis Sci 2005; 50: 153847.
  • 27
    Shiha G, Menesy A, Seef S, et al. Endoscopic ligation compared with propranolol and combined isosorbide mononitrate (ISMN) and propranolol to prevent recurrent variceal bleeding: a prospective randomized trial. Gut 2005; 54(Suppl. 2): A6.
  • 28
    Romero G, Kravetz D, Argonz J, et al. Comparative study between nadolol and 5-isosorbide mononitrate vs endoscopic band ligation plus slerotherapy in the prevention of variceal rebleeding in cirrhotic patients: a randomized controlled trial. Aliment Pharmacol Ther 2006; 24: 60111.
  • 29
    Ahmad I, Khan AA, Alam A, et al. Propranolol, isosorbide mononitrate and endoscopic band ligation – alone or in varying combinations for the prevention of esophageal variceal rebleeding. J Coll Physicians Surg Pak 2009; 19: 2836.
  • 30
    Sollano JD, Melchor RJ, Chan M, Ismael AE. Propranolol prevents rebleeding after variceal ligation. Gastroinets Endosc 2002; 53: AB143.
  • 31
    Villanueva C, Aracil C, Colomo A, et al. Clinical trial: a randomized controlled study on prevention of variceal rebleeding comparing nadolol+ligation vs hepatic venous pressure gradient-guided pharmacological therapy. Alimen Pharmacol Ther 2009; 29: 397408.
  • 32
    Cheng WF, Shi RH, Chen XX, Ding J. Effects of propranolol on rebleeding following endoscopic esophageal variceal ligation: a prospective, controlled trial. World Chin J Digestol 2009; 17: 835.
  • 33
    Abdel-Rhaim AY, Abdel-Ghany MS, El-Kholy B. Band ligation alone versus band ligation and propranolol in the management of bleeding esophageal varices. Am J Gastroenterol 2000; 95: 2442.
    Direct Link:
  • 34
    Ravipati M, Katragadda S, Swaminathan PD. Pharmacotherapy plus endoscopic intervention is more effective than pharmacotherapy or endoscopy alone in the secondary prevention of esophageal variceal bleeding: a meta-analysis of randomized, controlled trials. Gastrointes Endosc 2009; 70: 65864.
  • 35
    D'Amico G, Garcia-Pagan JC, Luca A, Bosch J. Hepatic vein pressure gradient reduction and prevention of variceal bleeding in cirrhosis: a systematic review. Gastroenterology 2006; 131: 161124.
  • 36
    Albillos A, Bañares R, González M, Ripoll C, Gonzalez R. Value of the hepatic venous pressure gradient to monitor drug therapy for portal hypertension: a meta-analysis. Am J Gastroenterol 2007; 102: 111626.
    Direct Link:
  • 37
    Abraldes JG, Tarantino I, Turnes J, et al. Hemodynamic response to pharmacological treatment of portal hypertension and long-term prognosis of cirrhosis. Hepatology 2003; 37: 9028.
  • 38
    Villanueva C, López-Balaguer JM, Aracil C, et al. Maintenance of hemodynamic response to treatment for portal hypertension and influence on complications of cirrhosis. J Hepatol 2004; 40: 75765.
  • 39
    Thalheimer U, Bosch J, Burroughs AK. How to prevent varices from bleeding: shades of grey – the case for nonselective β-blockers. Gastroenterology 2007; 133: 202936.
  • 40
    García-Pagán JC, Bosch J. Endoscopic band ligation in the treatment of portal hypertension. Nat Clin Pract Gastroenterol Hepatol 2005; 2: 52635.
  • 41
    Garcia-Tsao G, Bosch J, Groszmann RJ. Portal hypertension and variceal bleeding – unresolved issues. Summary of an American Association for the study of liver diseases and European Association for the study of the liver single-topic conference. Hepatology 2008; 47: 176472.
  • 42
    Bañares R, Moitinho E, Matilla A, et al. Randomized comparison of long-term carvedilol and propranolol administration in the treatment of portal hypertension in cirrhosis. Hepatoloy 2002; 36: 136773.
  • 43
    Bosch J, Abraldes JG, Fernández M, García-Pagán JC. Hepatic endothelial dysfunction and abnormal angiogenesis: new targets in the treatment of portal hypertension. J Hepatol 2010; 53: 55867.