SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Background : Controversy exists surrounding pharmacological therapy in acute variceal bleeding.

Aim : To determine the efficacy and safety of terlipressin.

Methods : Randomized trials were identified and duplicate, independent, review identified 20 randomized trials involving 1609 patients that compared terlipressin with placebo, balloon tamponade, endoscopic treatment, octreotide, somatostatin or vasopressin for treatment of acute oesophageal variceal haemorrhage.

Results : Meta-analysis showed that compared to placebo, terlipressin reduced mortality (relative risk 0.66, 95% CI 0.49–0.88), failure of haemostasis (relative risk 0.63, 95% CI 0.45–0.89) and the number of emergency procedures per patient required for uncontrolled bleeding or rebleeding (relative risk 0.72, 95% CI 0.55–0.93). When used as an adjuvant to endoscopic sclerotherapy, terlipressin reduced failure of haemostasis (relative risk 0.75, 95% CI 0.58–0.96), and had an effect on reducing mortality that approached statistical significance (relative risk 0.74, 95% CI 0.53–1.04). No significant difference was demonstrated between terlipressin and endoscopic sclerotherapy, balloon tamponade, somatostatin or vasopressin. Haemostasis was achieved more frequently with octreotide compared to terlipressin (relative risk 1.62, 95% CI 1.05–2.50), but this result was based on unblinded studies. Adverse events were similar between terlipressin and the other comparison groups except for vasopressin, which caused more withdrawals due to adverse events.

Conclusions : Terlipressin is a safe and effective treatment for acute oesophageal variceal bleeding, with or without adjuvant endoscopic sclerotherapy. Terlipressin appears to reduce mortality in acute oesophageal variceal bleeding compared to placebo, and is the only pharmacological agent shown to do so. Future studies will be required to detect potential mortality differences between terlipressin and other therapeutic approaches.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Acute oesophageal variceal bleeding is associated with a 20–50% in-hospital mortality1 and may be treated by pharmacological agents (somatostatin, octreotide, vasopressin and terlipressin), endoscopic treatment (sclerotherapy or band ligation), balloon tamponade, transjugular intrahepatic portosystemic stenting (TIPS) and surgery. Of these potential treatments, only pharmacological agents, which reduce portal pressure,2, 3 can be initiated early by nonspecialized personnel. Despite this attractive attribute, there is considerable controversy regarding their efficacy and wide variation in their use.

Terlipressin (triglycyl lysine vasopressin) is a synthetic analogue of vasopressin that can be given by intermittent intravenous injections rather than by continuous intravenous infusion (as is necessary for vasopressin). Although terlipressin is slowly cleaved in vivo to vasopressin, it is also believed to have an intrinsic vasoconstrictor effect of its own.1 Additionally, in contrast to vasopressin, terlipressin does not appear to have an effect on plasminogen activator, which is clearly undesirable in the setting of acute variceal haemorrhage.4 Indeed, terlipressin appears to have a very low incidence of severe adverse reactions.

The aim of this systematic review was to evaluate the efficacy and safety of terlipressin for treatment of acute oesophageal variceal bleeding compared to placebo, balloon tamponade, endoscopic treatments and other pharmacological agents.

Data identification

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Two reviewers (G.I. and J.D.) independently searched medical electronic databases (Medline, Embase, Cochrane Controlled Trials Register, Cochrane Hepato-Biliary Group Controlled Trials Register, Current Contents, Biosis) from 1966 to 2001. The bibliography of all identified studies was scanned and experts in the field and the manufacturers of terlipressin (Ferring Pharmaceuticals, Denmark) were contacted for additional published or unpublished studies. Trials published only as abstracts were identified mainly through the Cochrane databases and were included in an effort to reduce publication bias, provided the authors could be contacted for additional information.

Inclusion criteria

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Trials were included if they: (1) were randomized controlled trials, published in any language, either in full or in abstract; (2) included patients with suspected or documented acute oesophageal variceal bleeding; (3) compared terlipressin, any dose, route or duration to placebo, balloon tamponade, endoscopic treatment, octreotide, somatostatin or vasopressin; and (4) reported data on at least one of the outcome measures mentioned below.

Outcome measures

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

The following outcomes were abstracted from each study: (1) All cause mortality, in accordance with the Baveno consensus report5 in which any death up to 42 days after a variceal bleed was assumed to be caused by the bleed regardless of the mode of death. (2) Failure of haemostasis, defined in most studies as evidence of continued bleeding such as melena, haematemesis and bloody nasogastric aspirate, or unstable haemodynamic parameters and haemoglobin levels, or both. (3) Rebleeding evidenced by the same criteria used for failure of haemostasis. (4) Emergent procedures (balloon tamponade, sclerotherapy, surgery or TIPS) required for uncontrolled bleeding or rebleeding. Procedures done routinely as part of the protocol of a study were not included. (5) Blood transfused. (6) Length of hospitalization (days). (7) Adverse events that resulted in discontinuation of treatment or death.

Study selection, quality assessment and data collection

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Study selection, quality assessment and data collection were performed independently by two reviewers (G.I. and J.D.), one of whom was blinded with respect to the authors and the journal of publication. A copy of the data extraction form was sent to the Correspondence of each study, who was asked to correct erroneous data and provide missing information. We asked authors of abstracts to provide full information about methods and updated results. The Jadad scale was used to assess study quality, and was chosen a priori because of its ability to discriminate between ‘high quality’ (low bias) studies and ‘low quality’ (high bias) studies and because it has been validated in digestive disease studies.6–8 A brief description of the definitions and point score system follows:

  • 1
    Was the study described as randomized (this includes the use of words such as randomly, random and randomization)? One point.
  • 2
    Is the method of randomization appropriate? One point is added if the method to generate the sequence of randomization is described and it is appropriate (e.g. table of random numbers, computer generated). One point is deducted if the method to generate the sequence of randomization is described and it is inappropriate (e.g. date of birth). Zero points if the method of randomization is not described.
  • 3
    Was the study described as double-blind? One point.
  • 4
    Is the method used appropriate? One point is added if the method of masking is described and it is appropriate (e.g. identical placebo). One point is deducted if the method of masking is described and it is inappropriate(e.g. comparison of tablet vs. injection with no double dummy). Zero points if the method of masking is not described.
  • 5
    Is there a description of withdrawals and drop-outs? One point is given if the numbers and reasons for withdrawal in each group are stated. If there are no withdrawals, the report must say so. If there is no statement on withdrawals, this item is given no points.

Studies were rated ‘high quality’ if they scored three or more points.6–8

Data analysis

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Statistical analyses were performed using RevMan software, version 4.1. Intention-to-treat data were used for all included studies. The results are presented such that a relative risk less than one or a weighted mean difference less than zero favoured terlipressin. The Breslow-Day test was used to test for homogeneity under the null hypothesis that the relative risks were consistent across studies and homogeneity was rejected for P-values < 0.10.9 A random effects model was used for meta-analysis if there was evidence of significant heterogeneity.9 The Mantel–Haenszel method was used to calculate a combined relative risk for the dichotomous variables, whereas the inverse variance method was used to calculate the weighted mean difference for continuous variables.9

Study selection

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

A total of 346 articles were identified by computerized searches and an additional 32 by scanning their bibliographies, of which 162 were excluded because their content was not relevant to terlipressin. Of the remaining 216 articles, 185 were excluded for the following reasons: 105 were background/review articles, 57 were studies of treatments other that terlipressin and 23 were studies of the haemodynamic effects of terlipressin on portal and variceal pressure. From the remaining 31 studies, detailed information was abstracted regarding the eligibility of the study, its methodological quality and its results. Another 11 studies were excluded for the following reasons: no control group,10–13 nonrandomized,14–17 not acute variceal bleeding,18, 19 inadequate data.20 The remaining 20 studies (with a total of 1609 patients) were included in the meta-analysis.

Terlipressin vs. placebo

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Seven randomized, placebo-controlled trials21–27 with a total of 443 patients were identified (Table 1). Compared to placebo (Table 2), terlipressin led to significant reductions in mortality (relative risk of 0.66, 95% confidence interval (CI) 0.49–0.88), failure of initial haemostasis (relative risk of 0.66, 95% CI 0.53–0.82) and procedures performed to stop uncontrolled bleeding or rebleeding (relative risk of 0.72, 95% CI 0.55–0.93). Statistical heterogeneity existed only for the outcome of failure of initial haemostasis, and persisted even in an analysis limited to the five high quality studies.22–25, 27 Therefore, an analysis by the random effects model was performed, which yielded a combined relative risk of 0.63 (95% CI 0.45–0.89).

Table 1.  Characteristics of randomized trials included in the meta-analysis
Trial (year)Study quality characteristicsPatient and treatment characteristics
Jadad* score Double blindLosses to follow-up describedConcealment of allocationLength of follow-upNumber of patientsChild's C %Alcoholic cirrhosis % Terlipressin dose and durationNitroglycerine with terlipressinControl dose and durationInitial scleroth- erapy
  • Abstract; ND: not described; NA: not applicable.

  • * Jadad score:6, 7 Quality assessment scale, which awards one point for positive answers to each of the following: Study described as randomized? Adequate randomization used (e.g. table of random numbers, computer generated)? Study described as double-blind? Adequate blinding used (e.g. identical placebo)? Are the numbers and reasons for withdrawal in each group stated? Studies scoring 1–2/5 are termed low quality whereas studies scoring 3–5/5 are termed high quality.

  •  Sclerotherapy, which was performed routinely at the initial diagnostic endoscopy.

  • ‡ Concealment of allocation refers to methods employed in the allocation of participants to treatment or control arms, which ensure that the study coordinators do not know the next available allocation.47

Telripressin compared with placebo
Walker,  19863/5YesNoYesDischarge 5050%84%1–2 mg iv every  4 h for 36 hNoPlaceboNo
Freeman,  19893/5YesNoNoDischarge 3129%77%1–2 mg iv every  4 h for 48 hNoPlaceboNo
Soderlund,  19903/5YesNoNoDischarge 6033%82%2 mg iv every  4 h for 24–36 hNoPlaceboNo
Pauwels,  19942/5NoYesNo30 days 31NDND1–2 mg iv every  4–6 h for 24 hNoStandard  treatmentNo
Levacher,  19953/5YesNoYes42 days 8481%91%1–2 mg iv every  4 h for 8 hYesPlaceboYes
Brunati,  1996†NDNoNDNo5 days 5536%25%2 mg iv every  6 h for 48 hNoStandard  treatmentYes
Patch,  1999†3/5YesNDND42 days13362%ND1–2 mg iv every  4 h for 120 hNoPlaceboYes
Terlipressin compared with balloon tamponade
Colin,  19872/5NoYesNoDischarge 8122%ND1–2 mg iv every  6 h for 96 hNoBalloon  TamponadeNo
Fort,  19902/5NoNoNoDischarge 4755%68%1–2 mg iv every  6 h for 30 hYesBalloon  TamponadeNo
Garcia-C,  19972/5NoNoNo30 days 4040%88%1–2 mg iv every  4 h for 24 hNoBalloon  TamponadeNo
Terlipressin compared with sclerotherapy
Escorsell,  20003/5NoYesYes42 days21930%40%1–2 mg iv every  4 h for 168 hNoSlerotherapyNA
Terlipressin compared with octreotide
Silvain,  19932/5NoNoNo30 days 8747%91%1–2 mg iv every  4 h for 24 hYes25 µg/h for  12 h, 100 µg  at 12 h  and 18 hNo
Pedretti,  19942/5NoNoYes50 days 6012%33%2 mg iv every  4–6 h for 168 hNo100 µg once,  25 µg/h for  24 h, 100 µg  t.d.s. for 6dNo
Brunati,  1996†NDNoNDNo5 days 5636%25%2 mg iv every  6 h for 48 hNo100 µg every  8 h for 48 hYes
Terlipressin compared with somatostatin
Pauwels,  19942/5NoYesNo30 days 35NDND1–2 mg iv every  4–6 h for 24 hNo250 µg bolus, 250 µg/hNo
Walker,  19963/5YesNoYes30 days10612%74%1–2 mg iv every  4 h for 24 hNo250 µg bolus, 250 µg/hNo
Feu,  19965/5YesYesYes42 days16129%54%2 mg iv every  4 h for 48 hNo250 µg bolus, 250 µg/hNo
Terlipressin compared with vasopressin
Freeman,  19821/5NoNoNoDischarge 2115%48%1–2 mg iv every  6 for 24 hNo0.2–0.4 U/minNo
Desaint,  19871/5NoNoNoND 1643%ND1–2 mg iv every  6 for 24 hNo0.3 U/kg/hNo
Lee,  19881/5NoNoNo42 days 4527%ND1–2 mg iv every  6 for 24 hNo0.33–0.66 U/hNo
Chiu,  19902/5NoNoNoDischarge 5459%26%1–2 mg iv every  4 h for 32 hNo0.2–0.4 U/minNo
D'Amico,  19942/5NoNoYes30 days16519%19%2 mg iv every  6 h for 24 hNo0.4 U/minYes
Table 2.  Results of meta-analysis
 Number of high-quality studiesNumber of low-quality studiesNumber of patientsCombined relative risk (95% CI)Number needed to treat (95% CI)
  • * Results with a P-value < 0.05.

  • † Study quality was determined based on the Jadad scale.5, 6

  • ¶ 

    Values of relative risk less than 1 favour terlipressin.

  • ‡ Calculated for relative risk values with a confidence interval that excludes 1.

Telripressin compared with placebo
Mortality524430.66 (0.49–0.88)*8 (5–25)
Failure of haemostasis524430.66 (0.53–0.82)*6 (4–11)
Rebleeding311960.99 (0.60–1.61)NA
Procedures for uncontrolled bleeding513880.72 (0.55–0.93)*8 (5–33)
Terlipressin compared with balloon tamponade
Mortality031410.94 (0.50–1.77)NA
Failure of haemostasis031411.61 (0.69–3.74)NA
Rebleeding031410.84 (0.46–1.55)NA
Procedures for uncontrolled bleedingNANANANANA
Terlipressin compared with sclerotherapy
Mortality102191.49 (0.88–2.52)NA
Failure of haemostasis102191.09 (0.62–1.90)NA
Rebleeding102190.97 (0.62–1.54)NA
Procedures for uncontrolled bleeding102190.88 (0.49–1.57)NA
Terlipressin compared with octreotide
Mortality032031.20 (0.66–2.15)NA
Failure of haemostasis032031.62 (1.05–2.50)*7 (4–50) favours  octreotide
Rebleeding021470.52 (0.24–1.11)NA
Procedures for uncontrolled bleeding021471.39 (0.90–2.12)NA
Terlipressin compared with somatostatin
Mortality213020.95 (0.62–1.48)NA
Failure of haemostasis223021.05 (0.67–1.63)NA
Rebleeding213021.20 (0.80–1.81)NA
Procedures for uncontrolled bleeding213021.02 (0.63–1.64)NA
Terlipressin compared with vasopressin
Mortality053011.39 (0.96–2.00)NA
Failure of haemostasis053010.85 (0.65–1.13)NA
Rebleeding042561.44 (0.93–2.25)NA
Procedures for uncontrolled bleeding011650.58 (0.30–1.11)NA

In order to exclude the possibility that the observed beneficial results of terlipressin were driven by the low quality (high potential bias) studies, separate analyses were performed limited to the five22–25, 27 double-blinded studies that achieved a high quality score based on the Jadad scale. These analyses again demonstrated significant reductions in mortality (relative risk 0.61, 95% CI 0.45–0.84), failure of initial haemostasis (relative risk of 0.66, 95% CI 0.52–0.83) and procedures performed to stop uncontrolled bleeding or rebleeding (relative risk of 0.68, 95% CI 0.52–0.9).

Additionally, analyses were performed limited to the three studies23, 26, 27 in which patients were treated with endoscopic sclerotherapy at the initial diagnostic endoscopy, in order to investigate whether terlipressin improves the efficacy of sclerotherapy (Figures 1–3). The beneficial effect of terlipressin on mortality (Figure 1), failure of initial haemostasis (Figure 2) and procedures required for uncontrolled bleeding or rebleeding (Figure 3), was less among the three studies that employed initial sclerotherapy than among the four studies that did not. However, even among the three studies that employed initial sclerotherapy, addition of terlipressin was associated with beneficial effects that were either statistically significant (reduction in failure of initial haemostasis by a relative risk of 0.75, 95% CI 0.58–0.96), or approached statistical significance (reduction in mortality by a relative risk of 0.74, 95% CI 0.53–1.04, and reduction in procedures required to stop uncontrolled bleeding by a relative risk of 0.87, 95% CI 0.63–1.2).

image

Figure 1. Meta-analysis (fixed effects model) of randomized trials of terlipressin compared with placebo for acute variceal bleeding showing the effect of treatment on mortality. In studies that employed initial sclerotherapy, the test for overall effect was P=0.08, and the test for heterogeneity was P=0.76 (a P-value < 0.05 in the test of overall effect indicates a statistically significant effect and a P-value > 0.1 in the test of heterogeneity indicates lack of statistical heterogeneity). In studies not employing initial sclerotherapy, the test for overall effect was P=0.02, and the test for heterogeneity was P=0.27. In all studies, the test for overall effect was P=0.004, and the test for heterogeneity was P=0.53. *Relative risk less than one favours terlipressin; ¶n/N refers to the ratio of deaths to participants in each study or group of studies.

Download figure to PowerPoint

image

Figure 2. Meta-analysis (fixed effects model) of randomized trials of terlipressin compared with placebo for acute variceal bleeding showing the effect of treatment on failure of initial haemostasis. In studies that employed initial sclerotherapy, the test for overall effect was P=0.02, and the test for heterogeneity was P=0.14 (a P-value < 0.05 in the test of overall effect indicates a statistically significant effect and a P-value > 0.1 in the test of heterogeneity indicates lack of statistical heterogeneity). In studies not employing initial sclerotherapy, the test for overall effect was P=0.02, and the test for heterogeneity was P=0.27. In all studies, the test for overall effect was P=0.002, and the test for heterogeneity was P=0.093. *Relative risk less than one favours terlipressin; ¶n/N refers to the ratio of participants with uncontrolled bleeding to all participants in each study or group of studies.

Download figure to PowerPoint

image

Figure 3. Meta-analysis (fixed effects model) of randomized trials of terlipressin compared with placebo for acute variceal bleeding showing the effect of treatment on the number of emergency procedures (sclerotherapy, surgery, balloon tamponade or TIPS) required for uncontrolled bleeding or rebleeding. In studies that employed initial sclerotherapy, the test for overall effect was P=0.40, and the test for heterogeneity was P=0.67 (a P-value < 0.05 in the test of overall effect indicates a statistically significant effect and a P-value > 0.1 in the test of heterogeneity indicates lack of statistical heterogeneity). In studies not employing initial sclerotherapy, the test for overall effect was P=0.005, and the test for heterogeneity was P=0.069. In all studies, the test for overall effect was P=0.01, and the test for heterogeneity was P=0.098. *Relative risk less than one favours terlipressin; ¶n/N refers to the ratio of required emergency procedures to participants in each study or group of studies.

Download figure to PowerPoint

All seven trials of terlipressin vs. placebo demonstrated lower blood transfusion requirements in the terlipressin group compared to the placebo group (Table 3). However, because four of these studies22, 24, 26, 27 used a median and a range to summarize their results and one study23 reported mean units of blood transfused per patient per day, we could not combine the results in a meta-analysis. Only four studies reported rebleeding rates and found no statistically significant difference between terlipressin and placebo (relative risk 0.99, 95% CI 0.60–1.61). None of the studies reported data on length of hospitalization.

Table 3.  Comparison of blood transfusion requirements in the terlipressin and placebo groups
StudyBlood transfusions per patient in units of blood
TerlipressinPlacebo
Brunati, 1996Median=2Median=4
Freeman, 1989Median=3Median=4
Levacher, 1995Mean/day=0.79Mean/day=1.94
Patch, 1999Median=2.5Median=3
Pauwels, 1994Mean=4.9Mean=5.3
Soderlund, 1990Median=2Median=3
Walker, 1986Mean=5.4Mean=7.5

Terlipressin vs. balloon tamponade

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

The three randomized controlled trials28–30 comparing terlipressin to balloon tamponade (141 total patients) were graded ‘low quality’ (Table 1), in part due to the fact that none of these studies was double-blinded, a study design that is difficult, if not impossible, to achieve in a trial involving balloon tamponade. There was no statistically significant difference between terlipressin and balloon tamponade (Table 2) in any of the outcomes examined.

Terlipressin vs. endoscopic treatment

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

One high quality study involving 219 patients compared terlipressin to endoscopic sclerotherapy (Table 1).31 There were no statistically significant differences between the terlipressin and endoscopic sclerotherapy groups (Table 2) in any of the outcome measures examined.

Terlipressin vs. octreotide

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Three26, 32, 33 low quality, randomized controlled trials (Table 1), including 203 patients, compared terlipressin to octreotide. There was no statistically significant difference between terlipressin and octreotide (Table 2) in mortality, number of patients with rebleeding or procedures required for uncontrolled bleeding or rebleeding. There was a statistically significant difference for failure of initial haemostasis in favour of octreotide with a combined relative risk of 1.62 (95% CI 1.05–2.50). One study26 reported blood transfusions only as median and range: two units per patient were required in the terlipressin group (range 1–5) and three units in the octreotide group (range 1–8). The weighted mean difference for the other two studies favoured octreotide (0.68 units, 95% CI 0.09–1.28), however there was significant statistical heterogeneity (P=0.008). There was no statistically significant difference in a random effects model (0.91 units, 95% CI −0.75 to 2.58). No study provided data on length of hospital stay.

Adverse events

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

There was no statistically significant difference between terlipressin and any of the other groups (placebo, balloon tamponade, endoscopic treatment or other vasoactive agent) in the number of deaths that were thought to be secondary to adverse events. There were six deaths out of 798 treated patients in the terlipressin group: four were attributed to fibrinolysis,23 one to left ventricular failure33 and one was a ‘sudden death’ of unknown aetiology.28 There were five deaths out of 811 patients in the control groups: four were attributed to fibrinolysis, occurring in a placebo group23 and one to arm necrosis that led to shock and death, occurring in a vasopressin group.37

There was no statistically significant difference between terlipressin and any of the control groups, except vasopressin in adverse events requiring withdrawal of treatment. In the terlipressin group, five withdrawals of treatment out of 693 patients occurred: three were attributed to bradycardia,23, 32, 39 one to ventricular fibrillation34 and one to respiratory failure.28 In control groups, there were 11 withdrawals of treatment out of 697 patients: three were attributed to bleeding from compression ulcers due to balloon tamponade,28, 30 the other eight occurred in vasopressin groups and included myocardial ischaemia in one,38 severe headache in one,40 abdominal pain in three,40 bradycardia in one40 and hypertension in two.40 In the terlipressin vs. vasopressin analysis there were more withdrawals due to adverse events in the vasopressin group than in the terlipressin group (P < 0.05).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

Despite the publication of more than 50 randomized trials41 investigating pharmacological vasoactive agents for acute oesophageal variceal bleeding, significant controversy regarding their efficacy remains. Previous reviews have suggested that, among patients who were not routinely treated with endoscopic variceal sclerotherapy at presentation, terlipressin reduces mortality compared to placebo.42–44 Although no single study or meta-analysis has shown that somatostatin or octreotide reduce mortality in acute variceal bleeding,45, 46 a recent meta-analysis46 favoured the use of octreotide over vasopressin/terlipressin because its results suggested that octreotide improved control of bleeding compared to vasopressin/terlipressin. These controversies are reflected in the geographical variability in the use of vasoactive agents. Octreotide is the preferred agent in the United States (where terlipressin is not available), whereas terlipressin is commonly used in many European countries.

Part of the controversy regarding the use of vasoactive agents for variceal bleeding arises from the fact that many of the randomized trials were small, low quality, unblinded studies. This problem is compounded by the fact that some of the outcomes commonly used, such as control of bleeding and rebleeding, are relatively subjective and therefore may be subject to ascertainment bias, especially in nonblinded studies.7, 47 Additionally, despite the development of recent consensus statements5 most of the studies have used divergent definitions of control of bleeding and rebleeding, which makes it even harder to make direct comparisons. Finally, grouping all terlipressin/vasopressin studies46 or octreotide/somatostatin studies45 together in meta-analyses may be misleading because vasopressin and somatostatin have different properties from their synthetic analogues.1, 4

We have tried to address these potential problems by analysing separately high and low quality studies, by emphasizing mortality as an objective and clinically important outcome, and by analysing terlipressin as a specific group in this meta-analysis. Our results confirm the reduction in mortality associated with terlipressin treatment compared to placebo (relative risk of 0.66, 95% CI 0.49–0.88), in a larger number of studies and patients than previously reported.44 Assuming a mortality rate of 36% as was found in placebo control groups (80 out of 220 patients died), one would need to treat eight patients with terlipressin instead of placebo to prevent one death. The fact that more studies are now available allowed us to demonstrate that the beneficial effects of terlipressin on mortality, as well as failure of haemostasis and rebleeding, persisted in the sub-group of high quality, double-blinded studies, thus reducing the possibility that bias may be contributing to the observed effects.

Additionally, we were able to investigate whether terlipressin improves the efficacy of sclerotherapy, because three studies are now available in which patients who received sclerotherapy at their initial diagnostic endoscopy, were randomized to terlipressin or placebo. The data suggest that terlipressin, when added to sclerotherapy, leads to a statistically significant improvement in haemostasis (relative risk 0.75, 95% CI 0.58–0.96) as well as to reductions in mortality (relative risk 0.74, 95% CI 0.53–1.04) that approach statistical significance. Thus, terlipressin may be beneficial in the treatment of acute variceal bleeding both as a single agent and when added to endoscopic sclerotherapy. This finding is important because endoscopic treatment is now routinely performed for acute variceal bleeding in many centres.41

Whether any one specific vasoactive agent is clearly superior to other agents remains an open question. In this review we found no difference between terlipressin and somatostatin in any outcome, based on the results of three studies (two high quality and one low quality) with a total of 302 patients. Only three unblinded, low quality studies have compared terlipressin with octreotide; meta-analysis of these data suggested that octreotide may be superior to terlipressin (only) with respect to failure of initial haemostasis. Based on this finding, it has been suggested46 that octreotide should be favoured over terlipressin for the treatment of acute oesophageal variceal bleeding. However, this observation of a difference between terlipressin and octreotide with regard to haemostasis is based on small, unblinded, low quality studies, in which bias may arise, especially with respect to the ascertainment of a subjective outcome such as haemostasis. Additionally, compared to placebo, octreotide has not been shown to reduce mortality in single studies or in meta-analyses, despite the fact that over 500 patients have been studied.46 However, comparing the results of octreotide vs. placebo trials to the results of terlipressin vs. placebo trials may be difficult. First, in the octreotide trials, all patients received initial endoscopic therapy, which may make it harder to detect any additional benefit due to octreotide. In contrast, only three out of the seven terlipressin vs. placebo trials employed initial endoscopic sclerotherapy. It should be emphasized, however, that even in studies that employed initial sclerotherapy, terlipressin led to an improvement in failure of initial haemostasis (relative risk of 0.75, 95% CI 0.58–0.96) and a reduction in mortality that approached statistical significance (0.74, 95% CI 0.53–1.04). Second, in some of the trials involving octreotide, ligation – a more effective initial therapy than sclerotherapy – was used, making it even more difficult to demonstrate the efficacy of octreotide. Thus, future studies will be required to accurately define potential differences in efficacy between terlipressin and octreotide.

Whether terlipressin is as effective as endoscopic therapy is unknown. Only one randomized controlled study has compared terlipressin with endoscopic treatment,31 in this case sclerotherapy. Although this high quality, multicentre trial (and the largest single study in this review) did not demonstrate a statistically significant difference between terlipressin and sclerotherapy in any of the outcomes examined, a type II error can not be excluded – indeed, assuming a baseline mortality of 25% in the terlipressin group, this study was powered to detect an absolute difference in mortality between terlipressin and sclerotherapy of 14% (alpha=0.05, beta=0.80). Thus, larger studies will be required to evaluate whether a smaller, but significant, difference could exist between the two treatments. Currently, the optimal endoscopic approach is likely to be variceal band ligation48 and, unfortunately, data comparing terlipressin to band ligation are not available.

Terlipressin was initially developed as a synthetic analogue of vasopressin with the hope that it would be safer and easier to administer. We found no difference between terlipressin and any of the comparison groups in adverse events causing withdrawal or death except for the terlipressin vs. vasopressin group where more withdrawals occurred in the vasopressin group. Our results are in contrast to those of a recent meta-analysis46 which found that octreotide was associated with fewer adverse reactions than terlipressin or vasopressin grouped together. Unfortunately, this latter conclusion is misleading because the difference is entirely driven by the high adverse reaction rate of vasopressin, and no difference exists in a direct comparison between terlipressin and octreotide. The lack of an association between terlipressin and adverse reactions also suggests that routine co-administration of nitrates with terlipressin is probably unwarranted.

In conclusion, terlipressin is as safe as other commonly used treatments for acute oesophageal variceal bleeding. It significantly reduces the mortality of variceal bleeding compared to placebo, and this beneficial effect persists even when the analysis is limited to high quality studies. When used as an adjuvant to endoscopic sclerotherapy, terlipressin improves haemostasis, and has an effect on reducing mortality that approaches statistical significance. Therefore, these data support the use of terlipressin as initial treatment of acute variceal bleeding, with or without adjuvant endoscopic treatment. In contrast, octreotide does not appear to reduce mortality compared to placebo, although it may reduce the risk of failure of haemostasis. Future studies with enough power to detect potential mortality differences between terlipressin and octreotide/somatostatin in the setting of adjuvant endoscopic therapy are required.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES

We would like to thank all the authors of the primary studies who kindly provided additional and confirmatory data. This work was performed in collaboration with the Cochrane Hepatobiliary Group, and portions of it have been published in the 2001 edition of the Cochrane Library of systematic reviews. This work was supported by the Burroughs Welcome Fund (DCR is a recipient of the BWF Clinical Scientist Award). G.N.I. is a fellow at the Health Services Research and Development group of the Seattle Veterans Affairs Medical Center.

REFERENCES

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Data identification
  6. Inclusion criteria
  7. Outcome measures
  8. Study selection, quality assessment and data collection
  9. Data analysis
  10. Results
  11. Study selection
  12. Terlipressin vs. placebo
  13. Terlipressin vs. balloon tamponade
  14. Terlipressin vs. endoscopic treatment
  15. Terlipressin vs. octreotide
  16. Terlipressin vs. somatostatin
  17. Terlipressin vs. vasopressin
  18. Adverse events
  19. Discussion
  20. Acknowledgements
  21. REFERENCES
  • 1
    Burroughs AK. Pharmacological treatment of acute variceal bleeding. Digestion 1998; 59(Suppl. 2): 2836.
  • 2
    Lebrec D. Hemodynamic effects of terlipressin (Glypressin) in patients with cirrhosis. In: Lebrec, D, Blei, AT, eds. Vasopressin Analogs and Portal Hypertension. Paris: John Libbey Eurotext, 1987: 10512.
  • 3
    Escorsell A, Bandi JC, Moitinho E, et al. Time profile of haemodynamic effects of terlipressin in portal hypertension. J Hepatol 1997; 26: 6217.
  • 4
    Douglas JG, Forrest JA, Prowse CV, Cash JD, Finlayson ND. Effects of lysine vasopressin and glypressin on the fibrinolytic system in cirrhosis. Gut 1979; 20: 5657.
  • 5
    De Franchis R. Updating consensus in portal hypertension. Report of the Baveno III consensus workshop on definitions, methodology and therapeutic strategies in portal hypertension. J Hepatol 2000; 33: 86452.
  • 6
    Jadad AR. Meta-analysis of Randomized Clinical Trials in Pain Relief. PhD Thesis, University of Oxford, 1994.
  • 7
    Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: 112.
  • 8
    Moher D, Pham B, Jones A, et al. Does quality of reports of randomized trials affect estimates of intervention efficacy reported in meta-analysis? Lancet 1998; 352: 60913.
  • 9
    Petitti DB. Meta-Analysis, Decision Analysis, and Cost-Effectiveness Analysis. New York: Oxford University Press, 1994.
  • 10
    Aracidiacono R, Biraghi M, Bonomo GM, Fiaccadori F. Randomized controlled trial with terlipressin in cirrhotic patients with rebleeding esophageal varices: effects on precocious rebleeding and mortality rates. Curr Ther Res 1992; 52: 18694.
  • 11
    Fiaccadori F, Pedretti G, Biraghi M, Arcidiacono R. Terlypressin and endoscopic sclerotherapy control variceal bleeding and prevent early rebleeding in cirrhotic patients. Curr Ther Res Clin Exp 1993; 54: 51928.
  • 12
    Chang TT, Lee FY, Tsai YT, et al. A randomized controlled study of low-dose and high-dose terlipressin in the control of acute esophageal variceal hemorrhage. J Gastroenterol Hepatol 1991; 6: 4814.
  • 13
    De Franchis R, Arcidiacono PG, Carpinelli L, Andreoni B, Cestari L, Brunati S. Randomized controlled trial of desmopressin plus terlipressin vs. terlipressin alone for the treatment of acute variceal hemorrhage in cirrhotic patients: a multicenter, double-blind study. Hepatology 1993; 18: 11027.
  • 14
    Chen SC, Lin ZY, Lu SN, Chuang WL, Hsieh MY, Wang LY. The effect of pitressin and glypressin in variceal bleeding − a preliminary clinical trial. Kao-Hsiung i Hsueh Ko Hsueh Tsa Chih[Kaohsiung J Med Sci]1990; 6: 5515.
  • 15
    Heygere FD, Burroughs AK, Sherlock S, McIntyre N. Prospective evaluation of glypressin for variceal bleeding in cirrhotics. Gut 1986; 27: A1247(Abstract).
  • 16
    Olivero S, Mao P, Rozzio G, Foco A, Enrichens F, Bouzari H. Clinical experiment with glypressin in hemorrhage due to esophageal varices[Italian]. Minerva Dietologica e Gastroenterologica 1986; 32: 314.
  • 17
    Vosmik J, Jedlicka K, Mulder JL, Cort JH. Action of the triglycyl hormonogen of vasopressin (glypressin) in patients with liver cirrhosis and bleeding esophageal varices. Gastroenterology 1977; 72(4 Part 1): 6059.
  • 18
    Palazzi A, Laurenzi F, Tiburzi C, Ottaviani O. Usefulness of supplemental treatment with glypressin during the eradication phase of esophageal varices using endoscopic sclerotherapy[Italian]. Minerva Dietologica e Gastroenterologica 1989; 35: 25760.
  • 19
    Campisi C, Padula P, Peressini A, Boccardo F, Biraghi M, Casaccia M. Upper digestive hemorrhage. Comparison of terlipressin and octreotide[Italian]. Minerva Chirurgica 1993; 48: 10916.
  • 20
    Silk DBA, Mitchell KJ, MacDougal BDR, Williams R. Medical management of patients with portal hypertension bleeding from esophageal varices. In: Williams, R, Cantani, L, eds. Recenti Progressi Epatologi. Milan: Casa Editrice, 1979: 514.
  • 21
    Pauwels A, Florent C, Desaint B, Van Guivarch PH, Levy VG. Terlipressin and somatostatin in the treatment of hemorrhages from rupture of esophageal varices (letter)[French]. Gastroenterologie Clinique Biologique 1994; 18: 3889.
  • 22
    Freeman JG, Cobden I, Record CO. Placebo-controlled trial of terlipressin (glypressin) in the management of acute variceal bleeding. J Clin Gastroenterol 1989; 11: 5860.
  • 23
    Levacher S, Letoumelin P, Pateron D, Blaise M, Lapandry C, Pourriat JL. Early administration of terlipressin plus glyceryl trinitrate to control active upper gastrointestinal bleeding in cirrhotic patients. Lancet 1995; 346: 8658.
  • 24
    Soderlund C, Magnusson I, Torngren S, Lundell L. Terlipressin (triglycyl-lysine vasopressin) controls acute bleeding esophageal varices. A double-blind, randomized, placebo-controlled trial. Scand J Gastroenterol 1990; 25: 62230.
  • 25
    Walker S, Stiehl A, Raedsch R, Kommerell B. Terlipressin in bleeding esophageal varices: a placebo-controlled, double-blind study. Hepatology 1986; 6: 1125.
  • 26
    Brunati X, Ceriani R, Curioni R, Brunelli L, Repaci G, Morini L. Sclerotherapy alone vs scherotherapy plus terlipressin vs sclerotherapy plus octreotide in the treatment of acute variceal hemorrhage. Hepatology 1996; 24(4 part 2): 207A.
  • 27
    Patch D, Caplin M, Greenslade L, Burroughs A. Randomized double blind controlled trial of 5 day terlipressin vs placebo in acute variceal hemorrhage. J Hepatol 1999; 30(Suppl. 1): 55.
  • 28
    Colin R, Giuli N, Czernichow P, et al. Prospective comparison of glypressin, tamponade and their association in the treatment of bleeding esophageal varices. In: Lebrec, D, Blei, AT, eds. Vasopressin Analogs and Portal Hypertension. Paris: John Libbey Eurotext, 1987: 14953.
  • 29
    Fort E, Sautereau D, Silvain C, Ingrand P, Pillegand B, Beauchant M. A randomized trial of terlipressin plus nitroglycerin vs. balloon tamponade in the control of acute variceal hemorrhage. Hepatology 1990; 11: 67881.
  • 30
    Garcia-Compean D, Blanc P, Bories JM, Michel J, Desprez D, Pageaux GP. Treatment of active gastroesophageal variceal bleeding with terlipressin or hemostatic balloon in patients with cirrhosis: a randomized controlled trial. Arch Med Res 1997; 28: 2415.
  • 31
    Escorsell A, Ruiz del Arbol L, Planas R, Albillos A, Banares R, Bosch J. Multicenter randomized controlled trial of terlipressin versus sclerotherapy in the treatment of acute variceal bleeding: the TEST study. Hepatology 2000; 32: 4716.
  • 32
    Pedretti G, Elia G, Calzetti C, Magnani G, Fiaccadori F. Octreotide versus terlypressin in acute variceal hemorrhage in liver cirrhosis. Emergency control and prevention of early rebleeding. Clin Invest 1994; 72: 6539.
  • 33
    Silvain C, Carpentier S, Sautereau D, et al. Terlipressin plus transdermal nitroglycerin vs. octreotide in the control of acute bleeding from esophageal varices: a multicenter randomized trial. Hepatology 1993; 18: 615.
  • 34
    Feu F, Ruiz del Arbol L, Banares R, Planas R, Bosch J. Double-blind randomized controlled trial comparing terlipressin and somatostatin for acute variceal hemorrhage. Gastroenterology 1996; 111: 12919.
  • 35
    Walker S, Kreichgauer HP, Bode JC. Terlipressin (glypressin) versus somatostatin in the treatment of bleeding esophageal varices – final report of a placebo-controlled, double-blind study. Z Fur Gastroenterologie 1996; 34: 6928.
  • 36
    Freeman JG, Cobden I, Lishman AH, Record CO. Controlled trial of terlipressin (‘Glypressin’) versus vasopressin in the early treatment of esophageal varices. Lancet 1982; 2: 668.
  • 37
    Desaint B, Florent C, Levy VG. A randomized trial of triglycyl-lysine vasopressin versus lysine vasopressin in active cirrhotic variceal hemorrhage. In: Lebrec, D, Blei, AT, eds. Vasopressin Analogs and Portal Hypertension. Paris: John Libbey Eurotext, 1987: 1557.
  • 38
    Lee YF, Tsay YT, Lai KH, Lay CS, Chang TT, Malik U. A randomized controlled study of triglycyl-vasopressin and vasopressin plus nitrglycerin in the control of acute esophageal variceal hemorrhage. Chinese J Gastroenterol 1988; 5: 1318.
  • 39
    Chiu KW, Sheen IS, Liaw YF. A controlled study of glypressin versus vasopressin in the control of bleeding from esophageal varices. J Gastroenterol Hepatol 1990; 5: 54953.
  • 40
    D'Amico G, Traina M, Vizzini G, Tine F, Politi F, Montalbano L. Terlipressin or vasopressin plus transdermal nitroglycerin in a treatment strategy for digestive bleeding in cirrhosis. A randomized clinical trial. J Hepatol 1994; 20: 20612.
  • 41
    Sharara AI, Rockey DC. Gastroesophageal variceal hemorrhage. N Eng J Med 2001; 345(9): 66981.
  • 42
    D'Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995; 22(1): 33254.
  • 43
    Fano M, Biraghi M. Terlipressina (Glipressina) nel sanguinamento da varici esofagee: studio di metanalisi. Giorn Ital End Dig 1994; 17: 13541.
  • 44
    D'Amico G, Pagliaro L, Bosch J. Pharmacological treatment of portal hypertension: an evidence-based approach. Sem Liver Dis 1999; 19(4): 475505.
  • 45
    Gotzsche PC. Somatostatin or octreotide for acute bleeding esophageal varices. The Cochrane Database of Systematic Reviews 2000; Issue 2.
  • 46
    Corley DA, Cello JP, Adkisson W, Ko WF, Kerlikowske K. Octreotide for acute esophageal variceal bleeding: a meta-analysis. Gastroenterology 2001; 120(4): 94654.
  • 47
    Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. J Am Med Assoc 1995; 273(5): 40812.
  • 48
    Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding. Ann Intern Med 1995; 123: 2807.