Thirteen RCTs (1,077 patients) were included.
Overall mortality.
There was no significant difference between octreotide compared with any alternative intervention, for octreotide compared with vasopressin or terlipressin, or for octreotide compared with placebo or no intervention. Octreotide versus any alternative therapy (11 RCTs, 948 patients).
The RR was 0.89 (95% CI: 0.69, 1.14). The homogeneity p was 0.3.
Octreotide versus vasopressin or terlipressin (4 RCTs, 236 patients).
The RR was 0.80 (95% CI: 0.54, 1.19). The homogeneity p was 0.88.
Octreotide versus placebo or no therapy after all received initial sclerotherapy or banding (4 RCTs, 424 patients).
The RR was 0.81 (95% CI: 0.48, 1.35). The homogeneity p was 0.6.
Octreotide versus sclerotherapy (2 RCTs, 248 patients).
The RR was 1.1 (95% CI: 0.73, 1.66). There was evidence of heterogeneity (p=0.02), hence specific statements could not be made.
Sustained control of bleeding.
Octreotide was associated with a significant reduction in the risks of re-bleeding compared with any alternative intervention, for octreotide compared with vasopressin or terlipressin, and for octreotide compared with placebo or no treatment. There was no significant difference between octreotide and sclerotherapy. There was no evidence of heterogeneity for any of these comparisons.
Octreotide versus any alternative therapy (13 RCTs, 1,077 patients).
The RR was 0.63 (95% CI: 0.51, 0.77). The homogeneity p was 0.2.
Octreotide versus vasopressin or terlipressin (5 RCTs, 279 patients).
The RR was 0.58 (95% CI: 0.42, 0.81). The homogeneity p was 0.97.
Octreotide versus placebo or no therapy after all received initial sclerotherapy or banding (5 RCTs, 510 patients).
The RR was 0.46 (95% CI: 0.32, 0.67). The homogeneity p was 0.4.
Octreotide versus sclerotherapy (2 RCTs, 248 patients).
The RR was 0.94 (95% CI: 0.55, 1.62). The homogeneity p was 0.6.
Any complications.
Any complications were more common among those receiving vasopressin or terlipressin than octreotide. There was no significant difference between octreotide compared with any alternate therapy, for octreotide compared with placebo or no therapy, or for octreotide compared with sclerotherapy. There was evidence of significant heterogeneity among studies for octreotide versus any alternative therapy and for octreotide versus sclerotherapy, hence specific conclusions about these comparisons could not be made. Octreotide versus any alternative therapy (11 RCTs, 948 patients).
The RR was 0.77 (95% CI: 0.6, 1.00). The homogeneity p was less than 0.001. Octreotide versus vasopressin or terlipressin (4 RCTs, 236 patients).
The RR was 0.52 (95% CI: 0.33, 0.82). The homogeneity p was 0.13.
Octreotide versus placebo or no therapy after all received initial sclerotherapy or banding (4 RCTs, 424 patients).
The RR was 1.06 (95% CI: 0.72, 1.55). The homogeneity p was 1.
Octreotide versus sclerotherapy (2 RCTs, 248 patients).
The RR was 0.91 (95% CI: 0.5, 1.65). The homogeneity p was less than 0.001.
Major complications.
Major complications were more common for vasopressin or terlipressin than with octreotide, but not for any of the other comparisons.
Octreotide versus vasopressin or terlipressin (4 RCTs, 236 patients).
The RR was 0.31 (95% CI: 0.11, 0.87). The homogeneity p was 0.6.
All significant results found when using fixed-effect models were also significant when using random-effects models.
Sensitivity analyses.
Restricting the analyses to studies with adequate investigator blinding and adequate randomisation resulted in a barely significant reduction in mortality for octreotide versus any alternative therapy (previously, mortality was not significantly different between the groups), but it did not change the magnitude or significance of other summary results.
Octreotide versus any alternative therapy.
The RR was 0.76 (95% CI: 0.58, 1.0). The homogeneity p was 0.86.
After excluding 2 studies with a low proportion of patients with severe liver disease (less than 40% with Child C disease), there was no change in the magnitude or significance of summary estimates for sustained control of bleeding, major complications, or total mortality for octreotide versus any alternate therapy or for any other comparison.
The authors reported the following limitations of the review: the variation in alternative therapies, both between and within the subgroups; there were relatively few trials of each alternative intervention; the potential for publication bias; an inability to confirm whether complications were reported on a per-patient basis; the variability in the patient populations; the variability in the quality of the included studies; and inconsistent use of a standard definition of re-bleeding.