Background/Aims: Upper gastrointestinal hemorrhage (UGH) is a serious medical condition which affects a large number of individuals. Endoscopic therapy accompanied by medication is a standard approach that is used to improve the prognosis of UGH patients and a few medications have been developed including proton pump inhibitors (PPIs), histamine H2 receptor antagonist (H2RA), somatostatin analogues and tranexamic acid. This study is set to compare the efficacy and safety of various medical interventions that are used to manage upper gastrointestinal bleeding. Methods: We searched PubMed, Cochrane Library, and Embase for relevant articles. Eligible studies were determined by using both the inclusion and exclusion criteria. Both traditional pair-wise meta-analysis and net-work analysis were carried out to evaluate the corresponding interventions. Results and Conclusion: PPI is an effective medication for UGH patients and intravenous PPI exhibits equivalent effectiveness and safety in comparison to oral PPI. H2RA is not recommended for UGH patients as patients treated with H2RA are associated with an increased risk of adverse events including rebleeding, need for surgery and all-cause mortality. Moreover, patients treated with H2RA exhibit an increased length of average hospital stay and blood transfusion amount compared to those treated with PPI. Tranexamic acid is also considered as another promising medication for UGH.

Upper gastrointestinal hemorrhage (UGH) is a serious medical condition with an annual incidence of 89.8 per 100,000 populations. Males between 61 and 69 are more likely to suffer from this disease compared with other individuals and the overall mortality of UGH is about 8.2 cases per 100,000 populations per year [1]. For non-variceal UGH, the use of non-steroidal anti-inflammatory drugs (NSAIDs) and H. pylori infections are two common risk factors for upper gastrointestinal bleeding [2]. The development of endoscopic therapy reduces the incidence of rebleeding, need for surgery and the modality of UGH significantly. Endoscopic therapy is usually carried out in conjunction with proton pump inhibitors (PPIs), histamine H2 receptor antagonist (H2RA), somatostatin analogues and tranexamic acid. It is challenging for clinicians to determine the appropriate medication that should be accompanied with endoscopic therapy and thereby we designed this study to address this issue.

PPIs such as omeprazole, pantoprazole, tenatoprazole and lansoprazole have been widely used to treat peptic ulcer diseases and they are effective in managing bleeding resulted from endoscopic therapy. PPIs were developed in the 1980s [3] and they irreversibly block the H+/K+ ATPase which is known as the gastric proton pump of the gastric parietal cells [4]. PPIs are able to impede the terminal stage of gastric acid secretion and thereby healing duodenal ulcers and alleviating pain due to indigestion or heartburn. H2RAs including cimetidine, ranitidine, famotidine and nizatidine are another type of antagonists which target the H2 receptor of parietal cells. Histamine excreted by ECL cells contained in the stomach can bind to H2 receptors on parietal cells and thereby stimulating acid secretion [5]. As a result, the corresponding acid secretion is inhibited once H2RA is administered to UGH patients.

Somatostatin, also known as the growth hormone-inhibiting hormone, is an inhibitory hormone. The secretion of somatostatin in stomach is induced by low pH conditions. However, the effectiveness of somatostatin analogues is constrained under high pH conditions [6]. Apart from that, somatostatin is able to reduce the portal flow and hence it is usually used in emergency for managing bleeding oesophageal varices in patients with liver cirrhosis [7]. Besides that, somatostatin may suppress pepsin secretion and hence alleviate upper gastrointestinal bleeding. On the other hand, tranexamic acid is a medication which is specifically designed for preventing excessive blood loss due to trauma or surgery [8]. The corresponding mechanism of tranexamic acid is analogous to that of the amino acid lysine. Tranexamic acid is able to bind with receptors on both plasminogen and plasmin and it can act as an antifibrinolytic factor. As suggested by other studies, tranexamic acid is potentially effective for managing gastrointestinal hemorrhage [9].

Although studies have been conducted to evaluate different types of medications that are used accompanied with endoscopic therapy, comparing several medications simultaneously is rarely found in the current literature. Therefore, our study is set to compare the efficacy and safety of these medications by using the approach of net-work meta-analysis.

Search strategy

We searched PubMed, Cochrane Library, and Embase for relevant articles. The following terms were used for searching: ("Gastrointestinal Hemorrhage"[mh] OR upper digestive tract hemorrhage[tiab] OR upper gastrointestinal hemorrhage[tiab] OR UGIH[tiab] OR digestive bleeding[tiab]) AND ("randomized controlled trial"[pt] OR "controlled clinical trial"[pt] OR "placebos"[mh] OR "random allocation"[mh] OR "double-blind method"[mh] OR ((double[tw] OR treble[tw] OR triple[tw]) AND (mask* [tw] OR blind* [tw]))) AND ("proton pump inhibitors"[mh] OR "histamine H2 antagonists"[mh] OR "somatostatin analogues"[mh] OR "tranexamic acid"[mh]). We also reviewed the corresponding reference list of each retrieved articles in order to identify any relevant studies that may be neglected. Retrieved articles were qualified and screened by two independent reviewers (M. Jiang and P. Chen). The corresponding procedures with respect to systematic review are complied with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guideline [10]. Randomized controlled trials (RCTs) were enrolled for further selection if they assessed the performance of PPI, H2RA, somatostatin analogues or tranexamic acid with respect to upper digestive tract hemorrhage.

Selection criteria

In the current meta-analysis, studies were included if they were in accordance with the following criteria: 1) trials with randomized control trials; 2) all subjects were above the age of 18; 3) the diagnosis of upper gastrointestinal bleeding was confirmed by endoscope; 4) treatments were clearly defined including the type of medication, route of administration and doses; 5) studies included at least one of the following endpoints: rebleeding, all-cause mortality, need for surgery, hospital stay, blood transfusion amounts.

Data extraction

Relevant data from enrolled articles were extracted by two researchers (Mengyuan Jiang and Qiang Gao) independently. Name of the first author, year of publication, study design, sample size, duration of the follow-up period, mean age of enrolled subjects, gender of subjects, the corresponding percentage of duodenal ulcer and gastric ulcer in the sample, NSAID status, H. pylori infection status, medication details and clinical outcomes were documented for data analysis. Various clinical outcomes including rebreeding, all-cause mortality, need for surgery, hospital stay and blood transfusion amount were compared among different medications by using the net-work meta-analysis. Disagreement with respect to data extraction was resolved by a third reviewer (Qiang Gao).

Statistical analysis

In the present study, direct evidence from conventional meta-analysis was reinforced by net-work meta-analysis which synthesized indirect evidence within a net-work enclosed by various medications. The odds ratio (OR) and their corresponding 95%CI were used to compare different medications with respect to various clinical outcomes. Conventional meta-analysis compared one medication to the placebo and this process was iterated for each medication whereas net-work meta-analysis compared several medications simultaneously within the net-work. Heterogeneity among studies was assessed by using the Cochran's Q-statistic and I2 test. Significant heterogeneity was concluded if the corresponding P-value of the Cochran's Q test is less than 0.01 or I2 > 50%.

The fixed-effect model (Mantel-Haenszel method) was applied within the analysis if no significant heterogeneity was concluded whereas the random-effects model (DerSimonian-Laird method) was implemented in the case of significant heterogeneity. All statistical procedures were carried out by using STATA version 12.0 (StataCorp, College Station, TX, USA) software.

In contrary to the conventional meta-analysis, a random-effects model within a Bayesian framework using Markov chain Monte Carlo methods was implemented for the net-work meta-analysis and the corresponding procedures were carried out by using WinBUGS (MRC Bio-statistics Unit, Cambridge, UK). Medications were compared using the statistics of ORs and their 95%CrIs. The ranking probabilities with respect to each clinical outcome were also obtained by using the surface under the cumulative ranking curve (SUCRA) [11]. Besides that, the overall ORs and their corresponding interval were compared between the conventional meta-analysis and net-work meta-analysis in order to assess consistency between direct and indirect evidence.

Study characteristics

A total of 47 articles were enrolled in this meta-analysis [7,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57]. We subsequently divided the interventions into the following nine groups on the basis of medication type, administration route and dose: placebo/no treatment, H2RA, high-dose i.v. PPI, i.v. PPI, i.v. PPI+H2RA, p.o. PPI, p.o. PPI + tranexamic acid, tranexamic acid and somatostatin analogues. The following clinical outcomes we studied in this analysis: rebleeding, all-cause mortality, need for surgery, hospital stay and blood transfusion amount. In all, 9,528 subjects with UGH were enrolled in our analysis. Main characteristics of selected studies were summarized in Table 1. Besides, a net-work plot of relevant studies was illustrated in Fig. 1 which depicts the corresponding comparisons within the net-work.

Table 1

Main characteristics of included studies. d: day; w: week; m: month; DU: duodenal ulcer; GU: gastric ulcer; NSAID: non-steroidal anti-inflammatory drug; H. pylori: Helicobacter pylori; i.v.: intravenous; p.o.: oral; H2RA: H2 receptor antagonists; PPI: proton pump inhibitor

Main characteristics of included studies. d: day; w: week; m: month; DU: duodenal ulcer; GU: gastric ulcer; NSAID: non-steroidal anti-inflammatory drug; H. pylori: Helicobacter pylori; i.v.: intravenous; p.o.: oral; H2RA: H2 receptor antagonists; PPI: proton pump inhibitor
Main characteristics of included studies. d: day; w: week; m: month; DU: duodenal ulcer; GU: gastric ulcer; NSAID: non-steroidal anti-inflammatory drug; H. pylori: Helicobacter pylori; i.v.: intravenous; p.o.: oral; H2RA: H2 receptor antagonists; PPI: proton pump inhibitor
Fig. 1

The net-work plot of treatments under five clinical outcomes.

Fig. 1

The net-work plot of treatments under five clinical outcomes.

Close modal

Rebleeding

Comparisons among medications with respect to the event of rebleeding are illustrated in Table 2 and Fig. 2. As suggested by the pair-wise meta-analysis, the majority of medications with the exception of H2RA and somatostatin analogues can significantly reduce the risk of rebleeding (H2RA: OR = 1.21, 95%CI: 0.91 - 1.63; High-dose PPI (i.v.): OR = 0.70, 95%CI: 0.57 - 0.86, PPI (p.o.): OR = 0.37, 95%CI: 0.24 - 0.55; somatostatin analogues: OR = 1.22, 95%CI: 0.79 - 1.87; tranexamic acid: OR = 0.29, 95%CI: 0.18 - 0.46). However, the net-work meta-analysis suggests that only High-dose PPI (i.v.) and PPI (p.o.) are effective in reducing the risk of rebleeding (High-dose PPI (i.v.): OR = 2.24, 95%CrI: 1.02 - 4.90, PPI(p.o.): OR = 2.52, 95%CrI: 1.03 - 6.24). Moreover, pair-wise meta-analysis indicates that patients treated with high-dose PPI (i.v.), PPI (p.o.) and somatostatin analogues are associated with a lower risk of rebleeding compared to those treated by H2RA (High-dose PPI (i.v.): OR = 0.35, 95%CI: 0.23 - 0.53; PPI (p.o.): OR = 0.39, 95%CI: 0.21- 0.73; somatostatin analogues: OR = 0.18, 95%CI: 0.09 - 0.36). Similar results are also observed in the corresponding net-work meta-analysis (High-dose PPI (i.v.): OR = 3.67, 95%CrI: 1.68 - 8.24, PPI (p.o.): OR = 4.15, 95%CrI: 1.49 -11.58; PPI (i.v.): OR = 3.61, 95%CrI: 1.42 - 9.58).

Table 2

Comparison of different treatment in rebleeding between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment

Comparison of different treatment in rebleeding between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Comparison of different treatment in rebleeding between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Fig. 2

The forest plot of indirect comparisons of seven treatments under the outcome of rebleeding.

Fig. 2

The forest plot of indirect comparisons of seven treatments under the outcome of rebleeding.

Close modal

All-cause mortality

Comparisons among medications with respect to all-cause mortality are presented in Table 3 and Fig. 3. No significant difference in the risk of all-cause mortality was identified by the pair-wise meta-analysis. However, the net-work meta-analysis reveals that PPI (p.o.) is more effective than H2RA with respect to all-cause mortality (OR = 3.27, 95%CrI: 1.09 -11.12).

Table 3

Comparison of different treatment in all-cause mortality between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment

Comparison of different treatment in all-cause mortality between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Comparison of different treatment in all-cause mortality between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Fig. 3

The forest plot of indirect comparisons of seven treatments under the outcome of all-cause mortality.

Fig. 3

The forest plot of indirect comparisons of seven treatments under the outcome of all-cause mortality.

Close modal

Need for surgery

Results concerning the adverse event of need for surgery are illustrated in Table 4 and Fig. 4. In traditional pair-wise meta-analysis, we only observed that patients under the treatment of PPI (p.o.) and tranexamic acid had a lower risk of surgery compared to those with placebo (PPI (p.o.): OR = 0.32, 95%CI: 0.17 - 0.59; tranexamic acid: OR = 0.37, 95%CI: 0.20 - 0.67). In contrary, the net-work meta-analysis indicates that most medications except H2RA and somatostatin analogues are more effective than the placebo with respect to the need for surgery (H2RA: OR = 0.80, 95%CrI: 0.39 - 1.49; High-dose PPI (i.v.): OR = 1.62, 95%CrI: 1.07 - 2.97; PPI (i.v.): OR = 3.72, 95%CrI: 1.25 - 10.45; PPI (p.o.): OR = 2.83, 95%CrI: 1.38 - 5.70; somatostatin analogues: OR = 1.01, 95%CrI: 0.40 - 2.80; tranexamic acid: OR = 2.72, 95%CI: 1.27 - 6.05). In addition, high-dose PPI (i.v.), PPI (i.v.), PPI (p.o.) and tranexamic acid are more effective than H2RA with respect to this outcome (High-dose PPI (i.v.): OR = 2.03, 95%CrI: 1.15 - 4.51; PPI (i.v.): OR = 4.67, 95%CrI: 1.56 - 14.16; PPI (p.o.): OR = 3.60, 95%CrI: 1.47 - 8.87; somatostatin analogues: OR = 1.26, 95%CrI: 0.42 - 4.71; tranexamic acid: OR = 3.40, 95%CrI: 1.29 - 10.05).

Table 4

Comparison of different treatment in need for surgery between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment

Comparison of different treatment in need for surgery between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Comparison of different treatment in need for surgery between pair-wise meta-analysis and net-work meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with odds ratio and 95% credential interval. The column treatment is compared with the row treatment
Fig. 4

The forest plot of indirect comparisons of seven treatments under the outcome of need for surgery.

Fig. 4

The forest plot of indirect comparisons of seven treatments under the outcome of need for surgery.

Close modal

Hospital stay

Table 5 and Fig. 5 illustrate the corresponding comparisons of hospital stay duration for UGH patients managed by different medications. Patients treated with H2RA or PPI (p.o.) exhibit a decrease in the average hospital stay duration compared to those treated by placebo (H2RA: mean difference = -0.67, 95%CI: -0.80 to -0.53; PPI (p.o.): mean difference = -0.35, 95%CI: -0.67 to -0.03). Apart from that, both high-dose PPI (i.v.) and PPI (i.v.) are able to reduce the average length of hospital stay in comparison to those treated by H2RA (high-dose PPI (i.v.): mean difference = -0.63, 95%CI: -0.81 to -0.44; PPI (i.v.): mean difference = -0.32, 95%CI: -0.51 to -0.12). Besides, patients treated with PPI (i.v.) or somatostatin analogues are associated with an increase in the average hospital stay duration compared to those treated with high-dose PPI(i.v.) (PPI (i.v.): mean difference = 0.14, 95%CI: 0.01 - 0.26; somatostatin analogue: mean difference = 1.01, 95%CI: 0.68 - 1.33). Also, patients under the treatment of PPI (p.o.) exhibit longer average hospital stay period than those managed by PPI (i.v.) (mean difference = 0.66, 95%CI: 0.34 - 0.99).

Table 5

Comparison of different treatment in hospital stay between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment.

Comparison of different treatment in hospital stay between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment.
Comparison of different treatment in hospital stay between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment.
Fig. 5

The forest plot of indirect comparisons of six treatments under the outcome of average hospital stay.

Fig. 5

The forest plot of indirect comparisons of six treatments under the outcome of average hospital stay.

Close modal

Blood transfusion amount

Likewise, comparisons among medications with respect to the blood transfusion amount are displayed in Table 6 and Fig. 6. Patients treated with H2RA, PPI (p.o.) or somatostatin analogues exhibit smaller blood transfusion amount on average whereas patients treated with high-dose PPI(i.v.) appear to have a larger average larger blood transfusion amount in comparison to those treated by the placebo (H2RA: mean difference = -1.11, 95%CI:-1.25 to -0.97; High-dose PPI (i.v.): mean difference = 0.22, 95%CI: 0.10 - 0.33; PPI (i.v.): mean difference = -0.36, 95%CI:-0.68 to -0.03; PPI (p.o.): mean difference = -0.75, 95%CI:-1.20 to -0.30).

Table 6

Comparison of different treatment in blood transfusion amount between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment

Comparison of different treatment in blood transfusion amount between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment
Comparison of different treatment in blood transfusion amount between pair-wise meta-analysis and network meta-analysis. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral; The cells in blue are results of pair-wise meta-analysis in the first line, with assessment of heterogeneity as I2(%) and P value in the second line. On the other side, results of network meta-analysis are in red cells with standard mean difference and 95% credential interval. The column treatment is compared with the row treatment
Fig. 6

The forest plot of indirect comparisons of six treatments under the outcome of average blood transfusion amount.

Fig. 6

The forest plot of indirect comparisons of six treatments under the outcome of average blood transfusion amount.

Close modal

Ranking of seven medications

Net-work meta-analysis also provides a comprehensive ranking by using the corresponding SUCRA values. A larger SUCRA value indicates that one medication is superior to others and hence it will have a relatively high ranking (Table 7). As suggested by the SUCRA values, PPI (i.v.) and PPI (p.o.) appear to have strength in reducing the risk of all-cause mortality, average hospital stay and average blood transfusion amount. On the other hand, tranexamic acid is particularly effective for reducing the risk of rebleeding and it exhibit reliable performance with respect to all-cause mortality and need for surgery. By contrast, H2RA appears to have a relatively low ranking with respect to all of the five clinical outcomes.

Table 7

Relative ranking of six drugs assessed by SUCRA values. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral

Relative ranking of six drugs assessed by SUCRA values. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral
Relative ranking of six drugs assessed by SUCRA values. H2RA: H2 receptor antagonists; PPI: proton pump inhibitor; i.v.: intravenous; p.o.: oral

In our study, we performed both traditional pair-wise and net-work meta-analysis to evaluate the effectiveness of H2RA, PPI, somatostatin analogue and tranexamic acid for patients with UGH. According to the results, PPI (i.v.) and PPI (p.o.) exhibit great therapeutic performance since they not only effectively reduce the risk of rebleeding, all-cause mortality and need for surgery but also reduce the average hospital stay as well as the average blood transfusion amount. Tranexamic acid also ranks high with respect to rebleeding, all-cause mortality and need for surgery. H2RA exhibits a low rank with respect to these clinical outcomes and therefore its tolerability should be further verified.

This is a large-scale study which simultaneously evaluates the tolerability of various UGH medications. Our results exhibit strong consistency with those suggested by Zhang et al. who concluded that PPI is more effective than H2RA for UGH patients [58]. Another study indicates that high-dose PPI is as effective as low-dose PPI with respect to the risk of rebleeding, need for surgery and mortality [59,60]. As suggested by previous studies, the effectiveness and tolerability of PPI is not significantly affected by its administration routes. However, our results suggest that oral administration of PPI appears to be more tolerable than intravenous PPI since the corresponding risk of adverse effects is reduced [61].

H2RA is a medication which is usually recommended for managing bleeding gastric ulcer and its effectiveness in bleeding duodenal ulcer is still uncertain [62]. Furthermore, bleeding location appears to have influence on the effectiveness of H2RA but evidence with respect to this issue is very limited. The effect of tranexamic acid on upper gastrointestinal bleeding has also been mentioned [63], while further studies were still recommended to assess the effectiveness and tolerability of tranexamic acid.

The net-work meta-analysis enables us to combine both direct and indirect evidences so that various medications can be evaluated simultaneously. Results from traditional meta-analysis and net-work meta-analysis were in contradiction in some cases. Although the inclusion of indirect evidence significantly increased our sample size, the corresponding heterogeneity and bias might also be enlarged. Apart from that, we assume that patients with no treatments have the same effect as those who were treated by placebo and such an assumption should be validated. Moreover, the blood transfusion amount was reported in different units and we converted different units of blood transfusion amount into a common standard so that results can be comparable. Also, we did not perform any stratified analysis with respect to dose and administration route which are likely to be associated with the corresponding clinical outcomes. Finally, bleeding location of each patient was not considered in our study and this issue should be addressed by further studies.

Overall, PPI is an effective medication for UGH. Both intravenous and oral administration of PPI may provide equivalent tolerability for UGH patients. However, high dose PPI is associated with an increase in the risk of all cause death and need for surgery compared to low dose PPI. On the other hand, H2RA is not recommended for UGH patients since they do not provide effective solution for rebleeding, need for surgery and all-cause mortality. Tranexamic acid is also considered as a promising for UGH patients. Nevertheless, we encourage researchers and clinicians to work together for confirming the above conclusions since determining the appropriate medication for UGH patients is not simply a question to be solved by trial-and-error.

It's funded by a grant (No. 81370487 to Q. Gao) from National Natural Science Foundation of China.

The authors declare no commercial or financial conflict of interest.

1.
Loperfido S, Baldo V, Piovesana E, Bellina L, Rossi K, Groppo M, Caroli A, Dal Bo N, Monica F, Fabris L, Salvat HH, Bassi N, Okolicsanyi L: Changing trends in acute upper-GI bleeding: a population-based study. Gastrointest Endosc 2009;70:212-224.
2.
Crooks CJ, West J, Card TR: Comorbidities affect risk of nonvariceal upper gastrointestinal bleeding. Gastroenterology 2013;144:1384-1393, 1393.e1381-1382; quiz e1318-1389.
3.
Sachs G, Shin JM, Howden CW: Review article: the clinical pharmacology of proton pump inhibitors. Aliment Pharmacol Ther 2006;23 Suppl 2:2-8.
4.
Zajac P, Holbrook A, Super ME, Vogt M: An overview: Current clinical guidelines for the evaluation, diagnosis, treatment, and management of dyspepsia. Osteopathic Family Physician 2013;5:79-85.
5.
Panula P, Chazot PL, Cowart M, Gutzmer R, Leurs R, Liu WL, Stark H, Thurmond RL, Haas HL: International Union of Basic and Clinical Pharmacology. XCVIII. Histamine Receptors. Pharmacol Rev 2015;67:601-655.
6.
Avgerinos A, Sgouros S, Viazis N, Vlachogiannakos J, Papaxoinis K, Bergele C, Sklavos P, Raptis SA: Somatostatin inhibits gastric acid secretion more effectively than pantoprazole in patients with peptic ulcer bleeding: a prospective, randomized, placebo-controlled trial. Scand J Gastroenterol 2005;40:515-522.
7.
Gotzsche PC, Gjorup I, Bonnen H, Brahe NE, Becker U, Burcharth F: Somatostatin v placebo in bleeding oesophageal varices: randomised trial and meta-analysis. BMJ 1995;310:1495-1498.
8.
Napolitano LM, Cohen MJ, Cotton BA, Schreiber MA, Moore EE: Tranexamic acid in trauma: how should we use it? J Trauma Acute Care Surg 2013;74:1575-1586.
9.
Bennett C, Klingenberg SL, Langholz E, Gluud LL: Tranexamic acid for upper gastrointestinal bleeding. Cochrane Database Syst Rev 2014;11:CD006640.
10.
Knobloch K, Yoon U, Vogt PM: Preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement and publication bias. J Craniomaxillofac Surg 2011;39:91-92.
11.
Salanti G, Ades AE, Ioannidis JP: Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol 2011;64:163-171.
12.
Biggs JC, Hugh TB, Dodds AJ: Tranexamic acid and upper gastrointestinal haemorrhage-a double-blind trial. Gut 1976;17:729-734.
13.
von Holstein CC, Eriksson SB, Kallen R: Tranexamic acid as an aid to reducing blood transfusion requirements in gastric and duodenal bleeding. Br Med J (Clin Res Ed) 1987;294:7-10.
14.
Christiansen J, Ottenjann R, Von Arx F: Placebo-controlled trial with the somatostatin analogue SMS 201-995 in peptic ulcer bleeding. Gastroenterology 1989;97:568-574.
15.
Coraggio F, Bertini G, Catalano A, Scarpato P, Gualdieri L: Clinical, controlled trial of somatostatin with ranitidine and placebo in the control of peptic hemorrhage of the upper gastrointestinal tract. Digestion 1989;43:190-195.
16.
Valenzuela JE, Schubert T, Fogel MR, Strong RM, Levine J, Mills PR, Fabry TL, Taylor LW, Conn HO, Posillico JT: A multicenter, randomized, double-blind trial of somatostatin in the management of acute hemorrhage from esophageal varices. Hepatology 1989;10:958-961.
17.
Daneshmend TK, Hawkey CJ, Langman MJ, Logan RF, Long RG, Walt RP: Omeprazole versus placebo for acute upper gastrointestinal bleeding: randomised double blind controlled trial. BMJ 1992;304:143-147.
18.
Walt RP, Cottrell J, Mann SG, Freemantle NP, Langman MJ: Continuous intravenous famotidine for haemorrhage from peptic ulcer. Lancet 1992;340:1058-1062.
19.
Lanas A, Artal A, Blas JM, Arroyo MT, Lopez-Zaborras J, Sainz R: Effect of parenteral omeprazole and ranitidine on gastric pH and the outcome of bleeding peptic ulcer. J Clin Gastroenterol 1995;21:103-106.
20.
Villanueva C, Balanzo J, Torras X, Sainz S, Soriano G, Gonzalez D, Vilardell F: Omeprazole versus ranitidine as adjunct therapy to endoscopic injection in actively bleeding ulcers: a prospective and randomized study. Endoscopy 1995;27:308-312.
21.
Hasselgren G, Lind T, Lundell L, Aadland E, Efskind P, Falk A, Hyltander A, Soderlund C, Eriksson S, Fernstrom P: Continuous intravenous infusion of omeprazole in elderly patients with peptic ulcer bleeding. Results of a placebo-controlled multicenter study. Scand J Gastroenterol 1997;32:328-333.
22.
Khuroo MS, Yattoo GN, Javid G, Khan BA, Shah AA, Gulzar GM, Sodi JS: A comparison of omeprazole and placebo for bleeding peptic ulcer. N Engl J Med 1997;336:1054-1058.
23.
Schaffalitzky de Muckadell OB, Havelund T, Harling H, Boesby S, Snel P, Vreeburg EM, Eriksson S, Fernstrom P, Hasselgren G: Effect of omeprazole on the outcome of endoscopically treated bleeding peptic ulcers. Randomized double-blind placebo-controlled multicentre study. Scand J Gastroenterol 1997;32:320-327.
24.
Lin HJ, Lo WC, Lee FY, Perng CL, Tseng GY: A prospective randomized comparative trial showing that omeprazole prevents rebleeding in patients with bleeding peptic ulcer after successful endoscopic therapy. Arch Intern Med 1998;158:54-58.
25.
Lau JY, Sung JJ, Lee KK, Yung MY, Wong SK, Wu JC, Chan FK, Ng EK, You JH, Lee CW, Chan AC, Chung SC: Effect of intravenous omeprazole on recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. N Engl J Med 2000;343:310-316.
26.
Hawkey GM, Cole AT, McIntyre AS, Long RG, Hawkey CJ: Drug treatments in upper gastrointestinal bleeding: value of endoscopic findings as surrogate end points. Gut 2001;49:372-379.
27.
Javid G, Masoodi I, Zargar SA, Khan BA, Yatoo GN, Shah AH, Gulzar GM, Sodhi JS: Omeprazole as adjuvant therapy to endoscopic combination injection sclerotherapy for treating bleeding peptic ulcer. Am J Med 2001;111:280-284.
28.
Udd M, Miettinen P, Palmu A, Heikkinen M, Janatuinen E, Pasanen P, Tarvainen R, Kairaluoma MV, Lohman M, Mustonen H, Julkunen R: Regular-dose versus high-dose omeprazole in peptic ulcer bleeding: a prospective randomized double-blind study. Scand J Gastroenterol 2001;36:1332-1338.
29.
Sheu BS, Chi CH, Huang CC, Kao AW, Wang YL, Yang HB: Impact of intravenous omeprazole on Helicobacter pylori eradication by triple therapy in patients with peptic ulcer bleeding. Aliment Pharmacol Ther 2002;16:137-143.
30.
Kaviani MJ, Hashemi MR, Kazemifar AR, Roozitalab S, Mostaghni AA, Merat S, Alizadeh-Naini M, Yarmohammadi H: Effect of oral omeprazole in reducing re-bleeding in bleeding peptic ulcers: a prospective, double-blind, randomized, clinical trial. Aliment Pharmacol Ther 2003;17:211-216.
31.
Hsu PI, Lo GH, Lo CC, Lin CK, Chan HH, Wu CJ, Shie CB, Tsai PM, Wu DC, Wang WM, Lai KH: Intravenous pantoprazole versus ranitidine for prevention of rebleeding after endoscopic hemostasis of bleeding peptic ulcers. World J Gastroenterol 2004;10:3666-3669.
32.
Cheng HC, Kao AW, Chuang CH, Sheu BS: The efficacy of high- and low-dose intravenous omeprazole in preventing rebleeding for patients with bleeding peptic ulcers and comorbid illnesses. Dig Dis Sci 2005;50:1194-1201.
33.
Jensen DM, Pace SC, Soffer E, Comer GM, Study G: Continuous infusion of pantoprazole versus ranitidine for prevention of ulcer rebleeding: a U.S. multicenter randomized, double-blind study. Am J Gastroenterol 2006;101:1991-1999; quiz 2170.
34.
Lin HJ, Lo WC, Cheng YC, Perng CL: Role of intravenous omeprazole in patients with high-risk peptic ulcer bleeding after successful endoscopic epinephrine injection: a prospective randomized comparative trial. Am J Gastroenterol 2006;101:500-505.
35.
Yilmaz S, Bayan K, Tuzun Y, Dursun M, Canoruc F: A head to head comparison of oral vs intravenous omeprazole for patients with bleeding peptic ulcers with a clean base, flat spots and adherent clots. World J Gastroenterol 2006;12:7837-7843.
36.
Bajaj JS, Dua KS, Hanson K, Presberg K: Prospective, randomized trial comparing effect of oral versus intravenous pantoprazole on rebleeding after nonvariceal upper gastrointestinal bleeding: a pilot study. Dig Dis Sci 2007;52:2190-2194.
37.
Hung WK, Li VK, Chung CK, Ying MW, Loo CK, Liu CK, Lam BY, Chan MC: Randomized trial comparing pantoprazole infusion, bolus and no treatment on gastric pH and recurrent bleeding in peptic ulcers. ANZ J Surg 2007;77:677-681.
38.
Jeong HK, Park CH, Jun CH, Lee GH, Kim HI, Kim HS, Choi SK, Rew JS: A prospective randomized trial of either famotidine or pantoprazole for the prevention of bleeding after endoscopic submucosal dissection. J Korean Med Sci 2007;22:1055-1059.
39.
Lau JY, Leung WK, Wu JC, Chan FK, Wong VW, Chiu PW, Lee VW, Lee KK, Cheung FK, Siu P, Ng EK, Sung JJ: Omeprazole before endoscopy in patients with gastrointestinal bleeding. N Engl J Med 2007;356:1631-1640.
40.
Morales GF, Pereira Lima JC, Hornos AP, Marques DL, Costa CS, Lima Pereira L, Lopes CV, Raymondi R, Marroni CA: Octreotide for esophageal variceal bleeding treated with endoscopic sclerotherapy: a randomized, placebo-controlled trial. Hepatogastroenterology 2007;54:195-200.
41.
Tsibouris P, Zintzaras E, Lappas C, Moussia M, Tsianos G, Galeas T, Potamianos S: High-dose pantoprazole continuous infusion is superior to somatostatin after endoscopic hemostasis in patients with peptic ulcer bleeding. Am J Gastroenterol 2007;102:1192-1199.
42.
Uedo N, Takeuchi Y, Yamada T, Ishihara R, Ogiyama H, Yamamoto S, Kato M, Tatsumi K, Masuda E, Tamai C, Yamamoto S, Higashino K, Iishi H, Tatsuta M: Effect of a proton pump inhibitor or an H2-receptor antagonist on prevention of bleeding from ulcer after endoscopic submucosal dissection of early gastric cancer: a prospective randomized controlled trial. Am J Gastroenterol 2007;102:1610-1616.
43.
Andriulli A, Loperfido S, Focareta R, Leo P, Fornari F, Garripoli A, Tonti P, Peyre S, Spadaccini A, Marmo R, Merla A, Caroli A, Forte GB, Belmonte A, Aragona G, Imperiali G, Forte F, Monica F, Caruso N, Perri F: High- versus low-dose proton pump inhibitors after endoscopic hemostasis in patients with peptic ulcer bleeding: a multicentre, randomized study. Am J Gastroenterol 2008;103:3011-3018.
44.
Garrido A, Giraldez A, Trigo C, Leo E, Guil A, Marquez JL: [Intravenous proton-pump inhibitor for acute peptic ulcer bleeding-is profound acid suppression beneficial to reduce the risk of rebleeding?]. Rev Esp Enferm Dig 2008;100:466-469.
45.
Yuksel I, Ataseven H, Koklu S, Ertugrul I, Basar O, Odemis B, Ibis M, Sasmaz N, Sahin B: Intermittent versus continuous pantoprazole infusion in peptic ulcer bleeding: a prospective randomized study. Digestion 2008;78:39-43.
46.
Choi KD, Kim N, Jang IJ, Park YS, Cho JY, Kim JR, Shin JM, Jung HC, Song IS: Optimal dose of intravenous pantoprazole in patients with peptic ulcer bleeding requiring endoscopic hemostasis in Korea. J Gastroenterol Hepatol 2009;24:1617-1624.
47.
Javid G, Zargar SA, R US, Khan BA, Yatoo GN, Shah AH, Gulzar GM, Sodhi JS, Khan MA: Comparison of p.o. or i.v. proton pump inhibitors on 72-h intragastric pH in bleeding peptic ulcer. J Gastroenterol Hepatol 2009;24:1236-1243.
48.
Mesihovic R, Vanis N, Mehmedovic A, Gornjakovic S, Gribajcevic M: Proton pump inhibitors after endoscopic hemostasis in patients with peptic ulcer bleeding. Med Arh 2009;63:323-327.
49.
Tsai JJ, Hsu YC, Perng CL, Lin HJ: Oral or intravenous proton pump inhibitor in patients with peptic ulcer bleeding after successful endoscopic epinephrine injection. Br J Clin Pharmacol 2009;67:326-332.
50.
Hsu YC, Perng CL, Yang TH, Wang CS, Hsu WL, Wu HT, Cheng YC, Chiang MF, Lin HJ: A randomized controlled trial comparing two different dosages of infusional pantoprazole in peptic ulcer bleeding. Br J Clin Pharmacol 2010;69:245-251.
51.
Mostaghni AA, Hashemi SA, Heydari ST: Comparison of oral and intravenous proton pump inhibitor on patients with high risk bleeding peptic ulcers: a prospective, randomized, controlled clinical trial. Iran Red Crescent Med J 2011;13:458-463.
52.
Chen CC, Lee JY, Fang YJ, Hsu SJ, Han ML, Tseng PH, Liou JM, Hu FC, Lin TL, Wu MS, Wang HP, Lin JT: Randomised clinical trial: high-dose vs. standard-dose proton pump inhibitors for the prevention of recurrent haemorrhage after combined endoscopic haemostasis of bleeding peptic ulcers. Aliment Pharmacol Ther 2012;35:894-903.
53.
Kim HK, Kim JS, Kim TH, Kim CW, Cho YS, Kim SS, Chae HS, Han SW, Park YW, Son HS, Min JY, Cho GJ, Bag JS, Choi SO: Effect of high-dose oral rabeprazole on recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. Gastroenterol Res Pract 2012;2012:317125.
54.
Sakurada T, Kawashima J, Ariyama S, Kani K, Takabayashi H, Ohno S, Kato S, Yakabi K: Comparison of adjuvant therapies by an H2-receptor antagonist and a proton pump inhibitor after endoscopic treatment in hemostatic management of bleeding gastroduodenal ulcers. Dig Endosc 2012;24:93-99.
55.
Yen HH, Yang CW, Su WW, Soon MS, Wu SS, Lin HJ: Oral versus intravenous proton pump inhibitors in preventing re-bleeding for patients with peptic ulcer bleeding after successful endoscopic therapy. BMC Gastroenterol 2012;12:66.
56.
Demetrashvili ZM, Lashkhi IM, Ekaladze EN, Kamkamidze GK: Comparison of intravenous pantoprazole with intravenous ranitidine in peptic ulcer bleeding. Georgian Med News 2013;7-11.
57.
Sung JJ, Suen BY, Wu JC, Lau JY, Ching JY, Lee VW, Chiu PW, Tsoi KK, Chan FK: Effects of intravenous and oral esomeprazole in the prevention of recurrent bleeding from peptic ulcers after endoscopic therapy. Am J Gastroenterol 2014;109:1005-1010.
58.
Zhang YS, Li Q, He BS, Liu R, Li ZJ: Proton pump inhibitors therapy vs H2 receptor antagonists therapy for upper gastrointestinal bleeding after endoscopy: A meta-analysis. World J Gastroenterol 2015;21:6341-6351.
59.
Wu LC, Cao YF, Huang JH, Liao C, Gao F: High-dose vs low-dose proton pump inhibitors for upper gastrointestinal bleeding: a meta-analysis. World J Gastroenterol 2010;16:2558-2565.
60.
Wang CH, Ma MH, Chou HC, Yen ZS, Yang CW, Fang CC, Chen SC: High-dose vs non-high-dose proton pump inhibitors after endoscopic treatment in patients with bleeding peptic ulcer: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med 2010;170:751-758.
61.
Neumann I, Letelier LM, Rada G, Claro JC, Martin J, Howden CW, Yuan Y, Leontiadis GI: Comparison of different regimens of proton pump inhibitors for acute peptic ulcer bleeding. Cochrane Database Syst Rev 2013;6:CD007999.
62.
Levine JE, Leontiadis GI, Sharma VK, Howden CW: Meta-analysis: the efficacy of intravenous H2-receptor antagonists in bleeding peptic ulcer. Aliment Pharmacol Ther 2002;16:1137-1142.
63.
Gluud LL, Klingenberg SL, Langholz E: Tranexamic acid for upper gastrointestinal bleeding. Cochrane Database Syst Rev 2012;1:CD006640.
Open Access License / Drug Dosage / Disclaimer
This article is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND). Usage and distribution for commercial purposes as well as any distribution of modified material requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.