Is the era of intravenous proton pump inhibitors coming to an end in patients with bleeding peptic ulcers? Meta‐analysis of the published literature

Abstract

Aims

Oral and intravenous proton pump inhibitors (PPIs) are equipotent in raising gastric pH. However, it is not known whether oral PPIs can replace intravenous PPIs in patients with bleeding peptic ulcers.

Methods

We conducted a systematic review and meta‐analysis of randomized controlled trials to compare oral and intravenous PPIs among patients with peptic ulcer bleeding. A search of all major databases and relevant journals from inception to April 2015, without a restriction on languages, was performed.

Results

A total of 859 patients from seven randomized controlled trials were included in the meta‐analysis. Similar pooled outcome measures were demonstrated between the two groups in terms of oral PPIs vs. intravenous PPIs in the rate of recurrent bleeding within the 30‐day follow‐up period [risk ratio = 0.90; 95% confidence interval (CI): 0.58, 1.39; P = 0.62; I² = 0%). In terms of the rate of mortality, both oral and intravenous PPIs showed similar outcomes, and the pooled risk ratio was 0.88 (95% CI: 0.29, 2.71; P = 0.82; I² = 0%). Likewise, no significant difference was detected in the need for blood transfusion and length of hospital stay; the pooled mean differences were −0.14 (95% CI: –0.39, 0.12; P = 0.29; I² = 32%) and −0.60 (95% CI: –1.42, 0.23; P = 0.16; I² = 79%), respectively.

Conclusions

Our results suggest that oral PPIs are a feasible, safe alternative to intravenous PPIs in patients with bleeding peptic ulcers, and may be able to replace intravenous PPIs as the treatment of choice in these patients.

Introduction

Acute upper gastrointestinal bleeding is a potentially life‐threatening condition which remains among the most common reasons for emergency hospital admission. It has an annual incidence ranging from approximately 50 to 150 per 100 000 of the population. With significant morbidity and mortality, peptic ulcer disease is the most common cause of acute upper gastrointestinal bleeding, accounting for about half of episodes 1, 2, 3, 4, 5. Recurrent bleeding occurs in 14–36% of patients and could lead to end‐organ dysfunction and even death, which presents significant challenges for gastroenterologists after initial bleeding control by resuscitation and endoscopic therapy 6, 7, 8. Gastric acid disturbs the homeostasis of ulcers in the stomach and duodenum by inhibiting clot formation and promoting clot lysis; thus, the use of high‐dose intravenous proton pump inhibitors (PPIs) has become standard practice in the management of upper gastrointestinal bleeding 9.

Several studies have established the efficacy of the adjuvant use of both intravenous and oral PPIs in high‐risk bleeding ulcers after endoscopic therapy 10, 11, 12, 13, 14, 15, 16. However, the relative effectiveness of oral and intravenous routes of administration remains controversial. Studies have shown that the risk of rebleeding and continued bleeding from an ulcer is strongly associated with the stigmata seen at endoscopic examinations (major stigmata include spurting, oozing vessels, nonbleeding visible vessels or fresh adherent clots; old adherent clots were considered minor stigmata, and all adherent clots were considered as low‐risk stigmata) 10, 11, 17. It has been reported that the administration of oral PPIs can be appropriate and adequate for most patients who present with ulcer bleeding and have low‐risk stigmata with no requirement for intravenous PPI treatment 18. Overuse of intravenous PPIs in acute upper gastrointestinal bleeding is a common practice worldwide 10, 17. It is worth noting that, although the use of high‐dose intravenous PPIs is widely accepted, the efficacy of high‐dose oral PPIs in treating acute peptic ulcer bleeding has remained a topic of dispute 19. Due to the cost‐effectiveness of oral PPIs with respect to intravenous PPIs, the era of widespread oral PPI use for patients with high‐risk peptic ulcer bleeding may have arrived. Recently, a meta‐analysis demonstrated a similar effectiveness between oral and intravenous PPIs but the results were limited by the fact that the trials were open label and the relatively small sample size. Meanwhile, the trial of Sung et al., which included the largest patient sample size among the published studies, was published in 2014 19, 20. Furthermore, the previous studies came from South Asia, and the use of endoscopic therapy was not standardized. The present meta‐analysis was the most up to date, and compared the safety and efficacy of oral vs. intravenous PPI in high‐risk peptic ulcer bleeding when given with or without endoscopic intervention.

Methods

The methods of literature search, inclusion and exclusion criteria, outcome measures and methods of statistical analysis were defined in a protocol according to the Preferred Items for Systematic Reviews and Meta‐Analysis (PRISMA) recommendations for study reporting (see Supporting Information) 21.

Information sources and search strategy

The primary sources of the reviewed studies up to April 2015, without restriction on the languages of publication, were Pubmed, Embase, Cancerlit, Cochrane and ISI Web of Science. We combined the database‐specific search terms of proton pump inhibitors and PPI (pantoprazole, omeprazole, lansoprazole, esomeprazole, rabeprazole, dexlansoprazole) and intravenous and oral, as well as truncated search terms utilizing the wildcard (‘*’) character for patients with bleeding peptic ulcers after endoscopic haemostasis. The ‘related articles’ function was also used to broaden the search, and the computer search was supplemented with manual searches of reference lists for all retrieved original articles and review articles, primary studies and abstracts from meetings, to identify other studies not found in the computer search. Finally, only randomized controlled trials (RCTs) were given full consideration for analysis. Authors of relevant abstracts were contacted to obtain any unpublished data (if available). When the results of a single study were reported in more than one publication, only the most recent and complete data were included.

Eligibility criteria

All included clinical RCTs assessed the clinical effectiveness of oral PPIs vs. intravenous PPIs in patients older than 18 years with symptoms and signs of upper gastrointestinal bleeding, such as haematemesis, melena or the presence of blood in the nasogastric tube lavage. High‐risk peptic ulcer bleeding was defined as active bleeding (Forrest IA, IB) or having a nonbleeding protuberant vessel (Forrest IIA) 22. Haemostasis interventions by endoscopic therapy, haemoclips or heater probe were performed within 24 h of admission, with or without endoscopic injection. Exclusion criteria varied between studies but included the following: patients who were pregnant; did not obtain initial haemostasis with endoscopic therapy; had a history of chronic liver disease and portal hypertension; had a gastroduodenal malignancy; had undergone gastric surgery; had known adverse drug reactions to the trial drugs or were currently using antisecretory drugs, histamine type 2 receptor antagonists or PPIs; did not give written informed consent; had bleeding tendency; or had used PPIs within 14 days of enrolment. In addition, nonhuman studies, non‐experimental trials, review articles, editorials, letters/case reports and articles not reporting the outcomes of interest were excluded from the current meta‐analysis.

Data extraction and quality assessment

Two reviewers (JZX and JHS) independently considered the eligibility of potential abstracts and titles. Retrieval strategies were refined with a smaller set of reports. When there was a disagreement about a study or a lack of information for an accurate assessment of eligibility, the study was carried to the full‐text stage for evaluation. To ensure homogeneity of data collection and to rule out any subjective influence in data gathering and entry, data were extracted independently and in duplicate by another two reviewers (MTT and YZ), and discrepancies were resolved by iteration, discussion and reaching a consensus.

For each RCT, the following data were extracted: the baseline characteristics of number of patients, age, gender ratio, ulcer size, successful endoscopic therapy history, Forrest classification [stigmata of haemorrhage, a classification of upper gastrointestinal haemorrhage used for the purposes of comparison and in selecting patients for endoscopic treatment: Forrest I: acute haemorrhage (Forrest IA, spurting haemorrhage; Forrest IB, oozing haemorrhage); Forrest II: signs of recent haemorrhage (Forrest IIA, visible vessel; Forrest IIB, adherent clot; Forrest IIC, flat pigmented haematin on ulcer base); Forrest III: lesions without active bleeding (lesions without signs of recent haemorrhage or fibrin‐covered clean ulcer base)] and medication 22. The qualities of the included RCTs were assessed by the Cochrane Risk of Bias Tool 23. We reached complete concordance for all variables assessed. Moreover, authors of included published studies were contacted whenever we found that data essential for the meta‐analysis were missing or unclear. The effects of oral PPIs vs. intravenous PPI on rebleeding, transfusion requirements, length of hospital stay and mortality within 30 days of bleeding peptic ulcers were pooled from included RCTs.

Data synthesis and analysis

We performed the meta‐analysis using standard methods to evaluate the overall effect of oral and intravenous PPIs on recurrent bleeding, mortality, the need for blood transfusion and length of hospital stay according to an intent‐to‐treat (ITT) principle 23. The reported risk ratio (RR) and 95% confidence interval (CI) were used in the analysis. Medians were converted to means using the technique described by Hozo et al. 24

The fixed‐effect model (by the application of the generalized inverse variance method) was first used to pool the results; this assumes that all the studies share the same common (fixed or nonrandom) effect size. Variance was used to calculate the weight of each study. The standard heterogeneity test statistic, I ² , was used to assess the consistency of the effect sizes, which indicates the percentage of the variability in effect estimates that is due to true between‐study variance rather than within‐study variance. Heterogeneity was considered not to be statistically significant when the Cochrane Q test P value was more than 0.1. In cases of heterogeneity, a meta‐analysis was performed, applying the random‐effects model, which assumes that studies do not have the same effect size and assigns a weight to each study, taking into account both within‐ and between‐study variance based on the method of DerSimonian and Laird 25. In addition, an I ² value of less than 25% was defined as representing low heterogeneity, a value between 25% and 50% was defined as moderate heterogeneity and a value of greater than 50% was defined as high heterogeneity 23.

Subgroup analyses, which considered more homogeneous studies, were performed to identify subsets of patients more likely to benefit from this treatment and to assess the efficacy of different studies. To determine the extent to which the combined risk estimate might be affected by individual studies, sensitivity analysis was performed by consecutively omitting every study from the meta‐analysis (leave‐one‐out procedure). The mixed‐effect model was applied to obtain summary effects within and across subgroups, whereas subgroups were also compared by means of the I ² statistic 26. Funnel plots were used to screen for publication bias 27. Meta‐analysis was conducted using the Review Manager (RevMan) Meta‐Analysis software, version 5.1.6, and 95% CIs were calculated as estimates of precision for RR. The statistical tests were two sided, and P values <0.05 were considered to be statistically significant 28.

Results

Study characteristics

Table 1 lists the baseline characteristics of the seven included RCTs (a flowchart of publication search and selection is presented in Figure 1). All were single‐centre studies 19, 29, 30, 31, 32, 33, 34, with no significant difference between the groups in demographic information. The total number of patients per study ranged from 25 to 244. The majority of the patients were male and were randomized to receive oral or intravenous PPIs using a random number table or a computer‐generated sequence, or based on even and odd days of the month 33. Initial endoscopic haemostasis was defined as no visible haemorrhage with observation for 3 min. Ultimate haemostasis was defined as no rebleeding within 14 days after endoscopic therapy. Rebleeding was concluded if active bleeding, fresh blood or blood clots were found by emergent endoscopy, or if unstable vital signs, continuous tarry, bloody stool or a drop in haemoglobin level > 20 g l^–1 within 24 h were noted. Severity of bleeding was assessed by the Rockall scoring system in both groups 35. A blood transfusion was given if the haemoglobin level decreased to lower than 90 g l^–1 or if the patient's vital signs deteriorated. If a state of shock occurred, blood was transfused independently of haemoglobin levels. The patient's clinical status was monitored after discharge by telephone conversation if any rebleeding or death occurred within 30 days. It should be emphasized that most physicians do not prefer to administer intravenous PPIs in patients with low‐risk ulcers, as reported in the study by Yilmaz et al. 30. The other included RCTs all involved enrolled patients with high‐risk bleeding peptic ulcers after successful endoscopic interventions.

Table 1.

Baseline characteristics of the included RCTs

Authors/country	Comparison	No. of patients (male%)	Age (years, SD)	Ulcer size (cm, SD)	Successful endoscopic therapy history *	Forrest classification †			Medication
						IA/IB	IIA/IIB/IIC	III
Jae et al. 29	ORP	19(NA)	NA	NA	Yes	6	13	0	Pantoprazole 40 mg, po, bid for 5 days
/South Korea (2006)	IVP	19(NA)	NA	NA	Yes	4	15	0	Pantoprazole 80 mg, IV drip, qid for 3 days + pantoprazole 40 mg, po, qid for 8 weeks
Yilmaz et al. 30	ORP	99(67)	52.8(19.6)	1.06(0.6)	No	0	37	62	Omeprazole 40 mg, po, bid for 3 days + omeprazole 40 mg, po, qid for 6 weeks
/Turkey (2006)	IVP	112(71)	52.7(17.1)	1.05(0.4)	No	0	30	82	Omeprazole 80 mg, IV drip, qid for 3 d + omeprazole 40 mg, po, qid for 6 weeks
Bajaj et al. 31	ORP	12(50)	59.5(19.4)	NA	Yes	3	5	4	Pantoprazole 80 mg, po, bid for 3 d + pantoprazole 40 mg, po, bid for 30 days
/United States (2007)	IVP	13(77)	66.2(6.2)	NA	Yes	4	4	5	Pantoprazole 80 mg, IV drip, qid for 3 d + pantoprazole 40 mg, po, bid for 30 days
Tsai et al. 32	ORP	78(74)	67.9(15.2)	1.12(0.4)	Yes	33	44	0	Rabeprazole 20 mg, po, bid for 3days + rabeprazole 20 mg, po, qid for 2 months
/Taiwan (2008)	IVP	78(71)	69.4(15.0)	1.06(0.4)	Yes	31	47	0	Omeprazole 40 mg, IV drip, bid for 3 days + rabeprazole 20 mg, po, qid for 2 months/esomeprazole 40 mg, po, qid for 2 months
Mostaghni et al. 33	ORP	44(75)	57.3(16.5)	NA	Yes	14	30	0	Omeprazole 40 mg, po, bid for 3 days + omeprazole 20 mg, po, qid for 30 days
Iran (2011)	IVP	41(73)	61.7(17.2)	NA	Yes	12	29	0	Pantoprazole 80 mg, IV drip, qid for 2–3 days + omeprazole 20 mg, po, qid for 30 days
Yen et al. 34	ORP	50(68)	62.7(16.3)	1.60(1.4)	Yes	22	28	0	Lansoprazole 30 mg, po, qid for 3 days + lansoprazole 30 mg, po, qid for 2 months
Taiwan (2012)	IVP	50(74)	65.0(15.6)	1.40(1.4)	Yes	21	29	0	Esomeprazole 40 mg, IV drip, qid for 3 days + esomeprazole 40 mg, po, qid for 2 months
Sung et al. 19	ORP	126(76)	63.8(17.0)	2.00(12.7)‡	Yes	55	71	0	Placebo, IV drip, qid for 3 days + esomeprazole 40 mg, po, bid for 27 days
Hong Kong (2014)	IVP	118(72)	64.3(15.5)	2.00(13.6)‡	Yes	51	67	0	Esomeprazole 80 mg, IV drip, qid*3 days + placebo, po, bid for 27 days

^* Endoscopic successful haematemesis is considered successful when active bleeding (Forrest IA and IB) was stopped. For nonbleeding patients (Forrest IIA and IIB), the end point of endoscopy haematemesis was the successful application of haemoclips or a heater probe on the bleeding vessel after the removal of blood clot at the ulcer base.

^† Stigmata of haemorrhage, a classification of upper gastrointestinal haemorrhage used for purposes of comparison and in selecting patients for endoscopic treatment: Forrest I: acute haemorrhage (Forrest IA, spurting haemorrhage; Forrest IB, oozing haemorrhage); Forrest II: signs of recent haemorrhage (Forrest IIA, visible vessel; Forrest IIB, adherent clot; Forrest IIC, flat pigmented haematin on ulcer base); Forrest III: lesions without active bleeding (lesions without signs of recent haemorrhage or fibrin‐covered clean ulcer base).

^‡ Ulcer size >2 cm (with its percentage in each group in parentheses). Bid, twice daily; IV, intravenous; IVP, intravenous proton pump inhibitor; NA, not available; ORP, oral proton pump inhibitor; po, orally; qid, four times daily; SD, standard deviation.

Figure 1.

Flowchart of publication search and selection

Study quality assessment

Figure 2 summarizes the possible risks of bias of the trials included in the meta‐analysis. Selective reporting, baseline imbalance and sources of finding bias were avoided in all RCTs. Generation of adequate sequence was not reported in two studies 29, 31, and concealment of allocation was not fully implemented in another two trials 33, 34. Sample size estimation was performed in three studies 19, 32, 34. Yilmaz et al. conducted the RCT in a double‐blind manner as all treatment assignments were revealed at the end of the study. In that study, a person outside the study staff placed the two drug formulations into sealed, opaque envelopes and coded them based on random table numbers. In the trial by Sung et al., study medications were placed in sealed, consecutively numbered packages which would only be opened after informed consent was obtained by an endoscopist or research nurse 19. Other studies did not specify the blinding manner. In addition, ITT analysis was only applied by Sung et al.; however, their study was terminated prematurely owing to recruitment difficulties.

Figure 2.

Risks of bias in the trials included in the meta‐analysis

Meta‐analysis

As shown in Figure 3, similar pooled outcome measures were demonstrated between the two groups in terms of oral PPIs vs. intravenous PPIs in the rate of recurrent bleeding within a 30‐day follow‐up period (RR = 0.90; 95% CI: 0.58, 1.39; P = 0.62; I² = 0%). In subgroup analysis, there was no significant difference between the oral and intravenous PPI groups in the rate of recurrent bleeding within 72 h (2.4% vs. 5.1%, P = 0.26) and 7 days (5.6% vs. 6.8%, P = 0.68) 19. Two studies reported the 15‐day rebleeding result, and a nonsignificant difference pooled outcome was also demonstrated (RR = 1.07; 95% CI: 0.55, 2.10; P = 0.84; I² = 0%) 32, 34. In terms of the rate of mortality, both oral and intravenous PPIs showed similar outcomes, and the pooled RR was 0.88 (95% CI: 0.29, 2.71; P = 0.82), with a null heterogeneity (I² = 0%) (Figure 4). Likewise, no significant difference was detected in the need for blood transfusion or length of hospital stay, and the pooled mean differences were −0.14 (95% CI: –0.39, 0.12; P = 0.29; I² = 32%) and −0.60 (95% CI: –1.42, 0.23; P = 0.16; I² = 79%), respectively (Figures 5, 6). Additionally, with only seven studies, it is difficult to interpret publication bias for the present meta‐analysis. However, potential missing studies appeared to be those with statistically significant results, which is highly unlikely. Therefore, publication bias is unlikely to have been a concern in the present review (Figure 7). Sensitivity analysis also confirmed the stability of the pooled results.

Figure 3.

Oral proton pump inhibitor (PPI) vs. intravenous PPI in the rate of recurrent bleeding within a 30‐day follow‐up period. CI, confidence interval; IVP, intravenous PPI; ORP, oral PPI; M‐H, Mantel‐Haenszel method

Figure 4.

Oral proton pump inhibitor (PPI) vs. intravenous PPI in the rate of mortality within a 30‐day follow‐up period. CI, confidence interval; IVP, intravenous PPIs; ORP, oral PPIs; M‐H, Mantel‐Haenszel method

Figure 5.

Oral proton pump inhibitor (PPI) vs. intravenous PPI in blood transfusion within a 30‐day follow‐up period. CI, confidence interval; IV, intravenous; IVP, intravenous PPI; ORP, oral PPI

Figure 6.

Oral proton pump inhibitor (PPI) vs. intravenous PPI in length of hospital stay. CI, confidence interval; IV, intravenous; IVP, intravenous PPI; ORP, oral PPI

Figure 7.

Symmetry of Funnel plot depicted little publication bias in trials included in the meta‐analysis

Discussion

Based on these studies, we conducted the present meta‐analysis to clarify the clinical efficacy of high‐dose oral and intravenous PPIs after successful endoscopic therapy for patients with bleeding peptic ulcers. The most important finding in the present study was that adjuvant pharmacotherapy of oral administration of PPI was as effective as that of intravenous PPI in controlling bleeding peptic ulcers. Both oral and intravenous PPIs achieved similar success in inducing haemostasis in recurrent bleeding at 72 h, 7 days, 15 days and 30 days of follow‐up after the administration of standard endoscopic therapy. Other clinical outcomes, including the need for blood transfusion, surgery, length of hospital stay and mortality, were all comparable.

Rebleeding after initial control of acute nonvariceal upper gastrointestinal bleeding is an important clinical issue that continues to be a significant problem, requiring resuscitative and endoscopic therapy, and possibly resulting in end‐organ dysfunction and death 36, 37. Endoscopic therapy decreases, but does not eliminate, the risk of adverse outcomes for peptic ulcer bleeding. In addition, maintenance of an intragastric pH of > 6 has been considered to result in a lower rebleeding rate, so the use of PPIs following endoscopic therapy for bleeding peptic ulcers can help to promote platelet aggregation, stabilize blood clots and prevent fibrinolysis 7, 9, 38. In spite of this evidence, the optimal route and dosage of administration has remained controversial, although several recent studies have demonstrated the efficacy of intravenous PPIs in reducing the adverse outcome of peptic ulcer bleeding 39, 40. The administration of intravenous PPIs requires nursing supervision and hospital admission, leading to high costs, while oral administration is attractive due to widespread availability, ease of implementation and cost‐effectiveness. With this in mind, it would be reasonable to prescribe an oral PPI to patients with high‐risk bleeding ulcers, provided that it is as effective as its intravenous counterpart.

Gastric pH was used as a marker for the effectiveness of both oral and intravenous PPIs in the included RCTs. Based on previous studies showing that high doses of intravenous PPI followed by continuous infusion are able to sustain a higher intragastric pH, current guidelines suggest this clinical pathway as standard adjuvant pharmacotherapy for bleeding peptic ulcers, especially in Western countries 28, 41. Although high doses of intravenous PPI have been demonstrated to be superior to placebo 42, currently available evidence does not indicate that oral administration is inferior to intravenous administration with regard to the clinical outcome of bleeding peptic ulcers in patients following endoscopic therapy. The benefits of PPIs appear to be independent of the route of administration and dose, as reported by Andriulli et al., who evaluated 35 RCTs comparing PPI vs. placebo or histamine type 2 receptor antagonist 16. In addition, there is no evidence to indicate that bleeding, surgery or mortality would be influenced by the route of PPI administration, as reported in the meta‐analysis of Leontiadis et al. 43. More recently, Javid et al. demonstrated that there is no significant difference among various PPIs, administered via different routes, on raising gastric pH above 6 for 72 h after successful endoscopic haemostasis, and high doses of various oral and intravenous PPIs are equivalent in their ability to suppress gastric acid secretion 44. Furthermore, a similar effectiveness between oral and intravenous PPIs in raising intragastric pH has been demonstrated. After ingestion, the effect of an oral PPI would initiate 1 h, with the maximum plasma concentration achieved 2–3 h later; thus, oral administration is associated with high bioavailability 45.

It seems that oral PPIs have substantially equivalent efficacy to intravenous PPIs in decreasing the adverse outcomes of high‐risk bleeding ulcers, especially rebleeding. It has also been shown that high‐dose oral PPI following endoscopic treatment significantly decreases rebleeding, and probably mortality, when compared with placebo 39. From the included studies, the overall rebleeding rate reported by Tsai et al. 32 appeared to be higher in those receiving PPIs alone than in those receiving endoscopic intervention and PPIs; this may be because they adopted epinephrine injection as the primary haemostatic measure, which may be considered suboptimal for high‐risk bleeders 46. Tsai et al. had not expected to find that epinephrine injection was the best available therapy 32. However, their study revealed that oral and intravenous PPIs were similarly effective as adjuvant pharmacotherapy, even if the endoscopic therapy was limited to epinephrine injection. Notably, they conducted a head‐to‐head trial comparing a high dose of oral rabeprazole with a regular dose of intravenous omeprazole (40 mg intravenous infusion every 12 h), the latter being considered (in spite of controversial results) to be more effective 32. Therefore, Mostaghi et al. carried out a study to compare oral and intravenous PPIs directly 33. In the latter research, they not only achieved a lower rate of rebleeding in the omeprazole group, but also revealed that high‐dose oral omeprazole (40 mg twice daily) is equally effective as high‐dose intravenous pantoprazole (8 mg every hour) in reducing recurrent bleeding, the need for blood transfusion, length of hospital stay and mortality after successful endoscopic haemostasis.

Similar intragastric pH, in spite of different routes of administration, may suggest similar clinical outcomes. Yen et al. suggested that high doses of PPI (rather than regular doses) via the oral route may achieve comparable clinical outcomes to those found with high‐dose intravenous PPIs 34. Nevertheless, PPI dosage effects as they relate to clinical efficacy are another unsettled issue in the management of patients with peptic ulcer bleeding. The focus of the present study was to investigate the effects of route rather than the dosage. Hopefully, well‐designed future studies will help to resolve the PPI dosage controversy.

A previous meta‐analysis reported that PPI therapy was effective only in patients with high‐risk stigmata for rebleeding, which led Yilmaz et al. to design their study to make a comparison between oral and intravenous omeprazole in patients with bleeding peptic ulcers without high‐risk stigmata 30, 47. In doing so, they added stronger study‐supported evidence to the literature that patients with bleeding ulcers and low‐risk stigmata can be treated effectively with oral omeprazole. The present meta‐analysis showed that, compared with intravenous PPIs, the oral route of administration of PPIs is less challenging to implement, does not require frequent monitoring for infusion site reactions and can be more economical, as demonstrated by the lower cost of the oral administration method and the earlier discharge of patients receiving oral PPIs reported in the study by Yen et al. 34.

There were several limitations to the present study. First, some of the included RCTs were not designed as equivalence or non‐inferiority studies, which in retrospect would have been a more favourable study design. Second, the difficulty of planned patient enrolment led some studies to stop prematurely. Third, a large number of patients dropped out owing to the strict exclusion criteria which were imposed on various factors, resulting in the fact that the predefined sample size in each study might not have been large enough for detecting small differences.

In conclusion, the present results suggest that oral PPIs are a feasible, safe and effective alternative to intravenous PPIs in terms of rebleeding, the need for emergency surgery, transfusion requirements, length of hospital stay and mortality in patients with bleeding peptic ulcers. Oral PPIs may be able to replace intravenous PPIs as the treatment of choice in peptic ulcer bleeding.

Competing Interests

All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: YZ had support from 2015 Guangdong Medical Science and Technology Research Fund (A2015476) and 2015 Guangdong General Hospital, Guangdong Academy of Medical Sciences Doctoral Fund for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

Supporting information

S1 PRISMA 2009 Checklist

Supporting info item

Jian, Z. , Li, H. , Race, N. S. , Ma, T. , Jin, H. , and Yin, Z. (2016) Is the era of intravenous proton pump inhibitors coming to an end in patients with bleeding peptic ulcers? Meta‐analysis of the published literature. Br J Clin Pharmacol, 82: 880–889. doi: 10.1111/bcp.12866.