Which of Vonoprazan Alone or Intravenous Proton Pump Inhibitor Followed by Vonoprazan Is Optimal for Reducing Delayed Bleeding in Gastric Endoscopic Submucosal Dissection?

To date, endoscopic submucosal dissection (ESD) is accepted as a minimally invasive treatment procedure for early gastric cancer [1‒3]. Delayed bleeding is one of the major adverse events, accounting for 4.1%–5.9% in gastric ESD [4‒7]. Several methods, including polyglycolic acid sheet [8], closure with clips [9], and second-look endoscopy [10], were proposed to prevent delayed bleeding. However, these methods showed no significant effect in reducing delayed bleeding [11]. Meanwhile, acid-suppressive agents and discontinuation of antithrombotic (AT) agents have reduced effects of delayed bleeding [12, 13]. However, discontinuation of AT agents may increase the risk of thromboembolism and recent Japanese guidelines recommend continuing AT agents as much as possible when performing endoscopic procedures [14]. Therefore, acid-suppressive agent administration is the cornerstone for reducing delayed bleeding.

Among acid-suppressive agents, proton pump inhibitors (PPIs) are known to have reduced effects of delayed bleeding, compared with histamine-2 receptor antagonists, in gastric ESD [12]. Recently, vonoprazan is developed as a new oral potassium-competitive acid blocker, which has more effective acid suppression than PPI [15]. Our prior database study demonstrated that vonoprazan had a lower delayed bleeding risk than oral PPI in ESD for gastric tumors [16]. Another database study showed that intravenous PPI followed by vonoprazan had a lower bleeding risk than intravenous PPI followed by oral PPI, in gastric ESD [17]. In clinical practice, some clinicians use only oral acid-suppressive agents, whereas others use intravenous PPIs during the fasting period following gastric ESD [18, 19]. Based on these findings, the following question can be raised: which of vonoprazan alone or intravenous PPI followed by vonoprazan is an optimal acid-suppressive method for reducing delayed bleeding in gastric ESD? This study aimed to compare the delayed bleeding risk of vonoprazan with that of intravenous PPI followed by vonoprazan in gastric ESD.

This is a retrospective population-based study based on data retrieved from the Diagnosis Procedure Combination (DPC) database. The DPC database used in our study group includes over 1,000 acute care hospitals throughout Japan, covering approximately 90% of all tertiary hospitals and 50% of all acute care hospitalizations (7 million per year) [20, 21]. The database includes the following data in each patient: demographic information, age, sex, height, weight, drug use, diagnosis, past medical history, the day of admission and discharge, the day of treatment, endoscopic procedure, and complications during hospitalization, which were coded with the International Classification of Diseases and Related Health Problems, 10th Revision codes [22]. These datasets were linked using unique encoded identifiers. Patients’ hospital, procedures, and medications were coded by Japanese text data with the Medical Intervention Classification master code [23]. To protect patients’ confidentiality, all records were anonymized.

We initially identified adult patients (≥20 years) who were admitted to undergo gastric ESD between April 2014 and March 2021 (each indexed by the Japanese code) and who were administered intravenous PPIs and/or vonoprazan. In this study, gastric ESD was defined as gastroduodenal ESD for gastric tumors. Intravenous PPI administration is generally limited to 2 days (days 0–1) following gastric ESD [17, 24]; therefore, patients were excluded from this study (1) when intravenous PPI administration was on day 2 or later, except cases with intravenous PPI administration following delayed bleeding. Furthermore, to compare the delayed bleeding risk between vonoprazan alone and intravenous PPI followed by vonoprazan in gastric ESD, the following patients were excluded: (2) those with intravenous PPI administration but no vonoprazan prescription, (3) those in whom neither vonoprazan nor intravenous PPI was used before delayed bleeding, (4) those who were prescribed or injected histamine-2 receptor antagonists before delayed bleeding, and (5) those with missing data.

Patients were divided into the vonoprazan alone and intravenous PPI groups. Only vonoprazan was prescribed in the vonoprazan alone group, whereas intravenous PPI injection for 2 days (days 0–1) was followed by vonoprazan prescription in the intravenous PPI group. The standard dose of vonoprazan is 20 mg, which was set as the vonoprazan cutoff value; thus, vonoprazan ≥20 mg was regarded as the standard/high dose. In Japan, the standard doses of intravenous PPIs available are omeprazole at 20 mg and lansoprazole at 30 mg twice daily.

We collected data on age, sex, BMI, comorbidities, hemodialysis, concurrent medications, annual hospital volume, treatment year, and bleeding outcome. Comorbidity was measured using the Charlson comorbidity index [25]. Date during the hospitalization and, if any, those during rehospitalization until 30 days after ESD were collected. Annual hospital volume was categorized into the following quartiles according to the number of upper gastrointestinal treatment cases: low (<38 cases/year), intermediate (38–63 cases/year), high (64–113 cases/year), and very high (>113 cases/year). For medication, we accessed drugs that have the potential to be associated with bleeding. Antiplatelet agents (aspirin, P2Y12 receptor antagonist [P2Y12RA], cilostazol, and other antiplatelet agents), anticoagulants (warfarin, direct oral anticoagulant [DOAC], heparin, and other anticoagulants), nonsteroidal anti-inflammatory drugs, corticosteroids, and mucosal protective agents were accessed. Furthermore, we assessed data on hemodialysis because it has been reported to be one of the risk factors for delayed bleeding in gastric ESD [13]. Treatment year was categorized into quartiles (2014–2015, 2016–2017, 2018–2019, and 2020).

The primary outcome was delayed bleeding. According to previous studies [16, 26], delayed bleeding was defined as bleeding requiring endoscopic hemostasis and/or blood transfusion at ≥2 days following the treatment. First, a comparison of delayed bleeding risk between the vonoprazan alone and intravenous PPI groups was performed in all vonoprazan doses. Subsequently, the bleeding risk was compared when vonoprazan was restricted to the standard/high dose.

Propensity score matching (PSM) was performed to reduce bias by equating the two comparing groups based on the following 18 variables: age, sex, BMI, comorbidities, concurrent medications (aspirin, cilostazol, P2Y12RA, other antiplatelet agents, warfarin, DOAC, heparin, other anticoagulants, nonsteroidal anti-inflammatory drugs, mucosal protective agents, and corticosteroids), annual hospital volume, and treatment year. We used the nearest-neighbor 1:1 matching method with a 0.2 caliper width of the pooled standard deviation of the logit of the propensity score. Moreover, to assess the balance of covariates after matching, standardized differences (SDs) were calculated. SD ≤10% indicates good balance [27]. Following PSM, logistic regression analysis was used to compare bleeding outcomes. Imbalanced variables with SD >10% following PSM were adjusted for in the regression model [27]. A p < 0.05 was considered to be statistically significant. All statistical data were analyzed using Statistical Package for the Social Sciences (version 25, IBM, Armonk, NY, USA) and R version 3.6.1 for Windows software (The R Foundation for Statistical Computing, Vienna, Austria).

Sensitivity and subgroup analyses were performed to confirm the robustness of obtained results. First, to confirm the results of PSM analyses, inverse probability of treatment weighting (IPTW) analyses [28] were performed using the same variables. Second, subgroup analysis was conducted depending on age, sex, hemodialysis, and AT agents.

As a post hoc analysis to confirm the results of the subgroup analysis for the status of AT agents, we conducted additional PSMs to compare the delayed bleeding risk of the vonoprazan alone group with that of the intravenous PPI group in several relevant subgroups. First, we performed PSM in patients restricted to the use of AT agents. Furthermore, since aspirin, P2Y12RA, warfarin, and DOAC are known to have different risks for delayed bleeding [13], we conducted 4 additional PSMs in patients restricted to each agent separately. Within each subgroup, the propensity score was reestimated and patients were rematched on the newly estimated score using the nearest-neighbor 1:1 matching method with a 0.2 caliper width of the pooled standard deviation of the logit of the propensity score.

A total of 63,952 eligible patients who underwent gastric ESD were enrolled in this study with 27,761 and 36,191 patients in the vonoprazan alone and intravenous PPI groups, respectively (Fig. 1). Following PSM, 49,420 patients were analyzed with 24,710 patients each for the two groups. The baseline characteristics before and after PSM are shown in Table 1. All SDs were <10%, which showed good balance between the 2 groups.

Following PSM, delayed bleeding was observed in 5.9% (1,466 patients) and 5.9% (1,465 patients) in the vonoprazan alone and intravenous PPI groups, respectively (Fig. 2). The vonoprazan alone group had similar delayed bleeding risk to the intravenous PPI group (odds ratio [OR], 1.00; 95% confidence interval [CI], 0.93–1.08) (Table 2). In the IPTW analysis, a similar result was obtained (OR, 1.03; 95% CI, 0.96–1.10) (Table 2).

When vonoprazan was restricted to the standard/high dose (i.e., exclusion of patients with vonoprazan of <20 mg in daily dose), 24,105 patients each in the two groups were analyzed after PSM. The background factors were well balanced following PSM with a maximum SD of 4.5%. Similar to the results in the analysis of vonoprazan in all dose, delayed bleeding risk in the vonoprazan alone group showed no significant results (OR, 1.02; 95% CI, 0.95–1.11) in the analysis of vonoprazan restricted to the standard/high dose (Table 3). Additionally, IPTW analysis showed a similar result (OR, 1.03; 95% CI, 0.96–1.11).

In the subgroup analysis, the similarity of the vonoprazan alone group to the intravenous PPI group for delayed bleeding risk was not modified by sex, age, and hemodialysis; however, such similarity was modified by the status of AT agents (p for interaction = 0.029) (Table 4). Specifically, the delayed bleeding risk in the vonoprazan alone group was higher in patients with AT agents in contrast to no significant results in those without AT agents.

Based on the significant result in the subgroup analysis for the status of AT agents, we performed additional PSMs. In the analysis restricted to patients with AT agents, the vonoprazan alone group showed significantly higher delayed bleeding risk than the intravenous PPI group (OR, 1.15; 95% CI, 1.01–1.31) (Table 5). In the analyses restricted to those with each of AT agents, no significant results were acquired; however, their ORs were 1.11–1.24 (Table 5).

Vonoprazan has strong and sustained acid-inhibitory activity, and this drug is known to have a lower risk of delayed bleeding than oral PPI in gastric ESD [16]. However, no studies have compared the delayed bleeding risk between vonoprazan alone and intravenous PPI followed by vonoprazan in gastric ESD, despite that this issue has been paid more attention to in recent years [29]. Furthermore, a large number of cases may be required for reliable results because, even if there is a difference in the delayed bleeding risk between the 2 methods, the difference may be small based on a prior result in the comparison between vonoprazan and oral PPI in gastric ESD [16]. Therefore, we conducted this large-scale database study.

This study first revealed that the delayed bleeding risk of vonoprazan alone was similar to that of intravenous PPI followed by vonoprazan in gastric ESD. This result was consistent when the vonoprazan dose was restricted to the standard/high dose. Since vonoprazan can be orally administered, the use of vonoprazan or intravenous PPI in the fasting period following gastric ESD depended on the institutions [18, 19]. However, considering the higher cost of intravenous PPI than that of vonoprazan ($3.2–$5.8 vs. $1.85 in the standard daily dose in Japan), the use of vonoprazan including the fasting period following gastric ESD might be preferred in daily clinical practice.

Conversely, intravenous PPI followed by vonoprazan had a lower delayed bleeding risk than vonoprazan alone in gastric ESD when AT agents were administered. This association was almost consistent in the analysis of each AT agent. However, it should be noted that although the OR was statistically significant, it was relatively low (1.15). Therefore, whether this difference is clinically significant may be debatable. Nevertheless, intravenous PPI followed by vonoprazan might be favorable for reducing delayed bleeding in patients with AT agents.

According to the in vitro data [30, 31], an intragastric pH of >6 is required for promoting clot formation and stability. When lansoprazole 30 mg was intravenously administered twice daily to Chinese healthy subjects with Helicobacter pylori-negative, the least-squares mean percentages of time with the intragastric pH of >6 over the 24-h period were 63.03% and 77.58% on days 1 and 5, respectively [32]. Although no studies evaluated the percentage of time with the intragastric pH of >6 in taking vonoprazan, the mean 24-h pH of >5 holding time ratios in Japanese healthy subjects taking vonoprazan at 20 mg/day was 53.5% and 73.2% on days 1 and 7, respectively [33]. Although a direct comparison of time with the intragastric pH of >6 between vonoprazan and intravenous PPI has not been performed, the results in previous studies suggest that intravenous PPI may have a longer time with a pH of >6 than vonoprazan. In the present study, it is unclear why the vonoprazan alone and intravenous PPI groups had similar risks of delayed bleeding in patients without AT agents despite that the vonoprazan alone group had a higher risk of delayed bleeding than the intravenous PPI group in those with AT agents. One possible explanation is that, in those without AT agents, the acid-inhibitory effect of vonoprazan may be sufficient for reducing delayed bleeding in gastric ESD. However, since bleeding occurs more easily in those with AT agents, more strict acid inhibition may be demanded in such patients.

The strength in this study is the analysis of a numerous number of cases and events (63,952 patients with 2,540 bleeding events). Although concluding “similar” risks between the two groups are occasionally challenging, the OR value (1.00) with a narrow 95% CI, which was brought by the very large sample size, allowed us to conclude that the vonoprazan alone and intravenous PPI groups had similar delayed bleeding risks in gastric ESD.

However, this study also had several limitations. First, this was a retrospective study and was conducted without randomization. Thus, some unexamined confounders may exist. For example, potential confounders, including the number of tumors, tumor location and size, were not included; however, these lesion-related factors were reported to play a relatively small role in delayed bleeding in gastric ESD [13]. To confirm our results, a randomized controlled trial is desirable, although a very large number of cases may be required. Second, we used the initially prescribed vonoprazan dose. In this study, dose changes were not considered. Third, there is a possibility of misclassification bias in this study. Fourth, this study could not investigate the continuation or interruption of AT agents in gastric ESD. Fourth, some patients may have been unable to take vonoprazan for various reasons, including the residual effects of deep sedation and nausea. Lastly, some endoscopic methods, including the closure of ulcers and polyglycolic acid sheets, could not be investigated; however, these methods did not have preventive effects for delayed bleeding in gastric ESD [34, 35].

In conclusion, this large-scale population-based study first reported that vonoprazan alone and intravenous PPI followed by vonoprazan had similar delayed bleeding risks in gastric ESD. However, in patients with AT agents, vonoprazan alone had a higher risk of delayed bleeding than intravenous PPI followed by vonoprazan. Although both methods might be acceptable in gastric ESD when AT agents were not administered, the administration of only vonoprazan might be preferred in the aspect of the cost. Conversely, intravenous PPI followed by vonoprazan might be favorable in patients with AT agents.

The Ethics Committee of Tohoku University Graduate School of Medicine provided ethics approval for this study. Informed consent was waived .This study protocol was reviewed, and the need for written and informed consent was waived, due to the anonymity of data, by the Ethics Committee of Tohoku University Graduate School of Medicine, date of decision on February 14, 2024.

Waku Hatta was a member of the journal’s Editorial Board at the time of submission. The other authors have no conflicts of interest to declare.

This work was supported by the Ministry of Health, Labor, and Welfare (Grant No. 22AA2003).

Conceptualization and methodology: W.H.; collection of data: H.A. and K.T.; formal analysis, investigation, and writing – original draft preparation: H.A. and W.H.; writing – review and editing: K.T., N.T., Y.H., Y.O., M.S., X.J., T.K., A.I., S.H., K.F., and A.M.; funding acquisition: K.F.; supervision: K.F., K.F., and A.M.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.