Scandinavian Journal of Gastroenterology

ISSN: 0036-5521 (Print) 1502-7708 (Online) Journal homepage: https://www.tandfonline.com/loi/igas20

Short-term efficacy of potassium-competitive acid

blocker following gastric endoscopic submucosal
dissection: a propensity score analysis

Yohei Horikawa, Hiroya Mizutamari, Nobuya Mimori, Yuhei Kato, Saki

Fushimi, Sayaka Sato & Syunji Okubo

To cite this article: Yohei Horikawa, Hiroya Mizutamari, Nobuya Mimori, Yuhei Kato,
Saki Fushimi, Sayaka Sato & Syunji Okubo (2018) Short-term efficacy of potassium-
competitive acid blocker following gastric endoscopic submucosal dissection: a
propensity score analysis, Scandinavian Journal of Gastroenterology, 53:2, 243-251, DOI:
10.1080/00365521.2017.1410569

To link to this article: https://doi.org/10.1080/00365521.2017.1410569

Published online: 07 Dec 2017.

Submit your article to this journal

Article views: 451

View related articles

View Crossmark data

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=igas20
SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY, 2018
VOL. 53, NO. 2, 243–251
https://doi.org/10.1080/00365521.2017.1410569

ORIGINAL ARTICLE

Short-term efficacy of potassium-competitive acid blocker following gastric

endoscopic submucosal dissection: a propensity score analysis
Yohei Horikawa, Hiroya Mizutamari, Nobuya Mimori, Yuhei Kato, Saki Fushimi, Sayaka Sato and Syunji Okubo
Department of Gastroenterology, Hiraka General Hospital, Yokote Akita, Japan

ABSTRACT ARTICLE HISTORY

Background: Endoscopic submucosal dissection (ESD) is a promising method for the resection of Received 11 October 2017
superficial gastric neoplasms. Vonoprazan is a novel potassium-competitive acid blocker (P-CAB) that is Revised 21 November 2017
currently considered as a potential alternative to proton pump inhibitors (PPIs) for the treatment of Accepted 22 November 2017
acid-related diseases. However, the utility of vonoprazan in ESD-related ulcers is unclear. Therefore, we
compared the short-term efficacies of vonoprazan and the PPI lansoprazole in ESD-related ulcer healing KEYWORDS
during the first two weeks following intervention. Vonoprazan; short-term
Methods: This study included 115 superficial gastric neoplasms that were treated by ESD at Hiraka ulcer healing; granulation
General Hospital between April 2015 and January 2017. Patients treated with P-CAB (20 mg vonopra- tissue; gastric endoscopic
zan, n ¼ 62) or PPI (30 mg lansoprazole, n ¼ 53) were compared using propensity-score matching ana- submucosal dissection
lysis. Primary outcome was rate of ulcer reduction at two weeks after ESD. Secondary outcomes were
coverage ratio of ulcer base by granulation tissue and incidence of postoperative bleeding.
Results: The rate of ulcer reduction was significantly higher (median [range], 80.6% [67.6%–94.5%] vs.
62.7% [33.4%–85.2%]; p < .0001) and coverage ratio of the ulcer base by granulation tissue was
significantly accelerated (median [range], 84.1% [67.7%–95.3%] vs. 61.9% [12.1%–90.1%]; P < 0.0001) in
the P-CAB group compared with the PPI group. Postoperative bleeding was not observed in either
group.
Conclusions: P-CAB achieved rapid artificial ulcer healing with promotion of granulation tissue forma-
tion. However, conventional PPI with initial intravenous infusion might be sufficient for prevention of
postoperative bleeding following gastric ESD.

Introduction One distinct clinical consideration in treatment of artificial

Endoscopic submucosal dissection (ESD) is a promising
approach for the resection of superficial gastric neoplasms
[1–4]. The most serious adverse event in ESD is postoperative
bleeding that can occur within the first two weeks [5–7], for
which there are currently two established methods: post-ESD
coagulation of visible vessels (PEC) [8] and treatment with
acid secretion inhibitors following ESD [9]. PEC is widely
accepted and effectively performed. In addition, artificial
ulcers induced by gastric ESD are typically treated with pro-
ton pump inhibitors (PPIs) for 4–8 weeks based on the peptic
ulcer therapy protocol. However, the incidence of postopera-
tive bleeding with ESD remains at approximately 4.5% [10].
Vonoprazan is a novel, orally active potassium-competitive
acid blocker (P-CAB) that is considered as a potential alterna-
tive to PPIs for the treatment of acid-related diseases [11].
While the utility of vonoprazan in ESD-related ulcers is not
known, the more rapid, potent, and sustained anti-secretory
effect achieved with vonoprazan can be expected to sup-
press acid secretion and accelerate ulcer healing during the
first few weeks after ESD, compared with the commonly
used PPIs.
Numerous studies have investigated approaches to accel-
erate the healing of gastric ESD-derived ulcers [12–18].
ulcers is the prevention of postoperative bleeding.
Assessment of the extent of the endoscopic ulcer healing
might be a surrogate marker of postoperative bleeding.
However, whether accelerated ulcer healing might contribute
to the prevention of this postoperative bleeding during the
first few weeks after ESD remains unclear.
We therefore conducted a study to determine whether
vonoprazan promoted healing of ESD-related gastric ulcers
and prevented postoperative bleeding compared with
the available PPI lansoprazole during the first two weeks fol-
lowing intervention, using propensity score matching (PSM)
analysis [19].
Methods
Ethics
This study was performed in accordance with the Declaration
of Helsinki and was approved by the ethics committee of
Hiraka General Hospital. Written informed consent was
obtained from all patients or their families. The protocol for
this study was registered with the University Hospital Medical

CONTACT Yohei Horikawa horikawa_01@me.com Hiraka General Hospital, 3-1 Yatukuchi, Maego, Yokote City, Akita 013-0013, Japan
ß 2017 Informa UK Limited, trading as Taylor & Francis Group
244 Y. HORIKAWA ET AL.
Information Network (UMIN) Clinical Trials Registry (No. UMIN
000026391).
Participants
This study included a total of 118 consecutive patients with
125 superficial gastric neoplasms who were treated by ESD
at Hiraka General Hospital in Yokote, Japan between April
2015 and January 2017. Data were retrospectively reviewed
from the hospital database. The flow chart of patients
enrolled in this study is shown in Figure 1. The inclusion cri-
terion was cancers fulfilling the absolute and the expanded
criteria [20], which were obtained using endoscopic methods.
After excluding patients with a history of prior surgery
(n ¼ 2), a total of 114 patients with 121 lesions were included
in the study. All patients were intravenously infused with
omeprazole (20 mg b.i.d.) for the first two days. Thereafter,
patients were alternately assigned to treatment with 20 mg
vonoprazan (P-CAB group) or 30 mg lansoprazole (PPI group)
following the ESD procedure by endoscopy personnel, and
treatments were orally administered for 12 days. Patients
who had lesions with peptic ulcer scars (n ¼ 3), lesions with a
steroid injection (n ¼ 2), and postoperative bleeding on post-
ESD day one (n ¼ 1) in the PPI group were excluded.
Therefore, a total of 62 lesions treated by vonoprazan were
assessed and compared with 53 lesions treated by lansopra-
zole in the final analysis. Helicobacter pylori (H. pylori)
Figure 1. Flow chart of the study.
infection was assessed in all patients by at least one of the
following three methods: (1) the anti-Hp immunoglobulin G
serological test, (2) the rapid urease test, or (3) the 13C-urea
breath test. Atrophic gastritis patterns were evaluated using
the Kimura and Takemoto classification [21]. A follow-up
endoscopy was performed two weeks after ESD.
Study outcomes
The primary outcome of the present study was the rate of
artificial ulcer reduction at two weeks after ESD. The rate was
determined using the following formula: (1  ulcer area on
post-ESD day 14/ulcer area on post-ESD day 0)  100 (%).
The area of ulceration was calculated by multiplying the two
diameters (major axis  minor axis) (mm2), which were deter-
mined with a bendable endoscopic measuring device (M2-3;
Olympus, Tokyo, Japan). Secondary outcomes were coverage
ratio of the ulcer base by granulation tissue at two weeks
after ESD and incidence of postoperative bleeding. Coverage
ratio of the ulcer base area by granulation tissue was calcu-
lated using the following formula: (granulation tissue area on
post-ESD day 14/area of defective mucosa on post-ESD day
14)  100 (%). The granulation tissue was originally approxi-
mated based on the presence of a reddish cobblestone-like
appearance (Figure 2(A)); the area was then using the color
range detection function of the Photoshop CCV 2015 soft-
R
ware (Adobe) using images on a computer, which was
 SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY 245

confirmed by histological examination of the biopsy speci-
men from edge of ulcer base in some cases prior to the
study (Figure 2(B)). The relationship between the two param-
eters was evaluated by Spearman’s correlation analysis. All
data were collected from the images obtained by high-defin-
ition endoscopy. Three endoscopists certified by the
Japanese Gastrointestinal Endoscopy Society blinded to the
study drugs independently performed these assessments,
and the averages of the measurements were used in analy-
ses. Postoperative bleeding was defined as more than 2 g/dL
decrease in blood hemoglobin levels or clinical evidence of
bleeding due to ESD, in combined with the occurrence of
hematemesis and melena, presence of bleeding on
endoscopy, or a combination of unstable vital signs within
the first two weeks after the ESD procedure.
ESD
Low-dose aspirin (LDA) or non-steroidal anti-inflammatory
drugs (NSAIDs) were continued throughout the study period.
ESD techniques were described in detail elsewhere [4]. In brief,
patients were sedated with intravenous dexmedetomidine.
The resection areas were identified by magnified endoscopy,
and the normal mucosa surrounding the region (with a 5-mm
boundary) was marked by using a Flush knife BT (DK2618N;

Figure 2. (A) Endoscopic images at two weeks after gastric endoscopic submucosal dissection. Accelerated healing (a, b): Ulcer base is nearly completely covered
by granulation tissue, which can be observed as a cobblestone-like appearance (coverage ratio; 95.34% [a]; 91.01% [b]). Delayed healing (c, d): Ulcer base is covered
by a white coating and immature granulation tissue can be identified (coverage ratio; 40.53% [c]; 12.01% [d]). (B) Histological findings of the ulcer base at two
weeks after gastric endoscopic submucosal dissection. Accelerated healing (a): Microscopic view of the biopsied specimen shows the vascular-rich granulation tissue
with regenerative mucosa (hematoxylin & eosin staining, 20, biopsied from Figure 2(A)-a). Red arrows indicate the regenerative epithelium. Delayed healing (b):
Microscopic view shows debris with minimal granulation formation (hematoxylin & eosin staining, 20, biopsied from Figure 2(A)-c).
246 Y. HORIKAWA ET AL.
FUJIFILM Medical Co., ltd., Tokyo, Japan). A normal saline solu-
tion was then injected into the submucosal layer, just periph-
eral to the markings. A circumferential incision was made
followed by the dissection of the remaining submucosal layer
until the tumor was completely excised. The careful vessel
handling was performed during the ESD procedure using pre-
cut coagulation. The procedure time was recorded from the
time of marking until the end of tumor removal. Major bleed-
ing was defined as bleeding requiring a change to hemostatic
forceps (Coagrasper; FD-411QR, Olympus, Japan) for complete
hemostasis [22]. Endoscopy personnel counted the frequency
of major bleeding during ESD. After the procedure, the artifi-
cial ulcer base was carefully identified and exposed vessels
were coagulated, despite bleeding, with hemostatic forceps
[8]. Resected specimens were classified histopathologically
according to the Japanese Gastric Cancer Association criteria
[23] and evaluated for curability.
Follow-up
A second-look endoscopy was performed to identify and cau-
terize any visible vessels at the ulcer base and edges only in
patients who were at high risk for bleeding (i.e., extra-large
ulcer (>70 mm diameter) and antithrombotic therapy), using
a Coagrasper on post-ESD day 1. Patients who did not
experience any complications resumed food intake on post-
ESD day 2. All patients were discharged within seven days
following ESD. After a follow-up endoscopy at two weeks,
the selection of anti-secretory drugs (P-CAB, PPIs or
H2-Receptor antagonists and doses) additional two weeks
was entrusted with performed endoscopists.
Propensity-score matching
In this study, propensity-score matching was used to minim-
ize potentially confounding factors and selection biases and
to identify controls within the study-patient group. Among
the numerous potential predictive factors for delayed healing
in ESD that were previously reported are lesions associated
with severe atrophic gastritis [12], LDA/NSAIDs administration
[24], diabetes mellitus [13], lesion location in the middle third
of stomach [13], lesion circumference in lesser curvature [14],
initial ulcer size >40 mm [6,13,15], hemostasis with electro-
coagulation [13,16], and preoperative submucosal fibrosis
[17]. Therefore, nine possible confounders were used as
matching factors in the current study, based on previous
reports. These were age (years), gender (male/female), pat-
tern of atrophic gastritis (open type/others), LDA/NSAID
administration, diabetes mellitus, lesion location (middle/
upper third of stomach or lower third of stomach), lesion cir-
cumference (lesser curvature/others), initial ulcer size, use of
coagulation for major bleeding (0/others) were included as
independent variables in multivariable logistic regression
analysis, with the P-CAB therapy included as the dependent
variable. This model yielded an area under the receiver oper-
ating characteristic curve (AUC) of 0.74, which indicated a
slight predictive power. The propensity score for the P-CAB
therapy was calculated using a logistic regression analysis.
After the estimation of propensity scores, the patients who
received the PPI therapy were matched to those who
received the P-CAB therapy with a 1:1 optimal match using a
caliper width equal to one-fifth logit of the standard devi-
ation of the propensity score without replacement. The
standardized difference was used to measure covariate bal-
ance, whereby an absolute standardized difference above
10% represented meaningful imbalance.
Statistical analysis
Statistical analysis of the clinical and endoscopic data was
performed using the chi-square test or Fisher’s exact test for
categorical data and Student’s t test for numerical data for
univariate analysis, both before and after PSM. Multivariable
logistic regression analysis was used to determine the pro-
pensity score and multivariate analysis was used to deter-
mine factors associated with accelerated ulcer healing. The
relationship between the rate of ulcer reduction and cover-
age ratio of the ulcer base by granulation tissue was eval-
uated by Spearman’s correlation analysis using Student’s t
distribution. Both absolute differences and p values were
determined. Statistical significance was defined as p < .05.
All statistical analyses were performed using JMPV version
R
12.0 (SAS Institute, Cary, NC).
Results
Baseline characteristics and treatment outcomes
before PSM
The baseline characteristics for all patients before PSM are
shown in Table 1. Among the patient characteristics (age,
gender, H. pylori status, pattern of atrophic gastritis, LDA/
NSAID administration, antithrombotic therapy, oral steroid
therapy, comorbidities) and the lesion characteristics (lesion
location, lesion circumference, tumor depth, macroscopic
type, procedure time, initial ulcer size, histology, frequency of
coagulation for major bleeding, second-look endoscopy) did
not significantly differ between the two groups. Among the
study outcomes, the P-CAB group experienced a significant
reduction in the ulcer size at two weeks after ESD compared
with the PPI group (mean [SD], 80.97% [7.49%] vs. 57.63%
[19.29%], p < .0001; median [range], 81.4% [63.1%–97.0%] vs.
61.6% [7.8%–87.0%], p < .0001). In addition, coverage ratio of
the ulcer base by granulation tissue at two weeks after ESD
was also significantly accelerated in the P-CAB group com-
pared with the PPI group (mean [SD], 81.94% [8.61%] vs.
61.39% [16.51%], p < .0001; median [range], 83.4%
[50.3%–95.3%] vs. 65.3% [12.1%–90.2%], p < .0001). None of
the patients in the study cohort suffered adverse consequen-
ces, namely postoperative bleeding (Table 2).
Treatment outcomes in the P-CAB and PPI groups
after PSM
The matched variables in both groups after PSM are shown
in Table 3. Following PSM, the outcomes were compared
among 32 matched pairs (Table 4). The propensity score
 SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY 247

Table 1. Baseline characteristics of the 115 superficial gastric neoplasms that underwent ESD.
P-CAB group PPI group
(n ¼ 62) (n ¼ 53) p
Patient characteristics
Age, median [range] (y) 69.5 [47.0–84.0] 73.0 [60.0–86.0] .101
Gender
Male / Female 44/18 34/19 .483
H.pylori status
Positive/negative/post eradication 48/3/11 42/3/8 .918
Patterns of atrophic gastritis
Negative/closed type/open type 3/23/36 3/21/29 .932
LDA/NSAIDs administration
Positive/negative 7/55 9/44 .381
Antithrombotic therapy
Positive/negative 9/53 7/46 .401
Anticoagulant therapy 2/9 1/7 .221
Oral steroid therapy
Positive/negative 0/62 0/53 1.000
Co-morbidity (positive/negative)
Diabetes mellitus 17/45 9/44 .179
Chronic kidney disease (eGFR <30) 1/61 0/53 .265
Liver cirrhosis 0/62 0/53 1.000
Lesion characteristics
Lesion location
U/M/L 12/24/26 15/20/18 .483
Lesion circumference
A/P/G/L 10/14/14/24 4/13/8/28 .267
Tumor depth
M/SM 55/7 46/7 .754
Macroscopic type
Flat/Depressed 19/43 23/30 .215
Procedure time
Median [range] (min.) 69.0 [22.0–260.0] 60.0 [22.0–240.0] .855
Initial ulcer size
Median [range] (mm2) 1703.0 [609.0–6570.0] 1400.0 [450.0–4760.0] .234
Histology
Differentiated/pooly differentiated 51/11 49/4 .212
Coagulations for major bleeding during ESD
0 / others (times) 28/34 30/23 .221
Second-look endoscopy
Positive/negative 9/53 10/43 .332
Hemostasis positive/negative 8/9 10/10 .155
LDA: low dose aspirin; NSAIDs: Non-steroidal anti-inflammatory drugs; H.pylori: Helicobacter pylori; U: upper third;
M: middle third; L: lower third of the stomach; A: anterior wall; P: posterior wall; G: greater curvature; L: lower curvature;
M: mucosal cancer; SM: submucosal cancer.

Table 2. Outcomes of the 115 superficial gastric neoplasms that underwent ESD.
P-CAB group PPI group
(n ¼ 62) (n ¼ 53) p
The rate of ulcer reduction
Mean [SD] (%) 80.97 [7.49] 57.63 [19.29] <.0001
Median [range] (%) 81.4 [63.1–97.0] 61.6 [7.8–87.0]
Coverage ratio of ulcer base
Mean [SD] (%) 81.94 [8.61] 61.39 [16.51] <.0001
Median [range] (%) 83.4 [50.3–95.3] 65.3 [12.1–90.2]
Adverse events (positive/negative)
Postoperative bleeding (%) 0/62 (0) 0/53 (0) 1.000
SD: standard deviation.

model was well calibrated (AUC ¼0.74) and well matched
(caliper size: 0.089; standardized difference <0.1) between
the lesions of the P-CAB and the PPI groups. Similar to those
observed before matching, the P-CAB group experienced a
significant reduction in the ulcer size at two weeks after
ESD (mean [SD], 80.69% [7.46%] vs. 59.70% [17.64%],
p < .0001; median [range], 80.6% [67.6%–94.5%] vs. 62.7%
[33.4%–85.2%], p < .0001). In addition, coverage ratio of the
ulcer base by granulation tissue at two weeks after ESD was
also significantly accelerated in the P-CAB group (mean [SD],
83.82% [6.88%] vs. 60.13% [17.92%], p < .0001; median
[range], 84.1% [67.7%–95.3%] vs. 61.9% [12.1%–90.1%],
p < .0001). Postoperative bleeding was not observed in either
of the groups.
The relationship between the rate of ulcer reduction
and coverage ratio of the ulcer base
Figure 3 shows the scatter diagram of the overall relation-
ship between the ratio of ulcer reduction ratio and the
coverage ratio of ulcer base at two weeks after ESD.
The two parameters were closely correlated (Spearman’s
248 Y. HORIKAWA ET AL.

Table 3. Matching variables for propensity score matching.

P-CAB group PPI group
(n ¼ 32) (n ¼ 32) p
Matching variables
Age, median [range] (y) 71.0 [47.0–84.0] 71.5 [60.0–82.0] .101
Gender
Male/Female 25/7 26/6 .756
Patterns of atrophic gastritis
Open type/others 18/14 17/15 .802
LDA/NSAIDs administration
Positive/negative 4/28 4/28 1.000
Diabetes mellitus
Positive/negative 4/28 5/27 .719
Lesion location
M/U or L 11/21 10/22 .790
Lesion circumference
Lesser curvature / others 15/17 16/16 .803
Initial ulcer size
2
Median [range] (mm ) 1587.5 [726.0–3060.0] 1410.0 [527.0–4760.0] .234
Coagulations for major bleeding during ESD
0 / others (times) 16/16 17/15 .803
LDA: low dose aspirin; NSAIDs: Non-steroidal anti-inflammatory drugs; U: upper third; M: middle third; L: lower third of the stom-
ach; A: anterior wall; P: posterior wall; G: greater curvature; L: lower curvature.

Table 4. Outcomes after propensity score matching.

P-CAB group PPI group
(n ¼ 32) (n ¼ 32) p
The rate of ulcer reduction
Mean [SD] (%) 80.69 [7.46] 59.70 [17.64] <.0001
Median [range] (%) 80.6 [67.6–94.5] 62.7 [33.4–85.2]
Coverage ratio of ulcer base
Mean [SD] (%) 83.82 [6.88] 60.13 [17.92] <.0001
Median [range] (%) 84.1 [67.7–95.3] 61.9 [12.1–90.1]
Adverse events (positive/negative)
Postoperative bleeding (%) 0/32 (0) 0/32 (0) 1.000
SD: standard deviation.

Figure 3. The scatter diagram of the relationship between ulcer reduction ratio and ulcer base coverage ratio. Oval line shows the 95% probability ellipse. Two
parameters were closely correlated with each other. Spearman’s correlation coefficient (rS) ¼ 0.374, t ¼ 4.290 (df ¼113), p ¼ .000. The proton pump inhibitor (PPI)
group exhibited a significant variability in data and discrepancy between the two parameters. Furthermore, in the PPI group, there were significant outliers (n ¼ 6/
115: 5.2%) outside of the 95% confidence oval line.

correlation coefficient [Rs] ¼ 0.374, t ¼ 4.290 [df ¼ 113], two parameters. Furthermore, in the PPI group, there were
p ¼ 0.000). Some of the lesions in the PPI group indicated a significant outliers (n ¼ 6/115, 5.2%) outside of the 95%
marked variability in data and a discrepancy between the confidence oval line.
 SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY 249

Table 5. Multivariate analysis for ulcer healing accelerating factor.

Ulcer reduction ratio 76.0% Coverage ratio of ulcer base 74.6%
Variables Odds ratio 95% CI p Odds ratio 95% CI p
Age 1.00 .95–1.06 .921 .94 .88–1.00 .080
Gender (Male or Female) 1.63 .55–5.05 .379 1.05 .26–4.24 .950
Atrophic gastritis (others or open type) .47 .17–1.22 .124 .44 .14–1.33 .147
LDA/NSAIDs administration ( or þ) 1.09 .21–6.55 .917 3.05 .37–33.4 .308
Diabetes mellitus ( or þ) .60 .19–1.86 .375 1.56 .37–6.93 .545
Lesion location (others or M) 1.13 .42–3.04 .803 2.42 .75–8.30 .140
Lesion circumference (others or Lesser curvature) 1.89 .736–4.98 .185 .56 .17–1.72 .318
Initial ulcer size .70 .28–1.73 .438 .36 .11–1.12 .080
Coagulation (0 or not) 2.00 .75–5.72 .167 1.15 .37–3.63 .805
Drugs (P-CAB or PPI) 12.07 4.73–34.49 <.0001 41.80 12.7–177.0 <.0001
LDA: low dose aspirin; NSAIDs: Non-steroidal anti-inflammatory drugs; M: middle third of stomach; CI: confidence interval.

Factors associated with accelerated ulcer healing
The logistic regression analysis for ulcer reduction ratio
beyond the overall median of 76.0% or coverage ratio of ulcer
base beyond the overall median of 74.6% are indicated in
Table 5. The P-CAB therapy was the only factor strongly asso-
ciated with the ratio of ulcer reduction beyond 76.0% (odds
ratio, 12.07; 95% confidence interval, 4.73–34.49; p < .0001)
and coverage ratio of ulcer base beyond 74.6% (odds ratio,
41.80; 95% confidence interval, 12.7–177.0; p < .0001).
Discussion
The first step in ulcer healing is the formation of granulation
tissue with angiogenesis at the ulcer base, followed by epi-
thelial cell regeneration [25]. In the current study, we
assessed ESD-induced ulcer healing by determining the
reduction in ulcer size and the change in coverage of the
ulcer base by granulation tissue two weeks after the proced-
ure. It might be important for preventing the serious adverse
event that can occur within the first two weeks [5–7]. The
novel therapeutic agent vonoprazan demonstrated superior
efficacy in both outcomes, compared with the currently used
PPI lansoprazole, after propensity score matching using nine
possible confounders.
Based on recent reports that have raised concerns regard-
ing the efficacy of P-CAB–based H. pylori eradication
treatment [26,27], the major advantage of vonoprazan in
short-term artificial ulcer healing might be related to its abil-
ity to rapidly suppress gastric acid secretion [11]. Previous
studies revealed that there was a delay in sustained reduc-
tion in acid secretion with PPIs [28]. In contrast, vonoprazan
was reported to achieve steady-state acid levels by day one
[29]. In the current study, 20 mg omeprazole was adminis-
trated intravenously twice during the first two days, followed
by the oral delivery of the study drug from day three. This
change in the delivery of treatment from an intravenous to
an oral route might be responsible for a gap in acid suppres-
sion after day three. Therefore, the rapid onset of vonoprazan
might minimize decreases in gastric pH, which can translate
into shorter healing times.
In the current study, the reduction in ulcer size and the
healing of the ulcer base were closely correlated. Although
muscular contraction is the major healing mechanism in iat-
rogenic ulcers [5,13,18], ulcer base healing can be strongly
facilitated by anti-secretory agents [30]. The pharmacokinetics
and pharmacodynamics of PPIs are affected by cytochrome
p450 2C19 (CYP2C19) polymorphisms; in contrast, vonopra-
zan was shown to potently inhibit acid production independ-
ently of the CYP2C19 genotype [31]. We could not determine
the CYP2C19 genotype of the study patients due to the insti-
tutional limitations. However, the significant outliers in the
PPI group might indicate the presence of rapid metabolizers
among the subjects, which might contribute to the delayed
healing of artificial ulcers observed in the PPI group.
In the current study, vonoprazan exhibited a clear and
promising healing effect on post-ESD ulcers during the first
two weeks. However, there were no patients who experi-
enced postoperative bleeding, even in the PPI group with
comparably delayed healing. Possible explanations for this
prevention might be the careful vessel handling during
the ESD procedure [4,8] and effectiveness of hemostasis in
the second-look endoscopy at high risk patients of bleeding.
This finding suggests that there may not be a significant dif-
ference in the clinical outcomes associated with the use of
PPIs or P-CAB. Given that the aim of treatment with anti-
secretory drugs is the prevention of postoperative bleeding,
not rapid healing to achieve scarring, treatment by intraven-
ous PPIs plus conventional oral PPI administration might be
sufficient for clinical prophylaxis after ESD.
We performed a literature review to determine the rela-
tionship between ulcer healing and postoperative bleeding
(Table 6) [32–35]. Four studies investigated the efficacy of
vonoprazan in gastric ESD-derived ulcers. Although these
studies varied in the length of the assessment period, the
P-CAB therapy achieved comparable or superior efficacy in
ulcer healing, compared with the PPI therapy. In contrast,
studies on the efficacy of vonoprazan in postoperative bleed-
ing report conflicting findings. Only one study reported the
superiority of vonoprazan in prevention of postoperative
bleeding, compared with various PPIs at five weeks after ESD
[32]. However, the incidence of postoperative bleeding in the
PPI group was relatively high (10%) in that study. Since that
study used a 2:1 matched historical control cohort, these
results might reflect chronological biases due to the ESD
skills. Therefore, similar to the current study, previous studies
might have failed to prove the potential efficacy of vonopra-
zan in prevention of postoperative bleeding.
Omeprazole was used intravenously during the first two
days in the current study, which might have contributed to
 250 Y. HORIKAWA ET AL.

RCT: Randomized controlled trial; PSM: Propensity-score matching; VPZ: Vonoprazan; RPZ: Rabeprazole; EPZ: Esomeprazole; LPZ: Lansoprazole; PCAB: Potassium-competitive acid blocker; PPI: Proton-pump inhibitors; OMZ:
an increase in gastric pH and might have partially contrib-

1703.0 [609.0-6570.0] (PCAB)

uted to the observed rapid healing process. The ethical com-

1400.0 [450.0-4760.0](PPI)
mittee of our institution did not allow the study design to
This study
Retrospective (PSM)

59.7 ± 17.6% (PPI)

80.7 ± 7.3% (PCAB)
exclude the use of an intravenous PPI on fasting phase (POD

(for limited cases)

OMZ 40mg 2days

Included (13.9%)
62 (VPZ 20mg)
53 (LPZ 30mg)
0-1). However, a study using vonoprazan from the start
should be conducted for a strict comparison. Sakurai et al.

2 weeks
[29] reported that a gastric pH level above 4 was achieved at

POD 1
None

None
four hours after the first administration of vonoprazan.
Conversely, intravenous omeprazole was reported to increase
gastric pH to above 5 at eight hours after the first infusion
1256.0 [867.4-1628.9] (PCAB)
[36]. Further investigation is necessary to determine if vono-
1012.7 [588.8-1491.5] (PPI) prazan might be a potential alternative to intravenous PPIs.
This study has several limitations. First, this was a study
Tsuchiya et al.

1,2,4,6, and 8 weeks

2017 [35]

with a relatively small sample size, which was performed at a

OMZ 40mg 2days

Included (22.5%)
39 (VPZ 20mg)
41 (EPZ 20mg)

single center by only experienced endoscopists. However,

2.4% (PCAB) the consistent technical expertise and therapeutic strategies
7.3% (PPI)
facilitated the accurate assessment of ulcer healing and
Same
None

adverse events. Second, the CYP2C19 genotypes of the study

RCT

patients could not be determined due to the institutional

limitations. Future multicenter, large-sized, prospective
randomized controlled trials are warranted for detailed strati-
1262.6 [597.8–7322.3] (PPI)
1446.9 [605-3977.4] (PCAB)

fication of subjects to include analysis for the effect of the

CYP2C19 genotype in the efficacy of vonoprazan.
Takahashi et al.
2016 [34]

Included (30.7%)
LPZ 60mg 2days
14 (VPZ 20mg)
12 (LPZ 30mg)

Conclusion
P-CAB induced rapid artificial ulcer healing with the promo-
4 weeks

Same
None

None

None

tion of granulation tissue during the two weeks after ESD,

RCT
Table 6. Studies identified in a literature review that reported on effect of vonoprazan for gastric ESD-derived ulcer.

compared with the commonly used PPI lansoprazole.

However, none of the patients in either group suffered
adverse consequences, suggesting that conventional PPI with
initial intravenous infusion might be sufficient for the preven-
tion of postoperative bleeding following gastric ESD.
Muraoka et al.

Retrospective (PSM)
2016 [33]

Rebamipide 300mg

97.5 ± 3.2% (PCAB)

1260 ± 834 (PCAB)

94.8 ± 6.6% (PPI)

OMZ 40mg 2days

Disclosure statement
1251 ± 919 (PPI)
37 (VPZ 20mg)
38 (EPZ 20mg)

All authors (Yohei Horikawa, Hiroya Mizutamari, Nobuya Mimori, Yuhei

Excluded

4 weeks

Kato, Saki Fushimi, Sayaka Sato, and Syunji Okubo) declare that they
None

have no conflicts of interest or financial ties to disclose.

–

References
historical control cohort)
Retrospective (2:1 matched

[1] Ono H, Kondo H, Gotoda T, et al. Endoscopic mucosal resection

for treatment of early gastric cancer. Gut. 2001;48:225–229.
OMZ 40mg 2days (PPI)
Kagawa et al.

1140.1 ± 842.1 (PCAB)

Omeprazole; SD: standard deviation. significant difference.

[2] Yahagi N, Fujishiro M, Kakushima N, et al. Merits and demerits of

2016 [32]

Polaprezinc 150mg

963.1 ± 765.4 (PPI)

endoscopic submucosal dissection using a flex-knife. Stomach

Included (17.3%)
150 (RPZ 20mg)

5 weeks (PCAB)
75 (VPZ 20mg)

Intest. 2004;39:39–43.
8 weeks (PPI)
None (PCAB)

10.0% (PPI)
1.3% (PCAB)

[3] Oyama T, Tomori A, Hotta K, et al. Endoscopic submucosal dissec-

tion of early esophageal cancer. Clin Gastroenterol Hepatol.
POD 5

Same

2005;3(7 suppl 1):S67–S70.

[4] Toyonaga T, Nishino E, Man IM, et al. Principles of quality con-
trolled endoscopic submucosal dissection with appropriate dis-
section level and high quality resected specimen. Clin Endosc.
2012;45:362–374.
Mucosal protective agent use

mean ± SD or median [range]

[5] Kakushima N, Fujishiro M, Kodashima S, et al. Histopathologic

Antithrombotic agents use

characteristics of gastric ulcers created by endoscopic submucosal

Postoperative bleeding
Resection size (mm2)

dissection. Endoscopy. 2006;38:412–415.

2nd look endoscopy

Ulcer size reduction

[6] Lee SH, Lee CK, Chung IK, et al. Optimal duration of proton
pump inhibitor in the treatment of endoscopic submucosal dis-
Prior PPI use

Assessment

section-induced ulcers: a retrospective analysis and prospective

Number

validation study. Dig Dis Sci. 2012;57:429–434.

Author
Design

[7] Oda I, Gotoda T, Hamanaka H, et al. Endoscopic submucosal dis-

section of early gastric cancer: technical feasibility, operation

time, and complications from a large consecutive series. Dig
Endosc. 2005;17:54–58.
[8] Takizawa K, Oda I, Gotoda T, et al. Routine coagulation of visible
vessels may prevent delayed bleeding after endoscopic submuco-
sal dissection-an analysis of risk factors. Endoscopy. 2008;
40:179–183.
[9] Uedo N, Takeuchi Y, Yamada T, et al. 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.
[10] Park YM, Cho E, Kang HY, et al. The effectiveness and safety of
endoscopic submucosal dissection compared with endoscopic
mucosal resection for early gastric cancer: a systematic review
and metaanalysis. Surg Endosc. 2011;25:2666–2677.
[11] Jenkins H, Sakurai Y, Nishimura A, et al. Randomised clinical trial:
safety, tolerability, pharmacokinetics and pharmacodynamics of
repeated doses of TAK-438 (vonoprazan), a novel potassium-com-
petitive acid blocker, in healthy male subjects. Aliment Pharmacol
Ther. 2015;41:636–648.
[12] Fujiwara S, Morita Y, Toyonaga T, et al. A randomized controlled
trial of rebamipide plus rabeprazole for the healing of artificial
ulcers after endoscopic submucosal dissection. J Gastroenterol.
2011;46:595–602.
[13] Lim JH, Kim SG, Choi J, et al. Risk factors of delayed ulcer healing
after gastric endoscopic submucosal dissection. Surg Endosc.
2015;29:3666–3673.
[14] Niimi K, Fujishiro M, Goto O, et al. Prospective single-arm trial of
two-week rabeprazole treatment for ulcer healing after gastric
endoscopic submucosal dissection. Dig Endosc. 2012; 24:110–116.
[15] Kobayashi M, Takeuchi M, Hashimoto S, et al. Contributing factors
to gastric ulcer healing after endoscopic submucosal dissection
including the promoting effect of rebamipide. Dig Dis Sci.
2012;57:119–126.
[16] Horikawa Y, Mimori N, Mizutamari H, et al. Proper muscle layer
damage affects ulcer healing after gastric endoscopic submucosal
dissection. Dig Endosc. 2015;27:747–753.
[17] Kakushima N, Fujishiro M, Yahagi N, et al. Helicobactor pylori sta-
tus and the extent of gastric atrophy do not affect ulcer healing
after endoscopic submucosal dissection. J Gastroenterol Hepatol.
2006;21:1586–1589.
[18] Kakushima N, Yahagi N, Fujishiro M, et al. The healing process of
artificial ulcers after endoscopic submucosal dissection. Dig
Endosc. 2004;16:327–331.
[19] Ross ME, Kreider AR, Huang YS, et al. Propensity score methods
for analyzing observational data like randomized experiments:
challenges and solutions for rare outcomes and exposures. Am J
Epidemiol. 2015;181:989–995.
[20] Gotoda T, Yanagisawa A, Sasako M, et al. Incidence of lymph
node metastasis from early gastric cancer: estimation with a large
number of cases at two large centers. Gastric Cancer. 2000;
3:219–225.
SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY 251
[21] Kimura K, Takemoto T. An endoscopic recognition of the atrophic
border and its significance in chronic gastritis. Endoscopy.
1969;1:87–97.
[22] Horikawa Y, Toyonaga T, Mizutamari H, et al. Feasibility of knife-
coagulated cut in gastric endoscopic submucosal dissection: a
case-control study. Digestion. 2016; 94:192–198.
[23] Japanese Gastric Cancer Association. Japanese Classification of
Gastric Carcinoma 3rd English Edition. Gastric Cancer.
2011;14:101–112.
[24] Tarnawski AS, Ahluwalia A, Jones MK. Angiogenesis in gastric
mucosa: an important component of gastric erosion and ulcer
healing and its impairment in aging. J Gastroenterol Hepatol.
2014;29:112–123.
[25] Tarnawski AS. Cellular and molecular mechanisms of gastrointes-
tinal ulcer healing. Dig Dis Sci. 2005;50:S24–S33.
[26] Suzuki S, Gotoda T, Kusano C, et al. The efficacy and tolerability
of a triple therapy containing a potassium-competitive acid
blocker compared with a 7-day PPI-based low-dose clarithromycin
triple therapy. Am J Gastroenterol. 2016;111:949–956.
[27] Sakurai K, Suda H, Ido Y, et al. Comparative study: Vonoprazan
and proton pump inhibitors in Helicobacter pylori eradication
therapy. Wjg. 2017;23:668–675.
[28] Shimatani T, Inoue M, Kuroiwa T. Acid-suppressive effects of rabe-
prazole, omeprazole, and lansoprazole at reduced and standard
doses: a crossover comparative study in homozygous extensive
metabolizers of cytochrome P450 2C19. Clin Pharmacol Ther.
2006;79:144–152.
[29] Sakurai Y, Mori Y, Okamoto H, et al. Acid-inhibitory effects
of vonoprazan 20 mg compared with esomeprazole 20 mg or rabe-
prazole 10 mg in healthy adult male subjects-a randomised open-
label cross-over study. Aliment Pharmacol Ther. 2015;42:719–730.
[30] Malfertheiner P, Chan FK, McColl KE. Peptic ulcer disease. Lancet.
2009;374:1449–1461.
[31] Kagami T, Sahara S, Ichikawa H, et al. Potent acid inhibition by
vonoprazan in comparison with esomeprazole, with reference to
CYP2C19 genotype. Aliment Pharmacol Ther. 2016;43:1048–1059.
[32] Kagawa T, Iwamuro M, Ishikawa S, et al. Vonoprazan prevents
bleeding from endoscopic submucosal dissection-induced gastric
ulcers. Aliment Pharmacol Ther. 2016;44:583–591.
[33] Maruoka D, Arai M, Kasamatsu S, et al. Vonoprazan is superior to
proton pump inhibitors in healing artificial ulcers of the stomach
post-endoscopic submucosal dissection: a propensity score-
matching analysis. Dig Endosc. 2016;29:57–64.
[34] Takahashi K, Sato Y, Kohisa J, et al. Vonoprazan 20 mg vs lanso-
prazole 30 mg for endoscopic submucosal dissection-induced gas-
tric ulcers. Wjge. 2016;8:716–722.
[35] Tsuchiya I, Kato Y, Tanida E, et al. Effect of vonoprazan on the
treatment of artificial gastric ulcers after endoscopic submucosal
dissection: prospective randomized controlled trial. Dig Endosc.
2017;29:576–583.
[36] Miyoshi A, Aoki T, Matsuo Y, et al. The effects of Omeprazole on
upper gastrointestinal bleeding: pilot study (Japanese literature).
Jpn Pharmacol Ther. 1995;23:49–59.