Endoscopic submucosal dissection (ESD) has been recommended as the standard treatment for early gastric neoplasms.^1,2 Although the long-term outcome of ESD is excellent, procedure-related complications, such as bleeding, perforation, and post-procedural electrocoagulation syndrome, remain concerns.³ Post-ESD bleeding (PEB) is one of the most common of these complications, with an estimated incidence of 5% to 8.5%.^4-6 However, the incidence of PEB in high-risk patients, such as those administered concomitant antithrombotics, can be as high as 61.5%.^7-10 Therefore, several strategies have been developed to prevent PEB.^11-17 Among these, hemostatic powder (HP) is clinically used for gastrointestinal bleeding.^18,19 In addition, HPs have been shown to have a preventive effect against PEB.^20,21 HPs have the advantage of being easy to use without specialized endoscopic techniques because they are simply sprayed onto the surface of the post-ESD ulcer site through a catheter, regardless of the anatomical location. Moreover, they do not pose additional secondary tissue injury owing to their noncontact application. Several HPs are commercially available for clinical use. Recently developed HPs (Endospray and epidermal growth factor [EGF]-endospray) consist of super-absorbent polymer (sodium croscarmellose/sodium starch glycolate) and mucoadhesive polymer ingredients (hydroxyethyl cellulose) to form a hydrogel physical barrier. HPs rapidly absorb water from the blood, creating a high concentration of platelets, red blood cells, and coagulation proteins that accelerate the physiologic clotting cascade. In addition, interactions with the powders rapidly produce a gelled matrix that adheres to and seals the bleeding tissue. EGF-endospray contains the same materials as Endospray with additional EGF.²² EGF is one of several growth factors, including transforming growth factor alpha, basic fibroblast growth factor, and platelet-derived growth factor, known to heal ulcers by stimulating the reconstruction of damaged mucosal structures. Therefore, we investigated the hemostatic effects and mechanisms of HPs (Endospray and EGF-endospray) histologically after ESD in a porcine model.

1. Materials

Endospray (CGBio Co., Ltd., Seoul, Korea) is an endoscopic HP containing hydroxyethyl cellulose, with high hygroscopicity and mucosal adhesion properties. EGF-endospray (CGGEL^Ⓡ; CGBio Co., Ltd.) contains the same materials as Endospray with additional EGF.

2. Evaluation mechanism of HPs in vitro

The HPs were prepared for adhesion, waterproofing, water permeability, and water absorption tests (n=3 per group for each test). All the tests were conducted at 37℃±1℃. For gelation, the hemostatic agents and water were mixed at a ratio of 1:10. A litmus paper (Advantec UNIV; Toyo Roshi Kaisha Co. Ltd., Tokyo, Japan) was placed at the bottom of a Petri dish, and 10 g of the gelated HPs was placed over it to cover the litmus paper completely. The weight of the Petri dish in which the gel was placed was measured (W₁). For the adhesion test, the Petri dish was turned over after lid closure and stored in a constant-temperature chamber (37℃±1℃). We evaluated adhesion according to whether the gel remained intact on the surface of the Petri dish after 6 hours. For the waterproofing tests, approximately 25 mL of a pH 1.2 buffer (blue) was poured over the gelated HPs in the Petri dish, and the color of the litmus paper was assessed after 300±10 seconds. After completion of the waterproofing test, the sample was maintained at 37℃±1℃ for an additional 24 hours. The excess pH 1.2 buffer in the Petri dish was removed, and the weight was measured (W₂). Water permeability (g/m²∙hr) was then calculated using the following formula:

• Water permeability (g/m²∙hr)=(W₂–W₁)/(A×B)

  • W₁: initial sample weight (g)

  • W₂: sample weight after 24 hours (g)

  • A: area covered by gelated sample=0.001256 (m²)

  • B: time=24 (hours).

For the water absorption test, approximately 1 g of each HP was placed in a beaker, and the weight was measured (W₃). Next, 100 g of distilled water, pre-warmed to 37℃±1℃, was added to the sample. After the mixture reacted for 30 minutes in an incubator at 37℃±1℃, the sample weight was measured (W₄). The water absorption rate was then calculated using the following formula:

• Water absorption (%)=(W₄–W₃)/W₃×100

  • W₃: initial sample weight (g)

  • W₄: sample weight after 30 minutes (g)

3. Animals

Three male pigs weighing 50 to 60 kg (Sus scrofa; Cronex, Seoul, Korea) were used. All animal care and experimental procedures were performed in accordance with the regulations of the Experimental Animal Research Committee of the Republic of Korea (approval number: CRONEX-IACUC-201911002). The health of the animals was initially assessed. The animals were acclimated for 7 days after quarantine in the animal facility and were raised in individual spaces at 24℃±2℃ and 50%±10% relative humidity, with a ventilation frequency of 10 to 15 times/hr, 12 hours of lighting per day, and illumination intensity of 150 to 300 Lux. Dry food (500 g) was provided to each animal once a day, and the animals were fasted for 24 hours before the experiments. Water was freely provided after filtration with a reverse osmosis filter and an oil-water sterilizer and sterilization with ultraviolet light.

4. ESD-induced ulcer bleeding in porcine stomachs

ESD was performed in the porcine stomachs for the ulcer bleeding model. Weights and abnormalities were assessed before the experiment. Rompun (5 mg/kg; Bayer, Seoul, Korea), a muscle relaxant for animals, was mixed with Zoletil (15 mg/kg; VIRBAC, Carros, France), a general anesthetic agent, at a ratio of 6:4. Thereafter, the mixture was injected intramuscularly at 0.1 mL/kg to induce anesthesia. The animals were placed in the left lateral decubitus position on the operating table, and the same amount of anesthetic was injected intravenously. Inhalation anesthesia was administered using a 2:1 mixture of isoflurane (Troikaa Pharm. Ltd., Ahmedabad, India) and oxygen (Korea Gas Corp., Daegu, Korea). An endoscopic instrument (XL-200; Fujinon, Tokyo, Japan) was used for ESD, and three iatrogenic ulcers of 2 to 3 cm each were generated per animal. The target site was marked with a DualKnife (Olympus, Tokyo, Japan), and isotonic saline was injected into the submucosal layer. The ulcers were generated at least 5 cm apart to prevent mutual interference. One animal with three ulcers was then assigned to each group (control, Endospray, and EGF-endospray). To maximize bleeding, no additional hemostasis was performed after ESD (Supplementary Fig. 1).

5. Application of Endospray and EGF-endospray

Each of the three hemorrhagic ulcers in each porcine stomach was assigned to one of the three groups (control, Endospray, and EGF-endospray). In total, 3 g of each HP was applied using a delivery catheter and powder sprayer (Alto Shooter; Kaigen Co., Ltd., Osaka, Japan). The procedure was performed by a proficient endoscopist with more than 10 years of experience in ESD and endoscopic hemostasis. The control group did not receive any intervention for hemostasis during and after ESD. The animals were allowed to eat again 6 hours after the procedure. A total of 500 g of food was provided to each animal in the morning, and the animals were fasted for 12 hours before every endoscopic procedure. At 72 hours after ESD, the gastric tissue was extracted and fixed in 10% neutral buffered formalin.

6. Hemostatic function evaluation

The hemostatic effects of Endospray and EGF-endospray on the ESD site were observed through endoscopy immediately and at 5 minutes, 6 hours, and 72 hours after applying the HPs (Supplementary Fig. 2). Bleeding was defined as active bleeding lesions, including oozing or spurting bleeding. We evaluated the early hemostatic effects immediately and at 5 minutes after HP application. The hemostasis rate in the control group was evaluated as spontaneous hemostasis up to 5 minutes. Rebleeding was assessed at 6 and 72 hours after HP application. The success of initial hemostasis was assessed according to the hemostasis rate until 5 minutes after HP application. The rebleeding rate was evaluated according to the proportion of ulcers that rebled between 6 and 72 hours after HP application.

7. Histologic evaluation

The animals were sacrificed at 72 hours after HP application, and the stomachs were extracted to evaluate tissue regeneration. Samples fixed in 10% neutral buffered formalin (StatLab, McKinney, TX, USA) were dehydrated, made transparent, and penetrated using an automated tissue processor. Subsequently, the samples were embedded in paraffin. Using a microtome, the samples were cut into slices at a thickness of 4 μm. The sectioned samples were placed on separate slides to prepare them for staining. All the tissue samples were deparaffinized and hydrated. One tissue sample was stained with hematoxylin and eosin Y using Harris hematoxylin (YD Diagnostics, Yongin, Korea) and eosin Y (Showa, Gyoda, Japan). Hematoxylin and eosin images were used for the histological analysis of ulcer healing. The thickness of the superficial exudative zone was measured from the area where inflammatory cells aggregated on top of the granulation tissue and stained dark purple to the area of the surrounding necrotic tissue where it stained pale pink. The thickness of the granulation tissue was measured from below the superficial exudative zone to the muscle layer. The thickness of the superficial exudative zone and granulation tissue was measured at three locations centered on the ulcer, and the average value was analyzed. The number of new blood vessels was evaluated, and the average thereof was determined using 6 to 12 images of granulation tissue at 200x magnification.

8. Statistical analysis

Continuous variables are expressed as means and standard deviations, whereas categorical variables are expressed as frequencies and ratios. The results of the in vitro experiments were compared between the two HP groups (Endospray and EGF-endospray) using a two-sample t-test or Wilcoxon rank-sum test. The overall hemostatic rate and histological analysis data were analyzed using the one-way analysis of variance for comparisons across multiple groups (control, Endospray, and EGF-endospray), followed by the Tukey post hoc test. All the data were analyzed using Microsoft Excel. A p-value <0.05 was considered to indicate statistical significance.

1. Mechanisms of Endospray and EGF-endospray in vitro

The Endospray and EGF-endospray were gelated and adhered to the Petri dish surface for up to 6 hours (Fig. 1A). No color change in the pH measurement paper beneath the gel was observed because there was no penetration of the pH 1.2 buffer solution (Fig. 1B). This demonstrated that the HPs could completely block external water within 300 seconds. After the gelated Endospray and EGF-endospray were exposed to a pH 1.2 buffer (mimicking gastric acid) for 24 hours, the water permeability was 356.70±22.34 and 374.21±46.20 g/m²∙hr, respectively, and the difference between the groups was not significant (p=0.586). The water absorption rate (%) per unit weight (g) for both HPs was more than 20 times the initial weight, and the difference between the groups was not significant (Endospray, 2,376.75±64.78; EGF-endospray, 2,293.91±138.08; p=0.400) (Table 1).

Table 1 . In Vitro Characterization Results of Endospray and EGF-Endospray.

Test	Endospray	EGF-endospray	p-value
Adhesion, % (n/n)	100 (3/3)	100 (3/3)	NA
Waterproofing, % (n/n)	100 (3/3)	100 (3/3)	NA
Water permeability, mean±SD, g/m2∙hr	356.70±22.34	374.21±46.20	0.586*
Water absorption, mean±SD, %	2,376.75±64.78	2,293.91±138.08	0.400*

EGF, epidermal growth factor; NA, not available..

*Two-sample t-test..

Figure 1. Representative results of adhesion test (A) and waterproofing test (B).
EGF, epidermal growth factor.

2. Hemostatic function of Endospray and EGF-endospray

The hemostatic effects of Endospray and EGF-endospray were evaluated. All ESD procedures were performed in the lower third of the stomach, and the size of iatrogenic ulcers after ESD was 2 to 3 cm. The hemostasis rate immediately after application was 0% in the control group, while it was 66.7% (2/3) in both the Endospray and EGF-endospray groups. The hemostasis rates at 5 minutes after the application were 33.3% (1/3) in the control group and 100% (3/3) and 66.7% (2/3) in the Endospray and EGF-endospray groups, respectively (p=0.296). After 6 hours, the hemostasis rates were 33.3% (1/3) in the control and EGF-endospray groups and 66.7% (2/3) in the Endospray group (p=0.729). At 72 hours after application, hemostasis was achieved in all ulcers (Table 2). The initial hemostasis rates in the control, Endospray, and EGF-endospray groups were 33.3%, 100.0%, and 66.7%, respectively. In the control, Endospray, and EGF-endospray groups, the rebleeding rates were 33.3%, 16.7%, and 33.3%, respectively (Table 3).

Table 2 . The Number of Bleeding Ulcers That Achieved Hemostasis after Endoscopic Submucosal Dissection.

	Control (n=3)	Endospray (n=3)	EGF-endospray (n=3)
Immediately	0	2	2
5 Minutes	1	3	2
6 Hours	1	2	1
72 Hours	3	3	3

EGF, epidermal growth factor..

Table 3 . Initial Hemostasis Rate and Rebleeding Rates.

	Control	Endospray	EGF-endospray
Initial hemostasis rate	33.3	100	66.7
Rebleeding rate	33.3	16.7	33.3

EGF, epidermal growth factor..

3. Histologic evaluation of Endospray and EGF-endospray

Compared with the control group, Endospray and EGF-endospray groups demonstrated a greater thickness of regenerated mucosal and granulation tissue and a higher number of newly formed blood vessels (Fig. 2A). The mean thicknesses of the superficial exudative zone, in which necrotic tissue and inflammatory cells aggregated on the surface layer of the granulation tissue due to the wound-healing process, were 522.09±43.53 µm, 514.46±42.25 µm, and 680.35±22.73 µm in the control, Endospray, and EGF-endospray groups, respectively (Fig. 2B). The superficial exudate in the EGF-endospray group was 1.30 times thicker than that in the control group and 1.32 times thicker than that in the Endospray group (p=0.399). The mean thicknesses of the granulation tissue were 1,510.34±10.00 µm in the control group, 2,848.22±228.69 µm in the Endospray group, and 3,062.35±1,832.73 µm in the EGF-endospray group (Fig. 2C). Although there were no significant differences between the groups (p=0.404), the granulation tissue thickness was approximately 2.0 times greater in the Endospray and EGF-endospray groups than that in the control group. The numbers of new blood vessels per unit area were 15.27±8.74, 17.88±1.12, and 20.52±0.48 in the control, Endospray, and EGF-endospray groups, respectively (p=0.635) (Fig. 2D). The number of newly formed vessels was higher in the Endospray and EGF-endospray groups than that in the control group. Additionally, the number of newly formed blood vessels observed in the EGF-endospray group was 1.15 times greater than that in the Endospray group. No abnormal findings, such as complications due to the application of HPs, were observed before the animals were sacrificed.

Figure 2. Healing of gastric ulcers by Endospray or epidermal growth factor (EGF)-endospray treatment. (A) Representative microscopic images of hematoxylin and eosin staining of stomach sections at 72 hours in each group (×20). Histological evaluation of stomach sections for the thickness of (B) the superficial exudate zone, (C) granulation tissue, and (D) the number of new blood vessels per unit area.

With the expansion of ESD indications, various attempts have been made to reduce PEB, including topical HPs for endoscopic hemostasis. Topical HPs are non-traumatic and can be applied for diffuse and multifocal bleeding. Moreover, topical HPs have the advantage of being easy to use without specialized endoscopic techniques. Therefore, these powders can also induce hemostasis over a large area in a short period of time.^23,24 HPs immediately absorb body fluids, such as blood, and form a hydrogel barrier after being applied to bleeding ulcers, thereby inducing physical hemostatic effects.

The EGF-endospray used in this study was an endoscopic HP containing hydroxyethyl cellulose (mucosal adhesive polymer), sodium croscarmellose/sodium starch glycolate (super-absorbent polymer), and EGF (with wound-healing effects). According to the in vitro analysis of this study (Fig. 1), the water absorption of both HPs was more than 20 times (more than 2,000%) that of the initial weight of the sample within a short period of 30 minutes. Because both HPs had the same composition, except for the EGF, this result suggested that the addition of EGF does not influence the absorption capacity. Owing to their high water absorption rate, HPs can rapidly concentrate platelets and coagulation factors at the site of application through rapid body fluid absorption, thereby promoting hemostasis.²⁵ Although it was difficult to compare the study results with similar HPs directly, owing to the difference in the water exposure time and formulas used to calculate the water absorption,²⁶ this study demonstrated the excellent water absorption abilities of Endospray and EGF-endospray. Endospray and EGF-endospray achieve hemostasis by sealing the application site, in addition to promoting blood absorption. In this study, the gelated Endospray and EGF-endospray completely blocked water penetration into the application site within 300 seconds. In the adhesion test, the gelated HPs strongly adhered to the Petri dish surface for up to 6 hours.

In the animal study, the initial hemostasis rates were 83.3% and 66.7% in the Endospray and EGF-endospray groups, respectively. The hemostasis rate in the HP groups was approximately 4 to 5 times higher than that in the control group. These rates are similar to the hemostatic rate of Hemospray (Cook Medical, Bloomington, IN, USA), the most studied HP to date, in porcine gastric hemorrhagic ulcers.^27-29 However, there was a significant difference in the weight of the powder used to achieve hemostasis. In total, 20 to 50 g of Hemospray was required for hemostasis in spurting hemorrhage (Forrest grade Ia), and approximately 10 to 26 g of Hemospray was required for oozing hemorrhage (Forrest grade Ib).²⁸ In another study, an average of 100 g of Hemospray was required to achieve hemostasis in pulsatile bleeding.²⁹ Meanwhile, in our study, only 3 g of Endospray and EGF-endospray were applied to hemorrhagic ulcers with a 2 to 3 cm diameter. Rebleeding was observed in some ulcers treated with HP for up to 6 hours (Endospray, 16.7%; EGF-endospray, 33.3%); however, hemostasis was achieved in all ulcers after 72 hours. In this study, the rebleeding was higher in the EGF-endospray group at 6 hours. We think that the sample size of this study was too small to distinguish a difference in hemostatic effects according to the presence or absence of EGF.

Histological analysis (hematoxylin and eosin staining) was performed to determine the effect of EGF on gastrointestinal mucosal regeneration. The thickness of the superficial exudative zone was greater in the EGF-endospray group than that in the other two. This difference is likely because EGF increases gastric mucus secretion, attenuates gastric acid secretion, and stimulates cell migration in epithelial cell monolayers.³⁰ Therefore, EGF-endospray may stimulate physical hemostatic effects and release EGF to functionally promote cell proliferation and mucosal regeneration.^31,32 Submucosal regeneration was evaluated and compared by measuring the thickness of the granulation tissue. In the control group, the submucosal layer was not regenerated; thus, the granulation tissue was the thinnest in the three groups. However, the regenerated mucosal and submucosal layers in the EGF-endospray group were thicker than those in the other two groups. Moreover, the number of newly formed blood vessels in the Endospray and EGF-endospray groups was higher than that in the control group, indicating that the transition from the inflammatory to the proliferative phase was active during the wound-healing process. In a previous preclinical study using rabbits and pigs, the regenerated mucosa of the EGF group was much thicker, and the ulcer size much smaller than that of the control group.²² Since HP alone without EGF was not included, it was difficult to distinguish whether this healing effect was due to EGF or the mucoadhesive absorbable powder formulation. Therefore, we included a control group and HP-alone group without EGF. HPs had a positive effect on the regeneration and angiogenesis of the mucosal layer, and EGF had a positive effect on the regeneration of the submucosal layer. Currently, several HPs are commercially available for clinical use. They differ slightly in composition and action mechanisms. Few studies, however, have directly compared the efficacy of these HPs. Although we suspect that the efficacy of achieving hemostasis among HPs is likely similar, there are some technical differences among them. First, Hemospray is sprayed with high pressure and it has an advantage of covering a large area. However, high pressure application can cause further tissue injury in friable or inflamed mucosa, as well as perforation. In contrast, Endospray is sprayed at a much lower pressure, allowing for more sophisticated manipulation at a target area than Hemospray. In addition, a large amount of Hemospray is required to achieve hemostasis. However, in our study, only small amounts of Endospray and EGF-endospray were applied to hemorrhagic ulcers.

This study has several limitations. A small number of animals were included in the experiments and late rebleeding appearing after 72 hours was not assessed. Additionally, interaction between the three ulcers may have occurred because the three experimental groups were evaluated simultaneously in one animal. In addition, because all ESD procedures were performed in the lower third of the stomach, we were unable to evaluate differences according to anatomical location. The iatrogenic ulcers were planned to be created with a size of 2 to 3 cm before ESD. Even though the size of ulcers was confirmed in the target range with a size marker in all cases, the precise size of ulcers after procedure was not evaluated with a measuring instrument less than 1 cm. Also, while there was greater rebleeding in the EGF-endospray group at 6 hours compared to Endospray group, the sample size was too small to distinguish a difference in the hemostatic effect according to the presence or absence of EGF. Finally, because the gastric mucosa is different from Petri dish, future research using an in vivo model to evaluate HP mechanisms is warranted.

Despite these limitations, this study is meaningful as it analyzed the hemostatic effects and mechanisms of EGF-containing HP, thereby contributing to the data in this field. Although the addition of EGF did not significantly impact the hemostatic effect, histological findings indicated that EGF regenerates the defective mucosa, thickens the submucosal fibrous tissue, and promotes the formation of new blood vessels. In the future, large-scale studies that analyze late rebleeding rates, the optimal amount of HP required based on the amount of bleeding, systemic histological reactions based on the application of HP, and the long-term effects of ulcer healing are warranted.

In conclusion, Endospray and EGF-endospray demonstrated the ability to provide effective initial hemostasis and to exert histological ulcer-healing effects of EGF in an animal model of hemorrhagic gastric ulcer.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl220149.

This research was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute, funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI19C0534).

G.H.K. is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.

Study concept and design: B.M.M., J.C.P. Data acquisition: J.H.Y., B.M.M. Data analysis and interpretation: D.H.J., J.H.Y., B.M.M., J.S.K., H.L., G.H.K. Drafting of the manuscript: D.H.J., J.H.Y., J.C.P. Critical revision of the manuscript for important intellectual content: D.H.J., J.H.L., H.S.R., J.C.P. Study supervision: J.C.P. Approval of final manuscript: all authors.