Non-steroidal anti-inflammatory drugs inhibit the processes of mucosal cell proliferation associated with duodenal ulcer healing.

We demonstrate that, in patients taking non-steroidal anti-inflammatory drugs (NSAIDs), there is inhibition of the proliferation of mucosal cells that normally leads to healing of duodenal ulcers. A microdissection technique was used to quantitate mitosis in duodenal crypts at the ulcer edge, giving a regeneration index of mitotic rate at that site, as compared to nearby mucosa. In patients with duodenal ulcers occurring in the absence of NSAID therapy, there was a brisk regenerative response (median index 2.48, range 1.55-9.81, n = 8), significantly greater than in patients taking NSAIDs (median index 1.10, range 0.73-2.16, n = 10, p = 0.014). Inhibition of the process of epithelial cell division normally involved in duodenal ulcer healing could contribute to the delay in ulcer healing which may explain the higher complication rate for duodenal ulcer during NSAID therapy.     

Quantitative Evaluation of Mucosal Regeneration in Gastric Ulcers by Electronic Endoscopy —Comparison between Tractable and Intractable Ulcers—

Abstract: The electronic endoscope named TGS-50B (TOSHIBA) provides a fine resolvable image (Fig. 6) and flat white balance covering a wide range (Fig. 7). Using this electronic endoscope, the regenerative mucosal patterns in 15 cases with intractable and 30 cases with tractable gastric ulcers (Table 1) were observed at 2-week intervals until the healed stage or until 12 weeks after the start of H₂-blocker administration. To clarify the difference between intractable and tractable gastric ulcers, the electronic endoscopic images were fed into an image analyzer; SPICCA (AVIONICS). The regenerative mucosal pattern was converted into 20 binary digited particles by the image analyzer (Fig. 4, Color) and quantified by the oval specification (OSF) value (Fig. 5). The OSF value of the tractable ulcers significantly increased in the healing stage I (H₁) and healing stage 2 (H₂), in comparison with those of intractable ulcers (Fig. 8). Moreover, the OSF value of the tractable ulcer significantly increased at 4 weeks after the start of an H₂-blocker, comparing with that at 2 weeks. Consequently, the OSF values of the tractable ulcers were significantly high in comparison with those of the intractable ulcers, at 4, 6 and 8 weeks after the start of administration of H₂-blockers (Fig. 9). By histological study of the biopsy specimen, an increase in the OSF value suggested an extension of the regenerative mucosa by a well progressed annular contraction of the gastric wall. In conclusion, it is thought that intractability of gastric ulcers may be objectively predictable in its early healing stages by determination of the OSF value using the electronic endoscope and the image analyzer.     

Changes in Gastric Mucosal Blood Flow during Healing of EMR-Induced Ulcer

Abstract: We used the laser Doppler method to study the difference in gastric mucosal blood flow changes between peptic ulcer (65 cases) and artificial ulcer caused by endoscopic mucosal resection (35 cases) during their respective healing processes. At each endoscopic ulcer stage, blood flow at the ulcer margin and that in the surrounding mucosa were measured. In the artificial ulcer, which heals easily, blood flow at the ulcer margin was still high at the scarring stage as compared with that in the corresponding area of a peptic ulcer, which is prone to relapse. Moreover, the blood flow ratio (blood flow at the ulcer margin/blood flow in the surrounding mucosa) at the S1 stage in artificial ulcers was significantly higher than that in peptic ulcers (p<0.05). These results suggest that blood flow in the SI stage is an important aspect of ulcer healing and relapse.     

Mucosal cell proliferation in duodenal ulcer and duodenitis.

Mucosal cell proliferation in the first part of the duodenum was studied in 24 patients using a tissue culture technique in which endoscopic biopsies were subjected to autoradiography after exposure to tritiated thymidine. Eight patients had a normal duodenum, eight had duodenal ulcer, and eight had symptomatic chronic non-specific duodenitis. The mean crypt labelling index (LI) in normal duodenum was 8.8 0.4% (SEM). Increased labelling indices of 15.6 +/- 1.7% were found near the edge of duodenal ulcers and 17.8 1.8% in duodenitis. Treatment with cimetidine reduced both the severity of duodenitis and the mean crypt LI. The LI of histologically normal duodenal mucosa distal to ulcer of duodenitis was similar to that of the control subjects' mucosa. The increased mucosal cell proliferation seen in severe duodenitis, either alone or associated with duodenal ulceration, suggested that erosions and ulcers arose when the crypts passed into 'high output failure' and were unable to compensate for further epithelial cell loss. There was no evidence in out study for a generalised failure of mucosal cell proliferation in duodenal ulcer or duodenitis.     

S phase cells of regenerative epithelium in the healing process of gastric ulcer determined by an in vitro BrdU-anti BrdU method

In order to investigate the healing process of gastric ulcer, cell proliferation kinetics of the regenerative mucosa over gastric ulcers was studied by an in vitro bromodeoxyuridine (BrdU)-anti BrdU method. In addition, the effect of histamine H2 receptor antagonists on gastric cell kinetics was also studied. The BrdU labeling index (L.I.) at the site ulcer margin, which were determined by the ratio of labeled cells to epithelial cells of gastric mucosa, were significantly higher in cases of endoscopic A to S1 stage ulcer than in cases of the atrophic gastritis served as control. While the mean L.I. of surrounding mucosa, approximately 1 cm apart from the ulcer margin, in the stage A to S1 were also significantly higher than that of control, the value in the stage S2 was almost comparable to the value of control. There were no differences of L.I. between the cases treated with and without H2 receptor antagonists. These results suggest that in the healing process of ulcer not only reserved epithelial cells at the ulcer margin but also the cells apart from the margin participate in mucosal regeneration. And it seems advisable to consider the stage S2 rather than the stage S1 as the most precise healing stage of ulcer. H2 receptor antagonist may not give an influence on proliferative properties of regenerating mucosal epithelium.     

Hypergastrinemia increases proliferation of gastroduodenal epithelium during gastric ulcer healing in rats

We investigated if hypergastrinemia exerted any influence on the proliferation of gastroduodenal epithelium during the healing of ulcers in rats. A mucosal ulcer was induced in the corpus region of the stomach in three groups of rats, which were then given vehicle, omeprazole (400 mol/kg/day), or gastrin-17 (60 nmol/kg/day) for three or six days. A fourth group of unoperated rats served as controls. One hour before killing, [³H]thymidine was injected. The ulcer margin and corresponding control tissues were excised and processed for light microscopic determination of epithelial labeling index (LI), mitotic index, and apoptotic index. LI was also determined in other parts of the gastroduodenal mucosa. Three and six days after the ulcer operation, the LI in the vehicle-treated ulcer rats was significantly increased in the ulcer margin and in the duodenum, in comparison with the intact controls. In the ulcer margin, the mitotic index was significantly increased, in parallel with the LI; the apoptotic index remained at the control level. The LI in the ulcer margin was increased further after administration of omeprazole or gastrin-17, which elevated the plasma gastrin levels by 5–15 times. It is concluded that hypergastrinemia may increase cell proliferation in the ulcer margin, which may accelerate the rate of healing.     

In vivo cell kinetics studies of the duodenum during and after healing of cysteamine-induced duodenal ulcer in rats using bromodeoxyuridine incorporation

The present study investigated the defect of duodenal defence responsible for the tendency of recurrence of duodenal ulcer. Male Wistar rats were treated with cysteamine-HCl subcutaneously to induce duodenal ulcer, and bromodeoxyuridine (BrdU) was given intraperitoneally 1 h before laparotomy. Serial tissue sections and immunocytochemical staining of BrdU were done to study the cell kinetics during and after healing. Deep ulcers developed in the proximal part of the duodenum 24 h after cysteamine injection, and the ulcer was replaced by scar tissue 1 week later. BrdU-labelling index of normal duodenal mucosa in control rats was 30-34%. On the peripheral part of the regenerative mucosa, BrdU-labelling index increased from 1.5% to 26-27%, 1-4 weeks after injection and remained at this level thereafter. On the central part, except the most central area, the labelling index remained at 0% until 3 weeks, and was 15% 6 weeks after cysteamine treatment. It never achieved the level seen in control normal mucosa. The labelling rate of fibrous cells in scar tissue decreased from 28% to nearly 0% 1-4 weeks after the injection. It is concluded that both the ulcer scar and the regenerative mucosa do not achieve a mature state in the initial scarred stage; they need more time to reach a relatively mature condition. Moreover, the regenerative mucosa will be the weak, easily damaged part of the duodenum.     

Role of epidermal growth factor in healing of chronic gastroduodenal ulcers in rats

The healing of acetic acid-induced gastric and duodenal ulcers was examined together with biochemical indices of growth in gastric and duodenal mucosa in rats with intact or removed salivary glands after treatment with epidermal growth factor (EGF) or somatostatin, or both. After the extirpation of salivary glands, the healing rate of gastric and duodenal ulcerations was delayed and gastric content of immunoreactive EGF was reduced. This was accompanied by a significant decrease in the contents of deoxyribonucleic acid and ribonucleic acid in the gastric and duodenal mucosa. Repeated administration of EGF either subcutaneously or orally accelerated the healing of gastroduodenal ulcers in rats with intact salivary glands and completely reversed the delay in ulcer healing in sialoadenectomized animals. These effects were also accompanied by a significant increase in the growth parameters of gastric and duodenal mucosa. Administration of somatostatin, which prevented the growth-promoting action of subcutaneous EGF, resulted in a significant decrease in the EGF-stimulated healing of gastric and duodenal ulcerations in both intact and sialoadenectomized rats. Our findings suggest that cell proliferation is an important factor in healing of gastric and duodenal ulcerations and that EGF plays an important role in ulcer healing due to its mitogenic action.     

Effect of cell proliferation on healing of gastric and duodenal ulcers in rats

The healing of acetic acid-induced gastric and duodenal ulcers was examined together with the biochemical indices of growth in gastric and duodenal mucosa in the following three groups of rats: (a) chow-fed, (b) fed an isocaloric liquid diet, (c) fed the liquid diet plus pentagastrin injections (250 micrograms/kg, 3 times/day). Animals received the diet regimen for 10 days from 1 day after induction of ulcer (day 0). Following the feeding regimens, serum gastrin levels, oxyntic gland mucosal DNA synthesis, and gastric secretory function were significantly lowered in the rats fed liquid diets. DNA synthesis in the duodenal mucosa was not different from the pre-ulcer levels. Pentagastrin significantly restored the DNA synthetic and gastric secretory activity of the liquid diet-fed rats toward the levels in the chow-fed group. In the latter group, a significant increase in DNA synthesis and levels of serum gastrin was found at day 6 (after 5 days feeding), which corresponded with a rapid, spontaneous healing of ulcers. Feeding rats liquid diet significantly delayed the healing of gastric, but not duodenal ulcers. Repeated administration of pentagastrin accelerated gastric ulcer healing in the liquid diet group toward the rate observed in the chow-fed group, but had no effect on the healing of duodenal ulcers. These results indicate that cell proliferation is an important factor in the healing of gastric ulcers.     

Cholecystokinin-like activity in the duodenal mucosa of duodenal ulcer patients.

Cholecystokinin-like activity in the duodenal mucosa was measured by the bioassay method described by Ljungberg to elucidate its significance in 14 duodenal ulcer patients as well as in 13 normal subjects with no evidence of gastrointestinal diseases. The stage of duodenal ulceration was determined endoscopically according to the criterion of the Japanese Gastroenterological Endoscopic Society. The cholecystokinin-like activity in the duodenal mucosa of duodenal ulcer patients in active stage 1, which was considered as an early stage of active open duodenal ulceration, did not differ statistically from that of normal subjects, whereas that of duodenal ulcer patients in active stage 2 began to show a significant increase (p less than 0.05), and the cholecystokinin-like activity in the duodenal mucosa of duodenal ulcer patients in healing stage 1 or healing stage 2 was significantly higher than that in normal subjects (p less than 0.01). The cholecystokinin-like activity in the duodenal mucosa of duodenal ulcer patients in the scarring stage, however, returned to the normal range. It is concluded that cholecystokinin may act physiologically in the cure of duodenal ulcer.     

The Distribution of Helicobacter pylori in Human Gastric Mucosa in vivo

Abstract: To estimate the distribution of Helicobacter pylori in human gastric mucosa in vivo, the phenol red dye spraying endoscopy lias been successfully developed and is performed after an oral and/or intravenous premedication of famotidine 20mg. This technique was conducted on 25 patients with chronic, atrophic gastritis, on 21 patients with a gastric ulcer and on 14 patients with duodenal ulcers. The red color changes which occurred on the gastric mucosa were classified into three types; the diffuse staining type, the regional staining type and the patchy staining type. In the chronic gastritis group, the diffuse staining type, the regional staining type and the patchy staining type occurred in 11 patients (44%), 7 patients (28%) and 2 patients (8%), respectively. The remainder of the patients' mucosa was unstained. In addition, the regional staining type occurred most frequently in the gastric ulcer group, while the diffuse staining type was dominant in the duodenal ulcer group. Notably, a recurrent and intractable ulcer was surrounded by regional staining fields in the stomach, and showed a diffuse staining at the antrum of the duodenum. These facts suggest that Helicobacter pylori prevented ulcer healing. This concurs with the results of our previous experimental study which found that the low concentration NH₃ solution, produced by Helicobacter pylori, prevented an acetic acid ulcer healing in rats and resulted from the suppression of the cell kinetics of the regenerative epithelial cells and the fibroblasts in the connective tissues at the ulcer margins.     

Epithelial Cell Proliferation in Duodenal Ulcer

Epithelial cell proliferation in the duodenum was investigated in 50 patients by incubating mucosal biopsy samples with tritiated thymidine, followed by autoradiography. Fifteen patients had a normal duodenum, 15 duodenal ulcer undergoing elective surgery, 10 perforated duodenal ulcer, and 5 severe non-ulcer-associated duodenitis. The mean crypt cell labelling index in the duodenal bulb of controls was 8.8 ± 0.4% (mean ± SEM), at the edge of perforated ulcers 19.1 ± 2.0%, at the edge of elective ulcers 18.6 ± 1.4%, and in biopsy specimens from non-ulcer-associated duodenitis 14.0 ± 1.2%. The mean labelling index in the distal first part of duodenum of control patients was 9.1 ± 0.8 similar to the values found in histologically normal specimens distal to ulcer or duodenitis. The results indicate active epithelial cell proliferation in both duodenal ulcer and duodenitis. There was no evidence of impairment of epithelial cell proliferation in duodenal ulcer patients.     

Significance of Helicobacter Pylori Infection in Human Peptic Ulcers

Abstract: Experimental studies have suggested that the continuous administration of 0.02% NH, solution, induced by Helicobacter pylori (H, Pylori), leads to a glandular atrophy of the gastric mucosa, and adversely affects healing of acetic acid ulcers in rats, because of the suppression of cell kinetics of the regenerative epithelial cells and connective tissues at ulcer margins. To visualize the distribution of H. pylori in human gastric mucosa, a phenol red dye spraying endoscopy was performed in 45 patients with gastric ulcers, and 43 patients with duodenal ulcers, who were medicated with a full dose of H₂-blocker until ulcer healing, and with half doses thereafter. In the H. pylori negative cases, 8 (88.9%) of 9 gastric ulcers healed within 3 months after medication, with no relapse discernible up to 6 months after healing of the preceding ulcer. The relapse rate was 25% up to 12 months after ulcer healing. In contrast, only 22 (66.1%) of 36 gastric ulcers healed within 3 months after medication in the H. pylori positive cases. The relapse rate was 12.5% up to 3 months, 30.4% UP to 6 months and 63.6% up to 12 months after ulcer healing. In addition, all 6 duodenal ulcers healed within 2 months after medication in the H. pylori negative cases, with no relapse discernible up to 12 months after healing of the preceding ulcer. In contrast, in the H. pylon positive cases, 20 (53.1%) of 37 duodenal ulcers healed within 2 months, and the relapse rate was 14.3%, 33.3%, and 66.7% up to 3, 6 and 12 months respectively after healing of the preceding ulcer. These data suggest that H. pylori is likely to interfer with ulcer healing, and promotes peptic ulcer relapse.     

Reduction in index of oxygen saturation at margin of active duodenal ulcers may lead to slow healing

This study tested the hypothesis that reduced perfusion of a duodenal ulcer margin (ie, the mucosa 1–2 mm from the edge of the ulcer base) is associated with slow healing. Reflectance spectrophotometric measurement of indices of mucosal hemoglobin concentration (IHB) and mucosal hemoglobin oxygen saturation (ISO₂) were obtained endoscopically in 21 patients at the ulcer margin and the adjacent mucosa (ie, the mucosa 1–2 cm from the edge of the ulcer base). In 17 patients with adequate follow-up, stepwise multilinear regression analysis revealed a significantly negative correlation (r=s-0.69, P < 0.05) between ISO₂ at the ulcer margin minus ISO₂ at the adjacent mucosa (ISO)₂ and ulcer healing time. In addition, smoking, being black, and early relapse since the last ulcer attack were found to be associated with increased duration required for healing. The results of this pilot study suggest factors, in addition to smoking, that may have to be considered in future studies concerned with duodenal ulcer healing.This work was supported by the National Institute of Arthritis and Metabolism and Digestive Diseases Grant AM34840, American Society for Gastrointestinal Endoscopy Career Development Award H850208, Veterans Administration Medical Research Funds, and UCLA Academic Senate Grant 4063.     

Reproducible demonstration of blood flow at duodenal ulcer margins by endoscopic reflectance spectrophotometry

Changes in gastrointestinal mucosal blood flow were evaluated by index of oxygen saturation (ISO2) and index of hemoglobin concentration (IHB) measured with a reflectance spectrophotometer. This report examined the reproducibility of endoscopic measurements of ISO2 and IHB. Study 1: The everted stomachs of three anesthetized rats provided hands-on instruction (one teacher and three learners). Six sets of readings were obtained endoscopically (the mean calculated to give the measurement) at each level of gastric mucosal perfusion when gastric blood flow was varied by withdrawing blood from the carotid artery. Study 2: Fourteen duodenal ulcer patients with ulcer bleeding were transfused and stabilized. Two endoscopists (one teacher and one learner) took turns to obtain endoscopic ISO2 and IHB measurements at the margin of the ulcer and at the adjacent normal appearing mucosa. delta ISO2 was calculated as the ulcer margin value minus adjacent mucosa value. In study 1, the correlation coefficients between the ISO2 measurements of the experienced and those of the other three observers were 0.99, 0.97, and 0.97, respectively. In study 2, the correlation coefficients between the ISO2 measurements obtained at the ulcer margin and at the adjacent normal mucosa, and delta ISO2 obtained by the experienced observer and one of the three learners were 0.94, 0.97, and 0.94, respectively. Relative to the adjacent area, 79% of the duodenal ulcers studied had increased (+delta ISO2), and 21% had decreased blood flow (-delta ISO2) at the ulcer margins. IHB measurements were less reproducible, particularly at the ulcer margin. The measurements of ISO2 and delta ISO2 were reproducible in the everted rat stomach and in duodenal ulcer patients.(ABSTRACT TRUNCATED AT 250 WORDS)     

Role of apoptosis induced by Helicobacter pylori infection in the development of duodenal ulcer

BACKGROUND: Helicobacter pylori affects gastric epithelium integrity by acceleration of apoptosis. However, it remains unclear what product of the bacteria causes apoptosis, or whether or not the apoptosis is involved in the development of ulcers. AIMS: To elucidate the factor from H pylori that causes acceleration of apoptosis and the role of apoptosis in the development of duodenal ulcer in H pylori infection. PATIENTS: Five H pylori negative healthy volunteers, 47 H pylori positive patients with duodenal ulcer, and 35 H pylori positive patients with gastric ulcer. METHODS: An endoscopic examination was carried out to diagnose ulcers and determine their clinical stage. To analyse apoptosis, a cell cycle analysis was performed using biopsy specimens. RESULTS: There was a significant correlation between the urease activity of the H pylori strain and the level of apoptosis induced by this bacterial strain. Moreover, in duodenal ulcer patients infected with H pylori, the patients with an active ulcer exhibited a significantly higher level of apoptosis than those with ulcers at both the healing and scarring stages. CONCLUSION: These findings suggest that acceleration of apoptosis in the antral mucosa caused by the urease of H pylori plays a crucial role in the development of ulcers in the duodenum.     

Effect of sucralfate granules in suspension on endoscopic variceal sclerotherapy induced ulcer: analysis of the factors determining ulcer healing

Oesophageal ulcers commonly occur after endoscopic variceal sclerotherapy and usually cause complications and a delay in further sclerotherapy. The aims of this study are to investigate the effect of sucralfate granules in suspension on the treatment of endoscopic variceal sclerotherapy induced ulcer and analyse the factors determining the ulcer healing. Fifty-two patients with oesophageal variceal bleeding received elective endoscopic variceal sclerotherapy. After endoscopically proved oesophageal ulcers, they were randomized to receive either sucralfate granules in suspension (n = 22) or antacid (n = 23). Follow-up endoscopy was performed weekly. Ulcer healing rates were compared between the groups using the log-rank test. Forty-one ulcers receiving sucralfate and 48 ulcers receiving antacid treatment were evaluated. The clinical characteristics of the ulcers were similar in both groups. The ulcers in patients receiving sucralfate healed faster than those receiving antacid (P<0.02). On analysis of factors affecting ulcer healing, ulcers smaller than 1 cm2 (n = 59) appeared to heal faster than those larger than 1 cm2 (n = 30; P= 0.059) and shallow ulcers (n = 46) healed faster than deep ulcers (n = 43; P<0.001). On multifactorial analysis, ulcer depth was the only factor determining ulcer healing. The ulcer healing effects of sucralfate became more prominent when the ulcer was larger than 1.0cm2 (1.7+/-0.6 weeks vs 2.3+/-0.6 weeks, P= 0.011) and deep (1.7+/-0.7 weeks vs 2.5+/-1.0 weeks, P= 0.013) when compared with those receiving antacid. Sucralfate granules in suspension speed the healing of endoscopic variceal sclerotherapy induced ulcer, especially deep and large ulcers.     

Expression and immunolocalization of heat shock proteins in the healing of gastric ulcers in rats

BACKGROUND: Heat shock proteins are induced when cells are subjected to noxious stimuli. They afford cytoprotection and increase the resistance of the tissue to damage. However, their roles in the healing of gastric ulcers have not been well established. In this study, the expression and immunolocalization of three heat shock proteins (HSPs); namely inducible HSP70 (iHSP70), HSP47, and HSP32 during ulcer healing were investigated in rats with gastric ulcer. METHODS: Gastric ulcers (kissing ulcers) were induced by luminal application of acetic acid solution. Gastric tissue samples were obtained from the ulcer base, ulcer margin, and non-ulcerated area around the ulcer margin at different time intervals after ulcer induction. The protein levels and distributions of HSPs were analyzed with Western blotting and immunohistochemical methods, respectively. RESULTS: It was found that all HSPs were expressed in normal, non-ulcerated, and gastric ulcer tissues. HSP32 was elevated during inflammation (1-8 days after ulcer induction), while HSP47 expression was exacerbated at the ulcer base and margin during ulcer healing (3-12 days). Decreased expression of iHSP70 was observed at the ulcer base immediately after ulcer induction, but returned to normal level by the end of the healing stage (8-12 days). Inducible HSP70 was found distributed in the gastric glands and injured tissues in the inflamed areas. Wide distribution of HSP47 was detected in granulation tissues and collagen producing cells, while HSP32 was localized in the gastric glands and inflammatory cells. CONCLUSIONS: The findings indicate that iHSP70, HSP47, and HSP32 play different roles during ulcer healing. HSP32 seems to act as an inflammatory defensive factor, and HSP47 as a collagen-specific molecular chaperon contributing significantly to gastric ulcer healing. However, the role of iHSP70 in the ulcer healing process is still undefined.     

Salivary epidermal growth factor in patients with peptic ulcer]

Recent animal studies suggest salivary epidermal growth factor (EGF) has a cytoprotective effect in the upper GI tract and is one of the important factors to promote the healing of experimental ulcer. The present study was undertaken to clarify the role of salivary EGF in peptic ulcer patients. Saliva samples were collected from 129 endoscopically normal subjects and 232 peptic ulcer patients. Salivary EGF concentration was measured by RIA. Salivary EGF output in normal subjects was 5.26 +/- 0.26 (ng/5 min) (mean +/- SE). Those in patients with gastric ulcer (GU), duodenal ulcer (DU) and gastroduodenal ulcer (GDU) were 10.74 +/- 0.15, 8.13 +/- 0.83 and 9.79 +/- 0.91. EGF output in GU and GDU patients were higher than that in normal subjects respectively. Tractable GU patients (healed within 3 months with regular regimen) had higher EGF output than intractable GU patients. Among tractable GU patients, those who had healing within 8 weeks had higher output. EGF output in patients with recurrent GU was lower than that in non-recurrent GU patients. In 10 GU patients, EGF output became higher in healing stage than in active stage. Salivary EGF may promote the healing and prevent the recurrence of human gastric ulcer.