Quadruple therapy with ecabet sodium, omeprazole, amoxicillin and metronidazole is effective for eradication of Helicobacter pylori after failure of first-line therapy (KDOG0201 Study).

BACKGROUND AND OBJECT: An antiulcer agent, ecabet sodium, is active against Helicobacter pylori. The aim of the present study was to clinically examine whether eradication therapy, which includes ecabet sodium, is effective in eradication of H. pylori after failure of first-line therapy. METHODS: Patients with peptic ulcer who failed with first-line triple eradication therapy containing clarithromycin received quadruple therapy with omeprazole (20 mg, twice daily), amoxicillin (750 mg, twice daily), metronidazole (500 mg, twice daily) and ecabet sodium (1000 mg, twice daily) for 14 days. Eradication of H. pylori was judged by 13C-urea breath test 8 weeks later. RESULTS: Fifty-two patients (36 men and 16 women) were included. Their mean age was 51.4 years (range 28-73). One patient dropped out because of diarrhoea. The eradication rate was 98.0% (50/51) according to the per-protocol analysis and 96.2% (50/52) according to the intention-to-treat analysis. Side effects occurred in seven patients, but none were serious. CONCLUSIONS: Quadruple therapy including ecabet sodium is useful as second-line eradication treatment for H. pylori.

High Ki67, Bax, and thymidylate synthase expression well correlates with response to chemoradiation therapy in locally advanced rectal cancers: proposal of a logistic model for prediction.

BACKGROUND: Recently, preoperative chemoradiation therapy (CRT) for rectal cancer has been increasingly used as a neoadjuvant treatment. In the present study, the relation between histological response to CRT and immunohistochemical markers in biopsy specimens was investigated. METHODS: Biopsy specimens from a total of 60 patients were collected before preoperative CRT with S-1 and irinotecan, and liniac 45 Gy. Immunohistochemical staining for Ki67, Mcm3, Bax, Bcl-2, ssDNA, Grp78, thymidylate synthase (TS), dihydropyrimidine dehydrogenase (DPD), CD34, vascular endothelial growth factor, nestin, and L-type amino-acid transporter 1 was performed to allow comparison of the Ki67 labelling index (LI), Bax score, TS score, DPD score, microvessel density by CD34, and Grp78 score with cancer regression. RESULTS: When the cases were divided into responders (Dworak grades 3 and 4) and non-responders (grades 1 and 2) groups, good correlations were evident with Ki67 LI, Bax, Grp78, and TS expression. On multiple logistic regression analysis, Ki67 LI, Bax, and TS scores were found to be independent factors. With their use in a logistic model, P-values could predict responder cases with a sensitivity of 82.8% and a specificity of 83.9%. CONCLUSION: Using this system, treatment strategy for locally advanced rectal cancers can be determined before chemoradiation.

Mixed-histologic-type submucosal invasive gastric cancer as a risk factor for lymph node metastasis: feasibility of endoscopic submucosal dissection.

BACKGROUND AND STUDY AIMS: The clinicopathologic features of gastric cancers containing a mixture of differentiated-type and undifferentiated-type components remain uninvestigated. We evaluated the risk of lymph node metastasis and the feasibility of endoscopic submucosal dissection (ESD) for the treatment of mixed-histologic-type gastric cancers. PATIENT AND METHODS: We histologically classified 376 cases of gastric cancer with submucosal invasion into four types (differentiated type, differentiated-type-predominant mixed type, undifferentiated-type-predominant mixed type, and undifferentiated type) and studied the clinicopathologic relations of each type to lymph node metastasis. Lymphatic invasion was evaluated by D2-40 immunostaining. RESULTS: The overall prevalence of lymph node metastasis in gastric cancer with submucosal invasion was 16.5% (62/376). The prevalence of lymph node metastasis was 36.5% (23/63) in undifferentiated-type-predominant mixed type, which was significantly higher than those in the other three types (P < 0.001 vs. differentiated type, P = 0.013 vs. differentiated-type-predominant mixed type, and P = 0.003 vs. undifferentiated type). Lymphatic invasion, a depth of invasion of 500 microm or more from the lower margin of the muscularis mucosae (SM2), tumor size above 30 mm, and undifferentiated-type-predominant mixed histologic type were independent risk factors for lymph node metastasis. Submucosal cancers without these four risk factors were free of lymph node metastasis (0/41; 95 % confidence interval 0%-8.6%). CONCLUSIONS: Undifferentiated-type-predominant mixed-type gastric cancer with submucosal invasion carries a high risk of lymph node metastasis. ESD can be indicated for gastric cancer with submucosal invasion provided that the following conditions indicating a low risk of metastasis are met: a depth of invasion of no more than 500 microm or more from the lower margin of the muscularis mucosae (SM1), no lymphatic invasion, a tumor size of no more than 30 mm, and a proportion of undifferentiated components below 50%.

Successful treatment of S-1 + CDDP followed by salvage EMR for a case with metastatic Barrett's esophageal cancer.

A 62-year-old woman with Barrett's esophageal cancer was hospitalized. Abdominal CT confirmed metastases to the liver and lymph nodes, for which surgical excision and radiotherapy were not indicated. We started chemotherapy with a course of daily oral S-1 at a dose of 80 mg/m(2) for 21 days, with a 2-hour drip of cisplatin at 60 mg/m(2) on day 8. Breaks of 14 drug-free days were given between courses. After two courses, a repeat CT confirmed that the liver and lymph node metastases had disappeared; after three courses, another CT confirmed that the metastatic foci were still absent, so we judged the disease to be in complete remission. Endoscopy and upper GI series confirmed that the primary tumor was reduced, and endoscopic mucosal resection performed using the strip biopsy method. The excision specimen was well differentiated adenocarcinoma, and mucosal invasion, and the excision stump was negative. After two more courses of S-1 + cisplatin, chemotherapy has been suspended with the patient's consent, and in the 21 months after endoscopic mucosal resection, no recurrence has been observed. This is a rare case of metastatic Barrett's esophageal cancer in which the metastases were eradicated by S-1 + cisplatin, and the primary tumor successfully excised by endoscopic mucosal resection after downstaging.

Expansion of the indications for endoscopic mucosal resection in patients with superficial esophageal carcinoma.

BACKGROUND AND STUDY AIMS: Endoscopic mucosal resection (EMR) is a minimally invasive local treatment for superficial esophageal carcinoma (SEC). The use of EMR in patients with m3 or sm1 SEC remains controversial, however. The aim of this retrospective study was to evaluate the histopathological risk factors for lymph-node metastasis and recurrence in patients with m3 or sm1 SEC. PATIENTS AND METHODS: The study subjects were 43 patients with m3 or sm1 esophageal squamous-cell carcinomas: 23 patients were treated surgically (the surgery group), and 20 were treated by EMR (the EMR group). We assessed the following variables of the specimens resected by surgery or EMR: tumor depth, maximal surface diameter of the tumor (superficial size), maximum diameter of tumor invasion at the lamina muscularis mucosae (LMM invasion width), and lymphatic invasion. The relationships of these variables to lymph-node metastasis and recurrence were examined. RESULTS: In the surgery group, lymph-node metastasis was found in four patients, all of whom had tumors with lymphatic invasion, a superficial size of at least 25 mm, and an LMM invasion width of at least 2500 microm. In the EMR group, no patient met all three of these criteria, and there was no evidence of lymph-node metastasis or distant metastasis on follow-up after EMR (median follow-up 39 months). CONCLUSIONS: In patients with m3 or sm1 SEC, tumors that have lymphatic invasion, larger superficial size, and wider LMM invasion are associated with a high risk for lymph-node metastasis. EMR might be indicated for the treatment of patients with m3 or sm1 SECs without these characteristics.

Restoration by VIP of the carbachol-stimulated Cl- secretion in TTX-treated guinea pig distal colon.

To determine if vasoactive intestinal peptide (VIP) restores neural activity from tetrodotoxin (TTX) blockade, we studied the effects of VIP and related agents on carbachol (Cch)-induced Cl(-) secretion in control-isolated guinea pig distal colon and in that treated with TTX. The short circuit current (I(sc)) increased dose-dependently after serosal applications of Cch (10(-6) - 2 x 10(-5) M) and VIP (5 x 10(-9) - 10(-7) M). But no additive or synergistic increase in I(sc) was observed. Cch- and VIP-induced I(sc) was completely abolished by a serosal application of TTX (10(-6) M). However, a serosal application, not mucosal, of VIP (10(-7) M) and 8-bromo-cAMP (10(-3) M) restored the Cch-stimulated, TTX-inhibited I(sc) by 113% and 75.8%, respectively. Furthermore, mucosal and serosal applications of forskolin (aden late cyclase activator) restored the I(sc) by 43.9% and 65.3%, respectively. The restored I(sc) was completely abolished by atropine (muscarinic receptor antagonist). These results suggest that VIP may restore the cholinergic activity by increasing the level of intracellular cAMP, and that cholinergic neuron is very likely to be responsible for the regulation of Cl(-) secretion at neuroepithelial junctions. The exact mechanism of VIP's effect on the TTX-inhibited epithelial Cl(-) secretion, and its possible usefulness in the treatment of TTX-induced pathophysiological conditions, remain to be determined.

Phase 1/11 study of bi-weekly irinotecan plus cisplatin in the treatment of advanced gastric cancer.

OBJECTIVES: To conduct a phase I/II study of irinotecan with cisplatin to establish a recommended dose, and assess the safety, efficacy and feasibility of this regimen in unresectable advanced or recurrent gastric cancer. PATIENTS AND METHODS: In the phase I portion of the study, patients received a fixed dose of cisplatin (30 mg/m2) with escalating doses of irinotecan, ranging from 30 mg/m2 to 70 mg/m2, on days 1 and 15. In the phase II portion of the study, 40 patients were evaluated for response and safety at the recommended dose. RESULTS: Eighteen patients were enrolled in the phase I study. Dose-limiting toxicity (diarrhea and neutropenia) appeared at the irinotecan dose of 70 mg/m2. Therefore, the recommended irinotecan dose was 60 mg/m2. In the phase II study, 40 patients received cisplatin (30 mg/m2) plus irinotecan (60 mg/m2). Twenty-five out of 40 patients had received prior chemotherapy. The median number of cycles was 3.5. The response rate was 32.5% (13/40) overall, and 53.3% (8/15) in patients without prior chemotherapy. The median time to tumor progression (TTP) was 162 days. The median survival time was 288 days. Four patients (10%) developed grade 4 neutropenia and 3 patients (7.5%) developed grade 4 anemia. The only observed non-hematological toxicity at grade 3 or higher was diarrhea, seen in 2.5% (1/40) of the patients. CONCLUSION: Bi-weekly administration of irinotecan and cisplatin is safe and active for the management of unresectable advanced or recurrent gastric cancer.

Measuring tumor volume with three-dimensional endoscopic ultrasonography: an experimental and clinical study (including video).

BACKGROUND AND STUDY AIMS: The Olympus EU-IP2 three-dimensional endoscopic ultrasound (3D-EUS) imaging system makes it possible to display tumors in three dimensions and estimate their volume. MATERIALS AND METHODS: Experimental and clinical studies of the volume estimation function of the Olympus EU-IP2 system was carried out to evaluate its accuracy and assess the extent of tumor shrinkage caused by fixation, dehydration, and staining. Results. In the experimental studies, compared with the actual volume of a 1000-mm (3) gelatin column, the estimated volume was found to be equivalent to 114 +/- 1.8 % with the 3R probe and 143 +/- 0.8 % with the 2R probe (mean plus or minus standard deviation). The mean estimated volume of tumor models was 127 +/- 8.5 % with the 3R probe and 131 +/- 6.8 % with the 2R probe. Greater distance from the probe was associated with a greater degree of error than the target object's size, angle, or the number of traces of its outline made. In the clinical studies, compared with the histologically determined tumor volume (100 %), the mean estimated tumor volume was 178 +/- 48.2 % in situ, 168 +/- 31.3 % in resected specimens, and 137 +/- 31.5 % after fixation. Fixation, dehydration, and staining were thus associated with tumor shrinkage. CONCLUSIONS: The volume of gastrointestinal lesions can be estimated by 3D-EUS, although it is overestimated in comparison with actual values. 3D-EUS also allows direct comparisons to be made between the tumor volume before surgery and the volume of fixed pathological specimens, so that the rate of tumor shrinkage can be estimated.

Influence of endoscopic papillary balloon dilation and endoscopic sphincterotomy on sphincter of oddi function: a randomized controlled trial.

BACKGROUND AND STUDY AIM: Endoscopic papillary balloon dilation (EPBD) is assumed to have a theoretical advantage over endoscopic sphincterotomy (EST) in preserving sphincter of Oddi function because it does not involve cutting the biliary sphincter. Although attempts have been made to study the influence of EPBD and EST on sphincter of Oddi function, there is little agreement on this issue. In this study we used a method different from those described in previous reports to clarify whether EPBD or EST preserves sphincter of Oddi function better. PATIENTS AND METHODS: 200 patients with bile duct stones who met the eligibility criteria were randomly assigned to EPBD (n = 104) or an EST (n = 96) group. Sphincter of Oddi function was estimated by measurement of the activity of pancreatic enzymes in the common bile duct (CBD). Pure bile was collected immediately before EPBD or EST, at 1 week and at 1 year after the procedure. We also statistically investigated 14 factors other than EPBD or EST that might have the potential to affect sphincter of Oddi function. RESULTS: There was no significant difference between the baseline characteristics of the EPBD and EST groups. A total of 91 patients (46 in the EPBD group and 45 in the EST group) remained in the trial. Pure bile was collected from the CBD of 86 patients (43 EPBD and 43 EST) 1 week after the procedure. CBD stones were extracted successfully in all cases. Before the procedure, there were no significant differences in the levels of the five pancreatic enzymes between the EPBD and EST groups. At 1 week after the procedure, in both groups, there were significant increases in the levels of the five pancreatic enzymes. At 1 year after the procedure a complete series of pancreatic enzyme analyses was done in 33 patients (12 EPBD and 21 EST). There was no significant difference between the levels of the five pancreatic enzymes immediately before and 1 year after EPBD and EST. When the pancreatic enzyme levels of the two groups were directly compared, there was no significant difference at 1 year after the procedure, but the EPBD group had significantly higher levels at 1 week following the procedure. CONCLUSION: In this study it was found that in patients with CBD stones both EPBD and EST preserve sphincter of Oddi function.

Phase I/II study of S-1 combined with cisplatin in patients with advanced gastric cancer.

A dose-escalation study of cisplatin (CDDP) combined with S-1, a new oral dihydropyrimidine dehydrogenase inhibitory fluoropyrimidine, was performed to determine the maximum-tolerated dose (MTD), recommended dose (RD), dose-limiting toxicities (DLTs), and objective response rate (RR) in advanced gastric cancer (AGC). S-1 was given orally at 40 mg m(-2) b.i.d. for 21 consecutive days following a 2-week rest. CDDP was planned to be given intravenously on day 8, at a dose of 60, 70, or 80 mg m(-2) depending on the DLT. Treatment was repeated every 5 weeks, unless disease progression was observed. In the phase I portion, the MTD of CDDP was presumed to be 70 mg m(-2), because 33.3% of patients (2/6) developed DLTs, mainly neutropenia. Therefore, the RD of CDDP was estimated as 60 mg m(-2). In the phase II portion, 19 patients including six patients of the RD phase I portion were evaluated. The median administered courses was four (range: 1-8). The incidences of severe (grades 3-4) haematological and nonhaematological toxicities were 15.8 and 26.3%, respectively, but all were manageable. The RR was 74% (14/19, 95% confidence interval: 54.9-90.6%), and the median survival day was 383. This regimen is considered to be active against AGC with acceptable toxicity.

Endogenous prostaglandin I2 regulates the neural emergency system through release of calcitonin gene related peptide.

BACKGROUND: We previously reported that endogenous prostaglandin I(2), generated by a mild irritant, sensitised calcitonin gene related peptide (CGRP) containing sensory nerves and facilitated the release of CGRP and gastric mucosal protection against ethanol. Administration of capsaicin also inhibited ethanol induced gastric mucosal injury through immediate release of CGRP from primary sensory neurones, which is termed the neural emergency system. In the present study, we tested whether endogenous prostaglandin I(2) also modulates the cytoprotective action of capsaicin using prostaglandin I receptor knockout mice (IP(-/-)). METHODS: The stomachs of IP(-/-) or their wild-type counterparts (IP(+/+)), anaesthetised with urethane (1.225 g/kg), were doubly cannulated from the oesophageal and duodenal sides, and the gastric mucosa was perfused (1 ml/min) with physiological saline. Perfusate was changed to 50% ethanol alone, or 50% ethanol containing capsaicin (16 approximately 1600 micro M). The injured area was estimated at the end of each perfusion experiment. In some animals, CGRP-(8-37), a CGRP antagonist (0.3 mg/kg), or indomethacin (1 mg/kg) was intravenously injected before perfusion of 50% ethanol containing capsaicin. RESULTS: Capsaicin inhibited the injured area in a dose dependent manner. Fifty per cent ethanol containing capsaicin (480 micro M) immediately increased intragastric levels of CGRP although 50% ethanol alone did not. The protective action of capsaicin (480 micro M) against ethanol was completely abolished by intravenous injection of CGRP-(8-37). Indomethacin also inhibited the protective action of capsaicin, and this was accompanied by reduced levels of intragastric CGRP. Intragastric levels of prostaglandin E(2) were not increased by capsaicin treatment but those of 6-keto-prostaglandin F(1alpha), a metabolite of prostaglandin I(2), were markedly increased. No protective action of capsaicin was observed in IP(-/-) which lacked the ability to increase intragastric CGRP levels in response to ethanol containing capsaicin. The CGRP content of the stomach from untreated IP(-/-) did not differ from those in IP(+/+). Capsaicin (160 micro M) together with intragastric perfusion of beraprost sodium (PGI(2) analogue, 2.5 micro g/ml) showed enhanced protection against ethanol induced injury. This enhanced protection was completely blocked by intravenous injection of CGRP-(8-37). CONCLUSIONS: The present results suggest that endogenous prostaglandin I(2) enhances the protective action of the capsaicin mediated neural emergency system against ethanol induced gastric mucosal injury through enhancement of CGRP release.

Endoscopic mucosal resection in the management of gastric carcinoid tumors.

BACKGROUND AND STUDY AIMS: Gastric carcinoid tumors are a rare disease. Previously, total gastrectomy was regarded as the treatment of choice. However, differences in biological malignancy have recently led to the increased use of endoscopic mucosal resection (EMR) for treatment. We studied the outcome of EMR in patients with gastric carcinoids who were treated at our hospital and discuss the indications for endoscopic treatment. PATIENTS AND METHODS: Between 1986 and 1999 we carried out gastric mucosal resection in five patients with gastric carcinoid tumors. The procedure used for EMR was either strip biopsy or endoscopic aspiration mucosectomy. RESULTS: The carcinoid tumors measured 10 mm or less in four of the five patients. Two patients had type A gastritis, and all had hypergastrinemia. There was no evidence of recurrence during follow-up (range 6 - 66 months; mean 32.6 months). CONCLUSION: EMR is useful in the management of type 1 gastric carcinoids as classified by Rindi (hypergastrinemia; tumor diameter of 10 mm or less).

Serosal application of Ba(2+) induces oscillatory chloride secretion via activation of submucosal cholinergic neurones in guinea-pig distal colon.

The enteric nervous system regulates ion and fluid secretion in the mammalian intestine at both resting and stimulated conditions. To determine the type and activation mechanism of neurones involved, mucosa-submucosa sheets isolated from guinea-pig distal colon were studied in vitro in Ussing chambers. Serosal addition of 0.5-1 mM barium (Ba(2+)), a potassium (K(+)) channel inhibitor, caused oscillatory increases in short-circuit current (I(sc)). Mean values of the size and frequency of I(sc) were 369.1 microA cm(-2) and 2.3 min(-1). The oscillatory I(sc) induced by the low concentrations of Ba(2+) was blocked by either higher concentrations of Ba(2+) (2-5 mM) or other K(+) channel inhibitors, such as tetraethylammonium (TEA) (1 mM) and quinine (20 mM). The Ba(2+)-induced oscillatory I(sc) was also inhibited by tetrodotoxin (TTX) and atropine. In a nominally Ca(2+) free solution plus serosal addition of 0.1 mM ethylene glycol-bis (beta-aminoethyl ether) N,N,N',N'-tetraacetic acid (EGTA), a Ca(2+) chelator, the oscillatory I(sc) slowed and diminished. Further, the Ba(2+)-induced oscillatory I(sc) was partially inhibited by apical addition of 100 microM 5'-nitro-2-(3-phenylpropylamino)benzoic-acid (NPPB), a Cl(-) channel inhibitor, and completely disappeared in a low Cl(-) solution (11 mM) on both sides. On the other hand, application of either cimetidine, a histamine H(2) receptor antagonist, or hexamethonium, a nicotinic antagonist, to the serosal side did not affect the Ba(2+)-induced oscillatory I(sc). In conclusion, the Ba(2+)-induced oscillatory I(sc) is the transepithelial Cl(-) current which is stimulated by activation of cholinergic neurones in submucosal plexus of guinea-pig distal colon.

Thrombotic thrombocytopenic purpura in autoimmune hepatitis.

A 19-year-old woman presented with clinical manifestations of sudden, fulminant thrombotic thrombocytopenic purpura associated with autoimmune hepatitis and autoimmune thrombocytopenic purpura. Although thrombotic thrombocytopenic purpura may, rarely, be associated with systemic lupus erythematosus and other autoimmune diseases, to our knowledge, the syndrome has never been described in association with autoimmune hepatitis. In this patient, too, the etiology of thrombotic thrombocytopenic purpura associated with autoimmune disease remains elusive. The patient was treated with corticosteroid, which brought about no improvement in her condition, and she died of multiorgan failure. Diagnosis is challenging, but prompt diagnosis is necessary because thrombotic thrombocytopenic purpura is a life-threatening syndrome whose prognosis has been improved significantly by early plasmapheresis treatment.

Expression of a specific mucin type recognized by monoclonal antibodies in the rat gastric mucosa regenerating from acetic acid-induced ulcer.

BACKGROUND: Detailed research on the healing process of gastric mucosa from injury due to various necrotizing agents is important for regenerating medicine as well as for estimating the quality of ulcer healing. To elucidate this issue, we prepared monoclonal antibodies reacting with mucin molecules present in the specific region and layer of the gastrointestinal mucosa. METHODS: Acetic acid-induced gastric ulcers were prepared in 8-week-old male Wistar rats. Following 24-h fasting, the animals were killed at 20, 30 and 50 days after acid insult and their stomachs were removed immediately. Serial paraffin sections of the ulcer area were made and immunostained with three distinct monoclonal antibodies. RGM21 and HIK1083 react with mucins derived from the surface mucous cells of the corpus and the gland mucous cells of corpus and antrum, respectively. HCM31 stains sialomucin present in the small intestine and colonic mucosa of rat, but does not react with the intact gastric mucosa, except in the very narrow cardiac gland area. RESULTS: On the 20th day after ulcer preparation, the cells stained with RGM21 and HIK1083 were restricted only to the surface layer and the bottom layer of regenerating epithelia, respectively. HCN31 staining covered most of the regenerating epithelia. On the 30th day, the staining of HCM31 was enhanced in the regenerating area. Staining of RGM21 did not change, but the HIK1083 stained area increased in the lower part of the regenerating epithelia. On the 50th day, the staining with HCM31 weakened markedly in the lower part, and this area was occupied with HIK1083 positive cells. CONCLUSION: A notable but temporary expression of a kind of sialomucin specifically stained with HCM31 was observed in the regenerating epithelia during the healing stage of acetic acid-induced gastric damage.

Obliterative arteritis with nitric oxide synthase and HLA-DR expression in Crohn's colitis.

BACKGROUND/AIMS: To cast light on whether inflammatory vascular injury is a possible pathogenic mechanism in Crohn's disease, the histological characteristics of vascular lesions were investigated. METHODOLOGY: Affected vessels in surgically resected colons from 23 patients with Crohn's disease, 20 with ulcerative colitis, 7 with ischemic colitis, and 9 normal controls were analyzed by Victoria blue and hematoxylin and eosin staining as well as immunohistochemistry for HLA-DR, nitric oxid synthase, vascular endothelial growth factor and E-cadherin. RESULTS: Inflammatory-cell infiltrates affecting arteries, accompanied by obliterative intimal thickening, were more frequent in Crohn's disease cases than in the other groups (P < 0.05-0.0001). Crohn's disease activity was positively correlated with the degree of obliterative arteritis. Granulomatous vasculitis was found exclusively in Crohn's disease (10 cases; 43.5%). In addition, focally enhanced endothelial staining of HLA-DR, with expression in granulomas adjacent to vessels was occasionally observed. In the endothelium of affected vessels, strong expression of HLA-DR was more prevalent in Crohn's disease and/or ulcerative colitis as compared with the ischemic colitis and controls (P < 0.05-0.01). In the involved arteries, enhanced endothelial nitric oxide synthase expression was most common in Crohn's disease among the groups (P < 0.05). A few cases of Crohn's disease, ulcerative colitis and ischemic colitis were positive for inducible nitric oxide synthase, vascular endothelial growth factor or E-cadherin in the vessel walls. CONCLUSIONS: The presence of characteristic obliterative arteritis and granulomatous vasculitis, a possible cause of ischemic injury, supports, in part, a vascular hypothesis for the pathogenesis of Crohn's disease. Enhanced expression of endothelial nitric oxide synthase and HLA-DR possibly reflects compensatory endothelium-mediated vasodilation and amplification of the immune response, respectively.