Introduction

Behçet’s disease (BD) is a multisystem, chronic, relapsing vasculitis of unknown origin that affects nearly all organs and systems. While recurrent oral ulcerations are a “sine qua non” of BD, the frequency of extra-oral involvement of the gastrointestinal system varies widely in different countries. The most frequent extra-oral sites of gastrointestinal involvement are the ileocecal region and the colon. There are rare reports of gastric and duodenal ulcers in patients with BD in scientific journals. We had experience in treating a BD patient with repeated gastric ulcers in 1999. He was diagnosed as having Behçet’s disease in 1996 with manifestations of recurrent oral ulcers, genital ulcers, erythema nodosum, and folliculitis. Hematemesis, massive tarry stool, high fever, and oral ulcer were noted and required repeated transfusion. An endoscopy examination revealed active gastric ulcers. Due to treatment failure with H2 blockers in several previous bleeding ulcer episodes, methylprednisolone pulse therapy, followed by cyclophosphamide, was prescribed for rescue of active ulcer bleeding. Dramatic improvement was noted and no new gastric ulcer has been observed since then. To further evaluate the extent of gastric/duodenal ulcer involvement in Taiwan, BD patients diagnosed from 1999 to 2002 at our hospital were enrolled in this current study.

Material and patients

Patient with Behçet’s disease at rheumatology clinics of Dalin Tzuchi hospital from 1999 to 2002 were included in this study. BD was diagnosed according to the International Study Group criteria (ISG) [1]. A diagnosis of BD requires recurrent oral ulcerations plus two of the following additional findings: (1) recurrent genital ulcerations, (2) eye lesions consisting of an iritis, posterior uveitis, renal vessel occlusion, and/or optic neuritis, (3) skin lesions consisting of folliculitis, erythema nodosum, an acne-like exanthema, or a migratory thrombophlebitis, or (4) a positive pathergy test, which is an inflammatory skin reaction to an intradermal injection of saline. After admission, we obtained and recorded clinical and laboratory data including routine blood and urine analyses, erythrocyte sedimentation rate (ESR), C-reactive protein, and serum immunoglobulins from these patients. For evaluation of gastroduodenal involvement of BD in this local area, a diagnostic endoscopy was arranged for all patients. Helicobacter pylori infection by urease test was performed during endoscopy examination. Some patients underwent a second endoscopy examination 3 months later after treatment with systemic steroids (prednisolone 10 mg b.i.d. and cyclophosphamide 50 mg/day). Colonoscopy and barium enema were not routinely arranged but only for those patients with clinical manifestations of low abdomen pain or bloody stools. Considering the possibility of genetic loads in the pathogenesis of ulcer development, HLA tissue antigens were determined by a DNA typing technique with sequence-specific primers (SSP) (HLA-A.B.C.DR.DQ kit, Dynal Allset SSP, Dynal Biotech, Bromborough, Wirral, UK)

Results

Gastrointestinal ulcer was a relatively common manifestation in Chinese patients with BD in Taiwan. A total of 28 patients with BD were enrolled in this study. There were 14 patients who suffered from extra-oral gastrointestinal involvement determined by routine endoscope examinations (including 6 patients with combined gastric and duodenal ulcers, 3 with gastric ulcers, 3 with duodenal ulcers, 1 with anal ulcer, and 1 with a colon linear ulceration). The frequency of extra-oral GI involvement was 50% in our series (which might be an underestimate because no routine evaluation of rectocolon lesions was performed in this study). Most of our ulcer patients complained of dyspepsia, hunger pain, and epigastric pain. Two of them (patients 3 and 10) suffered severe hematemesis and bloody stools. Ten patients had previous episodes of peptic ulcers. The incidence of rectocolon lesions might be underestimated because routine colonoscopy examination was not arranged in the current study. No patient was found to have esophageal mucosa ulcerations during the examination.

Nine patients in this study had used a nonsteroidal anti-inflammatory drug (NSAID) intermittently for control of low back pain and joint pain (Table 1) at other local clinics. Only two of our ulcer patients had a history of NSAID use [patient 11 with meloxicam 7.5 mg/day for 6 months—the frequency in our series was at least 50% (no routine colonoscopy examination was performed)—and patient 20 with meloxicam 7.5 mg/day for 12 months]. It was noted, however, that duodenal ulcer and not gastric ulcer was found in patient 11, possibly not related to NSAID-induced gastric ulcer. Furthermore, only one of our ulcer patients (patient 26) had H. pylori infection.

Table 1 Demography of patients with Behçet’s disease. EN erythema nodosum, GU gastric ulcer, DU duodenal ulcer, DVT deep vein thrombosis, ND not done, patho microscopic examination of Helicobacter infection, CLO urease test for Helicobacter infection, MHC major histocompatibility complex
Full size table

In the current study, all of our BD patients had recurrent oral ulcers (100%). Twenty patients (71%) suffered from genital ulcers and 78% of our BD patients had one or more cutaneous lesions (folliculitis, erythema nodosum, and papulopustular lesion). On the contrary, only two patients had a history of iridocyclitis/uveitis. Besides, active inflammatory signs were also present simultaneously during episodes of gastric/duodenal ulcers. Leukocytosis, fever, neutrophilia, elevated ESR, and active cutaneous lesions were most frequently noted.

It was found that 7 of 12 patients (58%) with gastric/duodenal ulcerations carried an A2/B46/Cw1 or A11/B46/Cw1 genotype. On the other hand, none of the other 16 BD patients without gastric/duodenal ulcers had such genotypes. HLA-B51 genes were present in only five of our BD patients, and only one patient developed a gastric ulcer.

Discussion

Several criteria have been formed since 1969 in an attempt to aid in the diagnosis of BD. The various criteria include Mason and Bames, Japanese, O’Duffy, Zang, Dilsen, and most recently the International Study Group criteria [1]. However, there is no universally accepted definition of BD. The same patient may be classified as having complete BD by one set of criteria, while being considered to have an incomplete form of the disease by another set. The discriminatory performance in combination and in comparison with all other existing criteria sets are summarized by Kaklamani [2]. There are data that suggest that regardless of whether the BD is complete or incomplete, according to the criteria, the natural history of the disease is the same [3]. The ISG criteria compare favorably with the other criteria sets with a sensitivity 91%, specificity of 96%, and relative value of 187. All of our BD patients had oral ulcers (100%). Of 28 patients, 20 (71%) suffered from genital ulcer and 78% of our BD patients had one or more cutaneous lesions (folliculitis, erythema nodosum, and papulopustular lesion). The percentage of all the above of manifestations is similar to those described in the ISG criteria for BD. However, only two of our BD patients exhibited uveitis.

The frequency of gastrointestinal involvement varies in different countries, with a lower frequency in Turkey (2.8–5%), India (5.2%), and Israel (0%); a moderate frequency in France (14%), England (14%), Kuwait, and the United States (21%); and the highest frequency in Scotland (50%) and Japan (50–60%). The frequency in our series was at least 50% (no routine colonoscopy examination was performed). As for intestinal BD, some suggest it is a distinct entity of its own [4]. Ulcerations of the GI tract can be found throughout the intestine, but the most frequent area is the ileocecal region with extension to the ascending colon. Since this was not the major purpose of the current study, the true prevalence of ileocecal ulcers was not accurately estimated and needs further observations.

Compared to other parts of the gastrointestinal tract, the gastric mucosa appears to be the least frequently involved segment. Case reports of a Dieulafoy’s ulcer [5] and a gastric non-Hodgkin’s lymphoma associated with BD [6] have been described. Aphthous ulcers can occur in the duodenum. In two large autopsy series, a total of six patients with BD were found to have duodenal ulcers [7, 8]. Two cases of duodenal involvement in living patients have been reported [9, 10]. On the contrary, the prevalence of combined gastric/duodenal ulcers (6 of 28), simple duodenal ulcers (3 of 28), and simple gastric ulcers (3 of 28) was significantly higher in Chinese patients compared with previous reports.

In such a high prevalence of gastric/duodenal involvement of BD patients in this local area, it was found that 58% of our ulcer patients carried an A2/B46/Cw1 or A11/B46/Cw1 genotype (less than 4.6% from data of Tzu Chi Taiwan Marrow Donor Registry with over 150,000 prospective donors). On the other hand, none of the other 16 non-ulcer patients had these genotypes. It is possible that the presence of these distinct genotypes in patients might contribute to the development of gastric/duodenal ulcers in Taiwan. The role of HLA-B5(51) has been studied extensively during the past few years [11, 12]. In a recent report [13], allele HLA-B*5101 was shown to predispose to more severe diseases, such as uveitis and erythema nodosum. Japanese investigators [14] found that the BD gene is located near the HLA-B gene, but is not the HLA-B51 gene itself. Recent work suggests that polymorphism in tumor necrosis factor (TNF) and the MICA genes are noncontributory to the genetic load in BD. HLA-B51 genes were present in only five of our patients, and only one ulcer patient showed the B51 gene. It is suggested from our data that the HLA-B51 gene alone is not a contributory gene in BD and is not associated with the development of gastric/duodenal ulcerations in Chinese in Taiwan.

Only a few of our ulcer patients had risk factors of NSAID and H. pylori infection (two patients had histories of NSAID use but with negative survey in H. pylori infection; one patient had H. pylori infection but with NSAID history). Two studies assessing the frequency of H. pylori infection in Turkish patients with BD have been reported [15, 16]. In the first, the urease positivity rate was 65% in 34 patients with BD [15]. In the second report, a higher prevalence (85%) was noted [16]. In our report, only one patient showed evidence of H. pylori infection (patient 26). Furthermore, only two of our ulcer patients had histories of NSAID usage (patients 11 and 20). Data of NSAID history and H. pylori infection did not support a correlation between ulcer formation and these risk factors. It was suggested that vasculitis rather than H. pylori infection or NSAID might be responsible for the high prevalence of gastric/duodenal ulcers in our patients.

In conclusion, we have demonstrated that peptic ulcer is a common manifestation in Chinese BD patients in Taiwan, although rarely mentioned in other populations. A close relationship between A2/B46/Cw1 (or A11/B46/Cw1) and the development of gastric/duodenal ulcers was found. Successful treatment of refractory gastric ulcers with steroid and cyclophosphamide might suggest that the disease itself, rather than NSAID or H. pylori, was responsible for the development of gastroduodenal ulcer in BD disease.