Use of azathioprine for nongranulomatous ulcerative jejunoileitis

Multiple nonmalignant ulcerations of the small bowel are rare but frequently fatal complications of celiac sprue (1). The etiology of this ulceration is unknown. Patients with nongranulomatous ulcerative jejunoileitis (NGUJI) present with gastrointestinal complications including abdominal pain, malabsorption, bleeding, and small intestinal obstruction or perforation, despite adherence to a glutenfree diet. Therapy for this disorder includes initiation or maintenance of a strict gluten-free diet, resection of localized disease, total parenteral nutrition (TPN), and corticosteroid therapy in refractory cases. Even in patients treated aggressively, however, the mortality rate may approach 75% (2). Although cyclosporine and azathioprine have been used occasionally for the treatment of refractory sprue, they have not been used in patients with ulcerative jejunoileitis (3-5).

We present the first documented case of a patient with presumed celiac sprue who developed NGUJI that was treated successfully with azathioprine.

CASE PRESENTATION
A 56-year-old Caucasian woman presented with diarrhea and weight loss.Her stools were bulky, floating and pale.She had been thin all her life but was now feeling weak and losing weight despite adequate food intake.She had no history of abdominal pain or vomiting.Medical history included recurrent duodenal ulcers (documented on a barium study 10 years before presentation) and osteoarthritis.Her only medication was amitriptyline for fibromyalgia.Physical examination demonstrated an alert woman weighing 50 kg and 1.67 m tall.Other than muscle wasting, there were no abnormalities on examination.Findings of low hemoglobin, low iron, Howell-Jolly bodies on peripheral blood smear, elevated 24 h fecal fat, low magnesium and low albumin (Table 1) led to a presumptive diagnosis of celiac disease.A barium study showed diffuse edema of the small bowel.A small bowel biopsy (Figure 1) was consistent with the diagnosis of celiac disease, and the patient was placed on a gluten-free diet.
Despite a temporary improvement, she relapsed and was started on prednisone 20 mg/day with resolution of symp-toms.Over the next 10 years, she remained dependent on steroids at doses ranging from 7.5 to 20 mg/day.Despite prednisone therapy, she relapsed again and was hospitalized for weight loss, diarrhea and abdominal pain.Laboratory investigations are shown in Table 1.An upper gastrointestinal barium study showed a duodenal ulcer and mucosal thickening with slight irregularities on the surface of the mucosa in the jejunum.Serum gastrin was 42 ng/L (normal 0 to 100 ng/L).She was not taking potassium supplements or nonsteroidal anti-inflammatory medications.A computerized tomography scan demonstrated fluid-filled loops of small bowel, with enhancement of the wall of the bowel loops suggesting inflammation.An indium-labelled white blood cell scan showed diffuse, abnormal accumulation of radioactivity within the small bowel 4 h after injection.At gastroscopy, multiple small bowel ulcerations were visualized distal to the duodenum and were biopsied.Histology consisted of slight villous atrophy, inflammatory infiltration of the lamina propria and areas of ulceration of the villi.
The patient was started on TPN, and her prednisone dose was increased to 30 mg/day.To exclude the possibility of lymphomatous disorder, a laparotomy was performed.Several enterostomies were done intraoperatively in edematous, thickened small bowel, but only diffuse small bowel ulceration was seen with no evidence of lymphoma.A 4.5 cm seg-  ment of jejunum was resected and showed multiple small (0.5 cm) mucosal ulcerations extending through the muscularis mucosa into the underlying submucosa (Figure 2).Azathioprine 50 mg/day was given in addition to prednisone 30 mg/day.The patient remained on azathioprine for the next 42 months with marked improvement in her clinical and biochemical profile (see Table 1).Her hemoglobin, carotene and albumin normalized, and an indium scan and enteroscopy were normal.She was able to taper the prednisone to 7.5 mg/day and subsequently discontinue the azathioprine.Now, eight years following diagnosis of NGUJI, she is asymptomatic while receiving only vitamin D, calcium and prednisone 7.5 mg/day and maintaining a gluten-free diet.

DISCUSSION
Because of the histological (villous atrophy) and clinical (malabsorption) similarities, NGUJI is often considered to be a form of gluten-sensitive enteropathy.In many cases, the disorder appears to develop in compliant celiac patients with stable disease, but in other cases, the disorder occurs de novo, ie, in people with no known pre-existing disease.This leads some investigators to believe that it is a completely separate entity from celiac disease (6).The patient described in the present study had abnormal sucrose and lactulose/mannitol permeability studies indicative of a small bowel disorder (7-9) but lacked both endomysial antibodies and the typical human leukocyte antigen typing, giving support to the argument that her disorder is different from celiac sprue (10).
Of 70 previously reported patients with NGUJI, 51% had an underlying disease (6), the most common being celiac disease.The differential diagnosis of NGUJI includes hypogammaglobulinemia, Zollinger-Ellison syndrome, AIDS, Crohn's disease, ischemia, lymphoma, drug toxicity (eg, potassium tablets or nonsteroidal anti-inflammatory medications) and vasculitis (11,12).This patient had normal immunoglobulins; a normal gastrin level; a negative HIV test; no granulomata on her full-thickness biopsies; and no evidence of ischemia, lymphoma or vasculitis.She was not taking potassium or nonsteroidal anti-inflammatory medications.These facts support the diagnosis of NGUJI.
The pathogenesis of this disorder is unknown but it is presumably immunologically based involving antibodies with broad antiepithelial cell specificity, which leads to increased synthesis of immunoglobulin A, which is refractory to gluten withdrawal (10).The use of prednisone to treat NGUJI dates back to 1967 (1,13).The reported results using prednisone in small case series have been conflicting, with a 15% long term response reported by Baer et al (2) and a 78% long term benefit reported recently by Ruan et al (6).The benefit of corticosteroids may be related to suppression of elevated immunoglobulin A levels that are refractory to gluten withdrawal (unlike celiac patients without NGUJI) (10).The overall mortality rate is 50% to 75% (2,6).There are no previous reports of using azathioprine in the management of refractory NGUJI.This patient had required high dose prednisone for only partial relief of symptoms, and the long term side effect profile of corticosteroids is ominous -the use of azathioprine was therefore elected.Her symptoms, biochemical markers and radiological studies all demonstrated marked improvement with azathioprine.Perhaps, through T helper cell inhibition (5,14), azathioprine played a vital role in suppressing the inflammatory response in this patient with NGUJI.