In July 2004 the patient suffered of an episode of upper gastrointestinal (GI) bleeding of unknown origin at endoscopy, despite the presence of medium sized (F2) oesophageal varices. In the following 18 months she developed frequent diarrhoea, weight loss and severe anaemia, with faecal occult blood test repeatedly positive, reaching haemoglobin concentrations as low as 6.9 g/dl (normal range 13.6–17 g/dl), requiring multiple blood transfusions. Serology for cytomegalovirus (CMV) and Epstein–Barr virus (EBV) infections as well as focal culture and parasitological focal examinations were negative. At this stage maintenance immunosuppression was changed from tacrolimus to mycophenolate mofetil. An upper GI endoscopy confirmed the presence of oesophageal varices F2 and of a single sessile polyp in the stomach, 5 mm in diameter, which was removed. A colonoscopy and terminal ileoscopy were negative. Abdominal ultrasonography revealed slightly enlarged ileal walls (3.5 mm). The patient underwent multiple blood transfusions and intravenous infusion of iron and folates, resulting in only a transient improvement of the anaemia. Because of this clinical picture and a further episode of melena, a capsule endoscopy examination was performed in July 2006, which suggested the bleeding source to be in the jejunum; the recording showed, 17 minutes after starting (presumably in the proximal jejunum), a polypoid lesion, not bleeding at the moment of the exploration and, from the 25th to the 88th minute of recording, fresh blood in the absence of an identifiable mucosal lesion. The patient was therefore referred to undergo an SBE (Olympus, Tokyo, Japan), to confirm the diagnosis.

At admission to our hospital, the physical examination was negative and the haemoglobin concentration was 8.5 g/dl. Based on the results of the capsule endoscopy, the patient underwent push-and-pull enteroscopy (PPE) via the oral approach. The instrument consists of a high resolution videoendoscope 2 m long and a soft flexible overtube with one inflatable balloon attached to its distal end. Due to its extreme flexibility, the tip of the endoscope has a broad bending capability, which allows it to anchor to the small bowel wall without a balloon. Gripping the intestinal wall this way, the endoscope can be inserted further without forming redundant loops in the small bowel. In our patient, enteroscopic examination reached about 200 cm from the pylorus. On enteroscopy, gastric and duodenal mucosa appeared normal, while the small bowel had three polypoid lesions located in the jejunum, 5–15 mm in diameter. However, active bleeding was not observed from any of the polyps during the procedure. Only one polyp had ulcerations, which could account for the recent previous GI bleeding, thus requiring endoscopic treatment (fig 1). Treatment was achieved with an epinephrine injection and thermal coagulation with argon plasma coagulation (APC 2, Erbe, 40 W). No other lesions were detected in the length of small bowel that was explored. A biopsy sample taken from the hyperaemic mucosa of the polyps disclosed a pronounced eosinophilic infiltration (fig 2). Other specimens from the gastric antrum and duodenal bulb had no evidence of eosinophilic infiltration.

With regard to the first point, based on the clinical manifestations and gut wall involvement, three types of eosinophilic enteritis have been defined: mucosal disease, muscle disease, and serosal disease.⁷ The differential diagnosis requires distinction from Whipple’s disease, amyloidosis, lymphangiectasia,⁷ and Crohn’s disease.⁸ When eosinophilic infiltration is profound, resulting in rigid loops, it may mimic lymphoma and small bowel adenocarcinoma.⁹ According to Talley⁷ a diagnosis of eosinophilic GI disorder can be made when the following four criteria are fulfilled: (1) presence of GI symptoms; (2) demonstration of eosinophilic infiltration of the GI tract; (3) absence of eosinophilic involvement of other organs; and (4) absence of parasitic infection. All these criteria were fulfilled by the patient described in the present report, with the only exception of a concomitant eosinophilic liver involvement in the setting of a chronic allograft rejection. To our knowledge, the occurrence of eosinophilic enteritis has not yet been described in liver transplant recipients. It may be of significance that our patient received interferon α treatment because of recurrent post-transplant hepatitis C before the development of symptoms, later attributed to eosinophilic enteritis. Notably, eosinophilic enteritis has been already reported only in a single non–transplanted patient receiving interferon α therapy for hepatitis C,¹⁰ which suggests, however, that interferon induced immune stimulation may be capable of triggering an eosinophilic reaction. This matter is further complicated in liver transplant recipients by the need for maintenance immunosuppression to control cellular rejection. Our patient had histological evidence of an acute rejection episode (which determined peg-interferon withdrawal), as well as of initial chronic rejection, suggesting that the localised eosinophilic enteritis could have been an epiphenomenon of a more extended, systemic reaction. This hypothesis is reinforced by the fact that GI symptoms (diarrhoea and bleeding) worsened after May 2005, when the patient was switched from tacrolimus to mycophenolate treatment—in other words, when she was moved to a lower degree of maintenance immunosuppression. This case report therefore suggests that the development of unexplained anaemia and occult or overt GI bleeding of unknown origin in a under-immunosuppressed liver transplant recipient may indicate the presence of jejunal eosinophilic enteritis.

With regard to the second point, this case report emphasises the value of capsule endoscopy and of push enteroscopy in the diagnostic work-up and clinical management of this rare condition. In fact, radiographic changes associated with eosinophilic enteritis are variable and usually non-specific.^7,9 Capsule endoscopy, a non-invasive modality used to evaluate disorders of the small bowel, has already been used to evaluate intestinal complications caused by eosinophilic enteritis,¹¹ yet its use is limited due to the inability to obtain biopsy specimens. PPE allows specimen withdrawal, but it only spans the proximal jejunum.¹² Newly introduced double balloon enteroscopy (DBE)^1–4,13 and SBE display higher diagnostic yield and greater therapeutic possibilities compared to other diagnostic modalities. In the present case, SBE proved to be a useful technique for histological confirmation of the diagnosis, with no associated complications. A prompt diagnosis of this disorder avoids unnecessary laparoscopy and unnecessary treatments. To our knowledge, this is the first case of SBE diagnosis and treatment of eosinophilic enteritis presenting with bleeding arising from polypoid lesions of the small bowel.

In conclusion, eosinophilic enteritis is a rare disorder and can be difficult to diagnose, especially when accompanied with intestinal bleeding. We found SBE a useful tool for the timely and prompt diagnosis and treatment of this disorder.