Chronic Intestinal Pseudo-Obstruction Associated with Cytomegalovirus Infection in an Infant

Chronic intestinal pseudo-obstruction (CIP) comprises a heterogeneous group of disorders that result in ineffective intestinal contraction that produces symptoms and signs of intestinal obstruction in the absence of a mechanical lesion (1,2). The disorders vary in cause, severity, course, and response to treatment (3-5). As in adults, various forms of CIP have been recognized in the pediatric population (4,5). The pathology of CIP is still poorly understood. Cytomegalovirus (CMV) infection of the digestive tract is usually seen in immunodeficient patients; however, since 1964 several studies have reported GI disease due to CMV in patients without detectable immunodeficiency (6). Gastrointestinal CMV lesions are usually characterized by ulcerations, erosions, and mucosal hemorrhage. We herein present a case of CIP associated with CMV infection in an infant, which suggested a pathogenic role of CMV in the occurrence of CIP. CASE REPORT A 2-month-old white girl was admitted to our hospital in February 1992 because of a 3-week history of abdominal distension and vomiting. She was the first baby of nonconsanguineous parents, and she was born at term without any complication (weight, 2,950 g; length, 50 cm). When the child was 2 weeks of age, diarrhea occurred (10-12 stools/day), along with vomiting, weight loss, and major abdominal distension. On admission, the patient was dehydrated and had failure-to-thrive; her weight was 2,610 g (- 2 SD) and her length was 50 cm (- 1 SD). She also had respiratory signs characterized by cough and dyspnea without fever. Complete intestinal obstruction with several air-fluid levels was found on x-ray examination. Repeat chest x-ray film showed diffuse interstitial pneumonia. Mechanical obstruction was excluded by normal small bowel followthrough and barium enema. Total parenteral nutrition (TPN) by central venous catheter (Broviac) was initiated 3 days after admission because of patient intolerance to enteral nutrition; she was given a caseinhydrolyzed, lactose-free, MCT-enriched formula. Despite treatment, there was persistent abdominal distension. Anal manometric study showed normal rectoanal inhibitory reflex. Ileostomy was performed 12 cm proximal to ileocecal junction. Hirschsprung's disease was excluded by normal rectal surgical biopsy and appendiceal biopsy. Cystic fibrosis was also excluded by a normal sweat test. Thyroid hormone levels were normal, and the patient had had no immunodeficiency; she was HIV-negative and immunoglobulin levels were normal for age. A presumptive diagnosis of CIP was established. Thirty days after ileostomy, 4-h fasting antroduodenal manometry was performed by a catheter system with a low-compliance hydraulic capillary infusion pump connected to external pressure transducers. We observed a dysmotility pattern characterized by the absence of migrating motor complexes. On rectal and ileal biopsy specimens, there was no visceral myopathy or neurological alterations. The results of Smith's silver techniques were also normal. Because of the interstitial pneumopathy, CMV infection was suspected. There was no respiratory failure during the whole clinical course of the disease; thus, treatment by oxygen or mechanical ventilation was not necessary. CMV was cultured from an endoscopic small bowel biopsy performed when the patient was 2.5 months of age. The presence of IgM anti-CMV antibodies (1:80) at this age suggested recent infection; moreover, the titer of IgM anti-CMV antibodies rose to 1:480 at 3 months of age and fell to zero at 6 and 8 months of age. The pattern of IgG anti-CMV antibodies in our patient was as follows: none at 2 months of age, 1:160 at 3 months of age, and 1:1,280 at 6 and 8 months of age. At 6 and 8 months, the level was higher than the mother's level (1:960). IgM anti-CMV antibodies were not present in the mother (Fig. 1). We did not collect a sample from the mother for virus isolation, and no antenatal sample was available. Urine culture tests were intermittently positive for CMV in our patient. There was no viremia during the whole clinical course, and there was never any evidence of CMV cells with basophilic intranuclear inclusions on histopathological examination of the GI tract. Results of immunohistochemistry analysis were also negative. Moreover, there was no immaturity of ganglionic cells. During hospitalization, the patient had several occlusion episodes with no evidence of bacterial overgrowth (<105/ml of bacteria in duodenal fluid). No antiviral therapy was given. Jejunostomy was done at 2.5 months of age for occlusion despite ileostomy, and TPN was continued (Fig. 1). When the patient was 6 months of age, spontaneous remission occurred and the jejunostomy could be closed. Oral feeding with human milk was progressively introduced and well-tolerated. TPN was stopped when the patient was 8 months of age, and the ileostomy was closed at 9 months of age. The patient was discharged from the hospital in September of 1992. At the age of 42 months, this child remains well. DISCUSSION To the best of our knowledge, there has been only one case report associating CIP and CMV infection; Foucaud et al. described this combination in a 2-month-old boy (7). In that case, rectal biopsy revealed hypoganglionosis with thickening of nerve processes and intranuclear inclusions in some neurons. Suspected CMV infection was confirmed by viruria and specific IgM and IgG antibodies (7). The symptoms were much more severe in our patient than in the Foucaud et al. patient. The major difference between the two was the presence in our patient of signs of intestinal obstruction, leading to abdominal surgery. Because most patients previously exposed to CMV harbor latent virus in various organs with no evidence of organ damage, infection with CMV is more common than disease caused by CMV. A reasonable definition of CMV. intestinal disease is a process in the wall of the gut in which the presence of CMV is shown by routine histological examination, culture, or DNA staining in a patient in whom other diseases have been excluded (8). Our case fulfilled these criteria. The presence of IgM antibodies does not rule out maternofetal transmission of CMV. Nevertheless, the presence of CMV-specific IgM at a titer of 1:80 and the absence of maternal IgG when the child was 2.5 months old were indicative of infection within the first few weeks of life rather than congenital infection. This view was further supported by the subsequent rise and fall in IgM levels during the following months and the development of a significant IgG response. The characteristics of gastrointestinal CMV disease have been well-described in acquired immunodeficiency syndrome (9,10), organ transplant recipients (11,12), inflammatory bowel disease (13), during steroid therapy, and sometimes in healthy subjects (6). Of the different methods used to diagnose GI cytomegalovirus disease, serologic analysis with the presence of IgM anti-CMV antibodies suggests recent infection but does not establish the diagnosis of tissue-invasive disease. A positive test from blood or urine culture does not prove that GI signs or symptoms are due to CMV disease. In the same way, negative cultures from these sites do not exclude gastrointestinal CMV disease (14-16). However, multiple positive cultures from distant sites strongly suggest the diagnosis of gastrointestinal CMV disease when GI biopsy specimens are not available or negative (8). Most studies have confirmed the specificity of cytomegalic cells for diagnosing CMV because these cells were always associated with the presence of CMV antigen (17,18). In some proven gastrointestinal CMV disease, cytomegalic cells may be rare and very difficult to detect, requiring great time and effort by the pathologist. This lack of sensitivity is considered a disadvantage of routine histologic examination (8). Although cultures of intestinal biopsy specimens were positive in our patient, no cytomegalic cells could be found. Several techniques have been used to enhance diagnostic sensitivity, including immunochemical staining of histologic specimens and polymerase chain reaction to identify CMV DNA. However, even if not very sensitive, the presence of cytomegalic cells is more specific (8). In our case, the lack of histological abnormalities of the intestinal mucosa does not rule out the diagnosis of CIP according to the classification of Navarro et al. (19). Because gastrointestinal CMV disease in immunologically normal hosts is usually self-limited, no antiviral therapy was given to our patient, as suggested by Surawitz and Meyerson (20). Spontaneous recovery was indeed observed in our patient. We conclude that, in the presence of CIP, CMV infection should be investigated. Based on our patient and the patient described by Foucaud et al. (7), the prognosis of such a CMV infection seems to be good, as in immunocompetent patients. Nevertheless, nutritional support, either parenteral or enteral, may be transiently necessary.FIG. 1.: Clinical characteristics and evolution of anti-CMV antibodies.