Rectal bleeding is an alarming symptom and requires additional investigation. In infants it has been explained mainly by hypersensitivity. In addition to dietary antigens, intraluminal microbial agents challenge the immature gut mucosa. Although controlled in the mature gut, these antigens may induce inflammation in the developing gastrointestinal tract. The objectives of this study were to evaluate prospectively the clinical course of rectal bleeding and evaluate the impact of cow's milk allergy and aberrant gut microbiota on the condition. Because withdrawal of cow's milk antigens from the infants' diet is used as a first treatment without evidence of its efficacy, we also aimed to asses the effect of a cow's milk-elimination diet on the duration of rectal bleeding.
The study involved 40 consecutive infants (mean age: 2.7 months) with visible rectal bleeding during a 2-year period at the Tampere University Hospital Department of Pediatrics. Most of the infants (68%) were fully breastfed. At enrollment the infants were randomly allocated to receive a cow's milk-elimination diet (n = 19) or continue their previous diet (n = 21) for 1 month. Findings of colonoscopy, fecal bacterial culture, fluorescence in situ hybridization of selected gut genera, specific detection of fecal enteroviruses, rotaviruses, and adenoviruses, fecal electron microscopy for viruses, and mucosal electron microscopy for viruses were assessed. During each visit the severity of atopic eczema, if any, was assessed according to the SCORAD method. In evaluating the extent of sensitization, serum total immunoglobulin E (IgE) and specific IgE and skin-prick tests for cow's milk, egg, and wheat were studied. Cow's milk allergy was diagnosed by elimination and provocation testing. Five patients were hospitalized; all others were treated on an outpatient basis. The follow-up visits were scheduled 1 month later and at the age of 1 year. Sixty-four healthy reference infants were selected as controls according to the following criteria: age and timing of fecal sampling being identical to within 1 month.
Altogether, 32 (80%) infants manifested bloody stools during follow-up (mean [range]: 2.1 [1-15] per day). The mean number of days with rectal bleeding on follow-up was 6. Typically, bloody stools occurred irregularly, for which reason the mean time to the last occurrence of rectal bleeding was 24 (range: 1-85) days from admission. Atopic eczema at presentation or during follow-up was diagnosed in 38% of the infants. Increased specific IgE concentrations or a positive skin-prick test were uncommon. The growth of the infants was normal on admission and during follow-up. Colonoscopy revealed typically focal mucosal erythema and aphthous ulcerations. The mucosa appeared normal in less than half of the patients. No anorectal fissures or colonic polyps were found. Light microscopy revealed that the overall architecture of the mucosa was well maintained. Acute inflammation or postinflammatory state and focal infiltration of eosinophils in the lamina propria were the most common abnormalities. A cow's milk-elimination diet did not affect the duration of rectal bleeding. Cow's milk allergy was diagnosed in 7 (18%) patients. Virus-particle aggregates were found in the microvillus layer of the colon epithelium in 8 cases. The surface epithelium of the virus-positive colon biopsy specimens regularly showed degenerative changes in the microvillus layer and epithelial cells. Electron microscopy study of the colon biopsies disclosed virus particles (30 nm in diameter) on the surface of epithelial cells. Virus particles or RNA were present in feces in only a minority of the patients. All fecal cultures were negative for Salmonella, Shigella, and Yersinia. Campylobacter jejuni was found in the feces of 1 patient, and fecal cultures were positive for Clostridium difficile in 4 patients, Staphylococcus aureus in 8 patients, and yeast in 2 patients. Fluorescence in situ hybridization revealed that at the time of admission the total numbers of bacteria and the numbers of bifidobacteria and lactobacilli in feces were lower in the patients compared with controls. The fecal concentrations of microbes characterized in this study (Bacteroides, bifidobacteria, Clostridium, lactobacilli, and enterococci) did not differ significantly between the time of admission and the second visit in the patients or controls. At the age of 1 year, 7 patients still suffered from cow's milk allergy, 5 of whom also suffered from multiple food allergies. Atopic eczema and histopathologically confirmed inflammation of the colonic mucosa at presentation were associated with persistence of cow's milk allergy at the age of 1 year. No patients exhibited gastrointestinal complaints or visible blood in stools.
Rectal bleeding in infants is generally a benign and self-limiting disorder. Bloody stools occurred irregularly for only a few days during the following months. As in a previous report, most infants were exclusively breastfed. In the majority of the patients the cause of the condition remains unknown. An association with viruses can be seen in some patients. The microbes that commonly lead to bloody diarrhea in older children and adults, Salmonella, Shigella, and Yersinia, were absent in the present material. The low bifidobacterial numbers in fecal samples may indicate a significant aberrance that may provide a target for probiotic intervention to normalize gut microbiota. The gut microbiota overall seemed stable, because the numbers of major groups of microbiota tested did not change significantly between the time of admission and after 1 month. Cow's milk allergy among these patients is more uncommon than previously believed. Cow's milk challenge is thus essential in infants who become symptom-free during a cow's milk-free diet to reduce the number of false-positive cow's milk-allergy diagnoses.