Dietary Protein Proctocolitis

Eosinophilic proctocolitis is characterized by the presence of mucusy, bloody stools in an otherwise healthy infant. The disorder is attributed to an immune response directed, most commonly, against cow’s milk protein. The mean age at diagnosis is approximately 60 days, with a range of 1 day to 6 months. The bleeding is often mistakenly attributed to perirectal fissures, although bleeding associated with fissures tends to present with streaks of blood on hard, formed stool rather than mixed in frothy, mucusy stool, which is typical of proctocolitis. Failure to thrive is absent. About 60% of cases occur in breastfed infants where the immune response results from maternal ingestion of the food allergen, usually cow’s milk, which is passed in immunologically recognizable form into the breast milk. In formula-fed infants, the reaction is associated with cow’s milk or, less commonly, soy. Proctocolitis has rarely been described in infants fed hypoallergenic, extensively hydrolyzed formulas. Associated peripheral blood eosinophilia, hypoalbuminemia and/or anemia are uncommon. Markers of atopy such as atopic dermatitis or a positive family history of atopy are not significantly increased compared with the general population.

Endoscopic examination is usually not needed for diagnostic purposes but, when performed, shows patchy erythema, friability and a loss of vascularity generally limited to the rectum. High numbers of eosinophils (5 to 20 per high-power field) or eosinophilic abscesses are seen in the lamina propria, crypt epithelium and muscularis mucosa. The eosinophils are frequently associated with lymphoid nodules (lymphonodular hyperplasia) and rarely with granuloma formation. However, lymphonodular hyperplasia is not unique to this condition. The pathophysiology is unknown. Because inflammation is confined to the lower colon and is common in breastfed infants, it has been hypothesized that dietary antigens complexed to breast milk IgA may play a part in the activation of eosinophils and the distribution of the inflammatory process.

The frequency of food allergy causing rectal bleeding in infants has not been extensively studied. Xanthakos and colleagues performed colonoscopy and biopsy on 22 infants presenting with rectal bleeding, and proved eosinophilic colitis in 14 (64%). The remainder had normal biopsies (23%) or nonspecific colitis (14%). This group recommended dietary elimination for those with eosinophilic colitis and the majority had resolution within 1 to 3 weeks. However, the relationship of cow’s milk protein to symptoms was not proven by rechallenge. Arvola and colleagues examined 40 infants presenting with rectal bleeding. Infants were randomized to either avoid cow’s milk protein or maintain their current diet. The duration or severity of bleeding was no different between the two groups. During follow-up, cow’s milk allergy was diagnosed in 18% of the infants (based upon various criteria including flares of atopic dermatitis and urticaria upon food challenge as well as rectal bleeding) and for these infants, there was a reduced length of bleeding when they had been randomized to an elimination diet at study outset. Atopic dermatitis and inflammation of the colonic mucosa were associated with persistence of cow’s milk allergy to the age of 1 year. These studies indicate that food allergy may not be a common cause of rectal bleeding in infants unless there are additional signs of allergy, and that milk protein-induced proctocolits has to be differentiated from benign idiopathic neonatal transient eosinophilic colitis.

Food Protein-Induced Enterocolitis Syndrome (Dietary Protein Enterocolitis) ( Box 44-1 )

Food protein-induced enterocolitis syndrome [FPIES]) describes a symptom constellation of profuse vomiting, lethargy and diarrhea, usually diagnosed in the first months of life and most commonly attributable to an immune response to cow’s milk or soy. Inflammation involves both the small and large bowels. Unlike allergic proctocolitis, the majority of affected infants are asymptomatic while exclusively breastfed on an unrestricted maternal diet. When the causal protein remains in the diet (e.g. in young infants fed with cow’s milk or soy-based formulas), chronic symptoms can include watery or bloody diarrhea, poor growth, anemia, hypoalbuminemia and fecal leukocytes; the illness may progress to dehydration and hypotension over the course of days to weeks. Removal of the causal protein leads to resolution of symptoms but re-exposure results in a characteristic delayed (by about 1 to 3 hours) onset of repetitive, often projectile vomiting, lethargy, elevation of the peripheral blood polymorphonuclear leukocyte count and possibly reduced temperature, thrombocytosis, hypotension, diarrhea, dehydration, acidemia and methemoglobinemia. These reactions mimic sepsis.

Powell initially characterized the syndrome. She described nine infants with severe, protracted diarrhea and vomiting. The symptoms developed at 4 to 27 days after birth (mean, 11 days) in infants on a cow’s milk-based formula. Switching to a soy-based formula resulted in transient improvement, but symptoms generally recurred in 7 days. Seven of the nine infants were below birth weight, and eight of nine presented with dehydration. Eight of the infants appeared acutely ill and underwent sepsis evaluations that were negative. All infants were noted to have low serum albumin, elevated peripheral blood polymorphonuclear leukocyte counts, and stools that were positive for hemoglobin and reducing substances. The hospital course usually involved improvement while on intravenous fluids, followed by recurrence of dramatic symptoms with reintroduction of soy- or cow’s milk-based formula, including the development of shock in several infants. Follow-up with oral challenges was carried out with cow’s milk and soybean formulas at a mean age of 5.5 months, and 14 of the 18 challenges were positive. Ten of 14 challenges resulted in vomiting (onset 1 to 2.5 hours after ingestion; mean 2.1 hours) and all experienced diarrhea (onset 2 to 10 hours; mean 5 hours) with blood, polymorphonuclear leukocytes and eosinophils, and increased carbohydrate in the stool. There was a rise in peripheral blood polymorphonuclear cell counts in all positive challenges, peaking at 6 hours after ingestion, with a mean rise of 9900 cells/mm ³ (range 5,500 to 16,800 cells/mm ³ ). Only isolated gastrointestinal symptoms were reported.

The results of these studies led Powell to propose criteria for a positive oral challenge to diagnose food protein-induced enterocolitis of infancy. Confirmation of the allergy included a negative search for other causes, improvement when not ingesting the causal protein and a positive oral challenge resulting in vomiting/diarrhea, evidence of gastrointestinal inflammation through stool examination, and a rise in the peripheral polymorphonuclear leukocyte count to over 3,500 cells/mL.

Numerous foods, other than milk and soy, have subsequently been documented as triggers for FPIES, including rice, oat, meats, fish, fruits, vegetables and egg. The dramatic nature of the presentation often results in evaluations for sepsis or surgical diagnoses, and a delay in final diagnosis until more than one episode has occurred.

Since infantile FPIES is a diagnosis that can be made clinically, there are no large series in which biopsies have been performed solely in patients fulfilling Powell’s criteria. Regarding immunopathology, studies have focussed upon the role of T cells and the importance of tumor necrosis factor (TNF)-α. In a large cohort of Japanese infants with non-IgE-mediated food allergy, TNF-α, interleukin (IL)-6, and Th2 cytokines (IL-3, IL-5 and IL-13), but not interferon (IFN)-γ or IL-17 were increased in the supernatant from milk protein-stimulated peripheral blood mononuclear cell cultures of patients compared to nonallergic controls. Chung and colleagues examined the presence of TNF-α in duodenal biopsy specimens using immunostains in infants with FPIES. Semiquantitative analyses revealed higher staining for TNF-α in affected infants with villus atrophy compared to those without atrophy and in normal controls. Taken together, these studies support the notion that TNF-α plays a role in the acute and chronic symptoms of FPIES. It is also known that the regulatory cytokine transforming growth factor (TGF)-β1 is involved in the protection of the epithelial barrier of the gut from the penetration of foreign antigens. Chung and colleagues demonstrated that the type 1, but not type 2, receptor for TGF-β1 were decreased in duodenal biopsy specimens in patients with FPIES compared to controls. Analysis of humoral features in milk-induced enterocolitis showed milk-protein specific IgA, but very low levels of specific IgG1 and IgG4; this has been theorized to be pathogenic because IgG4 might otherwise block complement fixing antibodies. Specific IgE is sometimes noted as well, and may be a marker of persistence.

Dietary Protein Enteropathy

This disorder is characterized by protracted diarrhea, vomiting, malabsorption and failure to thrive. Additional features may include abdominal distention, early satiety, edema, hypoproteinemia and protein-losing enteropathy. Symptoms usually begin in the first several months of life, depending on the time of exposure to the causal proteins. The disorder was described primarily from the 1960s to the 1990s and was commonly attributed to cow’s milk protein. A decrease in prevalence was documented in Finland and Spain and attributed to a rise in breastfeeding and/or the use of adapted infant formula. There have been no clear reports of this diagnosis in the past several years, although presentations of eosinophilic gastroenteritis with protein-losing enteropathy share many features with previous descriptions of this disorder.

Unlike gluten-sensitive enteropathy (celiac disease), this enteropathy generally resolves in 1 to 2 years, and there is no increased threat of future malignancy.

Celiac Disease

Celiac disease, also termed celiac sprue or gluten-sensitive enteropathy and estimated to affect 1% of the population, is caused by an immune response triggered by wheat gluten or related rye and barley proteins that results in inflammatory injury to the small intestinal mucosa. The classic presentation occurs in infants after weaning, at the time when cereals are introduced into the diet. Early (<4 months of age) or delayed (>7 months) introduction of wheat may be a risk factor, and breastfeeding may be a related protective factor. Symptoms partly reflect malabsorption, with patients exhibiting failure to thrive, anemia and muscle wasting. Additional symptoms are varied and include diarrhea, abdominal pain, vomiting, osteoporosis, bone pain and aphthous stomatitis. Subclinical or minimal disease is possible, delaying diagnosis into adulthood. Chronic ingestion of gluten-containing grains in patients with celiac disease is associated with increased risk of enteropathy-associated T cell lymphoma. Celiac disease is associated with autoimmune disorders and IgA deficiency. Another associated disorder is dermatitis herpetiformis, a gluten-responsive dermatitis characterized by pruritic, erythematous papules, and/or vesicles distributed symmetrically on the extensor surfaces of the elbows and knees, and also on the face, buttocks, neck and trunk.

Endoscopy of the small bowel in active celiac disease typically reveals total villous atrophy and extensive cellular infiltrate. The disorder is caused by gliadin-specific T cell responses against deamidated gliadin produced by tissue transglutaminase. Gliadin stimulation of monocytes and macrophages may also contribute to the inflammatory response. Antigen presentation appears to be a central issue in the immunopathology because about 95% of patients are HLA-DQ2, with the remainder being HLA-DQ8. Gliadin is one of the few substrates for tissue transglutaminase, which deamidates specific glutamines within gliadin, creating epitopes that bind efficiently to DQ2 gut-derived T cells. The activation of DQ2- or DQ8-restricted T cells initiates the inflammatory response. Elimination of gliadin from the diet results in a down-regulation of the T cell-induced inflammatory process and normalization of the mucosal histology.