Results

Over the period 1990 to 1994, 42 infants who fulfilled the inclusion criteria of histologically proved non-infective colitis were investigated. Males were more common than females (ratio 1.7 to 1). Diarrhoea, bloody stools, and weight loss were the most common presenting symptoms. Abdominal pain, anorexia, and vomiting occurred in 41%, 39%, and 33% of the cases respectively (table 1). In 66% of infants the symptoms started before the age of 6 months, and the duration of the symptoms before presentation ranged from two weeks to 14 months (mean duration 3.5 months).

A personal history of atopy was present in 29% overall. This was commonest in those with allergic colitis (35%) being found in 20% of those with specific inflammatory bowel disease and 12% of those with autoimmune related colitis.

A positive family history of atopy/asthma was found in 48% of all the cases. Over 60% of the atopic relatives were grandparents, uncles and aunts, and cousins. A quarter of the infants themselves had a personal history of atopy or asthma or both. Thirty three per cent of the cases were exclusively fed on breast milk at the onset of the symptoms and another 31% were on breast and artificial milk. Recurrent upper respiratory infections were recorded in 33% of the cases.

Sixty two per cent of the cases had a weight less than the third centile for age and sex at presentation, and 36% a height less than the third centile.

All infants underwent colonoscopy and a total colonoscopic examination was achieved in 83% (35/42). Table 2 shows that during the endoscopic examination, macroscopic abnormalities of the intestine were evident in 83% of the patients. A pancolitis was found in 36% and inflammatory changes confined to the rectosigmoid 31%. The rectum was macroscopically abnormal in 78% of cases. The endoscopy was macroscopically normal in 17% of the cases, and all of these had allergic colitis on histology.

Allergic colitis accounted for 62% of cases of colitis in the infants, while non-specific colitis, autoimmune enteropathy, and ulcerative colitis each accounted for 10% of the cases. Crohn’s colitis was seen in only one male infant who was 18 months old (table 3).

Anaemia (haemoglobin < 105 g/l) was observed in 36% of the cases at presentation, with a mean haemoglobin concentration of 98 g/l. Thrombocytosis, increased erythrocyte sedimentation rate, and low albumin concentrations were observed in 56%, 49%, and 55% respectively, and eosinophilia in six cases (14%) (table 4). The nitroblue tetrazolium test and skin prick tests were normal in all the infants.

The T cell lymphocyte subset population showed a low CD3 and CD4 in two thirds of those with severe combined immunodeficiency (SCID) and three quarters of those with autoimmune enterocolitis, with a raised CD8 in two of the three patients with SCID who had graft versus host disease. An abnormal phytohaemagglutinin (PHA) response was observed in all the patients with autoimmune enterocolitis, and one third of patients with SCID. Both the T cell lymphocyte subsets and the PHA response were normal in those with allergic colitis and non-specific colitis. Immunoglobulin G concentrations were below the age related minima (< 3.1 g/l) in 26% of the infants, and low IgA concentrations (< 0.3 g/l) were found in 47%. Low IgG₂ (< 35 mg/dl) and IgG₄ (< 4.8 mg/dl) subclasses were observed in 59% and 48% of the infants studied respectively. This was due to a high prevalence in those with allergic colitis, where 71% had low IgG₂ and 50% low IgG₄, and in ulcerative colitis where three quarters of the cases had low IgG₂ and IgG₄ (table 5).

RESPONSE TO TREATMENT

Table 6 shows that 77% of the infants with allergic colitis had remission of their symptoms and started gaining weight on dietary manipulation within the first three months of treatment, while only 20% of those with specific inflammatory bowel disease (ulcerative colitis and Crohn’s disease) who were on steroids, salazopyrine, and dietary management went into remission in the first three months of treatment. Those with severe combined immunodeficiency and autoimmune enterocolitis responded to immunomodulation slowly, with only a quarter having a remission of the symptoms within six months of the start of treatment.

Discussion

This study showed that evidence of minor immunodeficiency is common in infants with non-infective colitis. In agreement with previous studies we have confirmed that in this age group the commonest cause of non-infective colitis is allergic colitis.1 5 6 9 11 14 15 Ulcerative colitis, non-specific colitis, autoimmune enterocolitis, and severe combined immunodeficiency accounted for only approximately one third of the cases, occurring with the same frequency.

Serum IgG₂ and IgG₄ deficiency commonly coexisted with deficiencies of IgA, raised IgE or both in this group of infants with colitis.16 17 These immunoglobulins have a role in health in protecting the individual against the uncontrolled absorption of whole proteins and complex antigens from the gastrointestinal lumen. Previous work has shown that the formation of immune complexes, particularly IgA specific immune complexes, may contribute to the release of goblet cell mucus onto the intestinal surface, thus preventing the penetration of antigens or the adherence of infectious agents.18 A deficiency of the serum immunoglobulins may lead to failure of antigen-immunoglobulin complex formation, resulting in intact antigen absorption and sensitisation of the individual. With re-exposure to dietary antigen, minute quantities of allergen may be absorbed and result in allergic symptoms. A vicious cycle of allergic reaction with increased allergen uptake across a disrupted mucosal surface may ensue. Uptake of non-specific intestinal antigen can in turn evoke an IgG mediated reaction which may predispose to chronic disease.19

Increased intestinal permeability, which is a genetically expressed defect that predisposes to the development of inflammatory bowel disease, has been reported.20 One of the genetic defects reported here is probably a deficiency of immunoglobulin A, IgG₂, and IgG₄ subclasses which predisposes individuals to the absorption of whole antigens, an IgE mediated immunological response, and the release of inflammatory mediators, resulting in an entero- colitis. The inflamed intestinal epithelium allows more antigens to cross the mucosa and the vicious cycle continues.

Anaphylatoxins stimulate macrophages to release prostaglandin E₂, thromboxane B₂, and leukotriene B₄,21 thus initiating and maintaining the inflammatory and tissue damaging process (type 2 and 3 hypersensitivity).

IgG₁ secreting plasma cells are the immunocyte class showing by far the greatest expansion in specific inflammatory disease.22 The upregulation of IgG₁ is probably a compensatory mechanism that can occur when the other subclasses are low, as seen in these infants.

A family history of atopy is very common in infants with colitis in this and other studies.9-11 14 23 These infants are prone to recurrent infections of the upper respiratory tract, which may be due to the IgG₂ and IgG₄ subclass deficiency16 as seen in our patients. The finding of a negative intradermal skin prick test in our patients is not unusual, since infants have a poor response to such tests.25

In this series, the T cell function was abnormal in patients with severe combined immunodeficiency, autoimmune enterocolitis, and ulcerative colitis, but largely normal in those with allergic and non-specific colitis. The low CD3 and CD4 T lymphocyte subset population with a raised CD8 in those with SCID probably reflects graft versus host disease (GVHD) due to immunocompetent lymphocytes from the mother. The endoscopic and histological findings in two of these patients were in keeping with GVHD. Those with autoimmune enterocolitis had plasma cell plus lymphocytic infiltration of the lamina propia and the mucosa, with positive smooth muscle and gut autoantibodies.

A third of the infants in this study were exclusively breast fed when they developed colitis. Previous studies in infants fed cows’ milk based formulas from birth have suggested that this group may be at a higher risk of developing allergic colitis. The condition is, however, described in children fed other whole protein milk formulas (soybean) and in infants fed on breast milk.14 24 In this latter group, exclusion of cows’ milk, egg, and soybean products from the maternal diet may alleviate the infant’s symptoms.

The outcome of treatment with exclusion diet alone was excellent in the children in this study with allergic colitis. Infants with other causes of colitis took longer to go into remission, requiring immunosuppression. This group frequently relapsed when intercurrent infections occurred. No case of non-specific colitis went on to develop specific inflammatory bowel disease during follow up within the time course of this study.

We have shown that infants with non-infective colitis commonly have an associated minor immunodeficiency. We suggest that in allergic colitis low levels of IgA, IgG₂, and IgG₄ might impair the function of mucosal defences against macromolecular passage across the mucosal barrier, thus predisposing to an allergic inflammatory response and mucosal injury. In ulcerative colitis it seems more likely that the minor immunodeficiency alluded to here is a result of an underlying immunodysregulation, although the findings of specific IgE reaginic antibodies to cow’s milk in as many as 30% of children with ulcerative colitis implies that loss of barrier function in these children might also predispose to allergic responses.

In this study population—as in others previously—allergic colitis was the commonest cause of non-infective colitis in infancy and this responded excellently to exclusion diet alone. A positive family and personal history of atopy was common in infants with allergic colitis and in this group the rectum was invariably affected. Macroscopic appearance at endoscopy did not discriminate between normality and allergic colitis, and for this reason mucosal biopsy is considered mandatory to make the diagnosis.