Purpose of the Study.

To determine the prevalence of allergic colitis (AC) in a cohort of healthy infants with rectal bleeding. A secondary purpose was to determine if bleeding would resolve in untreated infants with rectal bleeding without biopsy-proven AC.

Study Population.

There were 22 infants ≤6 months of age with rectal bleeding recruited from the referral area of Cincinnati Children’s Hospital Medical Center (Cincinnati, OH). All subjects had a negative history of bleeding disorders, negative stool cultures, positive hemoccult, and a negative history and physical examination for signs of infection, Hirschsprung disease, and inflammatory bowel disease.


AC was defined histologically as colonic mucosa with ≥6 eosinophils per high-power field and/or eosinophils in the colonic crypts or muscularis mucosae. Formula or maternal diet was changed only for infants with histologic findings of AC. Formula-fed infants were switched to an extensively hydrolyzed formula and were rebiopsied at 3 weeks. If the biopsy was normal, they were continued on the formula and managed clinically. Those with continued histologic evidence of colitis were changed to an amino acid–based formula and rebiopsied at 6 weeks. Breastfed infants continued breastfeeding while mothers followed a milk-protein–free diet. Those with resolution of bleeding and normal biopsies at 3 weeks continued with breastfeeding and a restricted maternal diet. Those with persistent histologic evidence of colitis were rebiopsied at 6 weeks with no further dietary change. Those with persistent bleeding were changed to hydrolysate and rebiopsied at 6 weeks. Those with persistent bleeding and histologic evidence of AC at 6 weeks were changed to an amino acid–based formula.


Of 22 subjects, 14 (63.6%) had histologic evidence of AC. Five had normal biopsies and 3 had nonspecific colitis. Seven of the 14 with AC were formula fed. Six of the 7 had resolution of bleeding, on average, in 1.8 weeks (range: 1–5 weeks). One of the 7 was changed to an amino acid formula at 3 weeks and had resolution of bleeding at 5 weeks. The remainder of the 14 were breastfed. Six were followed to completion of the study. One had a delayed diagnoses because of development of worsening rectal bleeding and an abnormal biopsy at week 3 despite a normal biopsy at the onset of the study. The infant failed to improve with hydrolyzed formula but had resolution of bleeding by week 8 after initiation of an amino acid formula. Of the remaining 5, 2 had normal histology at week 3 with maternal elimination of cow’s milk. Two had improvement by week 3, and 1 had no change. The average time for resolution in the breastfed group was 5.6 weeks (range: 2–8 weeks). For the 5 infants without histologic evidence of colitis, the average time for resolution of bleeding was 3.25 weeks. In those with nonspecific colitis, 2 had resolution by week 6, and the third was ultimately diagnosed with inflammatory bowel disease.


A significant proportion of infants with rectal bleeding may not have AC and may undergo unnecessary, expensive formula or maternal diet changes that may discourage breastfeeding.

Reviewer Comments.

This small study provides important insights about the prevalence and natural course of proctocolitis. A much larger prospective placebo-controlled study that compares treatment versus no treatment would be very helpful.

This content is only available via PDF.
You do not currently have access to this content.