Authors' Response.

Abstract

To the Editor: We greatly thank Hoffmann et al for their interest in our article (1) and would like to provide some clarifications regarding their questions. Food protein-induced proctocolitis, also known as allergic proctocolitis (AP), is characterized by the presence of mucoid, frothy, and bloody stools in an otherwise healthy infant (2). As rightly pointed out by Haufmann et al and emphasized in our article, rectal bleeding may have other causes and benign bleedings (idiopathic neonatal transient colitis [INTC]) that are not a result of food allergies and may spontaneously recover in infants, therefore an accurate differential diagnosis becomes essential (1,3). AP diagnosis was made on the basis of the following methodology in our study: mothers of infants whose symptoms resolved typically in 72 to 96 hours (with significant blood in stool, resolution took 1 week) after the elimination were recommended to retake causative foods as part of their normal daily diet after 3 weeks. If infants who had rectal bleeding, diarrhea, and mucus or blood in stool after reexposure to the offending food (via breast milk), they were diagnosed as having AP. Those patients (n:18) having had no symptoms after oral food challenge were excluded from our study (otherwise probable INTC cases). Under these considerations, OFC was applied to all (60/60) of our patients and AP was confirmed in all our patients. In addition, recurrence of symptoms was observed in the patients incidentally exposed to allergenic food during the follow-up period. Because all the patients were referred to our clinic with an AP prediagnosis and our study was essentially designed to provide clinicians more information on the natural course, prognosis, and offending foods of AP, our INTC or AP cases should not be taken as suggestion of a general ratio. The exact prevalence of AP is unknown; the estimated prevalence ranges from 0.16% to 64% of infants with isolated rectal bleeding (3–6). INTC and AP rectosigmoidoscopy both show nodular lymphoid hyperplasia with a pale mucosal surface and diffuse eosinophil infiltration in the lamina propria (7,8). Mori et al (7,9) indicated that microarray analysis enhanced expression of CCL11 (eotaxin-1) was more dominant in neonatal transient eosinophilic colitis, whereas the expression of CXCL13 was more dominant in food protein–induced proctocolitis. CXCL13 and its receptor CXCR5 control the organization of B cells within follicles of lymphoid tissues (10). The pathophysilogy of non–immunoglobulin-E gastrointestinal food allergies remains unknown. There is an urgent need for further studies to characterize mucosal inflammation. In agreement with Haufmann et al, rather than resorting to clinical suspicion and pathological diagnosis, which may lead to confusion, OFC should be conducted as was the case with our study.