Fecal Microbial Transplantation in Early‐Onset Colitis

A child, L.B., born at term, the third child whose older sister has juvenile idiopathic arthritis successfully treated with azathioprine, presented at 3 months with 6 to 8 bloody diarrheal stools per day, resolving spontaneously within 1 week. She was breast-fed for 4 months, when solid food was started. Formula was introduced at 6 months. She was asymptomatic between 4 and 10 months, but then presented again with bloody diarrhea, up to 10 stools per day, with mucus and bright red blood in her diapers. Upper endoscopy and histology were normal. Lower endoscopy showed pancolitis with loss of colonic haustrations, friability, absent vascular pattern, and some small ulcerations (Fig. 1). Histology was consistent with an ulcerative colitis (UC)–like phenotype; additionally, perinuclear anti-neutrophil cytoplasmic antibody was positive (1/160). Oral (50–60 mg · kg−1 · day−1) and rectal (333 mg/day) mesalamine was initiated. Following a good initial response, she relapsed 2 months later. At 13 months, azathioprine (2.0 mg · kg−1 · day−1) and prednisolone (2 mg · kg−1 · day−1) were initiated. Although she responded with decreased severity of symptom, stools remained frequent, bloody, and liquid. General condition, weight, and linear growth remained acceptable. The use of an amino acid–based formula for several months had no effect. Iron supplements failed to compensate for the rectal bleeding, and transfusion of packed red blood cells was needed every 3 to 4 weeks. Several probiotics, including VSL-3, proved unsuccessful. Infliximab treatment (5 mg/kg at 0–2–6–10 weeks) had no effect. An adverse reaction to sirolimus (Rapamune; Pfizer Pharmaceutical, New York, NY) made further administration impossible. Intravenous broad-spectrum antibiotics and total parental nutrition were administered for severe flares. Repeat biopsies with histology including electron microscopy did not reveal etiologies other than the "UC-like image." Antigen detection for Clostridium difficile was negative. An immunodeficiency was suspected because of L.B.'s young age at onset. Lymphocyte numbers and subsets were normal. Immunoglobulin levels were within the normal range for age. B-cell subsets for common variable immunodeficiency were analyzed and showed no decrease in switched memory B cells. Vaccination response against tetanus was adequate. Normal dihydrorhodamine flow cytometric assay made chronic granulomatous disease unlikely. Signaling-lymphocytic-activation-molecule–associated protein (SAP), X-linked inhibitor of apoptosis protein, and Wiskott-Aldrich syndrome protein deficiency was not tested as the child was female and the parents were not genetically related. Foxp3 was normal, excluding immunodysregulation polyendocrinopathy enteropathy X-linked syndrome. No mutations were detected in the NOD2/CARD15 or in the STAT3 genes. The IL-10 (interleukin 10) gene was not completely sequenced; however, evaluation of known single nucleotide polymorphisms in the IL-10 gene showed a wild-type sequence for exons 1–4 and the coding region in exon 5 and a homozygous single nucleotide polymorphism that is also found in the healthy population (rs3024496 in the 3′UTR in exon 5). Both the IL-10RA and IL-10RB gene were sequenced. A heterozygous unclassified variant was found in the IL-10RA gene. This heterozygous silent c.180G>A p.(=) is not predicted to lead to abnormal function of the protein. Coding exons 1–7 of the IL-10RB gene were sequenced and showed no mutations. Functional testing of the IL-10 pathway at 9 months showed normal function. In particular, in vitro T-cell stimulation with anti-CD3 yielded normal IL-17 secretion. At 20 hours after culture, IL-12p70, tumor necrosis factor-α, and IL-8 concentrations were normal in stimulation assays of peripheral blood mononuclear leukocytes with lipopolysaccharide; however, after lipopolysaccharide stimulation, IL-6 and IL-1-β were increased as observed in auto-inflammatory syndromes. Concentrations of IL-1-β, IL-6, and IL-12p70 were increased in plasma.FIGURE 1: Histology just before (A) and 4 months after (B) the last fecal transplantation. A, Colitis: distortion of the normal architecture; regenerative changes of the epithelium; presence of crypt abscesses; dense stromal inflammation. Absence of granulomas. A, Normal mucosa: crypts are arranged parallel to each other; they are covered by a normal epithelium; total restoration of the mucin content; the lamina propria contains no inflammatory infiltrate.The patient was seen almost every 2 weeks at the outpatient clinic and more frequently as needed (Fig. 2: growth chart). The attempted treatments did not result in a sufficient clinical improvement. The Pediatric Ulcerative Colitis Activity Index (PUCAI) (1) varied between 60 and 75 (Table 1).FIGURE 2: Growth chart of the patient. The growth chart illustrates faltering growth during the active disease and (ineffective) prednisolone treatment (growth below P3) and the recovery growth when the patient was asymptomatic and when prednisolone was stopped, again reaching P3 (genetically, she should be on P25). Weight has returned to P50, where it was at the beginning of the disease.TABLE 1: Evolution of the Pediatric Ulcerative Colitis Activity IndexBecause of limited experience with alternative medical treatment options and the invasiveness of total colectomy, the possibility of fecal transplantation was discussed with the parents, who consented for fecal transplantation. The approval of the ethical committee at UZ Brussel was obtained for the procedure. The first fecal microbial transplantation was attempted at the age of 18 months, 5 months after the start of prednisolone and azathioprine. Donor stools were screened (2). Approximately 100 g of fresh fecal material was homogenized in 100 mL of sterile saline and filtered through gauze to remove larger particles. The first 2 infusions were administered into the cecum through a colonoscope, whereas the next 5 were infused by nasoduodenal tube. The first 4 fecal transplantations, with fresh stools from an age-related niece, led to transitory improvement and normal defecation for 7 to 14 days. The last 3 infusions contained fresh stool from the patient's older brother. Interestingly, L.B. tolerated the fecal transplants from the niece well, but she reacted to the first fecal transplant from the brother with profuse sweating, vomiting, paleness, tachycardia (blood pressure remained at 70/40), and transitory fever for an hour. She recovered spontaneously within an hour, and no active intervention was needed. Normalization of the stools lasted for 1 month. Because of the observed but transient clinical improvement, further fecal transplant using the brother's stools was discussed with the parents. She had a similar but smaller reaction after the second transplant 6 weeks later, followed by remission for 2 months. A much lesser reaction occurred again at the third infusion, and 6 months later L.B. still had normal stools. Prednisolone and azathioprine were initially continued, but were gradually decreased during 6 weeks (first prednisolone, then azathioprine) when the patient had been symptom-free for 2 months. The histology of the last colonoscopy, performed during anesthesia to remove the Port-a-Cath (Smiths Medical, Gary, IN) 6 months after the last fecal infusion and approximately 3 months after stopping all medication, showed normal histology and thus a resting phase or inactive UC. To our knowledge, this is the first report of a successful fecal transplant in a child with early-onset colitis (3). Consideration of fecal transplant should be made in the context of whether it is safer than continuing ineffective regimens that may be detrimental to the patient. We observed that a patient receiving fecal transplantation may develop systemic reactions that may be related specifically to the fecal content of an individual donor. As this case documents potential intolerance to infusion of foreign fecal material, we recommend careful monitoring of these patients as we learn more about the safety and efficacy of this recently applied treatment. Bacteremia following fecal microbiota transplantation has been reported (4). Further investigations may help determine whether the host intestinal and/or systemic immune responses to infused feces are related to disease activity.