Abstract
BACKGROUND & AIMS Fecal microbiota transplantation (FMT) has been proved to be efficient in treating Clostridium difficile infection disease, yet its efficacy in treating Inflammatory bowel disease including Crohn’s Disease (CD) and Ulcerative Colitis (UC) at molecular level are blank.
METHODS We performed a parallel study of patients with moderate to severe CD (Harvey-Bradshaw Index ≥ 7) and UC (Montreal classification, S2 and S3). Patients were treated with single FMT (via mid-gut, from healthy donors; n = 15). All participants had their fecal samples collected and shotgun sequenced before FMT and during their follow-up visits. The primary outcome was clinical remission and that of CD is defined as a decrease of Harvey-Bradshaw > 3, clinical remission of UC is defined as a decrease of Mayo score > 3. To describe and quantify the change of gut microbiota of IBD patients after FMT, we monitored strain populations in 44 fecal samples. Besides, we built a machine learning model to predict the existence and abundance of post-FMT patients’ species compositions.
RESULTS Of all 15 patients, 3 days after FMT treatment, 8 out of 11 CD patients were relieved, 3 out of 4 UC patients were relieved (Table S1).
We observed the transfer of donor strains to recipient was more abundant in UC than in CD patients, persisting the follow-up time points. Besides, same-donor recipient differs in the degree of microbiota transfer. Furthermore, through building random forest classification and regression model, results showed that both the presence and abundance of some post-FMT patients’ species were predicable, indicating a possibility of precision engineering of the recipients’ gut microbiota under the FMT treatment.
CONCLUSIONS FMT treatment efficiency differed in CD and UC patients and post-FMT patients’ mOTU composition was predictable in our data set.