Introduction

Anti-TNF availability has been associated with a decrease or delay of the first intestinal resection in patients with Crohn’s disease (CD). However, few studies have evaluated the impact of availability of biological therapies on the change in postoperative recurrence (POR) preventive strategies and the need for second intestinal resection.

Aims & Methods

Aims: To assess time trends in the management strategies in the postoperative setting after a first intestinal resection for CD and the rate of second surgeries through the last three decades.
Methods: Adult patients with CD who underwent a first ileocolic resection with ileocolic anastomosis between 1990 and 2020 and at least one year of clinical follow-up, were identified from the Spanish ENEIDA registry. Patients in whom ileocolic resection was due to cancer were excluded. Patients were grouped by date of surgery (before biologicals -BB- 1990-2000; early biological era -EB- 2000-2010; widespread biological era -WB- 2011-2020). Medical prevention of POR was defined as any immunomodulator or biological agent started within the first 6 months after the ileocolic resection and maintained for at least 3 months. Medical treatment of POR was defined as any immunomodulator or biological agent started at least 6 months after the index surgery. Surgical POR was defined as the need for a second intestinal resection at least 6 months after the index surgery.

Results

A total of 5,084 patients were included (24% BB, 38% EB and 38% WB). Main characteristics of the whole cohort were as follows: 51% men, 39% active smokers at the first surgery, 40% penetrating and 40% stricturing behaviour and 25% perianal disease. 34% exposed to immunomodulators and 25% to biologicals before the first surgery. Median disease duration at first surgery of 3 years (range, 0-46), median age at index surgery was 35 years (range, 9-89), median follow-up until second resection or last visit of 125 months (range 6-408). 55% followed medical prophylaxis (33% thiopurines, 10% biological monotherapy, 12% combotherapy). Surgical POR occurred in 17%. A significant increase in medical prevention of POR with both thiopurines (BB 13%, EB 45%, WB 32%) and biological agents (BB 0%, EB 12%, WB 46%) was observed; moreover, an increasing proportion of patients started medical treatment with biologicals for POR (BB 35%, EB 48%, WB 43%) (P<0.001 for all comparisons). The cumulative surgical POR-free survival was significantly higher in both EB and WB cohorts as compared to BB cohort at 3, 5 and 10 years (BB 96%, 91%, 82%; EB 97%, 95%, 89%; WB 97%, 96%, 90%; P=<0.001).

Conclusion

A significant increase in the use of medical prevention for POR after a first intestinal resection for CD was observed in the last two decades. Moreover, almost half of the patients started biologicals to treat POR. These changes resulted in a significant decrease in the cumulative probability of a second intestinal resection in the long-term.

Introduction

Immune-mediated inflammatory diseases (IMID) involve immune dysregulation, potentially damaging various organs [1]. Gut microbiome regulates immune homeostasis, and abnormality of it is associated with various inflammatory diseases [2,3]. However, previously reported biomarkers have considerable limitations. Findings are often specific or inconsistent across studies, and while some studies assess common microbial patterns across diverse diseases, they often overlook disease pathobiology. To address this, we conducted a consensus-based approach to identify and dissect robust biomarkers common or unique to different IMIDs.

Aims & Methods

We collected and uniformly profiled public human gut metagenome data of IMIDs, using universal bacterial core gene-based taxonomic profiling [4] and pangenome-based functional profiling. A total of 2,252 samples from 18 studies of 28 case-control sub-cohorts of IMIDs (10 for Crohn’s disease;CD, 9 for ulcerative colitis;UC, 4 for ankylosing spondylitis;AS, 2 for chronic obstructive pulmonary disease;COPD, 2 for multiple sclerosis;MS, and 1 for psoriasis;PsO) were included as discovery set, and 253 samples from 3 studies for 7 case-control sub-cohort sets of CD or UC were included as external validation set. Cross-cohort and one-study-leave out validations were performed for the discovery set, and the external validation set was evaluated using the trained model from the IMID discovery set. For the individual marker analysis, significance scores were combined for each criteria, including disease diagnosis, and inflammatory bowel disease (IBD) and nonGI-IMIDs (MS, AS, COPD, and PsO). Additionally, multi cohort-based networks of IBD studies were assessed. Furthermore, to explore potential interpretation of the signatures in other clinical settings, we examined the IBD studies with disease activity as well as FMT study.

Results

Cross-cohort validation of IMID case-control sub-cohorts of the discovery set at species-level showed that intra-cohort performance had significantly higher area under curve (AUC) compared to inter-cohort (W=3701, p-value<0.001), and among the inter-cohort, intra-disease had significantly higher AUC than inter-disease (W=23278, p-value<0.001). When the external IBD sub-cohorts were tested, the trained model of the IMID set showed the average AUC 0.818 (SD=0.089). In individual signatures, there were several features commonly found to be highly scored in at least 5 IMIDs, including Enterocloster bolteae and probable pyridine nucleotide-disulfide oxidoreductase (K21739) in the case-side and Faecalibacterium sp. (CLG_10009380_s) and Oscillibacter sp. (MSSCM00948369_s) in the control-side. There were also disease or disease category-specific patterns such as Gemmiger formicilis and Alistipes shahii consistently found to be enriched in control-side in IBD but not in nonGI-IMIDs. Furthermore, we checked the signatures obtained from our analysis in the other clinical settings such as IBD samples with disease activity scores as well as FMT study. For example, Dysosmobacter welbionis was negatively correlated with Harvey Bradshaw Index in 2 CD sub-cohorts (Spearman’s rho=-0.5 and -0.26, p-value=0.009 and 0.039, respectively).

Conclusion

Our study analyzed multi-cohort gut metagenome data to identify consistent signatures across various IMIDs. Cross-cohort validation showed better intra-disease prediction compared to inter-disease, yet the performance of predictions between diseases was still comparable. Several microbial signatures were either shared or unique across IMIDs, with some showing potential clinical relevance in IBD.

References

1. McInnes IB, Gravallese EM. Immune-mediated inflammatory disease therapeutics: past, present and future. Nat Rev Immunol. 2021 Oct;21(10):680-686. doi: 10.1038/s41577-021-00603-1. Epub 2021 Sep 13. PMID: 34518662; PMCID: PMC8436867.
2. Forbes JD, Van Domselaar G, Bernstein CN. The Gut Microbiota in Immune-Mediated Inflammatory Diseases. Front Microbiol. 2016 Jul 11;7:1081. doi: 10.3389/fmicb.2016.01081. PMID: 27462309; PMCID: PMC4939298.
3. Forbes JD, Chen CY, Knox NC, Marrie RA, El-Gabalawy H, de Kievit T, Alfa M, Bernstein CN, Van Domselaar G. A comparative study of the gut microbiota in immune-mediated inflammatory diseases-does a common dysbiosis exist? Microbiome. 2018 Dec 13;6(1):221. doi: 10.1186/s40168-018-0603-4. PMID: 30545401; PMCID: PMC6292067.
4. Chalita M, Ha SM, Kim YO, Oh HS, Yoon SH, Chun J. Improved Metagenomic Taxonomic Profiling Using a Curated Core Gene-Based Bacterial Database Reveals Unrecognized Species in the Genus Streptococcus. Pathogens. 2020 Mar 10;9(3):204. doi: 10.3390/pathogens9030204. PMID: 32164338; PMCID: PMC7157611.

Disclosure

Y Jung, HS Oh, M Chalita, J Moon, J Yang, K Lee, S Park, YO Kim and J Chun are employees of CJ Bioscience, Inc.