2140. Biological Features of Response to VE303, a Defined Bacterial Consortium, in Patients with Clostridioides difficile Infection (CDI): Results from the Phase 2 CONSORTIUM Study

• Published in: Open forum infectious diseases 2023-11, Vol.10 (Supplement_2)
• Main Authors: Menon, Rajita, Crossette, Emily, Bhattarai, Shakti, Bucci, Vanni, Prince, Amanda L, Faith, Jeremiah, Olle, Bernat, Silber, Jeffrey L, Norman, Jason
• Format: Article
• Language: English
• ISSN: 2328-8957; 2328-8957
• DOI: 10.1093/ofid/ofad500.1763

Abstract Background Fecal transplant and other donor-derived treatments promote a gut environment resistant to CDI, but these treatments have inherently variable quality attributes, are difficult to scale, and can transfer harmful pathogens. VE303 is a rationally defined consortium of 8 purified, clonal strains, overcoming the limitations of donor-derived treatments. In the CONSORTIUM Study, high-dose VE303 was well tolerated, reduced the odds of recurrent CDI (rCDI) by > 80%compared with placebo, and led to both robust colonization of VE303 strains and early restoration of the native microbiota. VE303 strain detection was associated with clinical efficacy. Methods This was a double-blind, placebo-controlled phase 2 trial (Fig. 1). After completion of antibiotic treatment for a lab-confirmed CDI episode, 79 subjects were randomized to low-dose VE303, high-dose VE303, or placebo and dosed orally once daily for 14 days. Subjects were followed for 24 weeks to monitor safety and rCDI episodes; samples were collected during dosing and at weeks 4 and 8 for metagenomic, metabolomic, and immune analysis. Random forest classification models were used to identify variables that predicted strain colonization and response to VE303. Results On-study CDI recurrence in VE303-dosed subjects was predicted by higher primary bile acid (BA) levels, and lower levels of secondary BAs and short-chain fatty acids (SCFA) at screening. In particular, taurochenodeoxycholic acid predicted recurrence; lithocholic acid, deoxycholic acid, hexanoate and isovalerate predicted non-recurrence (Fig. 2). Advanced age was an important predictor of recurrence in all models (Fig. 2A, C). In models of VE303 colonization, particularly of the effector strain VE303-08 (Fig. 3A), low strain detection was predicted by high screening concentrations of primary BAs and inflammatory cytokines (Fig. 3B). Metabolite profiles at screening predict response to VE303. (A, C) Random Forest (RF) model importance plots show variables predicting on-study CDI recurrence in VE303-dosed subjects based on screening concentrations of BAs (top) and SCFAs (bottom). Variable importance is defined as the drop in predictive power when a variable is held out of the model. Bar color indicates the fraction of model iterations (out of 10) wherein the variable was found to be important. Vertical legend indicates association with recurrence. (B, D) Distributions of important BAs (top) and SCFAs (bottom) at screening in recurrent v