S1083 Association of Microbiome Composition and Health-Related Quality of Life in Patients with Recurrent Clostridioides Difficile Infection: Results from the PUNCH CD3 Phase 3, Randomized, Placebo-Controlled Clinical Trial

Introduction: Gut microbiome perturbation plays an important role in recurrent Clostridioides difficile infection (rCDI), which substantially reduces patients’ health-related quality of life (HRQL). This study explored the association between microbiome composition and HRQL by comparing gut microbiome profiles of patient groups with different levels of HRQL using data from REBYOTA’s (fecal microbiota, live-jslm [RBL]) phase 3 PUNCH CD3 trial (NCT03244644). Methods: The CDI-specific C. difficile Quality of Life Survey (Cdiff32) (total, physical, mental, social domain scores) and counts of gut microbiome taxa at class level were collected at baseline and weeks 1, 4, and 8. Data were analyzed in two ways for each Cdiff32 score. First, microbiota data from all visits regardless of treatment were divided into quartiles based on Cdiff32 score and average microbiome profiles were compared pairwise between quartiles. Second, patients were categorized as improved if Cdiff32 scores improved ≥ 10 points from baseline to week 8, unchanged otherwise. Average microbiome profiles were compared between baseline and week 8 within the improved and unchanged groups, respectively. All comparisons used the Dirichlet-multinomial model and differences were quantified via the Kullback-Leibler divergence. Results: A total of 176 out of 262 (67.2%) patients had Cdiff32 and microbiota data, comprising 501 unique data points (145 baseline, 144, 110, 102 at week 1, 4, 8; 69.2% RBL, 30.7% Placebo). The microbiome composition in the highest Cdiff32 score quartile (4th) was significantly different from most other groups with lower scores. Divergence between groups was smaller when average Cdiff32 scores were more similar (Table 1, Figure 1). Patients in the improved group had significantly different microbiome profiles between baseline and week 8 for all Cdiff32 scores; in the unchanged group, only for social domain score. Divergences within the improved group were 4 to 15 times larger than within the unchanged group. Patients in the improved group had healthier microbiota at week 8 vs baseline with higher relative abundances of Bacteroidia and Clostridia and lower Gammaproteobacteria and Bacilli. Conclusion: Our study found that patients with better HRQL had healthier microbiota, which points towards a direct link between microbiome and patient well-being in rCDI. This link could help explain the improvement in patient well-being observed after treatment with biotherapeutic products such as RBL that may restore the gut-brain axis. Table 1. - Mean and standard deviation (SD) of Cdiff32 scores in each quartile of patient visits Mean ± SD 1st quartile 2nd quartile 3rd quartile 4th quartile Total score 32.4 ± 9.6 53.3 ± 4.6 70.2 ± 6.0 89.6 ± 5.0 Physical domain score 39.1 ± 12.8 64.5 ± 5.8 82.4 ± 4.6 96.1 ± 3.0 Mental domain score 21.0 ± 7.9 38.9 ± 4.6 59.3 ± 7.2 83.9 ± 7.0 Social domain score 31.0 ± 10.2 55.8 ± 5.3 76.9 ± 7.0 97.3 ± 3.1 Note: [1] All scores range from 0 to 100, with 100 reflecting the best quality of life.Abbreviations: SD: standard deviation, Cdiff32: C. difficile Quality of Life Survey. Figure 1.: Pairwise Kullback-Leibler divergence between patient visit quartiles. Notes: [1] This analysis uses patient visits from all 176 patients. Patient visits in the 1st quartile have the lowest Cdiff32 scores and patient visits in the 4th quartile have the highest Cdiff32 scores. [2] Significance P<0.05 is denoted via * and was calculated using 1000 bootstrap repetitions. Abbreviations: Cdiff32: C. difficile Quality of Life Survey.