Randomized Controlled Trial of Oral Vancomycin Treatment in Clostridioides difficile-Colonized Patients

Abstract

Clostridioides difficile infection (CDI) is most commonly diagnosed using nucleic acid amplification tests (NAAT); the low positive predictive value of these assays results in patients colonized with C. difficile unnecessarily receiving CDI treatment antibiotics. The risks and benefits of antibiotic treatment in individuals with such cases are unknown. Fecal samples of NAAT-positive, toxin enzyme immunoassay (EIA)-negative patients were collected before, during, and after randomization to vancomycin (n = 8) or placebo (n = 7). C. difficile and antibiotic-resistant organisms (AROs) were selectively cultured from fecal and environmental samples. Shotgun metagenomics and comparative isolate genomics were used to understand the impact of oral vancomycin on the microbiome and environmental contamination. Overall, 80% of placebo patients and 71% of vancomycin patients were colonized with C. difficile posttreatment. One person randomized to placebo subsequently received treatment for CDI. In the vancomycin-treated group, beta-diversity (P = 0.0059) and macrolide-lincosamide-streptogramin (MLS) resistance genes (P = 0.037) increased after treatment; C. difficile and vancomycin-resistant enterococci (VRE) environmental contamination was found in 53% of patients and 26% of patients, respectively. We found that vancomycin alters the gut microbiota, does not permanently clear C. difficile, and is associated with VRE colonization/environmental contamination. (This study has been registered at ClinicalTrials.gov under registration no. NCT03388268.)IMPORTANCE A gold standard diagnostic for Clostridioides difficile infection (CDI) does not exist. An area of controversy is how to manage patients whose stool tests positive by nucleic acid amplification tests but negative by toxin enzyme immunoassay. Existing data suggest most of these patients do not have CDI, but most are treated with oral vancomycin. Potential benefits to treatment include a decreased risk for adverse outcomes if the patient does have CDI and the potential to decrease C. difficile shedding/transmission. However, oral vancomycin perturbs the intestinal microbiota and promotes antibiotic-resistant organism colonization/transmission. We conducted a double-blinded randomized controlled trial to assess the risk-benefit of oral vancomycin treatment in this population. Oral vancomycin did not result in long-term clearance of C. difficile, perturbed the microbiota, and was associated with colonization/shedding of vancomycin-resistant enterococci. This work underscores the need to better understand this population of patients in the context of C. difficile/ARO-related outcomes and transmission.

Keywords: C. difficile; vancomycin; vancomycin-resistant enterococci.

Copyright © 2021 Fishbein et al.

Figures

FIG 1

Vancomycin effect on C. difficile-colonized patient gut microbiomes. (a) Randomized control trial to test the effect of 10 days of oral vancomycin treatment on health-related outcomes in C. difficile-colonized patients. Patient stool and surfaces were sampled to examine patient microbiomes and environmental contamination of hospital environments. (b) Beta-diversity, as measured by Bray-Curtis dissimilarity, distributions of within-patient comparisons between placebo and vancomycin treatment groups. Dissimilarity was significantly different between treatment groups (**, P = 0.0057) as measured by a Wilcoxon rank sum test. (c) Measurement of alpha-diversity (richness) of microbial species in patient fecal samples over time due to vancomycin treatment. Richness was not significantly affected by vancomycin treatment (P = 0.23), as examined by a two-way analysis of variance. (d) Relative abundance of major antibiotic resistance (AMR) classes before (index) and after (week 8) treatment, averaged across patients. MLS, macrolide-lincosamide-streptogramin.

FIG 2

Patient shedding of C. difficile associated with environmental contamination. (a) Approximate maximum-likelihood tree of 75 C. difficile genomes isolated from patient stool and their environment. Each node represents an isolate found from a patient-time point, where the number before the dash represents the patient and letter after the dash represents the source of the isolate (S, stool; BR, bedrail; C, commode), followed by the number representing the time point of isolation. Isolates colored green represent those recovered from the environment. Color strips represent the outcome of NAAT testing for C. difficile toxins (tcdAB and cdtAB) and in silico MLST typing. (b) Distribution of pairwise single-nucleotide polymorphism (SNP) distances for each patient-isolate group. Distances are visually classfied in two ways: by time point, either between time point comparisons (green) or within time point comparisons (blue), and by source, either environmental-only comparisons (●), stool-to-environment comparisons (▲), or stool-only comparisons (■).

FIG 3

E. faecium isolates associated with VRE patient shedding/environmental contamination. (a) Pairwise average nucleotide identity (ANI) clustogram between Enterococcus isolates, where the color of the box indicates ANI between two isolate genomes. (b) Approximate maximum-likelihood phylogenetic tree of E. faecium isolates. Colored boxes indicate antimicrobial susceptibility testing (AST), where resistance status was determined in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines. (c) Pairwise SNP distances derived from the core genome alignment. P values were generated through a Wilcoxon rank sum test. ****, <0.0001; ***, <0.001.