Microbial markers for prediction of clinical outcomes in patients with gastrointestinal diseases

This thesis provides a broad overview of the current knowledge and potential of microbial markers in various gastrointestinal diseases, and of clinical markers in Clostridioides difficile infection (CDI). PART 1: Clinical and microbial markers of disease outcome in patients with CDI The first chapter is a systematic review about clinical prognostic factors for severe and recurrent CDI. 136 studies were included. We found that older patients and patients with multiple comorbidities were at risk for severe CDI. Risk factors for CDI recurrence were a higher age, healthcare-associated CDI, prior hospitalization, proton pump inhibitors started during or after CDI diagnosis, and previous rCDI. Besides the studies on the identification of clinical CDI predictors assessed in our review, several practical risk scores have been developed. Such tools could provide the next step towards an early identification of patients at risk of rCDI and patient-tailored treatment choices. In Chapter 2, we validated two promising prediction tools for rCDI in our own study cohort. Both risk scores performed poorly in our population, with area under the receiver operating characteristic curve scores of lower than 0.50. In addition to clinical factors, the strain type/ribotype of C. difficile is a frequently described risk factor for CDI recurrence and severity. In Chapter 3 we developed a ribotyping method that can be applied directly on fecal samples instead of cultured strains. With a sensitivity and specificity of 100%, this is a reliable and rapid technique that provides information on both the presence and the ribotype of C. difficile. Chapter 4 was a prospective, observational, multicenter study to develop a prediction model for rCDI based on both clinical and microbiological factors. We found that the microbiota composition was more important for the prediction of rCDI than clinical factors. Bacteroidetes abundance and diversity, and the difference in Proteobacteria diversity before- and after start of CDI treatment were the strongest predictors of rCDI. The sensitivity and specificity after cross-validation were 67% and 62%. PART 2: Microbial markers in other diseases Chapter 5 focused on patients with Inflammatory Bowel Disease (IBD), comprising ulcerative colitis (UC) and Crohn’s disease (CD). The objective of this study was to develop a microbiota-based test for discriminating between IBD type and monitoring IBD-related disease activity, to reduce the number of endoscopies in these patients. A microbiota-based model to classify patients by IBD type had a fair accuracy in mucosal biopsies (0.69), performed less in fecal samples (0.64), and poorly in rectal swabs (0.45). Discrimination of patients with active inflammation vs. remission was poor in all sample types. Chapter 6 focused on colonic diverticula. The microbiota composition and mucosal immunity were compared in colon biopsies of 19 patients with asymptomatic diverticulosis and 24 patients without diverticula. We found no significant differences between patients and controls, implying that mucosal inflammation and the colonic microbiota do not play a major role in the pathogenesis of diverticula formation. Chapter 7 is a systematic review of potential microbial markers for oncological outcomes in esophageal cancer. Only five articles were eligible for review, indicating the large knowledge gap regarding this subject. In three studies, presence of Fusobacterium nucleatum was associated with a poor tumor response and/or worse clinical outcomes. In another article, a low abundance of Proteobacteria and high abundances of Prevotella spp. and Streptococcus spp. were correlated with a shortened survival. One (poor quality) study assessed whether surgical outcomes after esophageal cancer surgery could be predicted by the esophageal microbiota, but no direct association was found. In summary, F. nucleatum appears to be a promising microbial marker for clinical outcome in esophageal cancer, but literature on this subject is scarce.