Association between butyrate-producing gut bacteria and the risk of infectious disease hospitalisation: results from two observational, population-based microbiome studies

Background: Microbiota alterations are common in patients hospitalised for severe infections, and preclinical models have shown that anaerobic butyrate-producing gut bacteria protect against systemic infections. However, the relationship between microbiota disruptions and increased susceptibility to severe infections in humans remains unclear. We investigated the relationship between gut microbiota and the risk of future infection-related hospitalisation in two large population-based cohorts. Methods: In this observational microbiome study, gut microbiota were characterised using 16S rRNA gene sequencing in independent population-based cohorts from the Netherlands (HELIUS study; derivation cohort) and Finland (FINRISK 2002 study; validation cohort). HELIUS was conducted in Amsterdam, Netherlands, and included adults (aged 18–70 years at inclusion) who were randomly sampled from the municipality register of Amsterdam. FINRISK 2002 was conducted in six regions in Finland and is a population survey that included a random sample of adults (aged 25–74 years). In both cohorts, participants completed questionnaires, underwent a physical examination, and provided a faecal sample at inclusion (Jan 3, 2013, to Nov 27, 2015, for HELIUS participants and Jan 21 to April 19, 2002, for FINRISK participants. For inclusion in our study, a faecal sample needed to be provided and successfully sequenced, and national registry data needed to be available. Primary predictor variables were microbiota composition, diversity, and relative abundance of butyrate-producing bacteria. Our primary outcome was hospitalisation or mortality due to any infectious disease during 5–7-year follow-up after faecal sample collection, based on national registry data. We examined associations between microbiota and infection risk using microbial ecology and Cox proportional hazards. Findings: We profiled gut microbiota from 10 699 participants (4248 [39·7%] from the derivation cohort and 6451 [60·3%] from the validation cohort). 602 (5·6%) participants (152 [3·6%] from the derivation cohort; 450 [7·0%] from the validation cohort) were hospitalised or died due to infections during follow-up. Gut microbiota composition of these participants differed from those without hospitalisation for infections (derivation p=0·041; validation p=0·0002). Specifically, higher relative abundance of butyrate-producing bacteria was associated with a reduced risk of hospitalisation for infections (derivation cohort cause-specific hazard ratio 0·75 [95% CI 0·60–0·94] per 10% increase in butyrate producers, p=0·013; validation cohort 0·86 [0·77–0·96] per 10% increase, p=0·0077). These associations remained unchanged following adjustment for demographics, lifestyle, antibiotic exposure, and comorbidities. Interpretation: Gut microbiota composition, specifically colonisation with butyrate-producing bacteria, was associated with protection against hospitalisation for infectious diseases in the general population across two independent European cohorts. Further studies should investigate whether modulation of the microbiome can reduce the risk of severe infections. Funding: Amsterdam UMC, Porticus, National Institutes of Health, Netherlands Organisation for Health Research and Development (ZonMw), and Leducq Foundation.