Abstract
AbstractBackgroundThe gut-lung axis is generally recognized, but there are few large studies of the gut microbiome and incident respiratory disease in adults.Objectives3To investigate the associations between gut microbiome and respiratory disease and to construct predictive models from baseline gut microbiome profiles for incident asthma or chronic obstructive pulmonary disease (COPD).MethodsShallow metagenomic sequencing was performed for stool samples from a prospective, population-based cohort (FINRISK02; N=7,115 adults) with linked national administrative health register derived classifications for incident asthma and COPD up to 15 years after baseline. Generalised linear models and Cox regressions were utilised to assess associations of microbial taxa and diversity with disease occurrence. Predictive models were constructed using machine learning with extreme gradient boosting. Models considered taxa abundances individually and in combination with other risk factors, including sex, age, body mass index and smoking status.ResultsA total of 695 and 392 significant microbial associations at different taxonomic levels were found with incident asthma and COPD, respectively. Gradient boosting decision trees of baseline gut microbiome predicted incident asthma and COPD with mean area under the curves of 0.608 and 0.780, respectively. For both incident asthma and COPD, the baseline gut microbiome had C-indices of 0.623 for asthma and 0.817 for COPD, which were more predictive than other conventional risk factors. The integration of gut microbiome and conventional risk factors further improved prediction capacities. Subgroup analyses indicated gut microbiome was significantly associated with incident COPD in both current smokers and non-smokers, as well as in individuals who reported never smoking.ConclusionsThe gut microbiome is a significant risk factor for incident asthma and incident COPD and is largely independent of conventional risk factors.
Publisher
Cold Spring Harbor Laboratory