Responsiveness to pulmonary rehabilitation is related with changes in oral microbiota of people with COPD

Abstract

Introduction: Pulmonary Rehabilitation (PR) is one of the most cost-effective therapies for chronic obstructive pulmonary disease (COPD), with proven benefits in domains such as exercise capacity and quality of life. Despite its clear benefits, patients are not equal-ly responsive to PR. Reasons behind that as well as the role of the airway microbiota in PR effectiveness are currently unknown.Objectives: Here, we explored for the first time, the effects of PR on oral microbiota and inflammatory markers and the link with responsiveness to PR. Study design: 76 participants were enrolled in this prospective cohort study, half of whom integrated a 12-week PR program. During the 6-month follow-up, a total of 417 saliva samples, and data on dyspnoea during exercise (mBorg), exercise capacity (6MWT) and impact of the disease (CAT) were collected. PR responsiveness was defined as overcoming the published minimal clinically important difference for mBorg (-1 point), 6MWT (25m) and CAT (-2 points).Results: PR modulated patients’ microbiota composition and dy-namics. Specifically, an enrichment of Proteobacteria (Haemophi-lus) and a depletion in Bacteroidetes (Prevotella), previously as-sociated with increased severity (Melo-Dias et al, Respir Res 2022), were observed upon PR. We also observed changes in the levels of IL-1b, TNF-a and IL-10. When separating patients in responders (R) and non-responders (NR), distinct patterns of bacteria/bacteria and bacteria/inflammatory marker longitudinal correlation were ob-served among the groups. In R, the increase in Prevotella nega-tively correlated with Lautropia (enriched in most severe cases of COPD (Melo-Dias et al., Respir Res 2022)). The opposite trend was observed in NR, with Lautropia showing a positive correlation with several pro-inflammatory markers. Conversely, in all groups of R, Rothia and Gemellaceae presented negative correlations with sev-eral pro-inflammatory markers. Conclusions: Overall, despite responsiveness to PR being multidi-mensional and heterogeneous, giving rise to a moderate overlap across domains in individual response, PR-induced changes in mi-crobiota revealed surprisingly consistent patterns among R and NR. Future studies should address the implications and stability of these findings.