The role of microbiota in varying response to inhaled antibiotics in bronchiectasis patients with chronic Pseudomonas aeruginosa infection: a post-hoc analysis of ORBIT3 and ORBIT4

<bold>Introduction:</bold> Replicate randomized trials of inhaled antibiotics in bronchiectasis have given conflicting results. We hypothesised that microbiota differences could explain this variation. <bold>Methods:</bold> ORBIT3 and ORBIT4 were global, double-blind, randomized placebo-controlled trials of inhaled liposomal ciprofloxacin in bronchiectasis patients with chronic P. aeruginosa infection. Random forest models searched for differences in relative abundance of species in baseline sputum, measured by full length 16S rRNA sequencing (LoopSeq), between trials and arms. A negative-binomial model for exacerbation frequency examined variation in treatment effect with selected taxa. <bold>Results:</bold> 377/598 patients across both trials had baseline microbiota data. Differences in taxa and diversity were relatively small. Pseudomonas abundance appeared lower in ORBIT4 vs ORBIT3 (p=0.159), and lower in the ORBIT4 treatment arm vs placebo (p=0.134). Rothia abundance appeared higher in the treatment arms vs placebo in ORBIT3 (p=0.157) and ORBIT4 (p=0.184), and higher in ORBIT4 vs ORBIT3 overall (p<10^-6). Modelling estimated greater treatment efficacy at lower Pseudomonas (p=0.180) and higher Rothia abundance (p=0.005), consistent with the larger treatment effect in ORBIT4. Using linear discriminant analysis to adjust for differences in the entire microbiota profile, the treatment estimate from both trials was similar (ORBIT3 RR=0.81, 95% CI 0.54-1.22; ORBIT4 RR=0.82, 95% CI 0.56-1.22). <bold>Conclusion:</bold> Differences in baseline microbiota may explain inconsistent efficacy estimates in replicate randomized clinical trials of inhaled antibiotics.