Same drug, same protocol, different results: is geography the hidden variable in ORBIT trials?

To the Editor Hull et al. (2025) emphasize the connection between microbiome-inflammation endotypes and responses to inhaled antibiotics in bronchiectasis patients with Pseudomonas aeruginosa.1 Although their study backs precision medicine, we contend that regional differences overlooked in the ORBIT-3/4 trials skewed treatment assessments. This is vital because inconsistent trial outcomes have obstructed inhaled antibiotic approvals, despite patient demand. Hull et al. examined twin trials (ORBIT-3 and ORBIT-4) with identical designs but differing results: ORBIT-4 showed a 37% reduction in annual exacerbation rate, while ORBIT-3 showed no significant improvement. The authors suggest this discrepancy is due to geographic differences, noting that patients from Central/Eastern Europe had higher microbiota diversity and less neutrophilic inflammation than those from Western regions. Some of the results of these studies have been previously reported in the form of abstracts.2,3 Endotypic differences affect bronchiectasis progression and drug response, but ORBIT-3/4 only used sex, smoking, and exacerbation history for stratified randomization, excluding geography. EMBARC registry data confirm regional variations: microbiology and inflammatory markers significantly differ across European nations.4 In post hoc analysis adjusting for location and microbiome, both trials converged toward ∼20% exacerbation reduction (ORBIT-3 RR 0.81; ORBIT-4 RR 0.82). Thus, ORBIT-4’s efficacy was inflated by ∼37% and ORBIT-3’s underestimated by ∼15%—solely due to unaccounted regional clustering. To address this bias, we suggest 2 evidence-based strategies: Stratified Randomization by Geography & Biomarkers Multi-center trials should stratify participants by region (eg, Western Europe, Eastern Europe, Asia-Pacific) to account for endotypic differences, while using biomarkers (eg, baseline P. aeruginosa load, neutrophil elastase, Rothia abundance) to define subgroups. The 2023 European endotyping study found distinct inflammatory clusters,5 indicating that biomarker-based stratification could predict antibiotic responses. This approach ensures even distribution of high-risk endotypes, like neutrophilic-predominant, across trial groups. Mixed-Effects Models Accounting for Clustering Statistical analyses should use mixed-effects models with geography as a fixed effect and trial centers as a random effect. For exacerbation endpoints, negative binomial regression models should adjust for regional confounders beforehand. Applying this to ORBIT-3/4 data would likely provide consistent efficacy estimates without needing post hoc adjustments. Reporting regional effect sizes (f2) helps quantify clinical relevance, aiding public health bodies in regionally allocating therapies where they are most effective. Failing to address geography undermines trial validity and may obscure a consistent ∼20% reduction in exacerbations across diverse populations. Geographic variation in pathogen prevalence and inflammatory response suggests that “one-size-fits-all” inhaled antibiotic strategies risk benefiting only subgroups while failing others.6 Transparent reporting of regional efficacy data could guide tailored guidelines—for example, targeting high-diversity, low-neutrophil-inflammation phenotypes for inhaled ciprofloxacin therapy. Hull et al. advance the field by linking microbiome-inflammation endotypes with treatment response. We urge that future inhaled antibiotic trials integrate geography robustly in both design and analysis, via region-stratified randomization and multilevel modeling. Such measures will reconcile trial inconsistencies, accelerate precision therapy development, and ensure equitable global respiratory care. The author thanks Hull RC and colleagues for their pioneering work. Dr Bin Deng drafted the manuscript. Dr Wenhua Liu critically revised the manuscript for intellectual content and approved the final version. Supplementary material is available at American Journal of ­Respiratory and Critical Care Medicine online. The author declares no conflicts of interest. Please see the ICMJE disclosure forms, which have been provided as supplementary material. None declared.