Prospective cohort study uncovers pre-diagnostic plasma proteomic signatures in adult participants, indicating potential blood biomarkers for early bronchiectasis risk stratification.
Key Points
To characterize circulating plasma proteomic profiles that precede an initial inpatient diagnosis of bronchiectasis and evaluate their potential utility for early risk assessment.
Assayed baseline plasma levels of 2,911 Olink Explore proteins among 52,916 prospective participants from the UK Biobank Pharma Proteomics Project (44,299 in the primary complete-case cohort).
Modeled associations with first recorded inpatient ICD-10 J47 diagnoses using multivariable Cox models with administrative censoring and a 5-year landmark framework.
Derived a 41-protein signature with elastic-net Cox regression and evaluated predictive performance via repeated nested cross-validation with out-of-fold predictions.
In primary complete-case analysis (601 events), 891 proteins were significantly associated with subsequent bronchiectasis (FDR < 0.05), led by markers such as WFDC2, LAMP3, and MSLN.
At 5 years, discriminative accuracy reached an AUC of 0.774 for the clinical model, 0.858 for the proteomic component, and 0.871 for the combined clinical-proteomic model.
Stratified analyses revealed 1,311 proteins associated with events within 5 years compared to 407 proteins beyond 5 years, demonstrating diverging proteomic trajectories that widen closer to diagnosis.