Xufeng Fu, Xiaohua Li, Xinqian Ma, Nan Wu, Wentao Ni, Hui Li
Background Whether a population-scale lifestyle-anchored inflammatory load index and its circulating metabolomic risk signature identify susceptibility to hospital-recorded acute respiratory infections (ARIs) has not been well established. This study examined these questions in the UK Biobank using the Circadian Imbalance Index (CII) and a nuclear magnetic resonance (NMR)–derived metabolomic risk signature (MRS). Methods This prospective cohort study included 295,679 UK Biobank participants after prespecified exclusions. The CII was derived from five baseline components: evening chronotype, atypical sleep duration, high neuroticism, atypical coffee intake, and low vitamin D, and was categorized as low (0-1), intermediate (2-3), or high (4-5). Cox proportional-hazards models evaluated associations between CII and incident hospital-recorded ARIs. Among 264,889 participants with complete baseline Nightingale NMR metabolomics, an MRS was derived from 249 standardized biomarkers by regressing standardized CII on the biomarker panel using elastic net regularized regression. Counterfactual mediation models estimated natural direct and indirect associations for lower respiratory tract infection (LRTI) and pneumonia. Results Higher CII was associated with greater risk across hospital-recorded ARI outcomes. In the fully adjusted model, high versus low CII was associated with higher risks of LRTI (hazard ratio [HR], 1.56; 95% confidence interval [CI], 1.49 to 1.63) and pneumonia (HR, 1.56; 95% CI, 1.48 to 1.64). Ten-year cumulative incidence of LRTI increased from 3.87% in the low CII group to 6.56% in the high CII group; corresponding pneumonia estimates were 2.69% and 4.52%. The MRS accounted for 18.5% (95% CI, 16.2% to 20.7%) of the continuous-CII association with LRTI and 19.0% (95% CI, 16.4% to 21.6%) of the association with pneumonia. Conclusion In a population-scale cohort, a lifestyle-anchored inflammatory load index showed a graded association with hospital-recorded ARIs, and a circulating NMR-derived MRS, dominated by fatty-acid-related biomarkers and including the established low-grade-inflammation marker glycoprotein acetylation (GlycA), accounted for approximately one-fifth of the association with LRTI.