A Europe-wide characterization of the external exposome: A spatio-temporal analysis

Background: Harmonised data on external environmental exposures are essential for multi-cohort exposome studies. This paper describes the development of fine-spatial resolution models and resulting exposure maps for 33 major exposome factors —including physical–chemical, built, social, and food environments—across Europe from 2000 to 2020, and examines their spatial and temporal interrelations. Methods: New fine spatial resolution Europe-wide models were developed for annual/monthly average air pollution, daily temperature, and annual-average road-traffic noise, combined with post-processing of existing data on green, blue, and grey spaces (imperviousness), walkability, light-at-night, and urbanicity. Exposure metrics relevant for epidemiological studies were developed for all exposome factors, with correlations calculated at European and country levels. Stability and trends over time were assessed for 18 factors. Results: At the European level, most environmental factors showed weak correlations (R < 0.4), except NO₂, which showed moderate to strong correlations with built environment factors. Country-level correlations varied. Annual average exposure surfaces were stable over time, with strong correlations between early and late time points for all factors except O₃​ (R = 0.66). Trends indicated decreases in air pollution and increases in temperature, green space, and imperviousness, while trends in light-at-night and O₃​ were mixed across Europe. Conclusions: This comprehensive analysis of the temporal and spatial relationships between external exposome factors across a large geographical area show low to moderate correlations between exposome factors. Annual average exposure surfaces were also stable over time across Europe. These findings support both the utility of multi-exposure epidemiological analyses, and that any modest temporal misalignment between exposure assessment and follow-up period of health studies is not critical. The data described in his paper are openly available to researchers.