Association of environmental factors and biological metrics with lung health: an environment-wide association study

Current understanding of respiratory health determinants remains fragmented and lacks comprehensiveness. We aimed to conduct an environmental-wide association study (EWAS) to systematically and unbiasedly identify environmental exposures and biological metrics associated with lung health. Eligible participants aged 40 years or older from the US National Health and Nutrition Examination Survey (1999-2018) were grouped into four overlapping cohorts based on outcomes of interest, including respiratory symptoms, spirometry metrics and patterns, and chronic lung diseases. Individuals in these cohorts were randomly split into training and testing sets, respectively. A total of 215 environmental factors and 167 biological phenotypes were included as candidate exposures. We performed survey-weighted logistic regression and linear models to investigate the association between the above candidates and different respiratory outcomes in the training sets (false discovery rate [FDR] < 5%), positive findings among which were sent for validation in the testing sets (P value < 0.05). Random forest (RF) models were used to determine independent effects and address multicollinearity among EWAS-identified variables. Further exploratory and sensitivity analyses were conducted to ensure the robustness of our findings. A total of 36,252 eligible participants aged 40 years or older were initially enrolled in this study, with four cohorts subsequently formed, varying in size from 8,324 to 36,062. After adjusting for potential confounders, we identified 17, 10, 14, and 31 factors associated with cough, sputum production, wheeze, and exertional dyspnea, respectively, in the training and testing datasets. Moreover, three factors were linked to an obstructive spirometry pattern, and 23 factors to a restrictive pattern. In terms of chronic lung conditions, 12, 7, 14, and 6 variables were correlated with asthma, chronic bronchitis, emphysema, and chronic obstructive pulmonary disease (COPD), respectively. Exploratory and sensitivity analyses validated the robustness of our findings and provided additional insights. Our systematic EWAS provided a comprehensive overview of the relationships of environmental and biological factors with various dimensions of lung health status. These findings added to our understanding of aspects influencing respiratory health and were expected to inform further research and clinical applications.