A Multilevel Model-Based Analysis Of The Association Between Atmospheric PM_2.5Exposure,Different Dietary Patterns And Dyslipidemia

Objective:To analyze the relationship between fine particulate matter(PM_2.5)exposure,different dietary patterns and dyslipidemia,and to further explore whether the two together influence the development of dyslipidemia.Methods:A baseline survey of 17,274residents aged 35-74 years in Urumqi,Ili and Hotan regions based on the"Xinjiang Multi-ethnic Natural Population Cohort Construction and Health Follow-up Study"was conducted.To understand the general condition of the study subjects and the prevalence of dyslipidemia by t-test and chi-square test,and to describe the PM_2.5 exposure;factor analysis was used to construct the dietary pattern of the study subjects and analyze the dietary characteristics of the residents.A Multilevel logistic regression models were used to analyze the association between PM_2.5 exposure,different dietary patterns and dyslipidemia and dyslipidemia fractions,using the street/village level as the second level and the individual level as the first level;in addition,a single-level logistic regression model was fitted and compared with a multilevel logistic regression model.Results:（1）The prevalence of dyslipidemia in the study population was 33.76%,with hypertriglyceridemia and hypoalphalipoproteinemia being the most common types.（2）Factor analysis was conducted on the diets of the study subjects to extract the traditional dietary pattern,the high quality protein dietary pattern,the savoury nut dietary pattern and the juice drink dietary pattern.（3）Fitting the null model indicated that the data from this study were suitable for multilevel analysis.（4）Multilevel logistic regression showed that each 10μg/m³ increase in atmospheric PM_2.5 was associated with a 7%increase in the risk of dyslipidemia（OR=1.07,95%CI=1.01～1.13）and a 14%increase in the risk of hypoalphalipoproteinemia（OR=1.14,95%CI=1.04～1.24）.（5）Multilevel logistic regression showed that the high-quality protein dietary pattern was associated with a low risk of dyslipidemia（T3 vs T1:OR=0.91,95%CI=0.83～0.99）,while the juice drink dietary pattern was associated with a high risk of dyslipidemia（T3 vs T1:OR=1.13,95%CI=1.04～1.23）.A multilevel logistic regression analysis of dietary patterns and dyslipidemia components showed that a high-quality protein dietary pattern was associated with a low risk of hyperbetalipoproteinemia（T3 vs T1:OR=0.72,95%CI=0.60～0.87）,hypoalphalipoproteinemia（T3 vs T1:OR=0.85,95%CI=0.76～0.95）,and Juice-drink dietary pattern was associated with high risk of hypertriglyceridemia（T3 vs T1:OR=1.21,95%CI=1.08～1.36）,hypercholesterolemia（T3 vs T1:OR=1.19,95%CI=1.04～1.35）,hyperbetalipoproteinemia（T3 vs T1:OR=1.31,95%CI=1.09～1.58）and hypoalphalipoproteinemia（T3 vs T1:OR=1.13,95%CI=1.01～1.26）were associated with a high risk.（6）The results of the multilevel model differed slightly from those of the single-level model in terms of the association and interaction between different dietary patterns and dyslipidemias,with the multilevel model reflecting the aggregation of dyslipidemia in geographical areas and fitting the data with a hierarchical structure better than the single-level model.Conclusion:PM_2.5 causes increased risk for dyslipidemia,especially for hypoalphalipoproteinemia;The high quality protein dietary pattern is a protective factor for dyslipidemia,especially for hypoalphalipoproteinemia;the juice drink dietary pattern is risk factor for dyslipidemia,with high risk for all types of dyslipidemia;these results still need to be explored in large prospective cohort studies.