Mediators or confounders:should obesity be adjusted to estimate the effects of trihalomethanes on thyroid function
Objective To explore whether obesity plays a confounding or mediating role in the effects of trihalometh-anes on thyroid function.Methods The cross-sectional data were obtained from the NHANES database.The continuous and categorical variables were compared with weighted rank sum test and χ2 tests,respectively.Variables with skewed distribution were subjected to natural ln transformation;THM was classified according to the quartile distribution of ln transformed concentration;directed acyclic graphs(DAGs)and Logistic regression were introduced to explore the po-tential role of body mass index(BMI)in the causal pathway of THM exposure and thyroid function.Results There was no significant effect on the association between THM exposure and thyroid function before and after additional adjustment for BMI as a categorical variable;however,when stratified analyses were performed based on obesity,THM was a risk factor for elevated TSH levels in obese patients.Conclusion Considering collider bias,the potential hetero-geneity observed in stratified analyses of obesity needs to be interpreted with caution,as well as the need for additional analyses of obesity as a confounder and/or mediator in THM studies.
NETL
NSTL
万方数据
维普
