Mixed exposure to phenol, parabens, pesticides, and phthalates and insulin resistance in NHANES: A mixture approach

Abstract

Purpose: The effects of environmental chemicals on insulin resistance have attracted extensive attention. Previous studies typically focused on the single chemical effects. This study adopted three different models to analyze the mixed effects of nine common chemicals (one phenol, two parabens, two chlorophenols and four phthalates) on insulin resistance.

Methods: Urinary concentrations of chemicals were extracted from National Health and Nutrition Examination Survey (NHANES) 2009-2016. Insulin resistance was assessed using homeostatic model assessment (HOMA) and defined as HOMA-IR >2.6. The generalized linear regression (GLM), weighted quantile sum regression (WQS) and Bayesian kernel machine regression models (BKMR) were applied to assess the relationship between chemical mixture and HOMA-IR or insulin resistance.

Results: Of the 2067 participants included, 872 (42.19 %) were identified as insulin resistant. In single-chemical GLM model, di-2-ethylhexyl phthalate (DEHP) had the highest parameter (β/OR, 95 % CIs) of 0.21 (quartile 4, 0.12- 0.29) and 1.95 (quartile 4, 1.39- 2.74). Similar results were observed in the multi-chemical models, with DEHP (quartile 4) showing the positive relationship with HOMA-IR (0.18, 0.08- 0.28) and insulin resistance (1.76, 1.17- 2.64). According to WQS models, the WQS indices were significantly positively correlated with both HOMA-IR (β: 0.07, 95 % CI: 0.03- 0.12) and insulin resistance (OR: 1.25, 95 % CI: 1.03- 1.53). DEHP was the top-weighted chemical positively correlated with both HOMA-IR and insulin resistance. In the BKMR model, the joint effect was also positively correlated with both outcomes. DEHP remained the main contributor to the joint effect, consistent with WQS analysis.

Conclusion: Our findings suggested that these chemical mixtures had the positive joint effects on both HOMA-IR and insulin resistance, with DEHP being the potentially predominant driver. The inter-validation of the three models may indicate that reducing the DEHP concentration could improve glucose homeostasis and reduce the risk of insulin resistance. However, further studies are recommended to deepen our findings and elucidate the mechanisms of insulin resistance and chemical mixture.

Keywords: Bayesian kernel machine regression; Chemical mixture; Insulin resistance; Weighted quantile sum regression.