Thyroid cancer and endocrine disruptive chemicals: a case-control study on per-fluoroalkyl substances and other persistent organic pollutants

Abstract

Objective: The aim was to evaluate the possible association between some endocrine disruptive chemicals and thyroid cancer (TC) in an Italian case-control cohort.

Methods: We enrolled 112 TC patients and 112 sex- and age-matched controls without known thyroid diseases. Per- and poly-fluoroalkyl substances (PFAS), poly-chlorinated biphenyls (PCBs), and dichlorodiphenyltrichloroethane (4,4'-DDT and 4,4'-DDE) were measured in the serum by liquid or gas chromatography-mass spectrometry. Unconditional logistic regression, Bayesan kernel machine regression and weighted quantile sum models were used to estimate the association between TC and pollutants' levels, considered individually or as mixture. BRAFV600E mutation was assessed by standard methods.

Results: The detection of perfluorodecanoic acid (PFDA) was positively correlated to TC (OR = 2.03, 95% CI: 1.10-3.75, P = 0.02), while a negative association was found with perfluorohexanesulfonic acid (PFHxS) levels (OR = 0.63, 95% CI: 0.41-0.98, P = 0.04). Moreover, perfluorononanoic acid (PFNA) was positively associated with the presence of thyroiditis, while PFHxS and perfluorooctane sulfonic acid (PFOS) with higher levels of presurgical thyroid-stimulating hormone (TSH). PFHxS, PFOS, PFNA, and PFDA were correlated with less aggressive TC, while poly-chlorinated biphenyls (PCB-105 and PCB-118) with larger and more aggressive tumors. Statistical models showed a negative association between pollutants' mixture and TC. BRAF V600E mutations were associated with PCB-153, PCB-138, and PCB-180.

Conclusion: Our study suggests, for the first time in a case-control population, that exposure to some PFAS and PCBs associates with TC and some clinical and molecular features. On the contrary, an inverse correlation was found with both PFHxS and pollutants' mixture, likely due to a potential reverse causality.

Keywords: BRAF; PCB; PFAS; endocrine disruptive chemicals; thyroid cancer.

Figure 1

Comparison of PFAS, PCBs, and 4,4′-DDE levels between cases and controls. Serum concentrations of perfluoroalkyl substances (PFAS) and of total lipids-adjusted polychlorobiphenyls (PCBs) and 4,4′-dichlorodiphenyldichloroethylene (4,4′-DDE), which reached a detection frequency greater than 20% in both cases and controls, are represented as boxplots. PFDA, perfluorodecanoic acid; PFHxS, perfluorohexanesulfonic acid; PFNA, perfluorononanoic acid; PFUdA, perfluoroundecanoic acid; PFOA, perfluorooctanoic acid; totPFOS, total perfluorooctane sulfonic acid; branPFOS, branched PFOS; linPFOS, linear PFOS.

Figure 2

Univariate exposure–response relationships (95% CI) between pollutants’ concentrations and thyroid cancer risk. The Bayesian kernel machine regression (BKMR) model was computed between ln-transformed serum concentrations of perfluoroalkyl substances (PFAS) and of total lipids-adjusted polychlorobiphenyls (PCBs) and 4,4′-DDE, and thyroid cancer fixing all other compounds at the median concentration. The model was adjusted for sex, age, BMI, familial history of thyroid diseases, and the use of nail polish/make up remover. PFHxS, perfluorohexanesulfonic acid; PFOA, perfluorooctanoic acid; PFNA, perfluorononanoic acid; linPFOS, linear perfluorooctane sulfonic acid; branPFOS, branched perfluorooctane sulfonic acid; PFDA, perfluorodecanoic acid; 4,4′-DDE, 4,4′-dichlorodiphenyldichloroethylene; PCB, polychlorobiphenyl.

Figure 3

Interactions among pollutants determining thyroid cancer risk in the BKMR model. Individual effect of pollutants on thyroid cancer when the concentration of all other compounds are fixed at their 25th percentile, their median or their 75th percentile. The Bayesian kernel machine regression (BKMR) model was computed with ln-transformed serum concentrations of perfluoroalkyl substances (PFAS), and of total lipids-adjusted polychlorobiphenyls (PCBs) and 4,4′-DDE. The model was adjusted for sex, age, BMI, familial history of thyroid diseases, and the use of nail polish/make up remover. Est, estimate; Q. fixed, quartile fixed; PCB, polychlorobiphenyl; 4,4′-DDE, 4,4′-dichlorodiphenyldichloroethylene; PFDA, perfluorodecanoic acid; branPFOS, branched perfluorooctane sulfonic acid; linear perfluorooctane sulfonic acid; PFNA, perfluorononanoic acid; PFOA, perfluorooctanoic acid; PFHxS, perfluorohexanesulfonic acid.

Figure 4

Schematic representation of possible roles of PFAS on thyroid cell and carcinogenesis. PFAS might affect hormone-dependent pathways increasing TSH levels. On the other hand, PFAS might affect hormone-independent pathways (A) increasing intracellular levels of reactive oxygen species (ROS), which lead to oxidative stress and DNA damage (B) inducing a suppression of cellular senescence (C) promoting an inflammatory status and/or modulating immune system response which finally leads to thyroiditis (figure created by BioRender.com – agreement number: HO26FQPCM9).