Environ. Sci. Technol. 2009, 43, 7909–7915
Organochlorine Compounds, Iodine Intake, and Thyroid Hormone Levels during Pregnancy M A R A L V A R E Z - P E D R E R O L , * ,†,‡,§,| MO ` N I C A G U X E N S , †,‡,§ ´ S I B A R L U Z E A , §,⊥ JESU MARISA REBAGLIATO,# A ` GUEDA RODRIGUEZ,∇ MERCEDES ESPADA,O FERNANDO GON ˜ I , §,⊥ M I K E L B A S T E R R E C H E A , §,⊥ A N D J O R D I S U N Y E R †,‡,§,| Centre for Research in Environmental Epidemiology, Doctor Aiguader 88, 08006 Barcelona, Spain, Municipal Institute of Medical Research (IMIM-Hospital del Mar), Doctor Aiguader 88, 08006 Barcelona, Spain, CIBER Epidemiologia y Salud Pu ´ blica (CIBERESP), Doctor Aiguader 88, 08006 Barcelona, Spain, Pompeu Fabra University, Doctor Aiguader 88, 08006 barcelona Barcelona, Spain, Subdireccio´n Salud Pu ´ blica Gipuzkoa, Departamento de Sanidad del Gobierno Vasco, Avenida Navarra 4, 20013 San Sebastia´n, Spain, Miguel Hernandez University, Avenida de la Universidad, 03202 Elche, Spain, Hospital Parc Taulı´, Parc Taulı´ 08208 Sabadell, Spain, and Laboratorio Normativo de Salud Pu ´ blica, Departamento de Sanidad Gobierno Vasco, Marı´a Dı´az de Haro 58, 48010 Bilbao, Spain
Received March 9, 2009. Revised manuscript received July 27, 2009. Accepted July 29, 2009.
The effect of organochlorine compounds (OCs) on thyroid function, as well as the potential confounding effect of iodine intake, was studied in a large sample of pregnant women from two population-based cohort studies in Sabadell (n ) 520) and Gipuzkoa (n ) 570), Spain. Thyroid hormones (free T4 and total T3), thyrotropin, and polychlorinated biphenyls (PCB congeners 118, 138, 153, and 180), hexachlorobenzene (HCB), betahexachlorocyclohexane (β-HCH), dichlorodiphenyl dichloroethylene (p′p′-DDE) and dichlorodiphenyl trichloroethane (p′pDDT) were measured in serum samples collected at first trimester of pregnancy. Urinary iodine concentration (UIC) was measured and iodine intake from diet, iodized salt, and supplements were estimated from a food frequency questionnaire. Levels of HCB and PCBs congeners 180, 153, and 138 were related to lower total T3 levels (adjusted coefficient (SE): -4.0(1.1), -6.1(1.6), -5.5(1.6), and 3.8(1.4), respectively) and higher free T4 levels (adjusted coefficient (SE): 0.013(0.005), 0.017(0.007), 0.016(0.007), and 0.007(0.006), respectively). These associations * Corresponding author phone: +34 93 214 73 04; fax: +34 93 214 73 01; e-mail:
[email protected]. † Centre for Research in Environmental Epidemiology. ‡ Municipal Institute of Medical Research (IMIM-Hospital del Mar). § CIBER Epidemiologia y Salud Pu ´ blica (CIBERESP). | Pompeu Fabra University. ⊥ Subdireccio´n Salud Pu ´blica Gipuzkoa, Departamento de Sanidad del Gobierno Vasco. # Miguel Hernandez University. ∇ Hospital Parc Taulı´. O Laboratorio Normativo de Salud Pu ´ blica, Departamento de Sanidad Gobierno Vasco. 10.1021/es9007273 CCC: $40.75
Published on Web 09/10/2009
2009 American Chemical Society
were homogeneous in both cohorts, especially for PCBs and total T3 (p-value for the interaction between cohorts >0.8). Iodine intake and UIC did not affect the association between OCs and thyroid hormones. Our results indicate that exposure to OCs during pregnancy can alter TH levels.
Introduction Organochlorine compounds (OCs) are persistent environmental contaminants that were intensively used in the past as pesticides (dichlorodiphenyl trichloroethane (p,p′-DDT), hexachlorobenzene (HCB), beta-hexachlorocyclohexane (βHCH)) or for a variety of applications, including dielectric fluids for capacitors and transformers, plastics, or reactive flame retardants (PCBs). HCB was also formed as an industrial byproduct in chlorination processes. Although the use of most OCs was severely restricted or banned in more than 160 countries during the 1970s, and most are not longer in production, some of these compounds are still released into the environment because of ongoing use in developing countries and improper storage or disposal in developed countries (1). Moreover, they persist in the environment and accumulate in the food chain and in human tissues, thus, in recent years, OCs have been detected in human milk, blood, and adipose tissue in the general population (2, 3). Some OCs have been shown to alter thyroid hormone (TH) concentrations in animals (4, 5) and humans (6). THs are essential for normal development of the human fetal brain and the fetus is dependent on maternal TH (7). Prenatal exposure to background concentrations of some OCs has been shown to have a neurological impact in infants (8), therefore thyroid function disruption has been proposed as a possible mechanism of action for the neurotoxicity of some OCs (9, 10). Although several studies in humans have shown an association between TH and OCs in humans, the results are controversial (6). In a previous study in children, we found that some OCs were negatively associated with total triiodothyronine (total T3), but we found little or no association with free thyroxine (free T4) or thyrotropin (TSH), respectively (11). There are only four studies in pregnant women (12-15), most of them with small sample sizes, nonetheless total T3 was measured in two of these studies and also found to be negatively related to OCs (12, 13). Given the importance of TH during pregnancy, thyroid toxicity of OCs during this period should be more intensively studied. Taking into account the four existing studies in pregnant women, T3 seems to be the most disrupted thyroid hormone during pregnancy, nevertheless further epidemiologic and experimental studies should be undertaken to clarify the mechanisms of toxicity. Moreover, no studies have evaluated the potential confounding effect of iodine in the relation between OCs and TH in humans. Iodine is essential for the synthesis of thyroid hormones (16), and marine fish is one of the main known dietary sources of iodine (17). Fish, especially fatty fish, is also known to be an important source of OCs (18), thus the association between OCs and TH could be confounded by the iodine intake. The objective of this study was to evaluate the effects of background exposure to some OCs on levels of TH (free T4 and total T3) and TSH, and the potential confounding effect of iodine intake among 1090 pregnant women from two general population-based mother and child cohort studies, with a particular interest in T3. VOL. 43, NO. 20, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7909
Methods Study Population. The INMA (INfancia y Medio Ambiente (Environment and Children)) project is a multicenter population-based mother and child cohort study established in seven areas of Spain. The aim of the project is to evaluate the impact of exposure to different contaminants on fetal and infant growth, health, and development (19). In this study, we analyzed the crossectional data obtained from pregnant women during the first trimester of pregnancy from two cohorts: Sabadell (Catalonia) and Gipuzkoa (Basque Country). Pregnant women (n ) 1297, 657 from Sabadell and 640 from Gipuzkoa) were recruited at their first routine antenatal care visit in the public health center or hospital of reference for defined geographical areas. The recruitment period was from July 2004 to July 2006 in Sabadell, and from April 2006 to January 2008 in Gipuzkoa. Among them, only those women who had been born in Spain, who had a serum sample available for TH and OC analysis, and questionnaire data were included in the study sample (n ) 553 in Sabadell and n ) 602 in Gipuzkoa). Questionnaires and biologic samples were obtained at the same time. Moreover, women who had a thyroid pathology (n ) 33 in Sabadell and n ) 31 in Gipuzkoa) and two women from Gipuzkoa with extreme levels of OCs were excluded from the analysis. Finally, 520 and 570 women from Sabadell and Gipuzkoa, respectively, were included in the subsequent analysis. Information on education, socioeconomic background, demographic factors, marital status, maternal health and obstetric history, parity, medication use, alcohol and smoking habits during pregnancy, anthropometric measures, and dietary intake was obtained through questionnaires administered during the first trimester of pregnancy. Informed consent was signed and the study was approved by the ethics committee of the Municipal Institute of Medical Research, Barcelona. Thyroid Hormones Measurement. TSH, total T3, and free T4 concentrations were measured in serum samples at the end of the first trimester of pregnancy (Normative Public Health Laboratory of Bilbao, Basque Country) using a solidphase, time-resolved sandwich fluoroimmunoassay (AutoDELFIA, PerkinElmer Life and Analytical Sciences, Wallac Oy, Turku, Finland) using a lanthanide metal europium (Eu) label. The between-assay coefficients of variation (CVs) at three levels of hormone concentration were 3.0, 3.1, and 2.6% for low, medium, and high concentrations of TSH, respectively, and 6.1, 4.1, and 4.0 for free T4. The intra-assay CVs were 7.7, 2.1, and 1.7% for TSH and 3.7, 3.0, and 3.3% for free T4. Samples were stored at -20 °C prior to the analyses. Iodine. Urinary iodine concentrations (UIC) were measured in spot urine samples during the first trimester in a total of 245 out of 520 and 569 out of 570 pregnant women included in the study from Sabadell and Gipuzkoa, respectively, using paired-ion reversed-phase high-performance liquid chromatography with electrochemical detection and a silver working electrode. Samples were collected at the same time as the blood samples and questionnaires, and stored at -20 °C prior to the analyses. In Sabadell UIC was only measured in 303 of the recruited women (245 of the women included in the subsequent analysis) given that the collection of urine samples for iodine quantification was begun at half recruitment. Those women who did not provide a urinary sample but were included in the analysis had higher T3 and T4 levels (p-value < 0.05) and they were older (pvalue < 0.05) than those women included in the analysis who provided a urine sample for iodine measurement. UIC were not adjusted for dilution since creatinine levels were not available. Food intake during the three months previous to the interview was assessed in all women by a food-frequency questionnaire (FFQ) administered by two trained nurses at 7910
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 20, 2009
the end of the first trimester of pregnancy. The frequency of food intake was reported per day, week, or month and portions were reported based on standard referent portion sizes, as pieces, glasses, cups, spoons, centiliters, or grams. The questionnaire is an adapted version of the FFQ developed by Willett (20), and it has been used and validated in a general population-based study in Valencia (Spain) (21). Iodine intake from food and salt was estimated and adjusted for total energy by the residual method. The type of salt consumed was asked for the time period beginning three months before conception and ending on the date of the interview. The production of iodized table salt at 60 mg of iodine per kg is regulated in Spain by law since 1982, but its use is not compulsory and its availability is not accompanied by a sustained educational campaign (22). In those women who reported consuming iodized salt, iodine content was taken into account using the FFQ item about “bits of salt” usually added at the table to raw and cooked foods, and also in the amount of salt assumed to be added in the cooking process of some meals. Information on dose and timing of consumption of specific potassium iodide supplements or vitamin/mineral preparations containing iodine was also obtained by a structured questionnaire. Daily iodine intake from food, salt, and supplements was dichotomized as adequate or inadequate according to the daily recommended iodine intake proposed for pregnant women by the World Health Organization (>250 µg/day) (23). Organochlorine Compounds Analysis. OCs (HCB, β-HCH, p,p′-DDT, p,p′-DDE, and PCB congeners 28, 118, 138, 153, and 180) were analyzed in serum using methods described elsewhere (24). Initial extraction was performed using 96well solid-phase extraction disk plates. Quantification was performed on a 5890 series II gas chromatograph equipped with split-splitless injector, ECD detector, and a 7673 autosampler (Agilent Technologies, Wilmington, DE). Quantitative and qualitative confirmation was carried out on a 6890 gas chromatograph equipped with EPC split-splitless injector, 5973 quadrupole MS detector, and a 7683 autosampler (Agilent Technologies, Wilmington, DE). The limit of detection was 0.071 ng/mL. All OCs serum concentrations were lipid-adjusted using the method described by Phyllips and colleagues (25). Cholesterol and triglicerides were measured using CHOD-POD and GPO-POD enzymatic colorimetric methods, respectively (Spinreact, Santa Coloma, Spain). Statistical Analyses. We conducted a cross-sectional analysis among pregnant women at first trimester of pregnancy to assess the relationships between TH and TSH levels (outcome variables) and OCs (exposure variables), and the potential confounding effect of iodine intake. The relationships were measured through the coefficients provided by the bivariate and multivariate linear regression models. Interquartile range increases (IQR) were also calculated in the multivariate models. Values of OCs below the detection limit were substituted with one-half of the detection limit. OCs and TSH showed a non-normal distribution and were log transformed (natural logarithm) before being included in the models as continuous variables. The analyses were first stratified by cohort, but due to nonheterogeneity (pvalue for the cross-product term of exposure vs cohort >0.10) between both cohorts in most of the bivariate models, the results of the two cohorts were pooled and adjusted for cohort in the multivariate models, except for the model including β-HCH, which was stratified by cohort given that only 46% of the subject from Gipuzkoa had detectable levels. Multivariate linear regression models were adjusted for those variables that changed the magnitude of one of the main effects by 10% or more (age, weight, parity, smoking habits during pregnancy, educational level, and gestational age at sampling). Multiple pollutant models were also performed and included more than one oranochlorine compound in
TABLE 1. Thyroid Hormone Concentrations and Organochlorine Levels at First Trimester of Pregnancy in the Cohort of Sabadell (n = 520) and the Cohort of Gipuzkoa (n = 570)a Sabadell % detected thyroid hormones and thyrotropin free T4 (ng/dL) T3 (ng/dL) TSH (mU/L) -
Gipuzkoa
PC25
PC50
PC75
% detected
PC25
PC50
PC75
0.76 139 0.80
0.82 156 1.20
0.89 173 1.74
-
0.72 141 0.83
0.77*** 156 1.28**
0.83 173 1.91
42.4 112.2