Article pubs.acs.org/est
Blood Concentrations of Persistent Organic Pollutants and Prediabetes and Diabetes in the General Population of Catalonia Magda Gasull,†,‡ José Pumarega,†,‡ María Téllez-Plaza,§,∥ Conxa Castell,⊥ Ricard Tresserras,⊥ Duk Hee Lee,# and Miquel Porta*,†,‡,▽ †
Hospital del Mar Research Institute-IMIM, Barcelona, Spain CIBER de Epidemiología y Salud Pública (CIBERESP), Spain § Department of Cardiovascular Epidemiology and Population Genetics, Centro Nacional de Investigaciones Cardiovasculares (CNIC), Madrid, Spain ∥ Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States ⊥ Department of Health, Generalitat de Catalunya, Spain # Department of Preventative Medicine, School of Medicine, Kyungpook National University, Daegu, Korea ▽ School of Medicine, Universitat Autònoma de Barcelona, Spain ‡
S Supporting Information *
ABSTRACT: The aim was to analyze the effects of body mass index (BMI), low-dose exposure, mixtures of persistent organic pollutants (POPs), and lipid adjustment on the relationship between POP concentrations and diabetes and prediabetes in the general adult population of Catalonia (Spain). Serum concentrations of POPs were measured by gas chromatography with electron-capture detection in 886 participants in a health interview survey. The highest concentrations of all POPs analyzed were found in subjects who had diabetes. Levels were also higher in individuals with prediabetes than in subjects without the disorder. In models adjusted by age, sex and BMI, the prevalence of diabetes and prediabetes increased in a dose-dependent manner across quartiles of PCBs 118, 138, 153, and 180, and HCB. When models were further adjusted for lipids, the associations were slightly lower and statistically significant, the ORs for the upper quartile ranging from 2.0 to 2.8 (all p-values for linear trend 85% of participants: p,p′-DDT, p,p′-DDE, PCB congeners 118, 138, 153, and 180, HCB and β-HCH.22 Total cholesterol and triglycerides levels were determined enzymatically (Txad-Pap and CIN-UV methods, respectively), using serum obtained in the health examination.22 Total serum lipids (TL) were calculated by the standard short formula (or Standard formula 2) based on total cholesterol and triglycerides,34,35 and POP concentrations were individually corrected for TL by dividing the crude serum POP concentration by TL. Statistical Analysis. Univariate statistics were computed as customary.36 Fisher’s exact test for homogeneity was applied to assess the relationship between two categorical variables. To assess differences on POP concentrations by diabetes status Student’s t-test and Mann−Whitney’s U-test were used. To estimate the magnitude of the associations between POPs and diabetes and prediabetes, multivariate-adjusted odds ratios (ORs) and their corresponding 95% confidence intervals (CI) were calculated by unconditional logistic regression37 with progressive degrees of adjustment, including (1) crude models, with unadjusted POP concentrations (in ng/mL); (2) models adjusted for socio-demographic variables (including age, sex, and BMI); (3) models further adjusted for total cholesterol and triglycerides; and (4) models using TL-corrected POPs (in ng/ g lipid). Including arterial pressure in the models (as mean value or as hypertension, dichotomous) did not significantly change the results; thus, it was not included in the final models. POP concentrations were introduced in the regression model as quartiles categories; we also evaluated the increase in the odds comparing the 80th and the 20th percentiles of POP concentrations from regression models with log-transformed POPs concentrations. We assessed linear relations by testing the log−linear POPs concentrations coefficients using the Wald test and the multivariate analogue of Mantel’s extension test for linear trend. We evaluated nonlinear relations using restricted quadratic splines with knots at the 20th, 50th, and 80th percentiles of each POP distribution.37 The p-value for nonlinear trend was estimated by testing the nonlinear spline terms using Wald’s test for multiple coefficients. The joint effects of POPs and BMI were explored graphically, in stratified analyses, and by including in the model the two main terms and the interaction term.36 In order to evaluate POPs mixtures we calculated the sum of orders or sum of category ranking of the eight most prevalent POPs mentioned above. Each POP was thus categorized in quartiles and the category number of each POP was summed, producing a value ranging between 8 (when all eight POPs had concentrations in the lowest quartile) and
examining in this population relationships that have been assessed in previous studies. Therefore, the aim of the present study was to analyze the relationship between POP serum concentrations and type 2 diabetes and prediabetes in the general adult population of Catalonia (Spain), as well as to elucidate the potential confounding or modifying effects on such relationship of BMI, lipids, low-dose exposure, and mixtures of POPs.
■
MATERIALS AND METHODS Study Population. The study population has been described in detail elsewhere.22 Briefly, participants in the Catalan Health Interview Survey (CHIS) aged 18−74 years were offered to take part in a health examination, which included a physical exam, a supplementary interview, and the collection of urine and blood samples. A total of 1374 individualswho gave specific written informed consent participated during 2002 in the health examination. Trained nurses recorded the weight and height, and the corresponding body mass index (BMI) was computed (measured weight (kg) divided by measured height squared (m2)).23 Information on blood concentrations of lipids and at least 1 mL of serum (for POP analyses) was available from 919 participants.22 The present report is based on 886 participants with data available on POP serum concentrations and on diabetes status (see below). There were no statistically significant differences between the 886 individuals and the remaining participants in the health examination with respect to age, sex, BMI, diabetes status, educational level, and occupational social class.22 Diabetes Status. A capillary blood sample obtained during the health examination was used to determine glucose concentration in whole blood. The fasting plasma glucose concentration was then calculated by multiplying the whole blood glucose by 1.12.24,25 Following previous epidemiologic studies,2,25−29 and recommendations of the American Diabetes Association,30 the World Health Organization,31 and a National Toxicology Program Workshop,1 participants were considered to have diabetes if (a) their fasting plasma glucose was ≥126 mg/dL, or (b) they answered affirmatively to the question “Do you have or did a doctor tell you that you have diabetes?” (only four individuals were classified as diabetic because they answered affirmatively to the question while they reported no current use of insulin or antidiabetic medication and their fasting plasma glucose was 100 mg/dL, and four pregnant women, leaving the mentioned 886 participants. Analysis of Serum Concentrations of POPs and Lipids. Laboratory methods have also previously been described in detail.22 The following 19 POPs were analyzed in serum: o,p′DDT (dichlorodiphenyltrichloroethane), p,p′-DDT, o,p′-DDE (dichlorodiphenyldichloroethene), p,p′-DDE, o,p′-DDD (dichlorodiphenyldichloroethane), p,p′-DDD; PCB (polychlorinated biphenyls) congeners 28, 52, 101, 118, 138, 153, and 180; PeCB (pentachlorobenzene), HCB (hexachlorobenzene), αHCH (hexachlorocyclohexane), β-HCH, γ-HCH, and δ-HCH. 7800
dx.doi.org/10.1021/es300712g | Environ. Sci. Technol. 2012, 46, 7799−7810
Environmental Science & Technology
Article
Table 1. Main Characteristics of the Study Subjects by Diabetes Status characteristic
total
normal subjects
subjects with prediabetes
subjects with diabetes
number of subjects (%) age (years)d median sex (% of males) body mass index (kg/m2)d median underweight (
