Changes in serum dioxin and PCB levels in residents around a municipal waste incinerator in Bilbao, Spain

Changes in serum dioxin and PCB levels in residents around a municipal waste incinerator in Bilbao, Spain MB Zubero, 1,2 E Eguiraun, 1 JJ Aurrekoetxea,1,2,3 A Lertxundi, 1,2,4 E Abad,5 J Parera, 5 F Goñi-Irigoyen ,2,3,4 J Ibarluzea,2,3,4 (Miren B Zubero,1,2 Elena Eguiraun,1 Juan J Aurrekoetxea,1,2,3 Aitana Lertxundi,1,2,4 Esteban Abad,5 Jordi Parera,5 Fernando Goñi-Irigoyen, 2,3,4 Jesus Ibarluzea,2,3,4,6) 1 Department of Preventative Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain 2 Biodonostia Health Research Institute, San Sebastian, Spain 3 Public Health Department, Basque Government, San Sebastian, Spain 4 Spanish Consortium for Research on Epidemiology and Public Health (CIBERESP), Spain 5 Laboratory of Dioxins, IDAEA-CSIC, Jordi Girona 18, 08034 Barcelona, Spain 6 Faculty of Psychology, University of the Basque Country UPV/EHU, San Sebastian, Spain Miren B Zubero, Department of Preventative Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain. Biodonostia Health Research Institute, San Sebastian, Spain; [email protected] Elena Eguiraun, Department of Preventive Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain. [email protected] Juan J Aurrekoetxea, Department of Preventive Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain; Public Health Department, Basque Government, San Sebastian, Spain; Biodonostia Health Research Institute, San Sebastian, Spain. [email protected] Aitana Lertxundi, Department of Preventive Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain; Biodonostia Health Research Institute, San Sebastian, Spain; Spanish Consortium for Research on Epidemiology and Public Health (CIBERESP), Spain. [email protected] Esteban Abad, Laboratory of Dioxins, IDAEA-CSIC, Jordi Girona 18, 08034 Barcelona, Spain. [email protected] Jordi Parera, Laboratory of Dioxins,IDAEA-CSIC, Jordi Girona 18, 08034 Barcelona, Spain. [email protected] Fernando Goñi-Irigoyen, Public Health Department, Basque Government, San Sebastian, Spain; Biodonostia Health Research Institute, San Sebastian, Spain; Spanish Consortium for Research on Epidemiology and Public Health (CIBERESP), Spain. [email protected] Jesus Ibarluzea, Public Health Department, Basque Government, San Sebastian, Spain; Biodonostia Health Research Institute, San Sebastian, Spain; Faculty of Psychology, University of the Basque Country UPV/EHU, San Sebastian, Spain; Spanish Consortium for Research on Epidemiology and Public Health (CIBERESP), Spain. [email protected] Corresponding author: Juan J Aurrekoetxea, Biodonostia Health Research Institute, San Sebastian, Spain; Department of Preventative Medicine and Public Health, University of the Basque Country (UPV/EHU), Leioa, Bizkaia, Spain; Public Health Department, Basque Government, San Sebastian, Spain. [email protected] Competing interests: None. Patient consent: Obtained. Ethics approval: This study was conducted with the approval of the Ethics Committee of the University of the Basque Country Funding: This study was funded by the MSWP Zabalgarbi S.A. Keywords: Organochlorine Compounds; Biological Monitoring; Incineration; Serum; Public Health Surveillance; Humans 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 This is the Accepted Manuscript version of a Published Work that appeared in final form in Environmental Research 156 : 738-746 (2017). To access the final edited and published work see https://doi.org/10.1016/j.envres.2017.04.039 © 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/ ABSTRACT There is a great concern in the Basque Country regarding emissions from waste incineration, in particular that of organochlorines (OCs), including dioxins, furans and polychlorinated biphenyls (PCBs), and their potential effect on human health. In 2005, a municipal solid waste plant (MSWP) started to operate in Bilbao, representing an opportunity to assess the exposure to the aforementioned pollutants among people living at various distances from the plant. In 2006 and 2008, we carried out two cross-sectional studies to quantify and assess changes in levels of these pollutants. The objective of this study was to describe the levels of OCs in the blood serum in 2013 of 127 adults of this prospective cohort, in four centres of population, near to and further away from the MSWP, and to study trends over time since it started to operate. This study shows the levels of OCs have decreased significantly, by 78.9% or more. Further, levels of OCs in areas near to the MSWP were not found to be higher than those in areas further afield. ABBREVIATIONS CI: confidence interval dl-PCB: dioxin-like PCB GM: geometric mean MSWP: municipal solid waste plant OC: organochlorine OCDD: octachlorodibenzodioxin PCB: polychlorinated biphenyl PCDD: polychlorinated dibenzodioxin PCDF: polychlorinated dibenzofuran TEQ: toxic equivalency factor DLC: dioxin-like compounds 1. INTRODUCTION The management of the solid waste we produce is among the great challenges facing society today. One of several methods for reducing the volume of this waste is incineration. However, municipal solid waste incineration is a potential source of pollution, given that during the process there is combustion of residues containing chlorine, as well as of heavy metals (Van Caneghem et al., 2010). During combustion processes, there is a release of various pollutants such as organochlorines, including the polychlorinated-pdibenzodioxins, the polychlorinated dibenzofurans, commonly known as dioxins and furans or simply dioxin like compounds (DLCs), and polychlorinated biphenyls (PCBs). Indeed, combustion is the main source of DLCs released into the environment (Domingo et al., 2000). Specifically, among thermal sources, solid waste 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 incineration has been identified as the principal source of DLCs during the 1980s and early 1990s (Jones et al., 1993; Eduljee et al., 1996; Quaß et al., 2000; Kim et al., 2001; Quaß et al., 2004; Burns et al., 2010). Since then, municipal solid waste plants (MSWPs) have become a more efficient method for treating solid waste and more effective at controlling DLC emissions (Ranzi et al., 2011). Nevertheless, there is still concern regarding the environmental and health consequences of waste incineration, and hence, it is important to assess its impact at the environmental level and on people living around MSWPs. A systematic review concluded that the level of evidence for MSWPs having an effect on health was limited for all types of cancers, soft tissue sarcoma, and reproductive and birth defects (orofacial and genitourinary) and inadequate for all the other health problems considered (Mattiello et al., 2013). The opening of a MSWP in Bilbao in 2005 represented a great opportunity to assess exposure to certain pollutants among people living at various distances from the plant. In 2006 and 2008, two cross-sectional studies were carried out to quantify and assess changes in the levels of OCs in a cohort of the general urban population, those with no work-related exposure and living in areas near to and further away from the MSWP (Zubero et al., 2009; Zubero et al., 2011). Given the high toxicity of dioxins and PCBs, a study was carried out to verify if there was a health risk to those close to the MSWP. The objective of studying this new cross-section of the cohort was to assess the levels of DLCs in people in four centres of population, near to and further away from the MSWP and study trends over time since it started to operate. 2. MATERIAL AND METHODS We carried out prospective cohort study with three cross-sectional analyses among people living in areas near to and further away from the MSWP. The sampling was performed in 2006 (Zubero et al., 2009), 2008 (Zubero et al., 2011) and in the third quarter of 2013. The first stage of the study was conducted when the MSWP started to operate, and the data collected then is considered the baseline for subsequent stages of the study. 2.1.STUDY POPULATION Regarding distance to the MSWP, we selected four areas: two areas nearby, Alonsotegi and Rekalde, and two areas further away, Santutxu and Balmaseda, (Zubero et al., 2011). The areas considered close to the MSWP were the town of Alonsotegi and the borough of Rekalde in Bilbao, both less than 2 km away from the plant and in urban environments with high traffic density. The areas further away, used as control areas, were the borough of Santutxu in Bilbao, also an urban environment with high traffic density (5 km away 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 from the plant) and a small town with low traffic density, Balmaseda (20 km away from the plant); both of these areas have little industrial activity and are out of the path of prevailing winds. In relation to the origin of the samples, in 2006, 36% of the participants came from the census in the areas close to the plant, and the rest were volunteers from the health centre. In remote areas, 25% were recruited from the census, and the rest were volunteers. In 2008, in nearby areas, 71% were previous participants, whereas in the remote areas this percentage was 68%. In 2013, we carried out the third stage of the study, recruiting people who had participated in the two previous studies, with the aim of recruiting 30 people from each study area, stratified by sex and age group at baseline (20-69 years old). The exclusion criteria at all stages were: having lived in the area for less than the previous 5 years, having worked in a company which carried out production processes involving potential exposure to DLCs in the previous 15 years, and being 70 years old or above. Having telephoned 223 candidate participants, we finally obtained blood samples from 127 individuals. The main reasons why participants in earlier stages of the study did not participate in this stage were work commitments, ill health/death, age, or declining further participation, as well as not being possible to contact them, or that they had moved away. All participants signed an informed consent form to take part in the study, which was approved by the Ethics Committee for Research and Teaching of the University of the Basque Country. 2.2.STUDY VARIABLES We interviewed all participants and collected data on anthropometric and socioeconomic characteristics; reproductive and breastfeeding history, intake of local food produce, and smoking and drinking habits. Weight gain or loss of more than 5 kg in the previous 5 years was noted. Regarding social class, participants were classified into one of two categories, manual or non-manual workers, according to their most recent occupation (not as possible source of DLC but as possible confounding factor), or that of the head of the family if they did not work, in line with the Spanish Adaptation of the British OPCS Classification of Occupations (social classes I+II+III and IV+V). Participants were also classified into two groups by level of education, depending on whether or not they had university qualifications. All participants were interviewed by the same researcher to avoid interviewer bias. We collected data on the frequency of intake of local food produce (fruit and vegetables, cow's milk and cheese, poultry and other meat), with four response options: never or less than once every 2 months, occasional (around once a month), around once a week, or daily (more than twice a week). These data were recoded to a dichotomous variable: never or occasionally versus weekly or daily intake. 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 2.3.BLOOD TEST ANALYSIS A 70-ml blood sample was collected from all study participants. They were not required to fast, but were recommended to avoid eating fatty foods the night before. Blood samples were collected in blood collection tubes (Vacutainer®) without anticoagulant and transferred immediately to a glass tube for centrifugation. To obtain the serum, the blood was left to clot for approximately 60-75 minutes at room temperature. The sample was then centrifuged at 1500 G for 15 minutes. The serum was separated from the clot with a Pasteur pipette. Unlike in previous stages of the study, in which the analysis was performed using pools of samples grouped by age and sex, the analysis in this study was based on individual samples. Samples were frozen at -20°C within 90 minutes after collection. Blood extraction, and serum preparation, storage and transport were carried out in accordance with the conditions established by the reference laboratories (Patterson et al., 1991). Next, workflow analysis is performed as follows: about 40 mL of serum was spiked with a 13C12 labelled dioxin and PCB standard mixture. Afterwards, analytes were extracted from the matrix by solid-phase extraction (SPE) approach using C18 cartridges (SPE Cartridges – C18; ISOLUTE, Charlotte, NC, USA). Elution was performed with n-hexane. Organic components, fat and other interfering substances were removed by treating the hexane extracts with silica gel (Supelco, Bellefonte, PA, USA) modified with sulphuric acid (44%). Further, the extracts were cleaned up by solid-liquid adsorption chromatography, using multilayer silica, basic alumina and activated carbon adsorbents. Purified extracts were rotary concentrated, transferred into vials and concentrated to dryness by a gentle stream of nitrogen prior to the mass spectrometry analysis. To evaluate the recovery rates, final extracts were reconstructed in a known amount of a mixture of labelled 13C12-PCDD/Fs (EPA-1613ISS, Wellington Laboratiores Inc., Guelph, Canada), 13C12-dl-PCBs (WP-ISS, Wellington Laboratories Inc, Guelph, Canada) and 13C12-marker PCBs (MBP-118, Wellington Laboratories Inc, Guelph, Canada). Instrumental analysis was based on the use of high resolution gas chromatography coupled to a high resolution mass spectrometry (HRGC-HRMS). Analyses were performed on an Agilent gas chromatograph (Agilent Technologies, Palo Alto, CA, USA) coupled to an AutoSpec Premier high resolution mass spectrometer (Waters, Manchester, UK) at 10,000 resolving power (10% valley definition). Gas chromatographic separation was performed on a DB-5ms fused silica column (60 m x 0.25 mm i.d. x 0.25-µm film thickness; J&W Scientific, CA, USA) for PCDD/Fs and dl-PCBs, whereas for marker PCBs, the analysis was carried out on a DB-XLB fused silica column (60 m x 0.25 mm i.d. x 0.25 µm-film thickness; J&W Scientific, CA, USA). Quantification was carried out by the isotopic dilution method. Relative response factors were measured for each individual compound by the analysis of six different calibration solutions for PCDDs/Fs, dl237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 PCBs and marker PCBs. Finally, the results were expressed as TEQ (toxic equivalent) values using WHO-TEQ (World Health Organization Toxic Equivalent). WHO-TEQ values were calculated in "upperbound", assuming the limit of detection (LOD) for those non-detected or below the LOD congeners. More detailed information on instrumental analysis is reported elsewhere (Abad et al., 2000). Enzymatic lipid determinations were performed using 100 µL serum sub-sample. Overall, four types of lipids were considered: triglycerides, total cholesterol, non-esterified (free) cholesterol and phospholipids. A welldocumented summation method was used to calculate the total lipid concentrations (Patterson et al., 1991). The total lipid content was expressed in g/L. For the conversion between volumetric and gravimetric data, a value of 1.026 g/mL was used for serum specific gravity. The requirements for ensuring high quality data include the application of quality assessment and quality control measures. A series of control blanks were analysed to assess laboratory cross-contamination. Detailed information on the analytical method is included as SI. 2.4.STATISTICAL METHODS The sample was described using the absolute and relative frequencies of the variables collected during the interview. The levels of congeners were described by mean, median, and range (minimum and maximum values), assessing their contribution to the WHO-TEQs. Pearson’s coefficient of correlation between the levels of congeners was calculated. Further, the geometric means (GMs) and 95% confidence intervals (CIs) of the lipid-corrected data were calculated for each group of pollutants and compared as a function of socioeconomic and reproductive variables. For the temporal analysis of the levels of PCDD/Fs and dl-PCBs, given that the previous samples were analysed in serum pools, the corresponding value of a new pool was calculated using the arithmetic mean of each individual value in 2013, because the measured concentration in a pooled sample is comparable to the arithmetic mean of individual specimens making up the pool (Heffernan et al., 2014). The pooled samples of 2006, 2008 and 2013 were adjusted to the normal distribution. Likewise, the distributions of the different groups of OCs, after logarithmic transformation, analyzed on individual samples were adjusted to the normal (Kolmogorov-Smirnov Z test p> 0.10). Changes in the arithmetic mean levels of dioxins, furans, and dl-PCBs in the pools from 2006 to 2008 and from 2006 to 2013 were analysed using a paired t-test. Similarly, changes in the geometric means of marker PCBs (individual samples) over these periods were analysed using a paired t-test. A multiple linear regression analysis was performed to identify the socioeconomic variables associated with each group of contaminants. All variables were included in any of the models with a p <0.10 according to the 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 likelihood ratio test. All the models included the geographical location, age and sex variables, since they were related to the selection of the sample in previous stages. Results are expressed as a percentage change [(exp(b)-1) * 100]. The relation between OCs and intake of local food produce was plotted, taking into account the distance to the MSWP. The criterion for statistical significance was p< 0.05. The statistical analysis was carried out using IBM SPSS Statistics for Windows, Version 20.0 (IBM SPSS, Armonk, NY, USA). 3. RESULTS Table 1 shows the characteristics of the sample. Slightly more than half of the 127 participants were women. Categorizing age into tertiles, approximately a third of participants were under 47 years of age, another third were between 47 and 60 and the last third over 60 years of age. The highest participation rate was observed in Balmaseda. A total of 29.1% of participants had normal weight, and a similar percentage was overweight, while 35.4% was obese. Nearly half of the sample reported weight change in the previous 5 years, 35.4% having put on weight and 14.2% having lost weight (≥ 5 kg). Regarding occupation and education, 42.5% of participants were manual workers and 28.3% had university qualifications. Overall, 20% were smokers at the time of the interview, 34.6% were ex-smokers and 43.3% were never-smokers. Among the women, more than half were in menopause, while 76.1% had had children and 77.8% had breastfed. Compared with previous years, in 2013, in addition to the corresponding increase in age, there was an increase in BMI and, also an increase in the number of women who had children. Social class showed non-manual workers to a greater extent in the 2013 sample than in the previous two. There were no significant differences in the characteristics of the rest of the variables. Table 2 shows the geometric mean, median, minimum and maximum levels of each of the congeners studied, as well as the WHO-TEQ values and the weight of each in the total level of substances with dioxin activity (DLC). Although a higher concentration of mono-ortho-PCBs was observed, especially PCB 118, they because of their lower toxicity, contributed a small part of the total WHO TEQ. Marker PCBs showed a clear decrease in levels, lower than DLCs. Three compounds, PCB 126, 12378-PeCDD (1,2,3,7,8pentachlorodibenzo-p-dioxin) and 23478-PeCDF (2,3,4,7,8-pentachlorodibenzofuran), accounted for 62.81% of the total WHO-TEQs. In contrast, other compounds with a high concentration in blood serum such as mono-ortho PCBs and to a lesser extent octachlorodibenzodioxin (OCDD) and PCB77 contributed very little to the WHO-TEQs, due to their low toxic equivalency factor. Supplementary Table 1 shows the correlations observed between the congeners; the coefficients being the highest among the PCBs and low between these and both dioxins and furans, while there was an intermediate correlation between dioxins and furans. 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 Table 3 shows a dramatic decrease in the mean values of OCs by 2013 compared to 2008 and 2006, in the total and in the four studied areas. There was not a significant decrease in mean values between 2006 and 2008. Indeed, in this period, we found a significant increase in the concentration of dioxin-like PCBs. Table 4 lists the GMs and 95% CIs for each pollutant with respect to the socioeconomic, reproductive and breastfeeding variables. Men had significantly higher levels of dl-PCBs than women. Among the four study areas, we found the highest mean OC levels in Santutxu, the difference being significant in all cases except for PCDD/Fs. We observed that individuals with university qualifications had significantly lower mean levels of dioxins, but higher levels of dl-PCBs. We did not observe significant differences between OCs and social class, body mass index (BMI) or weight change (data not shown). Levels of dl-PCBs were higher in women who had had children than those who had not, the difference being marginally significant; while those who had breastfed had lower levels of marker PCBs, the difference again being marginally significant. Postmenopausal women had significantly higher levels of dl-PCBs than other women. Multiple linear regression analysis (Table 5) showed that compared to the coefficients for Balmaseda, which is the least urban and industrialised of the areas studied and the furthest from the incinerator incinerator, the highest coefficients for all OCs were found for Santutxu, the differences being statistically significant for dl-PCBs, DLC and marker PCBs, but not for PCDD/Fs. In the other two zones, those closest to the MSWP, the coefficients did not show differences from those for Balmaseda, though Alonsotegi showed the lowest coefficients for all the OCs. Comparing the sexes, men did not show differences in marker PCB levels with women. Levels of both types of PCBs were positively associated with age; PCDD/Fs levels, however, were not significantly associated with age. Individuals who had only completed non-university education had significantly higher levels of PCDD/Fs than participants with university qualifications. The levels of PCDD/Fs and DLC were inversely related to weight gain in the previous 5 years, but the differences were not significant. Women who had breastfed had significantly lower levels of PCBs than the rest of all the participants. Figure 1 shows the relationship of OC levels with intake of local food produce and distance from the MSWP studied (Rekalde and Alonsotegi being nearby, while Balmaseda and Santutxu are further away). We did not observe any significant increases in the mean levels of OCs with intake of local food produce in areas near the MSWP. 4. DISCUSSION 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 In our study, individuals living near to a solid waste MSWP did not have higher blood levels of OCs than those living further afield. On the contrary, the highest levels of these pollutants were found in one the more distant areas, even after adjusting for confounding factors. Further, the intake of local food produce near the MSWP was not associated with an increase in the blood concentration of the pollutants analysed. These data are in agreement with those reported by Ranzi et al. (2011), who found that the ratio of emissions in 2008 to those in 1994-1996 for two Italian incinerators was about 0.0001 (10,000 times less) for dioxins and furans. Other studies analysing trends over time in areas close to MSWPs made similar observations, with no increase in OC levels in areas close to MSWPs (Deml et al., 1996; Evans et al., 2000; González et al., 2000; Chen et al., 2004; Huang et al., 2007; Reis et al., 2007; De Felip et al., 2008; Parera et al., 2013). In addition, we observed significant decreases in the levels of OCs, both in the areas near to and further away from the plant between 2006 or 2008 and 2013. Wittsiepe et al. (2000) also observed a decrease from 43.77 to 20.7 pg TEQ/g lipid between 1989 and 1998, and Humblet et al. (2011) a decrease from 36 to 25 pg TEQ/g lipid in women between 2000 and 2009. Similarly, Porta et al. (2012) reported OC levels to have decreased by 34 to 56% between 2002 and 2006, and Llobet et al. (2003) also found a decrease in OC levels over time. A general tendency for PCB levels to decrease was also observed by Linderholm et al. (2010) between 1990 and 2007 and Humblet et al. (2011), who found a decrease from 291 to 211 ng/g lipid in women. A systematic review of studies published up to 2010 on the blood levels of dioxins, furans and dlPCBs reported a mean annual decrease of 7% for dioxins and 6% for both furans and dl-PCBs (Consonni et al., 2012). A study based on the incinerator in Mataró, Spain (Parera et al., 2013) found a decrease in the levels of PCDD/Fs and marker PCBs, as in our study. A recent study showed also an important decrease of serum PCBs in a population living near a contaminated area after public health intervention (Magoni et al., 2016). Unlike various other studies (Deml et al., 1996; Evans et al., 2000; González et al., 2000; Reis et al., 2007), we did not find significant differences between the levels of PCDD/Fs by age. On the other hand, dl-PCB and marker PCB levels were positively associated with age. The aforementioned systematic review of blood OC levels, showed a 2% increase in levels for the three groups of OCs (PCDDs, PCDFs and dl-PCBs) with each year of age (Consonni et al., 2012). A previous study of samples collected from our study population between 2006 and 2008 (Zubero et al., 2015) showed an increase in the levels of OC pesticides and marker PCBs with age, consistent with a cohort effect, and a decrease between samples taken in the two periods (a decreasing period effect). This suggests that the higher levels found in older individuals are due to the fact that OC exposure has decreased over the years, and hence has less impact on younger cohorts, rather than a process of bioaccumulation over the years. 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 Supplementary Table 1: Pearson’s correlation coefficient between different congeners (dioxins, furans, dl-PCBs and marker PCBs). 2378TCDD 12378PeCDD 123478HxCDD 123678HxCDD 123789HxCDD 1234678HpCDD OCDD 2378TCDF 12378PeCDF 23478PeCDF 123478HxCDF 123678HxCDF 234678HxCDF 123789HxCDF 1234678HpCDF 1234789HpCDF OCDF PCB77 PCB81 PCB126 PCB169 PCB105 PCB114 PCB118 PCB123 PCB156 PCB157 PCB167 PCB189 PCB28 PCB52 PCB101 PCB138 PCB153 PCB180 2,3,7,8-TCDD 1 1,2,3,7,8-PeCDD .33 1 1,2,3,4,7,8-HxCDD .49 .50 1 1,2,3,6,7,8-HxCDD .26 .38 .31 1 1,2,3,7,8,9-HxCDD .39 .55 .61 .51 1 1,2,3,4,6,7,8-HpCDD .25 .45 .43 .71 .53 1 OCDD .18 .33 .29 .65 .44 .66 1 2,3,7,8-TCDF .41 .32 .27 .13 .16 .18 .12 1 1,2,3,7,8-PeCDF .34 .40 .35 .21 .29 .25 .18 .34 1 2,3,4,7,8-PeCDF .23 .50 .28 .69 .42 .59 .58 .18 .39 1 1,2,3,4,7,8-HxCDF .33 .51 .50 .44 .52 .54 .39 .33 .41 .40 1 1,2,3,6,7,8-HxCDF .32 .42 .42 .34 .41 .44 .31 .25 .46 .37 .67 1 2,3,4,6,7,8-HxCDF .34 .40 .44 .33 .45 .33 .28 .32 .53 .38 .56 .49 1 1,2,3,7,8,9-HxCDF .32 .40 .49 .19 .44 .20 .13 .16 .46 .24 .50 .42 .44 1 1,2,3,4,6,7,8-HpCDF .25 .47 .42 .55 .44 .62 .53 .31 .35 .60 .53 .45 .44 .39 1 1,2,3,4,7,8,9-HpCDF .40 .41 .46 .22 .44 .23 .15 .36 .50 .21 .44 .45 .61 .55 .36 1 OCDF .28 .45 .37 .19 .46 .20 .12 .28 .44 .19 .38 .29 .42 .49 .29 .61 1 PCB-77 .23 .13 .19 .26 .12 .27 .30 .13 .16 .24 .14 .25 .17 .07 .23 .13 -.08 1 PCB-81 .18 .10 .18 .15 .10 .07 .076 .04 .01 .13 .01 .05 .03 .11 .12 .11 -.03 .68 1 PCB-126 .14 .09 .21 .23 .06 .17 .17 -.01 -.11 .20 .01 .07 -.02 -.03 .13 -.05 -.10 .49 .67 1 PCB-169 .23 .16 .24 .45 .21 .27 .30 -.02 -.05 .39 .11 .11 .09 .13 .28 .07 -.08 .51 .79 .73 1 PCB-105 .19 .24 .12 .56 .23 .55 .56 .15 .10 .47 .27 .32 .08 -.04 .33 -.01 -.16 .66 .36 .44 .42 1 PCB-114 .25 .25 .15 .69 .29 .61 .64 .10 .10 .60 .28 .30 .12 .04 .42 .02 -.09 .58 .38 .49 .55 .91 1 PCB-118 .19 .23 .11 .60 .25 .57 .59 .13 .09 .50 .27 .32 .06 -.04 .36 -.02 -.16 .61 .33 .44 .44 .99 .93 1 PCB-123 .22 .25 .30 .46 .34 .47 .46 .03 .10 .40 .30 .34 .05 .14 .29 .05 -.03 .65 .64 .62 .64 .72 .74 .73 1 PCB-156 .20 .24 .11 .71 .30 .58 .67 .11 .07 .62 .27 .26 .07 .03 .45 -.01 -.04 .40 .22 .39 .48 .82 .92 .88 .60 1 PCB-157 .21 .25 .13 .71 .32 .60 .67 .10 .06 .62 .29 .26 .07 .06 .46 .00 -.04 .40 .25 .41 .51 .83 .94 .88 .63 .99 1 PCB-167 .18 .25 .11 .68 .28 .59 .65 .10 .03 .59 .27 .28 .05 -.02 .41 -.04 -.11 .44 .23 .41 .49 .86 .91 .91 .64 .96 .95 1 PCB-189 .25 .24 .17 .65 .29 .54 .64 .12 .06 .58 .27 .24 .05 .09 .44 .04 -.03 .41 .27 .44 .54 .75 .85 .79 .60 .92 .91 .87 1 PCB-28 .30 .20 .17 .47 .22 .48 .52 .21 .20 .48 .23 .31 .23 .03 .44 .09 -.10 .80 .39 .38 .42 .77 .70 .74 .59 .61 .60 .62 .59 1 PCB-52 .31 .16 .22 .39 .17 .38 .42 .17 .19 .39 .16 .23 .19 .03 .35 .07 -.12 .86 .51 .44 .46 .72 .65 .68 .61 .52 .52 .54 .53 .94 1 PCB-101 .25 .10 .17 .30 .13 .32 .37 .14 .16 .30 .12 .20 .14 -.03 .24 .04 -.15 .87 .52 .44 .41 .74 .63 .69 .62 .48 .48 .51 .48 .90 .97 1 PCB-138 .24 .23 .17 .64 .30 .54 .68 .10 .08 .57 .25 .27 .09 -.02 .38 .01 -.10 .50 .26 .40 .49 .86 .88 .89 .63 .90 .90 .90 .84 .73 .65 .64 1 PCB-153 .24 .22 .17 .65 .30 .54 .68 .09 .09 .59 .25 .27 .10 -.01 .40 .00 -.10 .48 .24 .40 .49 .84 .88 .88 .62 .92 .91 .91 .86 .73 .65 .63 .99 1 PCB-180 .24 .19 .17 .64 .30 .51 .68 .09 .09 .59 .21 .23 .08 .01 .40 -.01 -.07 .45 .25 .42 .51 .78 .85 .83 .60 .92 .91 .89 .88 .70 .63 .60 .96 .98 1 Bold: p< 0.05 1 Table 4: Mean and 95% confidence intervals for levels of PCDD/Fs, dl-PCBs, DLC (pg WHO-TEQ/g fat) and total marker PCBs (ng/g fat) expressed as pg WHO-TEQ/g fat and marker PCBs expressed as ng/g fat, as a function of socioeconomic characteristics. 2013 PCDD/Fs dl-PCBs DLC Marker PCBs N GM (CI) GM (CI) GM (CI) GM (CI) TOTAL 127 4.66 (4.22-5.14) 2.53 (2.15-2.98) 7.93 (7.21-8.73) 32.36 (25.75-40.66) Sex Male 56 4.75 (4.13-5.46) 3.17 (2.54-3.96) 8.66 (7.58-9.89) 35.00 (24.59-49.82) Female 71 4.58 (3.98-5.27) 2.12 (1.69-2.66) 7.40 (6.47-8.47) 30.41 (22.52-41.07) P .731 .016 .112 .551 Age, years < 47 42 4.69 (4.05-5.43) 2.29 (1.71-3.06) 7.69 (6.62-8.92) 26.43 (17.35-40.25) 47-60 43 4.65 (3.86-5.60) 2.54 (1.87-3.44) 7.82 (6.44-9.50) 35.23 (23.23-53.42) > 60 42 4.63 (3.85-5.56) 2.79 (2.17-3.59) 8.30 (7.14-9.66) 36.31 (25.68-51.35) p value .916 .341 .524 .271 Municipality Alonsotegi 30 4.48 (3.74-5.36) 1.44 (1.15-1.80) 6.23 (5.38-7.22) 19.27 (12.96-28.64) Bilbao - Rekalde 30 4.55 (3.83-5.40) 2.82 (2.12-3.76) 8.21 (7.07-9.54) 26.48 (17.74-39.53) Bilbao - Santutxu 30 4.96 (3.76-6.54) 6.18 (4.86-7.85) 11.55 (9.13-14.60) 93.75 (55.66-157.92) Balmaseda 37 4.65 (3.94-5.48) 1.77 (1.32-2.38) 6.91 (5.88-8.13) 24.46 (17.59-34.02) p value .910 .000 .000 .000 Body mass index, kg/m2 < 25 37 4.94 (4.06-6.02) 2.81 (2.05-3.86) 8.52 (7.01-10.35) 33.90 (21.69-52.98) 25-29 45 4.76 (4.06-5.59) 2.72 (2.07-3.56) 8.13 (6.87-9.62) 34.51 (22.88-52.03) > 30 45 4.33 (3.67-5.11) 2.16 (1.66-2.83) 7.30 (6.35-8.39) 29.20 (20.73-41.13) p value .291 .199 .203 .595 Weight loss ≥ 5 kg Yes 63 4.68 (4.19-5.24) 2.43 (2.03-2.92) 7.84 (7.02-8.75) 33.04 (25.50-42.80) (previous 5 years) No 64 4.48 (3.71-5.41) 3.21 (2.33-4.43) 8.52 (7.47-9.72) 28.52 (19.52-41.67) p value .764 .248 .555 .662 Weight gain ≥ 5 kg Yes 45 4.89 (4.34-5.50) 2.61 (2.15-3.16) 8.22 (7.35-9.21) 31.45 (23.87-41.43) (previous 5 years) No 82 4.26 (3.56-5.09) 2.40 (1.78-3.24) 7.42 (6.23-8.85) 34.07 (22.66-51.24) p value .195 .643 .318 .744 Level of education Pre-university 91 4.96 (4.43-5.56) 2.27 (1.87-2.76) 7.96 (7.14-8.89) 29.74 (22.71-38.94) University 36 3.96 (3.27-4.80) 3.32 (2.48-4.43) 7.85 (6.44-9.56) 40.03 (26.09-61.44) p value .044 .041 .891 .252 Social class Manual 54 4.87 (4.22-5.62) 2.54 (1.97-3.27) 8.26 (7.24-9.41) 29.68 (21.15-41.65) Non manual 73 4.50 (3.92-5.16) 2.52 (2.03-3.13) 7.70 (6.72-8.82) 34.49 (25.31-47.00) p value .439 .967 .483 .526 Intake of local produces Never, occasionally 37 4.64 (2.48-6.80) 2.57 (0.25-4.89) 8.02 (5.86-10.19) 33.09 (30.61-35.58) Weekly, daily 90 4.66 (2.58-6.74) 2.52 (0.36-4.67) 7.89 (5.82-9.97) 31.81 (29.57-34.05) p value .965 .907 .880 .878 Had had children No 17 4.72 (3.68-6.06) 2.09 (1.31-3.33) 7.34 (5.59-9.64) 31.83 (17.88-56.68) Yes 54 4.54 (3.84-5.37) 2.13 (1.64-2.76) 7.42 (6.35-8.67) 29.97 (21.05-42.69) p value .915 .055 .283 .827 Had breastfed No 12 4.22 (2.96-6.04) 2.79 (1.71-4.54) 7.43 (5.20-10.62) 55.30 (28.54-107.15) Yes 42 4.64 (3.82-5.62) 1.97 (1.45-2.67) 7.42 (6.23-8.83) 25.16 (16.84-37.60) p value .657 .283 .994 .069 Menopause No 29 4.31 (3.58-5.20) 1.71 (1.19-2.47) 6.64 (5.50-8.01) 24.87 (15.32-40.38) Yes 42 4.78 (3.91-5.84) 2.46 (1.85-3.26) 7.98 (6.63-9.60) 34.94 (23.86-51.15) p value .719 .014 .108 .474 Table 5: Multiple linear regression including variables associated with the blood levels of PCDD/Fs, dl-PCBs, DLC (pg WHO-TEQ/g fat) and marker PCBs (ng/g fat). Percentage change and 95% confidence intervals. PCDD/Fs dl-PCBs DLC Marker PCB Geographical areas Alonsotegi -7.6* (-30.6 to 22.9) -22.0 (-46.3 to 13.3) -12.8 (-32.0 to 11.9) -16.5 (-53.5 to 49.7) Bilbao - Rekalde -7.4 (-30.0 to 22.5) 48.1 (2.8 to 113.3) 12.0 (-12.2 to 42.9) 16.6 (-34.2 to 106.5) Bilbao - Santutxu 25.9 (-6.6 to 69.8) 263.0 (145.7 to 436.2) 90.3 (46.6 to 146.9) 357.3 (148.3 to 742.2) Balmaseda Reference Reference Reference Reference Sex Female Reference Reference Reference Reference Male 1.3 (-21.8 to 31.0) -7.7 (-34.1 to 29.2) -3.9 (-23.3 to 20.3) 49.3 (-11.9 to 153.0) Age (years) -0.1 (-1.1 to 0.8) 1.7 (0.5 to 2.9) 0.5 (-0.3 to 1.3) 2.6 (0.6 to 4.5) Level of education Non-university Reference Reference Reference Reference University -32.7 (-47.8 to -13.2) 0.9 (-27.6 to 40.7) -22.1 (-37.6 to -2.8) -16.7 (-50.5 to 40.0) Body mass index (kg/m2) -1.6 (-3.4 to 0.2) -2.3 (-4.6 to -0.01) -1.4 (-3.0 to 0.2) -4.6 (-8.2 to -1.0) Weight gain in previous 5 years No Reference Reference Reference Reference Yes -17.1 (-33.3 to 3.0) -19.4 (-39.3 to 7.1) -17.3 (-31.5 to 0.0) -11.0 (-42.9 to 38.8) Women who had breastfed No Reference Reference Reference Reference (compared to the all rest of subjects) Yes -2.6 (-26.0 to 28.2) -27.8 (-49.6 to 3.3) -7.4 (-27.1 to 17.6) -50.5 (-71.8 to -13.3) * % change= (exp(b)-1) * 100; percentage change in OC levels associated with a change in the reference of the current category or an increase of one unit in the continuous variables. Hihglights: - This study shows a dramatic decrease in the mean values of OCs from 2006 and 2008 to 2013 - We found differences in mean OC levels in relation to de area of sample - Men had significantly higher levels of dl-PCBs than women - Levels of both dl-PCBs and marker PCBs were positively associated with age - Individuals with lower level of education had higher levels of PCDD/Fs than participants with university qualifications - Women who had breastfed had significantly lower levels of PCBs than the rest of all the participants