Concentrations of persistent organic pollutants in blood of the Spanish population: Temporal trend

The present article reviews the human biomonitoring studies conducted in Spain to assess exposure to persistent organic pollutants (POPs) such as polychlorinated biphenyls (PCBs), polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs), polybrominated diphenyl ethers (PBDEs), and perand polyfluoroalkyl substances (PFAS). In general terms, important variations in POPs concentrations between Spanish regions and specific populations were observed, while no associations between exposure to POPs and adverse health outcomes were found. Moreover, occupational exposure seems not to be a risk factor with regards to POPs exposure in the Spanish population. The present review highlights the importance of conducting human biomonitoring studies to find possible associations between POPs and adverse health effects.

Human exposure to these type of compounds leads to adverse health outcomes, such as disruptive effects in the endocrine, reproductive or immune systems, as well as neurotoxicity and genotoxic effects (7)(8)(9). Due to these detrimental health effects, it is essential to know the current exposure levels of the general population, so actions can be taken by policy makers in order to ensure that the population's health is not at risk. The present paper aims to review the available scientific literature on the current exposure levels of POPs in the blood of the Spanish population, as well as to establish the temporal trend.

Polychlorinated biphenyls (PCBs)
PCBs are one of the most frequently analysed organochlorine compounds (OCs), mainly due to their widespread exposure and detection around the world. In Spain, several studies have quantified the levels of PCB congeners -especially 138, 153 and 180 -and determined their association with a number of diseases. The temporal variation of PCBs 138, 153 and 180 is summarized in Table I.

Table I
A summary of concentrations of various persistent organic pollutants (POPs) in the Spanish populations: Temporal variations.  (28, 52, 101, 118, 138, 153 and 180) in cord blood from pregnant women (n = 355) and in the blood from their children (n = 278) at 4 years of age. Congeners 28, 52 and 101 could not be quantified in any of the samples of either cord or whole blood. Median levels for congeners 118, 138, 153 and 180 were 0.06, 0.14, 0.18 and 0.13 ng/mL for cord blood, and 0.09, 0.18, 0.25 and 0.12 ng/mL for whole blood samples. Correlations with poor neuropsychological development were not found (11).

Tabela I
Recently, Abellan et al. (2019), included 1308 mother-child pairs and analysed the levels of PCB congeners 138, 153 and 180 in the cord blood and their possible association with the lung function. The median levels of PCB 138, 153 and 180 were 0.08, 0.11 and 0.08 ng/mL, respectively. However, they were not associated with a decreased lung function in children at 4 and 7 years of age. . A total of 1455 mother-children pairs were recruited between 2004 and 2008, and maternal serum was analysed for PCB-118, -138, -153 and -180. The highest PCB concentrations were found for congener 153 (45.1 ng/g lipid), followed by congener 180 (32.5 ng/g lipid), congener 138 (27.1 ng/g lipid) and congener 118 (6.4 ng/g lipid). However, no association between PCBs and lower respiratory tract infections was found (14). In the same three cohorts -but recruited between 2003 and 2008 - Gascon et al. (2013) assessed the prenatal exposure to PCB-153, which is the PCB congener with the highest levels and detection rate in the study. The geometric mean found in maternal serum samples was 39.81 ng/g lipid, which is similar to that found in their previous study (14). A year later, Gascon et al. (2014) assessed the correlation between exposure to PCBs and respiratory health. In this case, a cohort from Menorca (Balearic Islands) was used. It consisted of 482 mothers who visited antenatal care between 1997 and 1998. Participants were divided into two study populations (A and B), with levels of ∑PCBs in cord blood of 0.58 and 0.59 ng/mL, respectively. Notwithstanding, no associations between PCBs and altered respiratory function were found (16).  (20). The median levels were 0.062 ng/mL for PCB-118, 0.085 ng/mL for PCB-138, 0.113 ng/mL for PCB-153 and 0.082 ng/mL for PCB-180. However, there was no association between PCB exposure and the birth size of new-borns (20). In another study conducted by this research group, PCB congeners 138, 153 and 180 were analysed in 2369 samples of maternal serum, and in 1140 cord serum samples of four Spanish cohorts from the INMA project (Asturias, Gipuzkoa, Sabadell and Valencia) (21). In both kinds of samples, PCB-153 was the congener showing the highest concentrations (0.28 and 0.12 ng/mL for maternal and cord blood, respectively). PCB-180 showed lower levels than PCB-153 (0.20 and 0.08 ng/mL for maternal and cord blood, respectively), while PCB-138 showed similar concentrations as PCB-180 (0.17 and 0.08 ng/mL for maternal and cord blood, respectively) (21). Llop et al. (2017) measured the levels of PCBs in 1128 samples from pregnant women from the INMA cohorts of Valencia and Sabadell and assessed their association with the thyroid system. PCB-153 showed the highest concentration (0.22 ng/mL) again, followed by PCB-180 0.15 ng/mL and PCB-138 (0.13 ng/mL). No correlations between PCBs and thyroid hormone levels were found (22). Morales et al. (2013) analysed the levels of 4 congeners (118, 138, 153 and 180) in serum samples from 2031 pregnant women. Serum concentrations were 7.6 ng/g lipid for PCB-118, 24.8 ng/g lipid for PCB-138, 40.9 ng/g lipid for PCB-153, and 28.8 ng/g lipid for PCB-180. In a previous INMA Cohort-related study, the maternal levels of 5 PCB congeners (28, 118, 138, 153 and 180) were measured in 584 pregnant women from the Sabadell cohort (24). PCB-118 was the congener with the lowest median levels (12.1 ng/g lipid), while PCB-153 showed the highest concentrations (35.5 ng/g lipid). PCB-138 and -180 had concentrations of 21.1 and 25.9 ng/g lipid, respectively. None of the samples showed detectable levels of PCB-28 (24). In a cohort of 344 children, Valvi et al. (2012) associated the levels of 7 PCBs congeners (28, 52, 101, 118, 138, 153 and 180) in cord blood samples with overweight at 6.5 years. As in previous studies, PCB-28, -52, -101 and -118 had the lowest detection rates (5, 11, 28% and 76%, respectively) and the lowest mean levels (0.07, 0.05, 0.04 and 0.06 ng/mL, respectively), while PCB-138, -150 and -180 were the most detected congeners (95, 98 and 95%, respectively), and also those with the highest concentrations (0.14, 0.18 and 0.13 ng/mL, respectively). Moreover, authors found an association between childhood overweight and PCB exposure in girls (25). Taking into account that the most detected congeners are 138, 153 and 180, Valvi et al. Other Catalan studies were conducted in different population groups. In Flix (Tarragona, Catalonia), a hospital-based case-control study was performed to assess the correlation between the exposure to environmental contaminants and the risk of colorectal cancer (40). A total of 220 subjects were randomly selected from a cohort of 866 participants, which were recruited between 1996 and 1998. The levels of PCB-28 and PCB-52 were below the LODs, while those of PCB-118, PCB-138, PCB-153 and PCB-180 were 92, 308, 362 and 252 ng/g lipid, respectively. An association between the risk of colorectal cancer and PCB congeners 28 and 118 was observed (40). In a general population cohort, Wingfors et al. (2000) compared the levels of PCBs for Spain and Sweden. Mean values of the ∑PCBs were 1310 ng/g for the Swedish population and 1450 ng/g for the Spanish population, with congener profiles varying between countries. In general, the higher chlorinated congeners were more dominant in the Spanish group than in the Swedish one (41) Catalonia is an autonomous community that has a great industrial activity, including various municipal solid waste incinerators (MSWI) and a hazardous waste incinerator (HWI), which could be an important source of exposure to polychlorinated compounds such as PCBs. Various studies have been conducted in the area of influence of these incinerators in order to assess the exposure of the general population and the occupationally-exposed population. Domingo et al. (2001) assessed the baseline exposure to PCBs when the HWI (Constantí, Tarragona County) started its industrial activity. A total of 28 workers in the new facility were recruited and plasma samples were analysed. The geometric means of PCB congeners 28, 52, 101, 138, 153 and 180 were 12.9, 7.5, 7.1, 129, 182 and 158 ng/g lipid, respectively. According to the authors, the levels of exposure were of the same order of magnitude as those previously found in a nonoccupationally exposed population from Mataró (Catalonia) -where a MSWI is located -and whose PCBs concentration were analysed in 201 adults (47). The levels for congeners 101, 138, 153 and 180 were 0.04, 0.49, 0.67 and 0.59 ng/mL, respectively. Congeners 28 and 52 were not detected in any of the samples (47). Since the first study (46), periodical surveys were carried out to monitor the internal levels PCBs in workers from the HWI. Thus, Schuhmacher et al. (2002) assessed the levels of PCBs in plasma samples from 23 workers, one year after the HWI started its regular activities. No differences in PCB levels between the baseline study and one year after were found in any of the three workplaces considered (plant, laboratory and administration) (48). In contrast, after three years of regular operations, the levels of PCBs 28, 52, 101, 138 and 153 were significantly lower in plasma samples from 28 workers of the HWI when compared to those of the baseline study (49). A similar trend was also reported by than those found in the baseline study (52). All the studies conducted in the HWI located in Tarragona showed that there were no evident signs of occupational exposure to PCBs.
In Mataró, Gonzalez et al. (2000) monitored the levels of PCBs in residents living near the MSWI two years after regular operations. PCBs 28 and 52 were not detected, neither in 1995, nor in 1997, in any of the samples. In 1997, general PCB exposure increased in both groups of residents -living near and further away from the MSWI -by 25 and 12%, respectively. Furthermore, minimal changes were observed among workers of the MSWI (53). Parera et al. (2013) carried out a temporal study to monitor the PCB levels in blood samples from the general population -both exposed and non-exposed. In 2012, the potentially exposed group had a concentration of PCBs of 0.52 µg/g fat and the control groups from Mataró and Arenys de Mar (a village near Mataró) showed concentrations of 0.65 and 1.16 µg/g fat. Furthermore, the workers from the MSWI showed a concentration of 0.71 µg/g fat. No significant differences were observed between groups (54).
The PANKRAS II study is also a regional project that was conducted between 1992 and 1995 at five general hospitals of eastern Spain, where 185 patients with exocrine pancreatic cancer were recruited. In a study performed by Porta et al. (1999), a cohort of 51 subjects were selected among the PANKRAS II cohort to evaluate the association between selected PCBs and the mutations in codon 12 of the K-ras gene. The highest median values were found for PCB-180 (1.56 ng/mL), followed by PCB-138 (1.45 ng/mL) and PCB-153 (1.31 ng/mL). Patients with K-ras mutation had significantly higher concentrations of these three congeners when compared to the wild-type cases (55). Some years later, Porta et al. (2008) compared the levels of PCBs with the social class of 135 subjects. Authors found that these compounds were higher in the less affluent occupational classes, meaning that social class could not be independent of the risk of pancreatic cancer.
Regarding eastern Spain, other individual studies have been carried out. A population-based birth cohort was established in Menorca (Balearic Islands) that included 259 children born between 1997 and 1998 (Álvarez-Pedrerol et al., 2008). Median levels of PCB congeners 28, 52, 101, 118, 138, 153 and 180 were 0.01, 0.03, 0.08, 0.10, 0.17, 0.25 and 0.12 ng/mL, respectively. The authors suggested that exposure to PCB compounds might affect the thyroid system, specifically T3 levels (57). In Valencia, another eastern city, an association between thyroid levels and PCB exposure was assessed in pregnant women (n=157) (58). The sum of total PCBs was 0.85 ng/mL, which is similar to that reported by Álvarez-Pedrerol et al. (2008) (0.78 ng/mL). By contrast, no association was found between PCB exposure and thyroid hormone levels (58).
In southern Spain, exposure assessment studies to PCBs have also been conducted.  (59). The same authors also studied exposure to PCBs in 107 women diagnosed with gestational diabetes mellitus (60). The median values for congeners 138, 153 and 180 were 0.26, 0.73 and 0.31 ng/mL, respectively. The authors suggested that exposure to certain POPs was a modifiable risk factor that contributes to insulin resistance (60).

Per-and polyfluoroalkyl substances (PFAS)
PFAS are resistant to thermal, chemical and biological degradation due to their strong bonds. Their properties make them ideal for use in many consumer products. Therefore, their presence is widespread around the world. The temporal variation of PFOS, PFOA, PFNA and PFHxS is summarized in Table 1. In Spain, some national studies have been carried out to determine the levels of exposure to PFAS, specifically the BIOAMBIENT.ES project and the INMA project. The former includes two studies. Bartolomé  also evaluated the associations between prenatal exposure to PFAS and birth outcomes in 1202 mother-child pairs. Mean levels were 0.58 ng/mL for PFHxS, 6.05 ng/mL for PFOS, 2.35 ng/mL for PFOA and 0.66 ng/mL for PFNA. In general, PFAS levels were not associated with birth outcomes. Only PFHxS, PFOA and PFNA showed a weak, but non-significant correlation with reduced birth weight (70). The same authors assessed the associations between prenatal exposure to PFAS and cardiometabolic risk (71). Geometric mean was 0.61 ng/mL for PFHxS, 5.80 ng/mL for PFOS, 2.32 ng/mL for PFOA and 0.66 ng/mL for PFNA. Nonetheless, PFAS exposure was not associated with an increased cardiometabolic risk (71). In turn, Matilla-Santander et al. (2017), studied the association between PFAS and metabolic outcomes in 1240 pregnant women from three regions of the INMA project (Gipuzkoa, Valencia and Sabadell). The highest median values were found for PFOS (6.05 ng/mL), followed by PFOA (2.35 ng/mL), PFNA (0.65 ng/mL) and PFHxS (0.58 ng/mL). PFOS and PFHxS were positively associated with impaired glucose tolerance, as well as with gestational diabetes mellitus. PFOA was positively associated with total cholesterol, while PFOS and PFNA were negatively correlated with triglycerides (72). Recently, Costa et al. (2019) assessed the possible associations between PFAS exposure and foetal biometry in a subset of 1230 mother-child pairs from the INMA cohort. PFHxS, PFOA, PFOS and PFNA medians were the same than those reported by Matilla-Santander et al. (2017). In general terms, there were no associations in any trimester of pregnancy (73). Finally, using the same cohort and the same PFAS values, Carrizosa et al. (2021) associated PFAS exposure to the neuropsychological development of the foetus. Higher prenatal exposure to PFAS was associated with a worse motor development at 14 months, especially to PFHxS. In addition, a marginal positive correlation between general cognitive development at 4-5 years and PFOS, was observed (74). In a study conducted in Tarragona (Catalonia) belonging to the HEALS project, the levels of PFOS and PFOA in 50 pregnant women were measured (75). Mean plasma levels of PFOA were 0.45, 0.13 and 0.12 ng/mL at the first trimester, at delivery and in cord plasma, while those of PFOS were 2.93, 2.21, and 1.17 ng/mL, respectively. These results are lower than those reported by other authors in pregnant women (68)(69)(70)(71)(72)(73)(74).

Polybrominated biphenyl ethers (PBDEs)
PBDEs are compounds used as flame retardants additives, due to their capacity to react with gas-phase free radicals during combustion. They are chemically stable and might accumulate in the food chain. The temporal variation of ∑PBDEs is summarized in Table 1. Within the INMA project, several studies have been conducted to assess exposure to PBDEs in pregnant women and in children. Carrizo et al. (2007) analysed 13 PBDE congeners (17, 28, 47, 66, 71, 85, 99, 100, 138, 153, 154, 183 and 190) in 92 cord blood samples, and in serum from 4-year-old children. The average concentration of the sum of PBDEs was 6.2 and 4.3 ng/g lipid in new-borns and in 4-year-old children, respectively. BDE-47 showed the highest concentrations (2.8 and 2.9 ng/g lipid in cord blood serum and serum, respectively). Several years later, Gascon et al. (2011) also analysed the levels of three congeners of PBDEs (47, 99 and 100), which showed the highest concentrations in the study conducted by Carrizo et al. (2007). The median levels of BDE-47, BDE-99 and BDE-100 were 2.10, 0.38 and 0.38 ng/g lipid for cord blood, while for serum at age of 4 they were 0.12, 0.12 and 0.12 ng/g lipid, respectively. Results showed that postnatal exposure to BDE-47 was significantly related to an increased risk of symptoms of the attention deficit subscale of ADHD symptoms (77). Vizcaino et al. (2011b) determined the levels of PBDEs in a non-occupationally exposed mother-child pairs from the INMA-Valencia cohort (n=174). Total PBDE levels in umbilical cord serum and in maternal serum were 9.6 ng/g lipid for both matrices, being BDE-47 the congener with the highest concentration (2.3 ng/g lipid). The same authors evaluated the prenatal exposure to PBDEs in two Mediterranean cohorts from Valencia and Menorca (27). A total of 265 new-borns were recruited and the levels of 5 PBDEs (47, 99, 100, 153 and 183) were measured. Total concentrations of PBDEs in Valencia cohort were 6.5 ng/g lipid, while in the Menorca cohort they were 2.8 ng/g lipid. In both cohorts, BDE-47 was the highest contributor to PBDE exposure, with the results being similar to those reported in other studies (76)(77)(78).

Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs)
PCDD/Fs are probably the most hazardous pollutants released by incinerators. They are very persistent in the environment and in the tissues of the exposed population. Therefore, most of the studies have been conducted around incinerator plants. The temporal variation of ∑PCDD/Fs is summarized in Table 1. In Tarragona County (Catalonia), the first study was carried out in 1999 during the construction of a new HWI (46). The levels of PCDD/Fs in 28 workers of the HWI were analysed, with mean levels of 27.6 pg I-TEQ/g lipid (range: 13.4 -84 pg I-TEQ/g lipid) (46). One year after starting the plant activities, Schuhmacher et al. (2002), assessed the levels of PCDD/Fs in 19 HWI workers. The results showed a mean concentration of 16.9 pg I-TEQ/g lipid, which is similar to that of the baseline survey (48). Six years after regular operations of the HWI, Mari et al., (2007) measured the concentrations of PCDD/Fs in blood samples from 19 workers in order to establish temporal variations with respect to the baseline study. The mean PCDD/Fs concentration was significantly lower than that found in the first study (10.4 pg I-TEQ/g lipid) (50). Later, the same research group, also determined the levels of PCDD/Fs in blood samples from workers of the HWI after 8 years of operations (n=19) (51). The decrease in PCDD/Fs levels was even more substantial when compared to the baseline study (2.5 vs. 27.6 pg WHO-TEQ/g lipid). Results indicated that there are no evident signs of occupational exposure to PCDD/Fs (51). Twelve years since the start of the activities of the HWI, Mari et al. (2013) analysed the concentrations of PCDD/Fs in blood samples from 18 workers. Similar values with respect to the previous survey were found (4.6 pg I-TEQ/g lipid) (52). Concurrently with the assessment of occupational exposure, the same research group also studied the exposure of the residents living near this facility. Schuhmacher et al. (1999) measured the baseline levels of PCDD/Fs in 20 subjects living in the vicinity of the new HWI. The mean PCDD/Fs value was 27.0 I-TEQ/g lipid with a range between 14.8 and 48.9 pg I-TEQ/g lipid (82). Three years after regular operations, Agramunt et al. (2005), studied the temporal variation in comparison to the baseline study in subjects living near the HWI (n=20). The authors found a significant reduction in PCDD/Fs with respect to the first survey (27.0 vs. 15.7 pg I-TEQ/g lipid) (83). Some years later, Nadal et al. (2008) also measured the levels of PCDD/Fs of non-occupationally exposed population (n=20). The authors reported a significant decrease with respect to the baseline study (27.0 vs. 9.36 pg I-TEQ/g lipid) (84). In 2012, a follow-up study was carried out by the same research group, where levels of PCDD/Fs were assessed in blood samples from 40 subjects (85). There was a further decrease of the exposure (6.18 pg I-TEQ/g lipid) when compared to the previous surveys (84,85). Finally, the last survey was conducted in 2018, and blood samples from 40 subjects were obtained (86). The level of PCDD/Fs for the non-occupationally exposed population was 6.79 pg I-TEQ/g lipid, which was significantly lower than the baseline concentration (27.0 pg I-TEQ/g lipid) but similar to those found in the two previous surveys (9.36 and 6.18 pg I-TEQ/g lipid) (86).
In Mataró, another Catalan city, two studies to assess human exposure around a MSWI were conducted. González et al. (1998) recruited 198 adults living near and far from the incinerator and assessed the exposure to PCDD/Fs. Mean PCDD/F levels were 13.5 and 13.4 pg I-TEQ/g lipid for subjects living near and far from the MSWI, respectively (47). Two years after regular operations, the same research group analysed the levels of PCDD/Fs in the same population groups (53). There was an increase of the PCDD/Fs concentrations in both groups, with them being 16.7 pg I-TEQ/g lipid for residents living close to or further of the MSWI. However, according to the authors, the slight increase in the PCDD/Fs concentration was not due to the incinerator's emissions (53). Several years later, Parera et al. (2013) assessed the temporal trend of PCDD/Fs exposure around the MSWI in 104 exposed subjects and 97 non-exposed subjects randomly selected from Mataró. Mean levels of PCDD/Fs were 12.9 and 13.3 pg I-TEQ/g lipid for exposed and non-exposed populations (54).
In the Basque Country, various studies were also conducted to assess human exposure to PCDD/Fs around a MSWI. Zubero et al. (2009) analysed the levels of PCDD/Fs in pooled serum samples from residents living at a different distance from the plant. Median levels of PCDD/Fs were similar regardless of the zone: 24.3 and 27.3 pg WHO-TEQ/g lipid for Zone E1 and E2, which are located less than 2 km away from the plant; 21.3 pg WHO-TEQ/g lipid for Zone C1, which is located 5 km from the MSWI but with high-traffic density, and 18.8 pg WHO-TEQ/g lipid for Zone C2, which is an area located 20 km from the plant, and with minimal industrial and traffic activities. Total mean PCDD/Fs were 23.45 pg WHO-TEQ/g lipid (64). Two years later, the same research group, analysed the concentrations of PCDD/Fs in pooled serum samples again, and compared the results with those found in the previous survey (87). Mean PCDD/Fs levels in 2008 were 23.6 pg WHO-TEQ/g lipid, which are similar to those previously found (23.45 pg WHO-TEQ/g lipid) (87). Finally, in 2013, levels of PCDD/Fs were also measured in the same area (66). There was a significant decrease in the exposure to PCDD/Fs, with a mean level of 4.67 pg WHO-TEQ/g lipid (66).

Conclusions
The present article has reviewed the results published in the scientific literature regarding the levels of various POPs in the Spanish population, considering children, pregnant women, general population, as well as workers from industrial activities, basically waste incinerators. The results show that there is great variation among Spanish regions and populations, with no clear pattern of exposure and with the concentrations of PCBs being higher in the Northern regions than in the South. In general terms, there were no associations between POPs exposure and health effects. Only specific congeners were associated with some specific health outcomes, especially diabetes and obesity. Finally, there were no differences in POPs exposure -especially PCBs and PCDD/Fs -between occupationally and non-occupationally exposed populations. In summary, many human biomonitoring studies to assess POP exposure in the population have been conducted through the years in Spain, highlighting the importance of these surveys to disclose potential associations with several potential diseases.