Persistent organic pollutants in hospitalized individuals in the municipality of Petropolis, Brazil

Persistent organic pollutants (POPs) are highly lipophilic and can accumulate and biomagnify in food chains. Characterized as a public health problem, exposure to these compounds enables the development of diseases like cancer, cardiovascular disease, diabetes, and obesity. The objective of this study was to estimate the plasma levels of organochlorine pesticides and PCBs in 151 samples (97 women; 54 men) in hospitalized individuals in Petropolis, Brazil. Individuals over 18 years of age and residing for at least 2 years in the mountainous region of the State of Rio de Janeiro, Brazil, participated in a cross-sectional study. Interviews using a structured questionnaire and blood samples to estimate plasma levels of persistent organic pollutants provided data. Gas chromatography coupled with triple quadrupole mass spectrometry provided the levels of organochlorine pesticides and PCBs. Compared to data present in the literature, the concentration of POPs was lower, and individuals from 55 to 64 years of age (3.28 ng mL−1) and women (2.52 ng mL−1) presented a higher average concentration of organochlorine pesticides; men (0.05 ng mL−1) also presented a high concentration of PCBs. This is the first Brazilian study to estimate the concentration of several POPs in a hospital-based sample that includes men and women, thus contributing to the characterization of our population regarding environmental exposures relevant to health.


Introduction
Persistent organic pollutants (POPs), such as organochlorine pesticides and polychlorinated biphenyls (PCBs), are part of a group of environmental pollutants that are highly lipophilic, besides having the ability to accumulate and biomagnify in food chains (Arrebola et al. 2015). These compounds, identified as endocrine disruptors, relate to the development of diseases such as cancer, cardiovascular diseases, diabetes, and obesity (Fry and Power 2017;Han et al. 2019;Park et al. 2020;Wahlang 2018;Yang et al. 2017).
In 1985, there was a ban on the commercialization, distribution, and usage of organochlorine pesticides in agriculture. Nevertheless, studies report that residues of these pesticides are still present in foods that make up the primary diet of human beings, such as milk, cheese, and fish, being, in some cases, above the maximum acceptable limits (Botaro et al. 2011;Santos et al. 2015;Carneiro et al. 2015). Additionally, because they are lipophilic compounds, degrade slowly, with half-lives usually above 5 years, and bioaccumulate along the food chain, these compounds can persist in organisms and the environment for more than 30 years, even if their contamination sources are few (Peres and Moreira 2003;Flores et al. 2004). The present study was carried out in the city of Petropolis, which is part of the mountainous micro-region of the state of Rio de Janeiro. The city stands out for being part of the leading agricultural hub of the state, with 2120 ha of land devoted to agriculture and 29% of its establishments making use of pesticides, with an average use 18 times higher than that of the state (IBGE 2017;Peres and Moreira 2007). It is common knowledge that the use of organochlorine pesticides continues to occur in the region, possibly obtained through smuggling and illegal trade (de Souza Guida 2016; Meire et al. 2012;Peres and Moreira 2007). Produced in large quantities from the 1930s to the 1970s and used as thermal and electrical insulators in equipment such as transformers, PCBs suffered a worldwide ban after the Stockholm Conference due to their carcinogenic potential. However, given their high chemical stability, we can often find these substances in the environment due to incorrect disposal, leaks, or accidents involving them (Bloise 2018).
Some studies have already evaluated the serum concentrations of POPs in some populations (Fry and Power 2017;Park et al. 2020;Holmes et al. 2014;Wielsøe et al. 2017;Boada et al. 2012;Arrebola et al. 2015). However, in Brazil, most studies have observed the concentration of these substances in animals and the environment (Cascaes et al. 2014;De Souza Pereira et al. 2007;Dias et al. 2013;Sotão Neto et al. 2020;Yogui et al. 2010). Other studies evaluated serum concentrations only in women, pregnant or not, and umbilical cord samples (Eudge et al. 2012;Sarcinelli et al. 2003;Mendonça et al. 1999;Fróes-Asmus et al. 2021). Thus, the level of exposure of the Brazilian population is still uncertain. This study aims to estimate the plasma levels of organochlorines and PCBs in hospitalized individuals in Petropolis, Brazil.

Materials and methods
The present work is a descriptive and cross-sectional study carried out with all individuals recruited as controls of a hospital-based case-control study conducted in Petropolis to investigate the association between exposure to pesticides and the development of different types of cancer.
In the case-control study, the first hospitalized individual who met the inclusion criteria was selected as control in the same period that the cases were interviewed: from December 2014 to March 2020. Individuals with cognitive impairment that made it impossible to communicate or understand the research questionnaire questions, individuals hospitalized for treatment of any cancer or symptoms potentially associated with neoplasms, and individuals with a previous history of cancer were excluded. Thus, the present study has 151 individuals over 18 years of age admitted to Hospital Municipal Alcides Carneiro (Fig. 1), from December 2014 to March 2020, with no previous history of cancer.
Trained personnel used a structured questionnaire to interview the participants at the health facility. The Fig. 1 Location of the hospital were the study was conducted questionnaire included sociodemographic data, lifestyle, and previous contact with pesticides. Individuals involved with agricultural activities answered questions about the use of pesticides, protective equipment, characteristics of the property, and the crops produced on it.
The research instrument initially presented a series of questions about the use of pesticides in general (for example "have you ever used pesticides for rodent control?", "for commercial buildings?", "on agricultural crops?"), and, when the participant answered affirmatively to any of them, a list of commercial names for specific products was presented. For each product that the respondent reported having used, data on quantity and frequency over time were requested.
The analytical method was adapted from Sarcinelli et al. (2003). Briefly, the samples were thawed at room temperature and subsequently denatured, then diluted with equal parts of methanol and water, then mixed and extracted in previously conditioned 500 mg/6 mL C18 solid-phase extraction cartridges (JT Baker, USA), dried, and eluted with 7 mL of hexane afterward. A pre-conditioned 1 g/6 mL florisil cleanup cartridge received the eluate, which was eluted with hexane followed by dichloromethane/hexane (1:1 v/v), a solution of petroleum ether/hexane (85:15). The extracts evaporated to dryness under a nitrogen atmosphere. The volume was resuspended to 100 µL with hexane and analyzed by GC-MS/MS, using 5 µL of dibromophenol 1, 1′-biphenyl-4, 4′dibromine at 1 µg mL −1 as the internal standard.
A GC-MS/MS analysis using Thermo Scientific equipment, model TSQ 8000 EVO, containing a Trace GC 1310 chromatograph with AS 1310 autosampler with programmable temperature vaporization (PTV) injectors, operating in splitless mode with ThermoFisher Xcalibur™ and Trace-Finder™ software. An Agilent® DB-5MS column that contained phenylmethyl siloxane (30 m × 250 μm × 0.25 μm), with injection without flow splitting received 2 μL of the sample. Compounds were identified by selected individual reaction monitoring (SRM) adjusted for retention time. The National Institute of Standards and Technology (NIST) libraries, included in the Tracefinder™ software, defined the spectrometric conditions, and Accu Standard's high purity standards confirmed those conditions. Therefore, transitions were defined according to their high specificity combined with high abundance.
Plasma levels of each POP were presented as mean, median, and standard deviation (SD) in ng mL −1 , and ng g −1 of lipid to compare the results obtained in the literature. Since the lipid content was not measured individually in each sample, the average reference values for adults of fasting triglycerides (below 150 mg dL −1 ) and fasting total cholesterol (below 190 mg dL −1 ) served as the basis. The total lipid content was estimated using the formula established by Phillips et al. (1989), whose ratio is total lipids = 2.27 * total cholesterol + triglycerides + 0.623. As a result, the calculated total lipid content was 2.27 * 180 + 150 + 0.623 mg dL −1 , or 560 mg dL −1 .
The IBM-SPSS 20.0 program provided data analysis. Characterization of the study population was according to age group, gender, skin color, schooling, BMI, alcohol consumption, smoking, and variables related to exposure to pesticides. The use of the Mann-Whitney or Kruskal-Wallis U tests established the existence of a correlation between the plasma levels of POPs and these variables.
The procedures of this project followed the recommendations of Resolution No. 466/12 of the National Health Council. This project received the approval of the Research Ethics Committees of the Public Health National School (ENSP -Escola Nacional de Saúde Pública) (CAAE: 31,126,514.0.0000.5240) and the Alcides Carneiro Municipal Hospital.

Results
The calibration curves of the analytical method obtained were linear with correlation coefficients (R2) > 0.95 for OCs and PCBs, with accuracy and precision determined between 0.2 and 10 ng mL −1 as a working range. The limit of detection (LOD) ranged from 0.015 to 0.47 ng mL −1 for OCs and from 0.016 to 0.36 ng mL −1 for PCBs. The limit of quantification (LOQ) ranged from 0.049 to 1.42 ng mL −1 for OCs and from 0.045 to 1.08 ng mL −1 for PCBs. Average recoveries were 93% to 105% for OC pesticides and 87 to 107% for PCBs.
The average concentration of persistent organic pollutants in the analyzed samples was 2.19 ng mL −1 , corresponding to 391.07 ng g −1 of lipid, most of which are represented by organochlorines, with an average concentration of 2.15 ng mL −1 , equal to 383.93 ng g −1 of lipid. At the same time, PCBs had the lowest mean concentration, 0.04 ng mL −1 , and equivalent to 7.14 ng g −1 of lipid (Table 1). Table 2 shows the characterization of the population of this study, composed of 151 individuals, being 97 women (64.2%) and 54 men (35.8%), and the average concentrations of persistent organic pollutants by subgroups, according to selected epidemiological variables. The most of the individuals of the study had a domestic exposure (68.9%), while 3.7% made use of agricultural pesticides and 27.7% never used such compounds (data not shown). The population had average concentrations of POPs (p = 0.003) and organochlorines (p = 0.002), statistically different according to age groups, with individuals from 55 to 64 years presenting the highest concentrations of these compounds. The average age was 52.84 years, with a standard deviation of 15.6 years.
There was a statistically significant difference observed between the genders, with women presenting a higher concentration of POPs (p = 0.007) and organochlorines (p = 0.006) but lower of PCBs (p = 0.029). When evaluating the concentration of these compounds for different age groups, according to gender, it was observed that women continued to present higher average concentrations of POPs and organochlorines than men. The exception was for the stratum from 45 to 54 years old, which, for men, represents the stratum with the highest average concentration, thus differing from the general sample analysis (Table 2).
There were no statistically significant differences between the concentrations of these analyzed compounds (Table 2) regarding the variables of skin color, schooling, BMI, alcohol consumption, and smoking.

Discussion
This study is the first Brazilian study to estimate the concentration of several POPs in a sample that includes men and women. Thus, it was possible to observe statistically significant differences between the genders and differences in the age groups with the highest plasma concentrations of POPs in each one of them. These findings reinforce the need to include different population groups when characterizing Table 1 Plasma levels of persistent organic pollutants (ng mL −1 and ng g −1 of lipid) in hospitalized individuals in Petropolis City (n = 151), 2014-2020 The means of the datasets without outliers and the medians of the datasets with outliers are in bold ∑POPs sum of persistent organic pollutants, ∑DDT sum of DDT metabolites, ∑HCH sum of hexachlorocyclohexane metabolites, ∑OCs sum of organochlorines, ∑PCBs sum of polychlorinated biphenyls, SD standard deviation

POPs
Minimum-maximum ng mL −1 ng g −1 of lipid In observation, the analyzed population of Petropolis had lower average plasma concentrations of POPs than those found by other studies.
A case-control study with 121 cases and 103 controls recruited in Spain between 1999 and 2001 by Boada et al. (2012) detected an organochlorine concentration equal to 1173.60 ng g −1 of lipid in the blood samples of controls. In addition, the case-control study with 77 cases and 84 controls interviewed in the periods from 2000 to 2003 and 2011 to 2014 in Greenland estimated an average concentration of 1529.79 ng g −1 of lipid in the blood samples of controls (Wielsøe et al. 2017). Those studies contrast the average concentration of organochlorines of 383.93 ng g −1 of lipid found in our study.
The estimated average concentration of 83.93 ng g −1 of lipid provided the sum of DDT metabolites analyzed in our study. The hospital-based case-control study with 217 cases and 213 controls conducted between 1995 and 1997 in Canada reported, in tissue samples, mean concentrations of DDT equal to 19,300 ng g −1 of lipid for p,p′-DDT between controls (Aronson et al. 2000). Another case-control study, also carried out in Canada by Darvishian et al. (2022), found an average concentration equal to 102 ng g −1 of lipid for a DDT metabolite (p,p′-DDE) in blood samples from controls collected between 2012 and 2014. The hospital-based case-control study with 209 cases and 165 controls residing in China between 2014 and 2019, conducted by Huang et al. (2019), detected a concentration equal to 126.39 ng g −1 of lipid among the samples of adipose tissue from controls for the DDT metabolite p,p′-DDT. It is expected that the concentrations of POPs found in adipose tissue samples are higher than the plasma concentrations measured in the present work because they are lipophilic compounds that are in equilibrium with the plasma fraction. These compounds accumulate in the adipose tissue and release into the blood in specific situations, such as the increase of the lipolysis process due to stress, breastfeeding, or sudden weight loss (Crinnion 2009). However, the significant difference between the concentrations reported in the present work and the cited studies is noteworthy. When comparing the result of the sum of DDT with other studies that used plasma samples, it is still possible to observe that the population of this study presents low concentrations. The case-control study by Demers et al. (2000), with 315 cases and 219 hospital controls selected between 1994 and 1997 in Canada, estimated an average concentration of 12,500 ng g −1 of lipid among controls, only of p,p′-DDT. The cross-sectional study conducted in the USA by Ouidir et al. (2020) estimated an average concentration of p,p′-DDE equal to 85.26 ng g −1 of lipid in the 260 blood samples from pregnant women collected between 2009 and 2013. These results are similar to the ones in this study. Finally, the hospital-based case-control study with 403 cases and 403 controls conducted between 2001 and 2005 in Japan by Itoh et al. (2009) detected an average concentration of 9.9 ng g −1 of DDT in the blood samples of controls, much lower than the result of the present work.
Other studies that worked with the average concentration of DDE (DDT metabolite) also reported concentrations higher than the concentrations found in the present study (0.44 ng mL −1 -data not shown in the table). A case-control study with residents in Rio de Janeiro City recruited between 1995 and 1996 estimated an average concentration of DDE equal to 4.80 ng mL −1 in the analyzed blood samples for 331 controls (Mendonça et al. 1999). Wolff et al. (2000) reported an average concentration of DDE equal to 7.27 ng mL −1 in the nested case-control study with 148 cases and 245 controls from a cohort study in New York City. The study by Arrebola et al. (2015), with 69 cases and 54 controls selected in hospitals in Tunisia in 2012, which used blood samples, detected an average concentration equal to 1.07 ng mL −1 for DDE. Two studies carried out in Rio de Janeiro City with blood samples from pregnant women found average concentrations of DDE equal to 1.77 ng mL −1 in 62 samples collected from 1997 to 1998 (Sarcinelli et al. 2003) and 0.14 ng mL −1 in 31 samples collected from 2017 to 2018 for a cohort study (Fróes-Asmus et al. 2021). However, the cohort study by Fróes-Asmus et al. (2021), which analyzed blood samples from pregnant women and their umbilical cord, whose collection was between October and November 2017, estimated a concentration of DDE three times lower than the concentrations presented in this work. The low concentrations shown in the study cited and the present work are expectable because of the '80 s ban on organochlorines. Organochlorines keep for a long time in the environment, but their concentration has been decreasing over the years because DDT and its metabolites' half-life are about 7 years. Thus, studies have shown a tendency for the concentration of these compounds to decrease as they depurate, while the sources of new contamination are limited (Bloise 2018; Thomas et al. 2008 andToan et al. 2009). Unfortunately, even today, it is known that different populations are still exposed to new sources of contamination by organochlorine pesticides obtained through illegal trade Moreira 2007, Li;Macdonald 2005;Muzardo and Graciani 2015).
Regarding the average concentration for exposure to various PCBs, the present study found an average of 7.14 ng g −1 of lipid. The hospital-based case-control study of 209 cases and 165 controls residing in China between 2014 and 2019 by Huang et al. (2019) estimated a mean concentration equal to 57.44 ng g −1 for all PCBs between adipose tissue samples from controls. In the USA, a cross-sectional study conducted by Ouidir et al. (2020) reported an average concentration of PCBs equal to 14.13 ng g −1 of lipid in the 260 blood samples analyzed.
Another study, with 77 cases and 84 controls interviewed in two periods (2000 to 2003 and 2011 to 2014) in Greenland, detected average concentrations equal to 1156.90 ng g −1 for PCBs in blood samples from controls (Wielsøe et al. al., 2017). Also, in the hospital-based case-control study, with 217 cases and 213 controls, conducted between 1995 and 1997 in Canada, the lowest concentration of polychlorinated biphenyls found was for PCB 105, equaling 6300 ng g −1 for controls (Aronson et al. 2000). Another hospital-based case-control study, with 405 cases and 405 controls conducted between 2001 and 2005 in Japan, by Itoh et al. (2009), estimated an average concentration of 180 ng g −1 of lipid for PCBs for controls in blood samples.
The case-control study, with 69 cases and 56 controls selected from hospitals in Tunisia, in 2012, by Arrebola et al. (2015), used blood samples and reported a mean concentration of 0.17 ng mL −1 , 0.68 ng mL −1 , and 0.18 ng mL −1 for PCBs 138, 153, and 180, respectively, in controls. The present study shows that the average sum of all PCBs was equal to 0.04 ng mL −1 . This concentration is also lower than the one found by the nested case-control study, with 148 cases and 245 controls, from a cohort study in New York City, by Wolff et al. (2000), which observed an average concentration equal to 4.97 ng mL −1 for PCBs in blood samples from controls.
The limitation of the present cross-sectional study is measuring the concentration of POPs from blood samples instead of using tissue samples. As mentioned earlier, this biological matrix is not the most appropriate for this type of assessment since it only indicates the percentage of POPs circulating in the body, not being so faithful to the level of chronic exposure (Crinnion 2009).
When comparing the results of different studies, one must observe the matrix and the laboratory techniques used. Although most of the studies cited have evaluated the concentrations of POPs from blood samples, they used the gas chromatography technique by electron capture, which is considered more sensitive in analyzing the presence of organochlorine pesticides (Barbosa et al. 2012). Thus, the use of gas chromatography coupled to mass spectrometry, despite being unequivocal, is less sensitive than the first and may be considered a limitation of the present work. Data insufficiency for evaluating if patients had experienced an event that could influence the presence of POPs in the blood or the reduction of these compounds levels-weight loss or breastfeeding -close to the time the blood sample was collected is another limitation.
On the other hand, the present study has the strength of investigating exposure to persistent organic pollutants in a location not yet studied and with a broader population compared to other studies that measured exposure exclusively among women.
Additional studies, with a better assessment of the concentration of persistent organic pollutants, are necessary to elucidate the level and trend of exposure of the Brazilian population to these compounds.

Conclusion
The present study found higher concentrations of organochlorine pesticides in individuals aged 55 to 64 years (3.28 ng mL −1 ). The concentration of these pesticides was higher in women (2.52 ng mL −1 ), while that of PCBs was higher in men (0.05 ng mL −1 ). It is worth noting that the average concentrations of persistent organic pollutants are below those observed by other studies in the literature, which may be due to the matrix used to measure the concentration of these compounds and also the time in which the samples were collected, once the concentration of these pollutants decreases over time, and some of them were banned since the eighties. Our results show that more studies are necessary to investigate the relation of populations to healthrelevant environmental exposures.