Exposure to air pollution and ovarian reserve parameters

Exposure to air pollution is associated with many different health effects, especially cardiovascular and respiratory diseases. Additionally, highly significant links between exposure to air pollution and fertility, particularly male fertility was observed, however the studies regarding exposure to selected air pollutants and female fertility assessed by ovarian reserve are rare. Hence, the main aim of the study was to analyze relationship between exposure to ambient air pollution and ovarian reserve parameters among Polish women. The study population consisted of 511 women, who attended to infertility clinic because of diagnostic purposes. Participants filled in the questionnaire about social-demographic, lifestyle and health factors. Infertility specialists assessed ovarian parameters such as: antral follicle count (AFC) and concentration of hormones: Anti-Müllerian hormone (AMH), follicle stimulating hormone (FSH) and estradiol (E2). The air pollutants level (sulfur dioxide, nitrogen dioxide, carbon monoxide, ozone, particulate matters) were obtained via National Environmental Protection Inspectorate database. Significant negative association between PM2,5 and AMH (p = 0.032), as well as AFC (p = 0.044) was observed. Moreover, SO2 concentrations decrease AFC (p = 0.038). The results also suggest that PM10, PM2.5, SO2 exposure on antral follicle count may be more pronounced among women with a female factor infertility diagnosis. Additionally, exposure to PM2.5 and NOx on AFC and AMH was stronger among older women (> 35 years of age). To conclude, the present study found that air pollution could lead to decrease in follicle antral count and Anti-Müllerian hormone level, especially exposure to PM2,5 and SO2 thus the evidence suggest negative impact to ovarian reserve.


Study population
The study group include 511 women in reproductive age (25-39 years of age) recruited from fertility clinic.Only menstruating women with ovulatory cycles without co-existing chronic diseases (e.g., fragile X syndrome, adrenocortical insufficiency, abnormal karyotype) were included.Women with: three spontaneous miscarriages, more than three in vitro fertilization procedures, chemotherapy or radiotherapy of pelvis, premature ovarian failure, previous surgical treatment of the ovaries, polycystic ovary syndrome, cyst in the ovaries with endometrium, hyperprolactinemia, hypogonadotropic and hypogonadism were excluded.700 woman fulfill the inclusion criteria and were suitable for the research, however only 511 (73%) gave the approval to participate in the research.
All the participants filled the questionnaire about socio-demographic characteristics, lifestyle factors (e.g.smoking, alcohol drinking, physical activity), health conditions (e.g.co-existing diseases), and occupational exposures.
This study was approved by the Nofer Institute of Occupational Medicine Bioethical Committee Board (ethics approval number: Resolution 23/2016).All experiments were performed in accordance with relevant guidelines and regulations.Informed consent was obtained from all subjects.

Analysis of ovarian reserve parameters
In the study ovarian reserve parameters were assessed such by antral follicular count (AFC), and reproductive hormones level (AMH, FSH, and estradiol).
The count of the antral follicle was assessed by ultrasonography (USG) according to Broekmans et al. criteria 30 .To assessment considered antral follicle with dimension measurement extent from 2 to 10 mm.All procedure was performed by trained gynecologist.The antral follicle count was treated for the analysis as the sum of antral follicle in both ovaries.
The blood samples were collected in early follicular phase cycle of spontaneous menstruation (2-4 day of cycle).The concentrations of the following hormones were assessed in serum: AMH (Anti-Müllerian hormone), FSH (follicle stimulating hormone) and estradiol (E2).The blood samples were centrifuged and a serum collected to polypropylene test tubes and stored in − 80 °C to time of analysis.
An enzyme linked immunosorbent method was used to analyzed the level of AMH using Gen-II ELISA kits (Beckman Coulter, Inc., USA) based on manufacturer instruction.The chemiluminescence method was used to measure levels of FSH and estradiol from ITROS ECi Immunodiagnostic System with MicroWell technology using commercially available VITROS Reagent Packs and the VITROS Calibrators which was used based on manufacturer instruction (Ortho-Clinical Diagnostics Johnson & Johnson, UK).
For each pollutant, the average value for the 6 months which address the period of developmental stage from primary follicle to antral follicle, nearest monitoring sites was assigned to the women according to their ZIP code of residence.

Statistical methods
Descriptive statistics were calculated for subjects grouped by demographic characteristics, along with the layout of examined air pollutants, and AMH, E2 and FSH levels and AFC.Multiple least squares linear regression models were used to quantify the relationship of air pollution exposure (explanatory variables) with AFC and the concentrations of AMH, E2 and FSH as dependent variables.Multivariate regression models was used to explore an association between air pollutants levels, reproductive hormone concentrations and AFC.Two models were constructed.The first model was adjusted for following variables: BMI (kg/m 2 ), age (years), smoking (no/ yes), initial infertility diagnosis (male factor; female factor; unexplained) and the second model for age, BMI, smoking, infertility diagnosis, duration of infertility (1-3 years; 3-5 years; > 5 years); alcohol consumption (none or < 1 drink/week; 1-3 drinks /week; everyday).The covariates in the model were included on biological and statistical consideration.The air pollutants in the model were treated as the categorical variable (first to 25th percentile value, second-greater than the 25th percentile value to the median, third greater than the median to 75th percentile value, while the fourth group consisted of values greater than the 75th percentile or continuous variable.Effect modification of the association between exposure to air pollution and ovarian reserve parameters by age, BMI, current smoking, infertility diagnosis, duration of infertility, all well-known predictors of ovarian reserve, by adding a cross-product term to the final multivariate model was tested.R statistical software was used for the analysis (version 4.2.2).

Ethics approval and consent to participate
The Bioethical Committee in Lodz, Poland, approved the study (Resolution no 23/2016).All participants obtained and signed written informed consents prior to enrollment.

Ambient air pollution levels
The levels of ambient air pollution are presented in the Table 3

Parameters of ovarian reserve and exposure to air pollution
The Table 4 presented relationship between air pollution levels and parameters of ovarian reserve.The statistical model was adjusted for: age, BMI, smoking and infertility diagnosis.A significant associations between parameters of ovarian reserve and air pollution was observed in 4 th quartile of exposure to PM 2,5 for antral follicle count (AFC) (p = 0.044) and Anti-Müllerian Hormone (AMH) (p = 0.032) compared to the first quartile, when the exposure was treated as the categorical variable.Additionally, statistically significant relationship was found between SO 2 concentration and antral follicle count (AFC) (p = 0.038).No association was found between estradiol (E2) and follicle stimulating hormone (FSH) and any examined air pollutants (PM 2.5 , PM 10 , SO 2 , CO, NOx, ozone) in quartiles of exposure.When the air pollution exposure was treated in the model as a continuous variable no association were found between any air pollutants and any examined ovarian reserve parameters (Table 4).When the model was adjusted additionally for duration of infertility and alcohol consumption only exposure to PM 2,5 in 4th quartile was related to decrease in antral follicle count number (p = 0.045) and AMH concentration (p = 0.048) (Table 4).Other air pollutants (SO 2 , PM 10, ozone, NOx, CO) were not significantly associated with ovarian reserve parameters.
The estimated effect of PM 10 , PM 2.5 , SO 2 exposure on antral follicle count was stronger among women whose primary infertility diagnosis was attributable to a female cause compared to women with an unexplained or male factor diagnosis (p-for-interaction 0.04, 0.02, 0.04 respectively) (Table 5).Additionally the effect of exposure to PM 2.5 and NOx was more pronounced among women > 35 years of age compared to women < 35 years old (p-for-interaction 0.01, 0.03 respectively).PM 2.5 affect also AMH concentration among older women (p-forinteraction 0.03) (Table 5).

Discussion
According to our best knowledge this is one of first study which assessed ambient air pollution exposure and ovarian reserve in European population.Additionally, contrary to previous studies the ovarian reserve was examined in complex way based on different ovarian reserve parameters (AFC, AMH, FSH, and E2).Also different  www.nature.com/scientificreports/air pollutants levels were evaluated (ozone, PM 10 , PM 2.5 , SO 2 , NO x , CO).Negative association between exposure to PM 2.5 and AMH levels and AFC were found.Additionally, exposure to SO 2 in the 4th quartile of exposure compared to the first one decrease the AFC.The results also suggest that PM 10 , PM 2.5 , SO 2 exposure on antral follicle count may be more pronounced among women with a female factor infertility diagnosis.Additionally, exposure to PM 2.5 and NOx on AFC and AMH was stronger among older women (> 35 years of age).Four epidemiological studies were performed to assess the air pollution exposure and ovarian reserve [24][25][26][27] .The first study was performed by Gaskins et al. 24 , the author observed significant negative relationship between exposure to PM 2,5 and antral follicle count.The association of PM 2.5 with antral follicle count was stronger among women with female factor infertility and abnormal menstrual cycles.The observed results are similar to our study regarding to PM 2.5 exposure and more pronounced effect among female with female factor of infertility.Abareshi et al. 25 observed inverse relationship between exposure to air pollution (PM 1 , PM 2,5 ) and levels of Anti-Müllerian hormone (AMH).In the study, the author examined only particulate matters exposure (PM 1 , PM 2,5 and PM 10 ) in population of 67 women 25 .Another study performed in Korea found that exposure to PM 10 and PM 2,5 decrease ovarian reserve assessed only by AMH concentrations 26 which is in line with our results.In the study by Feng et al. 27 only the antral follicular count (AFC) and air pollution exposure were evaluated.In the study SO 2 exposure was associated with lower antral follicle count 27 .Similar relationship was observed in our study however the duration of exposure was different than in our study.The study covered of long-term exposure to ambient air (from 2013 to 2019), whereas in our study for each pollutant, the average value for the 6 months which address the period of developmental stage from primary follicle to antral follicle was assessed.The study by Feng et al. 27 evaluated only AFC, as the main parameter of ovarian reserve.
The study performed in animals demonstrated that exposure to PM may reduce the antral follicle count in mice 29 .Another study in mice found diminishing ovarian reserve 32 after exposure to diesel exhaust.Furthermore, Zhou et al. 33 observed ovarian dysfunction associated with chronic PM 2,5 exposure (4 months).The causal relationship that links air pollution to ovarian reserve has yet to be elucidated.Prior reports have suggested that folliculogenesis can be impaired by the increased oxidative stress and cellular apoptosis induced by ambient air containing a range of pollutants 34 .ROS can damage the oocytes and can lead to negative impact on female reproduction 35 .In results oxidative stress (OS) can lead telomere dysfunction (DNA damage or aberrant telomere recombination) and caused miscarriages and infertility 36 .Basically, there are three ways how the environmental pollutants can impact ovarian function: by causing endocrine disrupting effect, by induction of oxidative stress and by causing epigenetic modification 37 .Thus according to study performed by Styszko et al. 38 , PM 10 and PM 2,5 fractions of aerosols shows high oxidative potential (OP) and also included metals in composition such as: K, Ca, Ti, V, Mn, Fe, Pb, Cu, Cr, Zn, Hg, Ni, some of these metals can as endocrine disrupting chemicals so via the impact on the hormone level the ovarian reserve can be diminished 39 .Hence, endocrine disruptors can alter hormone receptor binding and action, such as: mainly on the aryl hydrocarbon receptor (AHR), estrogen receptors (ERs) or androgen receptor (AR), which can lead directly on disturb ovarian functions 40 .The studies in mice suggest that aryl hydrocarbon receptor (AHR) can play a function in the formation primordial follicles and also regulation of antral follicle count 41 .Other study reported that the AHR can regulate follicle growth through change estradiol biosynthesis pathway factors.The study performed by Revelli et al., suggest that estrogens and progesterone lead in the intraovarian regulation of follicle growth 42 .According to study Walters et al., the androgen receptor can lead to follicular growth and follicle and ovulation development 43 .
In present study the ovarian reserve was assessed by specialist in gynaecological endocrinology and reproduction, which result in the reliable assessment of antral follicle count.Moreover this is the first study which assess all parameters of ovarian reserve such as: Anti-Müllerian Hormone (AMH), follicle-stimulating hormone (FSH) and antral follicle count (AFC).AFC has been well-established biomarker for reflecting ovarian reserve from the aspect of visualization 44 .Also AMH is a reliable predictor for reflecting ovarian reserve due to its invariability with cyclical menstrual cycles 45 and is associated with follicular size 46 .
The study has several strengths such as: performing the study in one fertility center, biological samples were collected and analyzed using the same procedure.The qualified medical specialist (infertility specialist) was assessed ovarian reserve according to the same standardized protocol.All procedures were carried out at the 2-4 days of the cycle (beginning of follicular phase).Moreover all participants answered the same questionnaire about sociodemographic, lifestyle and medical risk factors thus, we allowed for controlling potential confounding factors in statistical model.
The study has also some limitations.First, the study was performed in fertility clinic, which may not lead to generalized the results directly to general population.Second,, air pollution was assessed based on the information about ZIP code of residence for each women and was linked to the nearest air monitoring station.This method may not constitute the actual levels of exposure because we are not staying all the time in our places of residents.More proper measurement would minimize such a potential measurement bias.Nevertheless this is the popular method used in epidemiological studies with big sample size, where the individual monitoring may be not comfortable for participants and expensive [24][25][26][27][47][48][49]

Conclusions
In summary, we observed that the exposure to PM 2,5 decrease antral follicle count and AMH concentration, whereas the SO 2 exposure was also negatively related only to antral follicle count.The results rise of importance to reduce the potential risk of ambient air pollution exposure and suggest the need for more strict regulations and controls forced by appropriate law.Additionally, further studies involving general population and underlying the mechanism of adverse impact on ovarian reserve are needed.

Table 2
presents ovarian reserve parameters among study population.For AFC arithmetic mean (± SD) was 12.73 ± 8.94.Reproductive hormones levels were: 1.17 ± 1.46 (ng/ml) for Anti-Müllerian Hormone, 6.38 ± 2.18 (IU/l) for follicle stimulating hormone and 93.74 ± 16.63 (pg/ml) for estradiol.According to the guidelines the AFC, FSH was in normal range (range for AFC: more than 4 antral follicles in ovary and range for FSH: lower

Table 1 .
Characteristics of the study population N = 511.

Table 4 .
The association between air pollution levels and parameters of ovarian reserve-multivariate analysis.Significant values are in bold.