Air pollution Exposure And Mammographic Breast Density In Tehran, Iran

Background: Air pollution is one of the major public health challenges in many parts of the world possibly has an association with breast cancer. However, the mechanism is still unclear. This study aimed to nd an association between exposure to six criteria ambient air pollutants (PM 2.5 , PM 10 , SO 2 , NO 2 , O 3 , and CO) and mammographic breast density (MBD), as one of the strongest predictors for developing breast cancer, in women living in Tehran, Iran. Methods: Participants were selected from women attending two university hospitals for screening mammography from 2019 to2021. Breast density was rated by two expert radiologists. Individual exposures to 3-years ambient air pollution levels at the residence were estimated. Results: The nal analysis in 791 eligible women showed that low and high breast density was detected in 34.8% and 62.2% of participants, respectively. Logistic regression analysis after considering all possible confounding factors represented that an increase in 10 units of NO 2 (ppb) exposure was associated with an increased risk of breast density with an OR equal to 1.47 (95% CI: 1.10 to 1.97). Furthermore, CO level was associated with a decreasing breast density (OR = 0.33, 95% CI = 0.17 to 0.64). None of the other pollutants were associated with breast density. Conclusion: MBD was associated with higher levels of ambient air NO 2 and lower levels of CO. Perhaps MBD monitoring as an available tool in each population, can help the prediction of future breast cancer occurrence and nding the high-risk geographic areas. data and inverse distance weighting (IDW) deviation; OR:


Background
Air pollution is one of the major public health challenges in many parts of the world and in large cities.
Ambient air pollution is a mixture of different pollutants originating from natural and anthropogenic sources and the International Agency for Research on Cancer (IARC) classi ed it as Group 1 carcinogenic to humans (1). Several studies showed that short-and log-term exposure to air pollution can cause many chronic and acute health effects. Numerous studies reported that long-term exposure to outdoor air pollution caused globally around 4.24 million premature deaths annually (2)(3)(4)(5).
Breast cancer is one of the worldwide leading causes of mortality and morbidity, and according to a report of GLOBOCAN in 2018 accounts for more than 11.6% of all female cancers; while the disease burden of breast cancer shows an increasing trend in some populations (6). Because this disease imposes a heavy burden on the health system, more preventive efforts are necessary and further investigation should explore the underlying reasons for these epidemiological trends. Ecologic studies suggest that breast cancer risk is elevated in urban areas with high levels of air pollution compared to rural areas (7,8). Air pollution contains many carcinogens and other compounds that may act as endocrine disruptors, and air pollution exposure has been globally linked to many cancers such as lung, breast, and bladder cancer (9). In 1979 Hill and Winder found that inhaled toxicants (nicotine and Loading [MathJax]/jax/output/CommonHTML/jax.js cotinine) were detectable in breast uid after 30 minutes of smoking (10). Thus, toxic chemicals can reach the breast tissue and have possibly some impacts on it.
A systematic review and meta-analysis in 2021 investigated whether high levels of air pollution exposure were related to increased breast cancer risk (11). This study showed that NO 2 had a "moderate level of evidence" and that PM 2.5 and PM 10 have an "inadequate level of evidence" for supporting their association with breast cancer risk. Also, the biological mechanism of the effects of air pollutants on breast cancer still remains unknown (11).
Mammographic breast density (MBD) is one of the strongest predictors and biomarkers for developing breast cancer (12). Growing evidence suggests that environmental risk factors such as xenoestrogens [(Bisphenol-A (BPA)] and metalloestrogens (lead, cobalt, and magnesium) have a direct association with MBD (13,14). However, one study has reported a reverse association between serum levels of Polychlorinated biphenyl (PCBs) as a xenoestrogen and MBD (15). Limited studies evaluated the association between MBD and air pollution exposure, which had also inconsistent results (16)(17)(18)(19). To draw risk-reducing strategies for breast cancer, studying the impacts of ambient air pollutants on breast density may provide valuable data for monitoring and etiologic factors. Previous studies had limitations in the assessment of the exposure, in adjusting confounding variables, and in outcome ascertainment to nd a causal relationship; therefore further studies have been recommended (11).
Tehran as the capital of Iran is a megacity with about 10 million residents and air pollution is a major environmental challenge in this city. The people of Tehran are more exposed to high levels of ambient air pollution, to the point where government and non-government o ces are sometimes closed due to the severity of air pollution (3,20). Therefore, the present study was designed to investigate whether there is an association between exposure to six criteria ambient air pollutants (Nitrogen dioxide (NO 2 ), Sulfur dioxide (SO 2 ), Carbon monoxide (CO), Ozone (O 3 ), and Particulate matter (PM) 2.5 , 10 ) and MBD in women living in Tehran, Iran.

Study design and participants
This study was designed as a cross-sectional study; participants were selected from women attending two university hospitals a liated to Tehran University of Medical Sciences, Tehran, Iran, for screening mammography from 2019 to 2021. The study was approved by the ethics committee of Tehran University of Medical Sciences (IR.TUMS.VCR.REC.1398.897), and all participants have signed an informed consent. All methods have been performed in accordance with the relevant principles of the Declaration of Helsinki.
Criteria for inclusion in the study were at least 3 years residency in the capital city of Iran (Tehran) and having the ability to ll questionnaires. Exclusion criteria included suspicion for malignancy in the current Loading [MathJax]/jax/output/CommonHTML/jax.js mammography and an imprecise address.

Data Collection
Participants were asked to ll a questionnaire that captured demographic information, self-reported age, weight, height, reproductive history, menopause status, smoking history (active and passive), history of oral contraceptive (OCP) use, current use of hormone replacement therapy, and familial history of breast and ovarian cancer. All women who either had a current or previous history of active and passive (second hand) smoking were de ned as having a positive exposure to smoke. Menopause was de ned as cessation of the menstrual period at least one year sooner, women were strati ed into premenopausal and postmenopausal status. Furthermore, we gathered information about current aspirin and metformin use and consumption duration in each woman. Routine use of supplements including vitamin D, calcium, Vitamin E, Omega 3, and Evening Primrose oil were also recorded.
One expert radiologist reported the breast density in each center. In order to decrease both intra-and interreader variability, the mammographic breast density was checked by a third radiologist. Radiologists rated MBD according to the American College of Radiology (ACR) Breast Imaging-Reporting and Data System (BI-RADS) classi cation into four categories: almost entirely fatty (BI-RADS 1), scattered areas of bro glandular density (BI-RADS 2), heterogeneously dense (BI-RADS 3), and extremely dense (BI-RADS 4) (21). We categorized MBD into low density (ACR 1 and 2) and high density (ACR 3 and 4).
The exact address of residence of the participants in the recent 3 years and the telephone number of that place were recorded. Also, in the employed women, the address of their place of work and the hours of their attendance to work were recorded.

Air pollution exposure assessment
In this study, estimating the exposure of participants to ambient criteria air pollution was done in the following three steps: 1. Outdoor air quality data gathering from xed monitoring stations belong to Tehran Air Quality Control Company.
2. Data cleaning of air quality monitoring stations in order to outlier data detection. 3. Individual long-term exposure assessment using air quality data and inverse distance weighting (IDW) approach.
Air quality data gathering Real time hourly ambient air quality in Tehran city is monitored by xed monitoring stations. In Tehran city, Air Quality Control Company (AQCC) a liated to the Tehran Municipality is responsible for monitoring criteria air pollutant (PM 2.5 , PM 10 , NO 2 , O 3 , SO 2 , and CO). At the time of this study in 2020, there were 22 monitoring stations in Tehran belonged to the AQCC. Considering that the data of the AQCC stations are available on an hourly basis and are publicly available online, and that these monitoring Loading [MathJax]/jax/output/CommonHTML/jax.js stations are spatially located in all districts of Tehran city, therefore we used the data obtained from air quality monitoring stations belonged to AQCC in this study.
Finally, hourly data of six outdoor criteria air pollutants for the 3-years residency of participants were obtained from the website of AQCC (Available at: http: // airnow .tehran.ir / home / DataArchive.aspx).
Air quality data processing Data quality control is the most important part of air quality studies and estimating health effects. Data quality assurance was performed according to international organization guidelines such as World Health Organization (WHO), Environmental Protection Agency (EPA), and the European Union (22)(23)(24). Due to numerous operational and calibration problems related to air pollutant measuring stations, outlier detection and data cleaning from monitoring stations is very important and the results would have insu cient scienti c validity if this step is omitted.
In the present study, rst the data of all monitoring stations were obtained and then hourly data coverage of each pollutant in each station during the three years was determined. Included monitoring stations were only stations with ≥75% completeness of the total hours during the study period (22,24). Then, in order to outlier data detection, the modi ed Z-score approach proposed by some researchers for this purpose was used (3,23,25,26). Brie y, in order to identify outlier data, the following steps were used: Calculating the Z score for each hourly data at each station using the following equation:

SDofannualconcentrations
Calculating the following four conditions: Finally, the air quality data were detected as an outlier data and removed if they meet four abovementioned conditions.
Individual long-term exposure assessment To determine the long-term exposure of each participant to ambient air pollutants, the exact address according to the area, place, street, and alley in each year was obtained. In working women, if the place of work and living were different, the area and time spent in that area were also considered. Then, according ( ) to the location of monitoring stations and the location of the study subject, three of the nearest included air quality monitoring stations were identi ed for each participant and using the average annual data and IDW method, the 3-year annual mean of exposure was estimated as long-term exposure for each study subject.

Statistical analysis
Data were presented with mean ± standard deviation (SD) for continuous and frequency (percentage) for categorical variables. The ANOVA, t-test, and chi-square test were used to compare variables between study groups in the simple analysis step. A multiple logistic regression analysis was done between all pollutants criteria as independent variables with breast density as a dichotomous dependent variable (low = 0 and high =1). This analysis was done via two non-stepwise and stepwise algorithms and the result of the stepwise algorithm was chosen as signi cant pollution criteria. After that, another logistic regression analysis was done to evaluate whether the effects of pollutants on breast density were independent or affected by potential confounding variables. This analysis was done as three models. For the rst one, only signi cant pollutants were considered in the model. In the second model, medical and demographic variables were considered in addition to pollutants. In the last model, the history of medicine and supplement use was added to the previous variables. In the logistic regression analysis, odds ratio (OR) with 95% con dence interval (CI) was reported in addition to the p-value.

Results
Based on inclusion criteria 813 women were screened in this study. We excluded participants who were suspicious for malignancy in the current mammography (n=14), and who had written an incomplete address (n=8); nally 791 eligible women were recruited.
The mean age was 50.14 ± 7.61 (38-80) years old. About half of the women (50.1%) were premenopausal, and half of them were menopause (49.9%) at time of recruitment in the study. In the mammographies, low breast density was reported in 34.8% (n = 299) and high breast density in 62.2% (n = 492).  In the rst step, in a univariate analysis using a t-test, all six pollutants criteria were compared between low and high breast density. In this comparison, except for ambient air CO, which was on the borderline statistically signi cant (P-value = 0.054), other variables were not signi cantly different between the two groups ( Figure 1 & Supplementary table 2). Due to an unclear trend of ambient criteria air pollutants between the four categories of MBD, the comparison has been conducted only between high and low breast densities; and the comparison between the 4 categories is presented in the supplementary table 3.
Unlike univariate analysis, multiple regression analysis between six ambient air pollutants and MBD showed that outdoor air NO 2 (P-value = 0.003) and CO (P-value = 0.001) had a signi cant relationship with breast density. Logistic regression analysis with stepwise algorithm and breast density as a dependent variable showed that an increase in each unit of NO 2 (ppb) exposure was associated with an increased risk of breast density with an OR equal to 1.04 (95% CI: 1.01 to 1.07); and an OR equal to 1.47 for each 10 unit increase in NO 2 . Furthermore, CO level was associated with a decreasing risk of breast density in each 1 ppm (OR = 0.33, 95% CI = 0.17 to 0.64). None of the other pollutants were associated with breast density (  In order to evaluate whether the effects of pollutants on MBD is independent or disappears under the in uences of confounding variables, two others multiple analyses were performed. In the rst model, basic and reproductive factors (age, body mass index (BMI), Smoking, history of OCP usage, parity, menopause, and history of breast disease) were entered into the model. In the second model, metformin and aspirin intake, vitamin D, and calcium consumption were also entered into the model. Table 3 illustrates the results of the three models. Finally, multiple logistic regression analysis showed that ambient air CO (P = 0.018) and NO 2 (P = 0.022) had independent effects on the breast density.

Discussion
The present study has evaluated the precise impact of long-term exposure to six criteria ambient air pollutants on MBD in Iranian women for the rst time. Actually, many known and unknown factors are involved in breast tissue changes and eventually in breast cancer and it's not possible to control all confounding factors in a single context. By the way, based on the available evidence, we tried to evaluate the effects of six criteria ambient air pollutants on breast density considering the factors that seem to have an impact on MBD (basic and reproductive factors, aspirin, metformin, and supplement intake). To the best of our knowledge, there is no study with this broad level of assessment.
Our results represented that outdoor air NO 2 and CO exposure had statistically signi cant impacts on MBD. We found that an increased level of NO 2, as a marker of tra c-related air pollution (27), is associated with a higher MBD. Furthermore, ambient air CO concentration was associated with a lower MBD, while other criteria air pollutants were not related to MBD.
Our present results about ambient air NO 2 and PMx (PM 2.5 & PM 10 ) concentration were consistent with a recent systematic study and meta-analysis that found an increased risk of breast cancer with an increase in each 10 unit in NO 2 exposure (Hazard ratio (HR) = 1.02, 95% CI = 1.01-1.04), while PM 2.5 and PM 10 revealed no statistically signi cant associations with breast cancer risk (11). The results of our study on the relationship between air pollutants and MBD seem to be in line with studies that have examined the relationship between these pollutants and breast cancer.
Limited studies have evaluated the association between criteria ambient air pollutants and MBD with inconsistent results (16, 17, 19). Similar to our study, Du Pre and their colleague's results in the Nurses' Health Study couldn't support that recent particulate matter (PM 2.5 , PM air 2.5−10 , PM 10 ) or roadway exposure in uenced breast density (16). Two other studies had contradictory results with the present study (17,19). The Danish Diet, Cancer and Health Cohort investigated the association between long-term exposure to tra c-related air pollution (NO 2 , NO x ) and MBD in a prospective cohort of women aged 50 and older. They found a reverse association between air NO 2 level and MBD (OR= 0.89, 95% CI: 0.80-0.89 per 10 µg/m 3 ) with no interaction with menopause, smoking, or obesity (17). In Yaghjyan et al. study, women higher than 40 years old with known residential zip codes and estimated PM 2.5 and O 3 levels for the year preceding the mammogram date were included. They found that women with extreme breast density had higher mean PM 2.5 and lower O 3 exposure levels (19).
Numerous studies in line with our study have investigated the relationship between endocrine-disrupting chemicals (EDCs) and heavy metals with MBD (13,14,18). In a cross-sectional study in 725 women (40-65 years old), a higher urinary level of magnesium was associated with a higher MBD (13). In postmenopausal women (n = 264), women with high serum levels of BPA and mono-ethyl phthalate had an elevated breast density (14). In a large-scale study (n= 222,581), the relation of the MBD of women who underwent a routine screening mammogram in 2011 and residential levels of ambient air polycyclic aromatic hydrocarbons (PAHs) and metals was assessed. Higher residential levels of arsenic, cobalt, lead, manganese, nickel, or PAHs were individually associated with breast density. Comparing the highest to the lowest quartile, higher odds for dense breasts were observed for cobalt (OR = 1.60, 95% CI 1.56-1.64) and lead (OR = 1.56, 95% CI 1.52-1.64). These associations were stronger in premenopausal women (18). An exception is one cross-sectional study of PCBs, which reported some PCB congeners' plasma levels were associated with lower MBD in postmenopausal women (15).
Consistent with the present study, a review study by White and colleagues that summarized eight casecontrol studies and nine cohort studies suggested little evidence to support an association between particulate matter and breast cancer risk. More consistent ndings have reported a relation between NO 2 or NO X level and breast cancer (28).
In this study, we found a reverse association between CO level and MBD. Two recent studies that evaluated the effect of ambient air CO on breast cancer had equivocal results. A Korean study reported that CO concentration was positively and signi cantly associated with breast cancer (OR= 1.08, 95% CI= 1.06-1.10) (29) and another cohort study in Taiwan found that women who had CO poisoning were at a lower risk of developing breast cancer (30). We did not nd any studies that examined the association between MBD and carbon monoxide. Since there is a positive correlation between MBD and breast cancer, the results of our study are consistent with the Taiwanese study.
It is very important to note that according to the evidence, environmental pollutants are associated with a higher risk of invasive breast cancer and increased mortality in these patients. Investigation on breast found to be signi cantly associated with BC morbidity (32). These ndings show the importance of research on the impact of pollutants on women's health, and public health professionals and policymakers should consider these characteristics to develop relevant interventions and prevention strategies that are more cost-effective and e cient.
The advantage of this study is, we considered all the possibly effective factors and known determinants of MBD, all of which are estrogen-related.
As it was demonstrated in Table 1, 91 (11.5%) and 106 (13.4%) women in the present study sample consumed metformin and aspirin, respectively. Numerous studies have evaluated the effects of aspirin and metformin on MBD with inconsistent results (33)(34)(35)(36). In addition some researches have shown that higher vitamin D and calcium intake are associated with decreased MBD (37)(38)(39).
According to the ndings of the mentioned studies and the high percentage of women who had taken metformin (11.5%), aspirin (13.4%), vitamin D (49.7%), and calcium (44.4%) in our study, it seems that without considering the use of these drugs, the results may not be expressed correctly. However, our ndings showed that even by considering these factors, the results did not change.
It should be noted that most of the sampling in this study coincided with the worldwide onset of COVID-19 pandemic. Therefore, it is possible that the participants during this period were women at higher risk of breast cancer who had been referred for screening despite the COVID-19 pandemic.

Conclusion
In conclusion, based on our results, MBD was associated with higher levels of ambient air NO 2 and lower levels of CO. Considering our results and other evidences, important decisions at the national level are essential to reduce environmental pollution, especially air, to achieve sustainable development. Perhaps MBD monitoring as an available tool in each population, can help the prediction of future breast cancer occurrence and nding the high-risk geographic areas. Availability of data and materials

List Of Abbreviations
The datasets used and analyzed during the current study are available on reasonable request from the corresponding author.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementarytablesFinal.docx