Relationship between air pollutants and spontaneous abortion in a coal resource valley city: a retrospective cohort study

Abstract Objective Pollutants in the atmosphere have been linked to poor pregnancy outcomes in women. However, such investigations are scarce in metropolitan northern China. The major exposure window of air pollution affecting pregnant women is also unknown. Methods For the analysis, this retrospective cohort study enrolled 6960 pregnant women recorded at Tongchuan People’s Hospital from January 2018 to December 2019. Pollutant concentration values from the nearest monitoring station to the pregnant women were used to estimate exposure doses for each exposure window. Logistic regression models were created to investigate the connection between pollutants and spontaneous abortion while controlling for confounding factors. Results PM2.5 was a risk factor for spontaneous abortion in T3 (30–60 days before the first day of the last menstrual period [LMP]), (OR: 1.305, 95% CI: 1.143–1.490) and T4 (60–90 days before the first day of the LMP),(OR: 1.450, 95% CI: 1.239–1.696) after controlling for covariates. In the same window, PM10 was a risk factor (OR: 1.308, 95% CI: 1.140–1.500), (OR: 1.386, 95% CI: 1.184–1.621). In T2 (30 days before the first day of the LMP), T3, and T4, SO2 was a risk factor for spontaneous abortion (OR: 1.185, 95% CI: 1.025–1.371), (OR: 1.219, 95% CI: 1.071–1.396), (OR: 1.202, 95% CI: 1.040–1.389). In T3 and T4, NO2 was a risk factor (OR: 1.171, 95% CI: 1.019– 1.346), (OR: 1.443, 95% CI: 1.259–1.655). In T1 (from the first day of the LMP to the date of abortion), O3 was found to be a risk factor (OR: 1.366, 95% CI: 1.226–1.521). Conclusion Exposure to high levels of air pollutants before and during pregnancy may be a risk factor for spontaneous abortion in pregnant women. This study further illustrates the importance of reducing air pollution emissions.


Introduction
Air pollution is a significant public health concern, with 1.24 million individuals in China dying of air pollution-related diseases in 2017 [1].Recent studies have elucidated the detrimental effects of air pollution on various physiological systems and disease development [2][3][4].At the same time, traces of air pollutants have been found in studies of the causes of adverse pregnancy outcomes in pregnant women.Air pollutants have been associated with not only premature birth [5,6] but also spontaneous abortion (SAB) [7,8] and birth defects [9,10].A growing body of research has confirmed a link between air pollutants and adverse pregnancy outcomes; however, this relationship and the main window of action of the pollutants show heterogeneous results.Contrarily, some studies showed no such association [11][12][13].
Several studies in China have investigated the relationship between air pollutants and adverse birth outcomes, focusing primarily on the effects of atmospheric pollution on preterm birth and low birth weight [14][15][16], whereas studies on atmospheric pollution and SAB are rare.SAB is a natural pregnancy loss that occurs before 20 gestational weeks.Numerous causes contribute to SAB [17,18], the two most prominent causes being chromosomal abnormalities [19][20][21] and infections [22,23].Some studies have found that particulate matter from air pollution can cross the placental barrier and impair placental function [24], as well as induce oxidative stress [25], mitochondrial and nuclear DNA methylation [26], and endocrine disruption, all of which can lead to adverse pregnancy outcomes [27], which may be the mechanisms by which pollutants cause SAB.Women who have experienced SAB have a higher risk of postpartum anxiety, depression, and suicide [28,29].Additionally, women who have experienced SAB have an increased risk of developing cardiovascular diseases [30,31], diabetes [32,33], and hypercholesterolemia later in life [34].The only studies on abortion in China were conducted in developed cities [15,35].
Therefore, this study aimed to evaluate the association between six air pollutants and SAB and identify the primary exposure window in Tongchuan, a typical underdeveloped city in northwestern China.

Study participants
Tongchuan has a coal-bearing area of 387.2 km 2 , and its vast coal reserves determine its economy and energy structure.The use of a large number of coal mines has further aggravated air pollution in Tongchuan.We collected data from 8,085 pregnant women admitted to the Obstetrics and Gynecology Department of Tongchuan People's Hospital and the Central Hospital of Tongchuan Mining Bureau between January 1, 2018, and December 31, 2019.The sample comprised patients aged �18 years, with no mental illness, living in a local residence for >1 year during pregnancy, and with a specific residential address.In total, 717 pregnant women were excluded because of premature birth, low birth weight, birth abnormalities, induced abortions, or stillbirths.Additionally, 407 pregnant women were excluded owing to missing information.Finally, 6,659 women with full-term birth and 301 with SAB were included.Data on demographic characteristics, reproductive history, occupational and residential history, birth outcomes, maternal complications, maternal comorbidities, and maternal history of hypertension were collected from the pregnant women.

Pollutant exposure assessment
Because the pregnancy of the women studied was no longer than 28 weeks owing to SAB, we divided the exposure window into four trimesters (T 1 , T 2 , T 3 , and T 4 ).T 1 : from the first day of the last menstrual period (LMP) to the date of abortion; T 2 : 30 days before the first day of the LMP; T 3 :60-30 days before the first day of the LMP; and T 4 : 90-60 days before the first day of the LMP.The daily pollution concentrations from 2017 to 2019 in Tongchuan were measured at four locations (Figure 1).Only ozone (O 3 ) had a maximum 8-h sliding average among the six pollutants [particulate matter] PM 10 , PM 2.5 , sulfur dioxide [SO 2 ], nitrogen dioxide [NO 2 ], carbon monoxide [CO], and O 3 , whereas others had 24-h sliding averages.Pollutant data was gathered from the Shaanxi Province's real-time air quality release system.Therefore, we chose the concentration at the nearest monitoring station to the pregnant women's addresses as the exposure concentration, and the average daily pollutant concentration from 2017 to 2019 was measured.

Statistical methods
Data were collated using EXCEL and analyzed using IBM SPSS 26.The K-S test was used to check the normality of the pollutant, and the mean, standard deviation, and interquartile spacing were used to describe the pollutant.Participant data were analyzed using the chi-squared test.Pollutant data were standardized using (actual-mean)/standard deviation.After controlling for possible confounders in the chi-square analyses, air pollutants were added as continuous variables to the single pollutant model, and odds ratios (ORs) and 95% confidence intervals (CI) were calculated.The relationship between standard unit concentrations (10 lg/m 3 ) of air pollutants and SAB was investigated.To examine the stability of the model, the effect of each pollutant on SAB was analyzed as a single factor and by adding other risk factors layer by layer.Sensitivity analysis was conducted in three steps: (1) Model 1: one-way analysis of each pollutant concentration and SAB.(2) Model 2: occupation and age were added to Model 1 to analyze the effect of each pollutant on SAB.(3) Model 3: Complicating diseases, comorbid diseases, and hypertension during pregnancy were added to Model 2 to analyze the effect of each pollutant on SAB.

General description of air pollution in Tongchuan City
PM 2.5 and PM 10 levels consistently surpassed the national secondary standard [36], whereas NO 2 and O 3  THE JOURNAL OF MATERNAL-FETAL & NEONATAL MEDICINE levels occasionally did so.Among the six pollutants, only CO satisfied the national threshold.The Chinese ambient air concentration standards (annual averages) are listed in Table 1.Every year, O 3 increased during summer, whereas the remaining five pollutants peaked during winter (Figure 2).

Demographic characteristics of study participants
The general data of the pregnant women and their relevant confounding factors are shown in Table 1.SAB at 25-30 years, pregnancy times >1, number of births �1, number of cesarean sections �1, occupation in agriculture, LMP in winter, regular menstrual cycle, absence of complicating diseases, presence of comorbid diseases, and absence of hypertension in pregnancy accounted for the majority of participants.Maternal age, pregnancy, delivery, number of cesarean sections, maternal occupation, season of last menstruation, menstrual cycle, complicating diseases, comorbid diseases, and gestational hypertension were all positively associated with the occurrence of SAB, according to the chi-square test (Table 2).

Comparison of pollutant concentrations during pregnancy
Compared with women with full-term pregnancies, pregnant women who experienced SAB at T 1 were exposed to higher concentrations of PM 2.5 , PM 10 , NO 2 , and O 3 ; those who experienced SAB at T 2 were exposed to higher concentrations of PM 2.5 , PM 10 , SO 2 , NO 2 , and CO; those who experienced SAB at T4 were exposed to higher concentrations of PM 2.5 , PM 10 , and NO 2 (Table 3).

Logistic regression analysis of SAB and air pollutants
Before adjustment, exposures to PM 2.5 , PM  3).

Sensitivity analysis
The relationship between cumulative exposure to each pollutant and SAB did not change significantly when the indicators were added layer by layer, indicating good stability of the main model (Table 4).

Discussion
This study analyzed the correlation between pregnant women and air pollutants in two tertiary hospitals in Tongchuan City, China.The chi-squared analysis showed that maternal age, pregnancy, delivery, number of cesarean sections, maternal occupation, LMP, menstrual cycle, complicating diseases, comorbid diseases, and hypertension during pregnancy were associated with SAB.After adjusting for these confounders, pre-pregnancy exposure to PM 2.5 , PM 10 , and SO 2 and early pregnancy exposure to O 3 significantly increased SAB risk.Furthermore, we found that CO exposure was negatively associated with SAB risk in four windows.The negative correlation between air pollutants in the exposure window and SAB may be due to methodological errors or unidentified confounding factors.SAB is one of the most common complications in pregnant women, with women of childbearing age having a 10% chance of experiencing it [37].Not only can SAB disrupt a pregnant woman's family harmony [38], it can also lead to uterine infections [39].Several studies reported findings similar to ours.Ha et al. identified a relationship between O 3 exposure and SAB (HR: 1.12, 95% CI: 1.07-1.17) in a prospective cohort study [40].They also found that long-term exposure was linked to SAB.According to a study conducted in Jiangsu, China, each 10 mg/m 3 increase in PM 2.5 was associated with a 43.3% increase in the likelihood of experiencing SAB [41].A study conducted in five cities found that, for every 10 lg/m 3 increase in PM 10 or O 3 concentration, SAB incidence increased by 19.7% and 33.6%, respectively [42].Wang found that NO 2 exposure significantly affected SAB incidence (OR: 1.68, 95% CI: 1.28-2.21),but CO exposure had no effect [35].However, some studies reported contradictory findings.Hou et al. found a link between SO 2 exposure in the first month of pregnancy and SAB at 14 weeks (OR: 19.76, 95% CI: 2.340-166.71)[43]; however, we found a stronger link between SO 2 exposure in the 3 months before the LMP and SAB.Di Ciaula et al. found that SAB was not related to NO 2 exposure, which contradicts our findings [42].These heterogeneous results are most likely due to differences in experimental designs and inconsistencies in exposure assessment methods.Dastoorpoor et al. found that the hysteresis impact of SO 2 was highest within 1 h of exposure (relative risk [RR]: 1.007, 95% CI: 1.003-1.012)and on the same day of exposure (RR: 1.001, 95% CI: 1.000-1.002)[44].Another study found that the delayed effect of NO 2 was highest within 5 h of exposure (OR: 3.289, 95% CI: 1.568-5.011),with women aged >30 years being more sensitive than those aged 18-29 years [45].The period of air pollutants of the lag effect varied among pollutants, providing a new area for future research.
Although the mechanisms of action between air pollutants and SAB are largely understood, there are some new findings.These new findings may explain the mechanisms linking pre-pregnancy exposure to air pollutants and SAB.PM from air pollution has been found to contain estrogenic, antiprogestational, and antiandrogenic substances that cause oxidative stress and can damage oocyte or sperm DNA, which is essential for oocyte fertilization, chromosome segregation during meiosis, sperm-egg fusion, and embryonic developmental competence [46].Preconception and pregnancy exposures have been associated with maternal steroid changes that may have some effect on the mother and the fetus [47].It has also been shown that exposure to air pollutants results in reduced growth and altered protein expression in early trophoblast cells [48].
The present study has some limitations.First, it is possible that a significant proportion of pregnant women who had SAB were managed outside the hospital or in hospital outpatient clinics and, therefore, were not included in the study cohort.This limitation may have affected the overall study results because of the non-inclusion of this subset of pregnant women.Therefore, we speculate that the lack of data from this population may have contributed to the null or protective outcomes between T 0 -phase air pollutants and SAB.Follow-up studies should comprehensively determine early pregnancy losses to fully elucidate the relationship between air pollutants and SAB.Second, the possibility of the lagged effects of exposure to air pollutants was not considered.That is, exposure to atmospheric pollutants on the first day may produce some effects not only on the first day but also on subsequent days.This may have affected our results.In addition, it is difficult to estimate individual exposure accurately.A method based on the nearest monitoring site may result in misclassification: if a pregnant woman works during her pregnancy, her exposure environment constantly changes between indoor and outdoor environments.When a pregnant woman's address changes throughout her pregnancy, obtaining information after the move is challenging, and the exposure level may be miscalculated.Finally, some possible influencing factors such as maternal nutritional status, socioeconomic status, prenatal care services, uterine physiological abnormalities, antiphospholipid syndrome, and autoimmune diseases were not included in the study.The lack of inclusion of these confounding factors may have limited the interpretation of the study results.
This study offers valuable insights for women who are planning conception or are currently pregnant, enabling them to safeguard their well-being and strategize their outdoor engagements.

Conclusion
Exposure to air pollutants during and before pregnancy increases the risk of SAB.PM 2.5 , PM 10 , SO 2 , NO 2 , and O 3 are risk factors for SAB.Therefore, it is crucial to reduce air pollutant emissions, arrange travel plans during pregnancy, and educate pregnant women regarding the potential adverse effects of air pollutants on the risk of SAB.

Figure 1 .
Figure 1.Distribution map of outdoor air quality monitoring stations in Tongchuan City.

Figure 2 .
Figure 2. Description of air pollutant concentrations in the Tongchuan urban area, from 2017 to 2019.

Figure 3 .
Figure 3. Logistic regression models of single and compound pollutants and spontaneous abortion.Note: Adjusted for maternal age, pregnancy, delivery, number of cesarean sections, maternal occupation, last menstrual season, menstrual cycle, Complicating diseases, comorbid diseases, and hypertension during pregnancy.

Table 2 .
General information analysis of pregnant women.
10, SO 2 , and NO 2 in T 2 , T 3 , and T 4 were risk factors for SAB.In addition, exposure to O 3 in T 1 was a risk factor for SAB.

Table 3 .
Comparison of pollutant exposure of pregnant women during diferent pregnancies.