Exposure to air pollution and risk of hypertensive disorders of pregnancy: a retrospective cohort study

Abstract

Background: There is a lack of epidemiological evidence on the association between air pollution exposure and HDP in pregnant women in small and medium-sized cities, and the air pollution situation in small and medium-sized cities cannot be ignored and the health of their inhabitants deserves to be considered.

Objective: To explore pollutants affecting the risk of HDP in small and medium-sized cities and to explore differences in the effects of air pollution on GH and PE.

Methods: A total of 9,820 women who delivered at Handan Maternal and Child Health Hospital in Hebei Province from February 2018 to July 2020 were included in the study. The effects of air pollution exposure on the risk of HDP during preconception and pregnancy were assessed using logistic regression models and principal component logistic regression models.

Results: In multi-pollutant models adjusted for covariates, each 20 μg/m³ increase in PM_2.5 and PM₁₀ and each 10 μg/m³ increase in NO₂ during the Pre_T period was associated with a 2.4% (OR=1.024, 95% CI: 1.010-1.039), 2.5% (OR=1.025, 95% CI: 1.012-1.037), and 2.0% (OR=1.020, 95% CI: 1.005-1.037) increase in the risk of HDP, respectively. PM_2.5, PM₁₀ and NO₂ exposure during the Pre_T period also increased the risk of GH and PE in pregnant women, and the risk of each pollutant to GH was lower than that of PE. In addition, O₃ exposure per 20 μg/m³ increment during the T period increased the risk of GH with an OR of 1.026 (95% CI: 1.002 to 1.050).

Conclusions: PM_2.5, PM₁₀, NO₂ and O₃ exposure had a significant effect on the increased risk of developing HDP in pregnant women, and the effects of pollutants on the risk of GH were different from those on PE.

1. Introduction

Hypertensive disorders in pregnancy (HDP) are usually divided into two categories, including gestational hypertension (GH) and pre-eclampsia (PE). It is not clear what the differences are that lead to the development of GH and PE in pregnant women during pregnancy, but the two conditions may have overlapping processes. The pathophysiological mechanisms underlying both disorders have not been elucidated. Current theories suggest that although the etiology of HDP is clearly multifactorial, the most widely accepted theory is that inadequate fetal cell trophoblast infiltration leads to inadequate maternal spiral arterial vascular remodeling (Michikawa et al., 2021; Naderi et al., 2017). HDP is detrimental to the short and long-term health of both the mother and her fetus. In pregnant women, HDP is associated not only with their endothelial abnormalities and liver and kidney dysfunction, but also with the development of cardiovascular disease, neurological changes, end-stage renal disease and type II diabetes in later life (Dall'Asta et al., 2021; Mielke et al., 2016; Theilen et al., 2016; Wang et al., 2013). With the development of GH into preeclampsia and even eclampsia, HDP significantly increases the risk of maternal death in the perinatal period (Dasari and Habeebullah, 2010; Un Nisa et al., 2019). In addition to its impact on the mother's health, HDP is a major risk factor for intrauterine growth restriction, preterm birth and low birth weight (Padula et al., 2019). Infants of affected mothers are also at greater risk of hospitalization for neurological, respiratory, endocrine and metabolic complications (Turbeville and Sasser, 2020; Xu et al., 2014). In addition, children of mothers with HDP are more likely to have adolescent hypertension and are at increased risk of stroke in adulthood (Assibey-Mensah et al., 2019).

While treatment may mitigate the adverse effects of HDP to some extent, prevention of the disease through modification of risk factors remains a central strategy to avoid its short-term and long-term adverse health effects. Currently known risk factors for HDP include maternal characteristics such as infertility, obesity, advanced maternal age, adolescent pregnancy, pre-pregnancy hypertension or diabetes, poor diet, and family history of complications during pregnancy; and pregnancy-related factors such as multiple pregnancies and placental abnormalities (Fleisch et al., 2016; Naderi et al., 2017; Rana et al., 2019). Although past evidence has provided insight into the importance of preventing complications in pregnancy, there are few modifiable risk factors or preventive factors from a public health perspective. In this light, the hypothesis that ambient air pollution is a risk factor for pregnancy complications deserves attention. Air pollution is an increasing problem in the context of rapid global urbanization and industrialization, especially in developing countries. Recent data from the World Health Organization show that air pollution levels remain dangerously high in many parts of the world, with 90% of the global population breathing air that exceeds the World Health Organization's guideline limits for high pollution levels, and that air pollution is particularly severe in low- and middle-income countries (2018a). Numerous epidemiological studies have shown that air pollutants can not only cause damage to the human respiratory and cardiovascular systems, but also have adverse effects on the nervous and reproductive systems, and some pollutants are mutagenic and carcinogenic (Bala et al., 2021; Gu et al., 2019; Lim and Thurston, 2019; Monrad et al., 2017; Nazar and Niedoszytko, 2022; Roberts et al., 2019; Yazdi et al., 2019).

Although the pathophysiological link between ambient air pollution and HDP is unclear, several possible mechanisms have been proposed, including systemic inflammation and oxidative stress (Brook et al., 2010; Kelly and Fussell, 2017). Several studies have investigated the association between air pollution and HDP. Due to the distinction between PE and GH, different HDP outcomes have been reported in different studies. Some studies have reported only PE (Mendola et al., 2016; Wang et al., 2018; Wu et al., 2009; Yorifuji et al., 2015; Zhai et al., 2012), others have reported only GH (Zhu et al., 2017), while others have reported both PE and GH and their aggregation (HDP) (Huang et al., 2015; Pedersen et al., 2017; Savitz et al., 2015). The relationship between air pollution exposure and the risk of developing HDP is inconclusive, with some studies suggesting that air pollution increases the risk of developing HDP in pregnant women (Mandakh et al., 2020; Michikawa et al., 2021; Pedersen et al., 2017; Wang et al., 2018) and some studies reaching the opposite conclusion (Choe et al., 2018; Melody et al., 2020), but others not finding a significant association between the two (Madsen et al., 2018). Most studies have been conducted in Europe and the United States in some countries with low pollution levels, and data on non-Western populations are very limited (Wang et al., 2018). China, as the world's most populous country, is facing a serious air pollution problem. Handan city is one of the most polluted cities in China in terms of air pollution due to multiple factors such as the concentration of heavy industries that consume a lot of energy such as iron and steel, high traffic flow, and meteorological conditions and geography that are not conducive to the diffusion of pollutants (2018b; 2019; 2020; 2021).

Accordingly, this retrospective cohort study was conducted in Handan City, Hebei Province. Concentrations of air pollutants, including inhalable particulate matter (PM₁₀), lung-entering particulate matter (PM_2.5), sulfur dioxide (SO₂), nitrogen dioxide (NO₂), carbon monoxide (CO) and ozone (O₃), were obtained from monitoring stations. Logistic regression models and principal component logistic regression models were used to assess the effects of each pollutant unadjusted and adjusted for the remaining five pollutants on the risk of HDP in pregnant women, respectively. The effects of air pollution on GH and PE were evaluated separately due to their differences. In addition, subgroup analyses of pregnant women of different ages, parities and education levels were conducted in this study to investigate the differences in the effects of air pollution on people with different characteristics.

2. Materials and methods

2.1. Study population and design

This birth cohort study was conducted in Handan City, Hebei Province, China. Handan city is in the south of Hebei province and has a warm temperate continental monsoon climate with four distinct seasons. It has 6 districts, 11 counties and 1 county-level city with a total area of 12,000 km². The total registered population is 10.57 million, the resident population is 9.55 million, and the birth rate is 11.80%. This is a retrospective cohort study, and the participants were women who delivered at the Handan Maternal and Child Health Hospital in Hebei Province from February 2018 to July 2020. The Handan Maternal and Child Health Hospital receives more than half of the maternal deliveries in Handan each year, with case data covering the entire city. Hospital records detail maternal and fetal characteristics as well as clinical data on pregnancy and delivery. Pregnant women with addresses outside the administrative division of Handan city, missing information on any variable, multiple deliveries, gestational age less than 26 weeks or gestational age greater than 44 weeks, and chronic hypertensive disease before pregnancy were excluded from this study. Ultimately, 9,820 pregnant women were included in the analysis. The screening details are shown in Fig. 1. This study was based on a de-identified dataset and therefore did not require review or approval.

2.2. Outcome and covariates

Handan Maternal and Child Health Hospital used the current International Classification of Diseases 10th Revision to name, classify and treat hypertensive disorders of pregnancy. Gestational hypertension was diagnosed as the first occurrence of systolic/diastolic blood pressure ≥ 140/90 mmHg after 20 weeks of gestation (at least twice at 6-hour intervals), the absence of proteinuria, and the return of blood pressure to normal at 12 weeks after delivery. Pre-eclampsia was defined as the presence of a systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg in a pregnant woman after 20 weeks of gestation with any of the following: urine protein quantification ≥ 0.3 g/24 h or urine protein/creatinine ratio ≥ 0.3. Pre-eclampsia was also diagnosed without proteinuria but with involvement of any of the following organs or systems: vital organs such as the heart, lungs, liver, or kidneys, or abnormal changes in the hematologic, digestive, or neurologic systems, and placental-fetal involvement.

We assessed the following potentially relevant covariates based on prior evidence(Umesawa and Kobashi, 2017): maternal age (< 25, 25 ~ 30, 30 ~ 35, > 35 years old), education (less than bachelor, bachelor or above), health insurance (none, employee medical insurance, urban and rural medical insurance), parity (nulliparous, multiparous), conception year (2017, 2018, 2019), conception season [spring (March-May), summer (June-August), autumn (September-November), winter (December-February)], and previous adverse pregnancy and childbirth (yes, no).

2.3. Exposure assessment

Data on the daily mean concentrations of PM₁₀, PM_2.5, SO₂, NO₂ and CO and maximum 8-hour average concentrations of O₃ in Handan were obtained from the Handan Ecology and Environment Bureau at 44 monitoring sites which covered all exposure windows of all participants. The distribution of pregnant women's addresses and monitoring stations is shown in Fig. 2, where the number of monitoring stations is sufficiently large and spatially evenly distributed. Therefore, after excluding participants who were 20 km away from the monitoring stations, the pollutant concentration data from the nearest monitoring station to the address were directly matched to pregnant women (Michikawa et al., 2015; Vinikoor-Imler et al., 2012). Afterwards, for each pollutant, the mean exposure was calculated for each pregnant woman based on her gestational weeks and delivery date for the following windows: (1) preconception (13 weeks before pregnancy, Pre_T); (2) first trimester (1–13 gestational weeks, T1); (3) second trimester (14–26 gestational weeks, T2); (4) the first two trimesters (T). To ensure that exposures were prior to outcomes, exposures after 26 weeks of gestation were not considered in this study.

2.4. Statistical analyses

Differences in selected characteristics between pregnant women with and without HDP were compared using the χ² test. A descriptive analysis of the air pollution exposure levels of pregnant women in different exposure windows was performed and the correlation between the average exposure levels of pollutants in different exposure windows was demonstrated by the Spearman correlation coefficient.

The effect of each air pollutant on HDP for each exposure window was assessed using a logistic regression model. To facilitate discussion of the results and comparison with other data in the literature, the pollutant concentrations were used as continuous variables to represent the findings in this study. The odds ratios (ORs) and 95% confidence intervals (95% CIs) of HDP with every 20 µg/m³ increase in PM₁₀, 20 µg/m³ increase in PM_2.5, 20 µg/m³ increase in O₃, 10 µg/m³ increase in SO₂, 10 µg/m³ increase in NO₂ and 0.5 mg/m³ increase in CO during each exposure window were estimated. Two single-pollutant logistic models were established. Model 1 was a crude model; model 2 was adjusted for all covariates. The detailed methodology is provided in Part I of the Supplementary Material.

Due to the simultaneous presence of multiple pollutants in the environment in which pregnant women live, adjustment for other pollutants is required to estimate the effects of each pollutant more accurately. Parameter estimation in the logistic regression models requires that the respective variables be independent of each other. The six pollutants are strongly correlated with each other, and putting all six pollutants into the model at the same time will lead to multicollinearity problem. The existence of multicollinearity will lead to inaccurate estimation of the regression coefficients, or even make the regression coefficients in opposite directions, which is detrimental to the interpretation of the results. Therefore, the principal component logistic regression method was used for the multi-pollutant model(Aguilera et al., 2006; Barker and Brown, 2001; Escabias et al., 2007). Two multi-pollutant models were established. Model 1 was a rough model with the first three principal component variables of six pollutants as independent variables; model 2 was also adjusted for all covariates, with the independent variables being the first eight principal component variables of the six pollutants and all covariates. The detailed methodology is provided in Part II of the Supplementary Material.

In addition, participants were divided by age, parity and childbirth and education level, and subgroup analyses were performed with single-pollutant and multi-pollutant models, respectively, adjusting for all covariates except the grouping variable. All statistical analyses were performed using R software, version 3.5.3, with a two-sided test with an alpha level of 0.05.

3. Results

^a Analysis of χ² test.

The participant characteristics are shown in Table 1 and Tables S.1-S.2. Among the 9,820 participants included in the analysis, the percentages of HDP, GH and PE patients were 22.37%, 7.89% and 15.84%, respectively. The age of pregnant women was mainly concentrated between 25–35 years old, with 38.66% and 35.58% of the total study population between 25–30 years old and between 30–35 years old, respectively. More than half of the pregnant women had university or higher education, accounting for about 51.40%. About 42.02% of the pregnant women had urban and rural health insurance types, and about 36.76% had employee health insurance types. More than half of the pregnant women had pregnancy experience, and about 30.48% of the pregnant women had a history of bad pregnancy and delivery.

The statistical description and Spearman's correlation analysis of the average exposure levels of pollutants in different exposure windows for pregnant women are shown in Fig.S.1 and Fig.S.2, respectively. The medians of PM₁₀, PM_2.5, SO₂, NO₂, CO and O₃ in study periods ranged from 119.79 to 125.30 µg/m³, 59.34 to 68.71 µg/m³, 21.11 to 23.31 µg/m³, 37.12 to 39.79 µg/m³, 1.20 to 1.28 mg/m³ and 102.22 to 123.86 µg/m³, respectively. From the upper and lower quartiles, there is a large variability in the group exposure levels of each pollutant in each window. According to Spearman's correlation coefficient, the correlation between pollutants in each window is strong, so it is not possible to include any two pollutants directly and simultaneously in the logistic regression model.

The ORs and 95% CIs for each pollutant in the single-pollutant model for HDP, GH, and PE are shown in Table 2, Table S.5, and Table S.6, respectively. In the single-pollutant model without adjustment for covariates, each 0.5 mg/m³ increment of CO during the Pre_T period had a significant effect on the risk of HDP, GH, and PE in pregnant women. In the single pollutant model adjusted for covariates, no pollutant was associated with the risk of HDP. O₃ exposure per 20 µg/m³ increment during T was significantly positively associated with the risk of GH (OR = 1.211, 95% CI: 1.094–1.340), and the effect of O₃ exposure during T1 was higher than that during T2, with ORs of 1.128 (95% CI: 1.046–1.218) and 1.093 (95% CI: 1.015–1.177), respectively. However, none of the pollutants had a significant effect on the risk of PE in pregnant women.

^a Crude models.

^b Models were additional adjusted for maternal age, education, health insurance, parity, conception year, conception season, and previous adverse pregnancy and childbirth.

* P < 0.05.

Table 3, Table S.7, and Table S.8 show the OR and 95% CI for each pollutant in the multi-pollutant model for the risk of HDP, GH, and PE, respectively. In the model without adjustment for covariates, no significant effect of any pollutant on the risk of developing HDP and GH was found. In the model adjusted for covariates, each 20 µg/m³ increase in PM_2.5 and PM₁₀ and each 10 µg/m³ increase in NO₂ during the Pre_T period was associated with a 2.4% (OR = 1.024, 95% CI: 1.010–1.039), 2.5% (OR = 1.025, 95% CI: 1.012–1.037), and 2.0% (OR = 1.020, 95% CI: 1.005–1.037) increase in the risk of HDP in pregnant women, respectively. PM_2.5, PM₁₀, and NO₂ exposure during the Pre_T period also increased the risk of GH and PE in pregnant women, and the risk of GH for each pollutant was less than the risk of PE. In addition, O₃ exposure per 20 µg/m³ increment during the T period increased the risk of GH with an OR of 1.026 (95% CI: 1.002–1.050).

^a Models were adjusted for all other pollutants.

^b Models were additional adjusted for maternal age, education, health insurance, parity, conception year, conception season, and previous adverse pregnancy and childbirth.

* P < 0.05.

In the multi-pollutant model, SO₂ and CO had no significant effect on the risk of HDP, GH and PE after adjusting for confounding variables, so only the results for the other four pollutants are presented in the stratified analysis. Results stratified by age showed that PM_2.5 exposure during T2 increased the risk of HDP in pregnant women ≤ 25 years (OR = 1.054, 95% CI: 1.005 to 1.105) and during T1 increased the risk of GH in pregnant women aged 25–30 years (OR = 1.030, 95% CI: 1.008 to 1.053) (Fig.S.3). O₃ exposure during the Pre_T period had a significant effect on the increased risk of HDP and GH in pregnant women aged 30–35 years, and O₃ exposure during the T period had a significant effect on the increased risk of GH in pregnant women > 35 years. There was no significant effect of any pollutant on the risk of PE in pregnant women of different ages.

For nulliparous women, PM_2.5 exposure during T1 increased the risk of HDP and NO₂ exposure during Pre_T increased the risk of GH, with ORs of 1.016 (95% CI: 1.002–1.030) and 1.034 (95% CI: 1.001–1.068), respectively (Fig.S.4). Moreover, PM_2.5 and PM₁₀ exposure during the Pre_T period increased the risk of PE by 7.4% (OR = 1.074, 95% CI: 1.019–1.131) and 8.0% (OR = 1.080, 95% CI: 1.024–1.139), respectively.

The effect of air pollution on the risk of HDP, GH and PE among pregnant women with different educational levels is shown in Figure S.5. PM_2.5 exposure during the T2 period had a significant effect on the risk of HDP among pregnant women with less than a bachelor's degree (OR = 1.025, 95% CI: 1.002–1.049). PM_2.5, PM₁₀ and NO₂ exposure during the Pre_T period also increased the risk of PE in pregnant women with less than a bachelor's degree, with ORs of 1.081 (95% CI: 1.026–1.138), 1.097 (95% CI: 1.038–1.159) and 1.083 (95% CI: 1.031–1.138), respectively.

4. Discussion

As a highly polluted city with a large population, the data from Handan are representative and generalized. In this study, we selected data from pregnant women who delivered at the Maternal and Child Health Hospital in Handan, Hebei Province, from February 2018 to July 2020 to investigate the effects of air pollutant exposure during preconception and pregnancy on the risk of hypertensive disorders in pregnancy. The results showed that PM_2.5, PM₁₀, NO₂ and O₃ exposure had a significant effect on the increased risk of HDP in pregnant women, and the effect of pollutants on the risk of GH was smaller than that of PE. Furthermore, the effects varied by age, gestational age and education of different pregnant women.

The results of the study included the effects of both adjusted and unadjusted confounding variables, which are in fact the main causes of HDP in pregnant women (Umesawa and Kobashi, 2017). Therefore, when exploring the effect of air pollution on HDP, the model results adjusting for confounding variables were closer to the actual effect. Results adjusted for other pollutants and covariates showed a significant effect of PM_2.5, PM₁₀ and NO₂ exposure during the Pre_T period on the risk of developing HDP in pregnant women. A study in Denmark showed that a 10 µg/m³ increase in NO₂ and a 10 dB increase in road traffic noise in the first trimester of pregnancy were associated with an increased risk of HDP with ORs of 1.07 (95% CI: 1.01–1.13) and 1.08 (95% CI: 1.02–1.15), respectively, demonstrating that road traffic exposure increases the risk of HDP (Pedersen et al., 2017). Two other studies also demonstrated that PM_2.5, PM₁₀ and NO₂ are risk factors for HDP risk (Huang et al., 2015; Savitz et al., 2015), which is consistent with our study. Preconception is also a critical exposure window, and preconception care focused on improving lifestyle and reducing adverse risk factors for women has been shown to prevent HDP (Mendola et al., 2016; Zhu et al., 2017). PM_2.5, PM₁₀ and NO₂ in the air in Handan mainly originate from traffic and industrial source emissions (Yang et al., 2018), so pregnant women should also be protected from air pollution during traffic and work during the preparation phase.

We found a statistically significant association between O₃ exposure during T period and increased risk of GH in pregnant women (Table S.5, Table S.7). Previous studies have demonstrated a possible strong association between exposure to air pollution and GH in pregnant women. Two US studies concluded that PM_2.5 and NO₂ exposures during pregnancy were positively associated with GH risk (Nobles et al., 2019; Savitz et al., 2015); another Dutch study found that exposure to PM₁₀ and CO was associated with an increased risk of GH (van den Hooven et al., 2011); a newly published meta-analysis showed that every 5 µg/m³ increment of PM_2.5, exposure during T1 and T2 and every 5 µg/m³ increment of PM₁₀ exposure during T1 significantly increased the incidence of GH with ORs of 1.11 (95% CI: 1.01–1.23), 1.16 (95% CI: 1.05–1.29), and 1.04 (95% CI: 1.02–1.07), respectively (Cao et al., 2021). The results of this study also showed a higher effect of O₃ exposure in the T1 period than in the T2 period (Table S.5). Early gestation may be a critical window of susceptibility for the occurrence of GH, possibly because air pollution during this period interferes with the maternal vascular remodeling process (Lee et al., 2013; Mobasher et al., 2013). This has been confirmed in other studies. One study found that O₃ exposure in first trimester had a greater effect on the risk of GH compared to second trimester (OR_T1=1.04, 95% CI: 1.03–1.06; OR_T2=1.03, 95% CI: 1.02–1.04) (Hu et al., 2017). Another systematic review also concluded that pregnant women are more susceptible to PM_2.5, in the first trimester of pregnancy and recommended increased protection for pregnant women during this period (Sun et al., 2020).

The results of the study showed that PM_2.5, PM₁₀ and NO₂ exposures during the Pre_T period were risk factors for PE risk (Table S.8). Previous exposure studies have also investigated the association between air pollution exposure during preconception and PE and showed that CO per 0.1 ppm increment was significantly associated with the risk of PE in pregnant women after adjusting for maternal characteristics and season of conception (OR = 1.07, 95% CI: 1.02–1.13) (Rudra et al., 2011). A study by Mendola et al (Mendola et al., 2016) also showed that per IQR increment of CO, NOx, O₃ and SO₂ exposure during preconception increased the risk of PE in pregnant women with asthma, although the results were not significant. It has been suggested that early-onset PE (diagnosed before 34 weeks of gestation) and late-onset PE (diagnosed after 34 weeks of gestation) and mild and severe PE have different prevalence and prognosis, with different possible pathogenesis and risk factors, and that air pollution may have different toxic effects depending on the time of exposure (Pedersen et al., 2017). However, due to the limitations of the level of detail of the information contained in the disease diagnostic results in the raw data, we were unable to distinguish different levels or periods of PE and study the effects of air pollution on them separately.

GH and PE are the two main types of HDP. It is unclear whether GH and PE are two different diseases or if they are somehow related. Unlike GH, PE is associated with significant proteinuria and other maternal organ dysfunction in addition to elevated blood pressure (Naderi et al., 2017). PE is a more serious disease and most studies have focused only on PE because of its clinical importance. Our study found that PM_2.5, PM₁₀ and NO₂ exposure during the Pre_T period were all less risky for GH than for PE, and that O₃ exposure was the risk factor but not for PE (Tables S.7-S.8). Another study found that higher levels of most criteria pollutants were associated with a lower risk of PE in preconception and early pregnancy, while higher levels of pollutants in mid-pregnancy were associated with a higher risk of GH (Nobles et al., 2019). There are some differences in the effects of air pollution on GH and PE in terms of pollutant class, exposure window, and effect size, which may be explained to some extent by potential differences in the etiology of GH and PE.

The effect of O₃ exposure during the T period on GH in pregnant women was lower in the multi-pollutant model compared to the single-pollutant model. Several studies have also shown that the pollutant effects in single-pollutant models are higher than those in two-pollutant models (Choe et al., 2019; Jo et al., 2019; Michikawa et al., 2015; Pan et al., 2017; Savitz et al., 2015). This may be because the single-pollutant model without adjusting for the remaining pollutants overestimates the pollutant effects. The coexistence of various air pollutants in the actual environment species requires adjusting for the effects of the remaining pollutants other than the target pollutant, and using a multi-pollutant model better reflects the effects of a single pollutant in a co-exposed environment. To improve their parameter estimation in the presence of multicollinearity, scholars have developed different statistical analysis methods such as principal component regression, partial least squares regression, and ridge regression (Aguilera et al., 2006; Barker and Brown, 2001; Escabias et al., 2007). In the future, more suitable methods need to be explored to address multicollinearity in multi-pollutant models to accurately reflect the effects of pollutants.

This study has some advantages. This study focused on the effect of air pollution on the risk of hypertensive disorders in pregnancy in small and medium-sized cities, using Handan city as an example. To identify the risk factors that really influence pregnant women developing hypertensive disorders of pregnancy, the effect of each pollutant after excluding other pollutants was skillfully evaluated using principal component logistic regression models. In addition, the effects of air pollution exposure on the risk of GH and PE were evaluated and compared separately because of the differences between GH and PE. At the same time, there are some shortcomings in this study. Due to missing information in the original data, it was not possible to adjust for some potentially important covariates, including maternal body mass index, maternal alcohol consumption, maternal stress, and noise exposure. Second, due to the limitations of the data obtained, there was no information on maternal mobility during pregnancy, and simply assigning exposure to a residential address to pregnant women did not allow for a careful assessment of maternal exposure at the workplace.

5. Conclusions

This study provides evidence for the relationship between air pollution and the prevalence of HDP. PM_2.5, PM₁₀, NO₂ and O₃ exposure during preconception and pregnancy are risk factors for HDP risk. Therefore, pregnant women should be concerned about air pollution not only during pregnancy, but also take protective measures during preconception. There are some differences in the effects of air pollution on GH and PE in terms of pollutant class, exposure window and effect size due to the differences between GH and PE. Further studies on the underlying pathophysiological mechanisms of GH and PE are needed in the future.

Declarations

Ethics approval and consent to participate

This study was based on a de-identified dataset and therefore did not require ethics approval and consent to participate.

Consent for publication

Not applicable.

Availability of data and materials

The datasets generated and/or analyzed during the current study are not publicly available due [REASON WHY DATA ARE NOT PUBLIC] but are available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Authors' contributions

Lei Cao: Methodology, Software, Writing - Original Draft. Ting Wang: Data curation, Software. Ruiping Diao: Validation, Visualization. Xuefeng Shi: Validation, Supervision, Writing - Review & Editing. Lu Cao and Zerui Gong: Investigation, Writing - Review & Editing. Hongjun Mao: Conceptualization, Supervision, Methodology.

Acknowledgements

Not applicable.

References

1. World Health Organization, 2018a. 9 out of 10 people worldwide breathe polluted air, but more countries are taking action. https://www.who.int/news/item/02-05-2018-9-out-of-10-people-worldwide-breathe-polluted-air-but-more-countries-are-taking-action.
2. Ministry of Ecology and Environment of the People's Republic of China, 2018b. 2017 China Ecological and Environmental Bulletin. https://www.mee.gov.cn/hjzl/sthjzk/zghjzkgb/.
3. Ministry of Ecology and Environment of the People's Republic of China, 2019. 2018 China Ecological and Environmental Bulletin. https://www.mee.gov.cn/hjzl/sthjzk/zghjzkgb/.
4. Ministry of Ecology and Environment of the People's Republic of China, 2020. 2019 China Ecological and Environmental Bulletin. https://www.mee.gov.cn/hjzl/sthjzk/zghjzkgb/.
5. Ministry of Ecology and Environment of the People's Republic of China, 2021. 2020 China Ecological and Environmental Bulletin. https://www.mee.gov.cn/hjzl/sthjzk/zghjzkgb/.
6. Aguilera AM, Escabias M, Valderrama MJ, 2006. Using principal components for estimating logistic regression with high-dimensional multicollinear data. Computational Statistics & Data Analysis. 50, 1905-1924.
7. Assibey-Mensah V, Glantz JC, Hopke PK, et al., 2019. Ambient wintertime particulate air pollution and hypertensive disorders of pregnancy in Monroe County, New York. Environmental Research. 168, 25-31.
8. Bala G-P, Rajnoveanu R-M, Tudorache E, et al., 2021. Air pollution exposure-the (in)visible risk factor for respiratory diseases. Environmental Science and Pollution Research. 28, 19615-19628.
9. Barker L, Brown C, 2001. Logistic regression when binary predictor variables are highly correlated. Stat Med. 20, 1431-42.
10. Brook RD, Rajagopalan S, Pope CA, III, et al., 2010. Particulate Matter Air Pollution and Cardiovascular Disease An Update to the Scientific Statement From the American Heart Association. Circulation. 121, 2331-2378.
11. Cao L, Wang L, Wu L, et al., 2021. Particulate matter and hypertensive disorders in pregnancy: systematic review and meta-analysis. Public Health. 200, 22-32.
12. Choe SA, Eliot MN, Savitz DA, et al., 2019. Ambient air pollution during pregnancy and risk of gestational diabetes in New York City. Environmental Research. 175, 414-420.
13. Choe SA, Kauderer S, Eliot MN, et al., 2018. Air pollution, land use, and complications of pregnancy. Science of the Total Environment. 645, 1057-1064.
14. Dall'Asta A, D'Antonio F, Saccone G, et al., 2021. Cardiovascular events following pregnancy complicated by pre-eclampsia with emphasis on comparison between early- and late-onset forms: systematic review and meta-analysis. Ultrasound in Obstetrics & Gynecology. 57, 698-709.
15. Dasari P, Habeebullah S, 2010. Maternal mortality due to hypertensive disorders of pregnancy in a tertiary care center in Southern India. International Journal of Gynecology & Obstetrics. 110, 271-273.
16. Escabias M, Aguilera AM, Valderrama MJ, 2007. Principal component estimation of functional logistic regression: discussion of two different approaches. Journal of Nonparametric Statistics. 16, 365-384.
17. Fleisch AF, Kloog I, Luttmann-Gibson H, et al., 2016. Air pollution exposure and gestational diabetes mellitus among pregnant women in Massachusetts: a cohort study. Environmental Health. 15, 40.
18. Gu X, Liu Q, Deng F, et al., 2019. Association between particulate matter air pollution and risk of depression and suicide: systematic review and meta-analysis. British Journal of Psychiatry. 215, 456-467.
19. Hu H, Ha S, Xu X, 2017. Ozone and hypertensive disorders of pregnancy in Florida: Identifying critical windows of exposure. Environmental Research. 153, 120-125.
20. Huang X, Qiu J, Zhang Y, et al., 2015. Ambient air pollutant PM10 and risk of pregnancy-induced hypertension in urban China. Environmental Research Letters. 10.
21. Jo H, Eckel SP, Chen JC, et al., 2019. Associations of gestational diabetes mellitus with residential air pollution exposure in a large Southern California pregnancy cohort. Environment International. 130, 104933.
22. Kelly FJ, Fussell JC, 2017. Role of oxidative stress in cardiovascular disease outcomes following exposure to ambient air pollution. Free Radical Biology and Medicine. 110, 345-367.
23. Lee PC, Roberts JM, Catov JM, et al., 2013. First trimester exposure to ambient air pollution, pregnancy complications and adverse birth outcomes in Allegheny County, PA. Maternal and Child Health Journal. 17, 545-55.
24. Lim CC, Thurston GD, 2019. Air Pollution, Oxidative Stress, and Diabetes: a Life Course Epidemiologic Perspective. Current Diabetes Reports. 19.
25. Madsen C, Haberg SE, Aamodt G, et al., 2018. Preeclampsia and Hypertension During Pregnancy in Areas with Relatively Low Levels of Traffic Air Pollution. Maternal and Child Health Journal. 22, 512-519.
26. Mandakh Y, Rittner R, Flanagan E, et al., 2020. Maternal Exposure to Ambient Air Pollution and Risk of Preeclampsia: A Population-Based Cohort Study in Scania, Sweden. International Journal of Environmental Research and Public Health. 17.
27. Melody SM, Wills K, Knibbs LD, et al., 2020. Maternal Exposure to Ambient Air Pollution and Pregnancy Complications in Victoria, Australia. International Journal of Environmental Research and Public Health. 17.
28. Mendola P, Wallace M, Liu D, et al., 2016. Air pollution exposure and preeclampsia among US women with and without asthma. Environmental Research. 148, 248-255.
29. Michikawa T, Morokuma S, Fukushima K, et al., 2015. A register-based study of the association between air pollutants and hypertensive disorders in pregnancy among the Japanese population. Environmental Research. 142, 644-50.
30. Michikawa T, Morokuma S, Yamazaki S, et al., 2021. Exposure to chemical components of fine particulate matter and ozone, and placenta-mediated pregnancy complications in Tokyo: a register-based study. Journal of Exposure Science and Environmental Epidemiology. 32, 135-145.
31. Mielke MM, Milic NM, Weissgerber TL, et al., 2016. Impaired Cognition and Brain Atrophy Decades After Hypertensive Pregnancy Disorders. Circulation-Cardiovascular Quality and Outcomes. 9, S70-S76.
32. Mobasher Z, Salam MT, Goodwin TM, et al., 2013. Associations between ambient air pollution and Hypertensive Disorders of Pregnancy. Environmental Research. 123, 9-16.
33. Monrad M, Sajadieh A, Christensen JS, et al., 2017. Long-Term Exposure to Traffic-Related Air Pollution and Risk of Incident Atrial Fibrillation: A Cohort Study. Environmental Health Perspectives. 125, 422-427.
34. Naderi S, Tsai SA, Khandelwal A, 2017. Hypertensive Disorders of Pregnancy. Current Atherosclerosis Reports. 19, 15.
35. Nazar W, Niedoszytko M, 2022. Air Pollution in Poland: A 2022 Narrative Review with Focus on Respiratory Diseases. International Journal of Environmental Research and Public Health. 19.
36. Nobles CJ, Williams A, Ouidir M, et al., 2019. Differential Effect of Ambient Air Pollution Exposure on Risk of Gestational Hypertension and Preeclampsia. Hypertension. 74, 384-390.
37. Padula AM, Yang W, Lurmann FW, et al., 2019. Prenatal exposure to air pollution, maternal diabetes and preterm birth. Environmental Research. 170, 160-167.
38. Pan SC, Huang CC, Lin SJ, et al., 2017. Gestational diabetes mellitus was related to ambient air pollutant nitric oxide during early gestation. Environ Res. 158, 318-323.
39. Pedersen M, Halldorsson TI, Olsen SF, et al., 2017. Impact of Road Traffic Pollution on Pre-eclampsia and Pregnancy-induced Hypertensive Disorders. Epidemiology. 28, 99-106.
40. Rana S, Lemoine E, Granger JP, et al., 2019. Preeclampsia: Pathophysiology, Challenges, and Perspectives. Circulation Research. 124, 1094-1112.
41. Roberts S, Arseneault L, Barratt B, et al., 2019. Exploration of NO2 and PM2.5 air pollution and mental health problems using high-resolution data in London-based children from a UK longitudinal cohort study. Psychiatry Research. 272, 8-17.
42. Rudra CB, Williams MA, Sheppard L, et al., 2011. Ambient carbon monoxide and fine particulate matter in relation to preeclampsia and preterm delivery in western Washington State. Environmental Health Perspectives. 119, 886-92.
43. Savitz DA, Elston B, Bobb JF, et al., 2015. Ambient Fine Particulate Matter, Nitrogen Dioxide, and Hypertensive Disorders of Pregnancy in New York City. Epidemiology. 26, 748-57.
44. Sun M, Yan W, Fang K, et al., 2020. The correlation between PM2.5 exposure and hypertensive disorders in pregnancy: A Meta-analysis. Science of the Total Environment. 703, 134985.
45. Theilen LH, Varner MW, Smith KR, et al., 2016. All-Cause and Cause-Specific Mortality After Hypertensive Disease of Pregnancy Reply. Obstetrics and Gynecology. 128, 1446-1446.
46. Turbeville HR, Sasser JM, 2020. Preeclampsia beyond pregnancy: long-term consequences for mother and child. American journal of physiology. Renal physiology. 318, F1315-F1326.
47. Umesawa M, Kobashi G, 2017. Epidemiology of hypertensive disorders in pregnancy: prevalence, risk factors, predictors and prognosis. Hypertension Research. 40, 213-220.
48. Un Nisa S, Shaikh AA, Kumar R, 2019. Maternal and Fetal Outcomes of Pregnancy-related Hypertensive Disorders in a Tertiary Care Hospital in Sukkur, Pakistan. Cureus. 11, e5507-e5507.
49. van den Hooven EH, de Kluizenaar Y, Pierik FH, et al., 2011. Air pollution, blood pressure, and the risk of hypertensive complications during pregnancy: the generation R study. Hypertension. 57, 406-12.
50. Vinikoor-Imler LC, Gray SC, Edwards SE, et al., 2012. The effects of exposure to particulate matter and neighbourhood deprivation on gestational hypertension. Paediatric and perinatal epidemiology. 26, 91-100.
51. Wang IK, Muo C-H, Chang Y-C, et al., 2013. Association between hypertensive disorders during pregnancy and end-stage renal disease: a population-based study. Canadian Medical Association Journal. 185, 207-213.
52. Wang Q, Zhang H, Liang Q, et al., 2018. Effects of prenatal exposure to air pollution on preeclampsia in Shenzhen, China. Environmental Pollution. 237, 18-27.
53. Wu J, Ren C, Delfino RJ, et al., 2009. Association between local traffic-generated air pollution and preeclampsia and preterm delivery in the south coast air basin of California. Environmental Health Perspectives. 117, 1773-9.
54. Xu X, Hu H, Ha S, et al., 2014. Ambient air pollution and hypertensive disorder of pregnancy. Journal of Epidemiology and Community Health. 68, 13-20.
55. Yang S, Ma YL, Duan FK, et al., 2018. Characteristics and formation of typical winter haze in Handan, one of the most polluted cities in China. Sci Total Environ. 613-614, 1367-1375.
56. Yazdi MD, Wang Y, Di Q, et al., 2019. Long-term exposure to PM2.5 and ozone and hospital admissions of Medicare participants in the Southeast USA. Environment International. 130.
57. Yorifuji T, Naruse H, Kashima S, et al., 2015. Residential proximity to major roads and obstetrical complications. Science of the Total Environment. 508, 188-92.
58. Zhai D, Guo Y, Smith G, et al., 2012. Maternal exposure to moderate ambient carbon monoxide is associated with decreased risk of preeclampsia. American Journal of Obstetrics and Gynecology. 207, 57 e1-9.
59. Zhu Y, Zhang C, Liu D, et al., 2017. Ambient Air Pollution and Risk of Gestational Hypertension. American Journal of Epidemiology. 186, 334-343.