DOI: https://doi.org/10.21203/rs.3.rs-2270323/v1
Background
Ambient particulate matter (PM) is a trigger factor for metabolic dysfunction. This study aimed to evaluate the associations between PM exposure and metabolic dysfunction in pregnancy.
Methods
This prospective, multicentre, observational, cohort study was conducted from January 2021 to June 2022. A total of 333 women with singleton pregnancies were recruited. For individual measurement of PM2.5 levels, an AirguardK® was installed inside the participants’ houses. Time-activity logs were recorded to predict each participant’s personal exposure to PM2.5. The participants were divided into groups according to the concentration of PM2.5 calculated by a time-weighted average model. We used the Student t-test and chi-squared test (or Fisher’s exact test) to evaluate changes in metabolic compound levels, blood pressure (BP), glucose intolerance, and lipid profiles, including serum triglyceride/high-density lipoprotein cholesterol (TG/HDL-C) ratios. Logistic and linear regression models were used to analyse the association between PM2.5 exposure and metabolic dysfunction, using odds ratios (ORs) and 95% confidence intervals (CIs).
Results
PM2.5 exposure during pregnancy worsened metabolic dysfunction. Third trimester BP was elevated in those participants exposed to higher levels of PM2.5 (p <0.05). The incidence of gestational diabetes mellitus (GDM) was constantly higher in those exposed to more PM2.5, regardless of the PM2.5 cut-off level (PM2.5 ≥10 µg/m3, 7.91 % vs. 16.09 %, p <0.05; PM2.5 ≥25 µg/m3, 7.91 % vs. 26.67 %, p <0.05). The proportion with a TG/HDL-C ratio ≥3.0 was significantly higher when PM2.5 was ≥10 µg/m3 (75.3 % vs. 83.0 %, p <0.05). Triglyceride levels were significantly higher in the PM2.5 ≥25 µg/m3 group (p=0.0171). We found an increased risk of elevated BP (adjusted OR [aOR]: 2.228, 95% CI: 1.115–4.449) and GDM (aOR 2.263, 95% CI 1.106–5.039) in the third trimester after adjusting for confounders.
Conclusion
Exposure to PM2.5 worsens metabolic dysfunction in pregnancy. Further studies are required to investigate the mechanisms by which ambient PM affects metabolic dysfunction in pregnancy.
Particulate matter 2.5 (PM2.5) is composed of aerodynamic particles with diameters ≤ 2.5 µm (1). Ambient particulate matter shows an increasing trend worldwide and has harmful effects on human health. Furthermore, the World Health Organization (WHO) listed air pollution and climate change among the top 10 global health hazards in 2019 (2).
Particulate matter can cause various health problems, and the foetus is more vulnerable to air pollution (3, 4). Previous studies showed that exposure to particulate matter is associated with adverse pregnancy outcomes and metabolic syndrome (5–8). However, the degree of harmful exposure to particulate matter is still inconclusive (9). Furthermore, animal studies suggest that particulate matter induces oxidative stress and systemic inflammation, which can cause metabolic dysfunction, including elevated blood pressure and altered lipid and glucose metabolism (10). Similar results were found in studies on general populations (11, 12).
Metabolic syndrome is a cluster of central obesity, glucose intolerance, hypertriglyceridemia, and low levels of high-density lipoprotein cholesterol (HDL-C). It can impact the national health status, causing complications that could progress to chronic diseases, such as cerebrovascular disease, diabetes mellitus, and cancer; thus, it requires comprehensive management. Additionally, the prevalence of metabolic syndrome in pregnancy is rising concomitantly with advanced maternal age, which has recently seen an increase. Previous studies have suggested that metabolic syndrome may be influenced by particulate matter.
To our knowledge, the association between particulate matter concentration and metabolic dysfunction in pregnancy has not been fully investigated. To explore the effect of particulate matter, an accurate individual measurement of particulate matter exposure is required. Currently, it is difficult to measure individual exposures to particulate matter because of limitations in research techniques, and most studies examining particulate matter are retrospective. Moreover, previous studies that investigated the association between particulate matter and pregnancy only examined particulate matter density outside the house or had limited evaluations of particulate matter inside the house (12–14).
Importantly, the source of particulate matter emission is not only from outside but also from inside the house. Modern people spend most of their time indoors, particularly in today's pandemic era. Therefore, particulate matter density inside the house should be considered when analysing its effects. Further, measurement of individual exposure to particulate matter, using a time-activity pattern, can provide a more accurate analysis of causal relationships with metabolic dysfunction in pregnancy. The objective of this study was to investigate the effect of exposure to PM2.5 on metabolic dysfunction in pregnant women via a personalized measurement method.
This prospective, multicentre, observational, cohort study was conducted from January 2021 to June 2022. We recruited pregnant women from the outpatient clinics of all participating institutions. We enrolled mothers aged ≥ 19 years and less than 28 weeks of gestation. We excluded patients with multiple pregnancies, chronic medical illness, other gynaecologic and obstetrical disease, and a history of cancer and smoking.
At each trimester, blood and urine samples were collected from the pregnant women. Pathological examination of the placenta and analysis of umbilical cord blood was performed at delivery. Recruitment of participants and sample collection processes are illustrated in Fig. 1.
AirguardK® (Kweather Co, Korea) is a small device that uses electricity and is equipped with a sensor that can detect air pollution levels based on a light scattering laser photometer technique (Fig. 2). To measure the air quality inside the house, the AirguardK® was placed for a minimum of one week per trimester. The air quality data, i.e. PM2.5, PM10, CO2, temperature, humidity, and indoor integrated index, obtained from the device, were collected through the internet with a 1-minute interval. The measured indoor air quality data were transmitted to the indoor air quality monitoring platform through the Long-Term Evolution (LTE) communication network to prevent memory loss, and data were collected and stored every minute. Measurement was carried out for a week every trimester, and the measured concentration value was obtained in real time using the Internet of Things and Information and Communication Technology.
Data on outdoor fine particle concentrations, i.e. PM10 and PM2.5, were collected from a nearby urban atmospheric measurement network, based on the addresses of the recruited women. The Urban Air Monitoring Station data used in this study were obtained from AirKorea (https://www.airkorea.or.kr/web) by The Korean Ministry of Environment. AirKorea has 614 stations in 162 cities and counties for measurement of air pollutants, including sulphur dioxide (SO2), nitrogen dioxide (NO2), ozone (O3), carbon monoxide (CO), PM10, and PM2.5, per annum in Korea.
To ascertain the degree of harmfulness of the particulate matter, human exposure, time-activity pattern, and indoor pollution were evaluated. A time-activity questionnaire for one week’s exposure was filled by the participants, and outdoor exposure was measured using the data from AirKorea and an application used to map/determine the time-activity pattern. The participants responded to a questionnaire on their socio-economic status and lifestyle, including dietary habits, activity grade, and house type. We applied a time-weighted average model to estimate the personal PM exposure of participants. The time-weighted average model is a concept that gives weight according to the activity area expressed by the following Eq. (15–17).
𝐶𝑝𝑒𝑟 = ∑( 𝑇𝑖𝑛𝑑𝑜𝑜𝑟 × 𝐶𝑖𝑛𝑑𝑜𝑜𝑟 + 𝑇𝑜𝑢𝑡𝑑𝑜𝑜𝑟 × 𝐶𝑜𝑢𝑡𝑑𝑜𝑜𝑟)
Where, 𝐶: predicted personal exposure
𝑇𝑖𝑛𝑑𝑜𝑜𝑟: Indoor activity time rate (indoor activity time/24 hr)
𝐶𝑖𝑛𝑑𝑜𝑜𝑟: indoor PM2.5 concentration
𝑇𝑜𝑢𝑡𝑑𝑜𝑜𝑟: outdoor activity time rate (outdoor activity time/24 hr)
𝐶𝑜𝑢𝑡𝑑𝑜𝑜𝑟: outdoor PM2.5 concentration
Based on the WHO’s annual average particulate matter concentration of 10 µm/m3, we divided the participants into the following groups: high-PM (PM2.5 ≥ 10 µg/m3), low-PM (PM2.5 < 10 µg/m3), and extremely high-PM (exthigh-PM) ( PM2.5 ≥ 25 µg/m3), based on the WHO’s daily average particulate matter concentration (18).
According to the metabolic syndrome criteria outlined by the International Diabetes Federation, metabolic syndrome is defined as a waist circumference ≥ 35 inches for women, in addition to abnormal metabolic parameters (elevated blood pressure, altered lipid metabolism, and glucose intolerance). Since pregnancy induces a pro-metabolic state, it was difficult to apply the original diagnostic criteria of the metabolic syndrome. Furthermore, the fasting state at the time of blood sampling and the registered trimester were inconsistent in this study. Thus, we replaced the evaluation of waist circumference with body mass index (BMI) before pregnancy; fasting glucose levels with glucose intolerance, including a diagnosis of gestational diabetes mellitus (GDM); and the lipid profile criteria were changed to TG ≥ 175 mg/dl and HDL < 40 mg/dl, based on previous research (19, 20).
To diagnose GDM, two approaches were used: the one-step approach, where a 75 g oral glucose tolerance test (OGTT) was performed; and the two-step approach, where a 50 g OGTT screening test plus a 100 g OGTT confirmation test were performed. All pregnant women underwent the GDM screening tests between 24 and 28 weeks of gestation. GDM was diagnosed when thresholds were met or exceeded (21). We defined abnormal diagnostic findings as glucose intolerance.
Lipid profiles were evaluated in the third trimester, reflecting PM2.5 concentrations throughout the pregnancy. Lipid profiles, including LDL cholesterol, TGs, and total cholesterol were measured through maternal blood samples, and HDL cholesterol was calculated using the traditional Friedewald Eq. (22). It is known that elevated TGs are a risk factor for insulin resistance while an increase in HDL is a protective factor, particularly since a TG/HDL-C ratio ≥ 3.0 mg/dl has a higher association with insulin resistance than TG and HDL alone. Thus, we evaluated the TG/HDL-C ratio as a marker of metabolic dysfunction in this study (23–25)
The following variables were treated as covariates based on previous studies: maternal age, pre-pregnancy BMI, and physical activity (8, 26).
The concentrations of PM2.5 were not normally distributed; thus, the data are presented as medians and interquartile ranges. Differences in continuous and categorical variables were analysed using the independent two sample t-test, ANOVA, and the chi-squared test (or Fisher’s exact test), as appropriate. Logistic and linear regression models were used to analyse the associations between exposure to PM and metabolic dysfunction using odds ratios (ORs) and 95% confidence intervals (CIs). Statistical analysis was conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA). A p-value under 0.05 was considered statistically significant.
Baseline characteristics of the study population are shown in Table 1. The mean age of the pregnant women in this study was 33.59 ± 4.09 years, and the majority were married (99.01%), had completed university (72.76%), and underwent natural pregnancy (86.2%). The proportion of pre-pregnancy BMI ≥ 25 kg/m2 was 15.3%, and the mean weight gain during pregnancy was 12.61 ± 5.01 kg. Furthermore, there was no difference in the percentages between income quantiles. The proportion of women who were exposed to grilling and frying was 70.53%, which were known cooking methods that emit a considerable amount of particulate matter (27). In general, pregnancy outcomes were not significantly different between the PM-stratified groups (Supplemental Table 1).
Characteristics |
All cohorts |
Low-PM (n = 215) |
High-PM (n = 87) |
p-value |
---|---|---|---|---|
Age (years) |
33.59 ± 4.09 |
33.26 ± 3.78 |
33.94 ± 4.62 |
0.2241 |
Pre-pregnancy BMI (kg/m2) |
22.22 ± 3.42 |
22.26 ± 3.56 |
21.96 ± 2.95 |
0.4480 |
Marital status (married) |
299 (99.01) |
215 (100.00) |
84 (96.55) |
0.0233 |
Educational status (University graduate) |
245 (91.36) |
199 (92.56) |
76 (88.37) |
0.4263 |
Occupation (yes) |
223 (67.20) |
112 (67.07) |
94 (69.12) |
0.7033 |
Income level (per month, Won) |
0.9480 |
|||
<400 (x104) |
77 (36.50) |
35 (36.08) |
35 (27.63) |
|
400 ~ 600 (x104) |
57 (27.00) |
27 (27.84) |
24 (25.81) |
|
≥600 (x104) |
77 (36.50) |
35 (36.08) |
34 (36.56) |
|
Gravidity (primigravida) |
148 (49.01) |
111 (51.63) |
37 (42.53) |
0.6128 |
Pregnancy route |
0.3537 |
|||
Natural |
257 (85.38) |
186 (86.51) |
71 (82.56) |
|
IUI |
3 (1.00) |
3 (1.40) |
0 (0.00) |
|
IVF-ET |
41 (13.62) |
26 (12.09) |
15 (17.44) |
|
Antioxidant medication |
138 (46.62) |
97 (45.97) |
41 (48.24) |
0.7239 |
Environmental factors |
||||
Cooking method (i.e., frying and grilling) |
213 (70.53) |
151 (70.23) |
62 (71.26) |
0.8586 |
Air cleaner (yes) |
246 (81.46) |
178 (82.79) |
68 (78.16) |
0.3485 |
Abbreviations: BMI (body mass index), IUI (intrauterine insemination), IVF-ET (in vitro fertilization-embryo transfer) |
The mean PM2.5 concentration during pregnancy was 9.34 ± 7.55 µg/m3. Table 2 shows the PM concentrations during the study. Based on the WHO’s annual average particulate matter concentration of 10 µg/m3, 24.6% of the overall study population was exposed to high PM2.5 levels. Moreover, 5.6%, 32.8%, and 28.0% of the study population were exposed to high PM2.5 in the first, second, and third trimesters, respectively.
PM 2.5 Concentration (µg/m3) |
Mean |
SD |
1st quartile |
Median |
3rd quartile |
Maximum |
---|---|---|---|---|---|---|
Entire pregnancy |
9.34 |
7.55 |
4.48 |
7.33 |
11.04 |
56.24 |
1st trimester |
6.52 |
4.14 |
3.56 |
5.93 |
8.56 |
26.24 |
2nd trimester |
47.45 |
8.51 |
3.67 |
6.13 |
9.87 |
47.45 |
3rd trimester |
81.58 |
12.17 |
4.45 |
8.21 |
16.5 |
81.58 |
Abbreviations: SD, standard deviation; PM, particulate matter |
The relationships between the particulate matter levels and metabolic parameters are presented in Tables 3 − 1 and 3 − 2. A PM2.5 concentration of 10 µg/m3 showed no significant difference in BP in the 2nd and 3rd trimesters between the high-PM (PM2.5 ≥10 µg/m3) and low-PM (PM2.5 <10 µg/m3) groups; however, the PM2.5 concentration appeared to influence glucose metabolism. Oral glucose test results in the high-PM group were higher than those in the low-PM group. Furthermore, GDM consistently showed significant results regardless of the cut-off value of PM2.5 concentration (PM2.5 ≥10 µg/m3, 7.91% vs. 16.09%, p < 0.05; PM2.5 ≥25 µg/m3, 7.91% vs. 26.67%, p < 0.05). PM2.5 concentration did not affect gestational weight gain. When metabolic syndrome was analysed, as previously defined (pre-pregnancy BMI ≥ 25 kg/m2 with two or more of the following: SBP > 135 mmHg or DBP > 85 mmHg in the third trimester, TG > 175 mg/dL, HDL-C < 40 mg/dL, or glucose intolerance), there was no difference according to particulate matter (high-PM, 69.23% [9/13] vs. 69.70% [23/33], p > 0.9999; Exthigh-PM, 100.00% [2/2] vs. 69.70% [23/33], p > 0.9999). However, BP, TG, and GDM showed significant differences between the high-PM and low-PM groups, while the proportion of HDL < 40 mg/dl showed no statistically significant difference. We observed that TG/HDL and the TG/HDL ratios over 3.0 were higher in the high-PM group than in the low-PM group; this difference was not statistically significant. (3.37 ± 37.46 vs. 6.43 ± 5.70, p = 0.0576; 75.3% vs. 83.0%, p = 0.2960, respectively). Subgroup analysis between PM2.5 concentrations under 10 and more than 25 µg/m3 is shown in Supplemental Table 2.
Metabolic parameters |
Low-PM (n = 215) |
High-PM (n = 87) |
p-value |
---|---|---|---|
2nd trimester |
|||
sBP (mmHg) |
111.76 ± 10.85 |
112.20 ± 10.27 |
0.7466 |
dBP (mmHg) |
65.37 ± 9.35 |
64.58 ± 8.79 |
0.5039 |
HbA1c (%) |
5.03 ± 0.35 |
5.03 ± 0.37 |
0.9642 |
50 g-OGTT (mg/dL) |
|||
1-h glucose |
121.34 ± 24.73 |
127.62 ± 31.36 |
0.1418 |
75 g-OGTT (mg/dL) |
|||
Fasting glucose |
83.71 ± 7.29 |
82.17 ± 8.77 |
0.4420 |
1-h glucose |
134.02 ± 32.77 |
139.67 ± 40.42 |
0.5345 |
2-h glucose |
123.52 ± 25.34 |
123.17 ± 35.38 |
0.9619 |
100 g-OGTT (mg/dL) |
|||
Fasting glucose |
83.00 ± 5.66 |
90.33 ± 6.88 |
0.3011 |
1-h glucose |
147.89 ± 33.98 |
172.70 ± 30.48 |
0.0496* |
2-h glucose |
136.93 ± 26.73 |
146.40 ± 19.49 |
0.3128 |
3-h glucose |
115.18 ± 24.88 |
136.20 ± 21.44 |
0.0232* |
3rd trimester |
|||
sBP (mmHg) |
115.22 ± 11.54 |
117.07 ± 11.98 |
0.2170 |
dBP (mmHg) |
70.54 ± 10.25 |
71.77 ± 9.86 |
0.3449 |
TG (mg/dL) |
315.73 ± 153.90 |
316.35 ± 129.46 |
0.9754 |
TC (mg/dL) |
275.56 ± 54.25 |
273.47 ± 49.69 |
0.7686 |
LDL-c (mg/dL) |
151.83 ± 42.66 |
145.49 ± 44.88 |
0.4010 |
HDL-c (mg/dL) |
64.43 ± 51.47 |
60.01 ± 16.91 |
0.4466 |
TG/HDL-C ratio |
3.37 ± 37.46 |
6.43 ± 5.70 |
0.4343 |
ΔWeight gain† (kg) |
12.89 ± 5.02 |
12.18 ± 5.89 |
0.2865 |
Metabolic syndrome |
|||
Elevated BP‡ |
25 (11.85) |
18 (20.93) |
0.0437* |
TG > 175 (mg/dL) |
169 (92.35) |
77 (98.72) |
0.0446* |
HDL-c < 40 (mg/dL) |
78 (80.41) |
43 (91.49) |
0.0889 |
TG/HDL-C ratio ≥ 3.0 |
73 (75.26) |
39 (82.98) |
0.2960 |
GDM |
17 (7.91) |
14 (16.09) |
0.0338* |
Glucose intolerance§ |
52 (24.19) |
27 (31.03) |
0.2201 |
* p-value < 0.05 | |||
†Body weight at delivery minus pre-pregnancy body weight | |||
‡SBP > 135 mmHg or DBP > 85 mmHg in the third trimester | |||
§Any abnormal glucose values on OGTT testing | |||
Abbreviations: sBP, systolic blood pressure; dBP, diastolic blood pressure; HbA1c, haemoglobin A1c; hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total cholesterol, LDL-c, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; GDM, gestational diabetes mellitus. | |||
Data are presented as Mean ± SD or N (%) |
Metabolic parameters |
PM < 25 µg/m3 (n = 215) |
ExtHigh-PM (n = 15) |
p-value |
---|---|---|---|
2nd trimester |
|||
sBP (mmHg) |
111.76 ± 10.85 |
114.00 ± 10.58 |
0.4390 |
dBP (mmHg) |
65.37 ± 9.35 |
64.87 ± 10.56 |
0.8420 |
HbA1c (%) |
5.03 ± 0.35 |
5.30 ± 0.52 |
0.0990 |
50 g-OGTT (mg/dL) |
|||
1-h glucose |
121.34 ± 24.73 |
140.17 ± 27.81 |
0.0123* |
75 g-OGTT (mg/dL) |
|||
Fasting glucose |
83.71 ± 7.29 |
84.00 ± 6.56 |
0.9459 |
1-h glucose |
134.02 ± 32.77 |
153.67 ± 50.90 |
0.3350 |
2-h glucose |
123.52 ± 25.34 |
127.33 ± 34.95 |
0.8059 |
100 g-OGTT (mg/dL) |
|||
Fasting glucose |
83.00 ± 5.66 |
85.00±. |
|
1-h glucose |
147.90 ± 33.98 |
173.67 ± 9.50 |
0.2072 |
2-h glucose |
136.93 ± 26.73 |
153.33 ± 15.89 |
0.3098 |
3-h glucose |
115.18 ± 24.88 |
136.67 ± 20.53 |
0.1612 |
3rd trimester |
|||
sBP (mmHg) |
115.22 ± 11.54 |
118.13 ± 12.05 |
0.3476 |
dBP (mmHg) |
70.54 ± 10.25 |
75.33 ± 7.71 |
0.0774 |
TG (mg/dL) |
315.73 ± 153.90 |
404.85 ± 183.90 |
0.0479* |
TC (mg/dL) |
275.56 ± 54.25 |
267.20 ± 46.26 |
0.5635 |
LDL-c (mg/dL) |
151.83 ± 42.66 |
132.00 ± 30.90 |
0.2647 |
HDL-c (mg/dL) |
64.43 ± 51.47 |
50.03 ± 20.15 |
0.1713 |
TG/HDL-C ratio |
3.37 ± 37.46 |
12.12 ± 8.10 |
0.0950 |
ΔWeight gain† (kg) |
12.89 ± 5.02 |
11.92 ± 7.45 |
0.6262 |
Metabolic syndrome |
|||
Elevated BP‡ |
25 (11.85) |
3 (20.00) |
0.4077 |
TG > 175 (mg/dL) |
169 (92.35) |
13 (100.00) |
0.6044 |
HDL-c < 40 (mg/dL) |
78 (80.41) |
5 (83.33) |
1.0000 |
TG/HDL-C ratio ≥ 3.0 |
73 (75.26) |
6 (100.00) |
0.3317 |
GDM |
17 (7.91) |
4 (26.67) |
0.0362* |
Glucose intolerance§ |
52 (24.19) |
7 (46.67) |
0.0676 |
* p-value < 0.05 | |||
†Body weight at delivery minus pre-pregnancy body weight | |||
‡SBP > 135 mmHg or DBP > 85 mmHg in the third trimester | |||
§Any abnormal glucose value on OGTT testing | |||
Abbreviations: sBP, systolic blood pressure; dBP, diastolic blood pressure; HbA1c, haemoglobin A1c; hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total cholesterol, LDL-c, low-density lipoprotein cholesterol; HDL-c, high-density lipoprotein cholesterol; GDM, gestational diabetes mellitus. | |||
Data are presented as Mean ± SD or N (%) |
Table 4 presents the extent to which particulate matter exposure in each trimester affected metabolic dysfunction. A high concentration of PM2.5 in the first trimester appeared to be associated with an elevated BP in the third trimester. Furthermore, the more exposed the pregnant woman is to particulate matter in the second trimester, the greater the impact on lipid metabolism, particularly TGs. Metabolic syndrome was not affected by trimester variance (1st trimester, p > 0.9999; 2nd trimester, p = 0.8695).
Metabolic parameters |
1st trimester |
2nd trimester |
||||||||
---|---|---|---|---|---|---|---|---|---|---|
Low-PM |
High-PM |
p-value |
Low-PM |
High-PM |
p-value |
|||||
2nd trimester |
||||||||||
sBP (mmHg) |
112.23 ± 10.83 |
112.67 ± 7.58 |
0.8695 |
112.44 ± 10.80 |
111.72 ± 10.25 |
0.5680 |
||||
dBP (mmHg) |
65.77 ± 9.07 |
63.22 ± 6.52 |
0.2566 |
65.48 ± 9.57 |
64.86 ± 8.37 |
0.5710 |
||||
HbA1c (%) |
4.99 ± 0.27 |
5.17 ± 0.39 |
0.0439* |
5.06 ± 0.36 |
4.98 ± 0.34 |
0.1122 |
||||
50 g-OGTT (mg/dL) |
||||||||||
1-h glucose |
119.98 ± 24.56 |
118.81 ± 28.05 |
0.8633 |
122.02 ± 24.59 |
126.06 ± 30.45 |
0.2910 |
||||
75 g-OGTT (mg/dL) |
||||||||||
Fasting glucose |
86.11 ± 7.75 |
81.67 ± 3.22 |
0.3687 |
83.90 ± 7.38 |
81.76 ± 8.45 |
0.2637 |
||||
1-h glucose |
134.56 ± 34.31 |
138.00 ± 15.72 |
0.8729 |
137.78 ± 33.57 |
133.62 ± 37.86 |
0.6298 |
||||
2-h glucose |
127.78 ± 24.65 |
97.33 ± 30.37 |
0.1083 |
124.05 ± 26.03 |
124.90 ± 32.94 |
0.9047 |
||||
100 g-OGTT (mg/dL) |
||||||||||
Fasting glucose |
88.00 ± NS |
98.00 ± NS |
NS |
79.00 ± NS |
95.20 ± 14.86 |
NS |
||||
1-h glucose |
148.08 ± 27.10 |
182.00 ± 24.04 |
0.1241 |
150.71 ± 35.49 |
168.13 ± 32.48 |
0.1321 |
||||
2-h glucose |
146.33 ± 23.26 |
144.00 ± 9.90 |
0.8940 |
139.83 ± 27.27 |
144.87 ± 36.33 |
0.6248 |
||||
3-h glucose |
123.67 ± 19.97 |
133.00 ± 48.08 |
0.6144 |
119.00 ± 24.81 |
126.20 ± 29.57 |
0.4180 |
||||
3rd trimester |
||||||||||
sBP (mmHg) |
113.57 ± 10.97 |
120.00 ± 12.04 |
0.0253* |
115.44 ± 11.59 |
116.18 ± 10.89 |
0.5871 |
||||
dBP (mmHg) |
69.33 ± 9.59 |
74.22 ± 8.70 |
0.0452* |
70.42 ± 10.72 |
71.64 ± 9.07 |
0.3124 |
||||
TG (mg/dL) |
298.84 ± 135.56 |
280.06 ± 94.03 |
0.5860 |
313.90 ± 156.00 |
324.83 ± 142.03 |
0.5718 |
||||
TC (mg/dL) |
274.57 ± 57.23 |
266.06 ± 46.54 |
0.5549 |
275.25 ± 53.01 |
274.58 ± 47.23 |
0.9186 |
||||
LDL-c (mg/dL) |
151.10 ± 40.38 |
146.67 ± 55.58 |
0.7380 |
151.36 ± 42.96 |
149.04 ± 44.42 |
0.7508 |
||||
HDL-c (mg/dL) |
63.63 ± 46.55 |
57.80 ± 12.52 |
0.3873 |
62.77 ± 46.67 |
58.76 ± 22.04 |
0.4873 |
||||
TG/HDL-C ratio |
7.45 ± 12.22 |
5.51 ± 2.73 |
0.6610 |
2.99 ± 37.41 |
8.30 ± 10.50 |
0.3369 |
||||
ΔWeight gain† (kg) |
12.84 ± 4.60 |
12.57 ± 5.58 |
0.8128 |
12.52 ± 5.05 |
12.78 ± 5.74 |
0.6818 |
||||
Metabolic syndrome |
||||||||||
Elevated BP‡ |
10 (9.52) |
4 (22.22) |
0.1244 |
24 (12.70) |
18 (15.93) |
0.4323 |
||||
TG > 175 (mg/dL) |
85 (92.39) |
16 (94.12) |
1.0000 |
153 (92.17) |
95 (97.94) |
0.0516 |
||||
HDL-c < 40 (mg/dL) |
58 (85.29) |
11 (91.67) |
1.0000 |
78 (82.11) |
41 (85.42) |
0.6168 |
||||
TG/HDL-C ratio ≥ 3.0 |
52 (76.47) |
10 (83.33) |
0.7352 |
76 (80.00) |
37 (77.08) |
0.6858 |
||||
GDM |
6 (5.61) |
3 (15.00) |
0.1503 |
16 (8.38) |
15 (13.04) |
0.1901 |
||||
Glucose intolerance§ |
21 (19.63) |
5 (25.00) |
0.5573 |
49 (25.65) |
33 (28.45) |
0.5917 |
||||
* p-value < 0.05 | ||||||||||
†Body weight at delivery minus pre-pregnancy body weight | ||||||||||
‡SBP > 135 mmHg or DBP > 85 mmHg in the third trimester | ||||||||||
§Any abnormal glucose value on OGTT testing | ||||||||||
Abbreviations: sBP, systolic blood pressure; dBP, diastolic blood pressure; HbA1c, haemoglobin A1c; hs-CRP, high-sensitivity C-reactive protein; TG, triglycerides; TC, total cholesterol, LDL-c, low-density lipoprotein cholesterol; HDL-c, high-density lipoprotein cholesterol; GDM, gestational diabetes mellitus. | ||||||||||
Data are presented as Mean ± SD or N (%) |
The risk of metabolic dysfunction was significantly higher in the high-PM group compared with the low-PM group, particularly BP in the third trimester and GDM after adjusting for maternal age and pre-pregnancy BMI (adjusted OR [aOR] [95% CI], 2.23 [1.12–4.45]; 2.26 [1.02–5.04], p < 0.05 for all). Effects on lipid profiles were marginally increased according to PM2.5 concentrations (aOR [95% CI], 6.38 [0.82–49.36], p = 0.0760; 2.62 [0.84–8.19], p = 0.0981). No significant associations were observed between PM2.5 exposure and HDL-C levels in pregnancy. We observed that PM2.5 exposure generally increased the risk of altered lipid metabolism even though statistical significance was scarce (Fig. 3).
This study aimed to investigate the association between PM2.5 exposure during pregnancy and metabolic dysfunction through personalized measurement of pollutant concentrations. The sources of particulate matter are diverse, and include industrial emissions, vehicles, and wildfires, with 25% coming from burning fuel. However, most activities were carried out indoors because of decreased mobility owing to pregnancy or the increased popularity of working from home; thus, the indoor sources of particulate matter should be considered when analysing the effect of particulate matter in pregnancy (2, 9).
This study found a positive association between PM2.5 exposure during pregnancy and metabolic dysfunction. Exposure to high PM2.5 altered baseline blood pressure, lipid metabolism, and glucose homeostasis. Pregnant women exposed to high PM2.5 concentrations showed elevated blood pressures in their third trimester, increased TGs, and had higher odds of developing GDM. Moreover, this study revealed that the greater the exposure to PM2.5, the higher the TG/HDL-C ratio was, which represents cardiovascular risk more effectively than does the lipid profile itself.
The results of this study corresponded well with those of earlier studies. A higher level of PM2.5 was associated with metabolic syndrome, hypertriglyceridemia, and high fasting blood glucose (8) (1, 28). Recent studies also showed the association between exposure to PM2.5 and GDM (5, 26). Researchers demonstrated that long-term exposure is a risk factor for dyslipidaemia and that a high TG/HDL ratio was associated with GDM (23, 29)
In contrast, the results of some studies were inconsistent with our findings. Exposure to PM2.5 during the pre-conception period and the 1st trimester did not increase the risk of GDM. Moreover, there was no association between PM2.5 and glucose intolerance, but HbA1c was affected by air pollution (4, 30–32). We suggest that the lack of prospective studies may be the reason for these discrepancies.
Although the exact biological mechanism of the effect of particulate matter on metabolic dysfunction has not yet been established, a hypothesis of oxidative stress is emerging. There are various sources of oxidative stress in pregnancy, including exposure to metals such as cadmium, mercury, methylmercury, lead, and chromium; tobacco; airborne particulate matter; and plastics (33). Previous studies found that inhalation of particulate matter affects oxidative stress generation and causes systemic inflammation, vascular dysfunction, atherosclerosis, and cardiovascular disease in non-pregnant populations (33, 34). Oxidative stress induces inflammation and causes metabolic syndrome beyond insulin resistance (35–37). Furthermore, studies have shown that reactive oxygen species and oxygen-centred free radicals induced by particulate matter are involved in the impairment of the insulin signalling pathway (35, 36, 38–42). In addition, oxidative stress induced by particulate matter exposure was associated with various obstetric complications, such as recurrent pregnancy loss, prematurity, intrauterine growth restriction (IUGR), diabetes, and preeclampsia (43–46). However, there were no significant differences between inflammatory markers in this study (Supplemental Table 3). Some studies attempted to prevent inflammation using antioxidants that contain flavonoids, arginine, vitamin C, vitamin E, carotenoids, resveratrol, and selenium. These studies have shown that antioxidants have a positive effect in preventing inflammation and metabolic syndrome (47–49).
The relationship between exposure to particulate matter and the risk of developing metabolic syndrome was consistently observed in the obese population (6, 50, 51). Moreover, obese patients also tended to have higher levels of inflammatory markers, including C-reactive protein (CRP). This pro-inflammation tendency is thought to increase the risk of various obstetric complications, such as GDM, preeclampsia, large for gestational age, and IUGR (52, 53). In this study, BMI > 25 kg/m2 before pregnancy was associated with a remarkably high CRP level; however, the effect of particulate matter in obese patients was not significant (Supplemental Table 4).
The importance of the in-utero environment of the foetus is emerging in aspects of foetal programming (54–56). The researchers suggested that external stimuli experienced by the foetus in the uterus can affect future health. In a retrospective study, particulate matter appeared to cause future foetal neurodevelopmental delay, which can be reasoned from an inflammatory perspective; however, this aspect is not fully understood (57, 58). This study showed us that exposure to particulate matter during pregnancy could influence maternal metabolic dysfunction. According to the foetal programming theory, particulate matter exposure could influence the intrauterine environment, which has an important role in future foetal health, beyond maternal well-being.
Our study had several strengths. First, this was the first prospective study examining the effects of particulate matter on pregnant women in South Korea. Second, this study focused on the individual effects of particulate matter by measuring personalized particulate matter concentrations. However, this study also had several limitations. First, it was limited by its small sample size. However, there is potential to increase the number of participants because this study is ongoing. Second, the study was performed during the 2019 coronavirus pandemic; hence, the outdoor concentration of PM2.5 might be poorly represented owing to reduced outdoor activities. Third, the PM2.5 data in the pre-conception period was absent; thus, this study could not represent the effect of PM2.5 before pregnancy. Last, the effect of climate factors, such as temperature, humidity, and seasonal variations, which could be a risk factor for adverse pregnancy outcomes, were not considered in this paper (59–61).
There was a positive relationship between PM2.5 and metabolic dysfunction in pregnancy. Exposure to PM2.5 during pregnancy was associated with higher rates of metabolic dysfunction. This study highlights the importance of managing PM2.5 concentrations that could alter metabolic parameters, followed by adverse pregnancy outcomes. However, further research is required to investigate particulate matter effects on metabolic dysfunction during pregnancy.
Particulate matter 2.5
Systolic blood pressure
Diastolic blood pressure
Haemoglobin A1c
Triglycerides
Total cholesterol
Low density lipoprotein cholesterol
High density lipoprotein cholesterol
Standard deviation
Adjusted odds ratio
Confidence intervals
Gestational diabetes mellitus
World Health Organization
Sulphur dioxide
Carbon dioxide
Nitrogen dioxide
Carbon monoxide
Long-Term Evolution:BMI:Body mass index
oral glucose tolerance test
C-reactive protein
Intrauterine growth restriction
Ethics approval and consent to participate
IRB approval: the present study was approved by the ethics committees of all the participating hospitals from where the pregnant women were recruited. All participants provided written informed consent prior to enrolment.
Consent for publication
Not applicable
Availability of data and materials
All data generated or analysed during this study are included in this published article [and its additional information files].
Competing interests
The authors declare that they have no competing interests.
Funding
This study was supported by the National Institute of Health research project (project No. #2021-ER1208-01)
Authors’ contributions
YSJ contributed to Conceptualization, Data curation; Formal analysis; Software; Visualization; Methodology; Writing – Original draft. SP contributed to Data curation; Software; Validation; Investigation. YMH contributed to Investigation. EK Data curation; Investigation. YAY contributed to Investigation. SMK contributed to Investigation. GL contributed to Investigation. KAL contributed to Investigation. SJK contributed to Investigation. GJC contributed to Investigation. MO contributed to Investigation. SHN contributed to Investigation. SL contributed to Investigation. JB contributed to Investigation. YK contributed to Investigation. SJL contributed to Investigation. NKK contributed to Investigation. YHK* Conceptualization; Funding acquisition; Writing – review & editing; Validation; Supervision. YJK* Conceptualization; Funding acquisition; Writing – review & editing; Supervision. All authors read and approved the final manuscript.
APPO study group
The APPO study group also includes the following current members: Jeong Eun Lee, Hwa Jeong Kim, So Jeong In, Hye Won Kim, and Bo Ra Kim.
Acknowledgments
This study was supported by the National Institute of Health research project (project No. #2021-ER1208-01). We are grateful to the APPO participants and staff at Yonsei University Hospital, Ewha Womans University Mokdong Hospital, Ewha Womans University Seoul Hospital, Korea University Guro Hospital, Kangwon National University Hospital, Keimyung University Dongsan Medical Center, and Ulsan University Hospital.
Authors’ information