Abstract
Background:
Polycyclic aromatic hydrocarbons (PAHs) are formed during incomplete burning of fossil fuels, wood, and tobacco products. High PAH exposure has been associated with low birth weight, intrauterine growth restriction, and preterm birth, but little is known about its impact on adverse outcomes in early pregnancy such as in-utero fetal death.
Objectives:
To examine associations between exposure to PAHs and missed abortion in which the embryo has died but a miscarriage has not yet occurred during early pregnancy in a Chinese population in Tianjin.
Methods:
A case–control study was conducted from April to November, 2007 in Tianjin, China. Cases experienced a missed abortion while controls underwent elective abortions before 14 weeks of pregnancy. Eighty-one cases were recruited from four hospitals, with the same number of controls matched on hospital, maternal age (± 8 years), gravidity (1 or >1), and gestational age (± 30 days). Two maternal measures of PAH exposures were obtained based on benzo[a]pyrene (BaP) DNA adducts in 1) aborted tissues and 2) maternal blood (for a subset of subjects). In addition, proxy measures for PAH exposures from different sources were derived from maternal interviews.
Results:
In conditional logistic regression analyses, we estimated more than 4-fold increase in risk of having experienced a missed abortion in women with above the median levels of blood BaP-DNA adducts (adjusted OR = 4.27; 95% CI, 1.41–12.99); but no increase with adduct levels in aborted tissues (adjusted OR = 0.76; 95% CI, 0.37–1.54). BaP-DNA adduct levels in maternal blood and aborted tissues were poorly correlated (r = −0.12; n = 102). Missed abortion risk also was higher among women reporting traffic congestion near the residence, commuting by walking, and performing regular cooking activities during pregnancy.
Conclusion:
High levels of maternal PAH exposures may contribute to an increased risk of experiencing a missed abortion during early pregnancy.
Keywords: Polycyclic aromatic hydrocarbon, DNA adduct, Air pollution, Pregnancy outcome, Abortion
1. Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a group of ubiquitous persistent organic pollutants released into the atmosphere during incomplete combustion and/or pyrolysis of fossil fuel, industrial or domestic coal, wood, cigarettes, and food items (Bostrom et al., 2002). Human exposure to PAHs occurs largely through inhalation and diet. In animal studies transplacental exposure to benzo[a]pyrene (BaP), one of the PAH species, has been associated with an increase in xenobiotic metabolism in the placental tissues (Sanyal, et al. 1994) and has been linked to fetal loss and a decrease in plasma hormone levels of progesterone, estrogen, and prolactin (Archibong et al., 2002). Exposure of rat fetuses to BaP during organogenesis reduced birth weight, crown-rump length, and placental proficiency as measured by protein weight and diameter (Sanyal and Li, 2007a).
In humans, studies conducted in the U.S. (Choi et al., 2006; Choi et al., 2008; Perera et al., 2003; Perera et al., 2004), Poland (Choi et al., 2006), and the Czech Republic (Dejmek et al., 2000) suggested that PAHs adversely impact birth outcomes and cognitive development in early childhood; including decreased head circumferences, birth length, and birth weight (Choi et al., 2006; Perera et al., 2003; Perera et al., 2005c; Perera et al., 1998; Tang et al., 2006), intrauterine growth retardation (Choi et al., 2008; Dejmek et al., 2000), preterm birth (Choi et al., 2008; Singh et al., 2008), and neurocognitive development in children (Perera et al., 2005b; Tang et al., 2006).
Most prior studies examined outcomes in live births only and were based on measures of PAH exposure taken in the later part of the pregnancy (e.g. the 3rd trimester). However, fetuses may be particularly sensitive to the impact of toxic PAHs during early pregnancy (e.g. the 1st trimester) because of the rapid development of fetal organs and the higher exposure per body weight of the fetus during this period compared to later pregnancy periods (Pinkerton and Joad, 2006; Schwartz, 2004). No study to date has investigated the potential impact of PAH exposure on outcomes such as abnormal fetal development and in-utero fetal death during early pregnancy. Missed abortion refers to an intrauterine pregnancy in which the fetus does not develop normally over a prolonged period of time, typically 6 weeks, or the fetus is already known to have died but the products of conception remain in-utero. Missed abortion is a complication of early pregnancy that occurs in up to 15% of all clinically recognized pregnancies (Daya, 2000), and can have severe social, emotional, and psychological impacts due to the failure of the pregnancy. Approximately 90% of all missed abortions occur before 14 weeks of gestation or in the 1st trimester.
We conducted a case–control study examining the influence of maternal exposure to PAHs on missed vs. induced abortions in Tianjin, an industrial city in northern China. Maternal PAH exposures were measured via biomarker in 1) maternal blood and 2) aborted tissues; we also conducted an interview survey to identify potential sources of environmental PAH exposures in these women.
2. Methods
2.1. Study area
The study was conducted in Tianjin, the second largest city in northern coastal China with a total area of 11,760 km2, and a population of 11.2 million in 2007. Tianjin is located at the lower reaches of the Haihe River and adjacent to the Bohai Sea, and is an important industrial center and a well-developed hub with a sea–land–air transportation network. As one of the fastest growing areas in coastal China, Tianjin measures atmospheric concentrations for PAHs much higher than many other cities worldwide (Wu et al., 2005). The major contributors to PAH pollution are coal combustion, vehicle emission, coking industry, and biomass burning (Zuo et al., 2007).
2.2. Subject selection
Using a case–control study design, between early April and end of November 2007 we enrolled cases with missed abortions and controls seeking induced abortion services in four hospitals, including the Main Hospital of Tianjin Medical University, the 2nd Hospital of Tianjin Medical University, the Hospital of Chinese People's Armed Police Forces, and Dongli Hospital. The study was approved by the Institutional Review Board of each of the four hospitals. Cases were pregnant women carrying an in-utero fetus that was confirmed as dead by ultrasound measurements before 14 weeks of gestation. Controls were women with normal pregnancies who requested an induction of an abortion due to an unplanned and unwanted pregnancy. Each case was assigned one control matched on hospital, maternal age (within ± 8 years), and gravidity (the total number of times a woman has been pregnant; one or more than one). Because no reliable data for cases were available to determine the exact date when the fetus died in-utero, we loosely matched the case control pairs on gestational age at time of abortion requiring the control's gestational age to be within 30 days of the case's. In the final matched dataset, cases had slightly longer gestational age (4 days) than controls because some case mothers wishing to maintain the pregnancy preferred to wait longer to confirm the death of the fetus. Yet, case and control fetuses may have been more similar in their actual gestational age at demise, i.e. the fetus of the missed abortion cases remained in-utero longer.
All study participants were screened to exclude smokers, women with chronic diseases (e.g. hypertension, heart disease, and diabetes) and pregnancy complications (e.g. vaginal infection), women with potentially high occupational exposures to PAH according to the literature (e.g. bus or taxi drivers, traffic police officers, coal plant workers, coke oven workers, cooks), and those who had resided in Tianjin for less than one year. All cases and controls were approached by trained research staff in the surgery room immediately following the abortion procedure. A total of 84 eligible cases were contacted and 81 of them agreed to participate in the study. Eight-one matched controls were recruited out of 89 eligible subjects approached.
2.3. Environmental and health survey
Approximately 1 h after the abortion procedure, we consented eligible subjects for the collection of biospecimen (aborted tissues and/or blood draws) and the administration of a 15–20min standardized interview administered by trained personnel. Each matched pair was interviewed by the same interviewer. We asked for three types of information: (i) demographics and socioeconomic information (e.g. age, home address, duration of residence in Tianjin, education, occupation, and income); (ii) reproductive history (e.g. previous abortion history, and number of previous birth to a fetus with a gestational age of 24 weeks or more — parity); and (iii) factors that may influence exposure to PAHs during pregnancy (e.g. time spend outdoors and in-transit, nearby industrial facilities, traffic activities near residence, cooking activities and fuels used, dietary PAH exposure via grilled, smoked, or barbecued foods, exposure to environmental tobacco smoke (ETS) etc.). We also ascertained information about other maternal risk factors suspected to affect pregnancy outcomes (e.g. virus infection, medication, X-ray exposure, and emotional disturbances); yet, very few cases (n = 0–6) and controls (n = 1–3) responded positive to these questions and excluding these participants did not change the main results of missed abortion (data not shown); thus we did not further consider these factors in our analyses.
2.4. Biomarker measurements
All four hospitals allowed us to collect the aborted tissues, but only two of the hospitals gave us permission for a maternal blood draw. Thus, maternal blood samples were collected for only a subset of the matched case–control pairs. The two hospitals that allowed blood draws were within 6km of the other two hospitals not permitting these procedures. All four hospitals employed the same diagnostic criteria, served a similar population, and were located within 5–13 km of each other.
Right after the interview, we collected 2 ml of maternal blood for the consented women by syringing blood into a heparinized EDTA-K2 glass tube (Nantong Fulilai Medical Instrument Co., Ltd., Jiangshu Province, China) to avoid clotting. The aborted tissue was analyzed in total because it is difficult to separate the villous from the embryo early in pregnancy, i.e. all samples contained both embryo and villous materials together. The aborted tissue was prewashed using 0.9% NaCl saline solution, and stored in 2 ml sample bottle (Vigorous Biotechnology, Inc., Beijing, China). All samples were put on ice and transported to the Medicinal Chemistry Laboratory of the Basic Medical Science Center at Tianjin Medical University immediately after collection for storage and processing. The whole blood samples were refrigerated at −20 °C, and the aborted tissues at −70 °C until analysis. Whole blood samples were extracted every week for genomic DNA using Human Whole Blood Genomic DNA Purification Kit (MagneSil, Promega Madison, WI, USA) and aborted tissues using Tissue/Cell Genomic DNA Purification Kit (GaiNing Biotechnology Co., Ltd., Beijing, China). BaP-DNA adducts in extracted DNA were analyzed using the high-performance liquid chromatography-fluorescence method (Alexandrov et al., 1992), which detects benzo[a] pyrene diol-epoxide (BPDE) tetrols. The assay is a sensitive and specific method for measuring BaP-DNA adducts in individuals (Bartsch and Hietanen, 1996). UV-260 spectrophotometer was used to analyze total DNA. Finally, standardized BaP-DNA adduct levels were calculated by dividing the BPDE concentrations by the total DNA concentrations.
2.5. Statistical analyses
All statistical analyses were conducted using R statistical software (version 2.6.1; The R Foundation for Statistical Computing, Vienna, Austria). We fitted multivariate conditional logistic regression models to our matched data to examine the influence of the PAH adduct levels in maternal blood and aborted tissue on the risk of missed abortion. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated for the association between missed abortions and BaP-DNA adduct levels and also some of the interview-derived environmental measures we considered surrogates for PAH exposure from different sources. We modeled BaP-DNA adduct levels as a continuous exposure (approximate lognormal distribution), and also compared subjects with higher to those with lower adduct levels using the median adduct level in the control group as the cut point. In addition, we adjusted in our analyses for confounding by maternal educational attainment and household income.
Interview-derived surrogates for PAH exposures were examined either as binary variables (e.g. traffic congestion near the residence, an industrial site or smoke stack near the residence) or categorical variables (e.g., self-reported residential distance to the nearest road, ETS, consumption of grilled, smoked, or barbecued foods, time spend outdoors/indoors/commuting, and frequency of cooking activities). Single- and multi-variable linear regression models were developed to predict BaP-DNA adduct levels in maternal blood and aborted tissues based on the questionnaire-based surrogate exposure variables while taking maternal educational attainment, household income, and gestational age into account. Employing a stepwise method we retained the variables found to be statistically significant (p<0.05) in our multi-variable linear prediction model.
3. Results
For all 81 eligible cases and the 81 matched controls enrolled in the study, we conducted interviews and collected aborted tissue samples, but only for 102 (51 pairs) we collected maternal blood samples (from two of four hospitals that gave permission). Table 1 shows the selected characteristics for the case and control groups. Most (90.1%) study participants were between 20 and 35 years of age. Missed abortion cases were less educated, reported lower monthly household income, and were somewhat younger than controls seeking induced abortions. Seventy-seven matched pairs were within 5 years of maternal age, the four remaining pairs were within 6–8 years. Differences between women with and without blood samples are illustrated in Supplemental Material (Table S1). The adduct levels in aborted tissues and most interview-derived exposures were similar in case and control women with and without a blood sample, i.e. women with blood samples did not seem to differ from those without blood samples by case status.
Table 1.
Selected characteristics of matched case–control subjects in the study population.
| Variables | Case (N = 81) | Control (N = 81) |
|---|---|---|
|
|
|
|
| n (%) | n (%) | |
|
| ||
| Maternal age (year) | ||
| <20 | 0 (0.0) | 3 (3.7) |
| 20–29 | 62 (76.5) | 50 (61.7) |
| 30–39 | 16 (19.8) | 24 (29.6) |
| ≥40 | 3 (3.7) | 4 (4.9) |
| Maternal education | ||
| Elementary school | 8 (9.9) | 1 (1.2) |
| Middle school | 38 (46.9) | 30 (37.0) |
| High school | 22 (27.2) | 33 (40.7) |
| College | 13 (16.0) | 15 (18.5) |
| Graduate school or higher | 0 (0.0) | 1 (1.2) |
| Unknown | 0 (0.0) | 1 (1.2) |
| Monthly household income (yuan) | ||
| <2000 | 40 (49.4) | 27 (33.3) |
| 2000–5000 | 36 (44.4) | 49 (60.5) |
| 5001+ | 3 (3.7) | 2 (2.5) |
| Unknown | 2 (2.5) | 3 (3.7) |
| Gestational age (days) | ||
| ≥49 | 2 (2.5) | 4 (4.9) |
| 50–59 | 12 (14.8) | 13 (16.0) |
| 60–69 | 25 (30.9) | 28 (34.6) |
| 70–79 | 15 (18.5) | 15 (18.5) |
| ≥80 | 27 (33.3) | 21 (25.9) |
| Gravidity | ||
| 1 | 33 (40.7) | 33 (40.7) |
| 2 | 21 (25.9) | 26 (32.1) |
| 3 | 18 (22.2) | 15 (18.5) |
| 4+ | 9 (11.1) | 7(8.6) |
| Parity | ||
| 0 | 55 (67.9) | 44 (54.3) |
| 1 | 22 (27.2) | 33 (40.7) |
| ≥2 | 4 (4.9) | 4 (4.9) |
Reproducibility and recovery studies were performed to examine the performance of the analytical methods in our lab. The within-day repeatability and between-day reproducibility yielded satisfactory results below 5%. For the sample spiked with BPDE standard concentrations (10, 100, and 800 ng/ml), the recovery rate ranged from 88% to 114%, with the relative standard deviation (RSD) generally below 10%. Our detection limit was 0.12 BaP-DNA adduct/108 nucleotides.
BaP-DNA adduct levels in maternal blood and in aborted tissues in cases and controls are listed in Table 2. BaP-DNA adducts in maternal blood correlated poorly and slightly negatively with the levels measured in aborted tissues (overall r = −0.12, n = 102; cases r = −0.02, n = 51; controls r = −0.21, n = 51). BaP-DNA adduct levels in maternal blood were significantly higher in the case group (6.0 adducts/108 nucleotides) than the control group (2.7 adducts/108 nucleotides) (p = 0.000 for difference). The mean number of BaP-DNA adducts in the aborted tissues was somewhat lower in the case (4.8 adducts/108 nucleotides) than the control group (6.0 adducts/108 nucleotides), however, this difference was not statistically significant (p = 0.29). In controls the adduct levels were on average 4.2 adducts/108 nucleotides higher in aborted tissues than in maternal blood (p<0.001 for difference; n = 51); while in cases the adduct levels were on average only 0.8 adducts/108 nucleotides higher in the aborted tissues than in the maternal blood (p = 0.48 for difference; n = 51).
Table 2.
Measured BaP-DNA adduct levels (/108 nucleotides) in maternal blood and aborted tissues.
| Mean | SDa | Min | Max | ||
|---|---|---|---|---|---|
|
| |||||
| Maternal blood | All | 4.3 | 4.0 | 0.1 | 21.8 |
| Controls | 2.7 | 2.2 | 0.2 | 12.8 | |
| Cases | 6.0 | 4.7 | 0.1 | 21.8 | |
| Aborted tissues | All | 5.4 | 6.7 | 0.2 | 33.2 |
| Controls | 6.0 | 7.4 | 0.2 | 33.2 | |
| Cases | 4.8 | 6.0 | 0.2 | 32.2 | |
SD: standard deviation.
After controlling for maternal education and household income, maternal blood BaP-DNA level (per adduct/108 nucleotides) increased the risk of having experienced a missed abortion (adjusted OR = 1.35; 95% CI, 1.11–1.64) (Table 3). When we categorized the number of BaP-DNA adducts according to the median in the control group, the risk of a missed abortion was more than 4-fold higher in those exposed above the 50th percentile (adjusted OR = 4.27; 95% CI, 1.41–12.99). Additionally adjusting for gestational age slightly increased the odds ratios; however, since the increments were less than 10% of the estimated effect size, we report the remaining results without adjusting for gestational age. Missed abortions were not associated with the BaP-DNA adducts in aborted tissues whether or not we treated adduct levels as a continuous variable (per adduct/108 nucleotides) (adjusted OR = 0.97; 95% CI, 0.93–1.02) or as a binary variable (median of the control group) (adjusted OR = 0.76; 95% CI, 0.37–1.54) or controlled for maternal education and household income. The association between adducts in aborted tissues and missed abortion changed minimally when we restricted analyses only to women for whom we also had obtained blood samples (n = 102) (continuous variable: adjusted OR = 0.96; 95% CI, 0.90–1.03 and binary variable: adjusted OR = 0.68; 95% CI, 0.27–1.73 after controlling for maternal education and household income).
Table 3.
Estimated odds ratios relating BaP-DNA adducts in maternal blood and aborted tissues with missed abortion using conditional logistic regression model.
| Exposure variable | Model | OR (95% CI) | p |
|---|---|---|---|
|
| |||
| Maternal blood (N = 102) | |||
| Continuous adduct levels (/108 nucleotides) | Unadjusted model | 1.31 (1.11–1.55) | 0.002 |
| Adjusted model 1: controlled for maternal education | 1.33 (1.10–1.60) | 0.003 | |
| Adjusted model 2: controlled for maternal education and household income | 1.35 (1.11–1.64) | 0.003 | |
| Adjusted model3: controlled for maternal education, household income, and gestational age | 1.37 (1.12–1.67) | 0.002 | |
| Categorical adduct levels (high vs. low) using the 50th percentile value of the controls as the reference group | Unadjusted model | 4.00 (1.50–10.66) | 0.006 |
| Adjusted model 1: controlled for maternal education | 3.34 (1.21–9.20) | 0.020 | |
| Adjusted model 2: controlled for maternal education and household income | 4.27 (1.41–12.99) | 0.011 | |
| Adjusted model 3: controlled for maternal education, household income, and gestational age | 4.56 (1.46–14.27) | 0.009 | |
| Aborted tissues (N = 162) | |||
| Continuous adduct levels (/108 nucleotides) | Unadjusted model | 0.98 (0.93–1.02) | 0.29 |
| Adjusted model 1: controlled for maternal education | 0.98 (0.93–1.02) | 0.29 | |
| Adjusted model 2: controlled for maternal education and household income | 0.97 (0.93–1.02) | 0.28 | |
| Adjusted model 3: controlled for maternal education, household income, and gestational age | 0.96 (0.92–1.02) | 0.16 | |
| Categorical adduct levels (high vs. low) using the 50th percentile value of the controls as the reference group | Unadjusted model | 0.86 (0.46–1.61) | 0.63 |
| Adjusted model 1: controlled for maternal education | 0.83 (0.43–1.60) | 0.58 | |
| Adjusted model 2: controlled for maternal education and household income | 0.76 (0.37–1.54) | 0.44 | |
| Adjusted model 3: controlled for maternal education, household income, and gestational age | 0.61 (0.28–1.32) | 0.21 | |
For interview-derived proxy measures for PAH sources, we observed a risk increase for missed abortion in women who reported commuting by walking (adjusted OR = 3.52; 95% CI, 1.44–8.57), traffic congestion near their residence (adjusted OR = 3.07; 95% CI, 1.31–7.16), living near an industrial site (adjusted OR = 3.21; 95% CI, 0.98–10.48), and frequent cooking activities during pregnancy (adjusted OR = 3.78; 95% CI, 1.11–12.87) (Table 4). Since most participants (91%) reported using only one type of transportation, the effect estimates changed little when we restricted analyses to those subjects with a single mode of transportation only (data not shown). Living close to a roadway and spending more time outdoors was also positively associated with missed abortion. Surprisingly, reporting the consumption of grilled, smoked, or barbecued foods seemed to decrease the risk of missed abortion even after we adjusted for socioeconomic factors (adjusted OR = 0.22; 95% CI, 0.06–0.86). However, these foods were mostly consumed by women who reported never or occasionally rather than routinely cooking (23%, 17%, and 9% among those never, occasionally or routinely cooking ate grilled, smoked, or barbecued foods). After controlling for cooking frequency, the consumption of grilled, smoked, or barbecued foods was no longer significantly associated with the risk of missed abortion (adjusted OR = 0.28; 95% CI, 0.07–1.13 adjusted also for maternal education and household income).
Table 4.
Odds ratios of single questionnaire exposure variables for missed abortion.
| Exposures | Cases (n) | Controls (n) | Unadjusted OR (95% CI) | Adjusted OR (95% CI)a |
|---|---|---|---|---|
|
| ||||
| ETS | ||||
| Spouse smoking | ||||
| Never | 32 | 36 | 1.00 | 1.00 |
| Sometimes | 16 | 15 | 1.21 (0.50–2.91) | 0.78 (0.29–2.11) |
| Often | 33 | 30 | 1.23 (0.62–2.42) | 1.10 (0.52–2.30) |
| Traffic-related air pollution | ||||
| Residential distance to the nearest road | ||||
| >100 m | 29 | 40 | 1.00 | 1.00 |
| ≤100 m | 51 | 40 | 1.71 (0.89–3.31) | 1.63 (0.75–3.56) |
| Traffic congestion near the residence | ||||
| No | 48 | 60 | 1.00 | 1.00 |
| Yes | 33 | 19 | 2.18 (1.07–4.45) | 3.07 (1.31–7.16) |
| Daily commute time | ||||
| ≤30 min | 44 | 45 | 1.00 | 1.00 |
| >30 min | 37 | 36 | 1.05 (0.57–1.94) | 1.06 (0.53–2.10) |
| Commute by walking | ||||
| No | 47 | 67 | 1.00 | 1.00 |
| Yes | 34 | 14 | 3.00 (1.47–6.14) | 3.52 (1.44–8.57) |
| Commute by biking | ||||
| No | 60 | 62 | 1.00 | 1.00 |
| Yes | 21 | 19 | 1.22 (0.51–2.95) | 0.92 (0.35–2.46) |
| Commute by passenger cars | ||||
| No | 67 | 67 | 1.00 | 1.00 |
| Yes | 14 | 14 | 1.00 (0.46–2.16) | 1.39 (0.57–3.38) |
| Commute by bus | ||||
| No | 64 | 58 | 1.00 | 1.00 |
| Yes | 17 | 23 | 0.65 (0.30–1.38) | 0.45 (0.18–1.10) |
| Daily average time outdoors | ||||
| ≤2 h | 43 | 54 | 1.00 | 1.00 |
| >2 h | 36 | 27 | 2.00 (0.97–4.12) | 1.92 (0.89–4.14) |
| Diet and cooking | ||||
| Consumption of grilled, smoked, and barbecued foods 1–2 times/week | ||||
| No | 73 | 64 | 1.00 | 1.00 |
| Yes | 8 | 16 | 0.27 (0.08–0.98) | 0.22 (0.06–0.86) |
| Cooking during pregnancy | ||||
| Never | 15 | 25 | 1.00 | 1.00 |
| Occasionally | 17 | 15 | 2.58 (0.85–7.86) | 2.89 (0.82–10.13) |
| Sometimes | 22 | 22 | 2.00 (0.77–5.20) | 1.73 (0.58–5.20) |
| Routinely | 27 | 19 | 3.39 (1.15–9.95) | 3.78 (1.11–12.87) |
| Industrial emission | ||||
| Industrial sites or stacks near the residence | ||||
| No | 68 | 76 | 1.00 | 1.00 |
| Yes | 13 | 5 | 3.00 (0.97–9.30) | 3.21 (0.98–10.48) |
Adjusted for maternal education and household income.
BaP-DNA adduct levels in maternal blood were significantly (p<0.05) and positively associated with several survey-derived environmental exposure surrogates including proximity of the residence to the nearest roadway, commuting by walking, traffic congestion near the residence, and daily average time outdoors, and significantly negatively associated with the consumption of grilled, smoked, or barbecued foods 1–2 times/week (see Supplemental Material, Table S2). When adjusting for case–control status, commuting by walking, traffic congestion near the residence, and daily average time outdoors remained significantly associated with maternal blood adduct levels (see Supplemental Material, Table S2). Consumption of grilled, smoked, or barbecued foods 1–2 times/week was negatively associated with maternal blood adduct levels (p = 0.08) but positively associated with chorionic adduct levels (p = 0.02) after adjustment for case–control status.
The linear regression prediction model we developed explained 26% of the variance in maternal blood adduct levels with only 3 variables in the model: BaP-DNA adduct level in maternal blood (/108 nucleotides) = 1.64 + 2.38 × case–control status (0 or 1) + 2.41 × commuting by walking during pregnancy (no or yes) + 1.68 × daily average time outdoors (≤2h or >2h) (n = 102; R2 = 0.26).
No meaningful multi-variable predictive model was obtained for BaP-DNA adduct levels in aborted tissues, i.e. the overall explained variance was never higher than 5%.
4. Discussion
We conducted a case–control study to examine the impact of PAH exposure on missed abortions in early pregnancy in a Chinese population exposed to high levels of ambient air pollution from combustion sources. Mothers' PAH exposures were measured relying mainly on two biomarkers i.e. PAH-DNA adducts in maternal blood and in aborted tissues; in addition, we collected some data on potential environmental sources of PAH exposures in interviews. To our knowledge, this is the first study to examine whether maternal blood and abortus-derived PAH adducts are associated with missed abortions.
Maternal blood but not aborted tissues BaP-DNA adduct levels was strongly associated with the risk of missed abortion and with three interview-derived environmental sources of PAHs, specifically self-reported traffic congestion near the residence, commuting by walking, and daily average time outdoors; self-reported traffic congestion and commuting by walking were also found to be associated with the risk of missed abortions while frequent cooking at home was associated with the risk of missed abortions but not the adduct levels in maternal blood.
The lack of an association between missed abortion and PAH adduct levels in aborted tissue and the absence of any correlation between maternal blood and aborted tissues DNA adduct levels in both cases and controls may be attributable to the more active and complex metabolism taking place in the rapidly proliferating placental and fetal cells of early pregnancy compared to the fully differentiated lymphocytes of the mother. Consistent with our data, prior studies have reported a lack of correlation between maternal blood and cord blood adducts (Perera et al., 2005a; Topinka et al., 2009; Whyatt et al., 2001) and between placental tissue and maternal and newborn blood adduct levels (Topinka et al., 2009). Topinka et al. (2009) proposed that the placenta might act as a protective barrier against some genotoxic components between the circulation of mother and child. PAH-DNA adduct levels in placental tissues have been found to be influenced by the activity of the enzymes cytochrome P450–1A1 and glutathione S-transferases (Whyatt et al., 2000). Cytochrome P4501A1 enzymatically activates c-PAHs that form genotoxic products (diol epoxides) and bind covalently to DNA; this enzyme is strongly induced by maternal cigarette smoking (Sanyal and Li, 2007b) and possibly by diet (Manchester et al., 1992). Induction of CYP1A1 in placentas might reduce the biologically effective dose to the fetus but also at the same time increase the number of DNA adducts in the placenta (Manchester et al., 1984; Topinka et al., 2009).
The overall average BaP-DNA adduct levels (including both cases and controls) were about 26% higher in the aborted tissue than in maternal blood. A study conducted in the Czech Republic found significantly higher BaP-like DNA adduct levels in maternal blood than placental tissue sampled at birth in women unexposed to tobacco smoke (plasma cotinine level ≤3 ng/ml) and similar levels of adducts in maternal blood and placenta in women exposed to tobacco smoke (Topinka et al., 2009). The higher adduct levels measured in aborted tissues than maternal blood in our study might be due to the aborted tissues having been collected early in pregnancy when the fetal/villous tissues exhibit a very high metabolism rate in order to form a fully functional placenta, different from the Czech Republic study that collected fully-developed placental tissues at time of delivery. Placental metabolism of and fetal susceptibility to PAH exposures might be different at early compared to late pregnancy, due to differences in rates of cell proliferation, ability to activate and detoxify carcinogens, DNA repair capacity, and the number of target cells at risk (Anderson et al., 2000). We were able to analyze some placental tissues at around 40 weeks of gestation (n = 11) at one of the two hospitals where we collected both maternal blood and aborted tissues. These women with a full term delivery lived in the same urban area and were of comparable socioeconomic status as those for whom we had sampled blood and abortus tissues. The measured mean BaP DNA adduct level was 1.3 ± 0.7/108 nucleotides (mean ± standard deviation (SD)) in the placental tissues at delivery, much lower than the average levels of 5.4 ± 6.7 adduct/108 nucleotides (mean ± SD) we observed in induced and missed aborted tissues collected in early pregnancy.
This study was conducted during a relatively warm period (April–November), reducing the potential for high PAH exposures from coal combustion during home heating in this city. Limited monitoring station data from the Tianjin Environmental Protection Agency showed particle-bounded BaP concentrations to be as high as 14.5 ng/m3 in January compared to only 1.3 ng/m3 in July of 2006. Another study found particle-bounded BaP concentrations in this city more than two times higher in the winter (December–February) than in autumn (September–October) (Wu et al., 2005). Thus, we suspect that maternal PAH exposures in Tianjin are even higher for women whose 1st trimester falls into the winter season.
The biologic mechanisms through which parental exposures to PAHs might affect early pregnancy loss are not well understood but may involve the induction of apoptosis after DNA damage from PAHs, antiestrogenic effects of PAHs, and binding to the human aryl hydrocarbon hydroxylase to induce P450 enzymes or to receptors for placental growth factors, altering trophoblast proliferation as well as its endocrine function and decreasing exchange of oxygen and nutrients between the fetus and the mother (Choi et al., 2006; Dejmek et al., 2000). In addition, fetotoxicity of PAHs has been associated with a genetic predisposition of the parents to aryl hydrocarbon hydroxylase induction (Perera et al., 1998; Shum et al., 1979; Sram et al., 1999).
Spontaneous abortions occur in 10–15% of pregnant women in China (Le, 2008). Based on our observations missed abortions account for approximately 30% of all spontaneous abortions seen at our hospitals, yet this rate does not take into account women who have spontaneous abortion but do not visit hospitals and the rate may vary by region or hospital. While our study design did not allow us to examine the impact of maternal PAH exposures on other types of spontaneous abortions, missed abortions are unique in that they make the sampling of villous materials feasible. In order to generalize our findings to all spontaneous abortions, we need to assume that PAHs exposures do not preferentially affect missed abortions. On the contrary, it would even be conceivable that the highest exposures lead to earlier spontaneous abortions that may or may not be clinically recognized by women; in that case, we would underestimate the adverse effects of PAHs on all abortions.
We found some of the environmental exposure sources ascertained in the maternal interview to be associated with an increased risk of missed abortion and/or predictors of elevated BaP-DNA adduct levels in maternal blood. One such environmental predictor variable was self-reported traffic congestion near the residence. Previous studies found weak correlations between maternal PAH-DNA adduct levels and estimated environmental PAH exposures at the individual level (e.g. PAHs in personal air, ETS, and dietary PAH consumption), probably due to the individual variations in biological responses (Perera et al., 2005a). Nevertheless several studies have linked biomarker-based personal PAH exposure levels in urine (e.g. urinary 1-hydroxypyrene) to traffic-related emissions, including estimated traffic-related air pollution as a surrogate biomarker for total PAH (Freire et al., 2009), on-road commute time (Gunier et al., 2006), and proximity to traffic (Tuntawiroon et al., 2007). In our study, commuting by walking but not by other means (i.e. passenger cars, bike, bus) was associated with maternal blood PAH-adduct levels. Also, while commuting by walking and biking was associated with lower maternal education and income, whereas commuting by vehicle and bus was linked to higher maternal education and income, adjustment for maternal education and household income did not influence the association between maternal adduct levels and transportation means. Biking was associated with shorter self-reported commute time (p = 0.01) and potentially less exposure to traffic-generated PAH exposures; self-reported commute time was not associated with walking, or using a motor vehicle or bus. Further investigations of the importance of walking compared to other means of transportation on maternal blood adduct levels are warranted since our results are based on very few crude questions and a relatively small sample size. The consumption of grilled, smoked, or barbecued foods paradoxically seemed to decrease the risk of missed abortions while cooking frequently at home increased risk. However, intake of possibly high PAH food items was found to be associated with a low frequency of cooking at home, i.e. the less a woman reported cooking at home the more she may have eaten outside the home; and while this Chinese population seldom cooks, grills, smokes, or barbecues food at home such foods are easily available from vendors outside the home.
While missed abortions were associated with frequent cooking activities, we did not find statistically significant correlation between cooking frequency and maternal blood PAH-DNA adduct levels, although the highest levels were observed among those reporting cooking more routinely. Consumption of grilled, smoked, and barbecued foods 1–2 times/week was negatively associated with PAH-DNA adduct levels in maternal blood (P = 0.08). It is possible that this question did not distinguish between food types (e.g. meat vs. vegetable) and was misinterpreted by the interviewees. For instance, it is not uncommon for the Chinese population to buy grilled or barbecued vegetables (e.g. sweet potato and chestnuts) from street vendors. The lack of association between self-reported cooking frequency and the adduct levels may reflect individual variation in adduct formation due to co-exposures, nutritional, and genetic factors (Perera et al., 2005a). Routinely cooking at home may also be associated with other risk factors (e.g. life style) for missed abortion that were not accounted for in our interview. Finally, living near an industrial site was associated with an increased risk of missed abortion (adjusted OR = 3.21; 95% CI, 0.98–10.48), but was not associated with the adduct levels in maternal blood and aborted tissues. This is likely attributable to the small percentage of subjects who reported living close to an industrial site (11%) and possibly also influenced by recall bias. In fact, a major limitation of our interview-based results is the retrospective nature of the exposure source assessment that allows for recall bias, such that women experiencing missed abortion might be more eager to find a causal explanation for this event than pregnant women who elected to have an abortion. Thus, we mainly relied on the results related to the biospecimens we collected and the BaP-DNA adduct levels in maternal blood and aborted tissues.
5. Conclusion
Using a case–control study design, we estimated that BAP-DNA adduct levels in maternal blood but not aborted tissues increased the risk for missed abortion in pregnant women living in Tianjin, China. BAP-DNA adduct levels in maternal blood and aborted tissues, however, were poorly correlated. A higher risk of missed abortion was also observed for maternal exposure to traffic-related pollution (i.e. commute by walking and traffic congestion near the residence) and routinely cooking during pregnancy. Maternal BAP-DNA adduct levels were moderately well predicted by case control status, commuting by walking, and daily average time spend outdoors (R2 = 0.26). Additional studies with larger sample sizes are needed to confirm the present findings and to examine underlying biological mechanisms.
Supplementary Material
Acknowledgements
The study was supported by Tianjin Science Council (No. 07JCZDJC07400) and the U.S. National Institute of Environmental Health Sciences (NIEHS, 1R21ES016379-01A1-Wu). The authors acknowledge the help from the Main Hospital of Tianjin Medical University, the 2nd Hospital of Tianjin Medical University, Hospital of Chinese People's Armed Police Forces, and Dongli Hospital of Tianjin. We especially thank Basic Research Center of the Main Hospital of Tianjin Medical University for processing the samples, and the researchers in the Epidemiology Division and in the Immunology Division of the Hospital of Chinese People's Armed Police Forces for helping analyzing the data.
Footnotes
The authors declare they have no competing financial interests.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at doi: 10.1016/j.scitotenv.2010.02.028.
References
- Alexandrov K, Rojas M, Geneste O, Castegnaro M, Camus AM, Petruzzelli S, et al. An improved fluorometric assay for dosimetry of benzo(a)pyrene diol-epoxide-DNA adducts in smokers' lung: comparisons with total bulky adducts and aryl hydrocarbon hydroxylase activity. Cancer Res 1992;52(22):6248–53. [PubMed] [Google Scholar]
- Anderson LM, Diwan BA, Fear NT, Roman E. Critical windows of exposure for children's health: cancer in human epidemiological studies and neoplasms in experimental animal models. Environ Health Perspect 2000;108(Suppl 3):573–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Archibong AE, Inyang F, Ramesh A, Greenwood M, Nayyar T, Kopsombut P, et al. Alteration of pregnancy related hormones and fetal survival in F-344 rats exposed by inhalation to benzo(a)pyrene. Reprod Toxicol 2002;16(6):801–8. [DOI] [PubMed] [Google Scholar]
- Bartsch H, Hietanen E. The role of individual susceptibility in cancer burden related to environmental exposure. Environ Health Perspect 1996;104(Suppl 3):569–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bostrom CE, Gerde P, Hanberg A, Jernstrom B, Johansson C, Kyrklund T, et al. Cancer risk assessment, indicators, and guidelines for polycyclic aromatic hydrocarbons in the ambient air. Environ Health Perspect 2002;110(Suppl 3):451–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi H, Jedrychowski W, Spengler J, Camann DE, Whyatt RM, Rauh V, et al. International studies of prenatal exposure to polycyclic aromatic hydrocarbons and fetal growth. Environ Health Perspect 2006;114(11):1744–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi H, Rauh V, Garfinkel R, Tu Y, Perera FP. Prenatal exposure to airborne polycyclic aromatic hydrocarbons and risk of intrauterine growth restriction. Environ Health Perspect 2008;116(5):658–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Daya S. Habitual abortion. In: Copeland LJ, editor. Textbook of gynecology. Philadelphia, PA: WB Saunders; 2000. p. 227–71. [Google Scholar]
- Dejmek J, Solansky I, Benes I, Lenicek J, Sram RJ. The impact of polycyclic aromatic hydrocarbons and fine particles on pregnancy outcome. Environ Health Perspect 2000;108(12):1159–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Freire C, Abril A, Fernandez MF, Ramos R, Estarlich M, Manrique A, et al. Urinary 1-hydroxypyrene and PAH exposure in 4-year-old Spanish children. Sci Total Environ 2009;407(5):1562–9. [DOI] [PubMed] [Google Scholar]
- Gunier RB, Reynolds P, Hurley SE, Yerabati S, Hertz A, Strickland P, et al. Estimating exposure to polycyclic aromatic hydrocarbons: a comparison of survey, biological monitoring, and geographic information system-based methods. Cancer Epidemiol Biomarkers Prev 2006;15(7):1376–81. [DOI] [PubMed] [Google Scholar]
- Le J. Obstetrics and gynecology 7th ed.. Beijing, P.R. China: People's Medical Publishing House; 2008. [Google Scholar]
- Manchester DK, Parker NB, Bowman CM. Maternal smoking increases xenobiotic metabolism in placenta but not umbilical vein endothelium. Pediatr Res 1984;18 (11):1071–5. [DOI] [PubMed] [Google Scholar]
- Manchester DK, Bowman ED, Parker NB, Caporaso NE, Weston A. Determinants of polycyclic aromatic hydrocarbon-DNA adducts in human placenta. Cancer Res 1992;52(6):1499–503. [PubMed] [Google Scholar]
- Perera FP, Whyatt RM, Jedrychowski W, Rauh V, Manchester D, Santella RM, et al. Recent developments in molecular epidemiology: a study of the effects of environmental polycyclic aromatic hydrocarbons on birth outcomes in Poland. Am J Epidemiol 1998;147(3):309–14. [DOI] [PubMed] [Google Scholar]
- Perera FP, Rauh V, Tsai WY, Kinney P, Camann D, Barr D, et al. Effects of transplacental exposure to environmental pollutants on birth outcomes in a multiethnic population. Environ Health Perspect 2003;111(2):201–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perera FP, Rauh V, Whyatt RM, Tsai WY, Bernert JT, Tu YH, et al. Molecular evidence of an interaction between prenatal environmental exposures and birth outcomes in a multiethnic population. Environ Health Perspect 2004;112(5):626–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perera F, Tang D, Whyatt R, Lederman SA, Jedrychowski W. DNA damage from polycyclic aromatic hydrocarbons measured by benzo[a]pyrene-DNA adducts in mothers and newborns from Northern Manhattan, the World Trade Center Area, Poland, and China. Cancer Epidemiol Biomarkers Prev 2005a;14(3):709–14. [DOI] [PubMed] [Google Scholar]
- Perera FP, Rauh V, Whyatt RM, Tang D, Tsai WY, Bernert JT, et al. A summary of recent findings on birth outcomes and developmental effects of prenatal ETS, PAH, and pesticide exposures. Neurotoxicology 2005b;26(4):573–87. [DOI] [PubMed] [Google Scholar]
- Perera FP, Tang D, Rauh V, Lester K, Tsai WY, Tu YH, et al. Relationships among polycyclic aromatic hydrocarbon-DNA adducts, proximity to the World Trade Center, and effects on fetal growth. Environ Health Perspect 2005c;113(8):1062–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pinkerton KE, Joad JP. Influence of air pollution on respiratory health during perinatal development. Clin Exp Pharmacol Physiol 2006;33(3):269–72. [DOI] [PubMed] [Google Scholar]
- Sanyal MK, Li YL. Deleterious effects of polynuclear aromatic hydrocarbon on blood vascular system of the rat fetus. Birth Defects Res B Dev Reprod Toxicol 2007a;80(5):367–73. [DOI] [PubMed] [Google Scholar]
- Sanyal MK, Li YL. Differential metabolism of benzo[alpha]pyrene in vitro by human placental tissues exposed to active maternal cigarette smoke. Birth Defects Res B Dev Reprod Toxicol 2007b;80(1):49–56. [DOI] [PubMed] [Google Scholar]
- Sanyal MK, Li YL, Belanger K. Metabolism of polynuclear aromatic hydrocarbon in human term placenta influenced by cigarette smoke exposure. Reprod Toxicol 1994;8(5):411–8. [DOI] [PubMed] [Google Scholar]
- Schwartz J. Air pollution and children's health. Pediatrics 2004;113(4 Suppl):1037–43. [PubMed] [Google Scholar]
- Shum S, Jensen NM, Nebert DW. The murine Ah locus: in utero toxicity and teratogenesis associated with genetic differences in benzo[a]pyrene metabolism. Teratology 1979;20(3):365–76. [DOI] [PubMed] [Google Scholar]
- Singh VK, Singh J, Anand M, Kumar P, Patel DK, Krishna Reddy MM, et al. Comparison of polycyclic aromatic hydrocarbon levels in placental tissues of Indian women with full- and preterm deliveries. Int J Hyg Environ Health 2008;211(5–6):639–47. [DOI] [PubMed] [Google Scholar]
- Sram RJ, Binkova B, Rossner P, Rubes J, Topinka J, Dejmek J. Adverse reproductive outcomes from exposure to environmental mutagens. Mutat Res 1999;428(1–2):203–15. [DOI] [PubMed] [Google Scholar]
- Tang D, Li TY, Liu JJ, Chen YH, Qu L, Perera F. PAH-DNA adducts in cord blood and fetal and child development in a Chinese cohort. Environ Health Perspect 2006;114(8):1297–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Topinka J, Milcova A, Libalova H, Novakova Z, Rossner P Jr, Balascak I, et al. Biomarkers of exposure to tobacco smoke and environmental pollutants in mothers and their transplacental transfer to the foetus. Part I: bulky DNA adducts. Mutat Res 2009;669(1–2):13–9. [DOI] [PubMed] [Google Scholar]
- Tuntawiroon J, Mahidol C, Navasumrit P, Autrup H, Ruchirawat M. Increased health risk in Bangkok children exposed to polycyclic aromatic hydrocarbons from traffic-related sources. Carcinogenesis 2007;28(4):816–22. [DOI] [PubMed] [Google Scholar]
- Whyatt RM, Perera FP, Jedrychowski W, Santella RM, Garte S, Bell DA. Association between polycyclic aromatic hydrocarbon-DNA adduct levels in maternal and newborn white blood cells and glutathione S-transferase P1 and CYP1A1 polymorphisms. Cancer Epidemiol Biomarkers Prev 2000;9(2):207–12. [PubMed] [Google Scholar]
- Whyatt RM, Jedrychowski W, Hemminki K, Santella RM, Tsai WY, Yang K, et al. Biomarkers of polycyclic aromatic hydrocarbon-DNA damage and cigarette smoke exposures in paired maternal and newborn blood samples as a measure of differential susceptibility. Cancer Epidemiol Biomarkers Prev 2001;10(6):581–8. [PubMed] [Google Scholar]
- Wu SP, Tao S, Zhang ZH, Lan T, Zuo Q. Distribution of particle-phase hydrocarbons, PAHs and OCPs in Tianjin, China. Atmos Environ 2005;39(38):7420–32. [Google Scholar]
- Zuo Q, Duan YH, Yang Y, Wang XJ, Tao S. Source apportionment of polycyclic aromatic hydrocarbons in surface soil in Tianjin, China. Environ Pollut 2007;147(2):303–10. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.