Postpartum psychological functioning impacts both women’s health and outcomes in children. Lower income, ethnic minority women may be at particular risk for adverse postpartum mental health outcomes. Studies link ambient air pollution exposure with psychological dysfunction in adults although this association has not been examined among postpartum women.
We studied associations between prenatal exposure to particulate matter with diameter ≤ 2.5 μm (PM2.5) and postpartum psychological functioning in a lower income, ethnically mixed sample of urban US women enrolled in a pregnancy cohort study. Analyses included 557 mothers who delivered at ≥37 weeks gestation. Daily estimates of residential PM2.5 over gestation were derived using a satellite-based spatio-temporally resolved model. Outcomes included the Edinburgh Postnatal Depression Scale (EPDS) score from 6 or 12 months postpartum and subscale scores for anhedonia, depressive and anxiety symptoms. Associations were also examined within racial/ethnic groups. Distributed lag models (DLMs) were implemented to identify windows of vulnerability during pregnancy.
Most mothers had less than a high school education (64%) and were primarily Hispanic (55%) and Black (29%). In the overall sample, a DLM adjusted for age, race, education, prenatal smoking, and season of delivery, we found significant associations between higher PM2.5 exposure in the second trimester and increased anhedonia subscale scores postpartum. In race stratified analyses, mid-pregnancy PM2.5 exposure was significantly associated with increased total EPDS scores as well as higher anhedonia and depressive symptom subscale scores among Black women.
Citation: Sheffield PE, Speranza R, Chiu Y-HM, Hsu H-HL, Curtin PC, Renzetti S, et al. (2018) Association between particulate air pollution exposure during pregnancy and postpartum maternal psychological functioning. PLoS ONE 13(4): e0195267. https://doi.org/10.1371/journal.pone.0195267
Editor: Roger A. Coulombe, Utah State University, UNITED STATES
Received: July 6, 2017; Accepted: March 19, 2018; Published: April 18, 2018
Copyright: © 2018 Sheffield et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All of the data underlying the findings of this study are available within the paper and its Supporting Information files.
Funding: The Asthma Coalition on Community, Environment, and Social Stress (ACCESS) project has been funded by the National Institutes of Health [grant numbers R01 ES010932, U01 HL072494, R21ES021318, and R01 HL080674 (Wright RJ, PI)], and phenotyping and biostatistical support was funded by the NIH [grant numbers P30 ES023515 and P30 ES000002]. During preparation of this manuscript, PS was supported by the NIH [grant number K23 ES024127]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Current estimates suggest that 10–20% of U.S. postpartum women experience depressive or anxiety disorders, impacting both women’s health and outcomes in their children.[1,2] Identifying modifiable risk factors for adverse postpartum psychological functioning is thus an important public health focus.
Notably, a diagnosis of postpartum depression includes the presence of either depressed mood reflecting high negative affect or anhedonia reflecting low positive affect.[3,4,5,6] Evidence increasingly demonstrates that while there is some overlap, the underlying neurobiological mechanisms and risk factors may differ between these specific psychiatric symptom profiles. Moreover, certain populations are disproportionately burdened. For example, epidemiological data demonstrate that prevalence rates of anxiety, depressed mood, and anhedonia, are elevated in urban ethnic communities, particularly among Black and Hispanic mothers,[5,7,8] and among women of lower socioeconomic status. Other data suggest that the subjective experience of affective states varies by culture and race/ethnicity. Taken together, these lines of research underscore the need to distinguish specific symptom profiles in order to more comprehensively assess risk, particularly among racially/ethnically diverse postpartum women.
In animals, sub-chronic and longer-term exposure to ambient pollution causes depressive and anxiety-like symptoms in adult mice. [11,12,13] In human non-pregnant adult samples, increased exposure to ambient particulate matter, nitrogen dioxide, and ozone have been associated with adverse neuropsychological functioning, including anxiety and depressive symptoms[14,15,16] and dispensing of psychiatric medication. Studies to date have mostly considered older adults and patients with chronic illness who may be particularly vulnerable. Exposure to ambient pollution in pregnancy can enhance systemic oxidative stress and impact endocrine and immune system disruption which in turn has implications for neuropsychological functioning in women during the perinatal period. No prior study has examined associations between prenatal ambient air pollution exposure and psychological functioning in postpartum women.
Furthermore, the dynamic physiology during pregnancy (e.g., changes in cortisol, other hormones, and neurotransmitters) [22,23] could affect susceptibility to toxic exposures during specific vulnerable windows over the prenatal period. Notably, our group has previously demonstrated that using exposure estimates of ambient air pollution averaged over pregnancy (or by trimester) to examine effects of air pollution on health outcomes reduces the power to detect effects or could result in missed effects altogether.
To begin to address this important research gap, we leveraged an ethnically diverse, lower income urban cohort of pregnant women to examine the association between ambient particulate matter with an aerodynamic diameter ≤2.5 μm (PM2.5) and postpartum psychological functioning. Our approach examines ambient air pollution exposure estimates by each week of pregnancy in order to use data driven methods to identify potential sensitive windows. Doing so allows us to explore potential sensitive windows of exposure since the effects of environmental toxins may vary over the course of the pregnancy. Given recent data demonstrating racial/ethnic differences in the prevalence of specific psychological symptoms (e.g., depressed mood, anhedonia, anxiety), we also examined associations in race stratified analyses.
Materials and methods
These analyses included women enrolled in the Asthma Coalition on Community, Environment, and Social Stress (ACCESS) project, a prospective pregnancy cohort originally established to examine the effects of perinatal stress and other environmental factors on urban childhood asthma risk. Briefly, N = 598 Spanish or English-speaking women receiving prenatal care at two academic medical centers and their affiliated community health centers were recruited from 2002 to 2007. Of these, 557 mothers delivered at ≥ 37 weeks gestation and had data on air pollution exposure and postpartum psychological functioning. Based on screening data, mothers who declined vs. enrolled were slightly less likely to be ethnic minorities (78.2% vs. 80.9% Hispanic or African American, respectively) or to have a high school education or less (57.2% vs. 60.8%, respectively) and slightly more likely to report an income level <20,000 annually (37.1% vs. 35.2%, respectively); there were no significant differences between groups on these covariates. The protocol was approved by the human studies committees at Brigham & Women’s Hospital and Boston Medical Center and written informed consent was obtained in the subjects’ primary language.
Daily prenatal PM2.5 exposure over gestation
Women’s daily exposure to PM2.5 over gestation was estimated based on their residence during pregnancy. This calculation was done using a validated satellite-based spatio-temporal model that predicts daily ambient concentrations of PM2.5, as detailed previously. In brief, the model uses a mixed model approach by regressing daily surface PM2.5 measurements, taken from the U.S. Environmental Protection Agency Air Quality System (AQS) and Interagency Monitoring of Protected Visual Environments Network, with daily aerosol optical depth (AOD) measurements, land-use terms (elevation, distance to major roads, percent open space, point emissions, and area emissions), and meteorological variables (temperature, wind speed, visibility). This PM2.5 model was trained with monitoring data collected by AQS and then validated with 10-fold cross-validation and used to predict daily PM2.5 at 1 km x 1 km resolution allowing for the estimations of PM2.5 to be calculated for each woman’s residential address during her pregnancy. The mean cross-validated R2 for the New England sub-region that includes the greater Boston area included in this study was 0.88. To reduce noise created by the day-to-day PM2.5 variation, women’s gestational exposure estimates were calculated by averaging daily predictions over each week of the pregnancy.
Postpartum psychological functioning
Women completed the 10-item Edinburgh Postnatal Depression Scale (EPDS)  at 6 and 12 months postpartum in a face-to-face interview. Women were asked about symptoms in the past 7 days including: 1: “able to laugh", 2: “looking forward", 3: “self-blaming", 4: “worrying", 5: “scared", 6: “things get on top of me” (overwhelmed), 7: “difficulty sleeping", 8: “feeling sad or miserable", 9: “crying", and 10: “thought of self-harming". Participants rated each item on a Likert scale ranging from never to very often (scored from 0 indicating the most favorable condition to 3 indicating the least favorable condition for each item).
While originally developed as a uni-dimensional scale for depression, a number of studies indicate that the EPDS captures multiple domains [28,29] with possible variation by race/ethnic subgroups.[30,31] As previously reported, the results from an exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) indicated that a three-factor model including depressive, anxiety, and anhedonia symptom subscales was the most optimal fit in our sample as a whole and across racial/ethnic groups with minimal differences. Item 10 “thought of self-harming" was omitted from the subscale analysis given rare frequency of endorsement (<1%). The anhedonia subscale includes items 1 and 2; the anxiety subscale includes items 3, 4, and 5; and the depression subscale includes items 7, 8, and 9. Item 6 (overwhelmed) grouped with anxiety for the Hispanic subgroup but with depressive symptoms for the Black/white/other subgroups, and was used accordingly in our stratified analyses. Values reported on each of these items are summed to determine subscale scores. EPDS scores among women with available data at 6 and 12 months (n = 288) were significantly correlated (rho = 0.44, p = <0.0001). The higher score from assessments at either 6 or 12 months postpartum was used in analyses. Analyses considered the total EPDS score and depressive, anxiety, and anhedonia symptom subscale scores.
Potential confounders previously identified as being related to maternal mental health and air pollution exposure were considered. Women's age, education level (up to high school education or more), and self-reported race/ethnicity (Black, Hispanic, white, other) were ascertained at enrollment. Additionally, we included variables that may co-vary with ambient air pollution exposure, such as prenatal smoking and season of delivery, and that have also been associated with maternal mental health.[33,34] Prenatal smoking status was determined by women’s self-report at enrollment and in the third trimester and was classified as positive if a woman reported smoking at either visit.
Because prenatal air pollution exposure is linked to preterm birth  and preterm birth is linked to postpartum depression, we restricted analyses to n = 557 mothers who delivered at ≥ 37 weeks to minimize over-adjustment for a pathway variable. The primary independent predictor was PM2.5, and the outcomes of separate models included total EPDS score (range 0–30 points) and “anhedonia” (range 0–6 for all subjects), “anxiety” (range 0–9 for Black and white subgroups; range 0–12 for Hispanic subgroup due to an additional item), and “depressive” (range 0–9 for Hispanic subgroup; range 0–12 for Black and white subgroups due to an additional item) symptom subscale scores.
In order to examine the time-varying association between prenatal PM2.5 exposure and postpartum mental health, we applied a data-driven statistical method to identify the sensitive prenatal exposure windows related to postpartum psychological outcomes. Specifically, we fit distributed lag models (DLMs)–an analytic method increasingly being used in pregnancy-related studies that can reduce potential bias from highly temporally resolved exposures (e.g., weekly over gestation) - to estimate the time-varying association between weekly estimated PM2.5 level during pregnancy and the outcomes, as detailed previously. In brief, this method incorporates data from all time points simultaneously and assumes that the association between the exposure and outcome at a given time point varies smoothly as a function of time. We fit the model Yi = β0+∑ [αjAPij]+β1x1i+…+βpxpi, where Yi is the measured EPDS score, APij is the estimated PM2.5 level for week j of pregnancy, and x1i, …, xpi are covariates for subject i. DLMs constrain the αj’s to vary smoothly over time, in our case by using B-splines to characterize the time dependence. The degrees of freedom of the spline were chosen using AIC. A significant sensitive window was identified when the pointwise 95% confidence bands did not contain 0. All models were adjusted for covariates noted previously. Further, to examine race-specific associations, we also conducted DLMs stratified by race/ethnicity. DLMs were implemented using the dlnm package version 2.2.6 in R Studio (Vienna, Austria) and other analyses were performed in SAS (v9.4, SAS Institute Inc., Cary, NC).
Table 1 shows the distribution of demographic characteristics, average PM2.5 exposure over the entire pregnancy, psychological symptom scores, and other covariates for the entire sample as well as by racial/ethnic groups. Most mothers had less than a high school education (64%); their mean age was 25.8 years; they were primarily Hispanic (55%) and Black (29%) and most did not smoke prenatally (86%).
Distributed lag models: Entire sample
Using AIC, we found the best fit from a distributed lag with 4 degrees of freedom. Fig 1 shows the association between prenatal ambient PM2.5 exposure levels and postpartum total EPDS and subscale scores in separate models using DLMs adjusted for race, education, age, prenatal smoking, and season of delivery. We observed a statistically significant sensitive window of PM2.5 exposure (estimated for a 10 μg/m3 increase in PM2.5 exposure) for elevated anhedonia subscale scores during mid-pregnancy, specifically gestational weeks 13 to 20. No statistically significant sensitive window of prenatal PM2.5 exposure was found in relationship to the total EPDS score or the other subscale scores.
This figure demonstrates the association between weekly averaged PM2.5 during pregnancy and postpartum EPDS total and subscale scores using a distributed lag model assuming week-specific effects, adjusting for race, education, age, prenatal smoking status and season of delivery. The y-axis shows the change in EPDS score in relation to a 10 μg/m3 increase in PM2.5 level; the x-axis depicts gestational weeks of the pregnancy. The solid line shows the predicted change, and the gray area indicates the 95% confidence interval. A sensitive window is identified when the estimated pointwise 95% confidence interval does not include 0.
Distributed lag models: Race-stratified analysis
Fig 2 shows the findings from DLMs examining prenatal ambient PM2.5 exposure and maternal postpartum EPDS total and subscale scores, stratified by race/ethnicity. In the Black subgroup, we found significant windows of PM2.5 exposure (estimated for a 10 μg/m3 increase in PM2.5 exposure) at 10–24 weeks gestation with anhedonia subscale scores, at 18–26 weeks gestation with depression subscale scores, and at 16–26 weeks gestation with EPDS total scores. The effects in Whites showed a similar, but not significant pattern. There was no significant pattern in Hispanic women; in addition, the shape of the distributed lag plot differed from that of Black and white women.
Within each stratum, this figure demonstrates the association between PM2.5 exposure during pregnancy and postpartum EPDS total and subscale scores using a distributed lag model assuming week-specific effects, adjusting for education, age, prenatal smoking status and season of delivery. The y-axis shows the change in EPDS score in relation to a 10 μg/m3 increase in PM2.5 level; the x-axis depicts gestational weeks of the pregnancy. The solid line shows the predicted score change, and the gray area indicates the 95% confidence interval. A sensitive window is identified when the estimated pointwise 95% confidence interval does not include 0.
This is the first study to examine associations between exposure to ambient fine particulate matter during pregnancy and maternal psychological functioning in the first year postpartum. We found associations between mid-pregnancy ambient fine particulate matter exposure and increased postpartum anhedonia symptoms when we considered the sample of ethnically diverse, lower income, urban mothers as a whole. In stratified analyses, effects were most evident in Blacks. Among Black women, increased exposure to PM2.5 was significantly associated with higher total postpartum EPDS scores as well as higher scores on the depressive and anhedonia symptom subscales with minor variability in the identified window of vulnerability in pregnancy across these outcome measures.
Notably, we leveraged weekly address-specific PM2.5 exposure estimates and DLMs in order to explore this concept of windows of vulnerability for maternal health during the prenatal period. Significant associations were observed primarily in the second trimester. Critical timing of exposure for environmental factors, or windows of vulnerability to exposures during pregnancy in relationship to later maternal health outcomes remains minimally explored and even the theoretical mechanism needs better refinement. A more definitive characterization of vulnerable windows may provide insight into underlying mechanisms when coupled with our understanding of psychopathology in the perinatal period and environmentally induced physiological changes that may underlie observed associations. Mechanisms that may underlie these associations were beyond the scope of the current analysis and will need to be examined in future research.
Additionally, a better understanding of risk and resilience factors is need to further explore the different patterns of associations between air pollution and psychological functioning by racial/ethnic groups. Among the factors to further explore are perinatal factors such as method of delivery, postpartum factors such as breast feeding, as well as larger community and societal factors such as religiosity, partner and larger social support networks, and stress. The pattern of association between PM2.5 and EPDS scores observed among Hispanic women differed from the patterns observed among Black and white women. This difference further supports the idea of different phenotypes of psychological functioning among different racial/ethnic groups and also suggests potentially different resilience factors such as social support, community expectations, and other protective factors that warrant further exploration.
Of the three subscales examined, only anhedonia and depressive symptoms were associated with increased exposure to particulate pollution in this study. This association is consistent with the emerging understanding of the role of inflammation in mediating anhedonic behaviors via impacts on neurotransmitter synthesis and accumulation of neurotoxic compounds. However, anxiety has also been linked to inflammatory cytokines and other markers of immune imbalance. In other research, prenatal exposure to particulate air pollution has been associated with decreased gene expression in brain-derived neurotrophic factor (BDNF)  and reduced levels of maternal BDNF prenatally has been implicated as a biomarker of postpartum depression. Future research should examine underlying mechanisms beyond inflammation (e.g., oxidative stress, altered gene expression). Moreover, it has been pointed out that symptoms of anhedonia may be commonly overlooked in clinical evaluations in comparison with more overt symptomatology such as depressed mood. Thus, consideration of these different presentations has implications both in research and in the clinical setting if we are to understand the full impact of environmental exposures on postpartum psychological functioning in women. A more nuanced understanding of the specific neuropsychological sequelae of air pollution remains an area of in need of increased investigation.
There are a number of notable strengths of the current analysis. These include the reasonably large sample of lower income ethnically mixed inner-city mothers, the high spatial resolution of the particulate matter exposure estimates coupled with the advanced statistical modeling that allowed us to use data driven methods to identify windows of vulnerability, and our broader assessment of psychological functioning using the EPDS subscales. We also acknowledge some limitations. The exposure estimate was unable to account for the actual amount of time a woman spent at her residence and while ambient air quality is a strong driver of indoor air quality, indoor PM2.5 concentrations may have differed resulting in exposure misclassification. Given the growing evidence that indoor air quality is often worse than outdoor air quality, our ambient air exposure estimates are likely conservative. We hypothesize that any misclassification is non-differential and would thus bias our results toward the null. However, indoor air quality is an important and potentially worsening exposure given the changing climate and should be the focus of further study to develop appropriate interventions. While we adjusted for a number of known confounders of the association between ambient particulate matter and maternal mental health, there may be unmeasured confounders we did not consider. For example, we were unable to adjust for pollutants other than PM2.5 or other concomitant chemical exposures in these data, specifically ones that may be correlated with PM2.5 exposure and on a causal pathway for maternal psychological functioning during the postpartum period. Additionally, while limiting to term births, we were not able to control for other birth outcomes or infant-specific health issues which could potentially be on the causal pathway between in-utero air pollutant exposures and maternal psychological functioning. Specific experiences that may differ based on race/ethnicity, for example racism, may moderate the effects of chemical pollutants on mental health outcomes which also may explain differences across racial groups. Future research in sociodemographically diverse samples should consider these additional factors with larger sample sizes to facilitate more formal examination of air pollution mixtures (multiple pollutants) as well as modifying factors. These findings may not be generalizable to populations with differing sociodemographic makeup nor directly comparable to studies that consider average air pollution exposures over the entire or trimesters of the pregnancy. However, our DLM analytic approach reduces the risk of potential bias or missed effects altogether when exposure windows are arbitrarily assigned. 
In summary, this study objectively elucidated sensitive prenatal windows and examined racial/ethnic differences for the association between PM2.5 during pregnancy and maternal postpartum psychological functioning. Our findings suggest that exposure to particulate air pollution during specific windows of pregnancy may be associated with higher risk of negative postpartum mental health outcomes in women–particularly anhedonia and depressive symptoms. Black women may be at higher risk. While this study adds to the growing literature linking ambient air pollution with mental health outcomes in adults, more studies are needed to replicate these findings in other populations, particularly with respect to potential windows of vulnerability for the mother during pregnancy, and to further elucidate underlying mechanisms.
- 1. Pawluski JL, Lonstein JS, Fleming AS. The Neurobiology of Postpartum Anxiety and Depression. Trends Neurosci. 2017 Feb;40(2):106–120. pmid:28129895.
- 2. Stewart DE, Vigod S. Postpartum Depression. N Engl J Med. 2016 Dec 1;375(22):2177–2186. pmid:27959754.
- 3. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing; 2013.
- 4. Treadway MT, Zald DH. Reconsidering anhedonia in depression: lessons from translational neuroscience. Neurosci Biobehav Rev. 2011;35:537–55. pmid:20603146
- 5. Liu CH, Tronick E. Prevalence and predictors of maternal postpartum depressed mood and anhedonia by race and ethnicity. Epidemiol Psychiatr Sci. 2014;23:201–9. pmid:23931673
- 6. Sibitz I, Berger P, Freidl M, Topitz A, Krautgartner M, Spiegel W, et al. ICD-10 or DSM-IV? Anhedonia, fatigue and depressed mood as screening symptoms for diagnosing a current depressive episode in physically ill patients in general hospital. J Affect Disord. 2010;126:245–51. pmid:20400184
- 7. Howell EA, Mora PA, Horowitz CR, Leventhal H. Racial and ethnic differences in factors associated with early postpartum depressive symptoms. Am Coll Ob Gyn. 2005;105:1442–50. pmid:15932842.
- 8. Clark C, Ryan L, Kawachi I, Canner MJ, Berkman L, Wright RJ. Witnessing community violence in residential neighborhoods: a mental health hazard for urban women. J Urban Health. 2008;85:22–38. pmid:17965940.
- 9. Rich-Edwards J, Kleinman K, Abrams A, Harlow BL, McLaughlin TJ, Joffe H, et al. Sociodemographic predictors of antenatal and postpartum depressive symptoms among women in a medical group practice. J Epidemiol Community Health. 2006 Mar;60(3):221–7. pmid:16476752.
- 10. Kanazawa A, White PM, Hampson SE. Ethnic variation in depressive symptoms in a community sample in Hawaii. Cultur Divers Ethnic Minor Psychol. 2007;13:35–44. pmid:17227175
- 11. Costa LG, Cole TB, Coburn J, Chang YC, Dao K, Roque P. Neurotoxicants are in the air: convergence of human, animal, and in vitro studies on the effects of air pollution on the brain. Biomed Res Int. 2014;2014:736385. pmid:24524086.
- 12. Mokoena ML, Harvey BH, Viljoen F, Ellis SM, Brink CB. Ozone exposure of Flinders Sensitive Line rats is a rodent translational model of neurobiological oxidative stress with relevance for depression and antidepressant response. Psychopharmacology (Berl). 2015 Aug;232(16):2921–38. Epub 2015 Apr 17. pmid:25877744.
- 13. Zhang X, Xu Y, Zhou L, Zhang C, Meng Q, Wu S, et al. Sex-Dependent Depression-Like Behavior Induced by Respiratory Administration of Aluminum Oxide Nanoparticles. Int J Environ Res Public Health. 2015 Dec 9;12(12):15692–705. pmid:26690197.
- 14. Pun VC, Manjourides J, Suh H. Association of Ambient Air Pollution with Depressive and Anxiety Symptoms in Older Adults: Results from the NSHAP Study. Environ Health Perspect. 2017 Mar;125(3):342–348. Epub 2016 Aug 12. pmid:27517877.
- 15. Power MC, Kioumourtzoglou MA, Hart JE, Okereke OI, Laden F, Weisskopf MG. The relation between past exposure to fine particulate air pollution and prevalent anxiety: observational cohort study. BMJ. 2015 Mar 24;350:h1111. pmid:25810495.
- 16. Lim YH, Kim H, Kim JH, Bae S, Park HY, Hong YC. Air pollution and symptoms of depression in elderly adults. Environ Health Perspect. 2012 Jul;120(7):1023–8. Epub 2012 Apr 18. pmid:22514209.
- 17. Oudin A, Bråbäck L, Åström DO, Strömgren M, Forsberg B. Association between neighbourhood air pollution concentrations and dispensed medication for psychiatric disorders in a large longitudinal cohort of Swedish children and adolescents. BMJ Open. 2016 Jun 3;6(6):e010004. pmid:27259522.
- 18. Cho J, Choi YJ, Suh M, Sohn J, Kim H, Cho SK, et al. Air pollution as a risk factor for depressive episode in patients with cardiovascular disease, diabetes mellitus, or asthma. J Affect Disord. 2014 Mar;157:45–51. Epub 2014 Jan 10. pmid:24581827.
- 19. Grevendonk L, Janssen BG, Vanpoucke C, Lefebvre W, Hoxha M, Bollati V, et al. Mitochondrial oxidative DNA damage and exposure to particulate air pollution in mother-newborn pairs. Environ Health. 2016 Jan 20;15:10. pmid:26792633.
- 20. Rier SE. Environmental immune disruption: a comorbidity factor for reproduction? Fertil Steril. 2008;89:e103–8. pmid:18308048
- 21. Posillico CK, Schwarz JM. An investigation into the effects of antenatal stressors on the postpartum neuroimmune profile and depressive-like behaviors. Behav Brain Res. 2016 Feb 1;298(Pt B):218–28. Epub 2015 Nov 14. pmid:26589802.
- 22. Entringer S, Buss C, Shirtcliff EA, Cammack AL, Yim IS, Chicz-DeMet A, Sandman CA, Wadhwa PD. Attenuation of maternal psychophysiological stress responses and the maternal cortisol awakening response over the course of human pregnancy. Stress. 2010 May;13(3):258–68. pmid:20067400
- 23. Weetman AP. Immunity, thyroid function and pregnancy: molecular mechanisms. Nat Rev Endocrinol. 2010 Jun;6(6):311–8. Epub 2010 Apr 27. pmid:20421883.
- 24. Wilson A, Chiu YM, Hsu HL, Wright RO, Wright RJ, Coull BA. Potential for Bias When Estimating Critical Windows for Air Pollution in Children's Health. Am J Epidemiol. 2017 Dec 1;186(11):1281–1289. pmid:29206986.
- 25. Wright RJ, Suglia SF, Levy J, Fortun K, Shields A, Subramanian S, et al. Transdisciplinary research strategies for understanding socially patterned disease: the Asthma Coalition on Community, Environment, and Social Stress (ACCESS) project as a case study. Cien Saude Colet. 2008 Nov-Dec;13(6):1729–42. pmid:18833350.
- 26. Kloog I, Chudnovsky AA, Just AC, Nordio F, Koutrakis P, Coull BA, et al. A new hybrid spatio-temporal model for estimating daily multi-year PM2.5 concentrations across northeastern USA using high resolution aerosol optical depth data. Atmos Environ. 2014;95:581–90.
- 27. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry. 1987;150:782–86. pmid:3651732
- 28. Cunningham NK, Brown PM, Page AC. Does the Edinburgh Postnatal Depression Scale measure the same constructs across time? Arch Womens Ment Health. 2015;18:793–804. pmid:25510935
- 29. Coates R, Ayers S, de Visser R. Factor Structure of the Edinburgh Postnatal Depression Scale in a Population-Based Sample. Psychol Assess. Epub 2016 Oct 13. pmid:27736124.
- 30. Hartley CM, Barroso N, Rey Y, Pettit JW, Bagner DM. Factor structure and psychometric properties of english and spanish versions of the Edinburgh postnatal depression scale among Hispanic women in a primary care setting. J Clin Psychol. 2014 Dec;70(12):1240–50. Epub 2014 May 7. pmid:24807217.
- 31. Lee King PA. Replicability of structural models of the Edinburgh Postnatal Depression Scale (EPDS) in a community sample of postpartum African American women with low socioeconomic status. Arch Womens Ment Health. 2012 Apr;15(2):77–86. pmid:22297555.
- 32. Chiu YH, Sheffield PE, Hsu HH, Curtin P, Goldstein J, Wright RJ. Subconstructs of the Edinburgh Postpartum Depression Scale in a Multi-Ethnic Inner-City Population in the U.S. Arch Womens Ment Health. 2017 2017 Aug 2. [Epub ahead of print] pmid:28770342.
- 33. Salimi S, Terplan M, Cheng D, Chisolm MS. The Relationship Between Postpartum Depression and Perinatal Cigarette Smoking: An Analysis of PRAMS Data. J Subst Abuse Treat. 2015 Sep;56:34–8. Epub 2015 Mar 25. pmid:25841705.
- 34. Sylvén SM, Papadopoulos FC, Olovsson M, Ekselius L, Poromaa IS, Skalkidou A. Seasonality patterns in postpartum depression. Am J Obstet Gynecol. 2011 May;204(5):413.e1-6. Epub 2011 Mar 24. pmid:21439544.
- 35. Backes CH, Nelin T, Gorr MW, Wold LE. Early life exposure to air pollution: how bad is it? Toxicol Lett. 2013 Jan 10;216(1):47–53. Epub 2012 Nov 16. pmid:23164674
- 36. Vigod SN, Villegas L, Dennis CL, Ross LE. Prevalence and risk factors for postpartum depression among women with preterm and low-birth-weight infants: a systematic review. BJOG. 2010 Apr;117(5):540–50. Epub 2010 Jan 29. pmid:20121831
- 37. Hsu HH, Chiu YH, Coull BA, Kloog I, Schwartz J, Lee A, et al. Prenatal Particulate Air Pollution and Asthma Onset in Urban Children. Identifying Sensitive Windows and Sex Differences. Am J Respir Crit Care Med. 2015 Nov 1;192(9):1052–9. pmid:26176842.
- 38. Gasparrini A. Distributed Lag Linear and Non-Linear Models in R: The Package dlnm. J Stat Softw. 2011;43:1–20.
- 39. Pan Z, Rosenblat JD, Swardfager W, McIntyre RS. Role of proinflammatory cytokines in dopaminergic system disturbances, implications for anhedonic features of MDD. Curr Pharm Des. Epub 2017 January 11.
- 40. Hou R, Garner M, Holmes C, Osmond C, Teeling J, Lau L, et al. Peripheral inflammatory cytokines and immune balance in Generalised Anxiety Disorder: Case-controlled study. Brain Behav Immun 2017 May;62:212–218. Epub 2017 Feb 1. pmid:28161475
- 41. Saenen ND, Plusquin M, Bijnens E, Janssen BG, Gyselaers W, Cox B, et al. In Utero Fine Particle Air Pollution and Placental Expression of Genes in the Brain-Derived Neurotrophic Factor Signaling Pathway: An ENVIRONAGE Birth Cohort Study. Environ Health Perspect. 2015 Aug;123(8):834–40. Epub 2015 Mar 27. pmid:25816123.
- 42. Fung J, Gelaye B, Zhong QY, Rondon MB, Sanchez SE, Barrios YV, et al. Association of decreased serum brain-derived neurotrophic factor (BDNF) concentrations in early pregnancy with antepartum depression. BMC Psychiatry. 2015 Mar 10;15:43. pmid:25886523.
- 43. IOM (Institute of Medicine). 2011. Climate Change, the Indoor Environment, and Health. Washington, DC: The National Academies Press.
- 44. Paradies Y, Ben J, Denson N, Elias A, Priest N, Pieterse A, et al. Racism as a Determinant of Health: A Systematic Review and Meta-Analysis. PLoS One. 2015 Sep 23;10(9):e0138511. pmid:26398658