Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of maternal depression and prenatal antidepressant exposure on child development : consideration… Park, Mina 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2019_september_park_mina.pdf [ 1.53MB ]
Metadata
JSON: 24-1.0379698.json
JSON-LD: 24-1.0379698-ld.json
RDF/XML (Pretty): 24-1.0379698-rdf.xml
RDF/JSON: 24-1.0379698-rdf.json
Turtle: 24-1.0379698-turtle.txt
N-Triples: 24-1.0379698-rdf-ntriples.txt
Original Record: 24-1.0379698-source.json
Full Text
24-1.0379698-fulltext.txt
Citation
24-1.0379698.ris

Full Text

THE IMPACT OF MATERNAL DEPRESSION AND PRENATAL ANTIDEPRESSANT EXPOSURE ON CHILD DEVELOPMENT: CONSIDERATION OF INTERNAL AND EXTERNAL CONTEXTUAL FACTORS  by Mina Park  B.A&Sc., McGill University, 2009 M.Sc., University of Toronto, 2012  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Population and Public Health)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2019 © Mina Park, 2019 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  The impact of maternal depression and prenatal antidepressant exposure on child development: Consideration of internal and external contextual factors  submitted by Mina Park in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Population and Public Health  Examining Committee: Tim Oberlander Supervisor  Michael Brauer Supervisory Committee Member  Jennifer Gardy Supervisory Committee Member KS Joseph University Examiner Evelyn Stewart University Examiner  Additional Supervisory Committee Members: Patricia Janssen Supervisory Committee Member iii  Abstract One in five women experience a depressive episode during pregnancy, with up to one in ten being treated with antidepressants. Despite an overall adverse developmental effect related to exposure to maternal depression and antidepressants during pregnancy, not all children are equally affected. Moreover, critical yet unanswered questions remain regarding the potential developmental risks and benefits of antidepressant treatment during pregnancy. The purpose of this thesis was to study the association between maternal depression and prenatal antidepressant exposure and child developmental outcomes while considering the role of key contextual factors that may influence these associations.   Depression severity, patterns of depressive symptoms over time, and underlying genetic risk were investigated for their role in influencing child outcomes. Small yet consistent associations were found between prenatal antidepressant exposure and vulnerability for anxiety and lower physical independence at kindergarten age, after stringently adjusting for confounding by maternal mental health. Genetic differences were found to underlay associations between prenatal maternal depression symptom levels and genome-wide differences in DNA methylation at age 18 years. Looking longitudinally, increasing maternal depression symptoms across the first three postpartum years were associated with worse child behavioral problems and executive functions at ages three and six years; a pattern of decreasing maternal depression symptoms over the same time period, despite initially higher depressive symptoms during pregnancy, was associated with no impairments.  iv  Collectively, the findings presented herein help explain the heterogeneity of child outcomes that are observed in relation to early developmental exposure to maternal depression and its treatment with antidepressants. This thesis highlights the impact of factors underlying and related to maternal mental health on children’s developmental outcomes, and suggest that investments made to improve maternal depression have the potential to benefit the health of both mothers and their children.   v  Lay Summary Maternal depression and prenatal antidepressant exposure have been associated with adverse child outcomes. However, not all children are equally affected by these exposures. This thesis investigates key factors that may influence the impact of these exposures on child development. Results from this body of work suggest that variations in maternal depression severity, patterns of maternal depression over time, and genetic differences associated with hereditary risk for depression may underlie variations in child developmental outcomes. vi  Preface This dissertation is my original work. The analyses reported herein are covered by UBC Research Ethics Certificate numbers (H10-03354, H08-01712). I was the lead investigator for all analyses, and responsible for the design of the research program, data analysis, and manuscript composition.   Chapters 2, 3, and 5 of the dissertation are each composed of manuscripts, which have either been published or submitted for publication in peer-reviewed journals. I was responsible for developing initial study designs and analytic approaches. I received guidance and assistance from my thesis supervisor, Dr. TF Oberlander, for these chapters. My research collaborators, Drs. GE Hanley and M Guhn, contributed to Chapter 2, and U Brain, Dr. RE Grunau, and Dr. A Diamond contributed to Chapter 5. I conducted all analyses, wrote all of the code for statistical analyses, and produced first drafts of all manuscripts. My supervisor and research collaborators made contributions to the study design, analysis and interpretation of data, and revised each article for intellectual content.   Research for Chapter 4 was conducted in the laboratory of Dr. MS Kobor. Chapter 4 was produced with additional input from my supervisor as well as research collaborators (S Goodman, J Armstrong, LM McEwen, J MacIsaac and Drs. M Essex and WT Boyce). I had a primary role in running wet-lab experiments, designed and conducted all analyses, and produced the final draft of this chapter. Dr. MS Kobor contributed to study design and data interpretation, and with Dr. M Essex, provided feedback on a previous draft of a manuscript from which this vii  chapter was adapted. Although not directly involved with this chapter, Dr. C Hertzman (deceased) was pivotal in shaping the research direction for this work.  Committee members, Drs. M Brauer and J Gardy, provided assistance with overall research directions, analytic approaches, and interpretation of results. Committee member Dr. Janssen helped assess the complete dissertation.   A version of Chapter 2 has been published in a peer-reviewed journal. Park M, Oberlander TF.  In-utero Selective Serotonin Reuptake Inhibitor Antidepressant Exposure: Fetal Programing and Developmental Interactions With Context. Clinical Pharmacology and Therapeutics, 2018 Oct; 104(4): 616-618.  A version of Chapter 3 has been submitted to a peer-reviewed journal. Park M, Hanley GE, Guhn M, Oberlander TF. Prenatal antidepressant exposure and child development at kindergarten age: a population-based retrospective cohort study.  A version of Chapter 5 has been published in a peer-reviewed journal. Park M, Brain U, Grunau RE, Diamond A, Oberlander TF. Maternal depression trajectories from pregnancy to 3 years postpartum are associated with children's behavior and executive functions at 3 and 6 years. Archives of Women’s Mental Health, 2018 Jun; 21(3):353-363.    viii  Table of Contents  Abstract ......................................................................................................................................... iii Lay Summary .................................................................................................................................v Preface ........................................................................................................................................... vi Table of Contents ....................................................................................................................... viii List of Tables .............................................................................................................................. xiii List of Figures ...............................................................................................................................xv List of Symbols .......................................................................................................................... xvii List of Abbreviations ............................................................................................................... xviii Acknowledgements .................................................................................................................... xix Dedication ................................................................................................................................... xxi Chapter 1: Introduction ................................................................................................................1 1.1 Maternal depression during pregnancy ........................................................................... 1 1.2 Antidepressant treatment during pregnancy ................................................................... 2 1.3 Associations between maternal depression and prenatal antidepressant exposure on child development ....................................................................................................................... 3 1.4 Study rationale ................................................................................................................ 4 1.4.1 Factors influencing the impact of maternal depression and prenatal antidepressant exposure on child development .............................................................................................. 5 1.5 Conceptual framework .................................................................................................... 6 1.6 Dissertation objectives .................................................................................................... 7 1.7 Data sources .................................................................................................................... 8 ix  1.8 Dissertation structure .................................................................................................... 10 Chapter 2: In-utero SSRI antidepressant exposure: Fetal programing via developmental interactions with context .............................................................................................................11 2.1 Synopsis ........................................................................................................................ 11 2.2 Prenatal SSRI exposure and developmental programing.............................................. 11 2.3 Altered serotonin signaling ........................................................................................... 12 2.4 Prenatal SSRI exposure: A differential susceptibility .................................................. 13 2.5 Maternal prenatal and postnatal context ....................................................................... 15 2.6 Conclusion .................................................................................................................... 16 Chapter 3: Prenatal antidepressant exposure and child development at kindergarten age .18 3.1 Synopsis ........................................................................................................................ 18 3.2 Introduction ................................................................................................................... 19 3.3 Methods......................................................................................................................... 21 3.3.1 Study design and data sources .................................................................................. 21 3.3.2 Study cohort .............................................................................................................. 22 3.3.3 Exposure classification ............................................................................................. 23 3.3.4 Outcome measures .................................................................................................... 23 3.3.5 High-dimensional propensity score .......................................................................... 24 3.3.6 Main analyses............................................................................................................ 25 3.3.7 Sensitivity analyses ................................................................................................... 26 3.4 Results ........................................................................................................................... 26 3.4.1 Demographic characteristics of study cohort ............................................................ 26 3.4.2 Results from main analyses ...................................................................................... 31 x  3.4.3 Results from sensitivity analyses .............................................................................. 33 3.5 Discussion ..................................................................................................................... 36 3.5.1 Comparison with previous findings .......................................................................... 36 3.5.2 Strengths and limitations........................................................................................... 38 3.5.3 Conclusions ............................................................................................................... 40 Chapter 4: Prenatal maternal depression symptomatology, genetic background, and DNA methylation patterns in adolescence...........................................................................................41 4.1 Synopsis ........................................................................................................................ 41 4.2 Introduction ................................................................................................................... 42 4.3 Methods......................................................................................................................... 44 4.3.1 Study design and study sample ................................................................................. 44 4.3.2 Measures ................................................................................................................... 45 4.3.2.1 Maternal depression .......................................................................................... 45 4.3.2.2 Demographic variables ..................................................................................... 46 4.3.2.3 Genetic ancestry ................................................................................................ 46 4.3.2.4 Outcomes at 18 years ........................................................................................ 47 4.3.3 DNA methylation data .............................................................................................. 47 4.3.4 Genotype data ........................................................................................................... 49 4.3.5 Pyrosequencing ......................................................................................................... 50 4.3.6 Statistical methods .................................................................................................... 50 4.3.6.1 Model building .................................................................................................. 50 4.3.6.2 Differential DNA methylation analyses............................................................ 52 4.3.6.3 Testing for effects of postnatal maternal depression ........................................ 52 xi  4.3.6.4 Testing for effects of genotype ......................................................................... 52 4.4 Results ........................................................................................................................... 53 4.4.1 Differentially methylated regions associated uniquely with exposure to prenatal maternal depression .............................................................................................................. 53 4.4.2 Differentially methylated regions showed unique associations with prenatal, and not postnatal, maternal depression levels .................................................................................... 59 4.4.3 Genotype influenced differentially methylated regions ............................................ 59 4.5 Discussion ..................................................................................................................... 60 Chapter 5: Maternal depression trajectories from pregnancy to three years postpartum and their associations with children’s behavior and executive functions in childhood .........63 5.1 Synopsis ........................................................................................................................ 63 5.2 Introduction ................................................................................................................... 64 5.3 Methods......................................................................................................................... 66 5.3.1 Study design .............................................................................................................. 66 5.3.2 Measures ................................................................................................................... 67 5.3.2.1 Maternal depression .......................................................................................... 67 5.3.2.2 Child behavior at three years ............................................................................ 68 5.3.2.3 Child mental health symptomatology at six years ............................................ 68 5.3.2.4 Child executive functions at six years .............................................................. 68 5.3.3 Cohort characteristics................................................................................................ 69 5.3.4 Statistical analyses .................................................................................................... 70 5.3.4.1 Analytic samples ............................................................................................... 70 5.3.4.2 Data imputation ................................................................................................. 71 xii  5.3.4.3 Growth mixture modelling ................................................................................ 71 5.3.4.4 Identifying factors associated with maternal depressive symptom trajectories 72 5.3.4.5 Multivariable linear regression analyses ........................................................... 72 5.3.4.6 Sensitivity analysis............................................................................................ 73 5.4 Results ........................................................................................................................... 73 5.4.1 Identification of maternal depressive symptom trajectories ..................................... 73 5.4.2 Maternal characteristics by trajectory group ............................................................. 75 5.4.3 Child characteristics by mother’s depressive symptom trajectory ........................... 78 5.4.4 Maternal depressive symptom trajectories and child behavior at three years .......... 80 5.4.5 Maternal depressive symptom trajectories, child behavior, and executive functions at 6 years 82 5.5 Discussion ..................................................................................................................... 83 Chapter 6: Discussion ..................................................................................................................89 6.1 Summary of findings..................................................................................................... 89 6.2 Strengths and limitations............................................................................................... 91 6.3 Significance and implications of thesis research .......................................................... 92 6.4 Future directions ........................................................................................................... 94 6.5 Conclusion .................................................................................................................... 95 References .....................................................................................................................................97 Appendix .....................................................................................................................................109  xiii  List of Tables  Table 3.1. Characteristics of mother-child dyads in the overall cohort and in a subset of exposed children matched to unexposed children using a high dimensional propensity score, by prenatal antidepressant exposure. All results are presented as n (%) unless otherwise stated. .................. 28 Table 3.2. Associations between prenatal antidepressant exposure and child development. Comparison groups were exposed children against unexposed children in the overall cohort, exposed children matched to unexposed children using a high dimensional propensity score, and children whose mothers had antidepressant dispensations in the 90 – 365 days prior to conception and continued versus discontinued treatment during pregnancy................................ 32 Table 3.3. Associations between prenatal antidepressant exposure and child development. Antidepressant exposure was restricted to pregnancies where exposure had to include the 2nd and/or 3rd trimesters. Comparison groups were exposed children matched to unexposed children using a high dimensional propensity score, and children whose mothers had antidepressant dispensations in the 90 – 365 days prior to conception and continued versus discontinued treatment during pregnancy. ......................................................................................................... 34 Table 4.1. Characteristics of mother-child dyads (n=178) in the present study........................... 44 Table 4.2. Regions of DNA methylation significantly associated with prenatal maternal depression symptom levels. .......................................................................................................... 54 Table 4.3. Overview of methylation quantitative trait loci within differentially methylated regions. .......................................................................................................................................... 60 Table 5.1. Depression scores by trajectory group from the 2nd trimester of pregnancy until three years postpartum. .......................................................................................................................... 74 xiv  Table 5.2. Maternal characteristics by depression trajectory group (n = 147). ............................ 76 Table 5.3. Child characteristics and outcomes by maternal depression trajectory group (n = 103)........................................................................................................................................................ 79 Table 5.4. Linear regression models estimating the effect of maternal depression trajectories on child outcomes at ages three and six years. .................................................................................. 81  xv  List of Figures  Figure 2.1. Diagram illustrating an integrative and contextual approach to understanding the impact of prenatal SSRI exposure on child developmental outcomes. Prenatal SSRI exposure may play a role in shaping early human brain development that sets pathways for both developmental risk and resiliency across childhood. However prenatal SSRI exposure may shape developmental variations via interactions with pharmacological, genetic and social-contextual factors that occur across prenatal and postnatal periods. .............................................................. 12 Figure 3.1. Flowchart detailing definition of comparison groups from the overall study cohort. Antidepressant use during pregnancy was defined as at least 2+ prescriptions during pregnancy or at least 1+ prescription during the 2nd or 3rd trimesters. Antidepressant use pre-conception was defined as any prescription in the 90 – 365 days before pregnancy. ............................................ 22 Figure 4.1. Scree plot showing the proportion of variance explained by the first 20 principal components of genotype data........................................................................................................ 47 Figure 4.2. PC1 scores plotted against PC2 scores per sample for the combined buccal and PBMC DNA methylation dataset.................................................................................................. 49 Figure 4.3. Scree plot showing the proportion of variance explained by the first 20 principal components of the DNA methylation data. ................................................................................... 51 Figure 4.4. Heat maps showing associations of potential confounding variables with each other and with the first 9 principal components of the DNA methylation data. .................................... 51 Figure 4.5. Histogram showing the expected number of overlapping probes between the most significant 350 CpG sites from probe-wise analyses and the 350 probes underlying DMRs, based on 10,000 simulations. The dotted red line represents the actual number of overlapping probes. 56 xvi  Figure 4.6. Regions of DNA methylation significantly associated with prenatal maternal depression exposure at 18 years of age. a) Plots showing the top 6 most statistically significant differentially methylated regions. Circles show each DNA methylation value per individual at each probe. For the purposes of regional visualization, individuals were divided into “high” or “low” prenatal maternal depression exposure based on a CES-D score of 16. Solid lines represent mean DNA methylation levels per group at each probe. * indicates sites that were represented in pyrosequencing experiments. b) Pyrosequencing verified DNA methylation levels in the CALCB gene locus. DNA methylation values against prenatal maternal depression scores are shown from pyrosequencing experiments (top) and from the 450K array (bottom). Solid blue lines represent unadjusted regression lines. .......................................................................................................... 58 Figure 5.1. Maternal depression trajectory groups. Trajectories were visualized using a locally weighted regression smoothing approach. Shaded bars represent 95% confidence intervals. ..... 74  xvii  List of Symbols 5HTT Serotonin transporter CALCB Calcitonin related polypeptide, beta HOXA2 Homeobox protein A2 HOXA5 Homeobox protein A5 SLC19A1 Solute carrier family 19 member 1 SETD6 SET domain containing 6  xviii  List of Abbreviations 450K Illumina Infinium HumanMethylation450 BeadChip array ADHD Attention-deficit/hyperactivity disorder ANOVA Analysis of Variance ATC Anatomical Therapeutic Classification BDI Beck Depression Inventory BEC Buccal epithelial cell BIC Bayesian information criterion BRIEF Behavior Rating Inventory of Executive Function CBCL Child Behavior Checklist CES-D Centre for Epidemiological Studies – Depression Scale CI Confidence interval CpG Cytosine-guanine dinucleotide unit DMR Differentially methylated region DNAm DNA methylation DSM Diagnostic and Statistical Manual EDI Early Development Instrument EF Executive functions EPDS Edinburgh Postnatal Depression Scale ESL English as a second language EWAS Epigenome-wide association study FDR False discovery rate GMM Growth mixture modelling HAMD Hamilton Depression scale HBQ Health and Behavior Questionnaire HDPS High-dimensional propensity score HF Hearts and Flowers ICD International Classification of Diseases meQTL Methylation quantitative trait loci OR Odds ratio PBMC Peripheral blood mononuclear cells PC Principal component PCA Principal component analysis PsychChip Illumina PsychChip SNP genotyping array SNP Single nucleotide polymorphism SSRI Selective serotonin reuptake inhibitor WSFW Wisconsin Study of Families and Work  xix  Acknowledgements I am grateful for the many people who have made this possible.  First, to my family. My parents were possessed of a prescience I will never fully understand and brought our family over to a new country to pursue unknown opportunities untold. Their struggle is what has allowed me to pursue this privileged path. I am and ever will be indebted to their sacrifice, love, and devotion. Throughout, I have been pushed along by the memory and histories of my grandmothers who impressed upon me the extent of this opportunity, one that the women of their generation never had. My siblings, my constant companions since the beginning – your stories are what motivated me to try to reach a deeper understanding, and to try to make a difference that matters.  Foremost, to my supervisor and mentor (for they are not always one and the same), who taught me the importance of scholarship but also through personal example, taught me deep lessons about integrity, compassion, and purpose. My mind has been indelibly shaped by his brilliance, and my horizons expanded beyond what I knew to be possible. Through him, I carry forward the awesome responsibility that it is to be a steward of knowledge. Collectively, to the community of scholars who have guided me along, thank you for your support and intervention at critical junctures in my journey. And to my friends and peers, thank you for your camaraderie, encouragement, and solidarity through it all. Finally, I am grateful for the experience of this PhD itself. It is through this pursuit that I was able to emerge from a time of profound personal loss, to find myself, transform my loss into meaning, and discover a way to move forward.   xx  This research was supported by a Canadian Institutes of Health Research Doctoral Research Award and a Four-Year Fellowship from the University of British Columbia. Drs. Oberlander and Kobor also jointly provided stipend support for one year.   xxi  Dedication For my family.1  Chapter 1: Introduction    1.1 Maternal depression during pregnancy  Depression is a common and chronic condition affecting women during their child-bearing years. Often described as a “silent killer”, the cardinal symptoms of depression are depressed mood and anhedonia; additional symptoms include feelings of guilt and worthlessness, fatigue, suicidal thoughts, and changes in sleep and activity 1. An episode of major depression is characterized by persistence of these symptoms for a period of at least two weeks 1. Depression is associated with significant losses in quality of life 2 and is commonly comorbid with other mood disorders, most notably anxiety, with up to one-half of depressed individuals also having an anxiety disorder 3. Though depressive symptoms can range in severity across and within individuals, and across the life course 4,5, in some cases, they can be life-threatening and require immediate intervention for the safety of the affected individual. Rates of depression and other mood disorders are notably higher in pregnant women than in the general population6,7. Population-based studies estimate that up to one in five women experience a depressive episode during pregnancy 6,8, compared to a past-year prevalence of one in twenty in the general female population 9, making it a leading complication associated with pregnancy. Prenatal depression is a major risk factor for an array of adverse health-related outcomes in both mothers and children. Affected women are at risk of inadequate prenatal care 10, comorbid psychiatric disorders, and obstetric complications 11. Women with depression during pregnancy are also at up to six-fold increased risk of depression during the postpartum 8,12.  2  Despite the high burden of depression among pregnant and postpartum women, depression and other mood disorders often remain under-diagnosed and untreated in this population. Several reasons contribute to under-diagnosis, including lack of knowledge of symptoms, symptoms being mistaken as normal fluctuations attributed to pregnancy or childbirth, symptoms being missed due to an emphasis on fetal or child health during the clinical encounter, and an inability or reluctance of both women and clinicians to discuss emotional needs 13. Yet, better identification of maternal depression during pregnancy and subsequent monitoring and engagement with treatment could improve long-term health outcomes for women and their children 14. Studies have shown that benefits in child behavioral problems following family-based interventions are mediated by improvement in maternal depression symptoms 15. Given the high prevalence of maternal depression during pregnancy and its associated complications throughout life, and impacts on children’s outcomes, such improvements could translate into population-level health benefits across generations as well as overall cost reductions to the healthcare system 16.  1.2 Antidepressant treatment during pregnancy An urgent clinical challenge is how to treat maternal depression during pregnancy. Use of psychotherapy can be an effective method for treating and preventing depression 17-19; however, it may not be an accessible treatment option, especially for socio-economically disadvantaged populations 20, and its efficacy in more severe presentations of depression may be limited 21. Another common first-line treatment for depression is antidepressant medications which include the drug classes selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), and others. The use of 3  antidepressants during pregnancy has increased drastically in the past twenty years, with one study showing prenatal antidepressant use in a representative North American population rising from 5.7% in 1999 to 13.4% in 2003 22; another study estimated similarly high use rates in subsequent years 23.  Pharmacological treatment of depression during pregnancy poses a significant treatment dilemma due to concerns about potential risks to the developing child. Despite high rates of antidepressant treatment at the point of conception, the proportion of women remaining on treatment throughout pregnancy drops by up to one-half 24-26. Importantly, discontinuation of medication may place women at risk of destabilization and relapsing symptomatology 27 and expose them and their offspring to the negative immediate and long-term impacts of untreated prenatal depression.   1.3 Associations between maternal depression and prenatal antidepressant exposure on child development Exposure to maternal depression and antidepressant medications during pregnancy have been extensively investigated for their effects on offspring. Offspring of mothers who are depressed, even at subclinical levels, are at risk of suboptimal developmental outcomes throughout life 28. Beyond the developmental effects of SSRIs, there is widely established evidence that prenatal maternal depression has an impact on risk of low birth weight and for being born preterm 29. Behaviorally, newborns and infants of depressed mothers have been found to have poorer motor and orientation skills 30; display greater irritability; and show less activity, attentiveness, and expressivity 31. Continued maternal depression throughout a child’s infancy and early years is associated with worse maternal-infant attachment 32,33 and lowered self-4  regulatory abilities 34. In childhood, children display worse socio-emotional development, delayed motor and language development, impaired executive functioning, and worse academic outcomes 31,35-38, with higher incidence of psychiatric disorders emerging during adolescence 39,40.  Studies investigating the impact of prenatal antidepressant exposure have similarly found adverse effects on children’s developmental outcomes across the life course 41,42. However, findings and effect sizes are challenging to interpret in light of confounding by indication by maternal mental health; namely, that the impact of antidepressant medications is difficult to differentiate from that of the underlying maternal mental health condition necessitating pharmacological treatment 43,44. Findings applying more stringent methodological approaches to control for confounding by indication have found attenuated associations between child outcomes and prenatal antidepressant exposure, though the investigated outcomes have largely remained clinical outcomes with an emphasis on autism spectrum disorder 43. This ambiguity and questions on the potential impact on child development more generally contributes to the ongoing dilemma on treatment decisions facing pregnant women and their clinicians. Correct attribution of the causal risk factors (ie, antidepressants or maternal depression) for adverse child developmental outcomes is critical to identify the relative risks and benefits of antidepressant treatment during pregnancy and inform clinical decision-making 44.  1.4 Study rationale The available epidemiological evidence to-date highlights the impact of prenatal exposure to maternal depression and antidepressant medications on offspring’s developmental outcomes – in other words, that children prenatally exposed to maternal depression and to 5  antidepressants are at risk of overall worse development has been well-established. However, an understanding of the factors that determine which children will be affected, how, and potential improvements associated with interactions between prenatal antidepressant exposure and changes in maternal mood, has not been as well-elucidated45,46. To address these knowledge gaps, in this dissertation, I consider the role of key factors related to prenatal maternal depression and antidepressant use during pregnancy that can potentially influence children’s developmental health outcomes.    1.4.1 Factors influencing the impact of maternal depression and prenatal antidepressant exposure on child development The factors under investigation in this dissertation relate to internal and external factors: namely, innate biological differences (genetic allelic and epigenetic variations) in the child, as well as external factors (differences in maternal depression severity and symptomatology over time) that can potentially influence the impact of maternal depression and prenatal antidepressant exposure on child development. These factors were specifically chosen for the following reasons: First, as previously mentioned, the use of antidepressants during pregnancy is associated with adverse developmental outcomes 41. However, the decision to continue or discontinue antidepressants during pregnancy is related to a number of factors, including differences in levels of maternal depressive symptoms (i.e. illness severity), that also affect developmental risk 42. Considering these contextual factors when examining the impact of prenatal antidepressant exposure on child outcomes is critical for correct attribution of developmental risk.  6  Second, children who are prenatally exposed to maternal depression are not equally affected. Innate biological differences may moderate or mediate the impact of prenatal exposures on later outcomes, through genetic allelic differences and epigenetic variations47. A greater understanding of biological pathways could contribute to an understanding of the differences in children’s outcomes that are observed. Third, though exposure to maternal depression prenatally has an independent effect on development, research suggests that subsequent ongoing postnatal exposure to maternal depression, and changes in maternal depressive symptoms over time, also has independent and synergistic effects 48,49. Taking a longitudinal perspective, rather than considering exposure at only one point in time, can help account for the variations in child outcomes that are seen across the life course.  Understanding the factors that influence the impact of maternal depression and prenatal antidepressant exposure on child development is necessary to better understand how and why some but not all children are affected. Findings from this dissertation should help illuminate reasons underlying developmental variations following these exposures. They should also help identify mothers and their children who may be at risk as well as those who may stand to benefit from differences in the investigated factors.   1.5 Conceptual framework This research study draws on the seminal theoretical framework provided by the “Integrative Model for the Transmission of Risk to Children of Depressed Mothers”50 (hereafter referred to as the “Transmission of Risk Framework”). This model outlines four distinct and interacting mechanisms by which transmission of development risk for psychopathology in 7  offspring of depressed mothers may occur. The proposed mechanisms in this model are: 1) heritability of depression; 2) innate neurobiological dysfunctions; 3) exposure to maternal behaviors, cognition, and affect; and 4) stressful context of children’s lives. The framework also takes a developmental perspective to the transmission of risk by critically considering the element of time: first, by considering the stage of development in the child and the associated underlying developmental processes at that stage, and second, by considering the differential impact of maternal depression at different developmental stages. Finally, this theory conceptualizes of offspring vulnerabilities as emerging across several domains of development encompassing psychobiology, behavior/interpersonal, cognition, and affect which together, may result in subclinical/clinical child outcomes. Importantly, this framework considers that various contextual factors, related to both innate and external influences, determine the impact of exposure to maternal depression on child developmental risk.  1.6 Dissertation objectives The primary purpose of this thesis was to investigate the role of key factors that can influence the relationship between prenatal exposure to maternal depression and antidepressant medication, and child development. In this pursuit, an additional purpose was to articulate an integrative perspective of how the considered internal and external factors can influence the relationship between prenatal maternal depression and antidepressant medication exposure and child development.     8  The specific objectives of this research were the following:  1) To present an integrative perspective on the role of key contextual factors in explaining why some but not all children may be affected by prenatal exposure to antidepressant medications and maternal depression.  2) To isolate the effect of prenatal antidepressant exposure on child developmental outcomes by addressing the role of underlying maternal mental health and other confounding factors.  3) To explore whether prenatal exposure to maternal depression is associated with genetic allelic and epigenetic variations. 4) To determine the effect of longitudinal patterns of maternal depression, spanning pre- and postnatal periods, on child developmental outcomes over time.   1.7 Data sources The data used in this research comes from administrative data from the province of British Columbia, Canada, as well as from two prospective birth cohort studies from British Columbia, Canada, and Wisconsin, United States. An overview on the different data sources is provided here, with information pertaining to each data source provided in greater detail within each chapter.  Population-level administrative data were obtained from several different population-based health and education databases from British Columbia, Canada. All data were anonymized and linked through an encrypted unique identifier to ensure anonymity and privacy. The province-wide databases included: 9  • Linked births data on mothers and their children through the Perinatal Data Registry51, providing information on ~99% of births in the province and allowing linkage of maternal-child dyads; • Physician billings data covering all fee-for-service claims from the Medical Services Plan52, including service dates, fee items, and associated diagnoses based on International Classification of Disease (ICD) codes; • Hospital discharge abstracts53, containing admission and discharge dates, procedure codes, and ICD diagnoses codes; • British Columbia Consolidation file54 providing demographic information on sex, age and dates of registration in the provincial healthcare plan; • Prescription drug information from PharmaNet55, capturing prescription drug dispensations from community pharmacies including dispensation date, drug name, and Drug Identification Numbers; and  • Early Development Instrument56 data accessed through the Human Early Learning Partnership that provides information on childhood developmental health.  One of the prospective birth cohort studies used in this dissertation was an ongoing study (Oberlander PI) focused on investigating the effects of prenatal maternal mood disorders and antidepressant use on children’s brain development, behaviour, and mental health. In this cohort, women were recruited from clinics and hospitals in Vancouver, British Columbia, and surrounding areas during their second trimester of pregnancy (n=191). The oldest children in this cohort were six years old at the time of analysis. Detailed phenotypic measurements were collected on mothers and children at various time points including clinician- and self-reports on 10  maternal depressive symptoms and maternal-reports and laboratory assessments of children’s mental health and behavior.  The Wisconsin Study of Families and Work (WSFW, Essex PI) is another prospective cohort study from which data was used for this dissertation. The WSFW recruited women in their 2nd trimester of pregnancy from obstetric and family practice clinics in Milwaukee and Madison, Wisconsin. The WSFW tracked parents and children from the 2nd trimester of pregnancy until children were 18 years old, and maternal depression was assessed by self-report. Biological samples, including buccal swabs of children, were collected when children were 18 years old. Participants in both birth cohort studies represent the middle-to-upper spectrum of socio-economic status and the majority identify as Caucasian.   1.8 Dissertation structure Including the introduction, this dissertation contains six chapters. The second chapter of this dissertation presents an integrative perspective outlining the importance of biological, contextual, and longitudinal factors in understanding the impact of prenatal exposure to maternal depression and antidepressant medications on child development. Chapter 3 investigates the effect of prenatal antidepressant exposure on child development at kindergarten, while accounting for the impact of underlying maternal mental health and other confounding factors. The fourth chapter attempts to explore genetic and epigenetic differences that could mediate the long-term impact of prenatal maternal depression exposure on child development. Chapter 5 investigates the impact of longitudinal patterns of maternal depression, spanning across prenatal and postnatal periods, on child development. Finally, the Discussion synthesizes the findings presented across the different studies, and highlights the significance of this thesis.  11  Chapter 2: In-utero SSRI antidepressant exposure: Fetal programing via developmental interactions with context   2.1 Synopsis Selective serotonin reuptake inhibitor (SSRIs) antidepressants are prescribed to manage mood disorders in pregnancy and research has focused on identifying affected developmental outcomes. Disentangling the impact of the medication from maternal mood (confounding by indication) remains a challenge and understanding why development among some but not all exposed children is at risk remains an urgent question.  This chapter examines prenatal SSRI exposure-related developmental variations related to innate, maternal, and longitudinal contextual influences.   2.2 Prenatal SSRI exposure and developmental programing  This chapter draws on the theoretical underpinnings of this research project and articulates an integrative perspective of how internal and external contextual factors can influence the impact of prenatal exposure to antidepressants and maternal depression on child development.  SSRIs are prescribed in pregnancy with the expectation that they confer a benefit to the mother and by extension her offspring through improvements inherent to enhanced maternal mood. With that in mind, two decades of research has primarily focused on isolating specific health or developmental risks associated with in-utero SSRI exposure. Increasingly, however, developmental and behavioral outcomes are being reported that illustrate variations in patterns of development. Human studies are challenged by a failure to isolate specific effects of 12  drug exposure from the effects of the underlying maternal mental illness (“confounding by indication”) and related genetic and environmental influences. An integrative perspective is emerging that is beginning to elucidate how molecular genetic factors, together with environment, influence relationships between prenatal exposures (SSRIs and maternal depression) and developmental outcomes (Figure 2.1).   Figure 2.1. Diagram illustrating an integrative and contextual approach to understanding the impact of prenatal SSRI exposure on child developmental outcomes. Prenatal SSRI exposure may play a role in shaping early human brain development that sets pathways for both developmental risk and resiliency across childhood. However prenatal SSRI exposure may shape developmental variations via interactions with pharmacological, genetic and social-contextual factors that occur across prenatal and postnatal periods. Adapted from 57   2.3 Altered serotonin signaling Central to an understanding of how in-utero SSRI exposure influences early brain development is the diverse role played by the neurotransmitter serotonin as a mediator between early life experiences and subsequent development. Serotonin plays critical roles during early 13  developmental periods. As a growth factor during prenatal periods, serotonin is involved in regulating the development of its own and related neural systems. Later in the mature brain, it shapes cognition, attention and stress responsivity. Given these dual roles, altering serotonin levels during early sensitive periods might have lasting developmental consequences, setting lifelong pathways that shape resiliency and adversity.  SSRIs are commonly used during pregnancy, and as they cross the placenta and the blood-brain barrier, altering fetal central serotonin signaling during critical periods of neurodevelopment, it is conceivable that such early exposure may alter early brain development. The effects of changes in early serotonin signaling are highlighted in a mouse model where the serotonin transporter (5HTT), which acts as a key regulator of extracellular and synaptic serotonin, is genetically absent as well as with pharmacological blockade of 5HTT using SSRI antidepressants. In this model, early shifts in serotonin signaling were linked with an increased risk in the offspring for affective disorders later in life 58. Mice with increased serotonin levels following fluoxetine exposure during an early postnatal period (akin to a human 3rd trimester) displayed depressive and anxiety-related behaviors in adulthood, mimicking the very effects of altered serotonin signaling associated with genetic 5HTT inactivation.   2.4 Prenatal SSRI exposure: A differential susceptibility   How SSRIs shape developmental pathways is a complex story that defies “main effects” findings and categorical outcomes.  Importantly, not all factors that shape serotonin signaling via SSRI exposure equally affect all children, and accounting for an ongoing interplay between biological (allelic variations), experiential (prenatal drug or maternal mood exposure) and contextual (postnatal care giving and environment) variables is essential.  Individuals at risk in 14  one environment may indeed benefit from a positive experience in another context, demonstrating both vulnerability and resilience to adversity. In fact, changes in the fetal environment may not only shape risk but also an adaptive response, whereby interactions among social, biological, and/or physical factors alter the capacity to react or adapt to subsequent life events, influencing subsequent pathways leading to health and adversity. This may reflect a process of  ‘prenatal programming of postnatal plasticity’ 59. Recent evidence raises the possibility that prenatal SSRI exposure and early alterations in 5HT signaling may function as a ‘plasticity’ factor.  Allelic variations that alter serotonin signaling in the context of both early (i.e. fetal) and ongoing (i.e. postnatal/childhood) life experience help illuminate how a “differential susceptibility” model 59 accounts for the diversity of child developmental outcomes that are observed. As a case in point, one study found that at six years of age, in-utero SSRI exposure appeared to offer a developmental 'benefit’ on cognitive performance in the presence of continued postnatal maternal depression. Similarly, children with a short allele of 5HTT (reflecting reduced levels of serotonin reuptake) showed resilience to high maternal depressive symptoms.  In contrast, cognitive capacity among children with two long alleles (reflecting greater availability of the serotonin transporter and therefore lower intrasynaptic serotonin) suffered where mothers had more severe depressive symptoms, suggesting higher sensitivity of the long allele to maternal depression 60.  In this study, children exposed to SSRIs in-utero, coupled with genetic variations that increase serotonin signaling (i.e., short alleles), appeared to have better preserved executive functioning, while children with long alleles appeared to benefit from stability or improvement in mother’s mood. Here, this differential susceptibility to environmental context illustrates how a 15  genetic variation (in this case, the short allele) appears to serve as a “plasticity allele” (in contrast to a “risk” allele), with outcomes dependent on environment. In this context, serotonergic tone, via either prenatal SSRI exposure or 5HTT allelic variations, appeared to affect a self-regulatory capacity that might heighten sensitivity to a world with a depressed mother. Highly vigilant individuals may therefore become either vulnerable or resilient, depending on the particular demands of their social environment.   2.5 Maternal prenatal and postnatal context Inherent to prenatal SSRI exposure are the underlying mood disturbances that led to treatment with antidepressants in pregnancy in the first place. Maternal mood disturbances themselves shape early developmental outcomes, however, and many maternal depression-related effects are often indistinguishable from the impact of antidepressants. Unanswered questions remain about links between SSRI exposure and developmental outcomes that may be associated with maternal or familial factors that led to SSRI use in the first place, raising key questions about the role SSRIs play and whether they may act as a “hitchhiker” in this setting, rather than a causal agent. In other words, there is a possibility that SSRI exposure represents a proxy exposure to more proximal roles for genetic and environmental factors as well as other drugs (ie, benzodiazepines) and associated disorders. Recent epidemiological work highlights the interconnectedness between maternal mental health and antidepressant use, showing attenuation of associations between prenatal SSRI exposure and autism spectrum disorder when adjusting for maternal mood disturbances and accompanying environmental factors 61.   Appreciating the impact of interactions between biological and contextual factors offers a critical perspective that reflects ongoing influences of daily life experience and individual 16  heterogeneity. Accordingly, studies investigating longitudinal trajectories of pre- and postnatal exposures provide insight into variations in developmental outcomes following divergent postnatal contextual circumstances. This has been highlighted in a recent study (Chapter 5) examining the role of ongoing postnatal experiences in relation to prenatal exposures 62. Using a data-driven approach to elucidate patterns of maternal depression across pregnancy and the first 3 years, the developmental outcomes of children exposed to high levels of maternal depressive symptoms depended on exposure to later life trajectories, such that continued exposure to high depressive symptoms throughout early childhood was associated with poorer emotional and cognitive development. In contrast, improvement in maternal depressive symptoms over the same time frame did not lead to worse developmental outcomes despite higher prenatal exposure to mother’s depression. In this sense, child development and behavior following prenatal exposures, with their inherent effects on fetal serotonin signaling, represents an ongoing interplay between psychological, pharmacological, genetic, and social factors inherent to both the mother and her child long after birth.  Using a life course perspective allows one to move from regarding prenatal SSRI exposure as a singular pharmacological exposure at one time, focused on one disorder (depression), to an opportunity that allows a consideration of ways to improve perinatal mental health, pregnancy and developmental outcomes in a longitudinal and iterative process across time.  2.6 Conclusion Prenatal SSRI exposure, possibly related to changes in fetal serotonin signaling, may play a critical role in shaping early human brain development, reflecting interactions between genetic, environmental, and social factors. Accordingly, the impact of prenatal influences that affect 17  developmental risk, potentially via alterations in central serotonin levels, needs to be considered within a longitudinal context of maternal prenatal and postnatal mental health, while also accounting for ongoing interactions with everyday childhood environment. Studying the impact of prenatal SSRI and maternal depression exposure offers an opportunity to examine fetal programming in a way that integrates ongoing maternal influences within a child developmental perspective (Figure 2.1). While early shifts in serotonin signaling may help set pathways for both developmental risk and resiliency, interactions between biological and experiential factors may ultimately shape a sensitivity to context that contributes a developmental benefit in some but not all social environments.  Accounting for a broad range of developmental outcomes is central to identifying children who are likely to benefit from maternal SSRI treatment and improved perinatal mental health. An integrative perspective allows for an appreciation of how biological factors influencing alterations in serotonin signaling may both heighten or lessen vulnerability in the context of maternal mental health (and its treatment) during and long after pregnancy. Such a perspective should enable the field to move beyond regarding prenatal SSRI exposure as a singularly adverse exposure. In this way, understanding the impact of early changes in serotonergic signaling in relation to factors associated with ongoing maternal and environmental contexts offers insights that should enable a better understanding of patterns of individual differences in childhood development associated with prenatal SSRI exposure, and optimize maternal and child outcomes throughout life.    18  Chapter 3: Prenatal antidepressant exposure and child development at kindergarten age  3.1 Synopsis Background: Prenatal antidepressant exposure has been associated with adverse child development. However, studies to-date have shown mixed results that could reflect incomplete control of confounding. The objective of this study was to determine associations between prenatal antidepressant exposure and child development. Methods: This was a population-based retrospective cohort study using administrative data on mothers and children from British Columbia, Canada (n = 94,712). Prenatal antidepressant exposure was defined as 2+ dispensations during pregnancy, or 1+ dispensations from the 2nd trimester of pregnancy until delivery. Child development was assessed using the teacher-reported Early Development Instrument at kindergarten. Vulnerability in physical health (readiness for school, independence, motor skills), emotional maturity (anxious, aggressive, and hyperactive behaviors), social competence (overall, respect, learning, readiness to explore), cognitive development (basic literacy, advanced literacy, numeracy, interest in literacy/numeracy), and communication skills were investigated. Children prenatally exposed to antidepressants were matched 1: maximum 4 to unexposed children using high-dimensional propensity scores (HDPS), a data-driven method that captures known and unobserved sources of confounding. Children whose mothers had had an antidepressant dispensation in the 90–365 days prior to conception and continued treatment during pregnancy were compared against those whose mothers discontinued antidepressants 19  during pregnancy. A sensitivity analysis was conducted restricting exposure as having had to include dispensations in the 2nd/3rd trimester. Results: 3.87% (n = 3,611) of children in the overall study cohort were prenatally exposed to antidepressants, and the strongest associations were seen with children’s risk for anxious behaviours. In the HDPS-matched cohort, prenatal antidepressant exposure was associated with increased vulnerability in physical independence (OR, 1.14; 95% CI, 1.00 – 1.30), social exploration (OR, 1.64; 95% CI, 1.23 – 2.20), and anxious behaviours (OR, 1.30; 95% CI, 1.01 – 1.66). Children of antidepressant continuers vs. discontinuers had higher anxious behaviours (OR, 1.32; 95% CI, 1.01 – 1.72). Sensitivity analyses also revealed increased vulnerability in physical independence in children of mothers who continued versus discontinued treatment (OR, 1.19; 95% CI, 1.03 – 1.38).  Conclusions: Prenatal antidepressant exposure was associated with increased odds of vulnerability in anxious behaviours and physical independence at kindergarten age. These findings likely reflect a combination of effects from drug exposure, underlying maternal illness severity, and maternal genetic propensity to have a mood disorder which could be conferred on her child.   3.2 Introduction As discussed in Section 1.3, a significant challenge in identifying the impact of prenatal antidepressant exposure is in distinguishing its effects from that of underlying maternal mental health problems. The decision to pursue or continue pharmacotherapy for depression during pregnancy raises challenging clinical questions for women and their clinicians due to concerns of adverse fetal and developmental effects. Because discontinuing antidepressant treatment may 20  increase the risk of relapsing depression during pregnancy 26,27, which in turn, carries risk for maternal and fetal health 63, elucidating the effects of prenatal antidepressant exposure on developmental outcomes to better understand potential risks remains an important research objective.  As described in Chapter 2, antidepressants readily cross the placenta and the fetal blood-brain barrier, leading to concerns that early shifts in neurotransmitter systems during developmentally sensitive periods could have adverse consequences on brain development 41. Increased risks associated with prenatal antidepressant exposure have been reported for outcomes including autism spectrum disorder 64, attentional disorders 65, and altered motor, emotional, and language development 66-69; yet, reports have been inconsistent and contradictory 43,70. Moreover, recent findings suggest that a broader range of neurodevelopmental outcomes encompassing overall risk of psychiatric disorders may be affected 42. The question of which, and to what extent, child developmental outcomes are affected by prenatal antidepressant exposure thus remains controversial.  Central to understanding the impact of prenatal antidepressant exposure is the need to address confounding by indication by disentangling the effects of antidepressant exposure from those of the underlying maternal mental illness 41. Untreated, maternal depression during pregnancy is an independent risk factor for adverse developmental outcomes including low birth weight, preterm birth, and socio-emotional disorders later in childhood 29,41. Moreover, the treatment pregnant women receive for their depression is also affected by contextual factors, such as socio-economic status 71, which also influence child development 72. Addressing these sources of confounding is critically important to isolate the unique impact of prenatal antidepressant exposure. Indeed, recent work using advanced methods to adjust for possible 21  measured and unmeasured factors related to indication for antidepressant medication during pregnancy suggest that mixed findings in the literature may reflect differences in control of confounding across different studies 61,73-75.  The present study was undertaken to investigate associations between prenatal antidepressant exposure and a spectrum of outcomes reflective of overall child development. Using population-level data, I sought to identify the impact of antidepressant exposure on children’s physical, emotional, social, cognitive, and communication development at kindergarten age, adjusting for confounding using stringent methods.   3.3 Methods 3.3.1 Study design and data sources This was a retrospective cohort study using administrative data from British Columbia, Canada. British Columbia collects information from all fee-for-service physician visits, hospitalizations, and prescription drug dispensations for registered residents, allowing linkage of individual-level data 52-55. Mothers and their offspring were identified through the British Columbia Perinatal Data Registry 51 that covers 99% of births province-wide and provides precise estimates of gestational age, which were used to calculate pregnancy duration. Child development outcomes were available through the Human Early Learning Partnership 56. Data across all sources were linked at Population Data BC using unique individual identifiers. This study received ethics approval from all Data Stewards and the University of British Columbia Children’s and Women’s Research Ethics Board. All inferences, opinions, and conclusions drawn are those of the authors and do not reflect the opinions or policies of the Data Stewards.  22  3.3.2 Study cohort The sampling frame for this study represented children with completed records of child development collected between 2003 – 2013. Singleton children born in-province to mothers registered in the provincial health plan for at least 275 days during the year of birth and year prior to birth were considered for the study. 275 days was chosen as the threshold, as women who were in the province for at least three-quarters of the year were likely to have had healthcare records. Only maternal-child dyads with complete data on covariates and outcomes were included (n = 94,712) (Figure 3.1).   Figure 3.1. Flowchart detailing definition of comparison groups from the overall study cohort. Antidepressant use during pregnancy was defined as at least 2+ prescriptions during pregnancy or at least 1+ prescription during the 2nd or 3rd trimesters. Antidepressant use pre-conception was defined as any prescription in the 90 – 365 days before pregnancy.  23  3.3.3 Exposure classification Dispensations of antidepressants during pregnancy were identified using Anatomical Therapeutic Chemical Classification System codes (Supplementary Table A.1). Pregnancy trimesters were defined as outlined in Supplementary Table A.1. Prenatal antidepressant exposure was defined as either two or more dispensations for antidepressants throughout pregnancy, or at least one dispensation from the 2nd trimester of pregnancy until delivery. Antidepressants belonging to different drug classes were included in the definition of exposure (see Supplementary Table A.1 for list and categorization of antidepressant drugs). Though antidepressants are generally prescribed for extended durations, including only single dispensations could include pregnancies where exposure ceased shortly after conception due to concerns of adverse fetal effects; however, single dispensations occurring during the 2nd trimester onward were included as they were likely filled with knowledge of pregnancy status.  3.3.4 Outcome measures Child development outcomes were assessed with the Early Development Instrument (EDI). The EDI is a validated measure of children’s school readiness that is completed by kindergarten teachers and has been used internationally in Canada and Australia 76. The EDI has been administered in British Columbia every one – three years since the 1999/2000 academic year. School and teacher participation in the EDI is voluntary and child participation is via passive parental consent; i.e., parents may withdraw their consent and their child will not be rated. The EDI collection captures the majority of the British Columbian kindergarten population (92% of the kindergarten population from all public-school districts), and approximately 88% of children in British Columbia are enrolled in public schools for kindergarten.  24  The EDI measures 16 subscales of early childhood development that fall into five distinct domains: physical health and well-being (gross and fine motor skills, physical independence, and physical readiness for the school day), emotional maturity (aggressive, anxious and fearful, hyperactive and inattentive, and prosocial and helping behaviors), social competence (overall social competence, approaches to learning, readiness to explore new things, and respect and responsibility), language and cognitive development (basic literacy, advanced literacy, basic numeracy, and interest in literacy, numeracy, and memory), and communication skills and general knowledge. Its psychometric properties and validity have been well-studied, including its associations with concurrent and later measures of developmental outcomes 77-79.  As responses are skewed and not readily amenable to transformation, children were classified as developmentally “vulnerable” if their scores fell below previously defined thresholds designed to capture meaningful differences in developmental status 80. The subscale measuring prosocial and helping behaviors was not used as it had >5% missing responses.    3.3.5 High-dimensional propensity score  Using high-dimensional administrative data to generate propensity scores represents a data-driven approach to selecting covariates reflective of both known and unknown sources of confounding. This approach can help capture unknown sources of confounding specifically by identifying unobserved confounding factors through empirical proxies identified using this data-driven method 81. High-dimensional propensity scores (HDPS) 81 were generated using the following data dimensions captured in the 365 days prior to conception: outpatient diagnoses and fee codes, inpatient diagnoses and procedures, and drug dispensations. The top 500 variables from these dimensions were determined first, based on baseline prevalence, and then based on 25  associations with exposure (Supplementary Table A.2). A number of pre-identified confounding factors were also included for HDPS estimation; these were child age at EDI, child sex, gestational age, child English as a Second Language status, EDI wave, maternal age at birth, parity, neighborhood income quintile 82, health service delivery area, diagnosis of a mood or anxiety disorder during pregnancy, and use of other psychotropic or antiepileptic medications during pregnancy. The HDPS was used for 1:4 (maximum) matching of exposed children to those whose mothers did not take any antidepressants during pregnancy, within a caliper of 0.2 standard deviations, per recommendations 83.    3.3.6 Main analyses Demographic characteristics between exposure groups were compared using standardized differences, which use standard deviations to compare group differences, making them appropriate for assessing covariate balance across variables measured using different units and in large study samples 84. A threshold of 0.1 was considered a meaningful difference 84. To provide a baseline comparison for inferential analyses, odds ratios and 95% confidence intervals of antidepressant exposure in the overall cohort using multivariable logistic regressions were estimated. Included covariates in these baseline analyses were child age at EDI, child sex, gestational age, child English as a Second Language status, EDI wave, maternal age at birth, neighborhood income quintile, health service delivery area, a mood or anxiety disorder diagnosis during pregnancy, and use of other psychotropic or antiepileptic medications during pregnancy. In the HDPS-matched cohort, conditional logistic regressions were conducted to estimate odds ratios and 95% confidence intervals. These analyses adjusted for other psychotropic or antiepileptic medications taken during pregnancy given that a standardized difference of 0.10 26  remained after matching 84 (Table 3.1). I also conducted analyses based on a subsample of mothers with an existing indication for antidepressants, determined based on dispensations occurring in the 90 – 365 days prior to conception. The developmental outcomes of children whose mothers continued antidepressant treatment during pregnancy were compared against those whose mothers discontinued treatment, adjusting for the previous full list of covariates. The definition of comparison groups from the overall study cohort is illustrated in Figure 3.1.   3.3.7 Sensitivity analyses I performed two sensitivity analyses. To address whether effects may have been influenced by misclassification of exposure, exposure was restricted as having had to include dispensations occurring in the 2nd and 3rd trimesters of pregnancy (i.e., eliminating exposures that occurred only in the first trimester). Second, to determine whether effects differed based on drug class, associations were estimated based on exposure to SSRIs only, non-SSRIs only, or both.. All data management and statistical analyses were conducted using SAS 9.4 85.   3.4 Results 3.4.1 Demographic characteristics of study cohort The study population included a total of 94,712 children born to unique mothers (Figure 1). Of the cohort, 3,611 children (3.87%) were exposed to antidepressants during pregnancy. Children were 48.7% female (n=46,086), with a mean gestational age of 38.9 weeks, and a mean age of 5.65 years at the time of EDI completion.  Mothers were a mean age of 30.5 years at birth and 55.3% (n=52,387) were multiparous.  27  Demographic characteristics of unexposed compared to exposed individuals for the overall and HDPS-matched cohorts are presented in Table 3.1. A higher proportion of mothers who took antidepressants during pregnancy were multiparous (62.1% vs 55.0%). Exposed children had a lower mean gestational age (38.4 vs 38.9 weeks) and fewer had ESL status (7.3% vs 21.7%) relative to unexposed children. There were significant differences between exposed and unexposed mothers in diagnoses of mood and anxiety disorders (64.2% vs 8.9%) and other psychotropic or antiepileptic medications taken during pregnancy (22.5% vs 3.5%).   28  Table 3.1. Characteristics of mother-child dyads in the overall cohort and in a subset of exposed children matched to unexposed children using a high dimensional propensity score, by prenatal antidepressant exposure. All results are presented as n (%) unless otherwise stated.  Overall Cohort (n = 94,712) Matched cohorta (n = 7,862)  Not exposed (n = 91,051) Exposed (n = 3,661) Standardized difference Not exposed (n = 5,609) Exposed (n = 2,253) Standardized difference Maternal characteristics  Maternal age at delivery, mean (SD) 30.51 (5.46) 30.58 (5.54) -0.01 30.26 (5.58) 30.30 (5.58) 0.01 Multiparous 50,115 (55.04) 2,272 (62.06) 0.12 3,476 (61.97) 1,381 (61.30) -0.01 Lowest neighborhood income quintile 19,425 (21.33) 822 (22.45) 0.05 1,256 (22.39) 508 (22.55) 0.01 Mood/anxiety disorders diagnosis during pregnancy 8,061 (8.85) 2,352 (64.24) 1.22 2,951 (52.61) 1,159 (51.44) -0.02 Other psychotropic or antiepileptic medications during pregnancy 3,156 (3.47) 822 (22.45) 1.21 886 (15.80) 394 (17.49) 0.10        Anti-depressant exposure during pregnancy NA   NA   SSRIs only  2,450 (66.92)   1,546 (68.62)  Non-SSRIs only  870 (23.76)   548 (24.32)  Both SSRI and non-SSRIs  341 (9.31)   159 (7.06)         2nd – 3rd trimester  3,221 (87.98)   1,954 (86.73)  Child characteristics 29  Child age at EDI, mean (SD) 5.65 (0.29) 5.65 (0.30) -0.01 5.65 (0.30) 5.65 (0.30) -0.01 Gestational age, mean (SD) 38.88 (1.79) 38.36 (1.89) -0.39 38.49 (2.18) 38.50 (1.79) 0.01 Female 44,327 (48.68) 1,759 (48.05) -0.01 2,756 (49.14) 1,092 (48.47) -0.01 English as a second language 19,708 (21.65) 266 (7.27) -0.92 551 (9.82) 205 (9.10) -0.08        Vulnerability in child development        Physical well-being       Readiness 3,266 (3.59) 222 (6.06) 0.55 297 (5.30) 137 (6.08) 0.15 Independence 12,531 (13.76) 731 (19.97) 0.37 935 (16.67) 428 (19.00) 0.13 Motor skills 29,245 (32.12) 1,448 (39.55) 0.21 2,051 (36.57) 866 (38.44) 0.05 Social competence       Competence 10,317 (11.33) 616 (16.83) 0.40 806 (14.37) 365 (16.20) 0.12 Learning  8,885 (9.76) 481 (13.14) 0.31 694 (12.37) 300 (13.32) 0.08 Exploring 2,254 (2.48) 131 (3.58) 0.43 133 (2.37) 84 (3.73) 0.52 Respect 5,816 (6.39) 339 (9.26) 0.39 473 (8.43) 209 (9.28) 0.10 Emotional maturity       Anxious 2,809 (3.09) 191 (5.22) 0.57 201 (3.58) 111 (4.93) 0.36 Aggressive 7,908 (8.69) 416 (11.36) 0.28 600 (10.70) 258 (11.45) 0.07 Hyperactive 13,148 (14.44) 664 (18.14) 0.23 997 (17.78) 398 (17.67) -0.01 Language/Cognitive development       Basic literacy 9,508 (10.44) 509 (13.90) 0.29 723 (12.89) 298 (13.23) 0.03 Interest 10,198 (11.20) 490 (13.38) 0.18 717 (12.78) 304 (13.49) 0.06 30  Advanced literacy 18,716 (20.56) 925 (25.27) 0.21 1,334 (23.78) 568 (25.21) 0.06 Basic numeracy 12,599 (13.84) 654 (17.86) 0.26 891 (15.89) 389 (17.27) 0.09 Communication skills 30,472 (33.47) 1,317 (35.97) 0.07 1,916 (34.16) 794 (35.24) 0.03        aBased on matching using a high-dimensional propensity score estimated for prenatal antidepressant use.  31   After matching based on the HDPS, demographic characteristics became more similar between exposed and unexposed groups; differences were no longer meaningful with the exception of other psychotropic or antiepileptic medications taken during pregnancy (17.5% vs 15.8%) (Table 3.1). Because a matching strategy restricted to a certain width of the HDPS was employed, not all exposed children were successfully matched. Demographic characteristics of unmatched and matched exposed mother-child dyads, and of antidepressant continuation and discontinuation groups can be found in Supplementary Tables A.3 and A.4.   3.4.2 Results from main analyses Results from main analyses are presented in Table 3.2, showing the odds of vulnerability on each developmental scale associated with prenatal antidepressant exposure. Children with prenatal antidepressant exposure were significantly more likely to be vulnerable on most developmental scales. These results were significantly attenuated in analyses that incorporated better control of confounding.   In the HDPS-matched cohort, prenatal antidepressant exposure was associated with developmental vulnerability in physical independence (OR 1.14; 95% CI, 1.00 – 1.30), social exploration (OR, 1.64; 95% CI, 1.23 – 2.20), and anxious behaviors (OR, 1.30; 95% CI, 1.01 – 1.66). Comparing developmental outcomes in children of mothers who continued antidepressant treatment during pregnancy relative to those who discontinued, only the association with anxious behaviors was significant (OR, 1.32; 95% CI, 1.01 – 1.72) (Table 3.2).   32  Table 3.2. Associations between prenatal antidepressant exposure and child development. Comparison groups were exposed children against unexposed children in the overall cohort, exposed children matched to unexposed children using a high dimensional propensity score, and children whose mothers had antidepressant dispensations in the 90 – 365 days prior to conception and continued versus discontinued treatment during pregnancy.  Overall cohort HDPS-matched cohort Antidepressant continuers vs. discontinuers Developmental vulnerability in: OR (95% CI)a OR (95% CI)b OR (95% CI)a Physical well-being    Readiness 1.32 (1.13 - 1.55) 1.13 (0.91 - 1.41) 1.09 (0.85 - 1.39) Independence 1.26 (1.15 - 1.38) 1.14 (1.00 - 1.30) 1.14 (0.99 - 1.32) Motor skills 1.20 (1.11 - 1.29) 1.06 (0.95 - 1.18) 1.04 (0.93 - 1.17) Social competence    Competence 1.27 (1.15 - 1.40) 1.11 (0.96 - 1.28) 1.03 (0.88 - 1.20) Learning  1.08 (0.97 - 1.21) 1.06 (0.91 - 1.23) 0.96 (0.81 - 1.14) Exploring 1.32 (1.08 - 1.61) 1.64 (1.23 - 2.20) 1.23 (0.89 - 1.70) Respect 1.14 (1.00 - 1.30) 1.07 (0.89 - 1.27) 0.93 (0.76 - 1.13) Emotional maturity    Anxious 1.41 (1.20 - 1.67) 1.30 (1.01 - 1.66) 1.32 (1.01 - 1.72) Aggressive 1.10 (0.98 - 1.24) 1.00 (0.86 - 1.18) 0.87 (0.73 - 1.04) Hyperactive 1.05 (0.96 - 1.16) 0.95 (0.83 - 1.08) 0.90 (0.77 - 1.04) Language/Cognitive development    Basic literacy 1.19 (1.07 - 1.33) 1.01 (0.87 - 1.17) 0.97 (0.82 - 1.15) Interest 1.11 (1.00 - 1.24) 1.05 (0.90 - 1.21) 1.00 (0.85 - 1.17) Advanced literacy 1.18 (1.08 - 1.28) 1.08 (0.96 - 1.22) 0.96 (0.84 - 1.10) Basic numeracy 1.20 (1.09 - 1.32) 1.07 (0.94 - 1.23) 1.00 (0.87 - 1.16) Communication skills 1.16 (1.07 - 1.25) 1.04 (0.93 - 1.16) 1.01 (0.89 - 1.14) Abbreviations: OR, Odds Ratio; CI, Confidence Interval; EDI, Early Development Instrument; ESL, English as a Second Language aAdjusted for maternal age at birth, neighborhood income quintile, parity, maternal mood or anxiety diagnoses during pregnancy, other psychotropic or antiepileptic medications dispensed during pregnancy, wave of EDI collection, gestational age at birth, child age at EDI, child sex, and child ESL status. bAdjusted for exposure to other psychotropic or antiepileptic medications during pregnancy.  33  3.4.3 Results from sensitivity analyses Results from analyses that did not consider children to be prenatally exposed to antidepressants if exposure occurred only in the first trimester of pregnancy were consistent with the analyses that considered exposure throughout pregnancy, with slightly elevated effect sizes (Table 3.3). However, a statistically significant effect in vulnerability in physical independence was observed among children whose mothers continued antidepressants during pregnancy compared with those whose mothers discontinued antidepressants (OR, 1.19; 95% CI, 1.03 – 1.38).    In drug class-specific analyses, directions of effect were consistent with those found in overall analyses, though statistical precision was reduced. Focusing on associations with physical independence, anxious behaviors, and social exploration, there were elevated point estimates with exposure to non-SSRIs and both SSRIs and non-SSRIs exposure, compared to SSRIs only, though results were not robust across analyses and confidence intervals were wide (Supplementary Tables A.5 and A.6).  34  Table 3.3. Associations between prenatal antidepressant exposure and child development. Antidepressant exposure was restricted to pregnancies where exposure had to include the 2nd and/or 3rd trimesters. Comparison groups were exposed children matched to unexposed children using a high dimensional propensity score, and children whose mothers had antidepressant dispensations in the 90 – 365 days prior to conception and continued versus discontinued treatment during pregnancy.  HDPS-matched cohort Antidepressant continuers vs. discontinuers Developmental vulnerability in: No. vulnerable (%) in exposed (n = 1,954) OR (95% CI)a No. vulnerable (%) in exposed (n = 2,384) OR (95% CI)b Physical well-being     Readiness 115 (5.89) 1.09 (0.86 - 1.38) 138 (5.79) 1.07 (0.82 - 1.38) Independence 382 (19.55) 1.15 (1.01 - 1.33) 497 (20.85) 1.19 (1.03 - 1.38) Motor skills 751 (38.43) 1.04 (0.93 - 1.16) 958 (40.18) 1.05 (0.93 - 1.18) Social competence     Competence 321 (16.43) 1.13 (0.97 - 1.31) 393 (16.48) 1.01 (0.86 - 1.19) Learning  266 (13.61) 1.08 (0.92 - 1.27) 309 (12.96) 0.97 (0.81 - 1.16) Exploring 74 (3.79) 1.73 (1.26 - 2.37) 82 (3.44) 1.29 (0.92 - 1.80) Respect 184 (9.42) 1.10 (0.91 - 1.33) 208 (8.72) 0.91 (0.74 - 1.12) Emotional maturity     Anxious 100 (5.12) 1.37 (1.05 - 1.79) 139 (5.83) 1.40 (1.07 - 1.83) Aggressive 224 (11.46) 1.01 (0.85 - 1.20) 262 (10.99) 0.84 (0.70 - 1.01) Hyperactive 350 (17.91) 0.96 (0.83 - 1.10) 431 (18.08) 0.90 (0.77 - 1.05) Language/Cognitive development     Basic literacy 267 (13.66) 1.03 (0.88 - 1.21) 336 (14.09) 1.00 (0.84 - 1.18) Interest 274 (14.02) 1.10 (0.94 - 1.29) 326 (13.67) 1.01 (0.85 - 1.20) Advanced literacy 491 (25.13) 1.10 (0.97 - 1.24) 591 (24.79) 0.95 (0.83 - 1.09) Basic numeracy 338 (17.30) 1.07 (0.92 - 1.24) 429 (17.99) 1.01 (0.86 - 1.17) Communication skills 678 (34.70) 1.00 (0.89 - 1.13) 850 (35.65) 1.00 (0.88 - 1.13) Abbreviations: OR, Odds Ratio; CI, Confidence Interval; EDI, Early Development Instrument; ESL, English as a Second Language aAdjusted for exposure to other psychotropic or antiepileptic medications during pregnancy. 35  bAdjusted for maternal age at birth, neighborhood income quintile, parity, maternal mood or anxiety diagnoses during pregnancy, other psychotropic or antiepileptic medications dispensed during pregnancy, wave of EDI collection, gestational age at birth, child age at EDI, child sex, and child ESL status.  36  3.5 Discussion In this population-based study, I investigated the impact of prenatal antidepressant exposure on children’s physical, emotional, social, cognitive, and communication development at kindergarten age. These assessments were based on normative and comprehensive measures of child development, reflective of typically developing children. This study reports that prenatal antidepressant exposure was associated with more anxious behaviors and worse physical independence, even after controlling for confounding by underlying maternal mental health using HDPS-matching and comparisons restricted to mothers with an existing indication for antidepressant treatment. However, it is important to emphasize the modest effect sizes that were found. Associations between prenatal antidepressant exposure and other child development outcomes were not statistically significant after robust control of confounding.     3.5.1 Comparison with previous findings These results concur with other studies reporting a selective impact of prenatal antidepressant exposure on children’s anxious behaviors. A previous study reporting associations between SSRI exposure in late pregnancy and children’s anxious and depressive behaviors at five years old similarly did not find the same relationship with externalizing behaviors, social problems, or other emotional problems 69. Also, a population-based study that controlled for confounding by familial factors using a sibling-adjusted study design found that prenatal antidepressant exposure was associated only with anxious symptoms at 36 months and not with other externalizing and internalizing behavioural outcomes 86. Another prospective cohort study also found that prenatal antidepressant exposure was associated with higher risk for anxious 37  behaviors at three years after adjusting for maternal depression symptoms during pregnancy and after birth, an effect that was sustained when children were six years old87.  Though detected effect sizes in the current and past studies were modest, the potential impact of prenatal antidepressant exposure on an anxiety-like phenotype in children is supported by biologically plausible mechanisms established through animal studies 58. Antidepressants typically act via manipulations of central levels of the key neurotransmitter serotonin, and developmental changes in serotonin may constrain the maturation of the serotonin system, explaining how increased intrasynaptic serotonin associated with antidepressant exposure may predispose to affective disorders later in life 41,58,88.  The association between antidepressant exposure and physical independence, and not motor skills, is inconsistent with previous studies reporting associations between prenatal antidepressant exposure and poorer motor development  66,89-91. This apparent discrepancy may be related to different methods used to assess physical developmental outcomes. It is important to note that the EDI is a behavioral assessment and thus, questions related to physical independence (i.e., “is well coordinated”) or motor skills (i.e., “levels of energy throughout the school day”) may capture overlapping constructs. It is possible that the significant association with physical independence may in fact be driven by differences in motor skills, and that elucidating more nuanced effects of prenatal antidepressant exposure on physical development require finer assessments than the EDI.  The lack of significant findings in other developmental domains (language development, cognitive skills, and communication skills) in analyses that employed stringent control of confounding suggests that observed associations in baseline analyses are likely confounded by indication or other unmeasured environmental differences, including those related to family 38  context. Similar attenuation of differences has been reported after control for confounding in prior studies investigating child developmental outcomes 61,65,73,92. These results reinforce the critical need to address maternal mental health when investigating associations between prenatal antidepressant exposure and child development in order to draw appropriate conclusions regarding the attributable causes of developmental risk.   3.5.2 Strengths and limitations This study is strengthened by its population-based nature and the relatively objective nature of the teacher as reporter, which overcomes potential biases by maternal mental health status or by maternal health seeking behaviors when outcomes are based on parental or clinical report. Further, despite inherent limitations, the unique and detailed developmental data included in the EDI allowed for exploration of potential impacts on a comprehensive suite of normative child development outcomes. This study is also strengthened by the methodology employed to tease apart effects due to confounding by indication and other unmeasured factors.  However, despite these efforts, a meaningful difference in use of other psychotropic or antiepileptic medications during pregnancy between the exposed and unexposed groups persisted. This difference may reflect an incongruity in maternal illness severity between groups; similarly, the decision to continue antidepressant treatment during pregnancy is likely related to illness severity or remittance coinciding with pregnancy.  I did not consider postnatal maternal mental state in analyses, as it is highly associated with prenatal maternal mental health and would also have been considered a mediating variable, but which may account for differences in child outcomes.  While relevant confounding factors including mood disorder diagnoses during pregnancy were adjusted for in these analyses, it is not clear how accurately these diagnoses reflect 39  differences in levels of psychiatric symptoms. Moreover, undiagnosed mood disorders or variations in symptoms and measures of daily function would not have been taken into account. I was unable to control for differences related to factors that were not represented in available data sources, such as education levels, ethnicity, immigrant status, or more refined neighborhood-level contextual factors. I was also unable to address the role of genetic confounding. While discordant-exposure sibling-based designs are one way to potentially address confounding by shared familial factors, including environmental and genetic risk, the interpretation of findings from such analyses are limited due to selection bias and confounding by non-shared factors 93. It is thus conceivable that prenatal antidepressant exposure may reflect a genetic propensity for a mother’s depression during pregnancy, which itself confers a risk for developmental risk in her child.  An inherent limitation of studies that rely on administrative data is misclassification of antidepressant exposure, as dispensations may not reflect whether women took their medications. As such, I used a relatively strict definition of antidepressant exposure. Results from sensitivity analyses that restricted exposure showed increased effect sizes without compromising statistical precision, supporting the use of a stricter definition of exposure.  The results of this study may not be generalizable to all women and children exposed to antidepressants during pregnancy. I intentionally chose to restrict analyses to exposed individuals who were able to be matched to comparable controls, or to those who were also exposed pre-conception, in order to generate appropriate comparison groups. Applying the results of this study to different populations who differ in terms of baseline characteristics or whose onset of antidepressant use occurs during pregnancy should be done with caution. The use of normative outcomes, reflective of typically developing children, also allows generalizability of findings to 40  general pediatric populations. However, a limitation of using the EDI is that associations with child behaviors or symptoms whose onset only occur later in childhood may have been missed. Selection bias in terms of who was included in the study is a concern; data were unlikely to have been missing at random and this study was restricted to individuals who had complete data thereby excluding women and children who represent the tail ends of the socio-economic distribution.   3.5.3 Conclusions To conclude, this study reports that prenatal antidepressant exposure was associated with modest increases in vulnerability in anxious behaviors and physical independence at kindergarten age, likely reflecting a combination of effects from drug exposure and underlying maternal mental illness severity. There were no associations with other domains of child development. Regardless of the specific areas of vulnerability, it is important to balance potential risks of antidepressant use during pregnancy against those of untreated maternal depression. Effective treatment of maternal depression during pregnancy is a priority for the health and well-being of mothers and their children.  41  Chapter 4: Prenatal maternal depression symptomatology, genetic background, and DNA methylation patterns in adolescence  4.1 Synopsis Background: Prenatal maternal depression is a common early life exposure with lasting effects on child development. Epigenetics has been proposed as a biological mechanism that may mediate the effects of prenatal exposures on long-term outcomes. However, shared genetic background between mothers and their offspring, reflecting heredity, may also contribute to observed effects. I sought to investigate whether prenatal maternal depression smptom levels are associated with long-term differences in children’s epigenetic marks, and to quantify the respective influence of genetic differences on observed associations.  Methods: I tested associations between prenatal maternal depression symptomatology and DNA methylation (DNAm) patterns in young adulthood from participants in a normative birth cohort with matched DNAm and genotype data (n=178). I explored the contribution of genotype to prenatal maternal depression symptomatology-associated DNAm patterns. Genome-wide DNAm patterns in buccal epithelial cells (BECs) at age 18 years were assessed using the Illumina HumanMethylation450 array and genotyping was performed with the Illumina PsychChip array. Results: 45 differentially methylated regions (DMRs) were significantly associated with prenatal maternal depression levels. However, of the 45 DMRs, a large proportion (44%) were under polygenic influence.  Conclusion: DNAm patterns in young adulthood that are associated with exposure to prenatal maternal depression levels are likely reflective of differences in genetic background; such 42  differences may reflect transmission of shared genetic risk for depression. These findings highlight the critical importance of incorporating the role of genetic factors in epigenetic studies investigating the impact of prenatal maternal depression symptomatology, in order to properly attribute the contribution of genetics on long-term epigenetic differences related to maternal depression  4.2 Introduction  Chapter 1 described biological differences that may moderate or mediate the risk of adverse outcomes following prenatal exposure to maternal depression and antidepressant medications. Although a number of biological processes have been proposed to explain how prenatal exposure to maternal depression affects offspring 94, the mechanisms underlying its lasting effects remain unclear 95. A recent avenue of research has been directed at the potential role that epigenetic processes may play in the transmission of developmental risk by maternal depression. Epigenetics refers to modifications of DNA that are often tightly linked to changes in gene expression but do not involve changes to the underlying genetic sequence 96. Of these modifications, DNA methylation (DNAm) is the best characterized in human populations 97. DNAm primarily refers to the addition of a methyl moiety onto the 5’ position of cytosine, most often in the context of CpG dinucleotides. The relationship between a gene’s DNAm and its expression is not straightforward 98; however, DNAm is an important regulator of gene activity 99. A notable example is that higher levels of DNAm at promoter sites are, in general, associated with repression of gene expression 100.   A number of additional characteristics about DNAm warrant attention. DNAm does not occur uniformly across the genome 99; in humans, the vast majority of methylated cytosines are 43  found within the context of cytosine-guanine (CpG) dinucleotide units, which themselves are located disproportionately in regions called CpG islands that mostly neighbor gene promoters 99.  Patterns of DNAm also exhibit wide variation within and across individuals. Distinct DNAm patterns occur in different tissue types 101 due to the role of epigenetic mechanisms in establishing and maintaining cellular identity 97. Moreover, changes in DNAm occur over the lifespan 102, with DNAm patterns undergoing the most rapid and dramatic changes during fetal development 103. Genotype also exerts widespread effects on DNAm, and individuals exhibit population-specific patterns of DNAm based on genetic ancestry 104. DNAm has been associated with prenatal environmental exposures, such as cigarette smoke 105,106. Though reported effect sizes in such studies have been small, they may represent biomarkers or even functional differences related to early life exposures 107. Combined with an understanding of the fetal period as a critical ontogenic window of development, these lines of evidence have opened up the prospect that DNAm may be involved in mediating the effects of prenatal maternal depression exposure on children’s biology and later outcomes 108,109.  Few studies have investigated the effect of prenatal exposure to maternal depressionon DNAm across the genome; yet, an epigenome-wide approach can lead to the discovery of novel associations with genes that were previously unconsidered, or reveal genome-wide patterns that would not be otherwise detected. One such epigenome-wide association study (EWAS) found 63 CpG sites in cord blood that were significantly associated with prenatal maternal depression 110. Notably, another study found overlapping DNAm differences across maternal blood, cord blood, and adult offspring’s hippocampi within genes involved in the immune system 111. However, null results have also been reported 112. 44  Importantly, work that has integrated genetic data found that most differentially methylated sites in neonates could be explained by an interaction of in-utero environment, including maternal depression, and genotype 113. A largely unaddressed question is the role that allelic differences play on the effect of prenatal exposures on DNAm. This is a highly pertinent concern given the noted impact of genotype on DNAm 108,113. Moreover, addressing genotype may also help untangle whether any observed effects may be due in part to hereditary influences 95. The purpose of the current study was thus to investigate whether prenatal maternal depression exposure is associated with DNAm patterns in young adults and to explore the contribution of underlying genetic influences in these associations.   4.3 Methods 4.3.1 Study design and study sample Data for this study comes from the Wisconsin Study of Families and Work (WSFW), a prospective birth cohort study based in Milwaukee and Madison, Wisconsin114. From the overall cohort, 178 mother-child dyads where children had data on DNAm at age 18 years and genotype were included for analysis in this study. Demographic characteristics of the study sample can be found in Table 4.1.   Table 4.1. Characteristics of mother-child dyads (n=178) in the present study. Maternal characteristics Age (years), mean (SD) 29.5 (4.3) Ethnic status, n (%)  Caucasian 166 (93.3%) Other 12 (6.7%) Education (years), mean (SD) 14.9 (2.2) 45  Income (in USD), median (MAD) 46,500 (17,791) Marital status, n (%)  Married 168 (94.4%) Divorced 3 (1.7%) Widowed 7 (3.9%) First child at time of study, n (%)  No 103 (57.9%) Yes 75 (42.1%) Prenatal depression score, mean (SD) 8.4 (7.6) Postnatal depression, n (%)  Low 158 (88.8) Moderate-fluctuating 10 (5.6) Moderate-increasing 10 (5.6) Child characteristics Sex, n (%)  Male  81 (45.5) Female 97 (54.5) Ethnic status, n (%)  Caucasian 161 (90.5) Other 17 (9.5) HBQ scores, mean (SD)  Internalizing 2.0 (0.6) Externalizing 1.8 (0.5) ADHD 2.4 (0.8) Global Physical Health 4.8 (0.9) ADHD, attention-deficit/hyperactivity disorder; HBQ, health and behavior questionnaire; USD, US dollars  4.3.2 Measures 4.3.2.1 Maternal depression Maternal depression symptom levels were measured using the Centre for Epidemiological Studies – Depression (CES-D) scale 115. The CES-D measures the presence and severity of depressive symptoms and ranges from 0 – 60; scores greater or equal to 16 are considered “at risk” for depression. Prenatal maternal depression levels were measured during 46  the second trimester of pregnancy. CES-D scores were also assessed at multiple postnatal time points when children were 1 month, 4 months, 12 months, 3.5 years, and 4.5 years old, and from kindergarten through grades 1, 3, 5, 7 and 9. Latent class mixture modelling using the “lcmm” package 116 was conducted on these scores to identify postnatal maternal depression trajectories. A model with three classes showed the best fit and was used to characterize postnatal maternal depression patterns in this study.    4.3.2.2 Demographic variables Demographic information on mothers was based on self-report measures collected at the time of recruitment into the study. Maternal age and years of formal education were recorded in complete years. Income was assessed as total family income, and ethnic status was categorized as Caucasian or Other.   4.3.2.3 Genetic ancestry Principal Component Analysis (PCA) of the genotype data was conducted; the first principal component accounted for more than 75% of the variance in the dataset and was used as a measure of genetic ancestry. The first principal component (PC) from a PCA of genotype data was used to measure genetic ancestry (Figure 4.1). 47   Figure 4.1. Scree plot showing the proportion of variance explained by the first 20 principal components of genotype data.   4.3.2.4 Outcomes at 18 years The self-report version of the Health and Behavior Questionnaire (HBQ) 117 was completed by child study participants during grade 12. The HBQ is an instrument that captures child functioning and impairment, and has been shown to be sensitive to internalizing disorders 118. A measure for global physical health is provided, as are subscale scores for mental health. One individual was missing data on the HBQ internalizing subscale; the mean score was used to impute data.   4.3.3 DNA methylation data  Buccal swabs were collected when children were 18 years old using MasterAmp Buccal Swabs (Epicentre Biotechnologies, Madison, WI). DNA was extracted using Isohelix Buccal DNA Isolation Kits (Cell Projects, Harrietsham, Kent, England) and purified and concentrated using DNA Clean & Concentrator (Zymo Research, Irvine, CA). 750ng of BEC DNA underwent 48  bisulfite conversion using EZ DNA Methylation Kit™ (Zymo Research, Irvine, CA). 180ng of bisulfite-converted DNA was processed with the Illumina Infinium HumanMethylation450 BeadChip (450K array) (Illumina Inc, San Diego, CA) per manufacturer's instructions. Microarray chips were scanned in two batches on an Illumina HiScan; a pair of technical replicates scanned across both batches showed a Pearson r > 0.99, indicating high reproducibility. Background-subtracted and color-corrected 450K array data were exported from Illumina’s GenomeStudio software; all subsequent processing steps and analyses were performed in R. Two outlier samples were detected using the “detectOutlier” function in the lumi package 119 and removed; investigation into these samples showed low quality DNA yielded during initial DNA extraction. 15 overlapping SNP probes between the 450K and PsychChip arrays were compared and showed a high correlation for all probes (Pearson |r| > 0.99), indicating no discrepant samples. Of the original 485,577 probes, a total of 73,502 were removed, corresponding to those that failed in at least 5% of samples, were not detected above background in at least 5% of samples, mapped to SNPs and the X/Y chromosomes, and were polymorphic for SNPs and cross-hybridizing probes 120, leaving 412,075 probes. Data were quantile- and SWAN-normalized to ensure inter-sample comparability and correct for probe-type differences, respectively 119,121. Finally, data were corrected for plate, chip position, and chip effects in successive iterations of ComBat 122.  Investigation of potential blood contamination on DNAm data was conducted using blood spike-in controls and PCA. 450K array data on the peripheral blood mononuclear cells (PBMC) of 21 individuals was obtained from the Gene Expression Omnibus (GSE58888). Samples were selected to be demographically comparable to the WSFW cohort; they came from 49  normal controls of Western European descent, ranging from 15 – 25 years of age. PCA was conducted on the combined PBMC and WSFW buccal cell DNAm data to detect potential blood contamination of WSFW buccal DNAm data. Examination of the top 2 principal components (PCs) from the PCA of the combined buccal and PBMC dataset revealed distinct clustering between the buccal epithelial cell (BEC) and PBMC samples, suggesting minimal blood contamination (Figure 4.2).    Figure 4.2. PC1 scores plotted against PC2 scores per sample for the combined buccal and PBMC DNA methylation dataset.   4.3.4 Genotype data DNA from saliva samples was extracted by ethanol precipitation. DNA was purified and concentrated using DNA Clean & Concentrator (Zymo Research, Irvine, CA). 200 ng of DNA were amplified and hybridized to the Illumina PsychChip SNP genotyping array (PsychChip) (Illumina Inc, San Diego, CA) per manufacturer's instructions. Chips were scanned in two batches using the Illumina iScan. 50  Genotype data was imported into Illumina’s GenomeStudio using default parameters. Genotype calls were generated using Illumina cluster definitions (PsychArray_A_ClusterFile.egt) with a Gencall threshold of 0.15.  Genotype data were filtered using PLINK. SNPs with a minor allele frequency <5%, >5% genotype data missing, and a Hardy-Weinberg equilibrium p-value <0.001 were removed, leaving a total of 256,934 out of the original 588,455 probes for further analysis.   4.3.5 Pyrosequencing  152 samples taken at age 18 with remaining BEC DNA underwent pyrosequencing at the CALCB locus to verify results from the 450K array technology. Primer sequences can be found in Supplementary Table A.7.  4.3.6 Statistical methods 4.3.6.1 Model building Child sex, genetic ancestry, and maternal postnatal depression trajectories were included a priori as covariates. To determine additional covariates, variables were examined for their associations with prenatal maternal depression symptomatology and the first 9 PCs of the DNAm data, as subsequent PCs accounted for little variance in the data (less than 1%) (Figure 4.3). Analyses were conducted using Pearson correlations or ANOVA for continuous variables and t-tests for categorical variables. Only internalizing symptoms at 18 years associated with both prenatal maternal depression levels and DNAm at a p<0.1 (Figure 4.4) and was further included in statistical analyses.  51   Figure 4.3. Scree plot showing the proportion of variance explained by the first 20 principal components of the DNA methylation data.    Figure 4.4. Heat maps showing associations of potential confounding variables with each other and with the first 9 principal components of the DNA methylation data.   52  4.3.6.2 Differential DNA methylation analyses Associations between prenatal maternal depression levels and regions of DNAm were tested using “DMRcate” 123. DMRcate is a method that agglomerates adjacent correlated CpGs using a Gaussian kernel smoothing approach 123. Differentially methylated regions (DMRs) were defined as those with a minimum of 2 CpG sites, at least one of which mapped to a gene based on the 450K UCSC-based annotation. A false discovery rate-corrected p-value (FDR) of 0.05 was used as the cut-off for statistical significance. Probe-wise linear regression analyses were also conducted with “limma” to supplement the regional analysis.    4.3.6.3 Testing for effects of postnatal maternal depression    To differentiate the effects of exposure to prenatal versus postnatal maternal depression levels on DNAm, a model with postnatal maternal depression symptomatology as the explanatory variable, adjusted for prenatal maternal depression and covariates, was also run using DMRcate 123. Underlying probes from separate analyses with prenatal and postnatal depression levels as explanatory variables were compared.   4.3.6.4 Testing for effects of genotype    Methylation quantitative trait loci (meQTL) refer to CpG sites where DNAm is associated with DNA sequence variation 124. MeQTL analysis was conducted based on proximity. Though wide variation in definitions exist, a CpG was considered to be regulated in cis if it was located within 1 million base pairs (bp) of a SNP; otherwise, CpGs were considered to be in trans 125. I modelled simple linear regressions of each DNAm-SNP pair to find meQTLs. 53  Bonferroni-corrected p-value cut-offs were used to determine significant meQTLs for both cis- (p<3.02 e-7) and trans-meQTL (p<5.57 e-10) analyses.   4.4 Results 4.4.1 Differentially methylated regions associated uniquely with exposure to prenatal maternal depression  Using a region-based approach, 45 DMRs were identified, corresponding to 350 unique CpGs, that significantly related to prenatal maternal depression levels in BECs (Table 4.2). Probe-wise analyses on the same dataset did not reveal any individual CpGs significantly associated with prenatal maternal depression symptomatology after stringent correction for multiple testing (FDR<0.05). However, the top 350 most significantly associated CpGs (p<0.0001) from the probe-wise analysis overlapped significantly with those represented in the DMRs (p<0.0001, Figure 4.5).  54  Table 4.2. Regions of DNA methylation significantly associated with prenatal maternal depression symptom levels. Associated gene name(s) Chromosome Start position Stop position Region size # of CpGs FDR-adjusted p-value Mean beta fold change (region) Max absolute delta beta (probe) PRSS22 * 16 2907517 2908934 1417 12 1.54E-07 0.0010 0.0365 CALCB *§ 11 15093613 15095978 2365 14 2.46E-07 0.0007 0.0111 DDO *§ 6 110720501 110721629 1128 6 1.20E-05 0.0020 0.0332 C1orf86,LOC100128003 *§ 1 2125049 2125979 930 9 1.21E-05 0.0012 0.0432 FN3K,TBCD *ǂ§ 17 80708279 80709150 871 5 4.96E-05 0.0023 0.0952 MOV10L1 22 50528179 50528752 573 12 4.96E-05 0.0012 0.0588 PLEC1,MIR661 8 145018816 145019839 1023 17 5.46E-05 0.0009 0.0392 HOXA2 7 27140797 27141388 591 6 5.96E-05 0.0021 0.0160 FAM134B *§ 5 16508920 16509323 403 5 9.03E-05 0.0037 0.0949 USP29 * 19 57630034 57630742 708 16 9.03E-05 0.0008 0.0336 BAT4,CSNK2B * 6 31631638 31632320 682 11 1.14E-04 0.0010 0.0469 SLC22A18AS,SLC22A18 § 11 2922051 2922852 801 13 2.95E-04 0.0009 0.0479 ZSCAN12L1 *§ 6 28058715 28059208 493 10 3.18E-04 0.0013 0.0175 COX4I2 20 30225325 30226046 721 9 3.43E-04 0.0012 0.0149 ZFYVE28 * 4 2341196 2341463 267 5 3.44E-04 0.0018 0.0402 SLC19A1 * 21 46961395 46961669 274 2 5.03E-04 0.0003 0.0153 DAXX ǂ§ 6 33288561 33289719 1158 35 6.68E-04 0.0008 0.0393 FLJ44606 *§ 5 126408756 126409553 797 13 9.32E-04 0.0012 0.0195 DVL1 § 1 1272398 1272559 161 4 0.0024 0.0005 0.0053 55  IGF2BP1 * 17 47113061 47113596 535 5 0.0024 0.0010 0.0325 SETD6 *§ 16 58548873 58549251 378 6 0.0027 0.0007 0.0200 NR1D1,THRA * 17 38249266 38249729 463 6 0.0046 0.0007 0.0171 SKIV2L,STK19 § 6 31936446 31937470 1024 20 0.0048 0.0004 0.0272 SGSM2 * 17 2265510 2265875 365 2 0.0071 0.0005 0.0109 LOC391322 * 22 24373484 24373618 134 2 0.0075 0.0016 0.0051 HEXIM2 17 43246883 43247290 407 4 0.0081 0.0009 0.0111 ITGAE 17 3697993 3698239 246 4 0.0108 0.0011 0.0183 CIZ1 § 7 158280410 158280939 529 4 0.0115 0.0018 0.0435 MEGF6 *ǂ 1 3407936 3408036 100 2 0.0118 -0.0008 0.0104 CIZ1 9 130955135 130955436 301 3 0.0126 0.0024 0.0253 ZAP70 ǂ 2 98351009 98351548 539 5 0.0126 0.0012 0.0200 MX2 21 42733397 42733894 497 7 0.0135 -0.0008 0.0244 PTPRN2 * 7 158059704 158060003 299 5 0.0135 0.0034 0.1423 GPR81 * 12 123214864 123215471 607 9 0.0136 0.0006 0.0149 HAS2AS,HAS2 * 8 122651271 122651575 304 7 0.0207 -0.0006 0.0164 C21orf125 21 44869258 44869885 627 6 0.0214 0.0008 0.0176 ZG16B § 16 2879393 2879544 151 2 0.0252 0.0020 0.0065 DPEP1 § 16 89686618 89687052 434 5 0.0268 0.0014 0.0253 GJC2 1 228337181 228337776 595 7 0.0281 0.0012 0.0753 PTPRN2 7 158075624 158075705 81 2 0.0307 0.0018 0.0313 IKZF1 7 50468399 50468722 323 2 0.0340 0.0017 0.0052 STK19,DOM3Z 6 31938744 31938891 147 8 0.0360 0.0007 0.0288 HOXA5 7 27183794 27184159 365 16 0.0451 0.0006 0.0345 DOK6 18 67069959 67070170 211 4 0.0460 -0.0007 0.0458 56  MIOX 22 50925232 50925337 105 3 0.0485 0.0005 0.0086 * DMR contains probe(s) represented in top 350 statistically significant sites from probe-wise analyses ǂ DMR contains probe(s) within DMRs significantly associated with postnatal depression § DMR contains probe(s) represented in meQTLs  CpG, cytosine-guanine dinucleotide; FDR, false discovery rate   Figure 4.5. Histogram showing the expected number of overlapping probes between the most significant 350 CpG sites from probe-wise analyses and the 350 probes underlying DMRs, based on 10,000 simulations. The dotted red line represents the actual number of overlapping probes. 57    A number of genes that map to DMRs are noteworthy for their known or putative functions. These include CALCB, a neurotransmitter whose closely related alpha analogue has been implicated in fetal programming of depression in a mouse model 126, as well as the HOXA2 and HOXA5 imprinted genes, which play important roles in central nervous system development 127. The presence of SLC19A1, a folate transporter, among the DMRs is interesting for two reasons: folate status during pregnancy is an important determinant of neonatal outcomes 128 and folate is also a required precursor for DNAm 129. Additionally, that SETD6 is among the DMRs is intriguing as its gene product is involved in methylating the histone variant H2A.Z, which has been shown to be responsive to environmental and developmental stimuli 99. Consistent with expectations 107, effects of prenatal maternal depression levels on DNAm were small (mean absolute difference per DMR for all probes = 0.03); however, a difference of up to 0.14 was seen, corresponding to a probe-level difference of 14%, from the lowest to the highest prenatal maternal depression score (Table 4.2). The majority of DMRs (91.1%) showed an overall positive direction of association between DNAm and prenatal maternal depression symptomatology. DNAm differences for the six most statistically significant DMRs are visually represented in Figure 4.6a. To verify technical robustness of DNAm values, pyrosequencing experiments were conducted on the CpGs represented in the CALCB locus; these values were consistent with measurements from the 450K array, suggesting technical differences do not explain the associations that were observed (Figure 4.6b). 58    Figure 4.6. Regions of DNA methylation significantly associated with prenatal maternal depression exposure at 18 years of age. a) Plots showing the top 6 most statistically significant differentially methylated regions. Circles show each DNA methylation value per individual at each probe. For the purposes of regional visualization, individuals were divided into “high” or “low” prenatal maternal depression exposure based on a CES-D score of 16. Solid lines represent mean DNA methylation levels per group at each probe. * indicates sites that were represented in pyrosequencing experiments. b) Pyrosequencing verified DNA methylation levels in the CALCB gene locus. DNA methylation values against prenatal maternal depression scores are shown from pyrosequencing experiments (top) and from the 450K array (bottom). Solid blue lines represent unadjusted regression lines.  59  4.4.2 Differentially methylated regions showed unique associations with prenatal, and not postnatal, maternal depression levels To assess whether exposure to prenatal maternal depression levelshad a unique DNAm signature that was independent of postnatal maternal depression symptoms, I compared the probes that were associated with prenatal versus postnatal exposures. Of the 45 DMRs, only four contained probes that were also significantly associated with postnatal maternal depression levels (Table 4.2) suggesting a unique association specific to prenatal exposure.   4.4.3 Genotype influenced differentially methylated regions I modelled associations between the 350 CpGs underlying DMRs with the 256,934 SNPs represented in the genotype data. 892 meQTLs were identified, mapping to 21 (44%) of the 45 DMRs. Of the 892 meQTLs, 483 (54.1%) and 409 (45.9%) were located in-cis and in-trans (Table 4.3). In total, meQTLs localized uniquely to 76 CpGs and 226 SNPs (Table 4.3); however, cis-meQTLs localized uniquely to 71 CpGs and 94 SNPs whereas trans-meQTLs localized uniquely to 15 CpGs and 132 SNPs (Table 4.3), suggesting greater polygenic influences on the latter. On average for cis-meQTLs, SNPs were located 141 kb away from CpGs with an effect size of -0.007; for trans-meQTLs, there was a mean effect size of 0.016 (Table 4.3).    60  Table 4.3. Overview of methylation quantitative trait loci within differentially methylated regions.  All Cis Trans meQTLs, n (%) 892 (100) 483 (45.9) 409 (54.1) Unique CpGs, n (%) 76 (100) 71 (93.4) 15 (19.7) Unique SNPs, n (%) 226 (100) 94 (41.6) 132 (58.4) Represented DMRs, n (%) 20 (100) 19 (95.0) 5 (25.0) Distance between CpG and SNPs (kb), mean (sd) NA 141.2 (216.2) NA Mean effect size 0.004 -0.007 0.016  4.5 Discussion I investigated whether fetal exposure to maternal depression symptomatology associates with DNAm 18 years later. 45 unique genomic regions were identified where BEC DNAm levels were significantly associated (FDR<0.05) with prenatal maternal depression symptoms, and a large proportion (44%) were under genetic influence. Collectively, these results suggest that associations between prenatal maternal depression levels and DNAm may be due to genetic factors, and suggest that heredity may contribute to associations that are observed between prenatal maternal depression levels and DNAm differences in offspring.  That a genetic influence on DNAm patterns associated with prenatal depression symptomatology was found is not unexpected, given the genetic heritability of depression and other mood disorders. However, this explanation requires further investigation within study designs that also collected parental genotype, both for mothers and fathers. Moreover, the fact that CpGs were selected for meQTL analysis based on their relation to prenatal maternal depression (which may already inherently incorporate a genetic component) may also explain the genetic influence that was found. Overall, current study findings are consistent with previous research that found that gene-by-in-utero-environment interactions explained a majority of the 61  variation seen in DNAm patterns 113. Future EWAS should endeavor to incorporate genotype in order to understand the respective influence of prenatal maternal depression exposure and genetics on DNAm 125,130 in addition to understanding the relative contributions of shared genetic background between mother, father, and child which could help clarify underlying biological reasons for differences in child outcomes.  Overall, detected associations with prenatal maternal depression exposure on DNAm were subtle, which is in line with expectations 107. However, a number of genes associated with DMRs were of particular interest due to their putative involvement in biologically relevant processes. Though it is tantalizing to speculate that differences in DNAm may mediate the observed relationship between prenatal maternal depression symptoms and internalizing symptoms, study results do not support this hypothesis given the large contribution of genetic background to observed differences in DNAm patterns.  The interpretation of current study findings is that detected BEC DNAm patterns may reflect genetic associations related to the heredity of depression rather than being functionally related to pathways to mental health. It is important to interpret study findings in light of a number of limitations. As with all observational studies, the results reported herein are associational rather than causal; it is possible that other unknown sources of confounding, such as environmental differences, may explain our results. As mentioned previously, the use of peripheral tissue is a limitation given that the brain represents the primary tissue of relevance for studies investigating the fetal programming of mental health, though BECs are believed to represent a better proxy for brain DNAm than blood 131.  Another limitation in the analytic strategy was that any potential interactive effects of genotype and prenatal maternal depression levels on DNAm (ie, a gene-by-environment 62  approach) were not investigated as other studies have done 113. Rather, the focus was on identifying main effects of prenatal maternal depression exposure on DNAm, then quantifying the proportion of DMRs influenced by genotype to help tease apart the respective effects of genetics and environment on DNAm 132. It is possible that alternative analytic strategies may yield different findings.  Finally, the approach taken in this study may be instructive for future studies. The methodology included an epigenome-wide approach (both regionally and probe-wise), careful examination of known confounders in EWAS (e.g. genotype, cell-type differences, concurrent behavioral outcomes), and incorporation of genetic information in relation to identified DNAm patterns. These aspects of the current study represent key strengths that provide a solid foundation for future research to build upon and replicate.    63  Chapter 5: Maternal depression trajectories from pregnancy to three years postpartum and their associations with children’s behavior and executive functions in childhood   5.1 Synopsis Background: The purpose of this study was to investigate how pre- and postnatal patterns of maternal depressive symptoms (from mid-pregnancy to 3 years postpartum) are associated with children's behavior at age 3 years and executive functions (EFs) and mental health indicators at age 6 years.  Methods: Data were collected from participants in a Vancouver-based longitudinal birth cohort. Maternal depressive symptoms were measured from mid-pregnancy to 3 years postpartum. Growth mixture modelling was used on standardized maternal depression scores (n=147) to identify trajectories. Children's behavioral problems and mental health symptomatology (internalizing, externalizing, and attention deficit hyperactivity disorder) were obtained at 3 and 6 years. EFs were assessed by a laboratory-based computerized task and maternal-report at 6 years. Multivariable linear regressions of children's outcomes against maternal depressive symptom trajectories were conducted (n=103).  Results: Three distinct patterns of maternal depressive symptom trajectories were identified: low (n=105), increasing (n=27), and decreasing (n=15). Children of mothers whose depressive symptoms increased reported more problem behaviors at 3 years, poorer EFs at 6 years as assessed by both instruments, but no significant differences in mental health symptomatology at 6 years, relative to those whose mothers had consistently low depressive symptoms. Children 64  whose mothers became less depressed over time had comparable levels of behavioral problems at age 3, executive functions and internalizing and externalizing scores at age 6, and fewer reported ADHD behaviors at age 6, than those whose mothers remained less depressed over time.  Conclusion: Children whose mothers' depressive symptoms improve over the first 3 years postpartum appear to have a comparable outlook to those whose mothers had consistently low depressive symptoms. However, these results require replication in independent cohorts.   5.2 Introduction   The previous two chapters have explored potential effects of prenatal maternal depression and antidepressant exposure on child development. However, evidence suggests that exposure to maternal depression, both pre- and post-natally, shapes pathways to health and disease for offspring throughout life. By taking a longitudinal perspective to characterize children’s exposures to maternal depression during pregnancy and the first three years postpartum, this chapter explores the impact of ongoing exposure to postnatal maternal mood on child developmental outcomes.   Despite strong evidence supporting an adverse developmental impact of perinatal maternal depression, disentangling the effect of prenatal versus postnatal maternal depressive symptoms remains challenging. A recent review of population-based longitudinal studies on maternal depression found a substantial number of individuals alternated between being classified as depressed or non-depressed across pre- and post-natal periods, and suggested that postpartum depression may represent a continuation of prenatal depression 8. Moreover, it may 65  be the severity and/or chronicity of maternal depression, rather than timing of exposure, that plays a critical role in influencing children’s mental health and development 133,134.  Traditional approaches to studying the effects of maternal depression on children have focused on depression at specific time points or have used predefined cut-points to assess severity and chronicity when using longitudinal data 133-135. However, these methods cannot separate the timing, severity, and chronicity of maternal depressive symptoms 136. To overcome these limitations, recent research efforts have identified trajectories of maternal depression using longitudinal data. Trajectory-based approaches are advantageous as they are able to capture the heterogeneity of depressive symptoms over time 137 and may thus better represent a child’s exposure to his or her mother’s depression across key developmental periods.   To date, several studies have modelled maternal mental health trajectories and their impact on children 136,138-145, and have demonstrated that modelling trajectories provides additional predictive value compared to traditional approaches. However, only a subset of these studies has examined trajectories starting during pregnancy 136,140-143, which as a critical period of development, should be included when determining the impact of maternal depression. Moreover, these studies have all assessed outcomes in childhood at just one point in time, although tracking outcomes as children age is important, as disturbances may arise differently depending on developmental status 146.  Conflicting findings have emerged from this body of work. Some studies have found that maternal mental health trajectories characterized by consistently high symptoms are especially harmful for children’s outcomes 136,141, while others have found that adverse effects on offspring are seen only when the timing of symptoms occurs during gestation 142,143. Finally, there have been no reports of the impact of maternal depression trajectories on children’s executive 66  functions despite the consistent finding that cognitive function is affected 35. Of important consideration regarding executive functions is that they may be particularly sensitive to maternal depression due to their vulnerability to environmental inputs 147.   The current study focused on identifying distinct trajectories of maternal depressive symptoms from pregnancy through the next 3 years, identifying whether maternal characteristics differ across different depressive trajectories, and investigating how these identified trajectories relate to children’s behavior at age three years and children’s mental health symptomatology and executive functions at six years of age. We expected that children of mothers whose trajectories reflected higher persistent or even variable levels of depressive symptoms relative to that with the least depressive symptoms would exhibit poorer behavioral, cognitive, and mental health outcomes at both three and six years.   5.3 Methods 5.3.1 Study design 191 women were recruited (from 258 approached) during their second trimester of pregnancy to participate in a prospective cohort study of the developmental impact of in utero selective serotonin reuptake inhibitor (SSRI) antidepressant exposure and maternal mental health, and both exposed and unexposed women were recruited. Women were physician- or self-referred from the Reproductive Mental Health Program at the British Columbia Women's Hospital and Health Centre, family physician practices, and community midwife clinics in the Vancouver Metropolitan area in British Columbia, Canada. The Reproductive Mental Health Program provides specialized perinatal psychiatric services for women who have been referred to the program by a physician, nurse practitioner, or midwife due to mental health concerns. Given 67  the convenience sampling strategy and the fact that the study aimed to investigate the effects of prenatal exposure to SSRIs and maternal mood disorders on offspring, study participants were over-sampled from women at high risk for mental disorders. Criteria for inclusion in the study included singleton pregnancy, confirmed gestational age at birth, and lack of fetal anomalies as detected by ultrasound. Exclusion criteria included the presence of maternal bipolar disorder or drug abuse, obstetrical complications (i.e. diabetes, hypertension) or fetal conditions (e.g. preterm birth, anencephaly). Data were collected over two waves of recruitment from January 2002 until April 2015.  5.3.2 Measures 5.3.2.1 Maternal depression  The Hamilton Depression Rating Scale 148 (HAMD) and the Edinburgh Postnatal Depression Scale 145 (EPDS) were used to assess maternal depressive symptoms during the second and third trimesters of pregnancy, and at six weeks, three months, six months, and 10 months. The Beck Depression Inventory 146 (BDI) was used at the three-year follow-up and the HAMD was used at the six-year follow-up. The HAMD is a 21-item clinician-rated scale with a range of zero to 63; all assessments were performed by trained research assistants. The BDI and EPDS are self-report questionnaires, the former with 21-items and a range of zero to 63, and the latter with 10-items with scores from zero to 30. All three instruments measure the existence and severity of depression symptoms. Cut-off scores signifying risk of meeting criteria for a major depressive episode for the HAMD, EPDS, and BDI, respectively, are 17, 13, and 13 61,149,150.  Different waves of the cohort underwent assessments for maternal depression at six weeks and 10 months.  68   5.3.2.2 Child behavior at three years The Child Behaviour Checklist 151 (CBCL) was completed by mothers when their children were three years old. The CBCL measures the degree of children’s emotional, behavioral, and social problems. Sex-specific standardized T-scores were used to represent the total score of problem behaviors and internalizing and externalizing subscales, with higher scores indicating more behavioral problems.   5.3.2.3 Child mental health symptomatology at six years The MacArthur Health and Behavior Questionnaire 152 (HBQ) yielded measures of children’s mental health symptomatology by maternal report at 6 years. Designed to map onto the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III diagnostic criteria 153, it has been validated as a sensitive instrument for detecting psychopathology in children 154.  Subscale scores for three domains of symptomatology are provided: internalizing (symptoms of depression, separation anxiety, and generalized anxiety), externalizing (oppositional defiance symptoms and conduct problems), and attention-deficit/hyperactivity disorder (ADHD) (inattention, impulsivity, and hyperactivity symptoms).   5.3.2.4 Child executive functions at six years Two instruments were used to measure executive functions. The parent-report version of the Behavior Rating Inventory of Executive Function 155 (BRIEF) was completed by mothers when their children were 6 years old. The BRIEF yields a score of global executive functions, with higher scores indicating worse outcomes. Children also completed the computerized Hearts 69  and Flowers (HF) task 156 at age 6 years. The HF task is conducted over three blocks and is designed to assess the executive function domains of inhibitory control, working memory, and cognitive flexibility. A stimulus appears to the right or left of a computer screen on every trial.  On Block 1 of the task (the congruent block), participants have only to do what comes naturally (i.e., pressing on the same side as the stimulus); executive functions are not taxed. On Block 2 (the incongruent block), participants have to resist that prepotent response and instead press on the side opposite the stimulus.  On Block 3 (the mixed block), the two types of trials are randomly intermixed, requiring remembering both rules, mentally translating “same [or opposite] side” into “right [or left] hand,” and flexibly switching between the two rules, inhibiting one to apply the other. Children arrived at the study centre in the morning and were oriented to the task and computerized set-up by performing a simple choice reaction time task. They were also given practice trials to make sure they understood the task demands.  Reaction time and percent accuracy were computed for each block, with calculations performed as previously described 60. Given past work suggesting that meaningful differences on the HF task among normative populations emerge in the most demanding block 60, percent accuracy in the third block was used as an outcome measure in this study.   5.3.3 Cohort characteristics  Maternal age, education, alcohol consumption, SSRI use, and child demographic factors were recorded by self-report. SSRI medications that were taken prenatally by mothers included paroxetine, fluoxetine, sertraline, venlafaxine, and citalopram. Minority status was coded as “yes” or “no” based on whether participants identified as Caucasian/white or another ethnicity. 70  Maternal history of depression was based on any prior diagnosis of major depression. Maternal history of mental disorders was based on any prior diagnoses of the following: major depression, substance abuse, mania, obsessive compulsive disorder, anxiety or panic attacks, eating disorders, and suicide attempts. Number of previous mental disorders was recorded as the total number of the previously listed conditions that had been diagnosed. Other psychotropic drug use was classified as “yes” or “no” based on whether mothers were taking other psychotropic medications in addition to SSRIs while pregnant, and included the following drugs: clonazepam, ativan, imovane, seroquel, stemetil, and loxepine.   5.3.4 Statistical analyses 5.3.4.1 Analytic samples From the overall cohort, two study samples were derived. The first analytic sample consisted of mothers with complete data on maternal depressive symptoms at the three-year time point (n=147), and this sample was used to compute maternal depression trajectories. The second and final analytic sample was composed of the subset of mother-child dyads with complete data on child outcomes at ages three and six years (n=103); this sample was used to calculate associations between maternal depression trajectories and child outcomes. Comparison of prenatal maternal depression scores from the original cohort (n=182 with HAMD scores, HAMD = 8.8) compared to the final analytic sample (n=103, HAMD=7.9) suggested that although selection bias may have been introduced by more depressed women dropping out of the study, differences in illness severity were minimal and not clinically meaningful. Comparisons of maternal age and education also showed minimal differences between the original cohort (n=182 71  with data, mean age=33.3 years,  education=17.1 years) and final analytic sample (n=103, mean age=33.2 years, education=17.3 years).  5.3.4.2 Data imputation Mothers with complete data on maternal depressive symptoms at the three-year time point (n=147) comprised the sample used to compute trajectories. Three mothers were missing information on alcohol consumption and one on maternal education; the means of these variables were used to impute and replace missing values. These imputed values were carried over into the next set of analyses conducted on the subset of mother-child dyads with complete data on maternal depressive symptom trajectories and child outcomes at three and six years (n=103).  5.3.4.3 Growth mixture modelling The first step was to compute trajectories of maternal depressive symptoms from the second trimester of pregnancy until three years post-birth. To meaningfully compare the different depression measures, scores at each time point were standardized to produce Z-scores and were averaged when there were multiple measurements per time point. I applied growth mixture modelling (GMM) to these scores using the “lcmm” package in R 116. GMM is a person-centered analytic method that allows for the detection of unobserved yet distinct trajectories within a population 137,157. The expectation in using GMM was to identify latent classes that captured the overall meaning and direction of unique trajectories within the data 157.  Missing data on maternal depressive symptoms (prenatal prior to three years post-partum) were handled under the missing at random assumption. The Bayesian Information Criterion (BIC) and mean posterior probabilities of group membership for all groups were 72  examined to guide model selection 158. Analyses were successively re-run with data omitted from the six-week and ten-month follow-up time points, as these data were only captured in alternating waves of the cohort, to assess the robustness of identified trajectories.    5.3.4.4 Identifying factors associated with maternal depressive symptom trajectories The second step was to identify characteristics associated with maternal depressive symptom trajectories. Differences between trajectory groups with respect to continuous and categorical variables were compared using analysis of variance (ANOVA) and chi-square tests; significant differences were analyzed using pairwise post-hoc tests (Tukey’s Honest Significant Difference test or pairwise chi-square tests with a family-wise error rate of p<0.05 to detect statistical significance).  5.3.4.5 Multivariable linear regression analyses The third stage of analysis was to estimate associations between maternal depressive symptom trajectories and child outcomes using multivariable linear regressions. First, the association between maternal trajectories and child behavioral problems at three years, then associations with children’s internalizing, externalizing, and ADHD symptoms, and executive functions on the BRIEF and HF task, at six years were estimated. Two models were fit for each outcome: unadjusted and adjusted. Adjusted regression models included potential confounding variables selected based on previous literature: child’s sex, age, gestational age at birth, birthweight, prenatal SSRI antidepressant exposure, maternal history of depression, maternal education, and maternal minority status. Concurrent maternal depressive symptoms were included as a covariate in all models examining six-year outcomes.  73   5.3.4.6 Sensitivity analysis Given the study’s small sample size and risk for overfitting based on correlations among confounding variables, regression analyses were fit using a more parsimonious model that excluded gestational age at birth and maternal history of depression as covariates.   5.4 Results 5.4.1 Identification of maternal depressive symptom trajectories A model with three trajectories of maternal depressive symptoms from the second trimester of pregnancy until three years postpartum demonstrated the best fit relative to models with two or four trajectories (BIC as the number of trajectories increased from two to four was 1905.3, 1882.7, and 1897.7). For the three-trajectory model, the mean posterior probability of being assigned to each trajectory was high for all groups (range 0.84 – 0.96) and above the recommended threshold of 0.8 159. Analysis of the dataset excluding the six-week and 10-month time points also yielded results consistent with a three-trajectory model demonstrating the best fit.  The three maternal depressive symptom trajectories are represented in Figure 5.1. Mean depressive scores at each time point per trajectory can be found in Table 5.1. The “low” trajectory represented 71.4% (n = 105) of mothers and was characterized by consistently low depressive symptom scores. The “increasing” trajectory consisted of 18.4% (n = 27) of mothers and reflected a pattern of moderate depressive symptoms during pregnancy with increasing symptomatology over time. The “decreasing” trajectory, comprising 10.2% of mothers (n = 15), was characterized by high levels of depressive symptomatology during pregnancy that 74  subsequently decreased.  Notably, most moms in the study sample were in the low group, a finding consistent with previous research 136,139. As expected, less women in the low trajectory were taking SSRIs prenatally and postnatally compared to the other trajectories (Table 5.1).   Figure 5.1. Maternal depression trajectory groups. Trajectories were visualized using a locally weighted regression smoothing approach. Shaded bars represent 95% confidence intervals.  Table 5.1. Depression scores by trajectory group from the 2nd trimester of pregnancy until three years postpartum.  Low n = 105 Increasing n = 27 Decreasing n = 15 Depression Z-score, mean (SD)    2nd trimester -0.52 (0.56) 0.58 (0.75) 1.63 (0.54) 3rd trimester -0.51 (0.55) 0.83 (0.64) 1.30 (0.87) 6 weeks -0.36 (0.57) 0.54 (0.60) 1.85 (1.46) 3 months -0.28 (0.75) 0.87 (0.84) 0.84 (0.89) 6 months -0.47 (0.53) 1.12 (0.72) 0.82 (0.63) 75  10 months -0.48 (0.48) 0.74 (1.01) 0.87 (0.55) 3 years -0.43 (0.52) 1.61 (0.77) 0.15 (0.96)  Raw depression score, mean (SD)    2nd trimester (EPDS) 3.63 (3.42) 10.08 (4.77) 15.87 (3.80) 3rd trimester (EPDS) 3.40 (2.91) 9.92 (3.67) 12.14 (4.88) 6 weeks (EPDS) 3.90 (3.12) 7.81 (3.12) 12.17 (5.78) 3 months (EPDS) 3.32 (3.37) 9.81 (4.83) 9.93 (4.33) 6 months (EPDS) 3.41 (3.05) 12.19 (4.30) 9.86 (3.32) 10 months (EPDS) 3.22 (2.74) 8.45 (4.91) 9.88 (2.03) 3 years (BDI) 4.62 (4.36) 21.78 (6.45) 9.53 (8.07) BDI, Beck Depression Inventory; EPDS, Edinburgh Postnatal Depression Score  5.4.2 Maternal characteristics by trajectory group  Women in the low trajectory group had completed more years of education than women in the increasing group (Table 5.2); they also had lower proportions of prior diagnoses of depression or history of mental disorders, plus the lowest mean number of previous mental disorders, relative to those in the increasing and decreasing groups (Table 5.2). Women in the low group were less likely to have taken SSRIs during pregnancy than those in the other groups. This difference was still present between women in the low and increasing groups at 3 years postpartum (Table 5.2). Mean maternal age at child’s birth, minority status, alcohol consumption during pregnancy, and method of delivery did not significantly differ across trajectory groups (Table 5.2).   76  Table 5.2. Maternal characteristics by depression trajectory group (n = 147).  LowA n = 105 IncreasingB n = 27 DecreasingC n = 15 Significant group differences* Overall test statistic**, p-value Maternal characteristics      Age at birth (years), mean (SD) 33.5 (±4.8) 32.9 (±5.9) 31.7 (±4.5)  0.9, 0.43  Education (years), mean (SD) 17.9 (±3.6) 15.6 (±2.3) 16.8 (±3.5) B<A 5.2, 0.007 Minority status, n (%) Yes No  19 (18.1%) 86 (81.9%)  5 (18.5%) 22 (81.5%)  3 (20%) 12 (80%)   0.03, 0.98 History of depression, n (%) Yes No  33 (31.4%) 72 (68.6%)  18 (66.7%) 9 (33.3%)  11 (73.3%) 4 (26.7%) A<BC 17.6, <0.001  History of mental disorders, n (%) Yes No  50 (47.6%) 55 (52.4%)  23 (85.2%) 4 (14.8%)  14 (93.3%) 1 (6.7%) A<BC 20.6, <0.0001   Number of previous mental disorders, mean (SD) 1.0 (±1.3) 2.0 (±1.3) 2.4 (±1.6) A<BC 11.1, <0.0001   Delivery mode, n (%) Vaginal Caesarean section  81 (77.1%) 24 (22.9%)  17 (63.0) 10 (37.0%)  10 (66.7%) 5 (33.3%)  2.6, 0.27 Alcohol consumption during pregnancy, mean (SD) 3.9 (±7.7) 2.4 (±5.2) 4.9 (±6.6)  0.7, 0.52 SSRI use prenatally, n (%) Yes No  30 (28.6%) 75 (71.4%)    20 (74.1%) 7 (25.9%)  9 (60.0%) 6 (40%) A<BC 21.3, <0.0001  SSRI use at 3 years, n (%) Yes No  30 (28.6%) 75 (71.4%)  18 (66.7%) 9 (33.3%)  7 (46.7%) 8 (53.3%) A<B 13.9, <0.001  Other psychotropic drug use, n (%) Yes No  2 (1.9%) 103 (98.1%)  8 (29.6%) 19 (70.4%)  5 (33.3%) 10 (66.7%) A<BC 27.8, <0.000001 77  *Family-wise error rate of p<0.05. **Test is for overall differences. For ANOVA, F-statistic; for chi-square test, χ2-statistic. 78  5.4.3 Child characteristics by mother’s depressive symptom trajectory Child sex, age, minority status, gestational age and birth weight were not differentially distributed among maternal depressive symptom trajectories (Table 5.3). However, children in the low group were exposed to fewer maternal depressive symptoms than those in either the increasing or decreasing groups (Table 5.3).  79  Table 5.3. Child characteristics and outcomes by maternal depression trajectory group (n = 103).  LowA n = 76 IncreasingB n = 18 DecreasingC n = 9 Significant group differences*  Overall test statistic**, p-value Child characteristics      Sex, n (%) Male Female  35 (46.1%) 41 (53.9%)  6 (33.3%) 12 (66.7%)  2 (22.2%) 7 (77.8%)  2.5, 0.28 Age at 3 years, mean (SD) 3.6 (±0.6) 3.8 (±0.8) 3.6 (±0.3)  1.3, 0.28 Age at 6 years, mean (SD) 5.9 (±0.6) 6.1 (±0.7) 6.1 (±0.5)  1.3, 0.29 Minority status, n (%) Yes No  16 (21.1%) 60 (78.9%)  4 (22.2%) 14 (77.8%)  1 (11.1%) 8 (88.9%)  0.5, 0.77 Gestational age at birth (weeks), mean (SD) 39.8 (±1.4) 39.5 (±1.8) 38.7 (±1.7)  2.1, 0.13 Birth weight (grams), mean (SD) 3451.3 (±489.4) 3217.6 (±490.9) 3239.8 (±426.0)  2.2, 0.12 Exposure to maternal depression at 6 years (HAMD), mean (SD) 5.8 (±5.2) 13.4 (±5.1) 14.3 (±7.4) A<BC 21.2, <0.0000001  Child outcomes – 3 years  Total CBCL score, mean (SD) 44.63 (9.76) 55.00 (10.20) 44.44 (4.25) AC<B 8.85, <0.001 CBCL score – Internalizing, mean (SD) 45.18 (10.66) 55.44 (9.90) 47.89 (4.37) A<B 7.43, <0.001 CBCL score – Externalizing, mean (SD) 44.75 (9.79) 52.89 (10.90) 44.44 (6.06) A<B 5.22, 0.007 Child outcomes – 6 years  HBQ – Internalizing, mean (SD) 0.30 (0.22) 0.39 (0.28) 0.29 (0.23)  1.34, 0.27 HBQ – Externalizing, mean (SD) 0.28 (0.23) 0.32 (0.19) 0.26 (0.13)  0.32, 0.73 HBQ – ADHD, mean (SD) 0.63 (0.35) 0.61 (0.42) 0.34 (0.21)  2.6, 0.08 Executive function – Global BRIEF scores, mean (SD) 52.47 (8.35) 58.78 (9.17) 47.44 (7.57) AC<B 6.30, 0.002 Executive function – HF % correct on block 3, mean (SD) 0.68 (0.23) 0.62 (0.23) 0.64 (0.27)  0.58, 0.56 *Family-wise error rate of p<0.05. **For ANOVA, F-statistic; for chi-square test, χ2-statistic. CBCL, Child Behavior Checklist; HAMD, Hamilton Depression Scale; HBQ, Health and Behavior questionnaire; HF, Hearts and Flowers task 80  5.4.4 Maternal depressive symptom trajectories and child behavior at three years  At three years of age, children of mothers who had increasing depressive symptoms over time showed more behavior problems overall and more evidence of both internalizing and externalizing behaviors than children of mothers in the low group (β=10.3, p<0.001; β=9.1, p<0.01; β=9.7, p<0.001, respectively; Table 5.4). In contrast, the level of problem behaviors at age three was comparable between children of mothers in the low and decreasing trajectory groups, as were their levels on the internalizing and externalizing subscales (Table 5.4). 81  Table 5.4. Linear regression models estimating the effect of maternal depression trajectories on child outcomes at ages three and six years.   Low n = 76 Increasing n = 18 Decreasing n = 9    β (95% CI) t-value p-value β (95% CI) t-value p-value Age 3         Total CBCL score Unadjusted  Reference 10.4 (5.4–15.3) 4.2 <0.0001 -0.2 (-6.9–6.5) -0.06 0.96 Adjusted* Reference 10.3 (4.9–15.7) 3.8 <0.001 -0.6 (-7.8–6.6)  -0.2 0.87 CBCL score – Internalizing Unadjusted  Reference 10.3 (5.0–15.5) 3.9 <0.001 2.7 (-4.4–9.8) 0.8 0.45 Adjusted* Reference 9.1 (3.4–14.9) 3.1 0.002 2.2 (-5.6–9.9) 0.6 0.58 CBCL score – Externalizing Unadjusted  Reference 8.1 (3.1–13.2) 3.2 0.001 -0.3 (-7.1–6.5) -0.09 0.93 Adjusted* Reference 9.7 (4.2–15.2) 3.5 <0.001 0.3 (-7.0–7.6) 0.07 0.94 Age 6         HBQ – Internalizing Unadjusted  Reference 0.1 (-0.02–0.2) 1.6 0.11 -0.002 (-0.2–0.2) -0.02 0.98 Adjusted** Reference 0.01 (-0.1–0.2)  0.2 0.81 -0.07 (-0.3–0.1) -0.9 0.39 HBQ – Externalizing Unadjusted  Reference 0.04 (-0.1–0.2) 0.7 0.48 -0.02 (-0.2–0.1) -0.2 0.79 Adjusted** Reference -0.003 (-0.1–0.1)  -0.04 0.96 -0.04 (-0.2–0.1) -0.5 0.64 HBQ – ADHD Unadjusted  Reference -0.02 (-0.2–0.2) -0.2 0.86 -0.3 (-0.5–-0.03) -2.3 0.03 Adjusted** Reference -0.1 (-0.3–0.1)  -1.1 0.29 -0.4 (-0.7–-0.1)  -2.8 0.006 Executive function – Global BRIEF scores Unadjusted  Reference 6.3 (1.9–10.7) 2.8 0.005 -5.0 (-10.9–0.9) -1.7 0.09 Adjusted** Reference 5.2 (-0.04–10.3) 2.0 0.05 -6.5 (-13.2–0.2) -1.9 0.06 Executive function – HF % correct on block 3 Unadjusted  Reference -0.06 (-0.2–0.1) -1.0 0.31 -0.03 (-0.2–0.1) -0.5 0.64 Adjusted** Reference -0.1 (-0.2–-0.009)  -2.1 0.03 -0.06 (-0.2–0.09) -0.9 0.39 *Adjusted for child sex, age, gestational age at birth, birth weight, prenatal SSRI exposure, maternal depression prior to pregnancy,  maternal minority status, and maternal education. **Adjusted for all of the above, and maternal depression at 6-year follow-up. BRIEF,  Behavior Rating Inventory of Executive Function; CBCL, Child Behavior Checklist; HAMD, Hamilton Depression Scale; HBQ, Health and Behavior questionnaire; HF, Hearts and Flowers task 82  5.4.5 Maternal depressive symptom trajectories, child behavior, and executive functions at 6 years At six years of age, maternal depressive symptom trajectories were unrelated to children’s internalizing and externalizing symptomatology. However, children whose mothers became less depressed over time had lower reported levels of ADHD symptoms than those whose mothers had consistently few depressive symptoms (β=-0.4, 95% CI; -0.7 – -0.1). ADHD symptomatology did not differ between children of mothers in the increasing and low trajectory groups (Table 5.4).  Turning to executive functions, children of mothers in the low and decreasing trajectory groups showed similar levels of executive functions as assessed by the BRIEF and HF task (Table 5.4). However, children of mothers in the increasing trajectory group had poorer executive functions at age six than those of mothers in the low group, as assessed by both the BRIEF (β=5.2, 95% CI; -0.04–10.3) and accuracy on the HF task (β=-0.1, 95% CI; -0.2–-0.009; Table 5.4).  Results from sensitivity analyses were consistent with those from fully adjusted models (data not shown).  83  5.5 Discussion Three distinct trajectories of maternal depressive symptoms from the second trimester of pregnancy through the first three years of a child’s life were identified that corresponded to low, increasing, and decreasing patterns over time. Children whose mothers’ moods improved during this period (the decreasing trajectory group) showed similar internalizing and externalizing scores at three and six years of age, comparable executive functions, and even less evidence of ADHD-related symptoms when they were six years old than children whose mothers had low depressive symptoms over time, despite mothers in the decreasing group remaining more symptomatic throughout than those who consistently showed few depressive symptoms. In contrast, children of mothers who had a trajectory of becoming increasingly depressive (i.e., the increasing group) showed more internalizing and externalizing behavior problems at age three years and poorer executive functions at age six years than children of mothers who had few depressive symptoms. Findings on executive functions were consistent across maternal report and an objective, laboratory-based measure and hence are unlikely to be due to maternal reporting bias. Moreover, analyses controlled for concurrent maternal depressive symptoms at age six years, suggesting a persistent effect of exposure to patterns of maternal depressive symptoms during the first three years of life on executive functions later in childhood. It is notable that at no time during the child’s first three years were the mean depression scores for mothers in the increasing or decreasing trajectory groups ever as low as for mothers in the low group. In fact, when I looked beyond the period of time covered by the trajectories, at six years postpartum, women in the decreasing group had even more depressive symptoms than women in the increasing group. This pattern of recurring symptoms concurs with the observation of a high relapse rate of depression among mothers 8. Not surprisingly, women in the low group 84  had the lowest rate of SSRI use. It is possible that the higher SSRI use observed among women in the increasing and decreasing trajectory groups reflects drug failure and raises the need to consider additional treatment modalities for women at-risk of depression.  Fewer trajectories were identified than in most previous studies; this is likely due to differences in analytic methods 157, sample sizes, and cohort characteristics. Nonetheless, there are areas of convergence between this study and others. Another study (n = 3,051) that modelled maternal depression trajectories from pregnancy until 27 years later 160 identified 3 groups of women: those with few or no symptoms or high-escalating depressive symptoms over time may correspond to our low and increasing groups. The principal difference was that their third group were women with occasional depressive symptoms and the third group in the current study was women with initially high then decreasing symptoms (though this difference may be reflective of the shorter duration of follow-up in our study). Moreover, the high-chronic trajectories reported in other research 136,140-143 may correspond to the increasing trajectory that was identified.  A strict “fetal programming” hypothesis 161 would have predicted that children of mothers in the decreasing trajectory group would look worse on outcome measures at 3 and 6 years than children of mothers with the least depressive symptoms. Instead, at both ages, children of mothers with decreasing depressive symptoms showed outcomes as good or better than those of mothers who consistently had few depressive symptoms over time. These findings stand in contrast to other research that has found that prenatal maternal depression independently predicts poorer cognitive development and mental health in offspring, even after adjustment for postnatal depression 135,162. These findings also differ from those of other trajectory-based studies that found children of mothers who had elevated depressive, anxious, and stress symptoms only during pregnancy had worse mental health in adolescence and young adulthood 142,143. These 85  differences in results may be due to methodological differences, specifically in the number of mental health symptoms included when computing the trajectories and the analytic approach taken to statistical modelling.  Indeed, the current study findings are in line with recent research that found significant effects of exposure to maternal depression in early childhood, during the first three years of life and not perinatally, on child behavioral problems after controlling for confounding by familial environment 163. Moreover, other studies that solely included depressive symptoms when modelling trajectories have also shown findings consistent with those of the current study regarding the influence of the timing of maternal mood: a maternal depression trajectory with high depressive symptoms during pregnancy alone was not associated with children’s behavior at age five years, 141 and an early postnatal trajectory characterized by high levels of maternal depression that subsided by 24 months did not predict poorer behavioral or mental health outcomes when children were 15 years old 139. The effect of postnatal, and not prenatal, maternal depressive symptoms on child outcomes is noteworthy. It is possible that depressed mothers are able to meet children’s needs in early life, but that deficits begin to emerge as children age and require more engaged parenting 163. Additional studies that model maternal depression trajectories longitudinally, beyond the time covered in our trajectories, and continue to track cognitive, behavioral, and mental health outcomes into childhood and beyond are warranted to clarify how the course of pre- and post-natal depression affects child development.  Overall, an adverse and persistent effect of an increasing trajectory of maternal depressive symptoms on children’s behavior is consistent with previous findings 136,139,141. This study’s results extend these findings by suggesting that early behavioral disturbances seen in relation to chronic, increasing maternal depression may subsequently manifest in later childhood 86  as disruptions in executive functions. One explanation for these results implicates the role of parental behaviors. Prior research has shown that postpartum depression results in poorer parenting behaviors specifically within domains that encourage the development of children’s emerging cognitive skills such as play, talking, singing, and reading 164,165. In turn, children exposed to poorer parenting during infancy and toddlerhood have worse executive functions starting as young as 18 months 166,167. Another possible pathway by which maternal depression may lead to worse executive functions in children is via the development of less secure attachment styles and poorer socio-emotional regulation 168,169.  As executive functions are especially vulnerable to adverse environmental impacts, such as stress or lack of social connection 170, it is plausible that increasing maternal depressive symptoms would selectively and disproportionately affect executive functions. As impairments in executive functions have also been widely observed in mental and physical illness 171, understanding the risk factors associated with poorer executive function and intervening to improve executive functions in preschoolers or earlier holds promise for reducing inequalities in health, wealth, and quality of life, and for positioning children to have better lifelong outcomes 172,173. A number of limitations should be mentioned in conjunction with study results. As the cohort was over-sampled from women at high risk for mental illness, study results may not be generalizable to other populations. While statistically significant differences in outcome measures were detected, the study’s sample size was limited and findings require replication in independent studies. The loss to follow-up in the study cohort is also a limitation as it may have introduced a form of selection bias, with more depressed mothers dropping out of the study. Though there was minimal evidence of selection bias due to illness severity, it is possible (and 87  likely) that there are sources of selection bias that we did not capture. A related limitation is that an assumption was made that data were missing at random in the computation of trajectories; yet, this may not be the case as more depressed women may be more likely to miss a study assessment. Maternal reporting bias is also a concern 174; however, it cannot explain this study’s results as mothers in the increasing and decreasing groups had similarly high levels of depressive symptoms when their children were six years old, yet the latter group reported better executive functions in their children than the former. Additional limitations are the use of different measures of depressive symptoms at different time points and that the influence of other parental figures on children was unable to be assessed. Finally, there is a possibility that study findings may reflect heritability, other sources of unmeasured confounding, or reverse causality. To conclude, this study reports that long after pregnancy, there are women who continue to have high and increasing depressive symptoms, and that these mood patterns have an impact on their children. The impact appears to be highest on internalizing and externalizing behaviors at age three years, which then emerge as poorer executive functions at age six years. This study also identified a group of women whose depressive symptoms decreased over pregnancy and the following three years. Even though these mothers’ moods remained consistently worse than those of low-depression trajectory mothers, associations between their children’s behavioral outcomes at three years, and mental health indicators and executive functions at six years, in relation to exposure to maternal depressive symptom trajectories, were comparable between both groups. These findings provide insight into the developmental sequelae of exposure to different trajectories of maternal depressive symptoms from pregnancy into early childhood. Importantly, this study suggests that if mothers with initially high levels of depressive symptoms experience even partial remission of symptoms during their child’s first three years of life, their children’s 88  emotional, cognitive, and behavioral development remain unaffected. Findings provide impetus to monitor and treat maternal depression, even years after birth, for the benefit of both mother and child.   89  Chapter 6: Discussion 6.1 Summary of findings The purpose of this dissertation was to investigate key factors that could influence the impact of maternal depression and prenatal antidepressant exposure on children’s developmental health outcomes. While maternal depression and prenatal antidepressant exposure have been associated with adverse developmental outcomes, understanding why there are variations in outcomes requires an integrative perspective that considers the impact of both time-sensitive and ongoing internal and external contextual factors (Chapter 2).   The study presented in Chapter 3 demonstrated an attenuation in the effect of prenatal antidepressant exposure on vulnerable child developmental outcomes at kindergarten age when differences in maternal mental health and related socio-demographic factors were controlled for. In unadjusted analyses, almost all domains of development (physical, emotional, language, cognitive, and communication) appeared to be adversely affected. After adjustment however, only a modest effect on children’s anxious behaviours and physical independence was identified. These results highlight the role of factors associated with prenatal antidepressant exposure, namely maternal mental health, that contribute to developmental risk and suggest that the effect on development attributable to antidepressant exposure may be modest and highly selective for certain outcomes. Moreover, this study also suggested that differences that were unable to be wholly teased apart due to illness severity and relatedly, due to associations with genetic/familial heritability, may contribute to observed effects. Overall, effects related to maternal mental illness may largely underlie apparent associations between prenatal antidepressant exposure and children’s developmental risk at kindergarten age. 90   Chapter 4 explored potential biological pathways that may mediate and moderate the long-term transmission of risk to children of mothers who were depressed during pregnancy. The findings from this chapter suggest that while small differences in epigenetic marks in late adolescence may be associated with prenatal maternal depression exposure, it is likely that these associations relate to genetic differences. These results indicate that the genetic, and thereby heritable, risk associated with depression may underlie potential long-term epigenetic differences associated with prenatal maternal depression exposure.    Chapter 5 considered longitudinal patterns of maternal depression, spanning both prenatal and postnatal periods, and their impact on children’s outcomes at ages three and six years. This study identified distinct populations of women based on their patterns of depressive symptomatology over time. A subpopulation of particular note were women who consistently had moderate to high levels of depressive symptoms prenatally and across the first three postpartum years, as their mood did not improve over time and their children were at risk of worse behavioral outcomes at age three years and poorer performance in executive functions at age six years. In contrast, children of women whose depressive symptoms improved over the same time frame had comparable outcomes as those whose mothers had consistently low depressive symptomatology. These findings indicate that improvements in maternal mood postnatally can translate into benefits for children’s outcomes and support the need for identification and effective treatment of maternal depression during pregnancy and beyond.  Guided by the Transmission of Risk Framework50, this dissertation investigated how innate individual and external contextual factors can shape children’s developmental outcomes following maternal depression and prenatal antidepressant exposure. The research in this dissertation presented an integrative perspective on this topic as well as three separate empirical 91  analyses of data from population-level and clinical birth cohort studies. Taken together, results from Chapters 3, 4, and 5 suggest a role for genetic heritable risk, maternal depression severity, and patterns of maternal depressive symptoms over time on the transmission of risk of children whose mothers experience depression during pregnancy. Overall, findings from this dissertation add to the growing body of literature highlighting the critical importance of perinatal and postnatal maternal mental health for the long-term health and development of mothers and their children.   6.2 Strengths and limitations This dissertation represents a unique contribution to the field of maternal mental health and child development by engaging in an integrative research perspective and investigating epigenetic, neuro-behavioural, and population-level normative or typical variations in early development. This dissertation employed an interdisciplinary research approach to seek a better understanding of how intrinsic and extrinsic contextual factors can contribute to variations in developmental risk following exposure to maternal depression and antidepressants during pregnancy. This approach facilitated an integration of the different types of studies (population-level epidemiology, behavioural birth cohort, and biological epigenetic/genetic) presented in this dissertation, while rooting epidemiological findings within a framework centered in biological plausibility (Figure 2.1).  There are a number of methodological strengths presented throughout this dissertation. An important strength is the longitudinal study designs that were employed across all studies which facilitated temporal analyses and ascertainment of putative cause-and-effect relationships. In addition to employing robust and stringent methods to control for confounding, internal 92  validity was also enhanced by the outcome measures used to assess normative child behavioral outcomes at different ages. Recall bias was removed as all data were collected prospectively.  As with all research, there are a number of limitations that should be noted. Beyond the limitations inherent to each individual study, which were previously discussed in each chapter, there were methodological considerations that are common across all studies. Selection bias is one such concern; the study populations in the prospectively collected cohort studies were not randomly sampled, and though Chapter 3 was a population-level study, the inclusion and exclusion criteria may have introduced selection bias. The generalizability of research findings may thus be limited. Although stringent attempts were made to control for a broad range of confounders, residual confounding may have biased study results. Also, given that all studies were observational and not randomized, causation cannot be determined from study results. Finally, though there was a focus on select contextual factors, there are many other important socio-environmental considerations that are critical in influencing child development but were not examined45,175.   6.3 Significance and implications of thesis research The prevalence and incidence of depression during pregnancy8 highlight the substantial burden of chronic mental illness among pregnant women while also raising concerns about the implications for health across two generations. Depression and other mood disorders are among the most common complications of childbirth. Yet, the peripartum is also an ideal period for screening and intervention efforts given the high levels of interaction with the healthcare system that occur starting from pregnancy and into the postnatal period.  Findings from this research contribute to a better understanding of factors influencing child development, and highlight the 93  critical importance of considering the maternal-child dyad in efforts aimed at improving child development. While both modifiable and unmodifiable factors were investigated, combined, findings from this dissertation triangulate upon maternal mental health, both pre- and postnatally, as a key factor influencing children’s developmental risk across the life course. They support targeting interventions specifically at sub-populations of women (i.e., those likely to have a more severe and persistent course of disease) as a way to improve maternal and child health as their children may also be at higher risk of being adversely impacted. With respect to treatment, results from this dissertation suggest that decisions regarding antidepressant use during pregnancy should focus on identifying efficacious ways to successfully manage perinatal maternal depression, while considering potential risks and benefits to both the mother and developing child.  Adverse and prolonged environmental experiences, such as those due to maternal depression, can have profound detrimental effects on a child’s health, well-being, and functioning throughout life. An optimal approach to reducing developmental risk would involve early intervention to prevent the consequences of adversity on the developing brain. Results from this dissertation suggest that improving maternal depression early in a child’s life could influence later developmental outcomes by altering the course of untreated maternal depression in terms of both illness severity and persistence. Intervention starting prenatally via the mothers may yield even greater societal returns on investment than those that commence later 176. Evidence supports the effectiveness of depression screening programs in routine perinatal care (during the second trimester to eight weeks postpartum) in improving depression outcomes, especially when paired with treatment interventions, with a potential reduction in depression prevalence of 2 – 9% at follow-up (up to 16 months postpartum) 177. The accrued benefits for both mothers and their 94  children14 could lead to improvements to the entire family unit, and positively shape outcomes across generations.   The results of this dissertation are directly useful to health care professionals and families affected by maternal depression. Findings contribute to the literature on risks and benefits of antidepressant use during pregnancy and can inform individual decision-making on treatment options. They also help identify which outcomes in children across life may be particularly vulnerable to the effects of maternal depression and prenatal antidepressant exposure. Further, study findings highlight a subpopulation of women who experience continued high levels of depressive symptoms after birth. They and their children may represent clinical priorities for close monitoring and intervention both pre- and postnatally. It is conceivable that some mothers and their children could benefit from prenatal treatment with an antidepressant and a key challenge ahead is to identify who benefits and why.   6.4 Future directions  Future research should aim at understanding additional modifiable contextual factors that contribute to attenuating developmental risk following maternal depression and prenatal antidepressant exposure. A better understanding of such factors that are amenable to change could help guide the development and implementation of future treatments and policy.  It is possible to conceptualize of more personalized treatment plans and specialized outreach programs that better identify the individuals who stand to benefit, and the conditions that can enhance improvements in populations that are at-risk 178. Research questions should also aim to identify specific aspects of treatment, such as focus on the mother-infant relationship179, and the 95  timing of interventions180, that can be specifically designed and targeted to minimize the adverse impact of mood disruptions on mothers, children, and the family unit.   A critical consideration is the mental health of the entire family unit which includes fathers as well as other parental figures; however, pursuit of this question represents a limitation of this dissertation. In addition to interventions specifically directed at the child, there is a need for family-centered supports that focus on parenting and other difficulties due to depression in the family. Both the mother’s and father’s mental health needs should be considered in order to ensure the optimal developmental health of children. Studies show that providing family-based psychoeducational prevention strategies in families where parents have depression improved parental functioning and decreased risk for internalizing disorders in children during the transition to adolescence 181, suggesting implementation of family-centered programs could confer benefits to the entire family unit.   To enhance the movement of interventions into healthcare policy, there is a need to study factors related to implementation of screening and treatment programs as well as the efficacy and cost-effectiveness of different intervention programs for different populations 182,183. Research should particularly focus on the impact of programs on disadvantaged and vulnerable populations, as they are known to be at higher risk of psychopathology, as well as on strategies to tailor programs for populations with high need 184.   6.5 Conclusion The research in this dissertation examined the developmental effects of maternal depression and prenatal antidepressant exposure, while highlighting the important role that related contextual factors play in influencing child outcomes. Guiding this work was the 96  Transmission of Risk Framework, which provided a model in which to consider both internal and external factors to understand the impact of maternal depression and prenatal antidepressant exposure on child development. The empirical studies in this dissertation investigated innate biological (genetic and epigenetic variation) and external (differences in maternal depressive symptoms associated with prenatal antidepressant use and differences in maternal depression symptomatology over time) factors related to these exposures that can influence developmental outcomes from early childhood to adolescence. Results suggested that differences related to genetic risk, depressive symptom severity during pregnancy, and patterns of depressive symptomatology over time are critical in determining the impact of maternal depression and prenatal antidepressant exposure on development.  To sum, though pregnancy is a period of heightened risk for depression and other mood disorders47, it is also a period in which women are in frequent contact with the healthcare system. As such, it represents an optimal time to commence preventative and treatment efforts for maternal depression, with potential sustained benefit for mothers, children, and their families. Taken together, findings from this dissertation highlight the need to recognize and address maternal depression during and following pregnancy. They also offer critical insight into how incorporating an understanding of factors related to prenatal maternal depression and antidepressant use can minimize associated harms and promote healthy maternal and child outcomes across life.  97  References 1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. Fourth edition. Text revision. Washington, DC.2000. 2. Brenes GA. Anxiety, Depression, and Quality of Life in Primary Care Patients. Prim. Care Companion J. Clin. Psychiatry. 2007;9(6):437-443. 3. Hirschfeld RM. The Comorbidity of Major Depression and Anxiety Disorders: Recognition and Management in Primary Care. Prim. Care Companion J. Clin. Psychiatry. Dec 2001;3(6):244-254. 4. Mezuk B, Kendler KS. Examining variation in depressive symptoms over the life course: a latent class analysis. Psychol. Med. 02/24 2012;42(10):2037-2046. 5. Sutin AR, Terracciano A, Milaneschi Y, An Y, Ferrucci L, Zonderman AB. The trajectory of depressive symptoms across the adult life span. JAMA psychiatry. Aug 2013;70(8):803-811. 6. Gavin NI, Gaynes BN, Lohr KN, Meltzer-Brody S, Gartlehner G, Swinson T. Perinatal depression: a systematic review of prevalence and incidence. Obstet. Gynecol. Nov 2005;106(5 Pt 1):1071-1083. 7. Vesga-Lopez O, Blanco C, Keyes K, Olfson M, Grant BF, Hasin DS. Psychiatric disorders in pregnant and postpartum women in the United States. Archives of general psychiatry. Jul 2008;65(7):805-815. 8. Underwood L, Waldie K, D'Souza S, Peterson ER, Morton S. A review of longitudinal studies on antenatal and postnatal depression. Archives of women's mental health. Apr 16 2016. 9. Patten SB, Williams JVA, Lavorato DH, Wang JL, McDonald K, Bulloch AGM. Descriptive epidemiology of major depressive disorder in Canada in 2012. Can. J. Psychiatry. 2015;60(1):23-30. 10. Dagher RK, Shenassa ED. Prenatal health behaviors and postpartum depression: is there an association? Archives of women's mental health. Feb 2012;15(1):31-37. 11. Malm H, Sourander A, Gissler M, et al. Pregnancy Complications Following Prenatal Exposure to SSRIs or Maternal Psychiatric Disorders: Results From Population-Based National Register Data. The American journal of psychiatry. Dec 2015;172(12):1224-1232. 12. Beck CT, Records K, Rice M. Further development of the Postpartum Depression Predictors Inventory-Revised. J. Obstet. Gynecol. Neonatal Nurs. Nov-Dec 2006;35(6):735-745. 13. Dennis CL, Chung-Lee L. Postpartum depression help-seeking barriers and maternal treatment preferences: a qualitative systematic review. Birth. Dec 2006;33(4):323-331. 14. Shaw DS, Connell A, Dishion TJ, Wilson MN, Gardner F. Improvements in maternal depression as a mediator of intervention effects on early childhood problem behavior. Dev. Psychopathol. Spring 2009;21(2):417-439. 15. Reuben JD, Shaw DS, Brennan LM, Dishion TJ, Wilson MN. A family-based intervention for improving children's emotional problems through effects on maternal depressive symptoms. Journal of consulting and clinical psychology. Dec 2015;83(6):1142-1148. 98  16. Furlong M, McGilloway S, Bywater T, Hutchings J, Smith SM, Donnelly M. Cochrane review: behavioural and cognitive-behavioural group-based parenting programmes for early-onset conduct problems in children aged 3 to 12 years (Review). Evidence-based child health : a Cochrane review journal. Mar 7 2013;8(2):318-692. 17. Sockol LE. A systematic review of the efficacy of cognitive behavioral therapy for treating and preventing perinatal depression. J. Affect. Disord. May 15 2015;177:7-21. 18. Hofmann SG, Curtiss J, Carpenter JK, Kind S. Effect of treatments for depression on quality of life: a meta-analysis. Cogn. Behav. Ther. Jun 2017;46(4):265-286. 19. Cuijpers P, van Straten A, Andersson G, van Oppen P. Psychotherapy for depression in adults: a meta-analysis of comparative outcome studies. Journal of consulting and clinical psychology. Dec 2008;76(6):909-922. 20. Hodgkinson S, Godoy L, Beers LS, Lewin A. Improving Mental Health Access for Low-Income Children and Families in the Primary Care Setting. Pediatrics. 10/10/accepted 2017;139(1):e20151175. 21. Cuijpers P, de Beurs DP, van Spijker BA, Berking M, Andersson G, Kerkhof AJ. The effects of psychotherapy for adult depression on suicidality and hopelessness: a systematic review and meta-analysis. J. Affect. Disord. Jan 25 2013;144(3):183-190. 22. Cooper WO, Willy ME, Pont SJ, Ray WA. Increasing use of antidepressants in pregnancy. Am. J. Obstet. Gynecol. Jun 2007;196(6):544 e541-545. 23. Hanley GE, Mintzes B. Patterns of psychotropic medicine use in pregnancy in the United States from 2006 to 2011 among women with private insurance. BMC Pregnancy Childbirth. Jul 22 2014;14:242. 24. Warburton W, Hertzman C, Oberlander TF. A register study of the impact of stopping third trimester selective serotonin reuptake inhibitor exposure on neonatal health. Acta Psychiatr. Scand. Jun 2010;121(6):471-479. 25. Smolina K, Hanley GE, Mintzes B, Oberlander TF, Morgan S. Trends and Determinants of Prescription Drug Use during Pregnancy and Postpartum in British Columbia, 2002-2011: A Population-Based Cohort Study. PloS one. 2015;10(5):e0128312. 26. Roca A, Imaz ML, Torres A, et al. Unplanned pregnancy and discontinuation of SSRIs in pregnant women with previously treated affective disorder. J. Affect. Disord. Sep 25 2013;150(3):807-813. 27. Cohen LS, Altshuler LL, Harlow BL, et al. Relapse of major depression during pregnancy in women who maintain or discontinue antidepressant treatment. JAMA : the journal of the American Medical Association. Feb 1 2006;295(5):499-507. 28. Gentile S. Untreated depression during pregnancy: Short- and long-term effects in offspring. A systematic review. Neuroscience. Sep 4 2015. 29. Jarde A, Morais M, Kingston D, et al. Neonatal Outcomes in Women With Untreated Antenatal Depression Compared With Women Without Depression: A Systematic Review and Meta-analysis. JAMA psychiatry. Jun 8 2016. 30. Piallini G, Brunoro S, Fenocchio C, et al. How Do Maternal Subclinical Symptoms Influence Infant Motor Development during the First Year of Life? Front. Psychol. 11/01 2016;7:1685. 31. Field T. Prenatal depression effects on early development: a review. Infant Behav. Dev. Feb 2011;34(1):1-14. 99  32. Behrendt HF, Konrad K, Goecke TW, Fakhrabadi R, Herpertz-Dahlmann B, Firk C. Postnatal Mother-to-Infant Attachment in Subclinically Depressed Mothers: Dyads at Risk? Psychopathology. 2016;49(4):269-276. 33. Martins C, Gaffan EA. Effects of early maternal depression on patterns of infant-mother attachment: a meta-analytic investigation. Journal of child psychology and psychiatry, and allied disciplines. Sep 2000;41(6):737-746. 34. Warnock FF, Craig KD, Bakeman R, Castral T, Mirlashari J. The relationship of prenatal maternal depression or anxiety to maternal caregiving behavior and infant behavior self-regulation during infant heel lance: an ethological time-based study of behavior. BMC Pregnancy Childbirth. 09/07 2016;16:264. 35. Sohr-Preston SL, Scaramella LV. Implications of timing of maternal depressive symptoms for early cognitive and language development. Clin. Child Fam. Psychol. Rev. Mar 2006;9(1):65-83. 36. Suri R, Lin AS, Cohen LS, Altshuler LL. Acute and long-term behavioral outcome of infants and children exposed in utero to either maternal depression or antidepressants: a review of the literature. J. Clin. Psychiatry. Oct 2014;75(10):e1142-1152. 37. Beck CT. Maternal depression and child behaviour problems: a meta-analysis. J. Adv. Nurs. Mar 1999;29(3):623-629. 38. Dahlen HM. The impact of maternal depression on child academic and socioemotional outcomes. Economics of Education Review. 2016/06/01/ 2016;52:77-90. 39. Capron LE, Glover V, Pearson RM, et al. Associations of maternal and paternal antenatal mood with offspring anxiety disorder at age 18 years. J. Affect. Disord. Nov 15 2015;187:20-26. 40. Pearson RM, Evans J, Kounali D, et al. Maternal depression during pregnancy and the postnatal period: risks and possible mechanisms for offspring depression at age 18 years. JAMA psychiatry. Dec 2013;70(12):1312-1319. 41. Rotem-Kohavi N, Oberlander TF. Variations in Neurodevelopmental Outcomes in Children with Prenatal SSRI Antidepressant Exposure. Birth defects research. Jul 17 2017;109(12):909-923. 42. Liu X, Agerbo E, Ingstrup KG, et al. Antidepressant use during pregnancy and psychiatric disorders in offspring: Danish nationwide register based cohort study. BMJ (Clinical research ed.). Sep 06 2017;358:j3668. 43. Brown HK, Hussain-Shamsy N, Lunsky Y, Dennis CE, Vigod SN. The Association Between Antenatal Exposure to Selective Serotonin Reuptake Inhibitors and Autism: A Systematic Review and Meta-Analysis. J. Clin. Psychiatry. Jan 2017;78(1):e48-e58. 44. Oberlander TF, Vigod SN. Developmental Effects of Prenatal Selective Serotonin Reuptake Inhibitor Exposure in Perspective: Are We Comparing Apples to Apples? Journal of the American Academy of Child and Adolescent Psychiatry. May 2016;55(5):351-352. 45. Goodman SH, Rouse MH, Connell AM, Broth MR, Hall CM, Heyward D. Maternal depression and child psychopathology: a meta-analytic review. Clin. Child Fam. Psychol. Rev. Mar 2011;14(1):1-27. 46. Cicchetti D, Rogosch FA, Toth SL. Maternal depressive disorder and contextual risk: contributions to the development of attachment insecurity and behavior problems in toddlerhood. Development and psychopathology. Spring 1998;10(2):283-300. 100  47. Meltzer-Brody S. New insights into perinatal depression: pathogenesis and treatment during pregnancy and postpartum. Dialogues Clin. Neurosci. 2011;13(1):89-100. 48. Diego MA, Field T, Hernandez-Reif M. Prepartum, postpartum and chronic depression effects on neonatal behavior. Infant Behavior and Development. 2005/06/01/ 2005;28(2):155-164. 49. Laurent HK, Leve LD, Neiderhiser JM, et al. Effects of Prenatal and Postnatal Parent Depressive Symptoms on Adopted Child HPA Regulation: Independent and Moderated Influences. Dev. Psychol. 06/11 2013;49(5):10.1037/a0028800. 50. Goodman SH, Gotlib IH. Risk for psychopathology in the children of depressed mothers: a developmental model for understanding mechanisms of transmission. Psychol. Rev. Jul 1999;106(3):458-490. 51. Perinatal Services BC (creator). British Columbia Perinatal Data Registry. Data Extract. Data Extract. Vancouver, BC, Canada: Population Data BC; 2015. 52. Ministry of Health (creator). Medical Services Plan (MSP) Payment Information File. Data Extract. Vancouver, BC, Canada: Population Data BC; 2015. 53. Canadian Institute for Health Information (creator). Discharge Abstract Database (Hospital Separations). Data Extract. Vancouver, BC, Canada: Population Data BC [publisher]; 2015. 54. Ministry of Health (creator). Consolidation File (MSP Registration & Premium Billing). Data Extract. Vancouver, BC, Canada: Population Data BC; 2015. 55. Ministry of Health (creator). PharmaNet. Data Extract. Vancouver, BC, Canada: Data Stewardship Committee; 2015. 56. Human Early Learning Partnership (creator). Early Development Instrument. Data Extract. Vancouver, BC, Canada: Population Data BC; 2015. 57. Park M, Oberlander TF. In-utero Selective Serotonin Reuptake Inhibitor Antidepressant Exposure: Fetal Programing and Developmental Interactions With Context. Clin. Pharmacol. Ther. 2018;104(4):616-618. 58. Ansorge MS, Zhou M, Lira A, Hen R, Gingrich JA. Early-life blockade of the 5-HT transporter alters emotional behavior in adult mice. Science. Oct 29 2004;306(5697):879-881. 59. Pluess M, Belsky J. Prenatal programming of postnatal plasticity? Development and psychopathology. Feb 2011;23(1):29-38. 60. Weikum WM, Brain U, Chau CM, et al. Prenatal serotonin reuptake inhibitor (SRI) antidepressant exposure and serotonin transporter promoter genotype (SLC6A4) influence executive functions at 6 years of age. Front. Cell. Neurosci. 2013;7:180. 61. Brown HK, Ray JG, Wilton AS, Lunsky Y, Gomes T, Vigod SN. Association Between Serotonergic Antidepressant Use During Pregnancy and Autism Spectrum Disorder in Children. JAMA : the journal of the American Medical Association. Apr 18 2017;317(15):1544-1552. 62. Park M, Brain U, Grunau RE, Diamond A, Oberlander TF. Maternal depression trajectories from pregnancy to 3 years postpartum are associated with children's behavior and executive functions at 3 and 6 years. Archives of women's mental health. Jan 16 2018. 63. Bonari L, Pinto N, Ahn E, Einarson A, Steiner M, Koren G. Perinatal risks of untreated depression during pregnancy. Can. J. Psychiatry. Nov 2004;49(11):726-735. 101  64. Boukhris T, Sheehy O, Mottron L, Berard A. Antidepressant Use During Pregnancy and the Risk of Autism Spectrum Disorder in Children. JAMA pediatrics. Feb 2016;170(2):117-124. 65. Man KKC, Chan EW, Ip P, et al. Prenatal antidepressant use and risk of attention-deficit/hyperactivity disorder in offspring: population based cohort study. BMJ (Clinical research ed.). May 31 2017;357:j2350. 66. Handal M, Skurtveit S, Furu K, et al. Motor development in children prenatally exposed to selective serotonin reuptake inhibitors: a large population-based pregnancy cohort study. BJOG. Nov 2016;123(12):1908-1917. 67. Skurtveit S, Selmer R, Roth C, Hernandez-Diaz S, Handal M. Prenatal exposure to antidepressants and language competence at age three: results from a large population-based pregnancy cohort in Norway. BJOG. Dec 2014;121(13):1621-1631. 68. Hurault-Delarue C, Damase-Michel C, Finotto L, et al. Psychomotor developmental effects of prenatal exposure to psychotropic drugs: a study in EFEMERIS database. Fundam. Clin. Pharmacol. Oct 2016;30(5):476-482. 69. Lupattelli A, Wood M, Ystrom E, Skurtveit S, Handal M, Nordeng H. Effect of Time-Dependent Selective Serotonin Reuptake Inhibitor Antidepressants During Pregnancy on Behavioral, Emotional, and Social Development in Preschool-Aged Children. J. Am. Acad. Child Adolesc. Psychiatry. 2018/03/01/ 2018;57(3):200-208. 70. Grzeskowiak LE, Morrison JL, Henriksen TB, et al. Prenatal antidepressant exposure and child behavioural outcomes at 7 years of age: a study within the Danish National Birth Cohort. BJOG. Nov 2016;123(12):1919-1928. 71. Bennett IM, Marcus SC, Palmer SC, Coyne JC. Pregnancy-related discontinuation of antidepressants and depression care visits among Medicaid recipients. Psychiatr. Serv. Apr 2010;61(4):386-391. 72. Cheng TL, Goodman E. Race, ethnicity, and socioeconomic status in research on child health. Pediatrics. Jan 2015;135(1):e225-237. 73. Sujan AC, Rickert ME, Oberg AS, et al. Associations of Maternal Antidepressant Use During the First Trimester of Pregnancy With Preterm Birth, Small for Gestational Age, Autism Spectrum Disorder, and Attention-Deficit/Hyperactivity Disorder in Offspring. JAMA : the journal of the American Medical Association. Apr 18 2017;317(15):1553-1562. 74. Hviid A, Melbye M, Pasternak B. Use of selective serotonin reuptake inhibitors during pregnancy and risk of autism. The New England journal of medicine. Dec 19 2013;369(25):2406-2415. 75. Viktorin A, Uher R, Reichenberg A, Levine SZ, Sandin S. Autism risk following antidepressant medication during pregnancy. Psychological medicine. May 22 2017:1-10. 76. Guhn M, Janus M, Enns J, et al. Examining the social determinants of children's developmental health: protocol for building a pan-Canadian population-based monitoring system for early childhood development. BMJ open. Apr 29 2016;6(4):e012020. 77. Janus M, Offord DR. Development and psychometric properties of the Early Development Instrument (EDI): A measure of children's school readiness. Canadian Journal of Behavioural Science / Revue canadienne des sciences du comportement. 2007;39(1):1-22. 102  78. Brinkman S, Gregory T, Harris J, Hart B, Blackmore S, Janus M. Associations Between the Early Development Instrument at Age 5, and Reading and Numeracy Skills at Ages 8, 10 and 12: a Prospective Linked Data Study. CHILD INDICATORS RESEARCH. 2013;6(4):695-708. 79. Forer B, Zumbo B. Validation of Multilevel Constructs: Validation Methods and Empirical Findings for the EDI. Vol 1032011. 80. Janus M, Walsh, C., & Duku, E.  . Early Development Instrument: Factor structure, sub-domains and Multiple Challenge Index. Hamilton, ON: Offord Centre for Child Studies, McMaster University;2005. 81. Schneeweiss S, Rassen JA, Glynn RJ, Avorn J, Mogun H, Brookhart MA. High-dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology. Jul 2009;20(4):512-522. 82. Statistics Canada (creator). Statistics Canada Income Band Data. Catalogue Number: 13C0016. Data Extract. Vancouver, BC, Canada: Population Data BC; 2015. 83. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm. Stat. Mar-Apr 2011;10(2):150-161. 84. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat. Med. Nov 10 2009;28(25):3083-3107. 85. SAS System for Windows 9.4 [computer program]. Cary, NC. 2014. 86. Brandlistuen RE, Ystrom E, Eberhard-Gran M, Nulman I, Koren G, Nordeng H. Behavioural effects of fetal antidepressant exposure in a Norwegian cohort of discordant siblings. International journal of epidemiology. Aug 2015;44(4):1397-1407. 87. Hanley GE, Brain U, Oberlander TF. Prenatal exposure to serotonin reuptake inhibitor antidepressants and childhood behavior. Pediatric research. Aug 2015;78(2):174-180. 88. Brummelte S, Mc Glanaghy E, Bonnin A, Oberlander TF. Developmental changes in serotonin signaling: Implications for early brain function, behavior and adaptation. Neuroscience. Feb 7 2017;342:212-231. 89. Galbally M, Lewis AJ, Buist A. Child developmental outcomes in preschool children following antidepressant exposure in pregnancy. Aust. N. Z. J. Psychiatry. Jul 2015;49(7):642-650. 90. Hanley GE, Brain U, Oberlander TF. Infant developmental outcomes following prenatal exposure to antidepressants, and maternal depressed mood and positive affect. Early Hum. Dev. Aug 2013;89(8):519-524. 91. Partridge MC, Salisbury AL, LaGasse LL. Fine Motor Differences and Prenatal Serotonin Reuptake Inhibitors Exposure. J. Pediatr. Aug 2016;175:144-149 e141. 92. Viktorin A, Uher R, Kolevzon A, Reichenberg A, Levine SZ, Sandin S. Association of Antidepressant Medication Use During Pregnancy With Intellectual Disability in Offspring. JAMA psychiatry. Jul 12 2017. 93. Frisell T, Oberg S, Kuja-Halkola R, Sjolander A. Sibling comparison designs: bias from non-shared confounders and measurement error. Epidemiology. Sep 2012;23(5):713-720. 94. Talge NM, Neal C, Glover V. Antenatal maternal stress and long-term effects on child neurodevelopment: how and why? Journal of child psychology and psychiatry, and allied disciplines. Mar-Apr 2007;48(3-4):245-261. 103  95. O'Donnell KJ, Meaney MJ. Fetal Origins of Mental Health: The Developmental Origins of Health and Disease Hypothesis. The American journal of psychiatry. Nov 14 2016:appiajp201616020138. 96. Bird A. Perceptions of epigenetics. Nature. May 24 2007;447(7143):396-398. 97. Goldberg AD, Allis CD, Bernstein E. Epigenetics: a landscape takes shape. Cell. Feb 23 2007;128(4):635-638. 98. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nature reviews. Genetics. Jul 2012;13(7):484-492. 99. Deaton AM, Bird A. CpG islands and the regulation of transcription. Genes & development. May 15 2011;25(10):1010-1022. 100. Bird A. DNA methylation patterns and epigenetic memory. Genes & development. Jan 1 2002;16(1):6-21. 101. Jaffe AE, Irizarry RA. Accounting for cellular heterogeneity is critical in epigenome-wide association studies. Genome biology. Feb 04 2014;15(2):R31. 102. Jones MJ, Goodman SJ, Kobor MS. DNA methylation and healthy human aging. Aging cell. Apr 25 2015. 103. Reik W, Dean W, Walter J. Epigenetic reprogramming in mammalian development. Science. Aug 10 2001;293(5532):1089-1093. 104. Fraser HB, Lam LL, Neumann SM, Kobor MS. Population-specificity of human DNA methylation. Genome biology. 2012;13(2):R8. 105. Joubert BR, Haberg SE, Nilsen RM, et al. 450K epigenome-wide scan identifies differential DNA methylation in newborns related to maternal smoking during pregnancy. Environ. Health Perspect. Oct 2012;120(10):1425-1431. 106. Feil R, Fraga MF. Epigenetics and the environment: emerging patterns and implications. Nature reviews. Genetics. Feb 2011;13(2):97-109. 107. Breton CV, Marsit CJ, Faustman E, et al. Small-Magnitude Effect Sizes in Epigenetic End Points are Important in Children's Environmental Health Studies: The Children's Environmental Health and Disease Prevention Research Center's Epigenetics Working Group. Environ Health Perspect. Apr 2017;125(4):511-526. 108. Boyce WT, Kobor MS. Development and the epigenome: the 'synapse' of gene-environment interplay. Dev Sci. Jan 2015;18(1):1-23. 109. Monk C, Spicer J, Champagne FA. Linking prenatal maternal adversity to developmental outcomes in infants: the role of epigenetic pathways. Development and psychopathology. Nov 2012;24(4):1361-1376. 110. Non AL, Binder AM, Kubzansky LD, Michels KB. Genome-wide DNA methylation in neonates exposed to maternal depression, anxiety, or SSRI medication during pregnancy. Epigenetics : official journal of the DNA Methylation Society. Jul 2014;9(7):964-972. 111. Nemoda Z, Massart R, Suderman M, et al. Maternal depression is associated with DNA methylation changes in cord blood T lymphocytes and adult hippocampi. Translational psychiatry. 2015;5:e545. 112. Schroeder JW, Smith AK, Brennan PA, et al. DNA methylation in neonates born to women receiving psychiatric care. Epigenetics : official journal of the DNA Methylation Society. Apr 2012;7(4):409-414. 113. Teh AL, Pan H, Chen L, et al. The effect of genotype and in utero environment on inter-individual variation in neonate DNA methylomes. Genome research. Apr 7 2014. 104  114. Clark R, Hyde JS, Essex MJ, Klein MH. Length of maternity leave and quality of mother-infant interactions. Child development. Apr 1997;68(2):364-383. 115. Radloff LS. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Appl. Psychol. Meas. 1977;1(3):385-401. 116. Proust-Lima C, Philipps V, Liquet B. Estimation of extended mixed models using latent classes and latent processes: the R package lcmm. ArXiv e-prints. 2015;1503.00890(May). 117. Essex MJ, Boyce WT, Goldstein LH, Armstrong JM, Kraemer HC, Kupfer DJ. The confluence of mental, physical, social, and academic difficulties in middle childhood. II: developing the Macarthur health and Behavior Questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry. May 2002;41(5):588-603. 118. Lemery-Chalfant K, Schreiber JE, Schmidt NL, Van Hulle CA, Essex MJ, Goldsmith HH. Assessing internalizing, externalizing, and attention problems in young children: validation of the MacArthur HBQ. Journal of the American Academy of Child and Adolescent Psychiatry. Oct 2007;46(10):1315-1323. 119. Du P, Kibbe WA, Lin SM. lumi: a pipeline for processing Illumina microarray. Bioinformatics. Jul 1 2008;24(13):1547-1548. 120. Price ME, Cotton AM, Lam LL, et al. Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array. Epigenetics & chromatin. 2013;6(1):4. 121. Maksimovic J, Gordon L, Oshlack A. SWAN: Subset-quantile within array normalization for illumina infinium HumanMethylation450 BeadChips. Genome biology. 2012;13(6):R44. 122. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. Jan 2007;8(1):118-127. 123. Peters TJ, Buckley MJ, Statham AL, et al. De novo identification of differentially methylated regions in the human genome. Epigenetics & chromatin. 2015;8:6. 124. Drong AW, Nicholson G, Hedman AK, et al. The presence of methylation quantitative trait loci indicates a direct genetic influence on the level of DNA methylation in adipose tissue. PloS one. 2013;8(2):e55923. 125. Gaunt TR, Shihab HA, Hemani G, et al. Systematic identification of genetic influences on methylation across the human life course. Genome biology. Mar 31 2016;17:61. 126. Jiao J, Opal MD, Dulawa SC. Gestational environment programs adult depression-like behavior through methylation of the calcitonin gene-related peptide gene. Molecular psychiatry. Dec 2013;18(12):1273-1280. 127. Tumpel S, Wiedemann LM, Krumlauf R. Hox genes and segmentation of the vertebrate hindbrain. Curr. Top. Dev. Biol. 2009;88:103-137. 128. Solanky N, Requena Jimenez A, D'Souza SW, Sibley CP, Glazier JD. Expression of folate transporters in human placenta and implications for homocysteine metabolism. Placenta. Feb 2010;31(2):134-143. 129. Crider KS, Yang TP, Berry RJ, Bailey LB. Folate and DNA methylation: a review of molecular mechanisms and the evidence for folate's role. Advances in nutrition (Bethesda, Md.). Jan 2012;3(1):21-38. 130. Lappalainen T, Greally JM. Associating cellular epigenetic models with human phenotypes. Nature reviews. Genetics. Jul 2017;18(7):441-451. 105  131. Smith AK, Kilaru V, Klengel T, et al. DNA extracted from saliva for methylation studies of psychiatric traits: evidence tissue specificity and relatedness to brain. Am. J. Med. Genet. B Neuropsychiatr. Genet. Jan 2015;168B(1):36-44. 132. Do C, Shearer A, Suzuki M, et al. Genetic-epigenetic interactions in cis: a major focus in the post-GWAS era. Genome biology. Jun 19 2017;18(1):120. 133. Brennan PA, Hammen C, Andersen MJ, Bor W, Najman JM, Williams GM. Chronicity, severity, and timing of maternal depressive symptoms: relationships with child outcomes at age 5. Dev. Psychol. Nov 2000;36(6):759-766. 134. Hammen C, Brennan PA. Severity, chronicity, and timing of maternal depression and risk for adolescent offspring diagnoses in a community sample. Archives of general psychiatry. Mar 2003;60(3):253-258. 135. O'Donnell KJ, Glover V, Barker ED, O'Connor TG. The persisting effect of maternal mood in pregnancy on childhood psychopathology. Development and psychopathology. May 2014;26(2):393-403. 136. Cents RA, Diamantopoulou S, Hudziak JJ, et al. Trajectories of maternal depressive symptoms predict child problem behaviour: the Generation R study. Psychological medicine. Jan 2013;43(1):13-25. 137. Jung T, Wickrama KAS. An Introduction to Latent Class Growth Analysis and Growth Mixture Modeling. Soc. Personal. Psychol. Compass. 2008;2(1):302-317. 138. Campbell SB, Matestic P, von Stauffenberg C, Mohan R, Kirchner T. Trajectories of maternal depressive symptoms, maternal sensitivity, and children's functioning at school entry. Dev. Psychol. Sep 2007;43(5):1202-1215. 139. Campbell SB, Morgan-Lopez AA, Cox MJ, McLoyd VC. A latent class analysis of maternal depressive symptoms over 12 years and offspring adjustment in adolescence. Journal of abnormal psychology. Aug 2009;118(3):479-493. 140. Guyon-Harris K, Huth-Bocks A, Lauterbach D, Janisse H. Trajectories of maternal depressive symptoms across the birth of a child: associations with toddler emotional development. Archives of women's mental health. Jul 17 2015. 141. van der Waerden J, Galera C, Larroque B, Saurel-Cubizolles MJ, Sutter-Dallay AL, Melchior M. Maternal Depression Trajectories and Children's Behavior at Age 5 Years. J. Pediatr. Jun 2015;166(6):1440-1448 e1441. 142. Betts KS, Williams GM, Najman JM, Alati R. Maternal depressive, anxious, and stress symptoms during pregnancy predict internalizing problems in adolescence. Depression and anxiety. Jan 2014;31(1):9-18. 143. Betts KS, Williams GM, Najman JM, Alati R. The relationship between maternal depressive, anxious, and stress symptoms during pregnancy and adult offspring behavioral and emotional problems. Depression and anxiety. Feb 2015;32(2):82-90. 144. Matijasevich A, Murray J, Cooper PJ, et al. Trajectories of maternal depression and offspring psychopathology at 6 years: 2004 Pelotas cohort study. J. Affect. Disord. Mar 15 2015;174:424-431. 145. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. The British journal of psychiatry : the journal of mental science. Jun 1987;150:782-786. 146. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Archives of general psychiatry. Jun 1961;4:561-571. 106  147. McEwen BS, Morrison JH. The brain on stress: vulnerability and plasticity of the prefrontal cortex over the life course. Neuron. Jul 10 2013;79(1):16-29. 148. Hamilton M. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry. Feb 1960;23:56-62. 149. Madigan S, Wade M, Plamondon A, Browne D, Jenkins JM. Birth Weight Variability and Language Development: Risk, Resilience, and Responsive Parenting. J. Pediatr. Psychol. Oct 2015;40(9):869-877. 150. Stewart DE, Vigod S. Postpartum Depression. The New England journal of medicine. Dec 01 2016;375(22):2177-2186. 151. Achenbach TM. The Child Behavior Profile: I. Boys aged 6–11. J. Consult. Clin. Psychol. 1978;46(3):478-488. 152. Boyce WT, Essex MJ, Woodward HR, Measelle JR, Ablow JC, Kupfer DJ. The confluence of mental, physical, social, and academic difficulties in middle childhood. I: exploring the "head waters" of early life morbidities. Journal of the American Academy of Child and Adolescent Psychiatry. May 2002;41(5):580-587. 153. Boyle MH, Offord DR, Racine Y, Fleming JE, Szatmari P, Sanford M. Evaluation of the revised Ontario Child Health Study scales. Journal of child psychology and psychiatry, and allied disciplines. Feb 1993;34(2):189-213. 154. Luby JL, Heffelfinger A, Measelle JR, et al. Differential performance of the macarthur HBQ and DISC-IV in identifying DSM-IV internalizing psychopathology in young children. Journal of the American Academy of Child and Adolescent Psychiatry. Apr 2002;41(4):458-466. 155. Gioia GA, Isquith PK, Guy SC, Kenworthy L. Behavior rating inventory of executive function. Child Neuropsychol. Sep 2000;6(3):235-238. 156. Davidson MC, Amso D, Anderson LC, Diamond A. Development of cognitive control and executive functions from 4 to 13 years: evidence from manipulations of memory, inhibition, and task switching. Neuropsychologia. 2006;44(11):2037-2078. 157. Muthén B, Muthén LK. Integrating Person-Centered and Variable-Centered Analyses: Growth Mixture Modeling With Latent Trajectory Classes. Alcoholism: Clinical and Experimental Research. 2000;24(6):882-891. 158. Nylund KL, Asparoutiov T, Muthen BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural equation modeling. 2007;14(4):535-569. 159. Wang M, Bodner TE. Growth mixture modeling: Identifying and predicting unobserved Subpopulations with longitudinal data. Organizational research methods. 01/2008 2007;11(1):197-197. 160. Najman JM, Plotnikova M, Williams GM, et al. Trajectories of maternal depression: a 27-year population-based prospective study. Epidemiology and psychiatric sciences. Jan 19 2016:1-10. 161. Glover V, O'Connor TG, O'Donnell K. Prenatal stress and the programming of the HPA axis. Neuroscience and biobehavioral reviews. Sep 2010;35(1):17-22. 162. Evans J, Melotti R, Heron J, et al. The timing of maternal depressive symptoms and child cognitive development: a longitudinal study. Journal of child psychology and psychiatry, and allied disciplines. Jun 2012;53(6):632-640. 107  163. Esteban-Cornejo I, Henriksson P, Cadenas-Sanchez C, et al. Early life programming of attention capacity in adolescents: The HELENA study. Maternal & child nutrition. Apr 11 2017. 164. Paulson JF, Dauber S, Leiferman JA. Individual and combined effects of postpartum depression in mothers and fathers on parenting behavior. Pediatrics. Aug 2006;118(2):659-668. 165. McLearn KT, Minkovitz CS, Strobino DM, Marks E, Hou W. Maternal depressive symptoms at 2 to 4 months post partum and early parenting practices. Arch. Pediatr. Adolesc. Med. Mar 2006;160(3):279-284. 166. Bernier A, Carlson SM, Whipple N. From external regulation to self-regulation: early parenting precursors of young children's executive functioning. Child development. Jan-Feb 2010;81(1):326-339. 167. Hughes CH, Ensor RA. How do families help or hinder the emergence of early executive function? New Dir. Child Adolesc. Dev. Spring 2009;2009(123):35-50. 168. Maughan A, Cicchetti D, Toth SL, Rogosch FA. Early-occurring maternal depression and maternal negativity in predicting young children's emotion regulation and socioemotional difficulties. J. Abnorm. Child Psychol. Oct 2007;35(5):685-703. 169. Toth SL, Rogosch FA, Sturge-Apple M, Cicchetti D. Maternal depression, children's attachment security, and representational development: an organizational perspective. Child development. Jan-Feb 2009;80(1):192-208. 170. Diamond A. Executive functions. Annual review of psychology. 2013;64:135-168. 171. Snyder HR. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: a meta-analysis and review. Psychological bulletin. Jan 2013;139(1):81-132. 172. Diamond A. Why assessing and improving executive functions early in life is critical. In: McCardle P, Freund, L. & Griffin, J. A., ed. Executive Function in Preschool-age children: Integrating Measurement, Neurodevelopment, and Translational research. Washington, DC: American Psychological Association; 2016:11 - 43. 173. Moffitt TE, Arseneault L, Belsky D, et al. A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences of the United States of America. Feb 15 2011;108(7):2693-2698. 174. Najman JM, Williams GM, Nikles J, et al. Bias influencing maternal reports of child behaviour and emotional state. Soc. Psychiatry Psychiatr. Epidemiol. Apr 2001;36(4):186-194. 175. Marmot M. Social determinants of health inequalities. Lancet. Mar 19-25 2005;365(9464):1099-1104. 176. Doyle O, Harmon CP, Heckman JJ, Tremblay RE. Investing in Early Human Development: Timing and Economic Efficiency. Econ. Hum. Biol. 01/21 2009;7(1):1-6. 177. O'Connor E, Rossom RC, Henninger M, Groom HC, Burda BU. Primary Care Screening for and Treatment of Depression in Pregnant and Postpartum Women: Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA : the journal of the American Medical Association. Jan 26 2016;315(4):388-406. 178. Stice E, Shaw H, Bohon C, Marti CN, Rohde P. A meta-analytic review of depression prevention programs for children and adolescents: factors that predict magnitude of 108  intervention effects. Journal of consulting and clinical psychology. Jun 2009;77(3):486-503. 179. Forman DR, O'Hara MW, Stuart S, Gorman LL, Larsen KE, Coy KC. Effective treatment for postpartum depression is not sufficient to improve the developing mother-child relationship. Development and psychopathology. Spring 2007;19(2):585-602. 180. Sockol LE, Epperson CN, Barber JP. Preventing postpartum depression: a meta-analytic review. Clin. Psychol. Rev. Dec 2013;33(8):1205-1217. 181. Beardslee WR, Gladstone TR, Wright EJ, Cooper AB. A family-based approach to the prevention of depressive symptoms in children at risk: evidence of parental and child change. Pediatrics. Aug 2003;112(2):e119-131. 182. Bee P, Bower P, Byford S, et al. The clinical effectiveness, cost-effectiveness and acceptability of community-based interventions aimed at improving or maintaining quality of life in children of parents with serious mental illness: a systematic review. Health Technol. Assess. Feb 2014;18(8):1-250. 183. Barlow J, Coren E. Parent-training programmes for improving maternal psychosocial health. The Cochrane database of systematic reviews. 2004(1):CD002020. 184. Shaw DS, Taraban LE. New Directions and Challenges in Preventing Conduct Problems in Early Childhood. Child development perspectives. 11/16 2017;11(2):85-89.  109  Appendix Supplementary Table A 1. Definitions used in the study. Variable Data source(s) Codes Definition Pregnancy trimesters    Conception date PDR  Final gestational age + 14 days 1st PDR  Conception date - 77 days 2nd PDR  Day 78 - 166 3rd PDR  Day 167 - birth     Drug exposures    Antidepressants PNet ATC: N06AA01, N06AA02, N06AA04, N06AA09, N06AA10, N06AA11, N06AA12, N06AB02, N06AB03, N06AB04, N06AB05, N06AB06, N06AB07, N06AB08, N06AB09, N06AB10, N06AF03, N06AF04, N06AX01, N06AX03, N06AX05, N06AX06, N06AX11, N06AX12, N06AX16, N06AX21, N06AX23, N06AX24 Pregnancy: 2+ during pregnancy/1+ from 2nd trimester onwards Preconception: 1+ during 365 - 90 days prior to conception SSRIs PNet ATC: N06AB02, N06AB03,  110  N06AB04, N06AB05, N06AB06, N06AB07, N06AB08, N06AB09, N06AB10 Other  PNet ATC: N06AA01, N06AA02, N06AA04, N06AA09, N06AA10, N06AA11, N06AA12, N06AF03, N06AF04, N06AX01, N06AX03, N06AX05, N06AX06, N06AX11, N06AX12, N06AX16, N06AX21, N06AX23, N06AX24  Other psychotropic medications PNet ATC: N05A, N05B 1+ during pregnancy Antiepileptic medications  PNet ATC: NO3A 1+ during pregnancy     Psychiatric disorders    Bipolar MSP, DAD ICD9: 296 ICD10: F30, F31, F34.0 1+ during pregnancy Depression MSP, DAD ICD9: 311, 300.4 ICD10: F32, F33, F34 (excluding F34.0), F38, F39 1+ during pregnancy Anxiety  MSP, DAD ICD9: 300 (excluding 300.4) ICD10: F40, F41, F42 1+ during pregnancy ATC, Anatomical Therapeutic Classifications; ICD, International Classification of Diseases; PNet, Pharmanet; DAD, Discharge Abstract Database; MSP, Medical Services Plan  111  Supplementary Table A 2. Top 500 variables used for HDPS estimation. Data source Variable* Type Exposure-variable ranking score PNet VENLAFAXINE Frequent 4.110742825 PNet PAROXETINE Frequent 4.068218983 PNet SERTRALINE Frequent 4.018622042 PNet FLUOXETINE Frequent 3.80470114 PNet CITALOPRAM Frequent 3.643383534 PNet VENLAFAXINE Sporadic 3.611543964 PNet PAROXETINE Sporadic 3.582296098 PNet SERTRALINE Sporadic 3.574334547 PNet QUETIAPINE Frequent 3.556628206 PNet FLUVOXAMINE Frequent 3.523838383 PNet TRAZODONE Frequent 3.520208615 PNet RISPERIDONE Frequent 3.360286929 PNet FLUVOXAMINE Sporadic 3.356784299 PNet CLONAZEPAM Frequent 3.345951424 PNet FLUOXETINE Sporadic 3.341516827 PNet QUETIAPINE Sporadic 3.16057363 DAD F322 Once 3.113599996 PNet CITALOPRAM Sporadic 3.098817696 PNet CLONAZEPAM Sporadic 3.0952657 PNet VENLAFAXINE Once 3.084866486 PNet QUETIAPINE Once 3.058030145 PNet SERTRALINE Once 3.020030332 DAD F329 Once 2.998572075 PNet RISPERIDONE Sporadic 2.995959971 PNet GABAPENTIN Frequent 2.957750081 PNet BUPROPION Frequent 2.944019888 PNet PAROXETINE Once 2.926291448 MSP 311 Frequent 2.859583841 PNet CITALOPRAM Once 2.765319001 PNet TRAZODONE Once 2.739589841 PNet RISPERIDONE Once 2.734946146 PNet FLUOXETINE Once 2.720164915 PNet CLONAZEPAM Once 2.705869882 MSP 296 Frequent 2.696101049 PNet FLUVOXAMINE Once 2.689954426 PNet VALPROIC ACID Frequent 2.674686954 MSP 311 Sporadic 2.653473431 MSP 610 Once 2.625601325 112  PNet GABAPENTIN Sporadic 2.604619392 MSP 4091 Frequent 2.591451514 MSP 296 Sporadic 2.585074796 PNet NORTRIPTYLINE Frequent 2.554437826 MSP 296 Once 2.541701086 PNet VALPROIC ACID Sporadic 2.534522516 PNet ALPRAZOLAM Frequent 2.532512465 PNet DIAZEPAM Frequent 2.520536274 PNet VALPROIC ACID Once 2.455531444 DAD X61 Once 2.433524897 PNet AMITRIPTYLINE Frequent 2.383335153 PNet LORAZEPAM Frequent 2.341255774 DAD 1FE57JA Once 2.297392723 PNet ZOPICLONE Frequent 2.280710223 PNet OXAZEPAM Frequent 2.253907611 DAD F432 Once 2.232854202 MSP 304 Frequent 2.188402439 PNet GABAPENTIN Once 2.176158858 PNet TEMAZEPAM Frequent 2.115071166 MSP 300 Frequent 2.114143092 MSP 4091 Sporadic 2.072898655 PNet ALPRAZOLAM Once 2.051822136 DAD 66331 Once 2.012063219 PNet MORPHINE Frequent 2.009710651 MSP 4091 Once 1.982047323 MSP 311 Once 1.942188144 PNet BUPROPION Once 1.941799445 MSP 304 Once 1.941451994 DAD U980 Once 1.910770703 DAD Y831 Once 1.901497066 MSP 9365 Sporadic 1.879189937 MSP 50B Frequent 1.878199646 DAD 5MD40JC Once 1.869948708 PNet ATENOLOL Frequent 1.855559971 PNet PINAVERIUM Frequent 1.845407599 DAD Z33 Once 1.827389094 PNet ZOPICLONE Once 1.797279292 DAD 1BN72LA Once 1.77859893 MSP 9365 Frequent 1.773341974 PNet ORLISTAT Frequent 1.773321873 PNet RABEPRAZOLE Frequent 1.772126415 113  DAD 1NT87LA Once 1.766764472 DAD 5MD54KM Once 1.766764472 DAD 1UC87WJ Once 1.766764472 MSP 9104 Frequent 1.738616262 PNet PETHIDINE Frequent 1.709606058 DAD U989 Once 1.697335966 MSP 3000 Frequent 1.693857701 PNet ATENOLOL Sporadic 1.69125692 PNet LORAZEPAM Once 1.691121244 PNet OXAZEPAM Once 1.686672042 DAD O081 Once 1.673238414 PNet AMITRIPTYLINE Once 1.658372503 DAD O99001 Once 1.621052661 PNet TEMAZEPAM Once 1.619749729 PNet NORTRIPTYLINE Once 1.587716241 MSP 300 Once 1.579804407 MSP 9365 Once 1.576582215 PNet DIAZEPAM Once 1.573608573 PNet LANSOPRAZOLE Frequent 1.562281343 MSP 309 Frequent 1.557879622 PNet LABETALOL Frequent 1.555455378 MSP 100 Once 1.546665454 DAD 2RM71DA Once 1.539707021 PNet ORLISTAT Once 1.534041284 DAD 1RF87LA Once 1.527284501 PNet PANTOPRAZOLE Frequent 1.520364059 PNet ZOLMITRIPTAN Frequent 1.519087734 DAD O99301 Once 1.508935363 DAD 1RN87CAAK Once 1.508935363 DAD 65221 Once 1.504815416 MSP 308 Frequent 1.503511095 PNet LABETALOL Sporadic 1.502893048 DAD 1YM78LAXXE Once 1.496031958 PNet PROPRANOLOL Frequent 1.490916857 MSP 4560 Once 1.478743176 DAD O26801 Once 1.47071415 DAD 3PH10XJ Once 1.4644836 MSP 120 Frequent 1.460123445 MSP 108 Frequent 1.421923986 DAD 2NA71BA Once 1.421923986 DAD 1RM59BAGX Once 1.421923986 114  PNet MORPHINE Once 1.421923986 PNet HYDROMORPHONE Once 1.41437678 DAD 6146 Once 1.401437062 DAD 65651 Once 1.401437062 DAD N62 Once 1.398393488 PNet MICONAZOLE Once 1.38377222 MSP 9349 Once 1.382277792 MSP 9104 Once 1.369427013 PNet DOXYLAMINE, COMBINATIONS Frequent 1.367856765 MSP 9100 Once 1.366438792 PNet OMEPRAZOLE Frequent 1.365228642 DAD 6320 Once 1.361027524 DAD 705 Once 1.341881278 PNet PROPRANOLOL Once 1.341881278 PNet FERROUS GLUCONATE Once 1.337840869 DAD O62001 Once 1.333370588 PNet PETHIDINE Once 1.323343157 MSP 109 Once 1.32299436 MSP 9231 Once 1.319683793 PNet PANTOPRAZOLE Once 1.317783727 MSP 3000 Once 1.317390258 DAD 1FE89WL Once 1.31656347 PNet PINAVERIUM Once 1.31656347 PNet RIZATRIPTAN Once 1.313443343 PNet CYANOCOBALAMIN Once 1.309446002 MSP 346 Frequent 1.30609217 MSP 9349 Frequent 1.303747817 PNet ESOMEPRAZOLE Frequent 1.289996443 DAD O47103 Once 1.288392593 PNet LABETALOL Once 1.28147215 PNet ATENOLOL Once 1.277342757 DAD O14001 Once 1.267773306 DAD D649 Once 1.267773306 DAD N301 Once 1.267773306 DAD 5MD53KK Once 1.267773306 DAD 1RN87CAGX Once 1.267773306 DAD 1HH59GPAW Once 1.267773306 PNet RABEPRAZOLE Once 1.267773306 PNet CARBAMAZEPINE Frequent 1.267773306 PNet SUMATRIPTAN Frequent 1.267773306 MSP 50B Once 1.258039868 115  PNet SALMETEROL AND FLUTICASONE Once 1.256256864 DAD Y838 Once 1.239602429 PNet CARBAMAZEPINE Sporadic 1.237620268 PNet HYDROCHLOROTHIAZIDE Frequent 1.236520762 DAD O42011 Once 1.232681986 DAD 65813 Once 1.232681986 DAD K011 Once 1.232681986 DAD O66901 Once 1.228967802 DAD 8501 Once 1.228967802 PNet PROCHLORPERAZINE Once 1.218583062 PNet CARBAMAZEPINE Once 1.217623522 DAD 1691 Once 1.217133344 DAD O61001 Once 1.204859481 MSP 9101 Once 1.203348289 DAD 2RM70CA Once 1.198780434 DAD 0134 Once 1.198780434 MSP 9013 Once 1.198487683 MSP 9232 Once 1.195471847 DAD 65631 Once 1.193035792 DAD 5LD07RAFX Once 1.193035792 DAD O36431 Once 1.191400327 PNet METOCLOPRAMIDE Once 1.183925333 DAD 5MD54NF Once 1.1807657 DAD 8421 Once 1.1807657 PNet SIBUTRAMINE Sporadic 1.172463126 MSP 9655 Once 1.171746473 MSP 309 Once 1.171097006 PNet FLUTICASONE Frequent 1.17042802 DAD 64501 Once 1.14941517 MSP 1054 Once 1.133363461 MSP 1051 Once 1.128940632 PNet ZOLMITRIPTAN Once 1.122407456 PNet ISOTRETINOIN Frequent 1.116542336 DAD K529 Once 1.103470255 PNet HYDROCHLOROTHIAZIDE Once 1.100340362 PNet SALMETEROL AND FLUTICASONE Frequent 1.085451749 PNet METFORMIN Frequent 1.085451749 DAD 64822 Once 1.083601951 DAD O41131 Once 1.08205616 DAD 8752 Once 1.076775986 PNet RIZATRIPTAN Frequent 1.073617292 116  PNet BUDESONIDE Frequent 1.073617292 DAD R104 Once 1.061921252 PNet FOLLITROPIN BETA Frequent 1.061921252 DAD 1PM50BAD2 Once 1.054199206 PNet SIBUTRAMINE Once 1.046536333 DAD 5AC30GUI2 Once 1.043270135 MSP 108 Once 1.035396079 DAD 2NM71BA Once 1.02576127 MSP 308 Once 1.019808965 MSP 91040 Once 1.018499695 DAD 857 Once 1.016793282 DAD D069 Once 1.016458878 DAD O30001 Once 1.016458878 DAD R102 Once 1.016458878 DAD 2NK71BA Once 1.016458878 DAD 1RD72DAGX Once 1.016458878 PNet SUMATRIPTAN Once 1.013752516 DAD J350 Once 1.00582425 MSP 354 Frequent 1.005409041 PNet LANSOPRAZOLE Once 0.995515704 DAD O13003 Once 0.989059903 MSP 7840 Frequent 0.988188444 PNet DICLOFENAC, COMBINATIONS Once 0.986605915 MSP 120 Once 0.985592821 MSP 9231 Frequent 0.982013602 PNet SALBUTAMOL Frequent 0.981520826 MSP 9230 Once 0.968366688 MSP 91326 Once 0.965533569 MSP 138 Once 0.965279745 DAD 2OT71DA Once 0.962391656 DAD 1FE89LA Once 0.962391656 PNet DOXYLAMINE, COMBINATIONS Once 0.961248146 PNet ACETAMINOPHEN, COMB EXCL PSYCHOLEPTICS Once 0.957311205 MSP 91126 Once 0.957035457 MSP 138 Frequent 0.95452792 PNet PHENTERMINE Once 0.946500289 DAD 1PQ50BABJ Once 0.944999914 DAD 2NF71BA Once 0.942350906 DAD O75803 Once 0.93641617 PNet NORELGESTROMIN AND ESTROGEN Frequent 0.93641617 117  PNet BROMOCRIPTINE Frequent 0.93641617 MSP 1822 Once 0.934943633 MSP 138 Sporadic 0.931301069 PNet PHENTERMINE Frequent 0.925487099 PNet OFLOXACIN Once 0.920230846 MSP 91725 Once 0.919545151 MSP 100 Sporadic 0.915143723 MSP 354 Once 0.913959594 MSP 537 Frequent 0.911098362 DAD O72204 Once 0.911098362 DAD 5PC20HAL6 Once 0.911098362 PNet TERBUTALINE Frequent 0.911098362 PNet CYCLOBENZAPRINE Once 0.9051379 MSP 346 Once 0.902842381 PNet SCOPOLAMINE Once 0.892749223 DAD O47003 Once 0.891295735 DAD O34201 Once 0.889119455 PNet ESTRADIOL Frequent 0.889119455 PNet SIBUTRAMINE Frequent 0.886405749 MSP 473 Once 0.884622688 DAD 66111 Once 0.880661107 PNet TERBUTALINE Once 0.880326703 MSP 13200 Frequent 0.879867039 DAD N871 Once 0.878308539 DAD O13001 Once 0.876412804 MSP 91655 Once 0.876268971 MSP 9244 Once 0.876105291 PNet ISOTRETINOIN Sporadic 0.874284389 PNet HEPATITIS A, INACTIVATED, WHOLE VIRUS Once 0.872292858 MSP 91245 Once 0.871877649 MSP 107 Once 0.865970392 DAD 1YS87LA Once 0.862308198 MSP 410 Once 0.854100217 MSP 01A Once 0.853783905 MSP 1811 Once 0.853397071 DAD Z371 Once 0.852829454 MSP 278 Frequent 0.848123563 DAD 65661 Once 0.846759556 MSP 91070 Once 0.843304276 MSP 128 Frequent 0.83215516 DAD Z292 Once 0.831961041 118  DAD 2NQ71BA Once 0.831055654 MSP 1831 Once 0.829925072 MSP 92231 Once 0.829696035 PNet CYPROTERONE AND ESTROGEN Frequent 0.826757214 MSP 92100 Once 0.82067021 MSP 1812 Once 0.819826779 MSP 349 Once 0.818183393 DAD 64121 Once 0.815788182 DAD N831 Once 0.815788182 DAD 1RF87DA Once 0.815788182 MSP 486 Once 0.805149784 MSP 90690 Once 0.803240966 DAD O72002 Once 0.797769677 DAD N390 Once 0.793315326 PNet EPINEPHRINE Once 0.792349609 MSP 401 Frequent 0.791690631 MSP 91210 Once 0.787870736 PNet METHYLPREDNISOLONE Once 0.787273012 MSP 13611 Once 0.783957538 PNet FERROUS SULPNetTE Once 0.78326499 MSP 91065 Once 0.783146854 DAD O24401 Once 0.78226549 PNet SALBUTAMOL Once 0.781218041 MSP 92368 Once 0.778931008 MSP 113 Once 0.777784384 MSP 1832 Once 0.775774595 MSP 91901 Once 0.77555287 DAD 650 Once 0.775300592 PNet LEVOTHYROXINE SODIUM Frequent 0.773360712 DAD N840 Once 0.77133642 DAD 6786 0Once 0.77133642 DAD 1RW52LA Once 0.77133642 DAD 5PC91HU Once 0.77133642 PNet METFORMIN Sporadic 0.764669728 DAD 5MD54KJ Once 0.762678357 MSP 7807 Once 0.761929013 DAD 4939 Once 0.756947682 DAD 1RN59CAAG Once 0.756947682 DAD 5AC30CKI2 Once 0.756947682 MSP 9902 Frequent 0.756608459 MSP 91421 Once 0.752935641 119  MSP 92450 Once 0.749635035 DAD 66201 Once 0.747129715 MSP 92376 Once 0.742492228 MSP 1821 Once 0.740877212 DAD N832 Once 0.740752996 PNet INSULIN (HUMAN) Frequent 0.739248105 PNet BUDESONIDE Once 0.738291625 PNet CEFUROXIME Once 0.736381405 MSP 90440 Once 0.731512785 DAD 6500 Once 0.731129373 MSP 128 Once 0.728776805 DAD O26803 Once 0.728776805 DAD 5AC30CK Once 0.728776805 DAD 5AB03JA Once 0.728776805 MSP 7840 Once 0.727042199 PNet MUPIROCIN Once 0.724720005 MSP 696 Frequent 0.724656859 PNet CEFIXIME Once 0.720871626 PNet BECLOMETHASONE Once 0.720266116 DAD 8798 Once 0.718142178 MSP 1841 Once 0.715302107 PNet SPIRONOLACTONE Frequent 0.714983483 PNet BECLOMETASONE Once 0.712647423 DAD 1RM87BAGX Once 0.712247503 PNet BENZOYL PEROXIDE Once 0.712247503 PNet FLUTICASONE Once 0.71159198 PNet INDOMETACIN Once 0.711077228 MSP 427 Once 0.710427667 MSP 9482 Once 0.709477295 DAD O45901 Once 0.708157518 PNet DIPHENOXYLATE Once 0.708157518 MSP 1842 Once 0.707401946 MSP 845 Frequent 0.706303949 PNet ROFECOXIB Once 0.702697092 DAD O42111 Once 0.698339551 DAD 841 Once 0.698339551 DAD O60001 Once 0.695986982 DAD 2NK70BA Once 0.695986982 PNet LEVOTHYROXINE SODIUM Sporadic 0.692616824 MSP 90305 Once 0.692075438 MSP 781 Frequent 0.691263029 120  PNet ISOTRETINOIN Once 0.691036477 PNet FUSIDIC ACID Once 0.685186369 PNet PHENAZOPYRIDINE Once 0.685032256 PNet CYPROTERONE AND ESTROGEN Sporadic 0.683881485 MSP 8470 Sporadic 0.683166294 MSP 100 Frequent 0.680554349 PNet KETOROLAC Once 0.678766385 MSP 530 Once 0.674709584 DAD 2PM70BA Once 0.674709584 DAD K8010 Once 0.672687339 PNet CLOTRIMAZOLE Once 0.672687339 MSP 8470 Frequent 0.670957234 DAD 5AC30AP Once 0.670957234 DAD 1OD89DA Once 0.668152183 PNet TACROLIMUS Once 0.668152183 PNet NORFLOXACIN Once 0.665802006 PNet VALACICLOVIR Frequent 0.665704202 DAD 5MD60JW Once 0.664238284 PNet DOMPERIDONE Once 0.66319716 MSP 90515 Once 0.662686693 PNet ACYCLOVIR Once 0.660873852 MSP 493 Frequent 0.659783934 DAD 2RM70BA Once 0.657317841 PNet IBUPROFEN Once 0.657172966 PNet CIMETIDINE Once 0.655646567 DAD 66411 Once 0.654039292 PNet ESOMEPRAZOLE Once 0.653016759 PNet METFORMIN Once 0.650873136 MSP 724 Frequent 0.649446602 DAD O75701 Once 0.648734098 DAD 1RB52DA Once 0.648734098 DAD 1EW52BA Once 0.648734098 DAD 5682 Once 0.648734098 DAD 1KR87WM Once 0.648734098 PNet MESALAZINE Frequent 0.648734098 PNet RANITIDINE Once 0.647898326 DAD Z223 Once 0.639164647 PNet BROMOCRIPTINE Once 0.635995072 MSP V242 Once 0.634097282 PNet METHOCARBAMOL, COMBINATIONS EXCL PSYCHOL Once 0.633466625 MSP 646 Frequent 0.630551778 121  PNet AMOXICILLIN AND ENZYME INHIBITOR Once 0.630336732 MSP 537 Once 0.630257821 PNet AZITHROMYCIN Once 0.630051155 DAD 1OT72DA Once 0.626994111 DAD 8789 2Once 0.625256694 DAD 2OT70DA Once 0.624732945 MSP 4008 Once 0.62341629 DAD O98801 Once 0.62341629 PNet VALDECOXIB Once 0.620296162 PNet TETRACYCLINE Once 0.618166934 MSP 92382 Once 0.615117487 PNet PHENTERMINE Sporadic 0.615117487 DAD 1RM87DAGX Once 0.61099377 DAD 6396 Once 0.61099377 PNet NAPROXEN Once 0.610158696 PNet CIPROFLOXACIN Once 0.610138494 DAD 8782 Once 0.604028871 MSP 92330 Once 0.60262552 DAD 65641 Once 0.598723677 PNet CODEINE, COMBINATIONS EXCL. PSYCHOLEPTIC Once 0.59864072 MSP 7245 Once 0.597584595 MSP 90205 Frequent 0.595873565 MSP 7245 Frequent 0.59452728 MSP 493 Once 0.594204288 DAD 787 4 Once 0.592979035 DAD J4590 Once 0.592644631 PNet ACICLOVIR Once 0.590541581 PNet MESALAZINE Sporadic 0.589014863 MSP 01H Once 0.587980202 PNet VALACICLOVIR Once 0.58708436 MSP 787 Frequent 0.587017858 PNet FLUCONAZOLE Frequent 0.586829667 PNet CLINDAMYCIN Once 0.586100434 MSP 790 Frequent 0.585264524 PNet CLARITHROMYCIN Once 0.583701627 DAD Z2238 Once 0.583594795 MSP 558 Once 0.581941447 PNet LEVOTHYROXINE SODIUM Once 0.58145209 DAD 8101 Once 0.579931319 PNet CELECOXIB Once 0.579564407 122  MSP 461 Once 0.578821202 MSP 8525 Once 0.577135537 MSP 034 Once 0.574626125 DAD O020 Once 0.574626125 DAD 2RM71BA Once 0.574626125 DAD 6619 Once 0.574626125 DAD 0116 Once 0.574626125 DAD 1NV89DA Once 0.574626125 DAD 1PE59KQAQ Once 0.574626125 DAD 5LD31AP Once 0.571456551 MSP 92165 Once 0.570873776 DAD O24411 Once 0.570506179 DAD 66421 Once 0.570506179 DAD 0289 Once 0.570506179 DAD 590 Once 0.570506179 PNet LANSOPRAZOLE, AMOXICILLIN AND CLARITHROM Once 0.570506179 PNet PREDNISONE Once 0.56914666 PNet NAFARELIN Once 0.567977733 MSP 90745 Once 0.567925933 MSP 3333 Once 0.567658203 MSP 715 Once 0.567508658 DAD 5LD20HAP1 Once 0.567508658 MSP 427 Frequent 0.566720946 PNet BETAHISTINE Once 0.566091223 DAD 65811 Once 0.56290158 DAD O24491 Once 0.562791668 PNet ESTRADIOL Once 0.562791668 DAD V270 Once 0.561373695 MSP 8573 Once 0.56062683 DAD 63492 Once 0.56044149 MSP 788 Frequent 0.558952891 MSP 995 Once 0.558743079 MSP 690 Once 0.556985432 DAD R1030 Once 0.556926548 PNet BUSERELIN Frequent 0.556926548 MSP 91000 Once 0.556272777 MSP 91745 Frequent 0.556003613 MSP 45A Frequent 0.554979128 MSP 845 Once 0.55479732 MSP 90315 Once 0.551095628 DAD 1RB87LA Once 0.551095628 123  MSP 190 Frequent 0.550600946 MSP 765 Once 0.550085016 PNet SPIRONOLACTONE Once 0.548983695 MSP 617 Once 0.548829338 DAD D24 Once 0.547516661 PNet IRON, MULTIVITAMINS AND MINERALS Frequent 0.546455248 MSP 310 Once 0.545838155 MSP 02A Frequent 0.543994122 MSP 90736 Once 0.54193839 PNet DESONIDE Once 0.54168574 MSP 90800 Once 0.541614241 MSP 90810 Once 0.541614241 MSP 478 Once 0.540238783 *Variables from Pharmanet refer to drug names, from MSP refer to ICD-9 codes and fee items, and from DAD refer to ICD-10 codes and procedure codes  DAD, Discharge Abstract Database; MSP, Medical Services Plan; PNet, Pharmanet;    124  Supplementary Table A 3. Demographic characteristics of exposed mother-child dyads, by whether they were matched or unmatched to a control based on a high dimensional propensity score.   Prenatally exposed mother-child dyads  Unmatched (n = 1,408) Matched (n=2,253) Standardized difference Maternal characteristics  Maternal age at delivery, mean (SD)  31.03 (5.44) 30.30 (5.58) -0.12 Multiparous 891 (63.28) 1,381 (61.30) -0.02 Lowest neighborhood income quintile 314 (22.30) 508 (22.55) 0.01 Mood/anxiety disorders diagnosis during pregnancy 1,193 (84.73)  1,159 (51.44) -0.27 Other psychotropic or antiepileptic medications during pregnancy 428 (30.40) 394 (17.49) -0.30     Anti-depressant exposure during pregnancy    SSRIs only 904 (64.20) 1,546 (68.62) 0.04 Non-SSRIs only 322 (22.87) 548 (24.32) 0.04 Both SSRI and non-SSRIs 182 (12.93) 159 (7.06) -0.34     2nd – 3rd trimester 1,267 (89.99) 1,954 (86.73) -0.02 Child characteristics Child age at EDI, mean (SD) 5.64 (0.31) 5.65 (0.30) 0.01 Gestational age, mean (SD) 38.14 (2.01) 38.50 (1.79) 0.16 Female 667 (47.37) 1,092 (48.47) 0.01 English as a second language 61 (4.33) 205 (9.10) 0.54    125  Supplementary Table A 4. Demographic characteristics of mother-child dyads among mothers who used antidepressants pre-conception and continued versus discontinued antidepressants treatment during pregnancy. Antidepressant use pre-conception was defined as any prescription dispensation in the 90 – 365 days before pregnancy.    Antidepressant continuers vs. discontinuers among women with antidepressant use pre-conception.   Antidepressant discontinuation (n=4,647) Antidepressant continuation (n=2,660) Standardized difference Maternal characteristics  Maternal age at delivery, mean (SD) 29.57 (5.61) 31.05 (5.43) 0.38 Multiparous 2,565 (55.20) 1,630 (61.28) 0.11 Lowest neighborhood income quintile 1,050 (22.60)  562 (21.13)  -0.07  Mood/anxiety disorders diagnosis during pregnancy 841 (18.10)  1,682 (63.23)  0.98  Other psychotropic or antiepileptic medications during pregnancy 378 (8.13)  584 (21.95)  0.86      Anti-depressant exposure during pregnancy NA   SSRIs only  1,735 (65.23)  Non-SSRIs only  641 (24.10)  Both SSRI and non-SSRIs  284 (10.68)      2nd – 3rd trimester  2,384 (89.62)  Child characteristics Child age at EDI, mean (SD) 5.66 (0.30) 5.65 (0.30) -0.07 Gestational age, mean (SD) 38.81 (1.93) 38.29 (1.94) -0.39 Female 2,277 (49.00) 1,273 (47.86) -0.02 English as a second language 578 (12.44) 152 (5.71) -0.73 126  Supplementary Table A 5. Associations between prenatal antidepressant exposure and child development by antidepressant drug class, comparing exposed children matched to unexposed children using a high dimensional propensity score.    SSRIs only Non-SSRIs only Both SSRIs and non-SSRIs  OR (95% CI)a OR (95% CI)a OR (95% CI)a Physical well-being    Readiness 0.90 (0.69 - 1.19) 1.58 (1.05 - 2.37) 2.25 (1.04 - 4.87) Independence 1.07 (0.91 - 1.26) 1.35 (1.05 - 1.73) 1.23 (0.74 - 2.04) Motor skills 1.01 (0.89 - 1.15) 1.16 (0.94 - 1.43) 1.24 (0.81 - 1.88) Social competence    Competence 1.02 (0.86 - 1.21) 1.34 (1.01 - 1.77) 1.37 (0.80 - 2.33) Learning  0.99 (0.83 - 1.20) 1.27 (0.95 - 1.70) 1.00 (0.56 - 1.77) Exploring 1.53 (1.06 - 2.21) 1.65 (0.95 - 2.88) 2.95 (0.88 - 9.85) Respect 0.95 (0.76 - 1.19) 1.32 (0.94 - 1.87) 1.34 (0.72 - 2.49) Emotional maturity    Anxious 1.25 (0.92 - 1.70) 1.33 (0.80 - 2.20) 1.70 (0.70 - 4.09) Aggressive 0.87 (0.72 - 1.07) 1.51 (1.11 - 2.06) 0.83 (0.45 - 1.52) Hyperactive 0.86 (0.73 - 1.02) 1.19 (0.92 - 1.55) 0.98 (0.59 - 1.63) Language/Cognitive development    Basic literacy 0.95 (0.79 - 1.14) 1.10 (0.82 - 1.48) 1.34 (0.76 - 2.36) Interest 1.05 (0.88 - 1.25) 1.01 (0.74 - 1.37) 1.15 (0.66 - 2.03) Advanced literacy 1.05 (0.91 - 1.21) 1.17 (0.93 - 1.47) 1.05 (0.65 - 1.68) Basic numeracy 1.09 (0.93 - 1.28) 1.09 (0.83 - 1.43) 0.87 (0.50 - 1.50) Communication skills 0.98 (0.86 - 1.11) 1.14 (0.92 - 1.41) 1.40 (0.93 - 2.09) Abbreviations: SSRI, Selective Serotonin Reuptake Inhibitor; OR, Odds Ratio; CI, Confidence Interval aAdjusted for exposure to other psychotropic or antiepileptic medications during pregnancy.    127  Supplementary Table A 6. Associations between prenatal antidepressant exposure and child development by drug class, comparing children whose mothers had antidepressant dispensations in the 90 – 365 days prior to conception and continued versus discontinued treatment during pregnancy.    SSRIs only Non-SSRIs only Both SSRIs and non-SSRIs  OR (95% CI)a OR (95% CI)a OR (95% CI)a Physical well-being    Readiness 0.96 (0.72 - 1.27) 1.27 (0.87 - 1.86) 1.66 (1.02 - 2.68) Independence 1.10 (0.94 - 1.30) 1.30 (1.04 - 1.63) 1.30 (0.93 - 1.80) Motor skills 1.08 (0.95 - 1.23) 1.04 (0.86 - 1.25) 1.33 (1.01 - 1.74) Social competence    Competence 0.95 (0.80 - 1.13) 1.32 (1.04 - 1.67) 1.22 (0.87 - 1.71) Learning  0.98 (0.81 - 1.18) 1.07 (0.82 - 1.40) 0.97 (0.65 - 1.44) Exploring 1.07 (0.73 - 1.57) 1.64 (1.01 - 2.65) 2.03 (1.08 - 3.79) Respect 0.84 (0.67 - 1.05) 1.08 (0.79 - 1.46) 1.29 (0.85 - 1.95) Emotional maturity    Anxious 1.24 (0.92 - 1.67) 1.58 (1.06 - 2.36) 2.03 (1.19 - 3.44) Aggressive 0.78 (0.63 - 0.95) 1.01 (0.77 - 1.33) 0.89 (0.59 - 1.33) Hyperactive 0.82 (0.69 - 0.97) 1.12 (0.89 - 1.41) 0.90 (0.64 - 1.27) Language/Cognitive development    Basic literacy 0.99 (0.82 - 1.20) 1.02 (0.78 - 1.34) 1.14 (0.79 - 1.66) Interest 1.04 (0.86 - 1.25) 0.98 (0.75 - 1.29) 1.02 (0.70 - 1.50) Advanced literacy 0.93 (0.80 - 1.08) 1.09 (0.88 - 1.35) 1.03 (0.75 - 1.39) Basic numeracy 1.03 (0.87 - 1.21) 1.00 (0.78 - 1.28) 1.19 (0.85 - 1.67) Communication skills 1.06 (0.92 - 1.21) 0.96 (0.79 - 1.17) 1.23 (0.93 - 1.63) Abbreviations: SSRI, Selective Serotonin Reuptake Inhibitor; OR, Odds Ratio; CI, Confidence Interval; ESL, English as a Second Language aAdjusted for maternal age at birth, neighborhood income quintile, parity, maternal mood or anxiety diagnoses during pregnancy, other psychotropic or antiepileptic medications dispensed during pregnancy, wave of EDI collection, gestational age at birth, child age at EDI, child sex, and child ESL status. 128  Supplementary Table A 7. Primer sequences used for pyrosequencing of the CALCB locus. Primer type Sequence Forward GAGAGGTAGATAGGTAGATGGAGTAAAT Reverse-biotin TATCCTCTCCCAAACTCTCCAACTTC Sequencing TGGAGATTAGAGTTTTAGGGTA Sequencing TGGAGATTAGAGTTTTAGGGTA Sequencing AATTTAAAGTGGTTAGGATTA   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0379698/manifest

Comment

Related Items