The courses of maternal and paternal depressive and anxiety symptoms during the prenatal period in the FinnBrain Birth Cohort study

Maternal prenatal symptoms of depression and anxiety have been suggested to impose differential effects on later offspring development, depending on their characteristics, such as timing, intensity and persistence. Paternal symptoms have been less investigated. While knowledge on these trajectory characteristics is essential for improved comprehension of prenatal stress, prospective studies including both expecting parents have been scarce. We aim at identifying and comparing the trajectories of prenatal depressive and anxiety symptoms in both parents in a pregnancy cohort design. The sample included 3202 mothers and 2076 fathers who were recruited to the FinnBrain Birth Cohort study (www.finnbrain.fi). Depressive symptoms were assessed using the Edinburgh Postnatal Depression Scale (EPDS) and general anxiety by the anxiety scale of the Symptom Checklist -90 (SCL-90) repeatedly at 14, 24, and 34 gestational weeks. Five differential depressive and four anxiety symptom trajectories were identified across pregnancy both in mothers and in fathers. The trajectories of consistently low depressive or anxiety symptoms were associated with higher educational level in both parents, and with nulliparity and non-smoking during pregnancy in mothers. Parents with consistently high or increasing levels of symptoms had more often prenatal SSRI medication. The congruences between elevated depressive and anxiety symptoms at any point in pregnancy, as well as parental trajectories within families were low. However, in this population-based sample, the self-reported symptom levels of both parents were generally very low. Variance in timing and persistence of parent-reported prenatal depressive and anxiety symptoms is potentially important, while symptom trajectories are very similar in mothers and fathers. These differential symptom trajectories and the significance of their correlates should be acknowledged when studying prenatal stress exposures and the related outcomes in children.


Introduction
Maternal psychiatric symptoms during pregnancy reportedly have diverse and often adverse influences on child development, and ultimately on offspring adult health [1][2][3][4][5]. Paternal symptoms during the pre-and postnatal period and their associations with offspring development have been investigated to a lesser extent [6][7][8]. Depression and anxiety are the most common psychiatric symptoms during the prenatal and early postnatal periods both in women as well as men. Women commonly present with more pronounced depressive and anxiety symptomatology than men, also during pregnancy [9,10]. Clinically significant prenatal depressive symptoms affect approximately 10-20% of mothers and 2-13% of fathers [8,[10][11][12][13][14][15][16]. The prevalence of prenatal anxiety symptoms in different studies varies from 9.5% to 29% in mothers and from 8% to 14% in fathers [9,10,15,17,18]. The variation in prevalence estimates is partly explained by the timing of the measurement and the method of assessment. Many background factors, such as history of previous psychiatric disorders, low socioeconomic status and pregnancy-related somatic risk factors increase the risk for prenatal depression or anxiety [19,20]. Comorbidity between depression and anxiety has shown to be high also during pregnancy [15,19,21]. However, depression and anxiety are also partially distinct conditions, supported by the notions of differential symptom course trajectories [22], clinical prognostic factors [22,23] and pathophysiology [24]. Furthermore, previous evidence suggests that prenatal anxiety and depressive symptoms may also have differences in their biological correlates [25] as well as distinct effects on fetal development [2,26]. Therefore, parental anxiety and depression should be studied independently, especially in the context of prenatal fetal exposures.
Previous reports on the longitudinal course of parental prenatal depressive and anxiety symptoms comprising both parents have been scarce and inconsistent. In several studies, depressive symptoms have shown to decline throughout the prenatal period and onwards, both in mothers and in fathers [9,10] However, one recent study reported highest levels of depressive symptoms during the first and third trimesters, in mothers [27]. Similarly, prenatal anxiety symptoms have shown to follow a U-shaped pattern with highest levels of symptoms during the first and third trimesters in both parents [9,10,27]. On the other hand, some studies have reported the peak in paternal anxiety symptoms to occur at mid-pregnancy, where after anxiety would decrease steadily and across the postpartum period [28,29]. As a result of these discrepancies, it has been suggested that the timing and duration of depressive and anxiety symptoms in the prenatal period do not follow a uniform course, implying considerable heterogeneity in symptoms between individuals [30,31]. This has important implications for the concept of prenatal stress exposure and for studies investigating offspring outcomes related to parental symptoms. By studying symptom trajectories, more detailed information on the individual differences in timing, intensity and consistency of parental symptoms could be identified.
Parental mental health trajectories have previously been studied mainly during the transition from the third trimester of pregnancy to postpartum or from the perinatal period into childhood years [32][33][34][35][36]. Studies based on general population samples concur in that most parents report only minimal or mild symptoms across the prenatal period and postpartum [33][34][35][36][37]. In addition, most studies have identified a group of continuously high depressive or anxiety symptom levels [32][33][34][35][36][37]. These studies have mainly included only one prenatal symptom assessment, usually during the third trimester [8,32,[34][35][36]. However, the importance of investigating the courses of depressive and anxiety symptoms across trimesters has been highlighted in several reports on the negative effects of prenatal psychiatric symptoms on child development and well-being [5,38]. Identifying differential courses of symptoms during pregnancy may provide a basis for a more detailed investigation of diverse offspring outcomes, related to variation in timing and chronicity of exposure to prenatal psychological distress.
Despite the strong evidence that fathers have a crucial role in child well-being [39], the research on parental mental health trajectories has mainly focused on mothers. A few studies have suggested that elevated paternal depressive or anxiety symptoms during transition from the third pregnancy trimester to the postpartum period increase the risk for later problems in parenting [26] and child socio-emotional development [7,8]. The mechanisms behind the possible effects of maternal and paternal psychiatric symptoms on child's outcome are at least partially different [8,40,41]. Maternal depression or anxiety during pregnancy has suggested to affect the child mainly through different intrauterine processes, including HPA-axis activity, functioning of the placenta, changes in the immunological milieu, or nutritional supply [38,42,43]. Instead, the effects of paternal psychiatric symptoms on the child may either be mediated by the mother experiencing higher psychosocial distress due to the father's psychological distress [40,41] or by potential epigenetic effects of in paternal sperm, resulting in DNA methylation in the offspring [44,45]. Moreover, direct genetic effects are among the potential mechanisms to explain the effects of prenatal distress on child outcomes, for both parents, as child characteristics and prenatal distress experienced by the parents may indicate a shared underlying genotype [46,47]. Overall, very few studies focus on paternal and maternal prenatal psychiatric symptoms simultaneously within the same study. In addition, as far as we know, there are no studies on the trajectories of paternal or spousal psychiatric symptoms covering most of the prenatal period by repeated prospective measurements.
The aim of this study was to analyze the trajectories of prenatal depressive and anxiety symptoms and the associations of these trajectories with sociodemographic factors in both mothers and fathers in a population-based pregnancy cohort. In addition, the aim was to evaluate the congruence of the anxiety and depression trajectories and the congruence between paternal and maternal symptom trajectories inside the families. To our knowledge, this is the first study evaluating the symptom trajectories across trimesters in both mothers and fathers. The study is part of the large FinnBrain Birth Cohort Study that prospectively investigates the long-term effects of early life stress exposures on e.g. brain development.

Sample
Recruitment for the FinnBrain Birth Cohort Study took place at maternal welfare clinics between December 2011 and April 2015 in the South-Western Hospital District and the Åland Islands in Finland. The study subjects gave their written informed consent. the Ethics Committee of the Hospital District of Southwest Finland approved the study protocol 14 th June 2011. The study population (N = 3808 families) comprises of women that attended the free-ofcharge ultrasound at the gestational week 12, their children-to-be-born and fathers of the children/partners of the mothers [48]. Verified pregnancy and sufficient knowledge of either Finnish or Swedish (the official languages in Finland) were required for participation. Of those informed about the study (N = 5790), a total of N = 3808 mothers and N = 2623 fathers or other partners of the mothers decided to participate. In this study, only reported fathers of the children participating were included, and other partners of mothers in the study were excluded. Mothers (N = 3202) and fathers of the children (N = 2076) responding to at least one symptom questionnaire during pregnancy were included. via postal mail. Prenatal depressive symptoms were evaluated in both parents using the Edinburgh Postnatal Depression Scale (EPDS, 34) at gwks 14, 24 and 34. EPDS is a self-report questionnaire consisting of 10 items. The parent is asked whether she or he has experienced the proposed symptoms during the last two weeks using a 4-point Likert scale. The EPDS has been studied extensively and it is thought to be a valid screen for both pre-and postnatal depression [49][50][51][52][53], also in fathers [54,55]. Prenatal anxiety of both parents was evaluated using the anxiety scale of Symptom Checklist 90 (SCL-90) at gwks 14, 24 and 34. The anxiety subscale of SCL-90 is a reliable and valid measure of anxiety symptoms in both clinical and research settings [56] and consists of 10 items rated from 0 to 5.
The sum scores of the EPDS and SCL-90 anxiety scales were used as continuous variables in the data analyses and they demonstrated good internal consistency (Cronbach's α's ranging from .82 to .84 in mothers and .79 to .81 in fathers for EPDS and .83 to .85 in mothers and .83 to .86 in fathers for SCL-90). The background variables i.e. relationship status, educational level, age, parity, and paternal use of serotonin reuptake inhibitors (SSRIs) during the index pregnancy were drawn from the questionnaires filled in at gwk 14. Educational level was categorized into four classes: secondary school (less than 9 years of schooling), high school/vocational education, university/polytechnics degree, or higher (doctoral/licenciate). No family income variable was available. Therefore, only variables describing maternal and paternal education were used as indicators of socioeconomic status. Other background variables i.e. maternal prenatal smoking, the household variable (living together) and maternal use of serotonin reuptake inhibitors (SSRIs) during the index pregnancy, were drawn from the questionnaires filled in at 34 gwks. In all, 80% of the mothers using SSRI at 34 gwks had already used SSRImedication at 14 gwks. Father's use of SSRI was measured only at 14 gwks.

Statistical analyses
The associations between the means of depressive and anxiety symptom scales at each time point in mothers and in fathers were investigated with zero-order Spearman correlations. As measures of effect sizes, Cohen's d and partial 2 were used. Structural Equation Modeling in Mplus 6, and latent growth mixture modeling, were conducted to examine the trajectories of maternal and paternal EPDS and SCL-90 symptoms [60]. Latent growth mixture modeling (LGMM) is a method for identifying multiple unobserved sub-populations (latent classes), describing the longitudinal change within each latent class and examining differences in the change among the latent classes [61]. The numbers of latent classes consisting of mothers' and fathers' EPDS and SCL-90 scores were determined by increasing the numbers of classes in separate analyses and observing the change in indices. The indices that were used in decisionmaking were Bayesian Information Criterion (BIC), Akaike Information Criterion (AIC) [60,62], Entropy value (with values closer to 1.0 indicating higher confidence of classification; [63] and posterior probabilities of class membership. Furthermore, the theoretical and clinical interpretability of the class solutions were used as a help in choosing the best model, a step considered necessary in the decision-making in LGMM [64]. Missing values were treated with the Expectation Maximization (EM) method in Mplus. The reliabilities of the final solutions were checked by replicating the analyses with non-imputed datasets (consisting complete data from 2157 mothers and 1475 fathers). The resultant model solutions (with and without EM) were alike. Next, the associations between trajectories and parental age, education, parity and SSRI medication use were investigated using ANOVA (age) and cross-tabulation with χ 2 test (education, parity and SSRIs).

Demographics and attrition
Parents with missing questionnaire data over the course of pregnancy had lower educational levels (χ 2 (3) = 38.76, p < .001) and were more often single or divorced (χ 2 (1) = 22.26, p < .001) than those who filled in the questionnaires at each of the three assessment points. No differences were seen in age, parity or in the level of depressive or anxiety symptoms at 14 gwks between those parents who filled in all questionnaires and the parents with missing questionnaire data. In the study sample, the mothers had higher educational levels and they were younger than the fathers. In addition, mothers reported more depressive and anxiety symptoms at each assessment point compared to fathers. Descriptive statistics and comparisons between mothers and fathers are displayed in Table 1. Maternal depressive (continuous variables) and anxiety symptoms (continuous variables) were intercorrelated at all assessment points during pregnancy (r = .58-.62, p < .001). Similarly, paternal depressive and anxiety symptoms were intercorrelated at all assessment points (r = .58-.59, p < .001).

The trajectories of maternal and paternal prenatal depressive and anxiety symptoms
Different class solutions were examined to obtain the best fit of the model for maternal and paternal prenatal depressive and anxiety symptoms. Regarding maternal depressive symptoms the four-and five-class-solutions resembled each other in terms of statistical indices, but the five-class solution had a better fit with the assumption of trajectories and it was thus chosen (Fig 1). The resultant fit indices are presented in Table 2 In fathers, the five-group solution yielded similar results as with mothers (Fig 2). The resultant fit indices are presented in Table 3. A large group of fathers with consistently low depressive symptom scores (N = 1475, 71%, M = 2.24; Estimate = -0.30, p < .001) was identified, and a smaller group with consistently high symptom scores (N = 66, 3%, M = 13.18; Estimate = -1.43, p < .001). Three other groups were divided into those with moderate and stable (N = 307, 15%, M = 5.03; Estimate = 0.90, p < .001), moderate and decreasing (N = 185; 9%; M = 8.61; Estimate = -0.29, p < .001), and moderate and increasing symptom scores (N = 42; 2%; M = 7.50; Estimate = 3.85, p < .001). Although this solution did not reach as optimal a fit based on statistical indices as the other solutions, it was chosen because it was replicated as completely similar with the non-imputed data. For the maternal anxiety symptom trajectories, the four-class solution was chosen, as the five-class solution did not provide any additional information on the course of symptoms ( Fig  3). The resultant fit indices are presented in Table 4 Similarly, to mothers, the five-class solution for the fathers did not provide any additional information on the course of anxiety symptoms, and the four-class solution was thus chosen (Fig 4). The resultant fit indices are presented in Table 5. One large group emerged with a con-

The associations between symptom trajectories and sociodemographic factors
The trajectories of maternal prenatal depressive symptoms differed in parity, household (living together) and smoking during pregnancy, and in educational level as well as in the use of SSRI   The trajectory groups of prenatal anxiety symptoms differed in parity, household (living together) and smoking during pregnancy in mothers and in age, educational level, and in the use of SSRI medication at 34 gestational weeks in both parents. Mothers with consistently low levels of anxiety were, on average, older (F (3, 3197) = 11.28, p < .001), were living more often with the father of their child (χ 2 (3) = 30.22, p < .001) and smoked less often during the third pregnancy trimester (χ 2 (3) = 58.9, p = < .001) than the mothers in all the other groups. In  addition, mothers in the groups with increasing or consistently low symptoms had higher educational levels than mothers in both the groups of consistently high as well as moderate and decreasing anxiety (χ 2 (9) = 52.58, p < .001). Moreover, mothers with consistently high symptom levels were more often multiparous (χ 2 (3) = 12.24, p < .01) compared to the mothers with increasing and decreasing symptoms. Fathers with decreasing anxiety symptoms were younger than the fathers in the other groups (F (3, 2066) = 5.02, p < .01). In addition, the groups with either consistently high, increasing or decreasing symptoms comprised of more fathers with lower education than the group with consistently low symptoms (χ 2 (9) = 45.60, p < .001). Furthermore, higher numbers of both mothers (26%, χ 2 (3) = 85.3, p < .001) and fathers (21%, χ 2 (3) = 35.1, p < .001) using SSRIs during pregnancy were seen in the group of consistently high symptoms compared to parents in all the other trajectory groups of anxiety. Descriptive values of the associations between trajectories and all background variables are shown in Table 6.

Congruence between depressive and anxiety symptom trajectories and between maternal and paternal symptom trajectories
The congruences between the trajectories of depressive and anxiety symptoms was studied using cross-tabulation, and the results are presented in Table 7. Most mothers and fathers with consistently low depressive symptoms belonged to the group of consistently low anxiety symptoms. Furthermore, the congruence between increasing depressive and anxiety symptom trajectories was relatively high. The most incongruent distributions found between the groups of consistently high levels of depressive and anxiety symptoms, both in mothers as well as in fathers.  The overlap of maternal and paternal trajectories within families was examined using cross tabulation (Tables 8 & 9). Notable congruence between mothers and fathers was found only between the consistently low symptom trajectories, i.e. the most prevalent categories in this study population. In families where mothers reported consistently low levels of depressive or anxiety symptoms throughout the pregnancy, the fathers also reported low levels of depressive or anxiety symptoms throughout the pregnancy. The congruence between mothers and fathers in the other depressive or anxiety symptom trajectory groups was generally low.

Discussion
In the present study, we identified five different trajectories describing the course of prenatal depressive symptoms both in mothers and in fathers. Regarding the prenatal anxiety symptoms, four different trajectories were identified, equally in both parents. The five trajectories of prenatal depressive symptoms included the following groups: consistently low levels of symptoms, consistently high levels of symptoms, moderate levels of symptoms, moderate and decreasing levels of symptoms and moderate and increasing levels of symptoms. Regarding prenatal anxiety symptoms, the four trajectories were: consistently low levels of symptoms, consistently high levels of symptoms, high and decreasing levels of symptoms and moderate and increasing levels of symptoms.
Our findings are in line with previous trajectory studies indicating 3 to 6 prenatal parental psychiatric symptom trajectories, including at least one trajectory of chronically high symptoms and one trajectory of consistently low symptom levels during the transition from  pregnancy to early childhood [32,34,37]. Our findings suggest that potentially important variance exists in the timing and persistence of parent-reported depressive and anxiety symptoms, already during the prenatal period. This should be considered when prenatal exposures are studied, and when child outcomes related to both maternal and paternal symptoms are investigated. Our findings also indicate that higher educational level and older age may be associated with consistently low levels of prenatal psychiatric symptoms in both parents. In mothers, consistently low symptom levels were related to primiparity, not smoking at the 3 rd trimester, and living together with the father of the child. Consistently elevated symptom levels were related to the use of SSRI medication in both parents. The congruence between the parents' trajectories inside families as well as between the trajectories of depressive and anxiety symptoms were relatively low suggesting that the trajectories of depressive and anxiety symptoms have partially independent courses within child-expecting families. However, the congruence in consistently low symptom levels between mothers and fathers within families, as well as between anxiety Parental prenatal anxiety and depressive symptoms and depressive symptoms, was notable. This could be explained by the preponderance of the low symptom category in this generally healthy population-based cohort sample. Approximately 85-90% of the parents in the FinnBrain Cohort reported low levels of depressive or anxiety symptoms at any trimester. This is in line with several previous studies showing that most parents in the general population report only minimal or mild symptoms during pregnancy and the postpartum period [32,33,35,36]. Mothers with low levels of depressive or anxiety symptoms were more often living with the father of the child, smoked less frequently, and were expecting their first child more often than mothers with high or changing levels of depressive or anxiety symptoms. While it has been recognized earlier that parental psychiatric symptoms often co-occur with other risk factors, our notion of several known sociodemographic risk factors accumulating especially in certain symptom trajectories deserves further attention. Thus, it may be that not only the chronicity of symptoms but also the affiliated risk factors play an important role when the offspring outcomes are considered.
Only 1-2% of the parents in our sample experienced consistently high levels of depressive or anxiety symptoms throughout the pregnancy. In line with previous studies, these parents were also more likely to have lower education and more previous children, suggesting that, among other risk factors, socioeconomic burden and the demands of everyday life are associated with the emergence of parental psychiatric symptoms not only during the postnatal but Parental prenatal anxiety and depressive symptoms also during the prenatal period [65,66]. These parents had also more often SSRI medication, which might reflect the primary severity of their symptoms. Clinically, it would be essential to identify these parents of consistently high levels of depressive or anxiety symptoms, as parents with chronically pronounced psychiatric symptoms reportedly form an evident risk for the offspring as well as for parenting problems and family adversity [5,67,68]. In addition, chronically elevated symptom levels during pregnancy may reflect prolonged (i.e. covering both prenatal and postnatal time periods) psychiatric problems and, based on previous research, strongly predict postpartum depression [69][70][71][72]. Further, 24% of mothers and 16% of fathers in our study reported moderate levels (i.e. differential trajectory to those with consistently low symptom levels) of depressive symptoms throughout pregnancy. In addition to identifying the chronically highly depressed and anxious parents, it would also be important to recognize the parents reporting subclinical levels of depressive or anxiety symptoms, especially if they appear as chronic. Recently, it has been emphasized that also parental subclinical, especially chronic, psychiatric symptoms might have negative associations with future child well-being [73][74]. Furthermore, a group of parents reporting decreasing symptom levels was identified, that presented high symptom scores during the first trimester but very low levels of depressive or anxiety symptoms during the third trimester. This decrease in symptoms may reflect a normative psychological adaptation to pregnancy and parenthood, as the first trimester often includes worries and stress due to the uncertainty of the pregnancy, whereas the second trimester is seen as a period of stability, and the third trimester as a period of positive expectations relating to the upcoming birth and parenthood [75]. By contrast, the groups with increasing levels of depressive or anxiety symptoms, including parents with a low level of symptoms in the first trimester but moderate or high levels of symptoms in the third trimester, Parental prenatal anxiety and depressive symptoms may reflect a specific psychiatric symptomatology relating to the upcoming birth, parenthood or concerns about the health of the child. High levels of symptoms before birth are reported to have a negative effect on the child through a prenatal programming effect [76][77][78], or through postnatal factors including more vulnerable parenting behavior transferring from the antenatal period [76,77,79]. Still, only few studies, to date, have considered the longitudinal course of parental symptoms as a factor in pre-and postnatal programming. Further research is needed to explore the mechanisms underlying the different trajectories and to identify how variation in timing and chronicity affect specified child and parent outcomes. Overall, the mothers in our sample reported more depressive and anxiety symptoms during pregnancy than the fathers, which is in line with previous studies reporting that in general, depressive and anxiety symptoms are more prevalent among females than males both in the prenatal period and overall [9,10]. On the other hand, within the groups of elevated anxiety symptom levels, mothers had lower symptom scores than fathers, suggesting that fathers experiencing psychological distress may represent a relatively severely anxious population. Interestingly, the trajectory solutions of depressive and anxiety symptoms were very similar between mothers and fathers, suggesting that the differences in trajectories are not sex-specific, but rather may depend on other factors that may be more related to the pathophysiologies and the courses of depression/anxiety per se [10,32,36].
The findings on the congruence of within-person depressive and anxiety symptom trajectories were only partly supportive of previous reports. Congruence was high between the trajectories of consistently low levels of depressive and anxiety symptoms (the same person is likely to report no symptoms at all), and moderate between the trajectories of increasing symptoms, both in mothers and in fathers. However, the congruence between depressive and anxiety symptoms was low in the groups of both consistently high symptoms and decreasing symptoms suggesting that these trajectories develop independently from one another, in both parents. This supports the previously presented notion that in the context of prenatal and early life stress, parental depressive and anxiety symptoms should be treated independently as stress exposures [47]. The congruence of maternal and paternal depressive and anxiety trajectories within families was also low, suggesting that the trajectories of depressive and anxiety symptoms during the prenatal period develop relatively independently inside the family. However, the congruence between maternal and paternal trajectories was high in the most prevalent categories of consistently low depressive and anxiety symptom levels, which suggests that in the majority of families, neither of the parents experience marked psychological distress. The differential trajectories within families, and the interactions between maternal and paternal symptom courses should be taken into consideration, when related offspring outcomes are investigated.
A few limitations should be pointed out in the present study. First, sample sizes in the groups of high depressive and/or anxiety symptoms levels were relatively low. Drop-outs may partially explain this small number of parents with very high symptoms levels in the study population. On the other hand, imputation in Mplus was used to diminish this bias. Another limitation was that the parents' symptoms were assessed using self-report questionnaires as opposed to diagnostic interviews that may have yielded more detailed information on mood and anxiety disorder diagnoses. However, the questionnaires employed in this study are widely used and reliable, and have been shown to be robustly in line with clinical assessments [50][51][52][53]56]. In the current study, we were not able to examine the mechanisms behind each trajectory, but this remains to be investigated in the follow-up reports after this description of typical symptom courses and their within-family occurrence. Additionally, the necessary next step should be to study the identified trajectories, and more specifically, the significance of timing, intensity and persistence of both maternal and paternal depressive and anxiety symptoms across the antenatal period in relation to offspring well-being and development. Both the independent maternal and paternal effects as well as the interactive effects of spousal symptom levels on child developmental outcomes need further attention.
To the best of our knowledge, this is the first study to evaluate the trajectories of parental depressive and anxiety symptoms during pregnancy in both mothers and in fathers as well as within families, in a large population-based birth cohort setting. Our findings indicate that the longitudinal development of parental depressive and anxiety symptoms during the prenatal period may represent processes independent of parental sex, but more characteristic of the specific psychiatric symptom type. As depressive and anxiety symptoms do not always overlap, they both should be independently targeted in research as well as in clinical work. The simultaneous assessment of both parents during gestation seems warranted also in future research. Importantly, the trajectories found in the present study provide a basis for investigating possible parental and offspring follow-up outcomes in more detail and in a more stratified manner, and for evaluating long-term prognoses of prenatal exposure to either maternal and/or paternal symptoms of depression and anxiety.