Trajectories of perinatal post-traumatic stress disorder scores in association with child’s behavior at 12 months

doi:10.21203/rs.3.rs-4714574/v1

Download PDF

Research Article

Trajectories of perinatal post-traumatic stress disorder scores in association with child’s behavior at 12 months

https://doi.org/10.21203/rs.3.rs-4714574/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Perinatal mental health is fundamental to a healthy society. The aim of this study was to describe the trajectories of women’s posttraumatic stress disorder (PTSD) symptoms during the perinatal period to assess their association with child behavior problems at 12 months. We designed an observational longitudinal study. Women were recruited through social media posting during the Coronavirus Disease 2019 (COVID-19) pandemic Italian national lockdown from April 8 to May 4, 2020, and contacted again at 6 and 12 months after the expected delivery date, collecting PTSD scores each time. Child behaviors were reported at 12 months postpartum. Inclusion criteria were residence in Italy, age over 18 years, and fluency in Italian. A total of 327 mother-child dyads were eligible for inclusion in the study. Clustering analysis suggested five groups of PTSD trajectories: a very low and stable (VL) group, 2 groups with decreasing PTSD symptoms over time (one high and decreasing (H-), one low and decreasing (L-)), and 2 groups with positive PTSD trajectories (one high and increasing (H+), one low and increasing (L+)). The H + and H- clusters had significantly higher risks (+ 58% and + 76% for H + and H-, respectively) for total child behavioral outcomes compared with the VL cluster, and higher risk for internalizing problems. Although many women had PTSD scores below the cut-off, we envision a significant risk for the children of mothers with elevated symptoms in pregnancy. Longitudinal modeling of perinatal PTSD symptoms is warranted for sensitive two-generation risk detection.

post-traumatic stress disorder

mental health

child development

maternal factors

perinatal

Mental health in the perinatal period is pivotal for the women’s health (Onoye et al., 2009; Seng et al., 2013), the caregiving (Radoš et al., 2020; Webb & Ayers, 2015), the child’s healthy development (Ayers et al., 2019; Erickson et al., 2019) and the society (Aizer et al., 2012). Particularly for the child, women (hereafter also referred to as “mother/maternal”) mental health in pregnancy hosts intergenerational cascading influences, including prenatal programming processes building the infant neurobiological systems (Glover, 2011; Glover et al., 2018) and the parental brain changes that contribute to shape the quality of postnatal caregiving functioning (Hoekzema et al., 2017; Swain et al., 2017). Whilst being so powerfully implicated in the psychobiology that sets for the offspring generation, women’s perinatal mental health is extremely sensitive to the quality of the social environment, with significant stressors, including psychosocial ones, placing the risk of a two-generation impact of maternal maladaptive adjustment and stress response to such exposure.

We focus this investigation on the impact of the pandemic Coronavirus Disease 2019 (COVID-19) outbreak experienced during pregnancy, as a massive psychosocial stress exposure, on the perinatal trajectories of post-traumatic stress disorder (PTSD) and these can differently and adversely impact child development, as is can be observed by describing early behavioral indexes.

Compared to depression and anxiety, PTSD is not commonly included in the conversation about perinatal mental health concerns (Moran Vozar et al., 2021), even if it may significantly undermine maternal and infant health (Van Sieleghem et al., 2022). Perinatal PTSD, mostly investigated as a result of traumatic childbirth, is associated with a higher risk of depression (Shahar et al., 2015), problems in the parent-infant relationship (Davies et al., 2008), and marital difficulties (Ayers et al., 2006), that may extend or impede delivery recovery (Dikmen-Yildiz et al., 2018), and further cascade into infants’ temperamental and behavioral problems (Van Sieleghem et al., 2022). PTSD in the perinatal period can be triggered not only by childbirth, but also by other traumatic or severely stressful events and environmental contexts occurring during pregnancy (Ayers, 2004; Durbano, 2013). Effects and consequences of the COVID-19 pandemic have been already described as a potential source of traumatic stress associated with an increase in the mental health burden for the general population (Penninx et al., 2022). Throughout 2020 and much of 2021, the pandemic produced very unexpected and unwelcomed changes in the individual functioning and the individual-society relationship, especially for perinatal women. Indeed, the pandemic was initially characterized by fear of attending public hospitals and for initial suspected risk of Severe Acute Respiratory Syndrome COronaVirus 2 (SARS-CoV-2) vertical transmission while carrying a pregnancy, and by continuous exposure to a wide range of tragic and stressful and inconsistent communication from governments and public press unlikely to contain fear and anxiety and to progressively restore a sense of personal safety in carrying daily activities. In this context, the COVID-19 pandemic particularly produced documented changes to the perinatal care (Hendrix et al., 2022), introducing an imposed disruption in social support access, likely inducting vulnerability and isolation during pregnancy, and delivery and the impossibility to share pregnancy-related life milestones (i.e., routine visits, bad news communication, labor). While facing such a scenario, on the individual level women’s health undergoes significant and dynamic changes to interesting neurophysiological and psychological systems supporting fetal growth, the transition to parenthood, and the emerging caregiving system (Grobman et al., 2024; McCormack et al., 2023; Sacchi et al., 2021). The alarming environment, the stressful life conditions, and the uncertainty of perinatal care management likely interfered with the pregnancy-related hormonal, emotional, and behavioral changes and with the mother-fetal psychophysiological exchanges, as well as the sense of chronic fatigue and the restricted postpartum social opportunities might have impacted the quality of parenting’s emotional experience and behavioral practices.

As a result, perinatal mental health secondary to pandemic is likely to be endangered. Indeed, the perinatal period is challenging for individual mental health as both the specific changes of pregnancy and the transition to parenthood may exacerbate new psychological distress, and some prior mental health difficulties may see worsening symptomatology. The same can be described, even in the general population, as a response to the dynamics of pandemic COVID-19. Therefore, it is critical to longitudinally address the relationship between the pandemic, women's stress-related mental health response, and child development in the perinatal period in order to identify interindividual variability in maternal stress response during the transition to parenthood and the resulting different risk to offspring development. This study aims to evaluate the association between women’s PTSD symptoms from pregnancy through 12 months postpartum and children's emotional-behavioral growth in early childhood. Given the longitudinal nature of the PTSD assessment, this study allowed the dynamic relationship between trajectories of women's PTSD symptoms and child behavioral outcomes to be explored. Indeed, based on the hypothesis advocated in the literature about the relationship between perinatal PTSD and adverse developmental outcomes for the child (Cook et al., 2018; Garthus-Niegel et al., 2017), we explored how this relationship may change according to the temporal dynamics of PTSD in a perinatal period exposed to a long-lasting health and psychosocial emergency. The paper is organized as follows: in Methods section we present our study data, derived from an observational study (Sacchi et al., 2023), the metrics used and the statistical analysis plan, including clustering of PTSD trajectories and regression models. In the Results section we present the main results, while in the last part we discuss the relative strengths and weaknesses of our proposal, suggesting an appropriate comparison of the results with those in the existing literature.

Procedure

Participants were recruited as a convenience sample of pregnant women during the pandemic COVID-19 onset in spring 2020 (Fig. 1). We performed three subsequent Qualtrics-hosted online surveys during pregnancy (t0), 6 months (t1), and 12 months (t2) postpartum. In each survey, participants provided the written consent form and explicitly agreed to participate. This study was part of a longitudinal project on perinatal maternal-infant health secondary to the COVID-19 pandemic (Sacchi et al., 2023) conducted at the University of Padua (Italy). The Institutional Review Board of the University of Padova approved the research (08/04/2020, approval n. 3545). The enrollment and the first survey (t0) was diffused via social media posting, confidential data were collected, and subsequent surveys (t1, t2) were emailed to the individual participants that agreed to be followed-up. Data were collected from April 8th, to May 4th, 2020 in the first survey (t0), from December 12th, 2020 to May 8th, 2021 in the second survey (t1), and from April 29th, 2021 to December 28th, 2021 in the third survey (t2). For the first survey (t0) participants reported their mental health symptoms and psychological stress due to the pandemic (i.e., subjective pandemic psychological distress). For the first postpartum (t1) assessment, participants completed a survey about current mental health symptoms and pandemic psychological stress; social support, prenatal and postpartum exposure to COVID-19 related stressful life events and birth outcomes (i.e., child’s biological sex, birth weight, gestational age at delivery) were also included. At the 12 months (t2) assessment, in addition to current mental health symptoms, the survey considered measures for parenting and child emotional-behavioral development.

Measures

This section describes the considered measures (more details are contained in the Supplement Material S1).

Post-traumatic stress disorder (PTSD) symptoms at t0, t1, and t2

we used the PTSD checklist for DSM-5, PCL-5 (Weathers et al., 2013) composed of a 20-item self-report measure that assesses the PTSD symptoms based on DSM-5 criteria (5th ed.; DSM–5; American Psychiatric Association, 2013). A cutoff score of 31 is indicated for probable current PTSD (Weathers et al., 2013).

Pandemic Psychological Stress (PPS) at t0: self-reported psychological stress secondary to the COVID-19 pandemic in pregnancy was investigated through a set of 7 questions. We produced a PPS score by means of regression scores of an exploratory factor analysis (EFA) (see Supplement Material: S2, Table S1).

COVID-19 Stressful Events Exposure (SEE) during pregnancy

a short checklist of questions was administered regarding the direct exposure to COVID-19 major stressful events during pregnancy measured at t1. Complete information about the scale has been previously published (Sacchi et al., 2023).

COVID-19 Stressful Events Exposure (SEE) postpartum

a short checklist of questions was administered regarding the direct exposure to COVID-19 major stressful events during the first 6 months postpartum measured at t1. Complete information about the scale has been previously published (Sacchi et al., 2023).

Social Support at t1

perceived social support has been investigated through the Multidimensional Scale for Perceived Social Support, MSPSS (Zimet et al., 1990); higher scores indicate greater social support.

Child emotional-behavioral problems at t2: The Child Behavior CheckList, CBCL/1½-5, (Achenbach, 1999) is a gold-standard parent-report questionnaire to assess emotional behavioral problems in children aged 1 ½- 5 years. The CBCL/1½-5 allows the evaluation of children's problems summarizing them in: 1) internalizing scale as the sum of scores on emotional reactivity, anxious/depression, somatic complaints, and withdrawn syndrome scales; 2) externalizing scale as the sum of scores on attention problems and aggressive problems syndromes scales; 3) total as the sum of internalizing and externalizing scales.

Socio-demographic, pregnancy and delivery information: at t0, a series of socio-demographic information was collected, including the mother’s age, work, education and marital status, town of residence, family income, gestational week, previous pregnancy (y/n), planned pregnancy (y/n), difficulty conceiving (y/n), and miscarriages (y/n). At t1, a set of questions was asked about the experience of delivery, such as: i) the type of birth (non-assisted vaginal birth, assisted vaginal birth, planned cesarean birth, emergency cesarean birth; ii) whether childbirth was experienced in isolation from the baby-partner, because of the pandemic containment measures; the child’s biological sex and gestational age at delivery.

Data analysis

Data imputation

A limited amount of missing data for each considered variable was reported (Table S2). Missing data were imputed by a process based on multiple imputations by chained equations (MICE) employing a classification and regression trees (CART) procedure that consent to handle non-monotonic regressions between variables in order to account for a more accurate prediction. The MICE process was iterated for 30 times; results stability was evaluated by visual inspection of the estimate chains.

Descriptive Statistical analysis

Data were summarized by frequency for categorical variables, and median and interquartile range (IQR) for continuous variables. Given the non-normal distribution of the continuous data, Wilcoxon rank-sum tests were computed to compare the distribution across two strata; with more than two strata, Kruskal-Wallis tests were considered. Association between categorical variables were assessed by chi-squared or Fisher's exact test if expected frequencies were less than 10. Statistical significance was assumed at the 5% level. Statistical analysis was performed using R.

PTSD Clustering – EM Algorithm

In order to regroup mothers with common trajectories of PTSD scores over the three assessment steps, we build an ad-hoc Expectation-Maximization (EM) algorithm (Dempster et al., 1977). The optimal number of clusters was chosen considering an elbow method based on the first difference in the BIC score and the first local minimum. More details were implemented in the supplementary materials (see Supplement Material S3).

Propensity score methods

All mothers were enrolled as pregnant during the COVID-19 pandemic, but not all reported a COVID-19-related stress event during pregnancy (Table 1). We use a propensity score method to reconstruct the probability of having a stress event during pregnancy based on the confounders measured at t0. We calculate the stabilized weight for each subject using an inverse probability of treatment weighting (IPTW) method (Austin & Stuart, 2015). We estimated this probability by means of a logistic regression model in which the dependent variable is the presence of a SEE during pregnancy, while the covariates are characteristics measured at t0: mother’s age, educational level,

marital status, residence in red region, job status, family income, parity, gestational week at enrollment, PPS. The results of the logistic regression are shown in Table S3, while the estimated weights by the presence of COVID-19 SEE during pregnancy are shown in Fig. S1.

Regression models

Given the presence of over-dispersion on CBCL scores (Fig. 2), negative binomial regression models were employed to evaluate the association between CBCL internalizing, externalizing and total problems scores and PTSD cluster membership controlling for a series of confounders: mother’s age (in continuous form), educational level (master’s degree or higher, bachelor’s degree, high school or lower), marital status (live-in partner, single, married), residence in a red region (Italian regions with pandemic restrictions; yes, no), job status (employed, not employed), family income (low < 0-12k €, low-Medium (12k-25k] €, medium (25k-50k] €, high > 50k €), parity (yes, no), gestational week at t0 (in continuous form), type of birth (non-programmed C-Section, programmed C-Section, vaginal delivery), children’s biological sex, child’s gestational age at birth (in continuous form), presence of at least COVID-19 SEE during pregnancy at t0 (yes, no), presence of at least COVID-19 SEE postpartum (yes, no), total MSPSS score (in continuous form), and PPS score (in continuous form). Further, the model incorporates the stabilized weights estimated by the propensity score model, presented in the previous subsection. The presence of outliers was assessed by means of quantile-quantile diagram of Pearson residuals, dropping observations far from the theoretical quantile line. The results of the estimated negative binomial regression models are reported using Incidence Rate Ratios (IRRs), by an exponential transformation of regression coefficients, with relative 95% Confidence Intervals (CIs). All R code and data associated with the real data application are available at https://osf.io/x6tev/.

Mothers were aged between 22 and 46 years (median 33), predominantly with a high educational level (41% master’s degree or higher). Most of the women was married (62%) or in a relationship (34%). About a half lived in a red region and they were almost all actively employed (93%). The most reported income was medium (25k-50k, 60%). About a quarter (23%) had a previous pregnancy, while the median gestational age at enrollment was 27 weeks. The predominant type of birth was vaginal delivery (78%) with a median gestational age at birth of 40 weeks. The presence of COVID-19 SEE during pregnancy was of 14%; this percentage increased during the postpartum period (31%). A low perceived support (MSPSS) and a high pandemic psychological stress (PPS) were related to high values of PTSD at t0 (both p < 0.001; Table 1).

Table 1

Main characteristics of the sample, overall and by PTSD higher than 31 during pregnancy
Characteristic	Overall,	PTSD at t0 > 31		p-value²
	N = 327¹	No, N = 274¹	Yes, N = 53¹
Mother’s age (years)	33 (30, 36)	33 (30, 36)	32 (30, 35)	0.473
Educational level				0.114
Bachelor’s degree	85 (26%)	74 (27%)	11 (21%)
High school or lower	107 (33%)	83 (30%)	24 (45%)
Master’s degree or higher	135 (41%)	117 (43%)	18 (34%)
Marital status				0.751
Live-in partner	111 (34%)	93 (34%)	19 (36%)
Single	14 (4.3%)	13 (4.7%)	1 (1.9%)
Married	202 (62%)	169 (61%)	33 (62%)
Residence in red region [Yes]	167 (52%)	149 (54%)	21 (40%)	0.052
Job status [Employed]	304 (93%)	256 (93%)	48 (91%)	0.555
Family income				0.055
Low ≤ 12k €	14 (4.3%)	8 (2.9%)	6 (11%)
Low-Medium (12k-25k] €	71 (22%)	59 (22%)	12 (23%)
Medium (25k-50k] €	195 (60%)	165 (60%)	30 (57%)
High > 50k €	47 (14%)	42 (15%)	5 (9.4%)
Parity [Yes]	75 (23%)	59 (22%)	16 (30%)	0.211
Gestational week at enrollment (weeks)	27 (19, 33)	26 (18, 33)	29 (24, 33)	0.060
Type of birth				0.315
Non-programmed C-Section	43 (13%)	34 (12%)	9 (17%)
Programmed C-Section	36 (11%)	33 (12%)	3 (5.7%)
Vaginal delivery	248 (78%)	207 (75%)	41 (78%)
Children’s sex [Female]	158 (48%)	136 (50%)	21 (40%)	0.229
Child’s gestational week at birth (weeks)	40 (38, 41)	40 (38, 41)	40 (38, 40)	0.457
COVID-19 SEE pregnancy [Yes]	47 (14%)	139 (14%)	8 (15%)	0.833
COVID-19 SEE postpartum [Yes]	101 (31%)	87 (32%)	14 (26%)	0.517
Total MSPSS	4.92 (4.33, 5.58)	5.08 (4.50, 5.58)	4.42 (3.75, 4.92)	< 0.001
CBCL total problems	9 (5, 16)	9 (5, 16)	9 (6, 16)	0.417
CBCL externalizing problems	6 (3, 11)	6 (3, 11)	7 (4, 12)	0.441
CBCL internalizing problems	3 (1, 6)	3 (1, 6)	3 (2, 8)	0.341
PPS score	0.05 (-0.60, 0.66)	-0.08 (-0.69, 0.59)	0.50 (-0.06, 0.97)	< 0.001
¹ Median (IQR) or Frequency (%)
² Wilcoxon rank sum test; Fisher's exact test

We observed a weak correlation between PTSD scores (mainly at t0) and CBCL outcomes (Figure S2). The clustering procedure identified 5 clusters (Figure S3). On the basis of the PTSD score at t0 (very low, low vs high) and the visual inspection of the temporal trend of PTSD trajectories (stable, increasing, and decreasing), we interpreted and labeled the 5 clusters as follows (Fig. 2): very low-and-stable (VL); low-and-decreasing (L-); low-and-increasing (L+); high-and-decreasing (H-), and high-and-increasing (H+).

The main characteristics of the sample by PTSD cluster membership are reported in Table 2. Distribution of educational level differs by cluster membership (p < 0.001) where the clusters VL, L- and H- reported a higher educational level than H + and L+. We observed a modest difference in the CBCL scores distribution across the identified clusters, with the highest values reported by clusters H- and H+, and the lowest ones among those in the VL group. MSPSS resulted to be the greatest among L cluster with a decreasing trend by cluster increasing. The highest PPS scores were reported by cluster H-, followed by H+. In the regression models, we removed three observations due to the presence of outliers; the final models were satisfactory (Figure S3). In Table 3, the regression model estimates reported a significant risk increase for CBCL total problems among clusters L-, L+, H-, and H + compared to the cluster VL taken as a reference, with an adjusted risk increase ranging from 46–76%. The risk increase was higher considering the internalizing problems, with cluster H- reported more than two-fold in comparison with cluster VL.

Table 2

Main characteristics of the sample by PTSD cluster membership.
Characteristic	PTSD cluster					p-value²
Characteristic	VL, N = 84¹	L-, N = 75¹	L+, N = 62¹	H-, N = 53¹	H+, N = 53¹
Mother’s age (years)	34 (32, 36)	33 (30, 36)	33 (30, 35)	33 (30, 37)	31 (29, 34)	0.067
Educational level						< 0.001
Bachelor’s degree	23 (27%)	21 (28%)	17 (27%)	6 (11%)	18 (34%)
High school or lower	15 (18%)	24 (32%)	25 (40%)	19 (36%)	24 (45%)
Master’s degree or higher	46 (55%)	30 (40%)	20 (32%)	28 (53%)	11 (21%)
Marital status						0.279
Live-in partner	31 (37%)	17 (23%)	26 (42%)	21 (40%)	16 (30%)
Single	2 (2.4%)	4 (5.3%)	2 (3.2%)	2 (3.8%)	4 (7.5%)
Married	51 (61%)	54 (72%)	34 (55%)	30 (57%)	33 (62%)
Residence in red region [Yes]	46 (55%)	46 (61%)	34 (55%)	26 (49%)	18 (34%)	0.040
Job status [Employed]	80 (95%)	67 (89%)	59 (95%)	51 (96%)	47 (89%)	0.311
Family income						0.193
Low ≤ 12k €	0 (0%)	3 (4.0%)	3 (4.8%)	3 (5.7%)	5 (9.4%)
Low-Medium (12k-25k] €	14 (17%)	13 (17%)	18 (29%)	12 (23%)	15 (28%)
Medium (25k-50k] €	58 (69%)	45 (60%)	33 (53%)	30 (57%)	28 (53%)
High > 50k €	12 (14%)	14 (19%)	8 (13%)	8 (15%)	5 (9.4%)
Parity [Yes] (%)	17 (20%)	18 (24%)	10 (16%)	17 (32%)	13 (25%)	0.345
Gestational week at enrollment (weeks)	26 (21, 32)	23 (18, 31)	27 (18, 35)	30 (22, 33)	27 (17, 33)	0.143
Type of birth						0.561
Non-programmed C-Section	10 (12%)	12 (16%)	6 (9.7%)	7 (13%)	8 (15%)
Programmed C-Section	15 (18%)	5 (6.7%)	5 (8.1%)	6 (11%)	5 (9.4%)
Vaginal delivery	59 (70%)	58 (77%)	51 (83%)	40 (76%)	40 (75%)
Children’s sex [Female]	44 (52%)	34 (45%)	26 (42%)	27 (51%)	26 (49%)	0.747
Child’s gestational week at birth (weeks)	40 (39, 41)	40 (39, 41)	40 (38, 40)	39 (38, 40)	40 (38, 40)	0.562
COVID-19 SEE pregnancy [Yes]	12 (14%)	10 (13%)	10 (16%)	8 (15%)	7 (13%)	0.990
COVID-19 SEE postpartum[Yes]	27 (32%)	21 (28%)	19 (31%)	16 (30%)	18 (34%)	0.963
Total MSPSS	5.17 (4.75, 5.67)	5.17 (4.58, 5.67)	4.92 (4.27, 5.50)	4.83 (4.17, 5.42)	4.42 (3.83, 5.00)	< 0.001
CBCL total problems	7 (3, 11)	10 (6, 17)	10 (5, 15)	10 (6, 21)	10 (6, 18)	0.002
CBCL externalizing problems	4 (2, 8)	7 (3, 11)	6 (3, 10)	7 (4, 15)	7 (4, 12)	0.010
CBCL internalizing problems	2 (1, 4)	3 (2, 6)	3 (2, 6)	4 (2, 8)	3 (2, 7)	0.002
PPS score	-0.34 (-1.03, 0.15)	0.02 (-0.77, 0.59)	0.03 (-0.59, 0.66)	0.50 (0.06, 0.93)	0.31 (-0.08, 0.74)	< 0.001
¹ Median (IQR) or Frequency (%)
² Kruskal-Wallis rank sum test; Pearson's Chi-squared test; Fisher's exact test

Table 3

Adjusted* IRR and 95%CI related to PTSD cluster membership for CBCL scores on internalizing, externalizing and total problems with respect to the reference cluster VL. Models incorporated the weights estimated by propensity score model.
	Internalizing problems			Externalizing problems			Total problems
	IRR¹	95% CI¹	p-value	IRR¹	95% CI¹	p-value	IRR¹	95% CI¹	p-value
Cluster
VL	Ref.	—	—	Ref.	—	—	Ref.	—	—
L-	1.67	1.21, 2.31	0.001	1.52	1.16, 2.00	0.002	1.62	1.24, 2.11	< 0.001
L+	1.52	1.08, 2.13	0.014	1.40	1.05, 1.85	0.019	1.46	1.11, 1.93	0.006
H-	2.02	1.42, 2.87	< 0.001	1.60	1.19, 2.17	0.002	1.76	1.31, 2.37	< 0.001
H+	1.72	1.20, 2.49	0.003	1.46	1.07, 2.00	0.014	1.58	1.16, 2.14	0.003
¹ IRR = Incidence Rate Ratio, CI = Confidence Interval

*adjusted for mother’s age (in continuous form), educational level (master’s degree or higher, bachelor’s degree, high school or lower), marital status (live-in partner, single, married), residence in red region (yes, no), job status (employed, not employed), family income (low ≤ 12k €, low-Medium (12k-25k] €, medium (25k-50k] €, high > 50k €), parity (yes, no), gestational week at enrollment (in continuous form), type of birth (non-programmed C-Section, programmed C-Section, vaginal delivery), children’s sex (male, female), child’s gestational age at birth (in continuous form), COVID-19 SEE pregnancy (yes, no), COVID-19 SEE postpartum (yes, no), total MSPSS score (in continuous form), and PPS (score in continuous form).

This study described trajectories of women’s perinatal PTSD symptoms and their association with children’s emotional-behavioral outcomes at 12 months of age. The clustering approach evidenced that women's PTSD response to the pandemic COVID-19 over their perinatal period is not homogeneous, identifying five different trajectories of PTSD symptoms. Considering the PTSD scores over time, the stable group (VL) showed very low and steadily low PTSD symptoms at each time point from pregnancy to 12 months post-delivery, which is consistent with a resilience pattern previously observed for birth-related PTSD (Dikmen-Yildiz et al., 2018). Two clusters (H- and L-) showed a decreasing trend of PTSD symptoms, which partially parallel the recovered pattern observed for birth-related PTSD (Dikmen-Yildiz et al., 2018). The main difference between them was that in the cluster H- the average PTSD score at t0 was above the cut-off point in pregnancy but not at 6 nor at 12 months postpartum, while the cluster L- presented PTSD scores in pregnancy below the threshold with a decreasing temporal trend, suggesting another resilience pattern with a slight PTSD vulnerability in pregnancy. The two remaining clusters (L + and H+) exhibited increasing PTSD trajectories for PTSD from pregnancy to postpartum. During pregnancy H + had higher PTSD scores than cluster L+, a peak 6 months after delivery and, a stabilization thereafter, partially presenting a pattern of chronic symptoms. The cluster L + reported a positive trend narrowing the cutoff score at 12 months’ postpartum, suggesting a delayed pattern for pandemic-related perinatal PTSD risk. These two last groups conceal two crucial potential risks of applying categorical approaches to perinatal mental health screening, as both groups would have been on average negative to a prenatal PTSD screening, but they both present the highest PTSD scores by 12 months postpartum. Since the observational nature of the study, the trajectory groups only partially align with longitudinal patterns seen in other studies on birth-related PTSD and trauma. Notably, in our sample the cluster VL (steadily very low PTSD symptoms) accounted for 25.7% of mothers, while the birth-related PTSD resilient group in the Dikmen-Yildiz and colleagues’ study (Dikmen-Yildiz et al., 2018) represented 61.9% of the study participants.

The observed associations with PTSD clusters were all statistically significant for both the externalizing and the internalizing domains, with stronger link observed for the latter. This consistently with both cohort evidence for prenatal stressful events and offspring conduct and hyperactivity symptoms risk (MacKinnon et al., 2018), and for maternal stress in pregnancy and child’s internalizing problems (Park et al., 2014). Externalizing behaviors are characterized by high activity levels, difficulty in inhibition, and/or aggressive behavior, while internalizing behaviors pertains depressive, anxiety, and somatic complaints. In our study, children whose mothers reported the highest PTSD scores in pregnancy (H+, H-) rated about two-fold higher risk in behavioral problems, in comparison with those in the VL cluster. Particularly, CBCL scores were the highest for children in the H- group, despite a negative trend in mothers’ PTSD symptoms over time. This suggests the centrality of maternal stress response during pregnancy for child development (Glover et al., 2018). We speculate that both fetal mechanisms for health and development programming and the parental neurobiological changes of the transition to parenthood might be severely impacted by the pandemic acute stress and the maternal post-traumatic response symptoms, producing a two-generation impact by simultaneously endangering the child’s neurobiological systems (e.g., deregulation of cortisol, cytokines, and serotonin functioning) (Glover, 2014) and the mother’s neurobehavioral resources available for parenting (e.g., sensitivity to infant signals, mother-fetus attachment) (Pearson et al., 2010; Rutherford et al., 2016; Sacchi et al., 2021). The second higher risk for child’s problem behaviors was observed in the H + group; this group presented sub-clinical levels of PTSD symptoms in pregnancy that reached a peak in early postpartum. The results were consistent with a process of stress accumulation that might have interested this group of women, where we also observed the highest levels of pandemic stress in pregnancy and the lowest perceived social support postpartum. Indeed, it might be that for these mothers a mental health vulnerability in pregnancy has become symptom-level significant only after prolonged exposure to the pandemic conditions (e.g., pandemic stress in pregnancy) and to the difficulties of postpartum and parenting with such diminished individual (i.e., mental health) and interpersonal (e.g., social support) resources. This partially consists with previous Italian evidence of a mediating role played by parental postpartum distress in the association between childbirth-related PTS symptoms and children’s internalizing behaviors (Di Blasio et al., 2017). Besides, hostile-reactive parenting during infancy has been also described to predict greater externalizing problems, and an indirect pathway from maternal postnatal distress to child attention-deficit/hyperactivity disorder has been described via parenting hostility. Notably, European data on parenting health during pandemic reported a higher level of harsh parenting during the COVID-19 lockdown (Sari et al., 2022). In addition, previous studies separately provide additional explanations for the association between perinatal maternal health and child’s behavioral outcomes. For instance, elevated maternal cortisol (as a response to environmental stressors) associates with elevated levels of testosterone in the uterine environment, which has been independently reported as associated with externalizing disorders; also, behavior inhibition and inattention are associated with dopamine receptor availability, and higher ratio of dopamine receptors and dopamine is observed in offspring of prenatally stressed mothers (Chapman et al., 2006; Knickmeyer et al., 2005; Lou et al., 2004; Scerbo & Kolko, 1994).

LIMITATIONS

This study has limitations that should be acknowledged for a correct interpretation. The use of self-report measures limits speculation to symptom-level PTSD risk. Alongside, information on women’s mental health and child outcomes were both obtained by maternal reports, thus the risk of reporter bias cannot be excluded, even if little psychometric evidence for maternal psychopathology biasing reports of child behavior problems has been found (Olino et al., 2021). Regional and socioeconomic distribution of the sample as well as the sample size, considering the longitudinal design and extraordinary life and research conditions, support the robustness of findings, however, the self-selection of participants should not be overlooked. In addition, another limitation of the study is that, beyond the general extraordinary conditions of psychosocial distress about which they are questioned, pregnant women are not directly asked to report a specific event in relation to their symptoms. To compensate for this limitation, respondents' answers were weighted, using a propensity score-based method derived from responses about the presence of COVID-19 Stressful Events Exposure during pregnancy reported at t1. It is important to highlight that we did not conduct neurobiological and parenting investigations, to deepen the understanding of underlying mechanisms for the mother-infant risk transmission. Along with this, several strengths support the evidence we presented. Timely assessment during the pandemic outbreak prevents recollection bias for maternal stress exposure; assessment of mother’s and child’s variables rely on gold-standard self-reported measures, that guaranteed faster, more effortless recruitment and longitudinal participation during sensitive life conditions, by also providing wide access to the survey for a portion of women limited by geographical and familiar constraints. Last, the innovative methodological approach of clustering participants for the unfolding of PTSD symptoms over time allows us to finely describe the variability of mental health dynamics over the transition to parenthood ultimately supporting the translational aim of improving the individualization of perinatal care.

Current results for perinatal PTSD in the context of the COVID-19 pandemic suggest the relevance of continuous and qualified attention toward mental health in perinatal setting, to timely detect individualized patterns of risk, including delayed vs recovery trajectories, chronic burden vs resilience vs vulnerable patterns in PTSD symptoms, and their different associations with second-generation risk for developmental psychopathology outcomes. Compared to single assessment, longitudinal modeling of perinatal PTSD symptoms allowed more sensitive two-generation risk detection. Preparedness of operators, awareness of perinatal post-traumatic stress disorder for timely assessment and intervention, and prevention of adverse maternal–infant outcomes especially with groups with recognized individual (i.e., childhood trauma, IPV, tocophobia) or collective (i.e., natural disasters, wars, and conflicts) risks for PTSD (Canfield & Silver, 2020; de Graaff et al., 2018) is warranted.

The study shows the significance of continuous monitoring of women mental health along the perinatal period to timely detect individualized patterns of risk, including delayed vs recovery trajectories, chronic burden vs resilience vs vulnerable patterns in PTSD symptoms, and their different associations with second-generation risk for developmental psychopathology outcomes. We provide support for the risks of applying categorical and single-time approaches to the screening for perinatal mental health. Longitudinal modeling of perinatal PTSD symptoms allowed more sensitive two-generation risk detection. Such an approach is warranted in the risk assessment and tracking of other perinatal mood and anxiety symptoms.

Competing interests. The authors declare no competing interests.

Ethical Approval. Ethical approval was obtained before data collection. [The Institutional Review Board of the University of Padova - 08/04/2020, approval n. 3545.]. Informed consent was obtained from all individual participants included in the study.

Data Availability. All R code and data associated with the real data application are available at https://osf.io/x6tev/.

Authors contribution. Conceptualization: CS, SV, PG; Data curation: PG; Formal analysis: PG; Investigation: CS, SV; Methodology: CS, SV; Writing - original draft: CS, SV, PG; Writing - review & editing: CS, PG.

Funding. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Achenbach TM (1999) The Child Behavior Checklist and related instruments. The use of psychological testing for treatment planning and outcomes assessment, 2nd edn. Lawrence Erlbaum Associates, pp 429–466
Aizer A, Stroud L, Buka S (2012) Maternal Stress and Child Outcomes: Evidence from Siblings (Working Paper 18422). National Bureau of Economic Research. https://doi.org/10.3386/w18422
Ayers S (2004) Delivery as a traumatic event: Prevalence, risk factors, and treatment for postnatal posttraumatic stress disorder. Clin Obstet Gynecol 47(3):552–567. https://doi.org/10.1097/01.grf.0000129919.00756.9c
Ayers S, Bond R, Webb R, Miller P, Bateson K (2019) Perinatal mental health and risk of child maltreatment: A systematic review and meta-analysis. Child Abuse Negl 98:104172. https://doi.org/10.1016/j.chiabu.2019.104172
Ayers S, Eagle A, Waring H (2006) The effects of childbirth-related post-traumatic stress disorder on women and their relationships: A qualitative study. Psychol Health Med 11(4):389–398. https://doi.org/10.1080/13548500600708409
Canfield D, Silver RM (2020) Detection and Prevention of Postpartum Posttraumatic Stress Disorder: A Call to Action. Obstet Gynecol 136(5):1030. https://doi.org/10.1097/AOG.0000000000004093
Chapman E, Baron-Cohen S, Auyeung B, Knickmeyer R, Taylor K, Hackett G (2006) Fetal testosterone and empathy: Evidence from the Empathy Quotient (EQ) and the Reading the Mind in the Eyes Test. Soc Neurosci 1(2):135–148. https://doi.org/10.1080/17470910600992239
Cook N, Ayers S, Horsch A (2018) Maternal posttraumatic stress disorder during the perinatal period and child outcomes: A systematic review. J Affect Disord 225:18–31. https://doi.org/10.1016/j.jad.2017.07.045
Davies J, Slade P, Wright I, Stewart P (2008) Posttraumatic stress symptoms following childbirth and mothers’ perceptions of their infants. Infant Mental Health J 29(6):537–554. https://doi.org/10.1002/imhj.20197
de Graaff LF, Honig A, van Pampus MG, Stramrood CAI (2018) Preventing post-traumatic stress disorder following childbirth and traumatic birth experiences: A systematic review. Acta Obstet Gynecol Scand 97(6):648–656. https://doi.org/10.1111/aogs.13291
Dempster AP, Laird NM, Rubin DB (1977) Maximum Likelihood from Incomplete Data Via the EM Algorithm. J Roy Stat Soc: Ser B (Methodol) 39(1):1–22. https://doi.org/10.1111/j.2517-6161.1977.tb01600.x
Di Blasio P, Camisasca E, Miragoli S, Ionio C, Milani L (2017) Does Maternal Parenting Stress Mediate the Association Between Postpartum PTS Symptoms and Children’s Internalizing and Externalizing Problems? A Longitudinal Perspective. Child Youth Care Forum 46(5):685–701. https://doi.org/10.1007/s10566-017-9400-7
Dikmen-Yildiz P, Ayers S, Phillips L (2018) Longitudinal trajectories of post-traumatic stress disorder (PTSD) after birth and associated risk factors. J Affect Disord 229:377–385. https://doi.org/10.1016/j.jad.2017.12.074
Durbano F (2013) New Insights into Anxiety Disorders. BoD – Books on Demand
Erickson N, Julian M, Muzik M (2019) Perinatal depression, PTSD, and trauma: Impact on mother–infant attachment and interventions to mitigate the transmission of risk. Int Rev Psychiatry 31(3):245–263. https://doi.org/10.1080/09540261.2018.1563529
Garthus-Niegel S, Ayers S, Martini J, von Soest T, Eberhard-Gran M (2017) The impact of postpartum post-traumatic stress disorder symptoms on child development: A population-based, 2-year follow-up study. Psychol Med 47(1):161–170. https://doi.org/10.1017/S003329171600235X
Glover V (2011) Annual Research Review: Prenatal stress and the origins of psychopathology: an evolutionary perspective. J Child Psychol Psychiatry 52(4):356–367. https://doi.org/10.1111/j.1469-7610.2011.02371.x
Glover V (2014) Maternal depression, anxiety and stress during pregnancy and child outcome; what needs to be done. Best Pract Res Clin Obstet Gynecol 28(1):25–35. https://doi.org/10.1016/j.bpobgyn.2013.08.017
Glover V, O’Donnell KJ, O’Connor TG, Fisher J (2018) Prenatal maternal stress, fetal programming, and mechanisms underlying later psychopathology-A global perspective. Dev Psychopathol 30(3):843–854. https://doi.org/10.1017/S095457941800038X
Grobman WA, Entringer S, Headen I, Janevic T, Kahn RS, Simhan H, Yee LM, Howell EA (2024) Social determinants of health and obstetric outcomes: A report and recommendations of the workshop of the Society for Maternal-Fetal Medicine. Am J Obstet Gynecol 230(2):B2–B16. https://doi.org/10.1016/j.ajog.2023.10.013
Hendrix CL, Werchan D, Lenniger C, Ablow JC, Amstadter AB, Austin A, Babineau V, Bogat GA, Cioffredi L-A, Conradt E, Crowell SE, Dumitriu D, Elliott AJ, Fifer W, Firestein M, Gao W, Gotlib I, Graham A, Gregory KD, Thomason ME (2022) Geotemporal analysis of perinatal care changes and maternal mental health: An example from the COVID-19 pandemic. Archives Women’s Mental Health 25(5):943–956. https://doi.org/10.1007/s00737-022-01252-6
Hoekzema E, Barba-Müller E, Pozzobon C, Picado M, Lucco F, García-García D, Soliva JC, Tobeña A, Desco M, Crone EA, Ballesteros A, Carmona S, Vilarroya O (2017) Pregnancy leads to long-lasting changes in human brain structure. Nat Neurosci 20(2):287–296. https://doi.org/10.1038/nn.4458
Knickmeyer R, Baron-Cohen S, Raggatt P, Taylor K (2005) Foetal testosterone, social relationships, and restricted interests in children. J Child Psychol Psychiatry 46(2):198–210. https://doi.org/10.1111/j.1469-7610.2004.00349.x
Lou HC, Rosa P, Pryds O, Karrebæk H, Lunding J, Cumming P, Gjedde A (2004) ADHD: Increased dopamine receptor availability linked to attention deficit and low neonatal cerebral blood flow. Dev Med Child Neurol 46(3):179–183. https://doi.org/10.1017/S0012162204000313
MacKinnon N, Kingsbury M, Mahedy L, Evans J, Colman I (2018) The Association Between Prenatal Stress and Externalizing Symptoms in Childhood: Evidence From the Avon Longitudinal Study of Parents and Children. Biol Psychiatry 83(2):100–108. https://doi.org/10.1016/j.biopsych.2017.07.010
McCormack C, Abuaish S, Monk C (2023) Is There an Inflammatory Profile of Perinatal Depression? Curr Psychiatry Rep 25(4):149–164. https://doi.org/10.1007/s11920-023-01414-y
Moran Vozar TE, Van Arsdale A, Gross LA, Hoff E, Pinch S (2021) The elephant in the delivery room: Enhancing awareness of the current literature and recommendations for perinatal PTSD. Pract Innovations 6(1):1–16. https://doi.org/10.1037/pri0000134
Olino TM, Michelini G, Mennies RJ, Kotov R, Klein DN (2021) Does maternal psychopathology bias reports of offspring symptoms? A study using moderated non-linear factor analysis. J Child Psychol Psychiatry 62(10):1195–1201. https://doi.org/10.1111/jcpp.13394
Onoye JM, Goebert D, Morland L, Matsu C, Wright T (2009) PTSD and postpartum mental health in a sample of Caucasian, Asian, and Pacific Islander women. Archives Women’s Mental Health 12(6):393–400. https://doi.org/10.1007/s00737-009-0087-0
Park S, Kim B-N, Kim J-W, Shin M-S, Yoo HJ, Lee J, Cho S-C (2014) Associations between maternal stress during pregnancy and offspring internalizing and externalizing problems in childhood. Int J Mental Health Syst 8(1):44. https://doi.org/10.1186/1752-4458-8-44
Pearson RM, Cooper RM, Penton-Voak IS, Lightman SL, Evans J (2010) Depressive symptoms in early pregnancy disrupt attentional processing of infant emotion. Psychol Med 40(4):621–631. https://doi.org/10.1017/S0033291709990961
Penninx BWJH, Benros ME, Klein RS, Vinkers CH (2022) How COVID-19 shaped mental health: From infection to pandemic effects. Nature Medicine, 28(10), Articolo 10. https://doi.org/10.1038/s41591-022-02028-2
Radoš SN, Matijaš M, Anđelinović M, Čartolovni A, Ayers S (2020) The role of posttraumatic stress and depression symptoms in mother-infant bonding. J Affect Disord 268:134–140. https://doi.org/10.1016/j.jad.2020.03.006
Rutherford HJV, Graber KM, Mayes LC (2016) Depression symptomatology and the neural correlates of infant face and cry perception during pregnancy. Soc Neurosci 11(4):467–474. https://doi.org/10.1080/17470919.2015.1108224
Sacchi C, Carli PD, Gregorini C, Monk C, Simonelli A (2023) In the pandemic from the womb. Prenatal exposure, maternal psychological stress and mental health in association with infant negative affect at 6 months of life. Dev Psychopathol 1–11. https://doi.org/10.1017/S0954579423000093
Sacchi C, Miscioscia M, Visentin S, Simonelli A (2021) Maternal–fetal attachment in pregnant Italian women: Multidimensional influences and the association with maternal caregiving in the infant’s first year of life. BMC Pregnancy Childbirth 21(1):488. https://doi.org/10.1186/s12884-021-03964-6
Sari NP, van IJzendoorn MH, Jansen P, Bakermans-Kranenburg M, Riem MME (2022) Higher Levels of Harsh Parenting During the COVID-19 Lockdown in the Netherlands. Child Maltreat 27(2):156–162. https://doi.org/10.1177/10775595211024748
Scerbo AS, Kolko DJ (1994) Salivary Testosterone and Cortisol in Disruptive Children: Relationship to Aggressive, Hyperactive, and Internalizing Behaviors. J Am Acad Child Adolesc Psychiatry 33(8):1174–1184. https://doi.org/10.1097/00004583-199410000-00013
Seng JS, Sperlich M, Low LK, Ronis DL, Muzik M, Liberzon I (2013) Childhood Abuse History, Posttraumatic Stress Disorder, Postpartum Mental Health, and Bonding: A Prospective Cohort Study. J Midwifery Women’s Health 58(1):57–68. https://doi.org/10.1111/j.1542-2011.2012.00237.x
Shahar G, Herishanu-Gilutz S, Holcberg G, Kofman O (2015) In first-time mothers, post-partum depressive symptom prospectively predict symptoms of post-traumatic stress. J Affect Disord 186:168–170. https://doi.org/10.1016/j.jad.2015.07.021
Swain JE, Ho SS, Rosenblum KL, Morelen D, Dayton CJ, Muzik M (2017) Parent–child intervention decreases stress and increases maternal brain activity and connectivity during own baby-cry: An exploratory study. Dev Psychopathol 29:535–553. https://doi.org/10.1017/S0954579417000165
Van Sieleghem S, Danckaerts M, Rieken R, Okkerse JME, de Jonge E, Bramer WM, van den Lambregtse MP (2022) Childbirth related PTSD and its association with infant outcome: A systematic review. Early Hum Dev 174:105667. https://doi.org/10.1016/j.earlhumdev.2022.105667
Weathers FW, Litz BT, Keane TM, Palmieri PA, Marx BP, Shnurr PP (2013) The PTSD Checklist for DSM-5 (PCL-5). Retrieved from https://www.ptsd.va.gov/professional/assessment/adult-sr/ptsd-checklist.asp
Webb R, Ayers S (2015) Cognitive biases in processing infant emotion by women with depression, anxiety and post-traumatic stress disorder in pregnancy or after birth: A systematic review. Cogn Emot 29(7):1278–1294. https://doi.org/10.1080/02699931.2014.977849
Zimet GD, Powell SS, Farley GK, Werkman S, Berkoff KA (1990) Psychometric Characteristics of the Multidimensional Scale of Perceived Social Support. J Pers Assess 55(3–4):610–617. https://doi.org/10.1080/00223891.1990.9674095

No competing interests reported.

PTSDtrajectoriesSM.docx

Download PDF

Editor assigned by journal
13 Jul, 2024
Submission checks completed at journal
10 Jul, 2024
First submitted to journal
09 Jul, 2024

You are reading this latest preprint version

Trajectories of perinatal post-traumatic stress disorder scores in association with child’s behavior at 12 months

Status:

Version 1

Abstract

Figures

INTRODUCTION

METHODS

Procedure

Measures

Data analysis

Data imputation

Descriptive Statistical analysis

PTSD Clustering – EM Algorithm

Propensity score methods

Regression models

RESULTS

DISCUSSION

LIMITATIONS

CONCLUSION

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1