Causal Pathways Between the Acute Experience of Violence During Pregnancy and Fetal Intrauterine Growth Restriction: A Cohort Study

Abstract

Introduction:

Violence during pregnancy (VDP) is a prevalent global issue with dire consequences for the mother and the developing fetus. These consequences include prematurity, low birthweight, and intrauterine growth restriction (IUGR), but its pathways remain elusive. This study investigated the causal pathways between VDP and IUGR using mediation analysis.

Methods:

A prospective population-based birth cohort was followed from the beginning of the third gestational trimester to the second year of life. IUGR was defined by the Kramer index, and information on VDP was collected using the WHO–Violence Against Women (WHO VAW) questionnaire. Cases were considered positive only when no other life episodes were reported. Ten different mediators were analyzed as possible pathways based on previous research. Path analysis was conducted to evaluate these relationships.

Results:

The path analysis model included 755 dyads and presented an adequate fit. Violence during pregnancy showed a direct effect (β = −0.195, p = 0.041) and a total effect (β = −0.276, p = 0.003) on IUGR. Violence was associated with gestational depression or anxiety, tobacco and alcohol consumption, changes in blood pressure, and the need for emergency care, but these did not constitute mediators of its effect on IUGR. The sum of the indirect effects, however, showed a significant association with IUGR (β = −0.081, p = 0.011).

Conclusion:

The acute experience of violence during pregnancy was associated with IUGR, primarily via a direct pathway. An indirect effect was also present but not mediated through the variables analyzed in this study. The robust strength of these associations underscores the negative health consequences of violence against women for the succeeding generation.

Introduction

Violence against women perpetrated by an intimate partner is a worldwide phenomenon, present in populations with different degrees of development and with different cultural characteristics. WHO data suggests its lifetime prevalence to be 30%,¹ whereas a recent review suggests its prevalence in pregnancy to be around 25% (95% confidence interval [CI]: 20.3, 30.5).²

Evidence suggests an array of negative health outcomes for the concept when violence is inflicted during pregnancy. A recent systematic review found varying evidence of its relationship to 29 different perinatal outcomes, including prematurity, low birthweight, miscarriage, hypertension, and bleeding.³ Some of these outcomes are more well-established, given their importance in infants’ morbimortality and later life consequences.^4

–7 Low birthweight, small size at birth, prematurity, and intrauterine growth restriction (IUGR) can be understood as proxies for the intrauterine environment condition and are thus related to the health and disease of most of the systems.⁸

A meta-analysis with 14 studies from developed countries found an odds ratio (OR) of 1.4 (95% CI: 1.1, 1.8) for low birthweight (LBW).⁹ Another meta-analysis, including 30 investigations, found an adjusted OR of 1.53 (95% CI: 1.29, 1.82) for LBW and that of 1.46 (95% CI: 1.27, 1.67) for prematurity.¹⁰ A further one, with 50 published studies included, found an OR for prematurity of 1.91 (95% CI 1.6, 2.29), that for LBW of 2.11 (95% CI: 1.68, 2.65), and a modest influence for small-for-gestational-age (SGA) infants (1.37, 95% CI: 1.02, 1.84).¹¹ A more recent meta-analysis with 19 studies estimated an OR for LBW of 1.18 (95% CI: 1.05, 1.31) and that for prematurity of 1.42 (95% CI: 1.21, 1.63).¹² A systematic review found an association of intimate partner violence and miscarriage ranging from 1.7 to 5.4 in 10 studies and reported a risk of 2.4–4.0 times higher for SGA in the case of physical violence. The evidence for the association between violence during pregnancy and IUGR is less clear, and the only available meta-analysis on the topic included a small number of studies in comparison with the other studied outcomes (four studies), finding no significant association.¹²

Given the challenges in tackling gender violence worldwide,^13,14 measures to mitigate its impact should be considered. In relation to its influence on the following generations, it is essential to have a profound comprehension of the involved causal mechanisms that link violence during pregnancy to perinatal and later-life negative consequences. Violence during pregnancy is associated with infectious processes and a higher hospitalization rate.¹⁵ Low self-esteem, depression, and anxiety disorders are more frequent in pregnant women who have suffered violence.^16,17 It is also known that impacted mental health compromises self-care. There is also an association with risky behaviors, such as a higher smoking rate, consumption of alcohol and other psychoactive substances, inadequate nutrition with small weight gain, and inadequate prenatal care.^15,16 There are also reports that intimate partner violence during pregnancy can exacerbate chronic health problems, such as hypertension and diabetes.^15,18 Furthermore, pregnancy in a violent relationship is commonly unplanned.^19,20

When reviewing this theme, Bailey²¹ suggested that “many of maternal factors are associated with intimate partner violence during pregnancy, including physical and mental problems, and negative health behaviors. Causality has not been established, and thus, the direction of these relations is not known.” Yount et al.,²² in another conceptual review of these causal pathways, recognized the necessity for clarifying the behavioral and biological pathways through which domestic violence influences infant growth. In the past decade, although a significant number of studies reported on the outcomes related to intimate partner violence during pregnancy, the study of the pathways remains scant, and still little is known about the contribution of each component of this complex chain.

This study aimed to identify and describe the causal pathways that explain the association of violence during pregnancy and newborn IUGR with mediation analyses.

Methods

Study design

A prospective population-based birth cohort²³ was followed from the beginning of the third gestational trimester to the second year of life. In the analysis here presented, we used data from the gestational and puerperium periods. All the mothers eligible for study participation were invited by means of a free and informed consent form, presented by a professional who was able to recognize reading or comprehension difficulties. The mothers read and signed the consent form after they demonstrated comprehension of the objective of the investigation. Hospital chart data were used only after obtaining the mothers’ consent. The research protocol was approved by the Research Committees of the Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo (Clinics Hospital of the Faculty of Medicine of the University of São Paulo; protocol number 0054/09) and of the Health Secretary of the City of São Paulo (protocol number 122/10).

Population and sample

All pregnant women who received care at five Basic Health Units (Unidades Básicas de Saúde [UBS]) in the Health District of Butantã, located in the western zone of the municipality of São Paulo, Brazil, between July 2010 and December 2012, were eligible for participation. Only the first five pregnant women with prenatal care appointments at each unit each week were included. Exclusion criteria for the analyses were non-singletons and children born with any condition related to developmental and/or growth delay.

The Health District of Butantã is covered by the Family Health Strategy (Estratégia Saúde da Família) Program, which guarantees access to prenatal care for all the families in the regions covered by the Basic Health Units by means of monthly domiciliary visits and early identification of pregnancies. This region is far from the center of the city, has a high populational density, has an absence of social amenities, and has significant social vulnerabilities. Violent environments with marked conflicts between drug traffickers and the police coexist with nurseries, schools, and churches. In this region, there is a mix of poverty and the recently formed middle class.

For our study, we calculated a sample size of 717 mothers, using the following assumptions: a prevalence of IUGR of 9% in the population; the ratio of nonexposed and exposed to violence of 4:1; a relative risk of 2.00; and a probability of α error of 5% and β error of 20%. Given the risk of follow-up loss, we chose to recruit 900 mothers.

Variables

The outcome chosen was being born with IUGR, defined by the Kramer index,²⁴ i.e., the ratio of birthweight to the 50% percentile of weight for gestational age and sex, with the Intergrowth-21 as the reference curve.

The gestational age definition was based on the ultrasound (US) examination conducted until the 12th week of pregnancy for 51.8% of the sample, whereas the rest had this examination later. For the second group, gestational age definition was based on the date of last menstruation. When the estimate by the Capurro Somatic Method and the report differed by >2 weeks, the estimate was used. This variable was standardized and used as a continuous variable.

The exposure variable chosen was violence perpetrated by the intimate partner. This information was collected at the beginning of the third trimester by trained psychologists during a prenatal care appointment. The chosen instrument was the World Health Organization Domestic Violence Questionnaire,²⁵ previously adapted to the Brazilian Portuguese language and to the Brazilian sociocultural context.²⁶ Questions 704 to 706 in the WHO instrument include seven types of physical abuse, four types of psychologic abuse and three types of sexual abuse perpetrated by the husband/partner in the past 12 months. As the pregnant women were in the sixth month of gestation at the time of the interview, the violence reported could have occurred immediately before or during conception. Cases were considered positive only when no other life episodes were reported. Exposure represented, then, the acute experience of violence. From these data, a continuous latent variable was created to represent acute violence during pregnancy.

To collect the information about violence, the team of psychologists received training based on the WHO guide “Putting Women First: Ethical and Safety Recommendations for Research on Domestic Violence Against Women”.²⁷ In addition, the training prioritized the skills needed to cope with the challenges that could arise in this setting. One field researcher had to leave the data collection because of being unable to face these challenges. Women identified as victims of violence were advised to seek the nearest Woman’s Defense Precinct and were suggested to talk to their family doctors.

Mental health disorders were measured using the Mini International Neuropsychiatric Interview (MINI), a structured diagnostic interview to evaluate current disorders.²⁸ This instrument was previously translated and adapted to the Brazilian context.²⁹ The MINI is a diagnostic instrument that can be administered by nonmedical professionals. For this study, depression and anxiety were considered. Current depressive disorder (ICD-10 F32.x) was labeled as “depression.” In the analysis, only women who reported the current episode as their first were included. The second category was labeled as “anxiety” and included cases of current panic disorder without agoraphobia (ICD F40.01), current panic disorder with agoraphobia (ICD F43.0), social anxiety (ICD F40.1), post-traumatic stress disorder (ICD-10 F43.1), generalized anxiety disorder (ICD-10 F41.1), and obsessive–compulsive disorder (ICD-10 F42.8). Women diagnosed with these disorders were counseled to talk to their family doctors. In rare cases, after the high-risk evaluation for the mother and her inability to follow the counseling, the researcher reported the information directly to the responsible doctor. Smoking and drinking were defined as any reported consumption during pregnancy.

The sociodemographic variables included were: socioeconomic class (SES), as proposed by the ABEP (Brazilian Association of Research Companies),³⁰ where “A” represents the highest-earning class and “E” the lowest-earning class; maternal education, in school years, grouped as 0–7 years (including women who did not complete mandatory school years), 8–10 years, and ≥11 years; maternal age, categorized as age <20 years (teenagers), 20–35 years, and ≥36 years; maternal migration, i.e., if the mother was born in the city of São Paulo or came from other regions; maternal skin color (white, black, brown, yellow/East Asian, or indigenous) and if the mother lives with her partner or not. In Brazil, brown skin color refers to a diverse ethnic background resulting from different historical processes of migration and mixture. For risk analyses, skin color was grouped into “white” and “non-white.” The emotions felt when pregnancy was discovered were also collected as a proxy for planned or intended pregnancy. Cases were grouped as experiencing good emotions, transient stress, moderate stress, great stress, or considering abortion.

For a variable to be considered a mediator, it needs to be associated with the exposure, must be temporally located between exposure and outcome (as a descendant of the exposure but ancestral to the outcome), and it must have a plausible explanation to be a pathway for the effects of the exposure to occur.³¹ Possible mediators included: the number of prenatal care appointments until recruitment (28 weeks); gestational weight gain, defined as the ratio between weight gain and expected weight gain for gestational age and body mass index; use of antibiotic medication, defined as having received antibiotic treatment during pregnancy, as a proxy for bacterial infection; alterations in arterial pressure, defined as a systolic pressure ≥140 mmHg or diastolic pressure ≥90 mmHg, nonprogrammed appointments (emergency care), vaginal bleeding, mental health disorders (anxiety and depression); and alcohol or tobacco consumption.

The identification of confounding variables (ancestors of both exposure and outcome) was made using the directed acyclic graph (DAG) methodology³² (see Annex I) and from available literature. The use of these graphs provides a robust method for identifying confounders and prevents over-adjustment for mediators. Variables identified as confounders included SES, maternal age, maternal education, and feelings about the pregnancy.

Data analysis

Information was collected from articles and later transferred to Microsoft Excel software (version 6.01). To avoid mistakes, double input and data verification were performed. Descriptive analyses and univariate and multivariable regressions were conducted using STATA software (version 14.0),³³ using generalized linear models. Path and mediation analyses were conducted in MPlus software,³⁴ with maximum likelihood estimation.

We used inverse probability weighting in the models to account for missing data. Cases with missing data were also considered by maximum likelihood estimation in MPlus. We investigated the association of violence during pregnancy and IUGR through linear regression models, testing if this relationship had a linear or quadratic behavior. Linear regression modeling was also performed to evaluate the relationship between each mediator and the chosen outcome. As for the path analysis, the model would be considered adequate if (a) p value >0.05 for the chi-square test (χ ²), (b) root mean square error of approximation (RMSEA) values <0.06, (c) comparative fit index and Tucker–Lewis index (CFI/TLI) values were ≥0.95, and (d) standardized root mean square residual (SRMR) values were ≤0.08.^35,36

The null hypothesis was rejected when the probability of the occurrence of a type 1 error was <5%.

Results

Table 1 shows the descriptive statistics for the subsamples followed and lost until puerperium. There were no statistically significant differences between these two groups. The followed group (n = 755) belonged mostly to economic category C (65.2%), with women of age between 20 and 34 years and education equivalent to completing middle school. Most women referred to the pregnancy as desired. Furthermore, 13.9% were diagnosed with depression and 19.2% with anxiety, whereas 16.6% reported smoking habits and 9.9% reported alcohol ingestion during pregnancy. Majority of these women reported having brown skin, were born in São Paulo, lived with their partners, and were not in their first pregnancy. Psychological violence was reported by 21.3% of them, whereas physical violence was reported by 12.5% and sexual violence by 2.1%.

Table 1.

Sample Description, Comparing the Followed Subsample to the Lost At Birth Subsample

Category	Lost subsample (N = 139)				Followed subsample (N = 755)				p value^a
Category	N	Frequency	95% CI		N	Frequency	95% CI		p value^a
Economic Class
A + B	139	12.9	8.3	19.7	755	18.1	15.6	21.1	0.311
C		70.5	62.3	77.6		65.2	61.7	68.5
D + E		16.5	11.2	23.8		16.7	14.2	19.5
Maternal Education
Incomplete primary	139	20.9	14.8	28.5	754	18.6	15.9	21.5	0.377
Primary		43.2	35.1	51.6		39.1	35.7	42.7
Secondary or tertiary		36.0	28.3	44.4		42.3	38.8	45.9
Maternal age
<20 years	139	26.6	19.9	34.7	755	21.5	18.7	24.5	0.060
20–34 years		69.8	61.5	76.9		69.5	66.2	72.7
>=35 years		3.6	1.5	8.4		9.0	7.2	11.3
Pregnancy
Much desired	139	41.7	33.7	50.2	752	45.9	42.3	49.5	0.813
Desired		23.7	17.3	31.6		22.6	19.8	25.7
Not expected		10.8	6.6	17.2		11.8	9.7	14.4
Not desired		15.8	10.6	23.0		13.0	10.8	15.6
Considered abortion		7.9	4.4	13.8		6.6	5.1	8.7
Depression
Yes	139	10.1	6.0	16.4	755	13.9	11.6	16.6	0.221
Anxiety
Yes	139	19.4	13.6	26.9	755	19.2	16.5	22.2	0.952
Tobacco consumption
Yes	139	18.0	12.4	25.4	753	16.6	14.1	19.4	0.688
Alcohol consumption
Yes	139	12.2	7.7	18.9	754	9.9	8.0	12.3	0.416
Physical violence
Yes	139	10.1	6.0	16.4	753	12.5	10.3	15.0	0.423
Psychological violence
Yes	137	19.7	13.8	27.3	748	21.3	18.5	24.3	0.683
Sexual violence
Yes	139	0.0	—	—	755	2.1	3.4	1.3	0.153^b
Maternal skin color
White	133	39.8	31.8	48.5	733	36.8	33.4	40.4	0.700
Black		18.0	12.3	25.6		16.6	14.1	19.5
Brown		41.4	33.2	50.0		44.7	41.2	48.4
Yellow (East Asian)		0.8	0.1	5.3		1.8	1.0	3.0
Migrant mother
Yes	138	40.6	32.6	49.1	750	44.1	40.6	47.7	0.439
Primiparous mother
Yes	139	52.5	44.1	60.8	754	46.7	43.1	50.3	0.206
Living with partner
Yes	139	70.5	62.3	77.6	747	74.0	70.8	77.1	0.387

χ ² test.

Fisher’s exact test.

95% CI, 95% confidence interval; G, gestational.

Table 2 summarizes the results of the linear regression models performed to investigate the relationship between the variables. Our models suggest that domestic violence was associated with IUGR by a quadratic relationship. Smoking, drinking, and inadequate gestational weight gain were also associated with IUGR.

Table 2.

Association Between Potential Predictors and Intrauterine Growth Restriction

	Crude analysis				Adjusted analysis^a
	Coefficient	p value	95% CI		Coefficient	p value	95% CI
Violence	−1.43	0.397	−4.72	1.87	−1.42	0.409	−4.79	1.95
Violence²	−11.05	0.006	−18.89	−3.21	−10.94	0.007	−18.84	−3.03
Smoking	−0.25	0.009	−0.45	−0.06	−0.25	0.012	−0.44	−0.06
Alcohol	−0.28	0.019	−0.52	−0.05	−0.26	0.033	−0.50	−0.02
Anxiety	−0.12	0.180	−0.30	0.06	−0.15	0.112	−0.33	0.03
Depression	−0.10	0.323	−0.31	0.10	−0.11	0.317	−0.31	0.10
Antibiotic Medication	−0.03	0.810	−0.25	0.19	−0.05	0.639	−0.27	0.17
Changes in BP	−0.08	0.364	−0.27	0.10	−0.11	0.241	−0.29	0.07
PN appointments	−0.20	0.134	−0.47	0.06	−0.19	0.171	−0.45	0.08
Gestational weight gain	0.17	0.002	0.06	0.28	0.16	0.004	0.05	0.27
Emergency care	−0.08	0.267	−0.22	0.06	−0.10	0.189	−0.24	0.05
Bleeding	−0.08	0.373	−0.26	0.10	−0.10	0.287	−0.27	0.08

Analysis adjusted for economic class, maternal education, maternal age, and unintended pregnancy.

95% CI, 95% confidence interval; BP, Blood Pressure; PN, Prenatal; G, gestational. Violence², quadratic term modeling.

Figure 1 presents the results of the path analysis. For evaluation of the model, we found a χ ² value of p of 0.110, an RMSEA of 0.019, a CFI of 0.970, a TLI of 0.913, and an SRMR value of 0.023, constituting good fit overall. Our model shows that violence was associated with being diagnosed with gestational depression or anxiety, smoking and drinking habits, blood pressure changes, and the need for emergency care, but these were not mediators of the effect of the exposure on the outcome. The violence showed a direct effect on IUGR (standardized regression coefficient β = −0.195; p = 0.041) and a total effect (β = −0.276; p = 0.003). Finally, even if separately, the mediator variables were not statistically associated, and the sum of their indirect effects shows a β of −0.081 (p = 0.011).

FIG. 1.

Path analysis for the effects of violence on intrauterine growth restriction. Coefficients for each variable were reciprocally adjusted and adjusted for the confounders (economic class, maternal education, maternal age, and desired pregnancy). GWG, gestational weight gain; Prenatal, number of prenatal care appointments; BP, changes in blood pressure during pregnancy; Emergency, need for emergency care during pregnancy; Bleeding, vaginal bleeding during pregnancy; IUGR, intrauterine growth restriction.

Discussion

Our findings suggest that violence during pregnancy is associated with IUGR. Violence was found to be associated with also changes in behavior (more intense tobacco and alcohol consumption and more frequent search for emergency services) and in health (higher prevalence of depression, anxiety, and perceived alterations in blood pressure). Nevertheless, mediation analysis suggests that the main effect on birthweight does not run through these pathways but through a direct one.

The definition of IUGR is somewhat problematic, and the term is wrongly used interchangeably with SGA.³⁷ IUGR is defined as an abnormal fetal growth pattern that is less than normal for the growth potential of the specific fetus, with possible placental, fetal, and maternal causes,^37,38 whereas an SGA infant is defined as an infant in the 10th percentile of growth for gestational age.³⁹ IUGR is, then, a clinical definition, applied to infants with evidence of malnutrition,⁴⁰ estimated to affect 8% to 10% of all pregnancies,⁴¹ with disproportional impact in the developing world.⁴² In the SGA group, approximately 60% of infants present evidence of growth restriction, whereas 40% constitute constitutively small individuals.⁴³ The diagnosis of IUGR during pregnancy can be made with the assistance of US imaging, with clinical screening being considered unreliable.^39,44 After birth, the ponderal index and the Kramer index can be used to identify growth restriction, with both providing a proportionality estimate and thus suggesting the severity of growth restriction.^24,45

The finding of a strong direct effect should draw attention to the negative consequences that violence during pregnancy can have on the newborn. IUGR is related to several perinatal and late-life outcomes, such as neurodevelopmental impairment, metabolic disorders, and chronic lung disease.⁴ In our analyses, we included what we consider the most important possible mediators of this relationship, covering different types of indirect pathways, and even so, we found a profound direct impact of violence on growth restriction. This suggests that there are biological mechanisms that account for these effects.

Changes in cortisol circulation and regulation resulting from the maternal stress response, associated with augmented sympathetic activity, can restrict uterine blood flow, and impair the supply of oxygen and nutrients to the developing fetus.^46,47 Prenatal stress may also reduce the activity of the 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) enzyme, responsible for converting cortisol to inactive cortisone,⁴⁸ thereby enhancing fetal corticoid exposure. These disruptions to the hypothalamic–pituitary–adrenal (HPA) axis may have important consequences for fetal HPA axis regulation and may lead to impaired fetal growth. A chronic hypoxic environment may also induce a proinflammatory state, which in turn may influence fetal development by impairing cellular differentiation and communication.⁴⁹

Mediation analyses offer important insights into the causal structures of associations, allowing researchers to identify possibilities for real-world interventions and gaps in scientific understanding.⁵⁰ In recent years, the number of available methods for mediation studies has increased, but they are still not widely employed, probably due to the complexities related to their application.⁵¹,⁵² An important aspect of causal mediation analysis that may also function as a barrier to its application is the fact that it depends heavily on the assumptions made by the researcher (e.g., which mediators are chosen or which confounders should be included), and so the evaluation of these models is not always straightforward.⁵³

Despite an overall significant effect of the indirect pathways from violence to IUGR, no included pathway was statistically significant on its own. Six of the chosen mediators were related only to the exposure (emergency medicine care, perceived changes in blood pressure, alcohol or tobacco consumption, anxiety, and depression), gestational weight gain was associated only with the outcome and vaginal bleeding during pregnancy, number of prenatal appointments, and infection (using antibiotic medication use as a proxy) was not related to either exposure or outcome. Considering the results of the association models shown in Table 2, these findings highlight the importance of the Structural Equation Modelling (SEM) methodology and how those variables are related to each other. The path analysis, by including the designated regressions all at the same time, may help identify associations when considering models that involve complex relationships between the included variables.

Our findings of a quadratic association between violence during pregnancy and IUGR suggest a strong impact of violence on fetal development. No previous studies have reported a quadratic association between violence and perinatal outcomes, which may suggest that this possibility is overlooked when a linear relationship is not found. We urge caution in evaluating the relationships between prenatal stress and defined outcomes, as given the complexities of fetal development, a linear relationship cannot always be reasonably expected.

The strengths of this study include the sampling process, with the active search for pregnant women in the regions covered by the health units possibly reducing selection bias related to better health behaviors leading to a bigger chance of inclusion in the study. All the variables included in this study were collected by trained psychologists and based on cross-culturally validated instruments, which adds more reliability to the results shown and allows for dialogue with other studies. In our mediation analyses, we included all the possible mediators recognized in the literature, covering an important range of indirect pathways.

However, there were also limitations. US examination for gestational age identification was available for a only little more than half of our sample, but we used two methods to estimating gestational age for the remaining women to reduce possible errors. Loss to follow-up could also be a limitation of our study, but loss rates remained at the expected level and no attrition bias or significant differences between the followed and lost groups were identified in our analyses. Additionally, neither the exact time of the occurrence of violence was determined nor were the subjects grouped by trimester of exposure. Considering the developmental windows in human pregnancy, violence occurring in the first trimester may have different and more pronounced effects when compared with violence occurring in the second or third trimester. Thus, not differentiating between exposure time could have undermined our power to detect these associations. Because our analyses provided positive results, we believe including timing of exposure would make it possible to better discern the effects of each time window, but our results remain valid without this information and point directions to further research.

Conclusions

Violence during pregnancy is associated with IUGR, a pathological condition with dire consequences for the offspring’s immediate and long-term health and survival. The primary mechanism underlying this association is a direct pathway, although many mediation pathways have previously been identified. The existence of a direct pathway and the robust strength of this association underscores the critical importance of implementing targeted policies specifically focused on addressing gender-based violence, for both the health of the mother and the next generation.

Footnotes

Acknowledgments

The authors would like to acknowledge all mothers and children who decided to take part in this study. The authors also would like to acknowledge the field workers and central staff.

Authors’ Contributions

L.B.: Conceptualization, Writing—Original Draft and Review; L.M.S.: Investigation, Data Curation; V.D.B.: Investigation, Data Curation; L.A.R.: Conceptualization, Investigation, Resources; G.V.P.: Research Administration, Investigation, Resources; E.C.M.: Conceptualization, Investigation, Fundraising; S.J.F.E.G.: Conceptualization, Investigation, Fundraising; B.F.-B.: Conceptualization, Investigation, Resources; A.A.F.: Conceptualization, Data Analysis, Writing—Original Draft and Review, Supervision.

Data Availability

All data are accessible under request by emailing ferraro@usp.br.

Author Disclosure Statement

L.A.R. has received grant or research support from, served as a consultant to, and served on the speakers’ bureau of Abdi Ibrahim, Abbott, Aché, Adium, Apsen, Bial, Knight Therapeutics, Medice, Novartis/Sandoz, Pfizer/Upjohn/Viatris, and Shire/Takeda in the past three years. The ADHD and Juvenile Bipolar Disorder Outpatient Programs chaired by L.A.R. have received unrestricted educational and research support from the following pharmaceutical companies in the past three years: Novartis/Sandoz and Shire/Takeda. L.A.R. has received authorship royalties from Oxford Press and ArtMed. The other authors declare that they have no conflicts of interest.

Funding Information

This study was funded by the CNPQ (National Council for Scientific and Technological Development) and FAPESP (Foundation for the Support of Research of the State of Sao Paulo, Brazil). CNPQ-National Council for Scientific and Technological Development (Grant 465550/2014-2); FAPESP—Foundation for the Support of Research of the State of Sao Paulo, Brazil (Grant 2014/50917-0).

Supplementary Material

Annex I

References

WHO, HRP. Intimate partner and sexual violence against women-Evidence Brief. 2019. WHO/RHR/19.16.

Román-Gálvez

, Martín-Peláez

, Fernández-Félix

, et al. Worldwide prevalence of intimate partner violence in pregnancy. A systematic review and meta-analysis. Front Public Health, 2021; 9:738459.

Pastor-Moreno

, Ruiz-Pérez

, Henarez-Montiel

, et al. Intimate partner violence and perinatal health: a systematic review. BJOG: An International Journal of Obstetrics, 2020; 127:537–547.

Armengaud

, Yzydorczyk

, Siddeek

, et al. Intrauterine growth restriction: Clinical consequences on health and disease at adulthood. Reprod Toxicol, 2021; 99:168–176.

Chehade

, Simeoni

, Guignard

, et al. Preterm birth: Long term cardiovascular and renal consequences. Curr Pediatr Rev, 2018; 14(4):219–226.

Eriksson

. Developmental origins of health and disease-67 from a small body size at birth to epigenetics. Ann Med, 2016; 48(6):456–467.

, Flatley

, Greer

, et al. Birth-weight centiles and the risk of serious adverse neonatal outcomes at term. J Perinat Med, 2018; 46(9):1048–1056.

Ashorn

, Ashorn

, Muthiani

, et al. Small vulnerable newborns-big potential for impact. Lancet, 2023; 401(10389):1692–1706.

Murphy

, Schei

, Myhr

, et al. Abuse: A risk factor for low birth weight? A systematic review and meta-analysis. CMAJ, 2001; 164(11):1567–1572.

10.

Shah

, Shah

, Knowledge Synthesis Group on Determinants of Preterm/LBW Births . Maternal exposure to domestic violence and pregnancy and birth outcomes: A systematic review and meta-analyses. J Womens Health, 2010; 19(11):2017–2031.

11.

Donovan

, Spracklen

, Schweizer

, et al. Intimate partner violence during pregnancy and the risk for adverse infant outcomes: A systematic review and meta-analysis. BJOG, 2016; 123(8):1289–1299.

12.

Hill

, Pallitto

, McCleary-Sills

, et al. A systematic review and meta-analysis of intimate partner violence during pregnancy and selected birth outcomes. Int J Gynaecol Obstet, 2016; 133(3):269–276.

13.

Moreira

, Pinto da Costa

. The impact of the Covid-19 pandemic in the precipitation of intimate partner violence. Int J Law Psychiatry, 2020; 71:101606.

14.

Ogunsiji

, Clisdell

. Intimate partner violence prevention and reduction: A review of literature. Health Care Women Int, 2017; 38(5):439–462.

15.

Chisholm

, Bullock

, Ferguson

JEJ

. Intimate partner violence and pregnancy: Epidemiology and impact. Am J Obstet Gynecol, 2017; 217(2):141–144.

16.

Alhusen

, Ray

, Sharps

, et al. Intimate partner violence during pregnancy: Maternal and neonatal outcomes. J Womens Health (Larchmt), 2015; 24(1):100–106.

17.

Sugg

. Intimate partner violence: Prevalence, health consequences, and intervention. Med Clin North Am, 2015; 99(3):629–649.

18.

Ferguson

JEJ

. Why doesn’t SOMEBODY do something? Am J Obstet Gynecol, 2010; 202(6):635–643.

19.

Baird

, Creedy

, Mitchell

. Intimate partner violence and pregnancy intentions: A qualitative study. J Clin Nurs, 2017; 26(15-16):2399–2408.

20.

Flaathen

EME

, Lukasse

, Cvancarova Småstuen

, et al. Intimate partner violence and the association of pregnancy intendedness - A cross-sectional study in southeastern Norway. Sex Reprod Healthc, 2021; 29:100651.

21.

Bailey

. Partner violence during pregnancy: Prevalence, effects, screening, and management. Int J Womens Health, 2010; 2:183–197.

22.

Yount

, DiGirolamo

, Ramakrishnan

. Impacts of domestic violence on child growth and nutrition: A conceptual review of the pathways of influence. Soc Sci Med, 2011; 72(9):1534–1554.

23.

Ferraro

, Rohde

, Polanczyk

, et al. The specific and combined role of domestic violence and mental health disorders during pregnancy on new-born health. BMC Pregnancy Childbirth, 2017; 17(1):257.

24.

Kramer

, McLean

, Olivier

, et al. Body proportionality and head and length “sparing” in growth-retarded neonates: A critical reappraisal. Pediatrics, 1989; 84(4):717–723.

25.

García-Moreno

, Jansen

, Ellsberg

, et al. WHO Multi-Country Study on Women’s Health and Domestic Violence against Women. World Health Organization; 2005.

26.

Schraiber

, Latorre

MdR

, França

Jr. , et al. Validity of the WHO VAW study instrument for estimating gender-based violence against women. Rev Saude Publica, 2010; 44(4):658–666.

27.

World Health Organization. Putting Women First: Ethical and Safety Recommendations for Research on Domestic Violence against Women. World Health Organization; 2001.

28.

Sheehan

, Lecrubier

, Sheehan

, et al. The Mini-International Neuropsychiatric Interview (MINI): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry, 1998; 59(20):22–33.

29.

Amorim

. Mini International Neuropsychiatric Interview (MINI): validação de entrevista breve para diagnóstico de transtornos mentais. Rev Bras Psiquiatr, 2000; 22(3):106–115.

30.

Brasil

. Critério de classificação econômica Brasil. Associação Brasileira de Empresas de Pesquisa (ABEP); 2008.

31.

Richiardi

, Bellocco

, Zugna

. Mediation analysis in epidemiology: Methods, interpretation and bias. Int J Epidemiol, 2013; 42(5):1511–1519.

32.

Textor

, Hardt

, Knüppel

. DAGitty: A graphical tool for analyzing causal diagrams. Epidemiology, 2011; 22(5):745.

33.

StataCorp. Stata Statistical Software: Release 14. StataCorp LP: College Station, TX; 2015.

34.

Muthén

, Muthén

. MPlus User’s Guide. 8th Edition. Muthén & Muthén: Los Angeles, CA; 1998.

35.

Wang

, Wang

. Structural Equation Modeling: Applications Using Mplus. John Wiley & Sons Ltd.; 2012.

36.

Kline

. Principles and Practice of Structural Equation Modeling. Guilford Press: New York, NY; 2016.

37.

Sharma

, Shastri

, Farahbakhsh

, et al. Intrauterine growth restriction-part 1. J Matern Fetal Neonatal Med, 2016; 29(24):3977–3987.

38.

Sacchi

, Marino

, Nosarti

, et al. Association of intrauterine growth restriction and small for gestational age status with childhood cognitive outcomes: A systematic review and meta-analysis. JAMA Pediatr, 2020; 174(8):772–781.

39.

Gordijn

, Beune

, Thilaganathan

, et al. Consensus definition of fetal growth restriction: A Delphi procedure. Ultrasound Obstet Gynecol, 2016; 48(3):333–339.

40.

Sharma

, Shastri

, Sharma

. Intrauterine growth restriction: Antenatal and postnatal aspects. Clin Med Insights Pediatr, 2016; 10:67–83.

41.

Alkalay

, Graham

, Pomerance

. Evaluation of neonates born with intrauterine growth retardation: Review and practice guidelines. J Perinatol, 1998; 18(2):142–151.

42.

Black

, Victora

, Walker

, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet, 2013; 382(9890):427–451.

43.

Figueras

, Gratacos

. An integrated approach to fetal growth restriction. Best Pract Res Clin Obstet Gynaecol, 2017; 38:48–58.

44.

Lausman

, Kingdom

; Maternal Fetal Medicine Committee . Intrauterine growth restriction: Screening, diagnosis, and management. J Obstet Gynaecol Can, 2013; 35(8):741–748.

45.

Landmann

, Reiss

, Misselwitz

, et al. Ponderal index for discrimination between symmetric and asymmetric growth restriction: Percentiles for neonates from 30 weeks to 43 weeks of gestation. J Matern Fetal Neonatal Med, 2006; 19(3):157–160.

46.

Lewis

, Austin

, Galbally

. Prenatal maternal mental health and fetal growth restriction: A systematic review. J Dev Orig Health Dis, 2016; 7(4):416–428.

47.

, Zhou

, Ge

, et al. Differential expression of placental 11β-HSD2 induced by high maternal glucocorticoid exposure mediates sex differences in placental and fetal development. Sci Total Environ, 2022; 827:154396.

48.

Jahnke

, Terán

, Murgueitio

, et al. Maternal stress, placental 11β-hydroxysteroid dehydrogenase type 2, and infant HPA axis development in humans: Psychosocial and physiological pathways. Placenta, 2021; 104:179–187.

49.

Wixey

, Chand

, Colditz

, et al. Review: Neuroinflammation in intrauterine growth restriction. Placenta, 2017; 54:117–124.

50.

Farland

, Correia

KFB

, Dodge

, et al. The importance of mediation in reproductive health studies. Hum Reprod, 2020; 35(6):1262–1266.

51.

Lapointe-Shaw

, Bouck

, Howell

, et al. Mediation analysis with a time-to-event outcome: A review of use and reporting in healthcare research. BMC Med Res Methodol, 2018; 18(1):118.

52.

Rijnhart

JJM

, Lamp

, Valente

, et al. Mediation analysis methods used in observational research: A scoping review and recommendations. BMC Med Res Methodol, 2021; 21(1):226.

53.

Agler

, De Boeck

. On the interpretation and use of mediation: Multiple perspectives on mediation analysis. Front Psychol, 2017; 8:1984.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.44 MB