Introduction

Childhood (defined as the time between infancy to the onset of puberty) and adolescence (between childhood and culturally/socially defined adulthood) are sensitive periods of growth and development that lay the foundation for later health status and health behaviors [1–3]. Adverse childhood experiences (ACEs) are traumatic events that occur during childhood and adolescence and undermine a child’s physical, social, and emotional development, learning, and the sense of safety and stability [4]. ACEs can have a profound effect on later development. Examples of ACEs include witnessing or experiencing abuse or violence (such as childhood sexual abuse [CSA]), the incarceration of a parent/guardian, and having parents/guardians with substance use or mental health problems [5]. Exposure to ACEs can contribute to harmful levels of stress hormones that can negatively affect brain development and immune responses [6–8]. In particular, ACEs have been studied as primary exposures contributing to adult morbidity and mortality [4]. A rich body of longitudinal research has investigated relationship between ACEs and adult health. ACEs are directly associated with a greater likelihood of chronic diseases (type 2 diabetes and cardiovascular disease), mental health outcomes (depression and attempted suicide), sexually transmitted infections, and substance abuse ranging from nicotine to intravenous drugs [9–17].

Interpersonal violence is the intentional use of physical force or power against another person to enact physical, sexual, psychological harm, including neglect or deprivation [18]. Recent estimates indicate that interpersonal violence is the fifth leading cause of death among youth 10–24 years old, representing 3.5% of disability adjusted life years globally [19]. However, the distribution of interpersonal violence and its forms is not random. In particular, girls are disproportionately affected by gender-based violence (GBV) [20], such as dating violence/intimate partner violence (IPV) (inclusive of physical, emotional, sexual, economic and controlling behaviors); GBV is defined as the intentional or threatened use of power or physical force, against another person or group of people based on their perceived gender or gender identity [20].

In addition to dating violence/IPV, GBV also includes CSA, child marriage, female genital cutting, and marital rape [21]. GBV can be perpetrated by intimate partners, family members, peers, and authority figures, and is underscored by a gender power hierarchy, wherein “masculinity” is privileged over “femininity” and cis-gender identities are presumed to be natural and superior in comparison to trans-identities [22]. Globally, 10% of females under the age of 20 have been forced to perform sexual acts [23], while 25% of young women who have been in relationships experience IPV by their mid-twenties [24, 25]. In sub-Saharan Africa, the prevalence of child marriage (defined as informal unions/marriages between a child under the age of 18 and another child or adult) is estimated to be 54.0%, ranging from 16.5% to 81.7% [26].

GBV perpetrated against females during childhood and adolescence is a specific ACE subtype (although not the only form of ACE) that can affect growth, development, and health [27]. Studies of ACEs have improved the knowledge base concerning the health impacts of GBV exposure during childhood. However, ACE studies don’t disentangle GBV from other ACEs and sometimes lack sex disaggregation. Further, the majority of research exploring the health sequala of GBV focuses on violence against women specifically and examines adult morbidity specifically [28–30] rather than focusing on exposures and outcomes that can be measured during childhood and adolescence. Further, even when CSA is the primary interest variable, health outcomes are often measured in adulthood. Overall, there is less research exploring the childhood and adolescent health consequences of GBV perpetrated against females during childhood and adolescence.

Nutrition, a critical contributor to health and wellbeing, is defined as the process of taking in food for growth, metabolism, and repair [31]. Moreover, childhood nutritional status can impact health outcomes across the life course, with micro nutrient insufficiencies such as iron or calcium contributing to cognitive impairments and osteopenia later in life [32]. Research examining the link between GBV exposure and nutrition has predominantly focused on forms of GBV perpetrated against adult women during and immediately after pregnancy [33–35]. For example, IPV victimization during pregnancy and lactation negatively affects breastfeeding practices and has been linked to premature birth and labor and low birth weight, known risk factors for later childhood low height and weight for age [36]. The evidence base exploring the impact of GBV on childhood nutrition outcomes has focused on GBV perpetrated against mothers and nutrition outcomes among their children [36–38] (Authors, under review). Less understood are the childhood and adolescent nutritional consequences of GBV among girls who have themselves experienced GBV during childhood and adolescence.

Emerging research indicates that some nutrition-related disorders of adulthood may begin to manifest during childhood and adolescence and that exposure to GBV may contribute to the development of these disorders. Research has postulated that exposure to GBV can result in trauma and stress-related physiological processes that can affect metabolism and appetite, the development of eating disorders, coping mechanisms that influence nutrition and eating behaviors, and weight gain [39–42]. Specifically, evidence indicates that disordered eating patterns and overweight/obesity among survivors of CSA and IPV/dating violence emerge in adolescence and have enduring effects in adulthood [43, 44]. Also, forms of GBV affecting girls, such as child marriage, the preferential feeding of men and boys, and controlling behaviors, may directly and indirectly impact determinants of nutrition—namely, access to nutrient-rich foods and nutrition-related services—potentially contributing to malnutrition in adolescence [45–48]. That we are aware, there is no existing evidence synthesis focused on summarizing the current body of knowledge surrounding the childhood and adolescence nutritional consequences of GBV.

Methods

The present review is part of a larger REA commissioned by the United Nations Children’s Fund (UNICEF). The larger REA had two research objectives: (1) Review and synthesize quantitative evidence describing the associations between GBV against girls and indicators of girls’ nutrition outcomes (Direct GBV exposure) and (2) Review and synthesize the quantitative evidence describing the associations between IPV against a maternal caregiver and children’s nutrition outcomes, with a particular focus on humanitarian contexts (Indirect GBV exposure). The present paper is focused specifically on the first objective of the larger REA, pertaining to the direct pathway. A separate publication details the indirect pathway (Authors, under review). There is no published review protocol. However, the protocol for the larger REA is presented in Appendix B of S2 Appendix.

We adapted the Preferred Reporting Items for Systematic Reviews and Meta Analysis standards (PRISMA) [49] for the purposes of conducting a REA. REAs adapt systematic review methodology to synthesize a body of literature, ensuring rapid completion of the review and applicability to policy and programmatic needs in a short timeframe [50]. The following changes to systematic review methodology were made: (i) title/ abstract screening was conducted by one of three reviewers and (ii) data extraction and quality assessment were conducted by a one or two team members per article, depending on whether the article warranted discussion. Furthermore, while inclusion and exclusion criteria were determined a-priori, the team iteratively refined the criteria based on the volume of sources and the overall research goal. We conducted the REA in 7 stages, which we describe below: (1) Preliminary scoping of the literature to clarify the research question and search strategy, (2) Search strategy execution, (3) Title and abstract screening, (4) Full text review, (5) Quality Assessment, (6) Data extraction, and (7) Synthesis.

As a preliminary step, we conducted a preliminary scoping of existing reviews and literature on the associations between girls’ direct exposure to GBV during childhood and nutrition outcomes. We searched Google Scholar, Cochrane Reviews and PubMed for reviews concerning direct childhood exposures to GBV and subsequent nutrition-related outcomes. Selected reviews and literature were examined to elucidate specific direct pathways, select forms of GBV to include in the review, and to help operationalize database search terms.

Our preliminary scoping review indicated a general lack of robust research synthesis detailing the effect of girls’ exposure to GBV on adverse nutrition outcomes during childhood and adolescence. In our preliminary scoping review, we identified the following forms of GBV experienced during childhood that are potentially related to later nutrition deficits among females: preferential feeding of boys, gendered child abuse (i.e., CSA), child marriage, and IPV/dating violence.

Eligibility criteria

Eligibility criteria were defined a-priori and then were iteratively refined to better reflect the existing body of literature and address the research question. Inclusion and exclusion criteria are displayed in Table 1. Two adjustments to the a-priori eligibility criteria were made. First, we originally wanted to focus on empirical evidence conducted in humanitarian settings. However, due to the limited number studies detailing girls’ direct exposure to GBV and nutrition related outcomes in humanitarian settings, we expanded the eligibility criteria to include studies conducted in any setting or country.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion criteria.

https://doi.org/10.1371/journal.pone.0281961.t001

Second, we originally conceptualized the population of interest to be females up to the age of 18. However, we noted that sensitive periods of overweight development following sexual abuse or IPV occur between adolescence and young adulthood. In our title and abstract screening of the direct pathway studies, we found several longitudinal studies also measured nutrition outcomes into adulthood. In an effort not to erroneously exclude samples of children followed into adulthood, and to be consistent with the WHO definition of adolescence which extends through age 19 years, we included nutrition outcomes also measured among girls who were up to 20 years old. Although many sources recruited samples that contained both males and females, all sources included in this REA provided sex-stratified measures of association. We used these to elucidate the association between exposure to gender-based violence and nutrition indicators in females during later childhood and adolescence.

Quality assessment and data extraction

Given the diverse quantitative study designs included, no single quality assessment tool was deemed appropriate. Accordingly, the team combined relevant quality assessment questions from the following quality assessment tools: National Institutes of Health (NIH) [51], Mixed Methods Assessment Tool (MMAT) [52], Appraisal Tool for Cross Sectional Studies (AXIS) [53], Newcastle-Ottawa Scale (NOS) [54], and the National Institutes of Health assessment tool for observational cohort and cross sectional studies [55]. The items included in the quality assessment are presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Quality assessment domains.

https://doi.org/10.1371/journal.pone.0281961.t002

The team designed a data extraction template in accordance with the research goals. Specifically, data were extracted according to general study characteristics, elements of the study design and sampling, violence exposure measurement, nutrition outcome measurement, and summary of results. The data extraction template was pilot tested for further refinement. Team members extracted information independently and extraction was reviewed by a senior member of the research team.

Results

The PRISMA flow diagram is presented in Fig 1. After executing the database search strategies, 6493 sources were imported to Covidence and 1516 duplicates were removed, yielding 4977 citations for screening. During title and abstract screening, 4461 sources were excluded, and 468 full texts were assessed for eligibility. During full text review, 450 sources were excluded for various reasons outlined in Fig 1. In total, only 18 represent the direct pathway and are included in this paper.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow chart.

https://doi.org/10.1371/journal.pone.0281961.g001

Study characteristics

Table 3 displays general characteristics for the 18 studies examining pathways between girls’ GBV exposure and indicators of nutrition outcomes during childhood and adolescence. Although more than half of studies were conducted in high-income settings, several studies reported outcomes in low-and-middle income countries. The specific countries included were Ethiopia (n = 1), Nepal (n = 1), Brazil (n = 2), Mexico (n = 1), as well as United States (n = 8), England (n = 2), and the Netherlands (n = 2), with one study combining data from the United Kingdom and Brazil (n = 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Study characteristics.

https://doi.org/10.1371/journal.pone.0281961.t003

Most sources sought to quantify associations between specific forms of sexual violence affecting children, including girls, and indicators of subsequent nutrition status in childhood and adolescence. Some studies conducted analyses that permit more nuanced understandings of causal mechanisms including use of community-based comparator groups [56–59]; moderation by cortisol reactivity [57]; moderation by HPA axis attenuation [41]; interaction of caregiver abuse and IPV/dating violence [60]; interaction of child marriage and age at first pregnancy [61]; and stratification by age [62] and developmental period [56].

Quality assessment

In Table 4 we present the quality assessment results. Key gaps in quality were due to high attrition/non-response, poor measurement of GBV, and limited external validity/generalizability due to the sampling of selected groups. All studies (100%) had clearly stated research questions and objectives. Studies varied in terms of external validity and representativeness, with 55.6% of studies being truly representative of the target population and 38.9% representing a selected group, for example, cases of child abuse reported to Child Protective Services. Less than half of studies (38.9%) achieved satisfactory response rates (meaning less than 10% attrition/non-response) and/or comparability between respondents and non-respondents. The measurement of girls’ exposure to gender-based violence was deemed rigorous in half of the included studies (50.0%) and the measurement of childhood nutrition-related outcomes was appropriately rigorous in 83.3% of the included studies. Rigorous measurement of GBV and nutrition was assessed using a variety of indicators including, whether the variables were clearly defined, accurately measured and applied consistently, measurements were justified and appropriate for answering the research question, measurements reflected what they are supposed to measure, measures were collected using validated and reliability tested tools/instruments/questions, variables were measured using ‘gold standard’ tools, and questionnaires were pre-tested prior to data collection. Two thirds (66.7%) of studies adequately controlled for confounding variables. Lastly, 22.2% of studies did not properly conduct or report on their statistical analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Quality assessment.

https://doi.org/10.1371/journal.pone.0281961.t004

Sampling

Table 5 presents participant recruitment and sample details. Gender-stratified samples ranged from under 100 females [56] to over 26,000 [63]. The range of sample sizes is a reflection of the study designs utilized; studies with smaller sample sizes focused on newly reported cases of CSA obtained from Child Protective Services [41, 56–59], or other similar organizations such as a clinic for victims of child abuse [64], and larger sample sizes pertained to secondary data analyses of nationally representative and/or longitudinal datasets of child and adolescent health [60, 62, 63, 65–71]. Sampling of CSA cases from Child Protective Services is a valid method for capturing this form of GBV given that cases of genital contact and/or perpetration by family members are investigated and substantiated according to national laws. In such studies, the survivor children and a non-abusing caregiver(s) are referred to the study within a specific time frame (i.e., 6 months) of the initial abuse disclosure. However, sampling of CSA cases from Child Protective Services may not represent the entire population of CSA cases given that only a proportion of children who experience this form of abuse formally enter the child protection system. Socio-economic status and racialization experiences (as well as other social determinants of health) may differ between CSA cases that formally enter protection systems and cases that are unreported, thereby potentially introducing sampling bias and limiting generalizability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Participant recruitment and sampling.

https://doi.org/10.1371/journal.pone.0281961.t005

The majority of studies included samples of adolescent participants or samples in which participants were adolescents/young adults at final data collection; although Martin-Martin et al. [64] specifically focused on children ages 0–11 years. Participant recruitment strategies varied across the sources and included both probability (n = 11) and non-probability samples (n = 7). Two predominant study designs were utilized: (i) longitudinal—including panel, prospective cohort, and cross cohort comparison designs—(n = 12) and (ii) cross sectional (n = 6). Thirteen studies conducted secondary analysis on existing datasets.

Violence exposures

The following forms of GBV were measured: child marriage [61, 68], sexual assault and rape [63, 65, 66], CSA perpetrated by parents/caregivers [41, 56–59, 62, 64, 67, 69–71], and dating violence/IPV [60] (Table 6). Some studies did not use validated survey instruments to measure GBV [63, 65–67, 70, 71], and in some cases relied on a single question to assess girls’ GBV exposure [63, 66, 67]. Clark et al., (2014) [60] employed the Conflict and Tactics Scale, which is a widely used and validated instrument. Martin-Martin et al., (2010) [64] used clinical records to document exposure to sexual abuse.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Violence exposure.

https://doi.org/10.1371/journal.pone.0281961.t006

Nutrition indicators

Nutrition indicators relied on anthropometric measures that were either observed or self-reported (Table 7). Body mass index (BMI) was the most used nutrition indicator included in the studies of older children and adolescents. In addition, other nutrition indicators that were assessed included height for age, weight for height, height for younger children, leg length, waist circumference, fat mass index, android fat percentage, mid-upper arm circumference (MUAC), and anemia among the older children and adolescents. Height and weight were assessed using either anthropometric measurements [41, 56–61, 64, 66, 67, 69–71] or self-reported height and weight (which may be affected by recall error or bias based on context dependent body ideals) [63, 69, 72]. Either weight for height or BMI were calculated as appropriate according to participants’ ages and operationalized for analysis using different methods [73, 74].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Nutrition outcomes.

https://doi.org/10.1371/journal.pone.0281961.t007

Height and leg length were assessed longitudinally throughout childhood and adolescence in Denholm et al., [62]. Multiple measures of adiposity were measured by Soares, et al., [67] using anthropometric measurements and dual-energy x-ray absorptiometry scans in adolescents. ‘Anemia’ during adolescence was measured in only one study using a point of care diagnostic lab test for hemoglobin and operationalized using WHO defined thresholds and presuming iron deficiency without documenting iron/ferritin levels [68].

Synthesis of findings

The 18 included studies addressed three pathways describing associations and potential mechanisms between exposure to GBV during childhood/adolescence and nutrition indicators. Study results are displayed in Table 8. Overall, findings indicate that sexual abuse (for example, CSA, IPV, dating violence) experienced during childhood and adolescence was positively and significantly associated with elevated BMI, overweight, obesity, and adiposity, particularly when studied longitudinally. We found more limited evidence with respect to the relationship between child marriage and undernutrition (anemia, BMI, MUAC, and chronic energy deficiency). We outline the three pathways below, with a focus on summarizing results and describing important methodological implications.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Results.

https://doi.org/10.1371/journal.pone.0281961.t008

Discussion

We conducted a REA to synthesize quantitative evidence describing associations between girls’ exposure to GBV and indicators of nutrition status during childhood and adolescence. Overall, the evidence base is limited: only 18 studies met our inclusion criteria and most studies examined either CSA or IPV/dating violence with later overweight/obesity/measures of adiposity. Our review indicates the lack of studies investigating the potential impacts of exposure to GBV on indicators of childhood and adolescence nutrition status among females; in particular, nutrition outcomes beyond those calculable with measures total body height and weight, indicators of key nutrient levels (including iron deficiency), or more sensitive indicators of growth during specific developmental periods. Beyond CSA, dating violence/IPV, and child marriage, no other forms of GBV were studied in relation to girls’ nutrition indicators. Another form of GBV that warrants empirical attention is the preferential feeding of boys which could be relevant in humanitarian/fragile settings where there may be high levels of food insecurity [46, 76–78]. Expansion of this evidence base in LMIC is also needed, with appropriate study designs and analytical procedures that account for high baseline levels of malnutrition and multi-level risk factors for malnutrition.

Our synthesis elucidated three potential pathways linking exposure to GBV and indicators of subsequent childhood and adolescence nutrition status among females:

1. Exposure to sexual abuse by a) parents/caregivers during childhood (CSA) and later elevated BMI (i.e., overweight, obesity) or increased adiposity, b) IPV/sexual violence during adolescence and later elevated BMI (i.e., overweight, obesity) or increased adiposity. The evidence base most clearly and consistently supported this pathway, given most studies analyzed sexual violence and BMI/obesity/overweight/increased adiposity outcomes.
2. Exposure to child marriage and undernutrition (amenia, chronic energy deficiency, and MUAC). There is emerging evidence supporting this pathway. However, additional research is needed to replicate the findings of the two studies investigating child marriage and undernutrition. Importantly, age at first pregnancy appears to be a consequence of girl child marriage or an impetus for girl child marriage that has negative nutritional consequences for adolescent females.
3. Exposure to childhood abuse (including sexual) and reduced height and leg length. The evidence base is inconclusive regarding this pathway, given only one study with insignificant results was found.

Among the included studies, the strongest and most consistent evidence supports the relationship between elevated BMI in late adolescence and early adulthood and exposure to sexual abuse during childhood (CSA) or adolescence (intimate partner violence/sexual assault/rape). Childhood and adolescent exposure to sexual violence, perpetrated by parents/caregivers and intimate partners, is an important predictor of nutrition status indicators in childhood and adolescence among females. Understanding the biological mechanisms that underlie this association may help guide further research regarding the consideration of intermediary variables, timing of causal effects, prevention efforts, and actionable periods of intervention.

CSA and exposure to IPV in adolescence are examples of ACEs, however health impacts are largely measured in adulthood and nutrition outcomes are less considered compared to chronic disease, mental health, and substance use. This is an important research gap given that findings from this REA indicate nutrition outcomes such as adolescent obesity are associated CSA, IPV, and sexual assault/rape exposure during childhood and adolescence. To further understand mechanisms that link GBV exposure in childhood/adolescence and the early emergence of nutrition disorders, additional research that considers childhood and adolescence nutrition outcomes is needed.

Our findings highlight that sexual violence, either perpetrated by parents/caregivers during early or middle childhood or by intimate partners or strangers later on in adolescence, is a stress-inducing form of trauma which engenders nutrition-related consequences in childhood and adolescence (as seen in this review), and into adulthood [15]. Previous studies have linked ACEs, including CSA and sexual violence, to stress response dysregulation [41, 79–81]. Like other ACEs, the biological mechanism linking sexual violence to elevated BMI likely involves the hypothalamic-pituitary-adrenocortical (HPA) axis (although, other pathways may also be involved). The HPA axis modulates the stress response by regulating the secretion of cortisol in response to environmental conditions perceived as being stressful and beyond an individual’s capacity to cope [7, 82, 83]. Exposure to sexual violence in early childhood and adolescence can serve to activate the HPA-axis in a manner that can negatively affect health outcomes related to nutrition, mental health, and immunity [79]. Long term or chronic activation of the HPA-axis from exposure to GBV and other adversities results in an attenuated diurnal decrease in cortisol or dampened/attenuated cortisol reactivity [57, 82], thus chronic exposures to GBV may be especially harmful.

Higher values of initial cortisol levels followed by a lowered rate of cortisol increase is characteristic of HPA axis dysregulation. Dampened or attenuated cortisol reactivity occurs when the HPA axis adapts to prolonged stress-induced cortisol secretion. The downregulation of basal cortisol secretion is a protective mechanism to mitigate the negative consequences of elevated cortisol. For adolescent girls, cortisol dampening/attenuation can negatively interact with normal pubertal physiological maturation, increasing the risk of developing overweight/obesity and adiposity. Biochemical pathways (metabolic disturbances, altered insulin signaling, inflammation induced damage to leptin and insulin receptors), behavioral pathways (increased impulsivity and related overeating), and/or the interaction between biochemical and behavioral pathways may explain how cortisol attenuation/dampening is related to overweight/obesity and adiposity [41, 84–86]. We did not locate any studies that considered eating patterns as intermediary variables between GBV and elevated BMI/overweight/obesity/adiposity, limiting a nuanced understanding of the behavioral pathway and its connection to the biochemical pathway.

Research focusing on childhood GBV has demonstrated that among girls who have experienced CSA, while initial cortisol levels following sexual abuse exposure may be high, there is a developmental transition to lower levels of cortisol in early adulthood via cortisol attenuation [87]. More recently, research supports associations between cortisol attenuation and CSA exposure in later childhood and early to mid-adolescence (ages 6–16) among females, wherein serum basal cortisol levels further decreased longitudinally into adulthood [41]. Cortisol attenuation over time (from childhood to adulthood) was also associated with an accelerated rate of BMI accumulation across development (late childhood, adolescence, and early adulthood) and completely mediated the effect of CSA on accelerated BMI accumulation over development [41]. In the articles we reviewed, exposure to sexual abuse in school aged girls was associated with elevated BMI in late childhood, peri-puberty (ages 6–11), and late adolescence into early adulthood. Importantly, we noted that in Martin and Martin, (2010) [64], Peckins et al., (2019) [57], Elsenburg et al., (2022) [71], and Souza Marques et al., (2022) [66], such relationships were only noted among girls, thereby suggesting the importance of conducting a sex-stratified analysis. Other studies have also examined the nutritional impact of adverse childhood life events, including one systematic review focused on the relationship between ACEs more generally and childhood obesity among both boys and girls, which highlighted that sexual abuse appears to have a greater effect on childhood obesity than other ACEs and particularly among girls [43].

The currently synthesized evidence base and the wider literature presents mixed results for sex and/or gender differences in HPA axis attenuation following exposure to chronic stressors and nutrition outcomes involving elevated BMI/obesity/overweight/adiposity. Additionally, there is paucity of tested biological and social mechanisms that account for any sex/gender differences noted. Compared to males, females may be more susceptible to stress-related disorders that impact eating behaviors [88]. Adolescent females also have estrogen-related elevated central adiposity due to hormonally regulated differences in fat distribution, with females developing higher subcutaneous and gluteofemoral fat [89]. In addition, GBV exposure distribution may differ by gender: girls are at greater risk for poly-victimization (cumulative exposure to CSA, sexual IPV/dating violence, and non-partner rape across childhood and adolescence) compared to boys, thereby translating to a dysfunctional stress responses characterized by cortisol attenuation and internalized mental health sequela [90, 91].

Future research investigating the relationship between exposure to CSA and IPV or dating violence among females could consider the inclusion of theoretically relevant intermediary variables. Other potential mechanisms beyond dampened/attenuated cortisol reactivity may also include mental health mediators, given that CSA can lead to mental health disorders and negative affect. For example, ‘emotional overeating’ at age 10 has been associated with exposure to adverse events in younger children [92]. Gaining weight by overeating may also be perceived as a means of deterring unwanted sexual attention/advances among girls previously exposed to CSA [41, 93]. It is important to also consider the role of eating disorders, PTSD and depression as mediator or effect modification variables. Only Elsenburg et al., (2022) considered the mediating role of major depressive disorder and generalized anxiety disorder on the relationship between CSA and BMI among young adults, reporting that CSA is positively associated with major depressive disorder which is in turn positively related to BMI for females but not males [71]. Other research has shown that among survivors of sexual violence, a transition from acute cortisol hyper-reactivity to prolonged hypo-reactivity is characteristic of post-traumatic stress disorder among females (PTSD) [79]. Depression is also of relevance: in a meta-analysis, Danese and Tan (2014) established that when estimates were adjusted for current depression, the relationship between childhood maltreatment (including sexual abuse) and obesity became insignificant [9]. Thus, studies can consider depression as a potential intermediary variable given it is likely a consequence of GBV and can affect nutrition outcomes. Lastly, research could also consider protective factors among children exposed to ACEs, including GBV, and how protective factors such as, quality education, presence of role models, and school/extracurricular engagement, and mental health supports [94, 95], can moderate the development of adverse nutrition outcomes.

We also noted that the effect of childhood exposures to sexual violence on elevated BMI, overweight/obesity/adiposity manifests during sensitive periods of adolescent development into early adulthood. Scholarship has focused mainly on elucidating at what developmental periods childhood maltreatment generally impacts elevated BMI and obesity, with late adolescence/early adulthood emerging as a sensitive period [9, 96]. For example, Sokol et al., (2018) [97] tested childhood exposures to physical, sexual, and emotional abuse and found that these yielded different BMI growth trajectories between the ages of 13 to 28 years for males and females. Importantly neither of these studies included examination of pre-pubertal growth trajectories, thus potential earlier impacts are unknown. In particular, girls’ exposure to sexual abuse between 13 and 17 years was positively associated with BMI in late adolescence to early adulthood, from 19 to 24.5 years of age [96]. A meta-analysis of over 40 studies reported a higher risk of obesity in adulthood, following exposure to child maltreatment, not childhood/adolescence [9]. However, the meta-analysis included few studies that examined exposure to sexual abuse with nutritional indicators measured during childhood and adolescence, as the majority of studies examined nutritional status measured in adulthood [9]. Thus, further research examining exposure to sexual violence with nutrition indicators measured during earlier developmental periods (including prior to and through adolescence into early adulthood) is needed.

Overall, we recommend further clarifying the linkage between GBV and nutrition during childhood and adolescence in females. There is a research-related gap in theorizing and empirically testing mechanisms that aim to explain how and why GBV against girls negatively impacts their nutritional status during later childhood and adolescence. The present REA found evidence to support associations between exposure to sexual violence during childhood and adolescence and elevated BMI/overweight and obesity later in childhood and adolescence. However, nuanced understandings of the causal mechanism underlying this pathway are limited. Given that adolescence represents a period of potential growth catch up and can coincide with pregnancy and child rearing [98], it is important that future studies consider measuring nutritional status among adolescent girls.

From a methodological standpoint, future research should specifically examine exposure to sexual violence when examining nutrition indicators in younger children. Sexual violence exposure can be disaggregated from other ACEs and measured by timing of exposure [infancy, early childhood, middle childhood, or adolescence; or by pubertal stage] and perpetrator type [family member/relative/caregiver, intimate partner, peer, stranger]). Important mediator variables include measures of behavioral changes (internalizing [emotional eating, low self esteem] and externalizing [aggression]), mental health symptoms and disorders (depression, PTSD, eating disorders), impulsivity, and dampened/attenuated cortisol reactivity. Further research can elucidate sensitive periods for anthropometric measures vis-à-vis exposure to sexual violence. Prospective longitudinal collection of informative biomarkers beginning early in childhood among potential at-risk populations will be key for improving our understanding of the relationship between exposure to sexual violence, a dysregulated stress response (including measures of behavior and mental health), and later elevated BMI. Research aiming to investigate the effect of exposure to sexual violence during childhood and adolescence on elevated BMI should consider longitudinal study designs that measure BMI into early adulthood to include sensitive periods of development. Sex-stratification of data could also elucidate sex and gender specific mechanisms, particularly given the sex-based differences in timing of developmental stages. Greater attention should be given to the temporal sequence of variables. For example, measuring mental health symptoms/disorders in parallel with nutrition outcomes can obscure the correct temporal sequence between mental health intermediaries and nutrition outcomes. Accurately capturing cortisol dampening/attenuation requires that investigators measure diurnal serum cortisol variation according to the circadian rhythm.

Our findings should be interpreted in light of some limitations. We conducted single screening of titles and abstracts (a single reviewer determined whether to include or exclude). Although in line with guidance on how to adapt systematic review methodology for an REA [99], single screening of titles and abstracts may have excluded relevant studies. Further, disparate use of outcome measures may mean that results from specific studies are not directly comparable. Lastly, we excluded articles in languages other than English and Spanish due to resources and time-constraints.

Conclusion

This REA identified a small evidence base of 18 quantitative studies investigating the association between GBV perpetrated against girls and subsequent child and adolescent nutrition outcomes. Only two studies investigated the relationship between child marriage and undernutrition, indicating the need for greater attention paid to the nutritional consequences of child marriage on girl brides and the importance of considering age at first pregnancy within the context of girl child marriage. Most included studies investigated the relationship between the outcomes of overweight/obesity/increased adiposity and the exposures of CSA, IPV, and dating violence. We recommend consideration of relevant intermediary variables and timing of effects, pertaining to the biologic mechanisms underlying the relationship between childhood GBV exposure and overweight/obesity. Future research is needed to not only better understand the mechanisms that underlie the relationship between childhood sexual violence exposure and overweight/obesity, but also consider other GBV exposures such as child marriage, the preferential feeding of boys on nutrition outcomes beyond overweight/obesity, and additional research in LMIC, resource poor and humanitarian settings.

Supporting information

Acknowledgments

The authors would like to acknowledge the support received by the project’s research assistant: Luna Johnstone.