Risks to the Unborn: An Umbrella Review on the Effects of Prenatal Maternal Stress Caused by Natural Disasters

1.1. Prenatal Stress and Early Human Development

From an evolutionary perspective, the foetus develops in such a way that it can cope in its future environment as best as possible. The prenatal period is considered to be a developmental phase which is especially sensitive to environmental factors. By being responsive to their mothers' conditions before birth, foetuses are supposed to be prepared for their lives after birth—forecasted by cues available during pregnancy, the so‐called predictive adaptive response (Bateson et al. 2004; Gluckman et al. 2005). The basic assumption is that physiological systems—like the brain in particular—are highly flexible at this early stage meaning that nerve connections are newly formed and rearranged depending on their use. Thus, foetuses are highly receptive to any prenatal cues and the resulting effects can be long‐lasting and far reaching—positive, but also negative ones. Specifically, on the one hand, the prenatal neural plasticity encourages rapid structural changes. Thereby, stress‐related cues during pregnancy strengthen the child's stress response by claiming it already during this adaptive developmental phase. Thus, prenatal stress exposure also results in an increased resilience to stress. However, on the other hand, neural plasticity also goes in line with an increased vulnerability. For example the negative effect of alcohol, cigarette, and drug consumption during pregnancy on the unborn child has been well researched and is no longer controversial (Avşar et al. 2021; Mamluk et al. 2017; Sundermann et al. 2019; Zhang et al. 2022). Moreover, aetiology models suggest that there are pre‐, peri‐, and postnatal influences affecting the mental as well as physical development over a lifespan, indicating that an increased risk of suffering from specific diseases in adulthood is already being set in the womb (Buss et al. 2012).

According to the underlying mechanisms of far‐reaching effects of prenatal influences, various studies demonstrate an effect of prenatal influences on genomic transcription, that is epigenetic effects. In detail, different studies demonstrate an effect of various prenatal influences ‐ such as maternal pregnancy‐related anxiety, maternal prenatal depression, or parental post‐traumatic stress disorder during pregnancy ‐ on the methylation of the promoter region of the glucocorticoid receptor (NR3C1), the receptor for the stress hormone cortisol (Glover et al. 2018). Besides that, although the human body is capable of adapting to stressors, any adaptation and response to acute stress is linked to numerous changes in the body including cardiovascular regulation, energy metabolism, and immune functioning (Ravi et al. 2021; Selye 1976). As the very first living space, the placenta develops specifically and individually for each pregnancy and is the only communication and exchange possibility between the mother and her baby (Wadhwa et al. 2011). Connected by the umbilical cord, mother and baby form one common blood circulation. Hence, if a pregnant woman is stressed, cortisol flows through her bloodstream. Importantly though, 11β‐Hydroxysteroid‐Dehydrogenase type 2 (11ß‐HSD‐2), an enzyme in the placenta, ensures that the mother's cortisol level does not unhindered pass into the child's bloodstream. Approximately 80%–90% of maternal cortisol is inactivated before reaching the child (Rakers et al. 2020) meaning that only 10%–20% achieve the child. However, animal and human studies indicate that this barrier function is lowered and the activity of 11ß‐HSD‐2 is down‐regulated when the maternal stress level is permanently increased (Jensen Peña et al. 2012; O’Donnell et al. 2012). As a consequence, a child's stress level can significantly be raised by external factors already during pregnancy. Moreover, maternal cortisol can affect the development of the foetal hypothalamic‐pituitary‐adrenal axis. As it changes its activity and alters its sensitivity level, this causes long‐lasting postnatal consequences (Moisiadis and Matthews 2014; Rakers et al. 2020). Maternal stress, along with malnutrition and maternal immune‐related factors, is one of the strongest early environmental factors that can influence the development and health of offspring later in life (Van Den Bergh et al. 2020).

However, so far, effects of prenatal maternal stress have been investigated in predominantly Caucasian samples from high‐income countries (Glover et al. 2018). Yet, it is very likely that many children are affected in low‐ and middle‐income countries as well because of a high risk of stress due to for example wars, interpersonal violence, poverty, or potential infections (Glover et al. 2018).

1.2. Natural Disasters as an Equalizing, External Stressor

At the time of writing this manuscript the western media landscape is filled with constant news on the 2025 Palisade Fires in California (USA), in which already many people have lost their lives and livelihood. While the debate about the exact reason for the fire continues, one thing is clear: climate change is at least partially to blame (Goss et al. 2020; Mansoor et al. 2022). Similarly, although potential problems in public administration and infrastructure did not help in overcoming or even avoiding the 2024 Spanish Floods, climate change also played a causal role. Although not all people may agree on all climate change related issues, most people would arguably agree that such disasters have long‐lasting negative consequences for the affected area and its people. Research shows that climate change is linked to an increase in the frequency and intensity of natural disasters, that is human emissions of greenhouse gases are altering the climate leading to more extreme weather events such as heatwaves, floods, and hurricanes (Banholzer et al. 2014; Van Aalst 2006). Observations since 1950 indicate a rise in extreme weather events, with predictions of further increases in the 21st century, including more frequent heatwaves and intense droughts (Banholzer et al. 2014; Coronese et al. 2019). The socio‐economic damages from natural disasters have been rising, with climate change contributing to this trend by increasing the intensity of such events (Coronese et al. 2019); affecting vulnerable populations disproportionately (Benevolenza and DeRigne 2019; Ibarrarán et al. 2009). Thus, if natural disasters cannot be prevented or contained in a timely manner, there is a distinct need for risk prevention and damage mitigation in order to enhance the societal resilience and stress resistance (Benevolenza and DeRigne 2019; Botzen and Van Den Bergh 2009; O’Brien et al. 2006).

Especially, for the most vulnerable among the victims of such disasters, the consequences can be life‐altering. One such vulnerable group are pregnant individuals ‐ the focus of this umbrella review. In that regard, natural disasters provide an opportunity to study the impact of prenatal stress with a largely independent random stressor. Hence, investigating the effects of exposure to natural disasters during pregnancy allows for an investigation of consequences of a stressor which the woman does not influence herself either by behavioural or genetic means, which then could have its own effects. Specifically, because exposure to natural disasters can occur to any person, natural disasters act somewhat like ‘natural experiments’, offering a form of approximately random assignment similar to controlled studies (Cao‐Lei et al. 2014; King et al. 2012; King and Laplante 2015). Importantly, however, one needs to state that exposure to natural disasters depends on for example socioeconomic status, ethnicity, housing conditions, and disability status (see e.g., Jardine et al. 2021; Thomson et al. 2021). Above that, resilience to disaster‐related stress is different between people meaning that the stress is perceived and coped differently by different people (Alves et al. 2021; Hajure et al. 2024; Langley‐Evans et al. 2022; SmithBattle and Phengnum 2023; Tuthill et al. 2021; Walker et al. 2022). In detail, resilience or vulnerability to stress ‐ due to any kind of stressor ‐ is different. Psychological factors such as emotional flexibility, locus of control, social problem solving, and cognitive skills, but also biological factors and their interplay with psychological and environmental factors contribute to stress resilience resulting in an inter‐ and intra‐individually pronounced resilience to stress in a particular situation (van der et al. 2013). However, using natural disasters to study the effects of prenatal maternal stress allows for the examination of the effects of stress on pregnant women and their offspring without the ethical concerns of inducing stress artificially. Moreover, even if different stress resilience does play a role, natural disasters affect broad populations simultaneously and the exposure to this particular stressor is less confounded with person‐specific factors than the exposure to other stressors like financial hardship or experience of physical violence. Thus, investigating the effects of stress due to exposure to natural disasters helps in isolating the effects of stress from person‐specific parameters as best as possible. Furthermore, this approach allows for the differentiation between objective stress (measurable hardship) and subjective stress (reported individual stress), providing insights into how each type affects foetal development during pregnancy and child development after birth (Cao‐Lei et al. 2014; King et al. 2012; St‐Pierre et al. 2018). Moreover, since the timing of a disaster is well documented and the time of exposure to the stressor is better assessable than for other stressors (like e.g., for stress due to financial hardship), the effects of stress at different stages of pregnancy can be investigated to a relatively large extent and tracking developmental outcomes over time in a longitudinal way is more precise (Heynen et al. 2023; Lafortune et al. 2021; Lequertier et al. 2019; Moss et al. 2017). For those reasons, in the current review in particular the effects of prenatal maternal stress caused by exposure to natural disasters are summarized.

1.3. The Scope of the Present Review

Experiencing a natural disaster in general is already potentially traumatizing but the consequences may be even more severe if the event coincides with a pregnancy. In the following, we summarize studies which report significant effects of a mother's exposure to a natural disaster on developmental variables in the child. When we talk about prenatal maternal stress below, we refer to severe or persistent stress which goes beyond everyday stress, that is goes beyond stress due to a busy everyday life or a stressful working day.

2.1. Summary of the Results

The results of the selected meta‐analysis and reviews highlight consistent effects of PNMS on various areas of development. Thus, an open‐ended summary of all effects, which were reported in the selected studies, indicate effects of PNMS on six developmental areas. The reported effects in the six developmental areas can be summarized as follows (for a synthesis of the results, see below).

2.2. Overlap in Results

In the published studies, there are overlaps which leads to an overrepresentation of certain natural disasters. In detail, four studies (King et al. 2012; Lafortune et al. 2021; Lamichhane et al. 2019; Matvienko‐Sikar et al. 2020) describe the effects of the snowstorm in Canada in 1998 and three studies describe the effects of the flood in Iowa in 2008 (Lafortune et al. 2021; Lamichhane et al. 2019; Matvienko‐Sikar et al. 2020). Thus the results may be seen as slightly biased towards the consequences of these disasters. However with that being said, in general, the degree of overlap between primary studies (and individuals in those studies) is minimal. This is also indicated by the Corrected Covered Area (CCA), a standard metric for quantifying overlap (using a citation matrix with rows representing unique studies and columns representing reviews). The resulting CCA was 1.9%, indicating only minimal overlap in primary studies across the included reviews and meta‐analyses.

2.3. Synthesis of the Results

A synthesis of the reported results indicates, on the one hand, that PNMS causes negative effects on all relevant developmental areas. This is illustrated by the differently coloured areas in a pie chart in Figure 2. Strength‐of‐evidence varies for the negative effects in different developmental areas as indicated by the different numbers of studies reporting negative effects in the particular developmental area (see Figure 2, indicated by the number of studies and different sizes of the coloured areas). On the other hand, a synthesis of the reported results indicates the interdependence of the negative effects. Specifically, against the background of developmental steps that build on one another, far‐reaching effects of early developmental difficulties on later developmental processes are highlighted (as indicated by the arrows in Figure 2). In detail, the identified effects on gestational duration which mostly take the form of preterm birth indicate overall limited development conditions due to the shortened time. Overall limited development conditions mean impairments in physical development as well as impairments in neurological development. Further, the identified impairments in physical development as indicated by reduced body size and weight, favour the development of physical illnesses and the identified impairments in neurological development favour restricted cognitive development. Restricted cognitive development then has far‐reaching consequences, including significantly impaired language acquisition. Limited language skills then mean significantly limited opportunities to acquire important social skills and emotional abilities. This again favours the emergence of critical internalizing and externalizing behaviour, that is, reduced expression of needs, that is being highly introverted or acting out aggressively and loudly. Behaviour like this per se indicates a child's emotional stress. Moreover, it favours the development of mental illness later in life.

Notably, one study investigated the effects of COVID‐19. While the pattern of results is comparable to other studies and disasters, COVID‐19 as a disaster sticks out because it is not a nature‐based and physical disaster like a blizzard or an earthquake. At the time of writing, COVID‐19 is a fairly recent disaster and the full effects of this prenatal stressor on birth outcomes and developmental trajectories cannot yet be fully understood.

3.1. Resulting Recommendations

By highlighting the potential negative consequences of PNMS for all developmental areas and the impairment of subsequent developmental steps conclusions can be drawn about specific recommendations for protective interventions. Importantly, to the best of our knowledge, those recommendations do not have to be limited or focussed on interventions for the case of a natural disaster. Rather, the here reported results depict effects of PNMS as best as possible independent of person‐specific factors in order to achieve a potentially high external validity. Nevertheless, authors have made recommendations on how to improve support for pregnant women in crisis situations. In particular, medical professionals working in disaster relief should be trained in the areas of maternal mental health and PNMS. Support programmes should include access to psychological support for women in the perinatal period in order to mitigate long‐term negative effects on their children (Sato et al. 2016). Psychological first aid according to WHO guidelines (World Health Organization and War Trauma Foundation and World Vision and International 2011) and cognitive behavioural therapy (CBT) are possible interventions at the population and individual level to reduce PNMS due to natural disasters (see King et al. 2021).

In line with etiological models, predictive adaptive response, and the foetal programming hypothesis, the here depicted results highlight aspects of pathogenesis but also of salutogenesis. First, the reported results highlight various negative impacts of PNMS for the child's health, which goes in line with theories like the ‘developmental origins of health and disease’ theory (DOHaD; Gillman 2005). According to the DOHaD hypothesis, both a child's development and susceptibility to disease later in life can be determined by environmental influences even before birth. PNMS, caused by a natural disaster, is an environmental factor that can affect foetal development during a sensitive phase. However, DOHaD traditionally focuses on the impacts of chemical exposures (e.g., air pollutants) or (mal)nutrition on health outcomes and the here reported results indicate that interactions of various factors should be taken into account. Thus, research should reflect real‐world scenarios by examining the interactions of multiple exposures including chemicals, nutrition, stress, infections, and drugs (Haugen et al. 2015) and should explore the broader impact of environmental chemicals on various organs and disease syndromes rather than focusing on single chemical‐disease relationships. As such, the assessment of natural disasters provides a possibility to assess a multi‐factorial stressor and thereby to fully grasp the integrated toxicity of exposures.

There also lies an opportunity in considering that a child's development can already be severely impaired during pregnancy. Considering the previously reported results, necessary assistance can be two‐folded. First, maternal stress should be addressed during prenatal checkups at hospitals or by gynaecologists as well as in midwife consultations or childbirth preparation courses. Additionally, Hentges et al. (2019) suggest a preventive program that can lower prenatal stress and lasts into the postnatal period so that negative effects are counteracted in the best possible way both prenatally and postnatally.

A second important chance in knowing about highly possible negative effects of PNMS lies in the possibility to address potential developmental difficulties after birth. Specifically, by knowing about maternal stress as a risk factor for foetal development and especially by knowing about the interdependence of the negative effects, a significant chance lies in targeted development support for the child after birth. Hence, the knowledge that a child's development may already be in danger during pregnancy is necessary to ensure that the resulting needs of the child can be taken into account after birth. Moreover, by focussing on changes in early development targeted early intervention is able to prevent later physical and mental diseases. In concrete terms, specific developmental support can be provided to the child if the extent to which its development was already restricted before birth is taken into account specifically. In this respect, knowledge of possible negative influences even before birth is essential to give the child the developmental support it needs so that developmental impairments can be counteracted in the best possible way. In this regard, one possibility also lies in sensitive and responsive parental care. Thus, sensitive and responsive parental care ‐ also by a non‐related but reliable caregiver ‐ is a powerful resilience factor. It helps the child to cope with upcoming developmental tasks (Hohm et al. 2017) and current reviews also indicate epigenetic, neurobiological, and behavioural consequences of sensitive and responsive maternal care claiming a prenatal‐postnatal interplay in shaping the child's development (Monk et al. 2012). Taken together, by addressing potential congenital developmental difficulties, the vicious circle of negative effects as depicted in Figure 2 can be interrupted and thereby otherwise occurring long‐lasting effects can be prevented.