Fetal programming provides a model for understanding the development of health and disease that is focused on prenatal conditions that impact the vulnerability of individuals to multiple pathologies (1, 2). Maternal exposure to environmental stressors during the prenatal period is one such condition that has been linked to various suboptimal birth outcomes such as preterm birth and low infant birthweight (3), as well as later neurodevelopmental impairments in childhood (4). The strength of the causal claim that maternal stress has a direct impact on fetal development is based on rigorous controlled experiments in animals, that distinguish prenatal from postpartum effects using methods such as cross-fostering or nursery rearing (5–8).

From a developmental and life-course perspective, however, stress exposure is unlikely to arise de novo following conception. In a similar vein, prenatal health is not independent of the health of the system prior to conception. For these reasons, the preconception period is emerging as an important focus for research on adverse birth outcomes and offspring development (9), as well as a model for understanding, and ultimately preventing, health disparities in pregnant women and their children (10, 11).

A small, but growing, evidence base in human studies provides preliminary support for the impact of preconception stress exposure on birth outcomes, although the majority of studies are based on maternal retrospective reports [see review (10)]. For example, at 9-months postpartum, adult participants in the United States’ Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) reported on stressful life events that occurred prior to conception. Results indicated that any exposure to preconception stress was associated with heightened risk for very low infant birthweight, and that the cumulative number of life events was inversely related to infant birthweight (12).

Results of the few prospective longitudinal studies lend some additional support for a link between preconception stress exposure and suboptimal birth outcomes. The National Child Development Study in Great Britain included measures of financial, parenting, family, and community stressors at birth, ages 7, 11 and/or 16 years (13). At ages 33 or 41 years, female participants (n ≈ 5,000, 96% White) recalled the outcomes of any pregnancies to date. Results indicated that exposure to stressors across childhood and adolescence was associated with higher rates of preterm birth and lower birthweight, even after accounting for prenatal stress exposure (14). In a follow-up to the National Longitudinal Study of Adolescent Health (Add Health), chronic stressors (e.g., parent receipt of public assistance, high neighborhood unemployment) during adolescence and young adulthood, but not stressful life events were inversely associated with offspring birthweight among the female participants (n > 5,000, 57% White, 24% Black, 18% Latina), and also partially explained racial/ethnic disparities in birthweight (15). In the Australian Longitudinal Study of Women's Health, participants' perceptions of stress measured in the three years prior to delivery were linked to state-based data on birthweight for 3,622 women (mostly partnered, highly educated) (16). Results showed no differences in offspring birthweight among women reporting none or minimal stress (survey scores 0 or 1) vs. stress deemed moderate/high. Finally, in a racially diverse sample (44% Black, 30% Latina or Hispanic, 27% White) of low-income women (n = 360), earlier GA at delivery was associated with high levels of stress appraisal (i.e., perceived stress and parenting stress) and both high and low levels of exposure to stressors (e.g., life events, financial strain, interpersonal violence) during the interpregnancy interval (17). In general, results of these prospective studies suggest that stress experienced prior to conception can negatively influence birth outcomes, but differences in sample characteristics, conceptualizations of stress and the timing and duration of the exposure window highlight the need for further study.

There are long-standing racial disparities in birth outcomes in the United States with Black women at elevated risk for delivering low birthweight and preterm infants compared to White women (18–20). Increased focus on social determinants of health points to systemic racism and other structural processes as major contributors to these persistent disparities (21–24) relative to individual-level factors (e.g., health behaviors, prenatal characteristics) (25, 26). Despite some evidence that preconception stress explains more variability in birth outcome inequities than stress experienced during pregnancy (27), little is known about the differential impact of type and timing of preconception stress exposure on birth outcomes among Black and White women.

Fetal growth and risk for adverse birth outcomes are known to differ by sex. Whereas birthweights for female infants are generally lower than for males, preterm birth and stillbirths occur more often in male gestations (28–30). Male fetuses may also be particularly susceptible to the negative effects of prenatal stress exposure (31–33), consistent with reports of a decrease in the male-to-female birth ratio in contexts of maternal stress (34–36). However, evidence also suggests that female fetuses show a reduction in growth rate in response to early gestational stress, a response that is considered adaptive (37). Given that most research has focused on stress experienced during pregnancy, the extent to which fetal sex moderates the impact of exposure to preconception stress on birth outcomes remains unclear.

The impact of stress exposure on health varies significantly as a function of type, timing, and chronicity, as shown in rodent and non-human primate models (38–42). Results from animal studies also demonstrate that when chronic stress is predictable, rather than unexpected or inconsistent, responses can become habituated and attenuated (43, 44). The study of type, severity and consistency of stress exposure in humans is clearly more complex, given the lack of experimental control and a less developed approach to a functional taxonomy compared to animal studies (45). Extant studies that have measured type and timing of stress in the immediate preconception period suggest that such dimensions explain variance in birth outcomes (14, 15, 17, 46). However, there is a need for longitudinal studies that begin in childhood, assess multiple domains of stress exposure across development, and follow participants through pregnancy and birth to rigorously test the impact of type, severity and consistency of preconception stress exposure on birth outcomes. The Pittsburgh Girls Study (PGS), an ongoing community-based longitudinal study, now in its 21st year of annual data collection, is one such study. Participants (n = 2,450) were enrolled at ages 5–8 years and have been interviewed each year about multiple aspects of health and development including exposure to stress. In prior analyses with this sample, we characterized the dynamic nature of three domains stress exposure (subsistence, safety, and caregiving stress) among all PGS participants between ages 7 to 17 years (47). These domains extended from a substantial literature based on animal models (39, 40) and work in humans (48). Analyses indicated variability in initial severity levels, in consistency over time, and timing of shifts in exposure level within- and across-domains. Moreover, group membership differed in terms of racial composition with Black participants over-represented in groups exposed to high and inconsistent levels, especially in the domains of subsistence and safety stress.

In the current study, we examine the influence of these three domains of preconception stress assessed between ages 7 to 17 years on later birth outcomes in PGS participants. We focus on gestational age at birth and infant birthweight given the extensive research on prenatal stress exposure and birth outcomes, the relevance for later health, and clear operational definitions. We hypothesize that high levels of stress exposure (severity) across childhood and adolescence would be inversely associated with gestational age and birthweight. In addition, we expect that changing levels of moderate-high stress (inconsistent exposures over time) would be associated with more adverse birth outcomes given lack of opportunity for adaptation/habituation. We hypothesize that these effects would be evident after controlling for several confounds including maternal age, parity and pre-pregnancy BMI. Finally, we hypothesize that preconception stress effects on birth outcomes would be greater for Black than White women. Because there was insufficient evidence for sex-specific effects of preconception stress exposures, we did not propose hypotheses, but conducted exploratory analyses of potential interactions with infant sex.

We used multiple waves of repeated data from a longitudinal study to examine the effects of type, timing and consistency of stress exposures during childhood and adolescence on later birth outcomes in a racially and socioeconomically diverse sample of urban-living women. We tested hypotheses that severity and inconsistent exposure to preconception stress would be associated with shorter GA at birth and lower birthweight. Results provided modest support for our hypotheses, showing negative effects of average-to-high levels of preconception subsistence stress (i.e., resource and housing stress) on offspring birthweight. This finding aligns with an established literature describing the pervasive effects of exposure to financial strain in childhood on health across the lifespan (61), as well as results from prior longitudinal and registry studies showing heightened risk for subsequent poor birth outcomes (14, 62, 63). The association is consistent with stress-sensitization and life-course models (64) contending that growing up in a household with income-related physical and social risks may potentiate physiological stress and alter stress regulatory systems (e.g., neuroendocrine, immunological, cardiometabolic) that, in turn, influence prenatal health, placental development and ultimately birth outcomes (65–68). Such a conceptualization integrates early programming (i.e., exposures most impactful during sensitive periods in childhood or adolescence), cumulative exposure (e.g., “wear and tear”) and specificity of type and timing models (10). Early life exposure to subsistence stress may also influence an individual's preparedness for pregnancy via long-term deficiencies in nutrition (e.g., dietary fatty acids, vitamin D), infection risk (69, 70), restricted access to quality health care, health screenings, and reproductive health preparations (e.g., folic acid supplementation) (71, 72) and via health behaviors such as smoking, disrupted sleep or depression that have also been linked to adverse birth outcomes (73–76). Important next steps will be research focused on subsistence stressors experienced before and during pregnancy that may influence fetal growth via additive, multiplicative or interactive effects, as well as factors that help explain heterogeneity in birth outcomes despite early adversity.

Stress related to subsistence or financial strain can be relatively stable across the life course (77), and this was evident for approximately 40% of the current sample. Our results, based on annual assessments showed a trend for inconsistent exposure to subsistence stress predicting shorter GA at birth, raising the intriguing possibility that unpredictability is an important feature of subsistence stress that could have implications for prenatal health (10). With a larger sample, future research could probe associations between patterns of inconsistent stress exposure (e.g., early vs. late increasing, childhood increasing and adolescent decreasing trajectories) on birth outcomes, and the extent to which unpredictability in stress exposure continues through pregnancy. In addition, dimensional measures may provide a more sensitive test of the impact of stress related to basic needs across development on perinatal health including birth outcomes.

Our findings also revealed an association between severity of preconception exposure to caregiving stress and GA that was moderated by infant sex. Although the magnitude of the effect was small (Cohen's f ≤ 0.25), the direction was consistent with prenatal stress studies that have highlighted male vulnerability (31, 32, 78, 79). Although not examined in the current study, it is possible that prenatal stressors related to early caregiving experiences mediated the effect of preconception stress exposure on birth outcomes. For example, early separation from a parent or exposure to caregiver depression may increase risk for interpersonal difficulties or could activate stressful feelings about parenting that emerge during pregnancy. Mechanistic studies are needed to extend these results and investigate the ways in which stress experienced across the lifespan could alter biological systems that support the healthy development of both male and female fetuses. Improving our understanding of potential selection effects of preconception stress on pregnancy status, pregnancy health and fetal outcomes are also important areas that warrant further research.

Our results showed further indication that the type of preconception stress has relevance for birth outcomes. We observed no effects of preconception exposure to neighborhood and domestic safety-related stress on birth outcomes. This result differs from studies documenting associations between exposure to violence in the prenatal period and risk for low birthweight and preterm birth (80) as well as retrospectively reported associations between preconception exposure to trauma (i.e., adult abuse and child maltreatment) on later birth outcomes (81). However, the literature on neighborhood safety is somewhat mixed and small but significant associations are most often reported in studies linking geocoded birth and police-recorded crime data [e.g. (82–84),]. In our prospective cohort study, it is possible that caregiver experiences of safety stress differed from those of the developing child. For example, some caregivers who perceive safety threats engage in higher levels of protective parenting behaviors, which appear to attenuate the impact of the exposure on child health and development [see (85) for a review]. Our dataset includes multiple measures of potential resilience factors (e.g., family support) that will be important for identifying the family and community contexts that mitigate the risks from stress exposure on birth outcomes.

Taken together, the observed patterns of association between type of preconception stress and birth outcomes suggest that there may be differential effects on health systems (e.g., maternal nutrition, endocrine, cardiovascular, or immune functioning) and/or epigenetic modification that warrants further investigation. In prior work with a different sample, we have shown that type of prenatal stress exposure is differentially related to neuroendocrine and cardiac response to a controlled stressor (86), and this may also be the case for preconception stress types. Increasing clarity in operational definitions of stress and a greater focus on the preconception period are critical steps towards filling these gaps and yielding precise, developmentally specific targets for effective preventive interventions.

Birthweight was lower in infants of Black compared with White mothers as has been observed in prior research (21, 22); in the current study, this effect was evident while also accounting for the type, severity and consistency of preconception stress exposures during childhood and adolescence. We previously documented racial differences in stress exposure trajectories based on the full PGS sample (47), and it is also widely recognized that cumulative wear and tear (e.g., weathering, accelerated aging) associated with systemic racism and other structural processes contributes to inequities in Black women's reproductive outcomes (87, 88). In the present PGS subsample of young women giving birth, differences in severity and consistency of stress exposures measured between ages 7–17 years did not account for the risk for low birthweight. Several explanations are possible. First, it is unlikely that our measures of stress, examined as categorical variables in separate models, fully captured the pervasive and multifaceted racial inequalities and lifelong exposure to discrimination experienced by Black women. Second, there may have been important unmeasured influences that occurred before age 7 (e.g., fetal programming, early trauma) and/or after age 17 (e.g., work, education, relationship stress, pregnancy complications and pregnancy-related stress, lack of access to quality prenatal care) that contributed to birth outcomes. Third, it is possible that specific developmental features of childhood-adolescent stress exposures (e.g., moderate or high increasing trajectories) that were associated with observed race differences in the full PGS could not be detected in the smaller childbearing subsample due to reduced statistical power. Large-scale studies with culturally sensitive, multi-level measures of stress and systemic racism across the lifespan are clearly warranted to understand the preconception and prenatal mechanisms that underlie these persistent racial disparities.

Despite some unique strengths of the current longitudinal study, several limitations should also be noted. First, we focused on stress exposures during the formative developmental periods of childhood and adolescence but did not model exposures through to the time of conception or through pregnancy. For example, it is possible that stress type, or changes in severity or consistency in the immediate preconception period contributed to the causal pathway. Results from European population-based registry studies suggest that exposure to stressful life events (i.e., death or serious illness in a close relative) in the 6- to 18-months prior to pregnancy may carry especially high risk as far as preterm birth and low infant birthweight are concerned (46, 89, 90). Understanding the salience and influence of stressors experienced during both the preconception and prenatal periods is critical for informing optimal timing of preventive interventions. Our findings contribute to this effort in demonstrating that, although effects were generally modest, stress exposures in childhood and adolescence may have implications for later birth outcomes and suggest potential benefits of preventive interventions even during the school years. Second, pre-pregnancy BMI was included as a proxy for overall health, but we recognize that BMI is a non-specific measure of perinatal health, especially for Black women (91). Moreover, chronic health conditions (e.g., hypertension or diabetes) that are known risk factors for low birthweight and GA (92) were not included. Third, several factors may have impacted the generalizability of our findings. For example, we focused on births occurring after age 18 to retain temporality between the independent and dependent variables. In doing so, however, we may have excluded births that occurred following especially high levels of preconception stress exposure (93), which probably also reduced our model estimates. Similarly, our focus on live births may have introduced sample bias, including potential bias related to infant sex given that male fetuses are often more vulnerable to loss. In addition, patterns of missing birth outcome data suggested that our results may be most relevant to Black or multi-racial, multiparous, women living in the Pittsburgh region. Fourth, the sizeable, but nonsignificant, birthweight coefficients suggest that the study may have been underpowered to detect causal effects. Finally, while the trajectory modeling approach used in the current study provided new information about severity and consistency of exposures across childhood and adolescence, applying the trajectories to the smaller PGS subsample of pregnant women necessitated reduction into relatively small binary groups that also prevented us from pinpointing developmental periods of heightened plasticity or vulnerability. For example, animal studies have shown that in the peripubertal period, plasticity in the hypothalamic-pituitary-adrenal (HPA) axis increases (94, 95), the effects of which may persist into adulthood (96). Human studies have also shown unique responses to stress around the time of puberty (97).

The current study supports a life-course perspective for understanding the impact of stress on women's reproductive health with relevance for the next generation. As in other areas of health such as cardiovascular health (98–100), testing the effects of the type, severity and consistency of stress across childhood and adolescence will help to incorporate life history in our understanding and prevention of health problems in pregnancy and the neonatal period. Such an approach is consistent with decade-long calls for the need for preconception physical health care [e.g. (101, 102)], in light of the often modest successes of pregnancy-specific behavioral health interventions (103, 104).