Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 1.
Published in final edited form as: Matern Child Health J. 2021 May 21;25(9):1455–1464. doi: 10.1007/s10995-021-03139-x

Adverse Pregnancy Outcomes Following the Assassination of John F. Kennedy in 1963

Alexa A Freedman 1,2, Gregory E Miller 2,3, Lauren S Keenan-Devlin 4, Britney P Smart 1, Janedelie Romero 1, Ann Borders 4,5,6, Linda M Ernst 7
PMCID: PMC8355166  NIHMSID: NIHMS1712499  PMID: 34021436

Abstract

Introduction

Women exposed to stressful events during pregnancy are thought to be at increased risk of adverse birth outcomes. However, studies investigating stressful events are often unable to control for important confounders, such as behavioral and genetic characteristics, or to isolate the impact of the stressor from other secondary effects. We used a discordant-sibling design, which provides stronger inferences about causality, to examine whether a widespread stressor with limited impact on day-to-day life (John F. Kennedy assassination) resulted in an increased risk of adverse birth outcomes.

Methods

Data were obtained from the Collaborative Perinatal Project, a prospective, multi-site cohort study conducted in the US from 1959 to 1965. Our analysis was restricted to singleton live births ≥24 weeks born before the assassination (n = 24,406) or in utero at the time (n = 5833). We also evaluated associations within siblings discordant for exposure (n = 1144). We used survival analysis to evaluate associations between exposure and preterm birth and marginal models to evaluate associations with birthweight and placental pathology.

Results

First trimester exposure was associated with preterm birth (hazard ratio (HR): 1.17; 95% CI: 1.05, 1.31). In the discordant-sibling model, the point estimate was similar (HR: 1.22; 95% CI: 0.36, 4.06). Third trimester exposure was associated with increased odds of fetal acute inflammation in the placenta (odds ratio (OR): 1.34, 95% CI: 1.05, 1.71).

Conclusions for Practice

First trimester exposure to an acute stressor was associated with preterm birth. We did not observe increased odds of placental pathology with first trimester exposure; however, stress may increase preterm birth risk through chronic placental inflammation, which was not evaluated in this sample.

Keywords: Stress, Pregnancy, Preterm birth, Placenta, Siblings

Introduction

Preterm birth (birth < 37 weeks gestation) is associated with increased risk of perinatal morbidity and mortality (Goldenberg et al., 2008; Qiu et al., 2012; Saigal & Doyle, 2008). In the United States, the rate of preterm birth is approximately 10% (Murphy et al., 2017). Morbidities associated with preterm birth also have implications for long-term health and development, including increased risk of neurodevelopmental disorders and cardiovascular disease (Chehade et al., 2018; Lundgren & Tuvemo, 2008; Saigal & Doyle, 2008).

Evidence indicates that women who experience stress during pregnancy are at increased risk of adverse birth outcomes. A comprehensive review of prenatal stress and gestational age at delivery reported that nine of fourteen studies evaluating discrete life events, such as loss of a family member, and six of nine studies evaluating catastrophic events, such as a terrorist attack or natural disaster, demonstrated significant associations with reduced gestational age at delivery (Dunkel Schetter & Glynn, 2011). A similar review of pregnancy outcomes following disasters reported evidence of associations with measures of fetal growth, including birthweight, but did not find consistent evidence to support an association with gestational age at delivery (E. Harville et al., 2010). Since the publication of these reviews, several studies have reported associations of life and catastrophic events with gestational age at delivery (Barrios et al., 2014; Palmeiro-Silva et al., 2018; Tegethoff et al., 2010a) and measures of fetal growth (E. W. Harville & Do, 2016; Palmeiro-Silva et al., 2018; Suzuki et al., 2016; Tegethoff et al., 2010a; Wainstock et al., 2013). Further, events during pregnancy have implications for child development and have been associated with infant temperament (Laplante et al., 2016; Zhang et al., 2018), autism and attention deficit/hyperactivity disorder (Ronald et al., 2010), and depressive symptoms in adolescence (Kingsbury et al., 2016). Pathways between stress and adverse birth and pediatric outcomes may involve alterations in the placenta, which plays a crucial role in regulating fetal growth and facilitating maternal–fetal communication (Bronson & Bale, 2016). Studies investigating links between stress and the placenta have reported associations of stress with measures of placental function (Dahlerup et al., 2018), placental inflammation (Ernst et al., 2013; Keenan-Devlin et al., 2017; Marinescu et al., 2014), placental weight (Tegethoff et al., 2010b), placental morphology (Lahti-Pulkkinen et al., 2018), and utero-placental blood flow (Shirazi et al., 2019; Teixeira et al., 1999).

Despite evidence of associations between stressful events and adverse pregnancy outcomes, important questions remain. Studies are generally unable to control for unmeasured or mis-measured confounders related to genetic, environmental, and behavioral characteristics, which could inflate the reported associations. Additionally, widely-experienced stressors, such as natural disasters and terrorist attacks, often have secondary effects with long-term implications for health, including psychological distress, environmental exposures, displacement, and reduced access to resources (Galea, 2007). These secondary effects may differentially affect the population and may make it challenging to disentangle the specific contribution of the acute event to adverse health outcomes. Further, there are conflicting hypotheses regarding sensitive windows for stress exposure. While there is evidence to suggest that greater susceptibility to stress in the first trimester may increase the likelihood of preterm birth, a competing hypothesis suggests that third trimester stress exposure is more pertinent for preterm birth and may affect parturition (Torche, 2011). Similarly, the mechanistic pathways between stress and adverse birth outcomes are poorly understood. Changes in prenatal care usage and medical intervention may play a role in the reported associations between stress and adverse birth outcomes (Azagba & Sharaf, 2011; Tudiver et al., 1995). Alternatively, the biologic stress response may affect the maternal–fetal interface and have a direct impact on parturition or fetal growth. The potential impact of maternal stress on placental physiology has been understudied in the context of stressful events during pregnancy and adverse birth outcomes.

To address these questions, we investigated the impact of a widely experienced stressor, the assassination of President John F. Kennedy (JFK) on November 22, 1963. The assassination of JFK shocked the nation and was often described as similar to the death of a parent or close friend (Banta, 1964; Sheatsley & Feldman, 1964), experiences that are associated with adverse pregnancy outcomes (Laszlo et al., 2016). Recognizing the potential implications of events like the JFK assassination, Catalano and Hartig (2001, p. 333) coined the term, “communal bereavement,” to describe the “widespread experience of distress among persons who never met the deceased.” Studies investigating similar instances of communal bereavement have reported subsequent increases in depressive symptoms among pregnant individuals (Truijens et al., 2015) and in the incidence of very low birthweight (Catalano & Hartig, 2001). We extend this work by utilizing a prospective pregnancy cohort to investigate the impact of the JFK assassination on pregnancy outcomes in a US-based sample.

Methods

Study Sample

Data were obtained from the Collaborative Perinatal Project, a prospective cohort study that enrolled subjects at their first prenatal care visit from 1959 to 1965 at twelve sites in the United States. Details of the study design have been published (Niswander & Gordon, 1972). Participants provided verbal consent, which was common at the time (Hardy, 2003). This analysis was exempt from IRB approval, as data are publicly available through the U.S. National Archives and Records Administration.

The exposure of interest (in utero at the time of the JFK assassination) was determined based on reported last menstrual period. Outcomes of interest included preterm birth, birthweight, and placental abnormalities. Preterm birth (defined as < 37 weeks gestation) was determined using last menstrual period and date of birth. Birthweight was abstracted from delivery records and z-scores based on gestational age and sex were calculated. Macroscopic and microscopic placental pathology exams were conducted for study participants following delivery. Variables from the placental examinations were grouped into four types of pathology based on Amsterdam criteria: maternal acute inflammation, fetal acute inflammation, maternal vascular malperfusion, and fetal vascular malperfusion (Khong et al., 2016). Maternal acute inflammation was defined as neutrophilic infiltration of the chorion and/or amnion. Fetal acute inflammation was defined as neutrophilic infiltration of the umbilical vein, fetal surface vessels, and/or umbilical artery. Maternal vascular malperfusion was defined as presence of one or more of the following: retroplacental hematoma, micro infarcts, intervillous thrombi and adjacent villous infarction (infarction hematoma), thrombosis, fibrinoid, or atheroma. Fetal vascular malperfusion was defined as thrombosis of fetal vessels.

Our analytic sample was restricted to singleton live births from 11 study sites (one study site stopped recruiting before the JFK assassination and was excluded). Additionally, anomalous births and births < 24 weeks gestation were excluded to account for viability in the 1960s and implausible gestational age values, resulting in an eligible sample size of 50,130 subjects (Fig. 1) (Seri & Evans, 2008). Of these subjects, 42,275 (84.3%) had complete information on placental pathology and covariates. Of these, 24,406 delivered before the JFK assassination (unexposed) and 5833 were pregnant at the time of the assassination (exposed). The sample also included 1144 women with both unexposed and subsequent exposed pregnancies for the sibling pairs analysis. Additionally, due to concerns regarding consistency of placental examinations across reviewers and institutions (over 100 pathologists at 11 different sites reviewed placentas), analyses of placental pathology were restricted to the Boston study site (n = 6959), where 92.6% of placentas were reviewed by the same pathologist.

Fig. 1.

Fig. 1

Open in a new tab

Study inclusion flowchart

Statistical Analysis

We used survival analysis to evaluate associations between exposure to the JFK assassination and preterm birth. Unlike conventional techniques, survival analysis can handle a change in exposure status during pregnancy (time-dependent exposure) (Ahrens et al., 2014). Among the full sample, we used extended Cox regression models to account for the time-dependent exposure. The time-dependent exposure variable was structured to ensure that pregnancies in utero at the time of the assassination were considered unexposed until the assassination and exposed after the assassination (e.g., a woman exposed in the 30th week was counted in the unexposed risk set until the 30th week and in the exposed risk set after). We also used a robust sandwich estimator to account for correlations among siblings (Allison, 2005). Gestational age was used as the timescale (last menstrual period considered time zero) with adjustment for left truncation based on gestational age at study enrollment. Study participants enrolled after 37 weeks were excluded from models of preterm birth. Models were adjusted for maternal age, race, parity, education, smoking status, fetal sex, and study site. In the full sample, we also conducted sensitivity analyses to account for potential inaccuracy of gestational age information (Kramer et al., 1988) and potential seasonality in conceptions leading to seasonality in preterm birth (Darrow et al., 2009). In the first, we excluded those with implausible birthweights for gestational age and sex, based on a z-score greater than three. In the second, we restricted the unexposed to reflect the same months of conception as the exposed (e.g., the stratified model of first trimester exposure was restricted to conceptions in August-November).

In the sibling subset, we used stratified Cox models to evaluate associations within sibling pairs discordant for exposure (fixed-effects analysis) (Allison, 2005). Sibling models were adjusted for maternal age, smoking status, and fetal sex, as these may differ across pregnancies. We were unable to control for birth order as the unexposed pregnancy occurred prior to the exposed pregnancy for all women (pregnancies conceived after the JFK assassination were excluded).

We also investigated associations between exposure and birthweight z-score as a measure of fetal growth using marginal models to account for correlations within siblings. Models were adjusted for maternal age, race, parity, education, body mass index, smoking status, and study site. Since the birthweight z-score incorporates gestational age and fetal sex, models were not adjusted for these variables. Sibling fixed effects models were operationalized by modeling the difference in birthweight z-score within a sibling pair using linear regression. Sibling models were adjusted for differences in maternal age, body mass index, and smoking status. We also used marginal models to investigate associations between exposure and dichotomous placental pathology variables. Models of placental pathology were adjusted for maternal age, race/ethnicity, smoking status, fetal sex, and gestational age. Models were stratified by trimester of exposure to facilitate detection of sensitive periods. Analyses were performed using SAS version 9.4 (SAS Institute INC., Cary, North Carolina) and an alpha level of 0.05 was used to determine statistical significance.

Results

Descriptive Characteristics

A majority of the subjects in this analysis were between 20 and 35 years old (68.5%), had a body mass index within the normal range (68.5%), were married (77.0%), and did not complete high school (58.9%) Table 1. Most identified as white (45.7%) or Black (47.7%) and 47.7% reported smoking during pregnancy. Most subjects went into spontaneous labor (90.5%), had a vaginal delivery (94.8%), and delivered at term (87.3%).

Table 1.

Descriptive characteristics for the analytic sample of the Collaborative Perinatal Project

Characteristic, N (%) Total Sample n = 30,239 Exposed n = 5833 Unexposed n = 24,406
Demographics
 Maternal age, years
 <20.0 7361 (24.3) 1420 (24.3) 5941 (24.3)
  20.0–34.9 20,715 (68.5) 4012 (68.8) 16,703 (68.4)
  35.0–39.9 1669 (5.5) 311 (5.3) 1358 (5.6)
 ≥40 494 (1.6) 90 (1.5) 404 (1.7)
 Race/ethnicity
  White 13,821 (45.7) 2622 (45.0) 11,199 (45.9)
  Black 14,422 (47.7) 2832 (48.6) 11,590 (47.5)
  Other 1996 (6.6) 379 (6.5) 1617 (6.6)
 BMI, kg/m2
 <18.5 2661 (9.6) 484 (9.2) 2177 (9.7)
  18.5–24.9 19,005 (68.5) 3559 (67.6) 15,446 (68.8)
  25.0–29.9 4273 (15.4) 865 (16.4) 3408 (15.2)
  30.0–34.9 1290 (4.7) 259 (4.9) 1031 (4.6)
 ≥35.0 505 (1.8) 101 (1.9) 404 (1.7)
 Married 23,283 (77.0) 4324 (74.1) 18,959 (77.7)
 Education, years
 < 12 (did not complete high school) 17,819 (58.9) 3397 (58.2) 14,422 (59.1)
  12 (completed high school) 8966 (29.7) 1713 (29.4) 7253 (29.7)
 >12 (more than high school) 3454 (11.4) 723 (12.4) 2731 (11.2)
 Smoked during pregnancy 14,376 (47.7) 2719 (46.7) 11,657 (48.0)
 Multiparous 21,279 (70.4) 4034 (69.2) 17,245 (70.7)
Pregnancy characteristics
 Gestational age, completed weeks
 <28 157 (0.5) 20 (0.3) 137 (0.5)
  28–31 542 (1.8) 98 (1.7) 444 (1.8)
  32–36 3138 (10.4) 586 (10.1) 2552 (10.5)
 ≥37 26,402 (87.3) 5129 (87.9) 21,273 (87.2)
 Low birthweight, <2500 g 2887 (9.6) 540 (9.3) 2347 (9.6)
 Sex, male 15,350 (50.8) 2924 (50.1) 11,980 (50.9)
 Spontaneous labor 27,197 (90.5) 5245 (90.4) 21,952 (90.5)
 Oxytocin administered 4174 (16.9) 979 (16.8) 3195 (16.9)
 Vaginal delivery 28,554 (94.8) 5506 (94.5) 23,048 (94.8)
Open in a new tab

Analytic Results

Among the full sample, first trimester exposure to the JFK assassination was associated with increased risk of preterm birth (hazard ratio (HR): 1.17, 95% confidence interval (CI): 1.05, 1.31) Table 2. Exposure in the second and third trimesters was not associated with preterm birth. Results were similar in a sensitivity analysis excluding those with implausible gestational age and birthweight combinations (HR: 1.17, 95% CI: 1.05, 1.31) and in a sensitivity analysis accounting for season of conception (HR: 1.16, 95% CI: 1.03, 1.31). The point estimate was also slightly stronger in the subset of sibling pairs discordant for exposure (HR: 1.22, 95% CI: 0.36, 4.06). First trimester exposure was also associated with a slight reduction in birthweight for gestational age and sex z-score (β: −0.05, 95% CI: −0.09, −0.01) Table 3. However, this finding was not present in the sibling analysis based on difference in z-score (β: 0.26, 95% CI: −0.02, 0.55).

Table 2.

Associations between exposure to the John F. Kennedy assassination and preterm birth in the total sample (n = 28,877) and sibling sample (n = 1144 pairs) using survival analysis

Unadjusted Adjusteda
HR 95% CI HR 95% CI
Total sampleb
 Unexposed Reference
 First trimester 1.19 1.07, 1.33 1.17 1.05, 1.31
 Second trimester 1.03 0.91, 1.17 1.04 0.91, 1.18
 Third trimester 0.92 0.75, 1.13 0.93 0.76, 1.14
Sensitivity analysis 1 c
 Unexposed Reference
 First trimester 1.19 1.07, 1.33 1.17 1.05, 1.31
 Second trimester 1.03 0.91, 1.17 1.04 0.91, 1.18
 Third trimester 0.92 0.75, 1.13 0.93 0.76, 1.14
Sensitivity Analysis 2d
 Unexposed Reference
 First trimester 1.20 1.07, 1.35 1.16 1.03, 1.31
 Second trimester 1.10 0.96, 1.27 1.09 0.95, 1.25
 Third trimester 0.89 0.72, 1.10 0.92 0.74, 1.14
Sibling sample
 Unexposed Reference
 First trimester 1.29 0.76, 2.20 1.22 0.36, 4.06
 Second trimester 1.33 0.75, 2.35 1.48 0.35, 6.37
 Third trimestere
Open in a new tab

CI confidence interval, HR hazard ratio

a

Total sample and sensitivity analysis models adjusted for maternal age, race, parity, education, smoking status, fetal sex, and study site. Sibling models adjusted for maternal age, body mass index, smoking status, and fetal sex

b

Participants enrolled in the study at or after 37 weeks were excluded from preterm birth models

c

Sensitivity analysis 1 excludes those with implausible gestational age based on a birthweight for gestational age and sex z-score > 3 (n = 171 excluded)

d

Sensitivity analysis 2 restricts the unexposed to the same months of conception as the exposed, based on trimester of exposure (model of first trimester exposure restricted to conceptions in August–November, second trimester exposure restricted to May–August conceptions, and third trimester exposure restricted to March–May conceptions)

e

Not evaluated due to bias introduced in sibling pairs exposed in the third trimester (exposed pregnancy likely to have greater gestational age than unexposed pregnancy because exposed pregnancy needed to continue long enough into the third trimester to become exposed and be included in the analysis)

Table 3.

Associations between exposure to the John F. Kennedy assassination and birthweight for gestational age and sex z-score in the total sample (n = 30,239) and sibling sample (n = 1144 pairs) using linear regression

Unadjusted Adjusteda
β 95% CI β 95% CI
Total Sample
 Unexposed Reference
 First trimester −0.05 −0.09, −0.01 −0.05 −0.09, −0.01
 Second trimester 0.01 −0.03, 0.06 0.00 −0.05, 0.05
 Third trimester −0.03 −0.08, 0.02 −0.03 −0.08, 0.02
Sibling sample
 Unexposed Reference
 First trimester 0.08 −0.01, 0.16 0.26 −0.02, 0.55
 Second trimester 0.21 0.12, 0.30 0.36 0.07, 0.65
 Third trimesterb
Open in a new tab

CI confidence interval

a

Total sample models adjusted for maternal age, race, parity, education, smoking status, body mass index, and study site. Sibling models adjusted for differences in maternal age, body mass index, and smoking status

b

Not evaluated due to bias introduced in sibling pairs exposed in the third trimester (exposed pregnancy likely to have greater gestational age than unexposed pregnancy because exposed pregnancy needed to continue long enough into the third trimester to become exposed and be included in the analysis)

Among participants at the Boston study site, first trimester exposure was unexpectedly associated with lower odds of fetal acute inflammation (odds ratio (OR): 0.61, 95% CI: 0.47, 0.80) and maternal vascular malperfusion (OR: 0.65, 95% CI: 0.50, 0.86) Table 4. Conversely, third trimester exposure was associated with increased odds of fetal acute inflammation (OR: 1.34, 95% CI: 1.05, 1.71). Estimates for exposure in the third trimester were also suggestive of increased odds of maternal acute inflammation (OR: 1.28, 95% CI: 0.95, 1.73) and maternal vascular malperfusion (OR: 1.27, 95% CI: 0.98, 1.65; Table 4). Fetal vascular malperfusion was rare (1.0%), though similar patterns were exhibited.

Table 4.

Associations between exposure to the John F. Kennedy assassination and placental abnormalities among participants from the Boston study site (n = 6959)

Prevalence Unadjusted Adjusteda
N (%) OR 95% CI OR 95% CI
Maternal acute inflammation 976 (14.0)
 Unexposed Reference
 First trimester 0.95 0.72, 1.27 0.95 0.72, 1.26
 Second trimester 0.82 0.61, 1.10 0.83 0.62, 1.11
 Third trimester 1.27 0.94, 1.71 1.28 0.95, 1.73
Fetal acute inflammation 1665 (23.9)
 Unexposed Reference
 First trimester 0.61 0.47, 0.79 0.61 0.47, 0.80
 Second trimester 0.79 0.63, 1.00 0.80 0.63, 1.01
 Third trimester 1.34 1.05, 1.71 1.34 1.05, 1.71
Maternal vascular malperfusion 1398 (20.1)
 Unexposed Reference
 First trimester 0.66 0.50, 0.86 0.65 0.50, 0.86
 Second trimester 0.83 0.65, 1.06 0.82 0.64, 1.05
 Third trimester 1.26 0.97, 1.64 1.27 0.98, 1.65
Fetal vascular malperfusion 68 (1.0)
 Unexposed Reference
 First trimesterb
 Second trimester 0.64 0.20, 2.05 0.66 0.20, 2.15
 Third trimester 1.88 0.80, 4.40 1.82 0.77, 4.30
Open in a new tab

CI confidence interval, OR odds ratio

a

Models adjusted for maternal age, race, gestational age, fetal sex, and smoking status

b

No exposed placentas with fetal vascular malperfusion in the first trimester

Discussion

Our results indicate that the impact of the JFK assassination on pregnancy outcomes depends on trimester of exposure. We observed that first trimester exposure was associated with increased hazard of preterm birth and a very slight reduction in birthweight for gestational age and sex z-score in the full sample. Exposure in the second or third trimester was not associated with preterm birth or birthweight z-score. In the discordant sibling sample, the point estimate for first trimester exposure and preterm birth was similar. Unlike in the full sample, there was no association between first trimester exposure and birthweight z-score in the sibling sample. However, parity is associated with increasing birthweight (Hinkle et al., 2014) and by design, the exposed sibling was born after the unexposed sibling and would be expected to have a higher birthweight. Thus, the null association with first trimester exposure may support a reduction in birthweight in the exposed sibling, consistent with the finding in the full sample.

Similar to our findings, another study has linked the JFK assassination to pregnancy outcomes, reporting a decline in the male/female birth ratio among non-white women (Grech, 2015).1 Additionally, a study investigating the prevalence of very low birthweight (<1500 g) following two stressors that also resulted in communal bereavement (the assassination of the prime minister of Sweden in 1986 and the sinking of a ferry that resulted in the death of over 800 people in 1994) reported increases in the quarterly incidence of very low birthweight in the Swedish population following each event (Catalano & Hartig, 2001). While the authors did not investigate preterm birth, very low birthweight is likely a consequence of preterm birth (Kramer, 1987). Several studies have evaluated terrorist attacks, which may also reflect a similar exposure to the JFK assassination, though terrorist attacks may have long-term impacts on psychological distress (Galea, 2007). Similar to our results, a study that evaluated the 2004 Madrid train bombings using a time-series analysis reported an elevated rate of preterm birth following the bombings (Sherrieb & Norris, 2013). Conversely, a study evaluating the impact of the September 11th terrorist attacks reported no association with preterm birth or measures of fetal growth (Endara et al., 2009). However, the study did not evaluate trimester of exposure or account for changing exposure status across pregnancy. These features of the analysis could have masked an association. Another study of September 11th in a Boston cohort that matched on gestational age at exposure and evaluated trimester-specific associations reported reduced odds of preterm birth with first trimester exposure (Rich-Edwards et al., 2005). The authors hypothesize that one explanation could be due to competing mortality.

Our findings of an association with first trimester exposure, but not second or third trimester exposure, supports early pregnancy as a particularly vulnerable period for stress exposure. An association between stress exposure in early pregnancy and preterm birth is consistent with studies investigating other forms of stress in early pregnancy (Shapiro et al., 2013). Additionally, studies investigating the biological impact of stress on pregnancy have demonstrated that the physiological response to stress is greater in early pregnancy as compared to late pregnancy (Wadhwa et al., 2011). Similarly, maternal perceptions of stress may decrease as pregnancy progresses (Glynn et al., 2004).

Our approach extends knowledge of the association between acute stress and preterm birth by employing modeling strategies to reduce bias and confounding. We modeled preterm birth using survival analysis with a time-dependent exposure, which improves classification of the exposed and unexposed periods of pregnancy. We also conducted sensitivity analyses to account for potential misclassification of preterm birth and seasonal patterns in conceptions. Additionally, we estimated associations among sibling pairs discordant for exposure, which enhances control for shared confounding factors, including stable maternal environmental and behavioral characteristics and partial control for genetic variation. While the confidence intervals from the sibling analyses were wide due to the smaller sample size of discordant siblings, the point estimate for first trimester exposure was similar to the main findings. Comparing the point estimates is useful for determining the direction and magnitude of the bias (Lash et al., 2009), thus our results suggest that the potential impact of uncontrolled confounding by fixed maternal effects may be small.

While the use of a cohort from the 1960s may limit external validity due to changes in pregnancy management and population characteristics (such as the prevalence of smoking), the reduced capacity for medical intervention and the nature of the stressor may provide a better picture of the true effect of stress on accelerating parturition. The majority of births were spontaneous and vaginal, which mitigates the impact of other potential mediators, such as changes in care seeking behavior and pregnancy management as a result of stress. The JFK assassination also did not have a long-term impact on daily life, as is often seen with other types of widely-experienced stressors. Thus, the observed associations cannot be attributed to secondary effects of the stressor, such as changes in access to resources or exposure to environmental toxins. Our analyses do not support placental pathology as a mediator of the association between stress exposure and preterm birth. First trimester exposure to the JFK assassination was associated with preterm birth risk, but not with increased risk of any of the placental pathology characteristics. Unexpectedly, we observed protective associations of first trimester exposure with maternal vascular malperfusion and fetal acute inflammation. This association may result from survivor bias; first trimester exposure may result in early pregnancy loss of the most severely affected pregnancies. Research has demonstrated that maternal stress is associated with early pregnancy loss (Qu et al., 2017), which may result in the prevalence of some placental abnormalities being lower among pregnancies exposed in the first trimester. The mean gestational age at enrollment for the analytic sample was 22.2 weeks, so we were unable to evaluate the impact of exposure on early pregnancy loss to explore this potential explanation. Though, one study reported that the male/female birth ratio declined among non-white women following the JFK assassination and the authors hypothesize that this was due to increased loss of male fetuses (Grech, 2015).

Interestingly, we did observe an association between third trimester exposure and increased risk of fetal acute inflammation, and our estimates were also suggestive of an association with maternal acute inflammation. These findings may reflect stress-related changes in host resistance to infections, which trigger acute inflammatory responses. However, as we have argued elsewhere, stress is more likely to increase preterm birth risk through chronic (rather than acute) inflammation at the maternal–fetal interface (Ernst et al., 2013; Keenan-Devlin et al., 2017). Unfortunately, evaluations of chronic placental inflammation, such as the presence of chronic chorioamnionitis, chronic deciduitis, and chronic villitis, were not included in the placental assessment. Future research should investigate chronic placental inflammation as a mediator of the association between stress and adverse birth outcomes.

Limitations of this work include the use of a cohort study that was conducted over 50 years ago. One concern in particular is the use of the use of last menstrual period to determine gestational age, which may result in misclassification of preterm birth (Kramer et al., 1988). As the vast majority of births are term, it is more likely that term births would be misclassified as preterm rather than preterm births misclassified as term births. When we excluded births with implausible birthweight for gestational age and sex, results were consistent. Additionally, while the CPP did include a thorough placental examination that assessed many characteristics still evaluated today, some methods for evaluating placental pathology have changed. Notably, measures indicative of chronic placental inflammation were not included in the examination. Strengths of the study included the large, diverse sample size of over 30,000 pregnancies, which facilitated evaluating trimester-specific exposure periods. We were also able to evaluate associations within sibling pairs discordant for exposure to improve control for shared confounders. Additionally, the CPP included a standardized placental examination, which allowed us to investigate potential mechanisms related to placental function.

Our results indicate that exposure to a widely-experienced stressor in the first trimester may increase the risk of preterm birth. Future studies should leverage similar widely-experienced stressors to determine if associations persist in modern cohorts. Additionally, studies should investigate potential mediation by the placenta using modern examination protocols and diagnostic criteria. In particular, studies should investigate chronic inflammation, as chronic inflammation is associated with both maternal stress and adverse birth outcomes and is a plausible mediator of the observed association between first trimester exposure and preterm birth. While stressors like the JFK assassination are often unavoidable and unpredictable, improving our understanding of how they impact pregnancy can inform clinical management, such as the potential need for additional mental health support.

Signifiance Statement.

What is already known on this subject?:

Many studies have reported associations between stressful events during pregnancy and adverse pregnancy outcomes. However, these studies are often unable to control for important confounders.

What this study adds?:

Exposure to an acute stressor in the first trimester was associated with an increased risk of preterm birth and the point estimate was similar in a subset of siblings discordant for exposure. Our results provide evidence to support an association between stress exposure in early pregnancy and preterm birth. Future research should investigate potential mechanisms of this association, including chronic placental inflammation.

Acknowledgements

We would like to acknowledge the US National Archives and Records Administration for providing access to study data from the Collaborative Perinatal Project.

Funding

This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (award number F32HD100076) and the National Institute on Minority Health and Health Disparities (award number R01MD011749). The funding sources were not involved in the study design, data analysis, or manuscript writing.

Footnotes

Conflict of Interest The authors declare that they have no conflict of interest.

1

A single work (unpublished thesis) has previously evaluated the impact of the JFK assassination on preterm birth using a sample that included data from the CPP (in addition to data from the Child Health and Development Study) The authors reported that exposure early in pregnancy was associated with preterm birth (HR: 1.16, 95% CI: 1.04, 1.29), but exposure in mid and late pregnancy were not associated with preterm birth. Our results differ slightly due different samples and modeling approaches (Kogut, 2004).

References

  1. Ahrens KA, Louik C, Kerr S, Mitchell AA, & Werler MM (2014). Seasonal influenza vaccination during pregnancy and the risks of preterm delivery and small for gestational age birth. Paediatric and Perinatal Epidemiology, 28(6), 498–509. 10.1111/ppe.12152 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allison PD (2005). Fixed effects regression methods for longitudinal data using SAS. SAS Institute. [Google Scholar]
  3. Azagba S, & Sharaf MF (2011). Psychosocial working conditions and the utilization of health care services. BMC Public Health, 11, 642. 10.1186/1471-2458-11-642 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Banta TJ (1964). The Kennedy assassination: Early thoughts and emotions. The Public Opinion Quarterly, 28(2), 216–224 [Google Scholar]
  5. Barrios YV, Sanchez SE, Qiu C, Gelaye B, & Williams MA (2014). Risk of spontaneous preterm birth in relation to maternal experience of serious life events during pregnancy. Int J Womens Health, 6, 249–257. 10.2147/IJWH.S54269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bronson SL, & Bale TL (2016). The placenta as a mediator of stress effects on neurodevelopmental reprogramming. Neuropsychopharmacology, 41(1), 207–218. 10.1038/npp.2015.231 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Catalano R, & Hartig T (2001). Communal bereavement and the incidence of very low birthweight in Sweden. J Health Soc Behav, 42(4), 333–341. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/11831135 [PubMed] [Google Scholar]
  8. Chehade H, Simeoni U, Guignard JP, & Boubred F (2018). Preterm birth: Long term cardiovascular and renal consequences. Current Pediatric Reviews. 10.2174/1573396314666180813121652 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dahlerup BR, Egsmose EL, Siersma V, Mortensen EL, Hedegaard M, Knudsen LE, & Mathiesen L (2018). Maternal stress and placental function, a study using questionnaires and biomarkers at birth. PLoS ONE, 13(11), e0207184. 10.1371/journal.pone.0207184 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Darrow LA, Strickland MJ, Klein M, Waller LA, Flanders WD, Correa A, Marcus M, & Tolbert PE (2009). Seasonality of birth and implications for temporal studies of preterm birth. Epidemiology, 20(5), 699–706. 10.1097/EDE.0b013e3181a66e96 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dunkel Schetter C, & Glynn LM (2011). Stress in pregnancy: Empirical evidence and theoretical issues to guide interdisciplinary research. In Contrada RJ & Baum A (Eds.), The handbook of stress science: Biology, psychology, and health. (pp. 321–343). Springer Pub. [Google Scholar]
  12. Endara SM, Ryan MA, Sevick CJ, Conlin AM, Macera CA, & Smith TC (2009). Does acute maternal stress in pregnancy affect infant health outcomes? Examination of a large cohort of infants born after the terrorist attacks of September 11, 2001. BMC Public Health, 9, 252. 10.1186/1471-2458-9-252 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ernst LM, Grobman WA, Wolfe K, Huang MH, McDade TW, Holl JL, & Borders AE (2013). Biological markers of stress in pregnancy: Associations with chronic placental inflammation at delivery. American Journal of Perinatology, 30(7), 557–564. 10.1055/s-0032-1329187 [DOI] [PubMed] [Google Scholar]
  14. Galea S (2007). The long-term health consequences of disasters and mass traumas. CMAJ, 176(9), 1293–1294. 10.1503/cmaj.070368 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Glynn LM, Schetter CD, Wadhwa PD, & Sandman CA (2004). Pregnancy affects appraisal of negative life events. Journal of Psychosomatic Research, 56(1), 47–52. 10.1016/S0022-3999(03)00133-8 [DOI] [PubMed] [Google Scholar]
  16. Goldenberg RL, Culhane JF, Iams JD, & Romero R (2008). Epidemiology and causes of preterm birth. Lancet, 371(9606), 75–84. 10.1016/S0140-6736(08)60074-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Grech V (2015). Ethnic differences in birth gender ratio responses in the United States after the September 11 Attacks and the President Kennedy assassination. Early Human Development, 91(12), 829–836. 10.1016/j.earlhumdev.2015.10.010 [DOI] [PubMed] [Google Scholar]
  18. Hardy JB (2003). The collaborative perinatal project: Lessons and legacy. Ann Epidemiol, 13(5), 303–311. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/12821268 [DOI] [PubMed] [Google Scholar]
  19. Harville E, Xiong X, & Buekens P (2010). Disasters and perinatal health: A systematic review. Obstetrical & Gynecological Survey, 65(11), 713–728. 10.1097/OGX.0b013e31820eddbe [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Harville EW, & Do M (2016). Reproductive and birth outcomes in haiti before and after the 2010 earthquake. Disaster Medicine and Public Health Preparedness, 10(1), 59–66. 10.1017/dmp.2015.69 [DOI] [PubMed] [Google Scholar]
  21. Hinkle SN, Albert PS, Mendola P, Sjaarda LA, Yeung E, Boghossian NS, & Laughon SK (2014). The association between parity and birthweight in a longitudinal consecutive pregnancy cohort. Paediatric and Perinatal Epidemiology, 28(2), 106–115. 10.1111/ppe.12099 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Keenan-Devlin LS, Ernst LM, Ross KM, Qadir S, Grobman WA, Holl JL, Crockett A, Miller G, & Borders AEB (2017). Maternal income during pregnancy is associated with chronic placental inflammation at birth. American Journal of Perinatology, 34(10), 1003–1010. 10.1055/s-0037-1601353 [DOI] [PubMed] [Google Scholar]
  23. Khong TY, Mooney EE, Ariel I, Balmus NC, Boyd TK, Brundler MA, Derricott H, Evans MJ, Faye-Petersen OM, Gillan JE, Heazell AEP, Heller DS, Jacques SM, Keating S, Kelehan P, Maes A, McKay EM, Morgan TK, Nikkels PGJ, … Gordijn SJ (2016). Sampling and definitions of placental lesions: Amsterdam placental workshop group consensus statement. Archives of Pathology and Laboratory Medicine, 140(7), 698–713. 10.5858/arpa.2015-0225-CC [DOI] [PubMed] [Google Scholar]
  24. Kingsbury M, Weeks M, MacKinnon N, Evans J, Mahedy L, Dykxhoorn J, & Colman I (2016). Stressful Life Events During Pregnancy and Offspring Depression: Evidence From a Prospective Cohort Study. J Am Acad Child Adolesc Psychiatry, 55(8), 709–716. 10.1016/j.jaac.2016.05.014 [DOI] [PubMed] [Google Scholar]
  25. Kogut KR (2004). Does a prenatal “shock” lead to preterm delivery? A study on the effect of John F. Kennedy’s assassination. (Master of Public Health). University of California, Berkeley, [Google Scholar]
  26. Kramer MS (1987). Determinants of low birth weight: methodological assessment and meta-analysis. Bull World Health Organ, 65(5), 663–737. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/3322602 [PMC free article] [PubMed] [Google Scholar]
  27. Kramer MS, McLean FH, Boyd ME, & Usher RH (1988). The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. JAMA, 260(22), 3306–3308. Retrieved from https://www.ncbi.nlm.nih.gov/pub-med/3054193 [PubMed] [Google Scholar]
  28. Lahti-Pulkkinen M, Cudmore MJ, Haeussner E, Schmitz C, Pesonen AK, Hamalainen E, Villa PM, Mehtälä S, Kajantie E, Laivuori H, Reynolds RM, Frank H-G, & Raikkonen K (2018). Placental morphology is associated with maternal depressive symptoms during pregnancy and toddler psychiatric problems. Science and Reports, 8(1), 791. 10.1038/s41598-017-19133-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Laplante DP, Brunet A, & King S (2016). The effects of maternal stress and illness during pregnancy on infant temperament: Project Ice Storm. Pediatric Research, 79(1–1), 107–113. 10.1038/pr.2015.177 [DOI] [PubMed] [Google Scholar]
  30. Lash TL, Fox MP, & Fink AK (2009). Chapter 10: Presentation and Inference. Applying Quantitative Bias Analysis to Epidemiologic Data. (pp. 175–181). Dordrecht: /Springer. [Google Scholar]
  31. Laszlo KD, Li J, Olsen J, Vestergaard M, Obel C, & Cnattingius S (2016). Maternal bereavement and the risk of preterm delivery: The importance of gestational age and of the precursor of preterm birth. Psychological Medicine, 46(6), 1163–1173. 10.1017/S0033291715002688 [DOI] [PubMed] [Google Scholar]
  32. Lundgren EM, & Tuvemo T (2008). Effects of being born small for gestational age on long-term intellectual performance. Best Practice & Research Clinical Endocrinology & Metabolism, 22(3), 477–488. 10.1016/j.beem.2008.01.014 [DOI] [PubMed] [Google Scholar]
  33. Marinescu IP, Foarfa MC, Pirlog MC, & Turculeanu A (2014). Prenatal depression and stress - Risk factors for placental pathology and spontaneous abortion. Rom J Morphol Embryol, 55(3 Suppl), 1155–1160. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/25607399 [PubMed] [Google Scholar]
  34. Murphy SL, Mathews TJ, Martin JA, Minkovitz CS, & Strobino DM (2017). Annual summary of vital statistics. Pediatrics. 10.1542/peds.2016-3239 [DOI] [PubMed] [Google Scholar]
  35. Niswander KR, Gordon M, & National Institute of Neurological Diseases and Stroke. (1972). The women and their pregnancies; the Collaborative Perinatal Study of the National Institute of Neurological Diseases and Stroke. Washington: National Institute of Health; for sale by the Supt. of Docs., U.S. Govt. Print. Off. [Google Scholar]
  36. Palmeiro-Silva YK, Orellana P, Venegas P, Monteiro L, Varas-Godoy M, Norwitz E, Rice G, Osorio E, & Illanes SE (2018). Effects of earthquake on perinatal outcomes: A Chilean register-based study. PLoS One, 13(2), e0191340. 10.1371/journal.pone.0191340 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Qiu X, Lodha A, Shah PS, Sankaran K, Seshia MM, Yee W, Jefferies A, & Canadian Neonatal N (2012). Neonatal outcomes of small for gestational age preterm infants in Canada. American Journal of Perinatology, 29(2), 87–94. 10.1055/s-0031-1295647 [DOI] [PubMed] [Google Scholar]
  38. Qu F, Wu Y, Zhu YH, Barry J, Ding T, Baio G, Muscat R, Todd BK, Wang F-F, & Hardiman PJ (2017). The association between psychological stress and miscarriage: A systematic review and meta-analysis. Science and Reports, 7(1), 1731. 10.1038/s41598-017-01792-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rich-Edwards JW, Kleinman KP, Strong EF, Oken E, & Gillman MW (2005). Preterm delivery in Boston before and after September 11th, 2001. Epidemiology, 16(3), 323–327. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/15824547 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Ronald A, Pennell CE, & Whitehouse AJ (2010). Prenatal maternal stress associated with ADHD and autistic traits in early childhood. Frontiers in Psychology, 1, 223. 10.3389/fpsyg.2010.00223 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Saigal S, & Doyle LW (2008). An overview of mortality and sequelae of preterm birth from infancy to adulthood. Lancet, 371(9608), 261–269. 10.1016/S0140-6736(08)60136-1 [DOI] [PubMed] [Google Scholar]
  42. Seri I, & Evans J (2008). Limits of viability: Definition of the gray zone. Journal of Perinatology, 28(Suppl 1), S4–8. 10.1038/jp.2008.42 [DOI] [PubMed] [Google Scholar]
  43. Shapiro GD, Fraser WD, Frasch MG, & Seguin JR (2013). Psychosocial stress in pregnancy and preterm birth: Associations and mechanisms. Journal of Perinatal Medicine, 41(6), 631–645. 10.1515/jpm-2012-0295 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sheatsley PB, & Feldman JJ (1964). The assassination of president Kennedy: A preliminary report on public reactions and behavior. The Public Opinion Quarterly, 28(2), 189–215 [Google Scholar]
  45. Sherrieb K, & Norris FH (2013). Public health consequences of terrorism on maternal-child health in New York City and Madrid. Journal of Urban Health, 90(3), 369–387. 10.1007/s11524-012-9769-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Shirazi M, Pooransari P, Rahimi Sharbaf F, Niromanesh SH, Sahebdel B, Shariat M, Pahlavan Z, & Ahmadian M (2019). Correlation of maternal stress because of positive aneuploidy screening serum analytes and uterine arteries’ Doppler ultrasound index: A prospective cohort study. International Journal of Fertility & Sterility, 12(4), 329–334. 10.22074/ijfs.2019.5359 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Suzuki K, Yamagata Z, Kawado M, & Hashimoto S (2016). Effects of the great East Japan earthquake on secondary sex ratio and perinatal outcomes. Journal of Epidemiology, 26(2), 76–83. 10.2188/jea.JE20150055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tegethoff M, Greene N, Olsen J, Meyer AH, & Meinlschmidt G (2010a). Maternal psychosocial adversity during pregnancy is associated with length of gestation and offspring size at birth: Evidence from a population-based cohort study. Psychosomatic Medicine, 72(4), 419–426. 10.1097/PSY.0b013e3181d2f0b0 [DOI] [PubMed] [Google Scholar]
  49. Tegethoff M, Greene N, Olsen J, Meyer AH, & Meinlschmidt G (2010b). Maternal psychosocial stress during pregnancy and placenta weight: Evidence from a national cohort study. PLoS One, 5(12), e14478. 10.1371/journal.pone.0014478 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Teixeira JM, Fisk NM, & Glover V (1999). Association between maternal anxiety in pregnancy and increased uterine artery resistance index: Cohort based study. BMJ, 318(7177), 153–157. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/9888905 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Torche F (2011). The effect of maternal stress on birth outcomes: Exploiting a natural experiment. Demography, 48(4), 1473–1491. 10.1007/s13524-011-0054-z [DOI] [PubMed] [Google Scholar]
  52. Truijens SE, Boerekamp CA, Spek V, van Son MJ, Oei SG, & Pop VJ (2015). Increased levels of depressive symptoms among pregnant women in The Netherlands after the crash of flight MH17. American Journal of Epidemiology, 182(5), 426–430. 10.1093/aje/kwv161 [DOI] [PubMed] [Google Scholar]
  53. Tudiver F, Permaul-Woods JA, Hilditch J, Harmina J, & Saini S (1995). Do widowers use the health care system differently? Does intervention make a difference? Can Fam Physician, 41, 392–400. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/7773026 [PMC free article] [PubMed] [Google Scholar]
  54. Wadhwa PD, Entringer S, Buss C, & Lu MC (2011). The contribution of maternal stress to preterm birth: Issues and considerations. Clinics in Perinatology, 38(3), 351–384. 10.1016/j.clp.2011.06.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wainstock T, Anteby E, Glasser S, Shoham-Vardi I, & Lerner-Geva L (2013). The association between prenatal maternal objective stress, perceived stress, preterm birth and low birthweight. The Journal of Maternal-Fetal & Neonatal Medicine, 26(10), 973–977. 10.3109/14767058.2013.766696 [DOI] [PubMed] [Google Scholar]
  56. Zhang W, Rajendran K, Ham J, Finik J, Buthmann J, Davey K, Pehme PM, Dana K, Pritchett A, Laws H, & Nomura Y (2018). Prenatal exposure to disaster-related traumatic stress and developmental trajectories of temperament in early childhood: Superstorm Sandy pregnancy study. Journal of Affective Disorders. 10.1016/j.jad.2018.02.067 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES