Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 1.
Published in final edited form as: Ann Epidemiol. 2024 Jan 23;91:30–36. doi: 10.1016/j.annepidem.2024.01.007

Understanding the role of childhood nurture, abuse, and stability on gestational diabetes in the Coronary Artery Risk Development in Young Adults study (CARDIA)

Kaitlyn K Stanhope a, Erica P Gunderson b, Shakira F Suglia c, Sheree L Boulet a, Denise J Jamieson a, Catarina I Kiefe d, Kiarri N Kershaw e
PMCID: PMC10922764  NIHMSID: NIHMS1961977  PMID: 38266664

Abstract

Background:

To estimate associations between facets of the maternal childhood family environment with gestational diabetes (GDM) and to test mediation by pre-pregnancy waist circumference.

Methods:

We used data from CARDIA, a cohort of individuals aged 18-30 years at baseline (1985-86), followed over 30 years (2016). We included participants with one or more pregnancies ≥ 20 weeks after baseline, without pre-pregnancy diabetes. The primary exposure was the Childhood Family Environment Scale (assessed year 15), including the total score and abuse, nurture, and stability subscales as continuous, separate exposures. The outcome was GDM (self-reported at each visit for each pregnancy). We fit log binomial models with generalized estimating equations to calculate risk ratios (RR) and 95% confidence intervals (CI), adjusting for age at delivery, parity, race (Black or White), and parental education. We used regression models with bootstrapped CIs to test mediation and effect modification by excess abdominal adiposity at the last preconception CARDIA visit (waist circumference ≥ 88 cm).

Results:

We included 1033 individuals (46% Black) with 1836 pregnancies. 130 pregnancies (7.1%) were complicated by GDM. For each 1 point increase on the abuse subscale (e.g., from “rarely or never” to “some or little of the time”) there was a 30% increased risk of GDM (RR: 1.3, 95% CI: 1.0, 1.7). There was evidence of effect modification but not mediation by preconception abdominal adiposity.

Conclusions:

A more adverse childhood family environment was associated with increased risk of GDM, with a stronger association among individuals with preconception waist circumference ≥ 88 cm.

Keywords: Diabetes, gestational, Adverse Childhood Experiences, Adiposity, Life Course Perspective, Mediation Analysis, Cohort study

Introduction

People who have been exposed to many stressful events across the life course may enter pregnancy at a greater biologic age, with ongoing vascular, neuroendocrine, and metabolic dysfunction that increases risk for poor maternal and infant outcomes.1–3 Early life stress, including deprivation and maltreatment, may have particularly negative effects on health prior to and during pregnancy. In order to develop prevention strategies, we need a clear understanding of the pathways through which preconception stressful experiences lead to poor perinatal outcomes.

In the United States, 5.8% of pregnant people are diagnosed with gestational diabetes.4–6 Gestational diabetes (GDM) is associated with elevated risk of maternal and infant complications.7 Following pregnancy, GDM is associated with cardiometabolic dysfunction for the birthing person (e.g., diabetes, dyslipidemia, hypertension) and, possibly, for the child (obesity).8–12 Preconception cardiometabolic and mental health may drive GDM risk with risk factors including adiposity, dysglycemia (i.e., prediabetes), hyperinsulinemia, and depression.13–15

Stressful experiences, particularly those occurring in early life, may also increase risk of GDM. Childhood adversity, including trauma and other stressful experiences, results in long-term physiologic changes, including metabolic dysfunction and obesity.16,17 In turn, these may result in worse outcomes during pregnancy, including GDM. Few studies have explored the relationship between childhood adversity and GDM; and evidence to date is inconsistent.18–20 Further, studies to date have not explored whether preconception health may account for all or some of the observed association between childhood adversity and GDM.

The goal of this study was to estimate the association of childhood family environment with gestational diabetes among female participants in the Coronary Artery Risk Development in Young Adults study (CARDIA) who had a pregnancy ≥ 20 weeks gestation over the study period (1985-2016) and to test whether the association is explained by effect modification or mediation by preconception abdominal adiposity.

Research Design and Methods

Study Population

This study is a secondary data analysis of data from CARDIA, an observational prospective cohort study of 5115 young adults recruited in 1985-1986 in Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California in the United States. Following enrollment and informed consent, the study followed young adults every 2-5 years through the present through in-person assessments and surveys (1985-2016). Over time, roughly 70% of individuals have completed follow-up visits (follow-up rates: Y2: 90%, Y5: 85%, Y7: 80%, Y10: 77%, Y15: 72%, Y20: 69%, Y25: 68%, Y30: 66%).

The research question and analytic plan for the current study was approved by the CARDIA publications committee prior to beginning the analysis. For the current study, we restricted the study population to women who had at least one pregnancy after the baseline exam (at ages 18-30), and complete data on childhood family environment, measured at year 15 (2000-2001) (Figure 1). The unit of analysis was pregnancy and we included all post-baseline pregnancies of 20 weeks gestation or greater with complete data on GDM (yes or no) among women without overt diabetes prior to the pregnancy. The CARDIA study received approval from the Institutional Review Boards of each study site and this analysis was reviewed and approved by the [deidentified] Institutional Review Board (00001397).

Figure 1.

Figure 1.

Open in a new tab

Creation of analytic dataset beginning with 1951 ever parous individuals in CARDIA.

Measures

The primary exposure was the Childhood Family Environment (CFE) scale, a 7-question measure of childhood adversity developed for CARDIA and based on the Adverse Childhood Experience Scale. As have prior authors,21 we categorized the CFE scale into three subscales: household stability, abuse, and nurture. For each subscale, we averaged the questions within the subscale to produce a value, 0-3 for the subscale. We also present the total CFE scale as a continuous measure. See table 1 for questions included in each subscale. For descriptive analyses, we divide the total CFE scale into tertiles.

Table 1.

Characteristics of individuals with at least one pregnancy > 20 weeks’ gestation following baseline (1985-2016) and follow-up through year 15 (2000-2001), by GDM status: Coronary Artery Risk Development in Young Adults study (CARDIA) (N=1,033)

Total Population No lifetime GDM births 1+ lifetime GDM births
N 1033 897 134
Characteristic n (%) or mean (SD) n (%) or mean (SD) n(%) or mean (SD)
White 53.8 (556) 53.2 (477) 58.1 (79)
Black 46.2 (477) 46.8 (420) 41.9 (57)
Childhood Family Environment Score*, mean (SD) 16.5 (4.2) 16.6 (4.2) 15.9 (4.2)
Abuse subscale mean score, mean (SD) 2.6 (0.7) 2.6 (0.7) 2.4 (0.8)
How often were you marked from getting hit?
   Rarely or none of the time 82.6 (853) 83.6 (750) 75.7 (103)
   Some or little of the time 10.8 (111) 10.7 (96) 11 (15)
   Occasionally or moderate amount of time 4.7 (49) 3.8 (34) 11 (15)
   Most or all of the time 1.9 (20) 1.9 (17) 2.2 (3)
How after were you sworn at or insulted?
   Rarely or none of the time 61.1 (631) 62.4 (560) 52.2 (71)
   Some or little of the time 19.7 (203) 19.3 (173) 22.1 (30)
   Occasionally or moderate amount of time 14.1 (146) 13.6 (122) 17.7 (24)
   Most or all of the time 5.1 (53) 4.7 (42) 8.1 (11)
Nurture mean score, mean (SD) 2.2 (0.9) 2.2 (0.8) 2.2 (0.9)
How often did you feel loved?
   Rarely or none of the time 3.7 (38) 4.0 (36) 1.5 (2)
   Some or little of the time 10.3 (106) 9.5 (85) 15.4 (21)
   Occasionally or moderate amount of time 21.5 (222) 21.4 (192) 22.1 (30)
   Most or all of the time 64.6 (667) 65.1 (584) 61 (83)
How often did you get physical affection?
   Rarely or none of the time 11.8 (122) 11.4 (102) 14.7 (20)
   Some or little of the time 18.9 (195) 18.4 (165) 22.1 (30)
   Occasionally or moderate amount of time 27.3 (282) 28.2 (253) 21.3 (29)
   Most or all of the time 42 (434) 42 (377) 41.9 (57)
Stability mean score, mean (SD) 3.5 (1.1) 3.5 (1.1) 3.4 (1.1)
Did family know what you were up to?
   Rarely or none of the time 6.4 (66) 6.0 (54) 8.8 (12)
   Some or little of the time 13.7 (141) 13.7 (123) 13.2 (18)
   Occasionally or moderate amount of time 24.1 (249) 23.5 (211) 27.9 (38)
   Most or all of the time 55.9 (577) 56.7 (509) 50.0 (68)
Was your house well-organized/managed?
   Rarely or none of the time 6.1 (63) 6.5 (58) 3.7 (5)
   Some or little of the time 12.4 (128) 11.9 (107) 15.4 (21)
   Occasionally or moderate amount of time 23.5 (243) 23.1 (207) 26.5 (36)
   Most or all of the time 58 (599) 58.5 (525) 54.4 (74)
How often did you live with an alcohol or drug abuser?
   Rarely or none of the time 69.7 (720) 70.1 (629) 66.9 (91)
   Some or little of the time 8.7 (90) 9 (81) 6.6 (9)
   Occasionally or moderate amount of time 7.5 (77) 7.1 (64) 9.6 (13)
   Most or all of the time 14.1 (146) 13.7 (123) 16.9 (23)
Maternal Educational Attainment
   <12 years 5.5 (57) 5.6 (50) 5.2 (7)
   12-15 years 60.9 (629) 61.5 (552) 56.6 (77)
   16+ years 33.6 (347) 32.9 (295) 38.2 (52)
Paternal Education Attainment
   <12 years 10.3 (106) 10.7 (96) 7.4 (10)
   12-15 years 51.4 (531) 50.7 (455) 55.9 (76)
   16+ years 38.3 (396) 38.6 (346) 36.8 (50)
Social Support, mean (SD) 1.5 (0.7) 1.5 (0.7) 1.6 (0.6)
John Henryism, continuous†, mean (SD) 48.8 (5.6) 49 (5.5) 47.6 (5.9)
Indicated active coping (>49)‡ 55.8 (574) 56.6 (506) 50.4 (68)
Marital status at baseline exam 26.9 (278) 26.5 (238) 29.4 (40)
Total pregnancies (lifetime), mean (SD) 2.3 (1.1) 2.2 (1.1) 2.3 (1)
Family history of diabetes 15.1 (137) 14.1 (111) 21.5 (26)
Open in a new tab
*

For childhood family environment scale and subscale, all are oriented so higher scores indicate a more positive context (e.g., less abuse, more affection)

†

James, 1994, Culture, Medicine and Psychiatry, Missing from 4 observations

The primary outcome was GDM. Participants self-reported GDM at each CARDIA visit for each pregnancy since the last study visit. In a validation sub-study comparing self-report to medical records, self-reported GDM was found to be highly accurate with 100% sensitivity and 92% specificity in the validation sample.22 Individuals were excluded if they developed overt diabetes prior to the first post-baseline pregnancy (n= 10). We also excluded subsequent pregnancies from the analysis if women developed overt diabetes after prior pregnancy (n=39). Diabetes classification was based on standardized biochemical testing for glucose tolerance at research visits every 5 years (i.e., fasting and 2-h 75 g OGTT serum glucose, HbA1c, or self-report of diabetes diagnosis and medication use).23

We considered the primary potential confounder of childhood adversity on GDM to be childhood socioeconomic status.24 To capture this, we control for parental education, measured at baseline, categorizing the highest level of education attained by either parent as less than high school (<12 years), high school graduate without college (12-15 years), or greater than high school (16+ years). We further control for age at pregnancy (binary, greater than 35 or not) and parity (0, 1, 2+ prior births) and race (self-reported Black or white) to improve precision.

The association of childhood adversity with adult outcomes may vary by cultural context, individual coping strategies, and adult life experiences.16 Thus, we considered two potential effect modifiers. First, individuals of different racial backgrounds may experience the impact of childhood adversity differently due to the excess burden of additional structural stressors experienced by Black people across the life course.3 Thus, we considered interaction by race (Black or white). Further, coping style may modify this relationship.25 The John Henryism scale measures the extent to which an individual uses high-effort or passive coping strategies to deal with stress. Prior research has suggested that a high-effort coping strategy in the face of persistent stressors and inadequate resources may contribute to poor cardiovascular health.25,26

We conducted a mediation analysis to explore whether some or the entire association between childhood family environment and GDM could be explained by an indirect association through preconception abdominal adiposity. We defined abdominal adiposity as waist circumference at the last CARDIA visit prior to the index pregnancy greater than or equal to 88 cm. To identify preconception visits, we chose the most recent CARDIA visit at least 38 weeks (266 days) prior to the delivery date. We selected waist circumference as an indicator of poor metabolic health due to its interpretability and strong association with mortality and diabetes across the life course.27 Childhood adversity is strongly associated with metabolic health across the life course;16 thus, it could be that individuals with more adverse childhood family environments have poor metabolic health prior to pregnancy that, in turn, elevates the risk of GDM. Testing a mediating pathway can inform secondary prevention strategies and our understanding of potential mechanisms.

Analysis

We conducted descriptive analysis of demographic, pregnancy, and preconception characteristics, stratified by tertile of CFE scale. To estimate associations between each CFE subscale (abuse, stability, nurture, total) and GDM, we fit log binomial models with each CFE subscale as the primary independent variable and GDM as the dependent variable. We accounted for repeated pregnancies for individuals by using generalized estimating equations (GEE) and robust standard errors. We fit sequential models for each subscale and the total scale – (1) unadjusted; (2) controlling for age (binary – maternal age > 35 yes/no), parity, and parental education attainment. We assessed interaction on the multiplicative scale by including interaction terms in the model and calculating strata-specific estimates separately for each potential effect modifier (race, coping style, abdominal adiposity). We calculated risk ratios and 95% confidence intervals. To facilitate comparison of associations between models, all risk ratios for the subscales represent the impact of a one unit change on the scale towards more adversity (e.g., 1 point higher on the abuse scale or 1 point lower on the nurture scale). For the overall scale, risk ratios represent a 1 standard deviation (4.2 points) change in the scale towards a more adverse environment. We conducted supplemental analyses: (1) excluding multiple gestation births to determine whether the inclusion of multiple gestations influenced the results and (2) with alternate (binary) exposure categorization to facilitate comparison with other literature.28

To test mediation by preconception abdominal adiposity, we used regression-based mediation analyses fitting separate models using GEE for the exposure mediator and exposure outcome relationships, using methods described by Vanderweele and Valeri.29,30 Using these models, we calculated natural direct and indirect effects of each CFE subscale and total scale as potentially mediated by abdominal adiposity (waist circumference ≥ 88). Further, we calculated controlled direct effects to determine the association of childhood family environment with GDM if no one had abdominal adiposity (waist circumference ≥ 88) at the last preconception CARDIA. Controlled direct effects measure both mediation and effect modification. We used the controlled direct effect, setting abdominal adiposity to absent, to estimate the percent of the total effect that would be eliminated if we were able to prevent abdominal adiposity. We used bootstrapping to create 95% confidence intervals for each parameter (1000 bootstrapped samples). We used SAS 9.4 for all analyses (SAS Institute Inc., Cary, NC).

Results

Our analytic sample included 1836 pregnancies greater than 20 weeks following CARDIA’s baseline visit to 1033 individuals (Figure 1). During the study follow up period (1985-2016), 13.1% (136) of all individuals ever experienced GDM at any pregnancy. Overall, individuals who experienced GDM in one or more pregnancies had a slightly lower (worse) childhood family environment score (mean: 15.9 (SD: 4.2) compared to 16.6 (SD: 4.2)) than those who never experienced GDM with similar patterns for the abuse, nurture, and stability subscales. Individuals who ever experienced GDM reported higher maternal educational attainment and slightly higher social support scores. The two groups reported similar marital status at baseline and similar lifetime parity.

We present characteristics of included pregnancies and cardiometabolic function at the last CARDIA visit prior to each pregnancy for 1836 included pregnancies greater than or equal to 20 weeks (Table 2). Overall, 7.1% (130) of pregnancies were complicated by GDM. Age at pregnancy, gestational age, and parity did not vary meaningfully by childhood family environment scale. Rates of multiple gestation were higher in pregnancies in the highest (most favorable) tertile of childhood family environment compared to the lowest. Overall, preconception indicators of cardiometabolic function were similar across tertiles of childhood family environment.

Table 2.

Characteristics of Pregnancies to individuals with at least one pregnancy > 20 weeks following baseline (1985-2016) and follow-up through year 15 (2000-2001), by tertiles of Childhood Family Environment (CFE): Coronary Artery Risk Development in Young Adults study (CARDIA)

Total Missing obs. Tertile of CFE
Low Medium High
N pregnancies 1836 585 697 554
N individuals 1033 331 394 308
Pregnancy characteristics Median (25th, 75th percentile) or % (n) Median (25th, 75th percentile) or % (n) Median (25th, 75th percentile) or % (n) Median (25th, 75th percentile) or % (n)
   Gestational diabetes, % (n) 7.1 (130) 0 7.5 (44) 8 (56) 5.4 (30)
   Black race, % (n) 44.1 (809) 43.1 (252) 49.2 (343) 38.6 (214)
   Age at pregnancy, median (25th, 75th percentile) 31 (28, 35) 0 31 (28, 34) 31 (28, 34) 31 (28, 35)
   Gestational age (weeks), median (25th, 75th percentile) 40 (38, 40) 9 40 (38, 40) 40 (38, 40) 40 (38, 40)
   Gravidity at last visit, median (25th, 75th percentile) 1 (1, 3) 0 2 (1, 3) 1 (1, 3) 1 (0, 2)
   Prior pregnancies (>20 weeks), median (25th, 75th percentile) 1 (0, 1) 0 1 (0, 2) 1 (0, 1) 0 (0, 1)
   Multiple gestation 2.5 (43) 88 2.2 (12) 2.1 (14) 3.2 (17)
Cardiometabolic function at last pre-pregnancy CARDIA visit
   Days between preconception visit and delivery 754 (496, 1053) 0 747 (485, 1037) 741 (486, 1008) 776 (535, 1112)
   BMI, median (25th, 75th percentile) 23.2 (21, 27.3) 0 23.5 (21.2, 27.5) 23.2 (21, 27.4) 22.8 (20.8, 27)
   Class 3 Obesity (40+), % (n) 51 (2.78, ) 0 18 (3.08, ) 19 (2.73, ) 14 (2.53, )
   Fasting glucose, median (25th, 75th percentile)* 81 (76, 86) 33 80 (76, 86) 80 (75, 85) 81 (77, 86)
   Fasting insulin, median (25th, 75th percentile) 7.8 (6.2, 10.2) 38 7.8 (6.3, 10.4) 7.8 (6.1, 10) 8 (6, 10.4)
   HOMA-IR, median (25th, 75th percentile)† 1.5 (1.2, 2.1) 47 1.5 (1.2, 2.1) 1.5 (1.2, 2) 1.6 (1.2, 2.1)
   Triglycerides, median (25th, 75th percentile) 60 (44, 84) 36 60 (45, 84) 59 (44, 85) 60 (44, 82)
   LDL cholesterol, median (25th, 75th percentile) 105 (85, 124) 36 102.5 (84, 124) 108 (86, 125) 104 (85, 123)
   HDL cholesterol, median (25th, 75th percentile) 56 (47, 66) 48 55 (47, 66) 55 (46, 67) 56 (49, 65)
   Waist circumference, median (25th, 75th percentile) 73 (67.5, 81.5) 0 74 (68.5, 82) 73 (67.3, 81.8) 72 (67, 80.3)
   Metabolic syndrome present†, % (n) 4.2 (74) 56 3.7 (21) 4.4 (30) 4.3 (23)
   Waist circumference ≥ 88, % (n) 15.7 (289) 0 16.2 (95) 17.2 (120) 13.4 (74)
Open in a new tab

Abbreviations CFE – childhood family environment, obs – observations, IQR – interquartile range

*

For pregnancies between visits 2-5 or 5-7, the baseline measure of fasting glucose and insulin was used.

†

(fasting insulin*fasting glucose)/405, a measure of insulin resistance (Matthews, 1985)

‡

At least three of the following risk factors prior to the index pregnancy: waist circumference greater than 88 cm, triglycerides greater than 150, HDL cholesterol below 50, blood pressure greater than 130/85, or fasting glucose greater than 100 (ATP-III Guidelines, 2001)

In both unadjusted and adjusted models, worse childhood family environment was associated with an increased risk of GDM, with the strongest pattern for the abuse subscale (Table 3). For the abuse subscale, the adjusted risk of GDM was 30% higher (total effect RR: 1.30, 95% CI: 1.02, 1.65) for a 1-unit worse childhood family environment (total scale range 0-3), adjusting for race, age at pregnancy, and parity at pregnancy. We estimated smaller, non-significant elevations in risk for the nurture and stability subscales, and the overall score. There was no evidence of heterogeneity of effect stratifying by race or John Henryism coping style (active v. passive) (Table S1). Restricting to singleton pregnancies resulted in similar results, albeit with slightly less precision (Table S2). Binary exposure definitions resulted in similar, albeit not statistically significant, estimates.

Table 3.

Estimated associations between childhood family environment total scale and subscales and mediation decomposition, including estimated total effect, controlled and natural direct effects, and natural indirect effect of childhood family environment scale and subscales on gestational diabetes as potentially mediated by abdominal adiposity at most recent CARDIA visit prior to the pregnancy, n = 1,836 pregnancies to 1,033 individuals

Total Effect NDE NIE CDE, WC ≥ 88 cm CDE, WC < 88 cm Percent Eliminated†
Subscale RR*, 95% CI RR*, 95% CI RR*, 95% CI RR*, 95% CI RR*, 95% CI %, 95% CI
Abuse subscale, 1-unit change 1.3 (1.02, 1.65) 1.31 (1.04, 1.57) 1.02 (1, 1.05) 1.58 (1.12, 2.07) 1.24 (0.97, 1.55) 8.14 (−3.93, 20.5)
Nurture subscale, 1-unit change 1.07 (0.88, 1.31) 1.1 (0.94, 1.29) 1.01 (0.99, 1.03) 1.29 (0.9, 1.64) 1.05 (0.87, 1.3) 3.38 (−10.09, 13.84)
Stability subscale, 1-unit change 1.07 (0.93, 1.24) 1.11 (0.95, 1.24) 1 (0.98, 1.05) 1.45 (1.09, 1.76) 0.99 (0.83, 1.14) 8.65 (−0.93, 16.9)
Total scale, 1-SD change 1.13 (0.96, 1.32) 1.16 (1, 1.33) 1.03 (1, 1.11) 1.35 (1.08, 1.6) 1.1 (1.04, 1.15) 8.57 (0.02, 16.72)
Binary exposures Any maltreatment 1.36 (0.94, 1.96) 1.26 (0.95, 1.98) 0.99 (1.07, 1.07) 1.93 (1.07, 3.91) 1.13 (0.82, 1.9) 0.17 (−0.06, 0.31)
4+ ACEs reported 1.14 (0.74, 1.75) 1.07 (0.7, 1.55) 1.04 (1, 1.11) 1.39 (0.64, 2.31) 0.98 (0.61, 1.57) 0.13 (−0.16, 0.35)
Open in a new tab

Abbreviations: ACE – adverse childhood experiences; CDE – controlled direct effect; WC – preconception waist circumference; NDE – natural direct effect; NIE – natural indirect effect; RR – risk ratio; CI – confidence interval; SD – standard deviation

*

Log binomial models fit with generalized estimating equations accounting for clustering by individual, control for parental education attainment, race, age at pregnancy, parity at pregnancy. Confidence intervals produced through bootstrapping (1000 samples)

†

The percent of the total estimated effect that would be eliminated if abdominal adiposity were absent in the entire population.

We conducted a mediation analysis to decompose the observed association into the direct, indirect, and controlled direct effect mediated by presence or absence of preconception abdominal adiposity (waist circumference ≥ 88) (Table 3). Overall, there was no evidence of mediation by preconception abdominal adiposity but there was evidence of effect modification. The estimated direct effects (risk ratios) were above one for all subscales and the total score, though attenuated compared to the total observed association. The indirect effect of childhood family environment was null (essentially 1) for all subscales and the total scale. We calculated controlled direct effects, setting abdominal adiposity to universally present (CDE, WC ≥ 88 cm) and universally absent (CDE, WC < 88 cm). We observed weaker associations if we set abdominal adiposity to absent (CDE, < 88 cm) and stronger associations if we set abdominal adiposity to present (CDE, WC ≥88 cm). For the stability subscale and total scale, while the overall estimated association did not indicate elevated risk, in individuals with abdominal adiposity (CDE, WC ≥88 cm), there was significantly elevated risk.

Discussion

In this analysis of 1836 pregnancies to 1033 participants in the CARDIA study, a more adverse childhood family environment, particularly that characterized by report of abuse, was directly associated with risk of gestational diabetes (GDM). The observed association between worse childhood family environment and GDM support the hypothesis that experiences across the early life course influence later pregnancy outcomes. Contrary to our hypothesis, we did not observe mediation by preconception abdominal adiposity. There was evidence of effect modification with preconception abdominal adiposity, suggesting that individuals with greater central obesity, possibly visceral fat, prior to pregnancy may be particularly vulnerable to manifestation of the negative effects of childhood adversity.

This study extends prior research on the relationship between childhood adversity and GDM. Two prior studies observed a positive association between childhood adversity and GDM only among individuals experiencing depression during pregnancy.18,19 Other studies have found no relationship between childhood adversity and GDM.20,31 These studies used a binary count of Adverse Childhood Experiences (more or less than four ACEs ever experienced or not) as the exposure definition. We incorporated the frequency of exposure and estimated different effects for different types of adversity, which may account for our positive results. Similar to prior studies of childhood adversity and perinatal outcomes,32,33 the childhood abuse subscale showed the strongest association with GDM. We observed evidence of effect modification but not mediation by preconception abdominal adiposity. The lack of mediation is surprising, as several previous studies have identified a positive association between childhood adversity and adiposity in young women and the association between pre-pregnancy adiposity and GDM is well known.13 One explanation may be the heterogeneity in biochemical risk profiles that have been associated with risk of GDM independent of obesity among women of reproductive age.13 It is also possible that individuals with adverse childhood family environments developed abdominal adiposity at younger ages than those with less adverse environments and that the long exposure drove increased GDM risk. Childhood adversity is associated with the development of greater adiposity at young ages.34 Prospective studies with follow-up beginning in childhood, including biochemical measures, may better disentangle this relationship.

Our research adds to the growing body of work that supports incorporating psychosocial assessments into preconception and prenatal care for risk stratification and referral to comprehensive medical and social services (e.g., support groups).35 Some have proposed screening and refer for childhood adversity.36 However, childhood adversity may explain only a small part of the variation in poor pregnancy outcomes, and it may be challenging to effectively refer patients based on this information alone.37 Alternately, health systems may consider incorporating principles of trauma-informed care,38 including universal screening for posttraumatic stress, anxiety, and depressive symptoms, and universal referral to support services as alternate secondary prevention methods.

This study offers several promising directions for future research. Prospective studies with follow-up beginning in childhood, including biochemical measures, may better disentangle the relationship between childhood adversity, adiposity, and GDM. Further, incorporation of preconception mental and metabolic health measures may help elucidate the interaction observed in other studies between depressive symptoms, childhood adversity, and gestational diabetes.19 Finally, researchers should develop and test interventions to mitigate the impact of childhood adversity on metabolic risk in pregnancy, including screening and potentially counseling.

The results of this analysis should be interpreted in light of several limitations. First, childhood family environment was assessed retrospectively. Previous research has shown inconsistencies between prospectively and retrospectively reported childhood adversity; however, both are associated with poor cardiometabolic health.16,39 Thus, retrospective report of childhood adversity appears to represent a salient construct for poor cardiometabolic health though possibly a different one than prospectively recorded childhood adversity. Second, there may be selection bias due to loss to follow-up prior to the year 15 assessment of childhood family environment. If individuals with worse childhood family environments were more likely to be lost to follow-up (possibly due to transient residence during adulthood, which may be more common among individuals who had experienced childhood adversity) and more likely to have poor pregnancy outcomes, including GDM (plausible, as individuals with poor pregnancy outcomes may be burdened by self-care or childcare and less likely to participate), this may bias observed results towards the null. Third, the measurement of childhood adversity, the childhood family environment scale, is unique to the CARDIA study, making the results challenging to compare with other studies using more common scales (e.g., the Adverse Childhood Experiences (ACEs) scale).40 Reassuringly, prior analyses using the childhood family environment scale have demonstrated similar associations with cardiometabolic health as those observed with ACEs, suggesting that it captures a similar construct.16,21,41,42 Fourth, our exposure definition treats each one unit change as the same (e.g., from “rarely or none” to “some or little” to “occasionally” to “all of the time”), which may not be true. We conducted sensitivity analyses with threshold-based definitions with similar results. Fifth, we do not have accurate prospective measurement of mental health, such as depression. Depression and other mental illnesses may mediate the relationship between childhood adversity and GDM.19 Future research should consider prospective preconception measures of depressive symptoms to explore this. Sixth, while the overall sample size was robust, the sample size for the mediation analysis was underpowered, as few people (15.7%, 289) had excess abdominal adiposity prior to pregnancy. Thus, we interpret those findings with caution and encourage replication in other, larger samples.

This research also leveraged several strengths. CARDIA comprises a diverse cohort with equal representation of Black and white individuals (by design), and a diversity of socioeconomic status and geography. Further, we were able to include high-quality, prospectively measured preconception measurement of cardiometabolic indicators.

Childhood experiences affect health across the life course, including mental and physical health during pregnancy. Primary prevention of childhood adversity, including social programs to prevent childhood abuse, are critical. For secondary prevention of negative sequelae of childhood adversity, clinicians should incorporate trauma-informed care principles into prenatal and preconception care, with particularly attention to individuals with preconception excess abdominal adiposity, who may experience additional risk.38,43 These best practices can facilitate optimal metabolic outcomes for people during pregnancy and throughout the life course and infants regardless of child trauma exposure.

Supplementary Material

1

Funding and Assistance.

KKS was supported by the National Heart, Lung, and Blood Institute (K99HL161355). The Coronary Artery Risk Development in Young Adults Study (CARDIA) is conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with the University of Alabama at Birmingham (HHSN268201800005I & HHSN268201800007I), Northwestern University (HHSN268201800003I), University of Minnesota (HHSN268201800006I), and Kaiser Foundation Research Institute (HHSN268201800004I). The derived pregnancy variables in CARDIA were developed under funding by The National Institute of Diabetes, Digestive, and Kidney Disease (NIDDK) (K01DK059944, R01DK090047, R01DK106201). This manuscript has been reviewed by CARDIA for scientific content.

Kaitlyn Stanhope reports financial support was provided by National Heart Lung and Blood Institute. Erica Gunderson reports financial support was provided by National Heart Lung and Blood Institute.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  • 1.Kramer MR, Hogue CR. What causes racial disparities in very preterm birth? A biosocial perspective. Epidemiol Rev. 2009;31:84–98. 10.1093/ajerev/mxp003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Kramer MR, Strahan AE, Preslar J, et al. Changing the conversation: applying a health equity framework to maternal mortality reviews. American Journal of Obstetrics and Gynecology. 2019;221(6):609.e1–609.e9. doi: 10.1016/j.ajog.2019.08.057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Geronimus AT. The weathering hypothesis and the health of African-American women and infants: evidence and speculations. Ethnicity & disease. 1992;2(3):207–221. [PubMed] [Google Scholar]
  • 4.Selen DJ, Thaweethai T, Schulte CCM, et al. Gestational Glucose Intolerance and Risk of Future Diabetes. Diabetes Care. 2022;46(1):83–91. doi: 10.2337/dc22-1390 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Sacks DA, Hadden DR, Maresh M, et al. Frequency of gestational diabetes mellitus at collaborating centers based on IADPSG consensus panel–recommended criteria: the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study. Diabetes care. 2012;35(3):526–528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Deputy NP, Kim SY, Conrey EJ, Bullard KM. Prevalence and Changes in Preexisting Diabetes and Gestational Diabetes Among Women Who Had a Live Birth - United States, 2012-2016. MMWR Morb Mortal Wkly Rep. 2018;67(43):1201–1207. doi: 10.15585/mmwr.mm6743a2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Dickens LT, Thomas CC. Updates in Gestational Diabetes Prevalence, Treatment, and Health Policy. Curr Diab Rep. 2019;19(6):33. doi: 10.1007/s11892-019-1147-0 [DOI] [PubMed] [Google Scholar]
  • 8.Damm P, Houshmand-Oeregaard A, Kelstrup L, Lauenborg J, Mathiesen ER, Clausen TD. Gestational diabetes mellitus and long-term consequences for mother and offspring: a view from Denmark. Diabetologia. 2016;59(7):1396–1399. doi: 10.1007/s00125-016-3985-5 [DOI] [PubMed] [Google Scholar]
  • 9.Tobias DK, Stuart JJ, Li S, et al. Association of History of Gestational Diabetes With Long-term Cardiovascular Disease Risk in a Large Prospective Cohort of US Women. JAMA Intern Med. 2017;177(12):1735–1742. doi: 10.1001/jamainternmed.2017.2790 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Gunderson EP, Chiang V, Pletcher MJ, et al. History of gestational diabetes mellitus and future risk of atherosclerosis in mid-life: the Coronary Artery Risk Development in Young Adults study. Journal of the American Heart Association. 2014;3(2):e000490. doi: 10.1161/JAHA.113.000490 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Gunderson EP, Sun B, Catov JM, et al. Gestational Diabetes History and Glucose Tolerance After Pregnancy Associated With Coronary Artery Calcium in Women During Midlife: The CARDIA Study. Circulation. Published online February 2021:CIRCULATIONAHA.120.047320. doi: 10.1161/CIRCULATIONAHA.120.047320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Catov JM, Sun B, Bertolet M, et al. Changes in Cardiometabolic Risk Factors Before and After Gestational Diabetes: A Prospective Life-Course Analysis in CARDIA Women. Obesity (Silver Spring). 2020;28(8):1397–1404. doi: 10.1002/oby.22848 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Gunderson EP, Quesenberry CP, Jacobs DR, et al. Longitudinal study of prepregnancy cardiometabolic risk factors and subsequent risk of gestational diabetes mellitus: The CARDIA study. American journal of epidemiology. 2010;172(10):1131–1143. doi: 10.1093/aje/kwq267 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Giannakou K, Evangelou E, Yiallouros P, et al. Risk factors for gestational diabetes: An umbrella review of meta-analyses of observational studies. PLoS One. 2019;14(4). doi: 10.1371/journal.pone.0215372 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.OuYang H, Chen B, Abdulrahman AM, Li L, Wu N. Associations between Gestational Diabetes and Anxiety or Depression: A Systematic Review. J Diabetes Res. 2021;2021:9959779. doi: 10.1155/2021/9959779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Suglia SF, Koenen KC, Boynton-Jarrett R, et al. Childhood and Adolescent Adversity and Cardiometabolic Outcomes: A Scientific Statement From the American Heart Association. Circulation. 2018;137(5):e15–e28. doi: 10.1161/CIR.0000000000000536 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Wiss DA, Brewerton TD. Adverse Childhood Experiences and Adult Obesity: A Systematic Review of Plausible Mechanisms and Meta-Analysis of Cross-Sectional Studies. Physiology & Behavior. 2020;223:112964. doi: 10.1016/j.physbeh.2020.112964 [DOI] [PubMed] [Google Scholar]
  • 18.Versteegen M, Bozlak CT, Larkin H, Appleton AA. Maternal depression, adverse childhood experiences, and social support in relation to gestational diabetes risk: results from the Albany Infant and Mother Study (AIMS). BMC Pregnancy Childbirth. 2021;21:335. doi: 10.1186/s12884-021-03814-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Schoenaker DAJM, Callaway LK, Mishra GD. The Role of Childhood Adversity in the Development of Gestational Diabetes. American Journal of Preventive Medicine. 2019;57(3):302–310. doi: 10.1016/j.amepre.2019.04.028 [DOI] [PubMed] [Google Scholar]
  • 20.Stanhope KK, Cammack AL, Perreira KM, et al. Adverse childhood experiences and lifetime adverse maternal outcomes (gestational diabetes and hypertensive disorders of pregnancy) in the Hispanic Community Health Study/Study of Latinos. Annals of Epidemiology. 2020;50:1–6. doi: 10.1016/j.annepidem.2020.08.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Aguayo L, Chirinos DA, Heard-Garris N, et al. Association of Exposure to Abuse, Nurture, and Household Organization in Childhood With 4 Cardiovascular Disease Risks Factors Among Participants in the CARDIA Study. J Am Heart Assoc. 2022;11(9):e023244. doi: 10.1161/JAHA.121.023244 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Gunderson EP, Lewis CE, Tsai AL, et al. A 20-year prospective study of childbearing and incidence of diabetes in young women, controlling for glycemia before conception: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Diabetes. 2007;56(12):2990–2996. doi: 10.2337/db07-1024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gunderson EP, Lewis CE, Lin Y, et al. Lactation duration and progression to diabetes in women across the childbearing years the 30-year CARDIA Study. JAMA Internal Medicine. 2018;178(3):328–337. doi: 10.1001/jamainternmed.2017.7978 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Walsh D, McCartney G, Smith M, Armour G. Relationship between childhood socioeconomic position and adverse childhood experiences (ACEs): a systematic review. J Epidemiol Community Health. Published online September 28, 2019. doi: 10.1136/jech-2019-212738 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.James SA. John Henryism and the health of African-Americans. Cult Med Psych. 1994;18(2):163–182. doi: 10.1007/BF01379448 [DOI] [PubMed] [Google Scholar]
  • 26.Felix AS, Shisler R, Nolan TS, et al. High-Effort Coping and Cardiovascular Disease among Women: A Systematic Review of the John Henryism Hypothesis. J Urban Health. 2019;96(1):12–22. doi: 10.1007/s11524-018-00333-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Jayedi A, Soltani S, Motlagh SZT, et al. Anthropometric and adiposity indicators and risk of type 2 diabetes: systematic review and dose-response meta-analysis of cohort studies. BMJ. 2022;376:e067516. doi: 10.1136/bmj-2021-067516 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Kim K, Yaffe K, Rehkopf DH, et al. Association of Adverse Childhood Experiences With Accelerated Epigenetic Aging in Midlife. JAMA Netw Open. 2023;6(6):e2317987. doi: 10.1001/jamanetworkopen.2023.17987 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Valeri L, VanderWeele TJ. Mediation analysis allowing for exposure–mediator interactions and causal interpretation: Theoretical assumptions and implementation with SAS and SPSS macros. Psychological Methods. 2013;18(2):137–150. doi: 10.1037/a0031034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.VanderWeele TJ. Mediation Analysis: A Practitioner’s Guide. Annual Review of Public Health. 2016;37(1):17–32. doi: 10.1146/annurev-publhealth-032315-021402 [DOI] [PubMed] [Google Scholar]
  • 31.Miller ES, Fleming O, Ekpe EE, Grobman WA, Heard-Garris N. Association between adverse childhood experiences and adverse pregnancy outcomes. Obstetrics & Gynecology. 2021;138(5):770–776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Cammack AL, Hogue CJ, Drews-Botsch CD, Kramer MR, Pearce BD. Associations Between Maternal Exposure to Child Abuse, Preterm Birth, and Very Preterm Birth in Young, Nulliparous Women. Matern Child Health J. 2019;23(6):847–857. doi: 10.1007/sl0995-018-02709-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Freedman AA, Cammack AL, Temple JR, et al. Maternal exposure to childhood maltreatment and risk of stillbirth. Annals of Epidemiology. 2017;27(8):459–465.e2. doi: 10.1016/j.annepidem.2017.07.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Elsenburg LK, van Wijk KJ, Liefbroer AC, Smidt N. Accumulation of adverse childhood events and overweight in children: a systematic review and meta-analysis. Obesity. 2017;25(5):820–832. [DOI] [PubMed] [Google Scholar]
  • 35.Carmichael SL, Abrams B, Ayadi AE, et al. Ways Forward in Preventing Severe Maternal Morbidity and Maternal Health Inequities: Conceptual Frameworks, Definitions, and Data, from a Population Health Perspective. Womens Health Issues. 2021;0(0). doi: 10.1016/j.whi.2021.11.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Gentry SV, Paterson BA. Does screening or routine enquiry for adverse childhood experiences (ACEs) meet criteria for a screening programme? A rapid evidence summary. Journal of Public Health. Published online July 7, 2021:fdab238. doi: 10.1093/pubmed/fdab238 [DOI] [PubMed] [Google Scholar]
  • 37.Racine NM, Madigan SL, Plamondon AR, McDonald SW, Tough SC. Differential Associations of Adverse Childhood Experience on Maternal Health. American Journal of Preventive Medicine. 2018;54(3):368–375. doi: 10.1016/j.amepre.2017.10.028 [DOI] [PubMed] [Google Scholar]
  • 38.Sperlich M, Seng JS, Li Y, Taylor J, Bradbury-Jones C. Integrating Trauma-Informed Care Into Maternity Care Practice: Conceptual and Practical Issues. Journal of Midwifery & Women’s Health. 2017;62(6):661–672. doi: 10.1111/jmwh.12674 [DOI] [PubMed] [Google Scholar]
  • 39.Reuben A, Moffitt TE, Caspi A, et al. Lest we forget: comparing retrospective and prospective assessments of adverse childhood experiences in the prediction of adult health. Journal of Child Psychology and Psychiatry. 2016;57(10):1103–1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of Childhood Abuse and Household Dysfunction to Many of the Leading Causes of Death in Adults. American Journal of Preventive Medicine. 1998;14(4):245–258. doi: 10.1016/S0749-3797(98)00017-8 [DOI] [PubMed] [Google Scholar]
  • 41.Pierce JB, Kershaw KN, Kiefe CI, et al. Association of Childhood Psychosocial Environment With 30-Year Cardiovascular Disease Incidence and Mortality in Middle Age. J Am Heart Assoc. 2020;9(9):e015326. doi: 10.1161/JAHA.119.015326 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Loucks EB, Taylor SE, Polak JF, Wilhelm A, Kalra P, Matthews KA. CHILDHOOD FAMILY PSYCHOSOCIAL ENVIRONMENT AND CAROTID INTIMA MEDIA THICKNESS: THE CARDIA STUDY. Soc Sci Med. 2014;104:15–22. doi: 10.1016/j.socscimed.2013.12.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Owens L, Terrell S, Low LK, et al. Universal precautions: the case for consistently trauma-informed reproductive healthcare. American Journal of Obstetrics & Gynecology. 2022;226(5):671–677. doi: 10.1016/j.ajog.2021.08.012 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1961977-supplement-1.docx (89.9KB, docx)

RESOURCES