Intergenerational Occurrence of Premature Birth and Reproductive Health in Prematurely-Born Women in the Women’s Health Initiative

Objective To compare reproductive history and postmenopausal health by birth status (preterm vs. full term) in a U.S. longitudinal study of postmenopausal women. Birth status was examined according to region of residence, household, and neighborhood socioeconomic status (SES). Methods In the Women’s Health Initiative Observational Study, 2271 women were born prematurely (< 37 weeks). ANOVA and Chi-square determined birth status differences of reproductive history, pregnancy, and postmenopausal health. Odds ratios were calculated using either binary logistic or multinomial logistic regression. SES and U.S. region of residence were examined as potential effect modifiers. Results Preterm-born women compared to term-born women had higher risk of delivering a premature infant (aOR 1.68, 95% CI [1.46, 1.93]), higher odds of later-age first pregnancy (aOR 1.27 95% CI [1.02, 1.58]), longer duration to become pregnant (> 1 year to pregnancy) (aOR 1.10 95% CI [1.01, 1.21]), more miscarriages (aOR 1.23 95% CI [1.11, 1.37]), and more pregnancy complications including hypertension (aOR 1.58 95% CI (1.13, 2.21)], preeclampsia (aOR 1.64 95% CI [1.24, 2.16]), and gestational diabetes (aOR 1.68 95% CI [1.11, 2.53]). Preterm-born women had higher odds of menopause before age 50 (aOR 1.09 95% CI [1.05, 1.14]). Post-menopause, they had higher rates of diabetes (p = .01), hypertension (p = .01), hysterectomy (p = .045), and higher Charlson Comorbidity Index scores (p = .01). Conclusions Preterm-born women had higher reproductive and pregnancy risks which when coupled with early menopause, may indicate a shorter childbearing period than term-born women. Guidelines for integration of preterm history in women’s health care across the life course are needed to identify and manage their higher risk.


Introduction
The stress endured by an infant born before 37 weeks gestation has been recognized among other early life stress factors to have an influence on health across the life course (Halfon et al., 2018).The U.S. rate of preterm birth (PTB; < 37 weeks gestation) increased in 2021 to 10.49% with rates varying by race (Black, 14.75%, Hispanic 10.23%, White, 9.50%, non-Hispanic Asian 9.23%) and more than 95% survive into adulthood (Osterman et al., 2023;Raju et al., 2017).Systematic research around the globe has reported an array of poorer adult outcomes for preterm-born individuals including cardiovascular, metabolic, renal, and respiratory disease, mental health conditions (Crump, 2020), poorer neurodevelopment (Allotey et al., 2018), and difficulty reaching adult milestones (Saigal et al., 2016).In a multinational study of 6.2 million individuals, preterm birth was associated with higher all-cause mortality with higher odds in lower gestations.This risk was higher in premature-born women compared to premature-born men (Risnes et al., 2021).
One area that has received little attention in pretermadult investigations is women's reproductive health including pregnancies and pregnancy complications, infant health, and the incidence of reproductive comorbidities.This is an important issue given the economic and social cost of prematurity, the higher risk of pregnancy complications from hypertension and preeclampsia and long-term consequences, and higher risk of offspring morbidity and mortality.
There is evidence of intergenerational occurrence of preterm birth (Swamy et al., 2008;Boivin et al., 2015;Ncube et al., 2017).From Norway's Medical Birth Registry of 1,167,506 singleton births between 19671,167,506 singleton births between and 19881,167,506 singleton births between , Swamy et al.,(2008) ) found a reduction in the rate of reproduction for those born preterm with an increased risk of preterm birth directly corresponding with lower gestational age categories.In Montreal, Canada, over 7000 prematurelyborn women had higher risk for preterm offspring if they were born before 32 weeks gestation [odds ratio (OR) 1.63 for < 32 weeks; OR 1.41 if 32-36 weeks] independent of gestational hypertension and diabetes (Boivin et al., 2015).In a U.S. investigation of 6592 maternal PTB-infant PTB associations, PTB women had increased odds for having a preterm infant (aOR = 1.46, 95% CI 1.0-1.9)with higher risks for earlier gestations versus later gestation (Ncube et al., 2017).Recent U.S. studies have focused on residential environments revealing social disparities and their contribution to premature birth and generational impacts, though the research is inconsistent.Ncube et al. (2017) reported more preterm birth with non-Hispanic Blacks of both the birth mother and offspring generations, or mothers who moved to neighborhoods with a higher percentage of non-Hispanic Black residents.(Castrillio et al., 2014) using Illinois birth records, found no intergenerational trend for prematurity in non-Hispanic Whites and African American women born small for gestational age.In contrast, Smid et al. (2017) examined preterm birth intergenerational occurrence in 71,676 non-Hispanic White and non-Hispanic Black women in Virginia.After adjustment for confounders, only non-Hispanic Black mothers had increased odds of an early preterm infant.No reports were found on the postmenopausal health of women born prematurely including timing of menopause, problems with breast and/or gynecological health.
Given these gaps in the literature, the first aim of this study was to compare women's self-reported reproductive history including age of menarche, pregnancy and birth histories including prematurely born offspring, pregnancy complications, and postmenopausal health by birth status (preterm vs. full term).Secondly, we examined the association of birth status with region of birth, region of residence, household and neighborhood socioeconomic status (SES) (Griffin et al., 2013).The sample is from the Observational Study (OS) of the Women's Health Initiative (WHI), a U.S. longitudinal cohort study of postmenopausal women of race, ethnic and social diversity (Anderson et al 2003;Langer et al 2003).The WHI-OS offers a unique opportunity to answer questions of life-long reproductive health from childbearing to post-menopause of prematurely-born compared to termborn women across the U.S.

Methods
The WHI longitudinal Observational Study (WHI-OS, n = 93,676) was designed to examine a natural history of risk factors for chronic diseases and death in postmenopausal women (Anderson et al., 2003;Langer et al., 2003).Between 1993 and 1998, women from 40 clinical centers across the United States were enrolled.The WHI-OS was the natural control for three clinical trial arms.

Sample
Postmenopausal women ages 50 to 79 found to be ineligible or unwilling to participate in the clinical trials were invited to be one of 100,000 women enrolled in the OS.Enrollment of a diverse sample of women of racial/ethnic groups was affected.Women were excluded if they had medical conditions of less than 3-year survival, complicating conditions of alcohol, drug dependency or dementia, or planned to move from the area within 3 years.

Procedure
Standardized, self-report questionnaires included demographic and risk exposure data including reproductive history, family, and medical history.Certified staff obtained physical measurements, including blood pressure, height and weight, and blood samples at the clinic visit.Uniform procedures at all 40 study sites were maintained through a standardized written protocol, centralized training of local clinic staff, local quality assurance activities, and periodic quality assurance visits by the Clinical Coordinating Center (Anderson et al., 2003;Langer et al., 2003).Questionnaires were mailed annually with in-person assessment every three years.All participants provided informed consent using ethics approvals by each center's Institutional Review Board.

Measures
The exposure of interest, preterm birth, was determined at study enrollment when participant's gestational age at birth was queried as full term (pregnancy lasted about 9 months) or preterm (born 4 or more weeks premature).The primary outcomes of interest included age at menarche, pregnancy complications (preeclampsia, gestational diabetes, hypertension during pregnancy, and infant weighing < 5.5 pounds), birth history (age at first birth, age at last birth, number of live births, number of times ever pregnant, duration to become pregnant, miscarriages, stillbirths, child born more than 3 weeks early, and breastfeeding for one month or more) and postmenopausal gynecological health (breast disease, reproductive surgery, depression, depression treatment, menopause before age of 50, hormone therapy).Variables considered as potential model covariates included birth cohort (estimated from 10-year age groups), race, ethnicity, education, income, Body Mass Index (BMI), smoking, Charlson Comorbidity Index (Austin et al., 2015;Charlson et al., 1994), and health insurance.Additionally, region of residence, region of birth, household SES, and neighborhood SES (Griffin et al., 2013) were examined as potential effect modifiers.Adjudication of study outcomes was completed at the local center by an MD who coded test reports, labs, and self-report documents according to WHI definitions (Curb et al., 2003).

Analysis
Of 93,676 participants in the WHI-OS, gestational age and birthweights were cross checked leaving 88,311 with reliable gestational age and birthweights.Of these 2.6% (n = 2,271) reported being born 4 or more weeks premature (Fig. 1).For the first aim, analyses were performed by birth status (preterm vs. full term).Descriptive statistics by birth status summarized socio-demographic characteristics, reproductive history, pregnancy outcome and postmenopausal gynecological health.Statistical differences by birth status were determined by ANOVA for continuous variables and Chisquare analysis for categorical variables.Odds ratios, both crude and variable-adjusted, were calculated using either binary logistic or multinomial logistic regression depending upon the outcome variable examined.Most outcome variables had two levels and hence binary logistic regression was utilized.However, multinomial logistic regression was used for outcomes of age at menarche, number of times pregnant, number of times pregnant for 6 months or more, age at first birth, age at last birth, length of breastfeeding, and hormone therapy.Covariates of birth cohort (estimated from 10-year age groups), race/ethnicity, education, income, BMI, cigarette smoking, comorbidity index (Charlson Comorbidity Index (Austin et al., 2015;Charlson et al., 1994), and any health insurance were included in full adjusted models.
For the second aim, the potential effect modification of region of residence and SES on birth status, residential region (Northeast, Midwest, West, South), household SES, and neighborhood SES variables were included in the adjusted models as an interaction term with birth status.Household SES was dichotomized into low (education less than high school or household income less than $20K/ year) and not low (high school education or more and household income > $20K/year).Neighborhood SES, an index summary measure developed for the WHI by Griffin et al. (Griffin et al., 2013), was divided into two levels (above, below) based on median neighborhood SES.U.S. region of residence was based on U.S. Census definition.Statistical differences in region of residence and SES variables were determined by the interaction p-value from the multiple variable models.Reproductive health variables by birth status were stratified by SES and region variables.Binary or multinomial logistic regression was used to calculate the odds ratios for birth status within the levels of SES or region variables.Missing data were less than 5%, with most variables having < 1% missingness.Income had the highest amount of missing data (4.3%),therefore the level "missing" was added.Analyses were conducted using SAS v9.4 (Cary, NC).A significance level of p ≤ 0.05 was used for all analyses unless otherwise noted.

Effect Modification Models
There were significant effect modifications for SES and preterm birth.Women born preterm with low SES had lower odds of ever being pregnant (p = 0.0026) compared to their

Discussion
In this U.S. longitudinal study comparing preterm-born and term-born women's reproductive health from pre-conception to menopause we found worse pregnancy outcomes, higher likelihood of birthing a preterm infant, earlier menopause and a trend for higher odds of hysterectomy.Preterm-born women had higher risk for a later first pregnancy, a longer time to pregnancy, more miscarriages, and more pregnancy complications of hypertension, preeclampsia (a multisystem disorder including hypertension), and gestational diabetes.These pregnancy risks coupled with early menopause, indicate that preterm-born women have a shorter childbearing period than their term-born peers.For bilateral oophorectomy, breast disease, or menopausal hormone therapy, the preterm and term-born groups were comparable.This study extends prior findings to later ages and shows the significant risks for preterm-born women's reproductive and gynecological health from menarche to menopause.The higher rates of pregnancy complications for pretermborn women likely contribute to the risk for delivering a preterm infant.Our findings of higher rates of preeclampsia and gestational diabetes have been reported in younger cohorts.A 1958 British cohort investigation of intergenerational effects on preterm delivery (≤ 36 weeks) identified hypertension history of mother, father and maternal grandmother each independently increased risk of preterm birth (OR 1.7, 2.0, and 1.5 respectively) (Hennessy and Alberman, 1998).This risk was very high for hypertensive preterm-born mothers compared to those born after 36 weeks (21% vs 9%).These pregnancy complications bring higher risk for later morbidity and mortality.In a WHI study of 48,113 women, four pregnancy outcomes (gestational diabetes, hypertensive disorders of pregnancy, low birth weight, preterm delivery) were independently associated with later heart disease (Sondergaard et al., 2020).In our older cohort of postmenopausal women, the health of those preterm-born align with these reports as shown by higher odds for gestational diabetes and hypertension, and higher Charlson Index scores.The utility of the Charlson Index in longitudinal studies is its demonstrated validity to estimate relative risk of death from clinical comorbidities (Austin et al., 2015;Charlson et al., 1994).In a diverse U.S. Collaborative Perinatal Project study of 46,551 women (born 1959-1966), pregnancy complications of preterm delivery, hypertensive disorders of pregnancy, and gestational diabetes were associated with higher mortality 50 years later with higher rates among Black than White participants (Hinkle et al., 2023).Gestational hypertension is associated with a greater risk of later overall cardiovascular disease, coronary heart disease, and heart failure (Lo et al., 2020).Thus, our findings add to the evidence that pregnancy complications experienced by preterm-born women have immediate risk for the health of mother and infant, but also confers long-term risk for chronic disease and early death.
Our findings of higher risk (68%) for preterm-born women to deliver a preterm offspring are comparable to two later cohorts, suggesting some continuity across decades.A 41-63% higher risk for preterm birth < 32 to 36 weeks gestation was reported in a Canadian cohort (Boivin et al., 2015).In Pennsylvania, 46% increase was reported inclusive of birth weight subgroups (Ncube et al., 2017).In Smid et al.'s Virginia cohort (2017), higher intergenerational risk for preterm birth was found for non-Hispanic Black mothers but not non-Hispanic White mothers (i.e., non-Hispanic White, (aOR 1.28, 95% CI [0.71, 2.31]), non-Hispanic Black (aOR 3.26,95% CI [1.77,6.02]).In contrast, Dorner et al. (2017) found an association for intergenerational preterm birth of < 30 weeks gestation, but no association for preterm births of 34-36 weeks for non-Latino Whites and African-Americans.We did not find higher risk among racialized groups in contrast to findings by Castillio and Smid, though the WHI cohort may have lower power for racial comparisons.
An explanation for higher reproductive risk unexplored in our study is lower social functioning in preterm-born adults.In an international individual participant data (IPD) meta-analysis of 5 preterm-born adult cohorts, lower friend relationships were found, but ratings on partner and family relationships were comparable (Ni et al., 2021).In a small Ontario cohort (n = 100 preterm, 89 = term) at age 40, the preterm-born women (birth weight 501-1000 g; birth years [1977][1978][1979][1980][1981][1982] were less likely to have been pregnant (38.3% vs 61.5%) or have children (OR,0.44;95% CI [0.20,0.96]).No differences were found for age of menarche, fertility, miscarriage rate or preterm birth, though their investigation may have been limited by sample size.Behaviorally, both women and men had less dating, marriage or cohabitation which may signal a possible explanation for these pregnancy findings (Saigal et al., 2016).
Socioeconomic factors have been shown to play a role in pregnancy.We found SES to have an effect on pregnancy and pregnancy complications but not postmenopausal health.In an older cohort of individuals born in 1958, the influence of maternal grandmother and mother's premature status, along with social class were strong predictors of infant birth weight but not gestational age (Emanuel et al., 1992).One population-based study of the heritability of spontaneous  (Wu et al., 2015).Household environments may have common patterns across generations that affect women and gestation.The background in pregnancy and neonatal medical care practices in the U.S. in the mid-1900s is notable for the WHI sample.Women were between 50-79 years of age, with birth dates in the 1920s, 1930s, and 1940s when they enrolled between 1993 and 1998.Between the 1920s and 1940s, hospital births in large cities grew from ~ 50 to 90%.High infant mortality led to the establishment of newborn nurseries but the care of premature infants received little attention before 1950 (Lussky et al., 2005).By the 1950s, a few pediatricians were beginning to create the specialty of neonatology.Incubators for premature babies, originally developed in Europe and Russia, appeared in U.S. exhibits as early as 1898 and showed improved survival, but the carnival atmosphere surrounding exhibitions and misplaced concerns about infections delayed their acceptance by medical professionals and the public (Barr, 1995;Lussky et al., 2005).Thus, if a WHI-OS preterm-born woman was born in a hospital, she likely did not receive specialty care and may have faced a variety of pulmonary complications, thermal irregularities, and nutritional challenges.It is also likely that as premature infants, the WHI participants survived because they had higher birth weights and fewer illnesses than the preterm infants who survive today.While more infants < 32 weeks gestation survive today, for decades the largest group of preterm infant survivors are those born between 34-36 weeks representing more than 70% of preterm infants (Osterman et al., 2023).Thus, our findings are relevant and associated with today's preterm-born adults, though examination in more racially diverse samples are needed.Future research priorities should include long-term follow-up of SES and racially diverse cohorts.
Our findings align with reports of higher risk indicators for chronic disease as shown by higher odds for gestational diabetes and hypertension, and higher Charlson Index scores.Recently, significant associations were found between preterm birth and prevalent hypertension, earlier onset hypertension, and hypertensive medication use in the WHI-OS participants of the present study (Brewer et al., 2023).Recent calls urge the integration of preterm history in adult care; however there are currently no guidelines addressing this (D'Agata et al., 2022;Luu et al., 2016).Birth history should be included as part of medical records beyond infancy and early childhood.Healthcare providers and the preterm-born individuals themselves need to be aware of their history and long-term risks to begin pro-active intervention.

Limitations
Limitations of this study include recall bias and self-report.In the WHI-OS, data quality assurance methods were incorporated at several levels and across study sites (Anderson et al., 2003).As an additional step in this study, we crossreferenced self-reported prematurity status with birth weight to define the preterm group.While birth records are generally preferred, the self-report birth data is valid.Birth weight recall by middle-aged and older women was accurately reported by 74% and birth weight category by 87% in a national study (Wodskou et al., 2010).The study outcomes were self-reported and not all were adjudicated which may result in some misclassification.The strengths of this study include its U.S. regional representation, large sample size, extensive baseline data collection, adjudication procedures, and the ability to account for potential confounding variables in analyses.The preterm group size for the not conceiving sub-analysis was small and cautiously interpreted.The details for early birth and newborn care, including family history, social and genetic predispositions are unknown.

Conclusions for Practice
The Women's Health Initiative (WHI) is a landmark study which has changed women's health and medical practice around the world.The WHI helps women and their healthcare providers make more personalized and informed decisions about women's health.Reproductive and gynecological health are central to women across their lifespan.Given the growing evidence for later adult health risks for prematurely born people, birth history should be included in medical records beyond infancy and early childhood.Healthcare providers and the preterm-born themselves need to be aware of the long-term risks to begin pro-active intervention (D'Agata et al., 2022;Kelly et al., 2021;Luu et al., 2016).

Table 1
Characteristics of WHI-OS participants by birth status + systolic blood pressure ≥ 140 mm Hg and/or diastolic blood pressure ≥ 90 mm Hg or being on medication for high blood pressure

Table 2
Reproductive history & gynecological health by birth status

Table 2
for secondary birth outcomes adjusted for birth cohort and race/ethnicity OR's for pregnancy, births, and pregnancy complication outcomes adjusted for birth cohort, race/ethnicity, and education OR's for gynecological health outcomes adjusted for birth cohort, race/ethnicity, education, income, BMI, smoking, comorbidity index, and any health insurance Fig.2a Pregnancy historyodds ratios (95% CI) for preterm vs. full term.b Pregnancy complications-odds ratios (95% CI) for preterm vs. full term.c Gynecological health-odds ratios (95% CI) for preterm vs.

Table 3
Reproductive history & gynecological health by SES