Associations of gestation length and offspring birthweight for gestational age with menopausal symptoms and age of natural menopause at midlife among women enrolled in a prebirth longitudinal cohort.

The current study included women enrolled in Project Viva, a longitudinal pre-birth cohort of mother-child dyads in eastern Massachusetts that enrolled participants in early pregnancy from April 1999 to July 2002 21 and has attempted to continue following them since. All Project Viva participants provided written informed consent, and the Project Viva study protocol was approved by the Institutional Review Board of Harvard Pilgrim Health Care. The study population has been described in detail elsewhere, 22 as has the population of women attending the Mid-Life Visit (2017–2021) when we assessed menopause symptoms. 23 Briefly, Project Viva enrolled 2100 women at their initial prenatal visits at a multispecialty group practice in suburban and urban eastern Massachusetts. Eligible women must have been able to communicate in English, bearing a singleton pregnancy, planning to deliver at one of the two study hospitals, and <22 weeks gestation at enrollment. The current study included women who had data available on menopausal symptoms at the Mid-Life Visit. We further limited our sample to women aged ≥45 years or <45 and who had reached menopause by the Mid-Life Visit to mitigate potential bias resulting from the inclusion of women at different stages of the menopausal transition in our analyses, resulting in a final analytic sample size of 691 women. For each woman’s index pregnancy, we calculated gestation length by subtracting the date of the woman’s last menstrual period from the infant’s date of birth, which was extracted from hospital medical records. We subsequently compared this calculated value to the estimated gestation length based on an ultrasound during the second trimester. If the values differed by more than ten days, then we estimated gestation length based on the ultrasound. We obtained data on the infant’s birthweight at the index pregnancy from hospital medical records. 24 We calculated sex-specific birthweight for gestational age z-scores at the index pregnancy using a US national reference. 25 In our analyses, we assessed gestation length in the index pregnancy as a continuous exposure and a categorical exposure divided into three categories (<34 weeks, 34 to <37 weeks, and ≥37 weeks) as well as a binary exposure dichotomized based on whether the infant was born preterm (<37 weeks vs. ≥37 weeks). We selected 34 weeks as a cut point for the variable with three categories because this is the threshold that is typically used to define an early preterm birth. 26 We treated the “≥37 weeks” group as the reference group in all analyses with categorical gestation length as an exposure. We also assessed sex-specific birthweight for gestational age z-score at the index pregnancy continuously and categorically based on whether the infant was born small for gestational age (SGA) (<10 th percentile), appropriate for gestational age (AGA) (10 th to <90 th percentile), or large for gestational age (LGA) (≥90 th percentile). Furthermore, we assessed a lifetime history of delivering a preterm infant as an additional exposure, as we had data available on the gestation length for all of each woman’s prior pregnancies, reported at the Mid-Life Visit. However, we did not have detailed birthweight information to also allow for the calculation of birthweight for gestational age for all lifetime pregnancies. Project Viva women completed the Menopause Rating Scale (MRS) 27 during the Mid-Life Visit at a mean (SD) age of 52.1 (3.8) years. The MRS is a validated tool for assessing symptoms commonly experienced around the time of menopause 27 that includes 11 items, which are scored on a 5-point Likert scale with the following categories: (0) “No symptoms”, (1) “Mild symptoms”, (2) “Moderate symptoms”, (3) “Severe symptoms”, and (4) “Very severe symptoms”. The scores on the individual items are summed to calculate the total score, resulting in a range of 0–44 for the total score, as well as sub-scores to assess somatic (hot flashes, sleep problems, joint/muscular discomfort, and heart discomfort), psychological (depressive symptoms, irritability, anxiety, and physical/mental exhaustion), and urogenital (vaginal dryness, sexual problems, and bladder problems) symptoms. In our analyses, we assessed the MRS total score and all three sub-scores as continuous outcomes, and the responses to the individual questions on the MRS as dichotomous outcomes based on whether the woman reported any or none of the symptoms queried. We also collected data on menopausal status at the Mid-Life Visit, in addition to age of onset of menopause for women who reported having reached menopause as well as whether their periods had stopped naturally or secondary to medical intervention. For women who had not reached menopause at the Mid-Life Visit, we again assessed menopausal status using the same approach on the Year 19 Questionnaire, administered approximately 1 year following the Mid-Life Visit. We also assessed menopausal status using this same approach at the Women’s Health Visit. In our analyses, we used menopausal status and age of onset of menopause from the Mid-Life Visit. If a woman had not reached menopause by the Mid-Life Visit, we used menopausal status and age of onset of menopause from the Year 19 Questionnaire. If a woman had not reached menopause by the Year 19 Questionnaire, we used menopausal status and age of onset of menopause from the Women’s Health Visit. We collected self-reported data at enrollment on date of birth, education, parity, annual household income (which we subsequently dichotomized as ≤$70,000 vs. >$70,000), pre-pregnancy weight, and height. We calculated pre-pregnancy BMI using the self-reported values of pre-pregnancy weight and height. We also collected self-reported data on race and ethnicity on the Year 19 Questionnaire, and we substituted missing values with race and ethnicity reported at enrollment. We tabulated summary statistics for all demographic characteristics, birth characteristics, and MRS scores for the entire analytic cohort of 691 women and by index pregnancy birth size. We evaluated the associations of index pregnancy birth characteristics with menopausal symptoms in midlife, assessed via the MRS total score and sub-scores, using linear regression. We assessed index pregnancy birth characteristics in relation to specific menopausal symptoms based on the individual MRS questions using binomial regression models with a log link function to estimate the relative risk (RR) (95% CI) of experiencing any symptoms vs. not experiencing any symptoms. Additionally, we assessed age at onset of natural menopause as a survival outcome using Cox proportional hazards models. We fitted unadjusted models and models adjusted for age at enrollment, education, parity, annual household income, and pre-pregnancy BMI. The Cox proportional hazards models for age of onset of natural menopause were not adjusted for age at enrollment, given this variable’s strong correlation with whether a woman reached natural menopause by the time of our Mid-Life Visit. In addition to the main analysis, we ran some sensitivity analyses. First, given potential differences in associations of interest by race and ethnicity, we performed sensitivity analyses in which we reran our models among only non-Hispanic White participants as the sample size was too small for stratum-specific models within other racial and ethnic groups. We also ran formal tests for statistical interaction to assess effect modification by race and ethnicity. Additionally, we ran sensitivity analyses adjusted for maternal gestational diabetes mellitus (GDM) status at the index pregnancy, in addition to the aforementioned covariates. All analyses were complete case analyses. For all models, we confirmed assumptions of linear regression via assessment of residuals for normality. We performed our analyses in SAS Enterprise Guide (version 8.3).

Mean (SD) age at enrollment of the 691 Project Viva women in this analysis was 33.7 (3.8) years ( Table 1 ). The majority of women identified as non-Hispanic White (N=506; 73%), had a college degree (N=550; 80%), and had an annual household income >$70,000/year (N=443; 68%). Project Viva women who were excluded from the present analyses were younger at enrollment, were more likely to be non-Hispanic Black, and were less likely to be a college graduate ( Supplementary Table 1 ). Mean (SD) gestation length was 39.5 (1.8) weeks and birthweight for gestational age z-score was 0.26 (0.96); 33 (5%) resulted in an SGA infant, 550 (80%) an AGA infant, and 108 (16%) an LGA infant. At the Mid-Life Visit, mean (SD) total MRS score was 7.9 (5.8) points; the psychological sub-score was 2.9 (2.7) points, the somatic sub-score 3.4 (2.5) points, and the urogenital sub-score 1.7 (2.0) points. Among 392 (57%) of women who had reached natural menopause, mean (SD) age of onset of natural menopause was 50.8 (3.7) years. In response to the individual questions on the MRS, 318 (46%) women reported having experienced any depressive mood symptoms, 351 (51%) reported having experienced any irritability symptoms, and 355 (52%) reported having experienced any anxiety symptoms ( Supplementary Table 2 ). Additionally, 418 (61%) of women reported having experienced hot flashes or sweating and 475 (69%) reported having sleep problems. We did not observe any associations of any of the studied index pregnancy birth characteristics with MRS total score or the three sub-scores in either unadjusted ( Supplementary Table 3 ) or adjusted ( Table 2 ) models. For example, gestation length (β=−0.07 points per week, 95% CI −0.32 to0.18) and birthweight for gestational age z-score (β=0.16 points per z-score, 95% CI −0.31 to0.63) were not associated with the MRS total score in adjusted models. We also assessed associations of index pregnancy gestation length and birthweight for gestational age with specific menopausal symptoms in midlife in unadjusted ( Supplementary Table 4 ) and adjusted ( Table 3 ) models. We observed an inverse association of gestation length with experiencing heart discomfort symptoms in women at midlife (adjusted RR [aRR]=0.92 per week, 95% CI 0.88–0.98). When assessing gestation length discretely, we found that in comparison to women with gestation length ≥37 weeks, those with a gestation length <34 weeks had a greater likelihood of experiencing heart discomfort symptoms in midlife (aRR=2.59, 95% CI 1.66–4.03). Similarly, women with a lifetime history of preterm birth were 48% more likely to experience heart discomfort symptoms (aRR=1.48, 95% CI 1.10–1.99). A one unit increase in offspring birthweight for gestational age z-score was associated with a 10% greater risk of experiencing hot flashes or sweating symptoms (aRR=1.10 per z-score, 95% CI 1.03–1.18) and a 7% greater risk of experiencing physical and mental exhaustion (aRR=1.07 per z-score, 95% CI 1.00–1.15). Women who delivered an SGA infant were 41% less likely to experience hot flashes or sweating symptoms compared to women who delivered an AGA infant (aRR=0.59, 95% CI 0.37–0.96), and the likelihood of experiencing sleep problem symptoms was greater in women who delivered an LGA infant compared to those who delivered an AGA infant (aRR=1.17, 95% CI 1.03–1.33). We did not observe any associations of gestation length or birthweight for gestational age in relation to depressive mood symptoms, irritability symptoms, anxiety symptoms, joint or muscular discomfort symptoms, sexual problems, bladder problems, or vaginal dryness in our adjusted models. We did not observe any associations of gestation length (adjusted HR [aHR]=0.99 per week, 95% CI 0.94–1.06) or birthweight for gestational age z-score (aHR=1.06 per z-score, 95% CI 0.94–1.19) with age of onset of natural menopause ( Table 4 ). Similarly, we did not observe any associations for the categorical preterm birth with this outcome. We assessed effect modification by race and ethnicity via formal tests for interaction in our unadjusted and adjusted models for the MRS total score and the sub-scores. We did not observe any evidence of effect modification in these analyses, as all of the interaction terms were non-significant. These included the interaction terms for gestation length (p=0.16) and birthweight for gestational age z-score (p=0.83) in relation to the MRS total score. We observed similar results in our analyses assessing associations of the birth characteristic histories with the MRS total score and its sub-scores among only non-Hispanic White participants, as we observed mostly null associations in these analyses ( Supplementary Table 5 ). We also did not observe any associations with age of onset of natural menopause ( Supplementary Table 6 ). The association between gestation length and heart discomfort symptoms among non-Hispanic White participants was consistent with the association from the overall sample when assessing gestation length as a continuous exposure (aRR=0.88 per week, 95% CI 0.82–0.94) and when comparing women with a gestation length <34 weeks to those with a gestation length ≥37 weeks (aRR=3.00, 95% CI 1.65–5.46) ( Supplementary Table 7 ). Heart discomfort symptoms were also more likely among women who delivered a preterm infant at the index pregnancy (aRR=1.77, 95% CI 1.14–2.76) and among women with a lifetime history of delivering a preterm infant (aRR=1.71, 95% CI 1.22–2.40) when we restricted our analyses to non-Hispanic White participants. We also observed associations between gestation length and experiencing depressive mood symptoms (aRR=0.95 per week, 95% CI 0.90–1.00) and irritability symptoms (aRR=0.95 per week, 95% CI 0.91–1.00) among non-Hispanic White participants. Delivering a preterm infant at the index pregnancy was associated with a 44% greater risk of depressive mood symptoms among non-Hispanic White participants (aRR=1.44, 95% CI 1.04–1.99). In addition, physical and mental exhaustion was more prevalent among non-Hispanic White women with a gestation length <34 weeks compared to those with a gestation length ≥37 weeks (aRR=1.51, 95% CI 1.06–2.15). Among non-Hispanic White women, birthweight for gestational age z-score was associated with a greater risk of experiencing hot flashes or sweating symptoms (aRR=1.08 per z-score, 95% CI 1.00–1.17) and a greater risk of joint or muscular discomfort symptoms (aRR=1.09 per z-score, 95% CI 1.00–1.20). Sleep problems were more common among non-Hispanic White women delivering an LGA infant at the index pregnancy compared to those delivering an AGA infant (aRR=1.16, 95% CI 1.01–1.33). In addition, non-Hispanic White women with longer gestation lengths were more likely to experience sexual problems (aRR=1.07 per week, 95% CI 1.00–1.15), and those with a lifetime history of delivering a preterm infant were 39% less likely to experience sexual problems (aRR=0.61, 95% CI 0.41–0.92). As in our main analyses, we observed null associations in our analyses for the MRS total score and its sub-scores after additionally adjusting for maternal GDM status at the index pregnancy ( Supplementary Table 8 ). When we assessed the specific menopausal symptoms in these analyses ( Supplementary Table 9 ), we observed similar results as in the main analyses for heart discomfort symptoms and hot flashes or sweating symptoms. We also observed associations between gestation length and sexual problem symptoms (aRR=1.08 per week, 95% CI 1.01–1.15) and bladder problem symptoms (aRR=1.07 per week, 95% CI 1.00–1.15) in these analyses. In addition, women with a lifetime history of delivering a preterm infant were at a lower risk of experiencing sexual problems (aRR=0.69, 95% CI 0.49–0.95).

In this prospective cohort study of women recruited in early pregnancy and followed for almost two decades, we assessed associations of gestation length and birthweight for gestational age z-scores at the index pregnancy with menopausal symptoms and age at natural menopause in midlife. Gestation length and birthweight for gestational age z-scores were not associated with the overall MRS scores or its sub-scores. The current study is among the first to evaluate a history of birth outcomes from a prior pregnancy in relation to menopausal symptoms, including heart discomfort symptoms. Previous studies have evaluated vasomotor symptoms, but we are not aware of any prior work that has assessed psychological or urogenital menopausal symptoms or age at natural menopause as outcomes, for which we observed mostly null results. One prior study observed associations between having a history of delivering a preterm or SGA infant and more severe vasomotor symptoms during menopause, 15 whereas we observed a reduced risk of experiencing hot flashes among women with a history of delivering an SGA infant compared to those who delivered an AGA infant. Due to the scarce literature on this topic, future studies are needed to corroborate our findings. There are potential mechanisms that may explain our finding that prior birth characteristic histories are associated with heart discomfort symptoms and vasomotor symptoms. Oxidative stress has been linked to the natural decline in estrogen that occurs during the menopausal transition, 28 which may be an underlying mechanism. Inflammation during pregnancy plays a strong role in the induction of spontaneous preterm deliveries 29 , 30 and has been found to be associated with an increased risk of maternal cardiovascular disease in rats. 31 Additionally, the presentation of hot flashes in menopausal women is associated with endothelial dysfunction. 32 Rapid arterial stiffening, which occurs during the menopausal transition 33 and is more common in women with a history of iatrogenic preterm birth, 34 may also contribute to the observed associations. Levels of stress hormones also may play a role, as corticotropin-releasing hormone promotes the production of prostaglandin and thereby induces delivery of the infant. 35 Higher maternal cortisol levels are associated with an increased risk of prematurity and low birthweight, 36 primarily via inadequate activation of placental 11-beta-hydroxysteroid dehydrogenase type 2. 37 This is notable because stress hormones may induce the onset of cardiovascular complications. 38 – 40 Furthermore, our finding that women with a history of delivering an LGA infant are at a greater risk of experiencing sleep problems is notable, given the important role of cortisol in regulating an individual’s circadian rhythm. 41 Shared common causes of unfavorable birth outcomes and menopausal experiences, such as maternal obesity 42 – 44 and GDM, 42 , 45 may also explain our findings. The current study has both strengths and limitations. Its main strength is the prospective design with almost two decades of follow up, which reduces the concern of reverse causation or recall bias. Having data available on age of onset of natural menopause also allowed us to assess time to onset of natural menopause as a survival outcome. In addition, the rich covariate database in Project Viva allowed us to account for potential confounding by several important factors including age at enrollment, parity, annual household income, and pre-pregnancy BMI. However, one limitation is the fact that our study participants were likely at different stages of menopause during the study, and we did not have data available on each woman’s specific menopausal transition stage. This is a limitation because menopausal symptoms have been shown to vary across stages of the menopausal transition. 46 However, our decision to restrict our analyses to women aged ≥45 years or already menopausal likely mitigated this limitation. Additionally, we were unable to conduct analyses with the inclusion of only women identifying as other than non-Hispanic White due to the small sample size. Therefore, future studies should be conducted in more racially and ethnically diverse populations. Furthermore, we conducted many statistical tests, potentially resulting in false positives. Therefore, results in this paper may be due to chance given the amount of statistical analyses performed. However, the current study was hypothesis driven, so we opted not to apply any multiple testing corrections in our analyses. Lastly, we were unable to conduct sensitivity analyses including only women for whom the index pregnancy was their only pregnancy and sensitivity analyses including only women for whom the index pregnancy was their last pregnancy due to sample size limitations. This sample size restriction prevented us from being able to assess the extent to which our null results in the main analyses may have been attributed to exposure misclassification bias. Correcting for such bias should be a key component of future studies examining prior birth characteristic histories in relation to women’s health outcomes in midlife.

In conclusion, we found no associations between gestation length of a prior pregnancy and offspring birthweight for gestational age z-scores with total menopausal symptoms or age at natural menopause during midlife. Additional studies examining these research questions in other cohorts should be conducted given the scare literature on the topic.

Menopause is a female reproductive milestone characterized by the cessation of menstruation. 1 Onset of menopause in women is often accompanied by numerous bothersome health symptoms, including vasomotor complications, 2 – 6 sleep problems, 4 , 5 and psychiatric comorbidities. 4 , 6 These symptoms can be attributed at least in part to hormonal changes, such as a reduction in estrogen and an increase in follicle stimulating hormone (FSH) levels. 1 For many women, the menopausal transition is both physically and emotionally challenging. 7 , 8 Most importantly, the unpleasant symptoms that many women experience during menopause may be associated with development of adverse health conditions later in life, such as cardiovascular disease 9 , 10 and impaired cognitive functioning. 11 Examining associations between women’s reproductive histories and menopausal symptoms would improve our understanding of which women may have more unfavorable experiences than others during the menopausal transition and pave the way for future work to investigate why. Emerging evidence links a history of reproductive complications, such as endometriosis and infertility, to higher risk for menopausal symptoms and an earlier age of onset of natural menopause. 12 – 14 Although prior studies have examined associations of women’s reproductive histories with midlife outcomes including vasomotor symptoms 15 and perimenopausal bleeding, 16 we have not identified any studies that have examined psychological and urogenital menopausal symptoms. Assessing these outcomes is important given their associations with low sexual functioning 17 and unfavorable longer-term health outcomes including cardiovascular disease. 18 – 20 Additionally, few studies have examined lifetime birth characteristic histories as predictors of menopausal symptoms, which is important to consider the effects of multiple births. The objective of the current study was to assess the extent to which gestation length and birthweight for gestational age, two indicators of pregnancy health and quality of the intrauterine environment, are associated with menopausal symptoms in midlife and age of onset of natural menopause. Secondarily, we also sought to assess the extent to which having a lifetime history of delivering a preterm infant is associated with menopausal symptoms in midlife and age of onset of natural menopause. We hypothesized that having a history of delivering an infant with a below or above appropriate birthweight for gestational age would be associated with a greater risk of experiencing adverse menopausal symptoms and an earlier age of onset of natural menopause. We also hypothesized that having a history of delivering a preterm infant at the index pregnancy or at any point during an individual’s lifetime would be associated with a greater risk of experiencing adverse menopausal symptoms and an earlier age of onset of natural menopause.

Supplementary Table 1. Demographic Characteristics of Included vs. Excluded Project Viva Women in the Current Analyses Supplementary Table 2. Distribution of Specific Menopausal Symptoms Assessed via the Menopause Rating Scale Supplementary Table 3. Unadjusted Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Menopausal Symptoms in Midlife Supplementary Table 4. Unadjusted Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Specific Menopausal Symptoms in Midlife Supplementary Table 5. Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Menopausal Symptoms in Midlife Including Only non-Hispanic White Participants a Supplementary Table 6. Associations of Gestational Age and Infant Birthweight for Gestational Age with Time to Onset of Natural Menopause Including Only non-Hispanic White Participants Supplementary Table 7. Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Specific Menopausal Symptoms in Midlife Including Only non-Hispanic White Participants Supplementary Table 8. Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Menopausal Symptoms in Midlife Additionally Adjusted for Maternal GDM Status at the Index Pregnancy Supplementary Table 9. Adjusted Associations of Gestation Length and Infant Birthweight for Gestational Age at Index Pregnancy with Specific Menopausal Symptoms in Midlife Additionally Adjusted for Maternal GDM Status at the Index Pregnancy