Longitudinal associations of psychosocial stressors with menopausal symptoms and well-being among women in midlife

Abstract

Objective:

We examined longitudinal associations of psychosocial stressors with menopausal symptoms and well-being of women in midlife in a longitudinal cohort.

Methods:

This study is based on 682 women from Project Viva, a prospective cohort enrolled in 1999-2002 during pregnancy (median age=33.3 years) and followed for almost 2 decades. In pregnancy, women self-reported psychosocial stressors (history of physical and sexual abuse and financial instability, from childhood to the current pregnancy). In 2017-2021 (median age=51.6 years), they reported their menopausal symptoms (0-44 point scale) and well-being (general health [good/fair/poor vs. excellent/v.good], generalized anxiety symptoms, and depressive symptoms [both – more than minimal levels vs. none/minimal]). We performed multivariable and logistic regression models to examine associations of psychosocial stressors with outcomes, adjusting for covariates.

Results:

History of physical abuse (reported by 37.3%) was associated with worse menopausal symptoms in the somato-vegetative (0.46 points; 95%CI 0.04 to 0.87) and psychological (0.52 points; 0.07 to 0.97) domains and with worse general health (OR=1.73; 1.17 to 2.55) and greater depressive symptoms (OR=1.74; 1.05 to 2.87). History of sexual abuse (7.7%) was associated with worse menopausal symptoms (2.81 points; 1.05 to 4.56) and worse general health (OR=2.04; 1.04 to 4.03) but not with depressive symptoms. History of financial instability (10.8%) was associated with worse menopausal symptoms (1.92 points; 0.49 to 3.34), worse general health (OR=2.16; 1.24 to 3.75) and greater depressive symptoms (OR=2.68; 1.44 to 4.98). We observed no association between psychosocial stressors and generalized anxiety symptoms assessed at midlife.

Conclusions:

Psychosocial stressors were associated with worse menopausal symptoms and well-being decades after initial report.

1. Introduction

Women in midlife may experience biological and psychological changes that affect their well-being, often related to the perimenopausal transition. In this life stage, many women experience physical symptoms such as hot flashes, sleep disturbance, and sexual dysfunction, which may adversely influence their quality of life.^1,2 During midlife, women are also susceptible to report psychological problems including anxiety and depressive symptoms.^3,4

Psychosocial correlates of menopausal symptoms and well-being include current and past history of adverse experiences such as history of abuse and financial hardship.^5-8 These stressors have been associated with a higher risk of physical and psychological problems among women in midlife, including greater menopausal symptoms and mood disorders.^5,9 Specifically, in cross-sectional studies of women in midlife, a retrospective report of adverse experiences in childhood was associated with higher levels of menopausal symptoms.^10,11 A systematic review also observed that individuals reporting adverse childhood experiences are at greater risk of health problems including sleep disturbance, depression and anxiety.¹² However, current knowledge related to the midlife period mainly relies on studies with small sample sizes, studies focusing on concurrent psychosocial stressors in the midlife period, or using a retrospective report of prior stressors^7,9-11,13-15 which are vulnerable to reverse causation and recall bias.

To address those gaps, we examined the associations of history of psychosocial stressors, reported at enrollment in early pregnancy in the Project Viva cohort in 1999-2002, with menopausal symptoms and well-being assessed in 2017-2021. We hypothesized that prior history of physical or sexual abuse or financial instability would be associated with greater menopausal symptoms and with worse well-being among women in midlife.

2. Materials and methods

2.1. Study design

Project Viva is a prospective cohort of mothers and children with enrollment in pregnancy and continued ongoing follow-up.¹⁶ Project Viva was initially established to examine prenatal health factors in relation to maternal and child health. Ongoing follow-up over the subsequent two decades has addressed a number of aims, including women’s health in midlife. Women were initially recruited between 1999 and 2002 from eight obstetric offices of Atrius Harvard Vanguard Medical Associates, a multispecialty group practice in eastern Massachusetts. Exclusion criteria included multiple gestation, inability to answer questions in English, gestational age ≥ 22 weeks at recruitment, and plans to move away from the study area before delivery. All women provided written informed consent. Institutional review boards reviewed and approved the project in line with ethical standards established by the Declaration of Helsinki. Of 2100 women with live singleton births, we included 682 women in the present analysis who provided data on psychosocial stressors reported in the index pregnancy and with menopausal symptoms and/or well-being values reported in midlife at a study visit in 2017-2021 (mean (standard deviation, SD) age = 52.0 (3.9) years). We restricted the analytic dataset to women who were postmenopausal or aged ≥ 45 years at the midlife follow-up, since we were specifically focusing on the peri/post-menopausal life stage. Supplemental Figure 1 presents the participant flow chart.

2.2. Exposures: Psychosocial stressors

2.2.1. History of physical and sexual abuse

At the mid-pregnancy visit (median 28 weeks of gestation), women reported history of past and recent physical and sexual abuse using the Personal Safety Questionnaire (PSQ),¹⁷ developed for this study and based on the Conflicts Tactics Scales.¹⁸ The PSQ queries the occurrence of specific incidents related to history of physical and sexual abuse by asking about people who harmed the respondent intentionally. Physical abuse was measured using five items: “push, grab or shove you,” “kick, bite or punch you,” “hit with something that hurt your body,” “choke or burn you,” and “physically attack you in some other way.” Sexual abuse was evaluated using one item “force you to have sexual activities.” The PSQ refers to incidents that occurred in four life phases: childhood (up to the age of 11 years by “anyone who was at least 5 years older than you”), adolescence (aged 12-17 years), adulthood (aged 18 years until this pregnancy) and in the current pregnancy. For our primary exposure, we categorized women reporting an experience of abuse at any period of life as having had a history of physical abuse ever, or sexual abuse ever, with the reference category of no reports of any physical or sexual abuse, respectively, for each category of abuse. We considered the experience of physical or sexual abuse at each separate life period as secondary exposures, with the reference categories of no abuse during that period.

2.2.2. History of financial instability

At the early pregnancy visit, women reported their past and recent history of financial instability using the question “Did you ever receive public assistance, receive welfare, or lack basic necessities (such as food, rent, or medical care) during the following periods? a) Before you were 18 years old b) From when you were 18 until this pregnancy c) During this pregnancy.” For our primary measure for analyses of this exposure, we considered participants as having experienced financial instability ever if they reported receiving public assistance, welfare or lacking basic necessities at any time point with the reference category of no report of financial difficulties in any life period. We considered the experience of financial instability at each separate life period as secondary exposures, with the reference categories of no financial instability during that period.

2.3. Outcomes: Menopausal symptoms and well-being

2.3.1. Menopausal symptoms

Women self-reported their menopausal symptoms using the Menopause Rating Scale (MRS).¹⁹ The MRS queries 11 menopausal symptoms over the past year on a 5-point scale of 0. None, 1. Mild, 2. Moderate, 3. Severe, or 4. Very Severe. The MRS provides three subscales and one total score of symptoms. The somato-vegetative scale (0-16 points) includes four categories of symptoms: hot flashes, heart discomfort, sleeping disorders, and joint and muscle complaints. The psychological scale (0-16 points) includes depressive complaints, irritability, anxiety complaints, and physical and mental exhaustion. The urogenital scale (0-12 points) includes sexual problems, bladder problems, and vaginal dryness. The total score of symptoms (0-44 points) is the sum of the three subscales. Total score was computed if at least 10 of the 11 items were available. Higher score indicates greater levels of menopausal symptoms. We observed good internal consistency for the overall MRS total score in our study population (Cronbach α = 0.83) with lower alphas in some subscales (psychological scale: Cronbach α = 0.85; somato-vegetative scale: Cronbach α = 0.59; urogenital scale: Cronbach α = 0.66).

2.3.2. General health

Women reported their general health using the question “In the past month, in general, would you say your health is” on a 5-point scale of 1. Excellent, 2. Very Good, 3. Good, 4. Fair, and 5. Poor. We dichotomized the general health self-evaluation as Good, Fair, or Poor vs. Excellent or Very Good.

2.3.3. Generalized anxiety symptoms

Women reported generalized anxiety symptoms using the Generalized Anxiety Disorder-7 (GAD-7).²⁰ The GAD-7 measures seven anxiety symptoms using the question “Over the last 2 weeks, how often have you been bothered by any of the following problems?.” Examples include “Feeling nervous, anxious, or on edge” or “Not being able to stop or control worrying.” Answers are on a 4-point scale of 0. Not at all, 1. Several days, 2. More than half the days, and 3. Nearly every day, with a possible total score of 0 to 21 points. Greater score indicates higher levels of generalized anxiety symptoms. The scale has been categorized into four levels: None or minimal (0-4 points), mild (5-9 points), moderate (10-14 points) and severe (15-21).²⁰ Given the skewed distribution of generalized anxiety scores in our sample, we dichotomized the variable as Mild, Moderate, or Severe vs. None or Minimal.

2.3.4. Depressive symptoms

Participants self-reported depressive symptoms using the Patient Health Questionnaire-9 (PHQ-9).²¹ The PHQ-9 measures nine depressive symptoms using the question “Over the last 2 weeks, how often have you been bothered by the following problems?.” Examples of symptoms include “Little interest or pleasure in doing things” or “Feeling down, depressed, or hopeless”. Answers are on a 4-point scale of 0. Not at all, 1. Several days, 2. More than half the days, and 3. Nearly every day, with a possible score of 0 to 27 points. Higher score indicates greater depressive symptoms. The scale allows five levels: None or minimal (0-4 points), mild (5-9 points), moderate (10-14 points), moderately severe (15-19), and severe (20-27).²¹ Given the skewed distribution of the depressive symptoms scores in our sample, we dichotomized the variable as Mild, Moderate, Moderately severe or Severe vs. None or Minimal.

2.4. Covariates

Using a self-administered questionnaire and interview at enrollment in early pregnancy, we collected information about the women’s age, race/ethnicity, education level, marital status, household income, and smoking status. We also assessed mid-pregnancy depressive symptoms using the Edinburgh Postnatal Depression Scale (EPDS);²² a score ≥13 indicates probable depression.²³

2.5. Statistical analysis

We analyzed each psychosocial exposure as a dichotomous variable (exposure to physical abuse, sexual abuse or financial instability, each yes vs. no). We categorized women reporting an experience of psychosocial stressors at any period of life as having had a history of physical abuse, sexual abuse, or financial instability ever. We also examined exposures within each life stage as additional analyses: in childhood, adolescence and adulthood (until the index pregnancy) for history of physical and sexual abuse; and in childhood or adolescence and in adulthood (until the index pregnancy) for financial instability (with the reference category of no report of stressors in the life period). We did not examine psychosocial stressors “in the current pregnancy” as separate exposures because only three women (0.5%) reported financial instability, eight women (1.4%) reported physical abuse, and none reported sexual abuse during the index pregnancy. Nevertheless, women reporting exposure to psychosocial stressors “in the current pregnancy” were considered as ever having an exposure to these stressors and were included in our main analyses.

We examined associations of a history of psychosocial stressors, reported at enrollment, with menopausal symptoms and well-being in midlife using linear regression for continuous outcomes and multivariable logistic regression for dichotomous outcomes. The first model (Model 1) was adjusted for race/ethnicity (categorized as white, Black, Asian, Hispanic, or “More than one race/ethnicity or other”), to account for differences in women’s experiences that may relate to both exposures and outcomes, and as such we consider it to be the only factor existing prior to potential exposures (true confounding) in our current analyses. The “More than one race/ethnicity or other” category includes 23 individuals reporting more than one race/ethnicity and one reporting “other” race/ethnicity. This grouping was for sample size purpose only and is not intended to undermine the distinct experiences of these individuals. We recognize that race is a social construct that does not necessarily reflect biological or cultural differences between groups.^24,25 Model 2 was further adjusted for sociodemographic and lifestyle characteristics measured at enrollment in the study (during the index pregnancy) and includes age at enrollment, education, marital status, household income, smoking status and probable prenatal depression (EPDS ≥ 13 in mid-pregnancy). We included these covariates in an additional model because they may be precision covariates or confounders via backdoor paths. Because these were assessed at study enrollment, they are temporally “downstream” of the psychosocial stressors we assessed as exposures, and thus may be at least partially mediators of the relationships of exposures prior to study enrollment with midlife outcomes.

For additional analyses examining associations of exposure to abuse in specific life periods with midlife outcomes, Model 2 additionally included prior experience of the same type of abuse. For example, when examining physical abuse in adulthood as the primary exposure of interest, we adjusted for physical abuse in childhood and in adolescence. Similarly, for analyses examining associations of financial instability in adulthood with midlife outcomes, we included experience of financial instability in childhood or adolescence (before 18 years) in Model 2.

To account for missing data, we performed multiple imputation for all 2100 women in Project Viva. We then limited the analyses to the 682 with non-missing exposure and outcome data. We used SAS (PROC MI) to impute 50 values for each missing observation and combined multivariable modeling estimates by using PROC MIANALYZE in SAS version 9.4 (SAS Institute). An alternative approach, including only participants with all covariate data (complete cases), yielded similar results (data not shown).

3. Results

3.1. Participants

At study enrollment, mean (SD) age of women was 33.7 (3.8) years, 80.3% had at least a college degree and 73.9% reported being non-Hispanic white (Table 1). A total of 37.3% of women reported a lifetime history of physical abuse, 7.7% a history of sexual abuse and 10.8% a history of financial instability. At the midlife visit, mean (SD) age was 52.0 (3.9) years. The mean (SD) MRS total score was 8.0 (5.7) points, the somato-vegetative scale was 3.4 (2.5) points, the urogenital scale was 1.7 (2.0) points and the psychological scale was 2.9 (2.7) points. A total of 31.3% of women reported good, fair, or poor general health. Generalized anxiety symptoms over the minimal threshold were reported by 12.2% and depressive symptoms by 16.0% of women.

Table 1:

Sociodemographic characteristics of the 682 included women in the Project Viva cohort

	Mean (SD) or N (%)
Characteristics at the index pregnancy
Age at enrollment, years	33.7 (3.8)
College degree, %	547 (80.3%)
Married or cohabiting, %	652 (95.9%)
Household income > $70,000/year, %	444 (68.1%)
Smoking during pregnancy, %	54 (7.9%)
Race/ethnicity, %
Black	76 (11.2%)
Hispanic	37 (5.4%)
Asian	41 (6.0%)
White	503 (73.9%)
More than one race/ethnicity or othera	24 (3.5%)
History of physical abuse, %
Ever	213 (37.3%)
In childhood	137 (24.0%)
In adolescence	113 (19.8%)
In adulthood (up to the index pregnancy)	81 (14.2%)
In the index pregnancy	8 (1.4%)
History of sexual abuse, %
Ever	44 (7.7%)
In childhood	16 (2.8%)
In adolescence	18 (3.2%)
In adulthood (up to the index pregnancy)	15 (2.6%)
In the index pregnancy	0 (0.0%)
History of financial instability, %
Ever	70 (10.8%)
In childhood or adolescence	52 (8.0%)
In adulthood (up to the index pregnancy)	23 (3.5%)
In the index pregnancy	3 (0.5%)
EPDS ≥13 in pregnancy, %	42 (7.0%)
Characteristics at the midlife period
Menopause Rating Scale
Somato-vegetative scale	3.4 (2.5)
Urogenital scale	1.7 (2.0)
Psychological scale	2.9 (2.7)
Total score of symptoms	8.0 (5.7)
General health, %
Excellent or very good	459 (68.7%)
Good, fair, or poor	209 (31.3%)
Generalized anxiety symptoms, %
None or minimal	570 (87.8%)
Mild, moderate, or severe	79 (12.2%)
Depressive symptoms, %
None or minimal	539 (84.0%)
Mild, moderate, moderately severe or severe	103 (16.0%)

Note.

^aThe “More than one race/ethnicity or other” category includes 23 individuals reporting more than one race/ethnicity and 1 individual reporting “other” race/ethnicity. EPDS: Edinburgh Postnatal Depression Scale; SD: Standard deviation.

Compared to the 682 included women, the 1418 participants excluded from our analyses were slightly younger at enrollment (30.9 vs 33.7 years), less likely to have a college degree (56.8% vs 80.3%) and a household income > $70,000/year (56.8% vs 68.1%) (Supplemental Table 1). The excluded women reported similar rates of sexual abuse compared to the included sample (9.5% vs 7.7%), physical abuse (35.8% vs 37.3%) and financial instability (12.3% vs 10.8%).

3.2. Regression analyses results

History of physical abuse

In Model 1, which was adjusted for race/ethnicity, women reporting a history of physical abuse had greater mean difference in MRS total symptom score (β = 1.13, 95% CI [0.14 to 2.11]), somato-vegetative score (0.51, 95% CI [0.09 to 0.93]), and psychological score (0.62, 95% CI [0.16 to 1.09]) (see Table 2). Moreover, a history of physical abuse was associated with greater odds of worse general health (odds ratio [OR] = 1.80, 95% CI [1.23 to 2.64]) and depressive symptoms over the minimal threshold (OR = 1.81, 95% CI [1.11 to 2.96]) (see Table 2). Model 2, accounting for additional covariates assessed at enrollment, provided a slightly weaker estimate with CI crossing the null for total MRS symptom score (0.96, 95% CI [−0.02 to 1.93]); all other results were similar in Model 2 compared with Model 1 (Table 2). History of physical abuse explained a small amount of variation in the outcomes in fully adjusted models (R² range = 0.03 to 0.16). History of physical abuse in childhood or adolescence, but not in adulthood, was related to worse menopausal symptoms with stronger estimates in the somato-vegetative and psychological scales (Supplemental Table 2).

Table 2:

Associations of psychosocial stressors over the lifetime with menopausal symptoms and well-being in midlife

	Menopause Rating Scale				Well-being
	Somato- vegetative	Urogenital	Psychological	Total score	Worse general health	Generalized anxiety symptoms	Depressive symptoms
	β (95% CI)				OR (95% CI)
History of physical abuse
N exposed / Total N	211/565				210/559	202/542	201/537
Model 1	0.51 (0.09 to 0.93)	−0.01 (−0.36 to 0.33)	0.62 (0.16 to 1.09)	1.13 (0.14 to 2.11)	1.80 (1.23 to 2.64)	1.41 (0.83 to 2.39)	1.81 (1.11 to 2.96)
Model 2	0.46 (0.04 to 0.87)	−0.02 (−0.36 to 0.33)	0.52 (0.07 to 0.97)	0.96 (−0.02 to 1.93)	1.73 (1.17 to 2.55)	1.39 (0.82 to 2.38)	1.74 (1.05 to 2.87)
History of sexual abuse
N exposed / Total N	44/565				43/559	39/542	41/537
Model 1	1.13 (0.38 to 1.88)	0.17 (−0.45 to 0.79)	1.62 (0.79 to 2.45)	2.92 (1.16 to 4.68)	1.97 (1.03 to 3.80)	2.16 (0.97 to 4.83)	1.72 (0.79 to 3.74)
Model 2	1.06 (0.31 to 1.81)	0.23 (−0.40 to 0.85)	1.52 (0.70 to 2.33)	2.81 (1.05 to 4.56)	2.04 (1.04 to 4.03)	2.06 (0.91 to 4.66)	1.65 (0.74 to 3.70)
History of financial instability
N exposed / Total N	69/641				68/636	69/618	69/609
Model 1	0.70 (0.09 to 1.32)	0.62 (0.12 to 1.12)	0.60 (−0.06 to 1.25)	1.92 (0.50 to 3.34)	2.29 (1.35 to 3.88)	1.38 (0.67 to 2.83)	2.64 (1.47 to 4.74)
Model 2	0.64 (0.02 to 1.26)	0.69 (0.19 to 1.20)	0.58 (−0.08 to 1.23)	1.92 (0.49 to 3.34)	2.16 (1.24 to 3.75)	1.30 (0.62 to 2.72)	2.68 (1.44 to 4.98)

Note. We dichotomized psychosocial stressors as 1 = history ever vs. 0 = no history (reference), general health as 1 = good, fair, or poor vs. 0 = excellent or very good (reference), and generalized anxiety and depressive symptoms scores as 1 = mild, moderate, or severe vs. 0 = none or minimal (reference).

β: Mean difference; CI: Confidence intervals; OR: Odds ratio.

Model 1: Adjusted for race/ethnicity.

Model 2: Model 1 + covariates at enrollment: age, education, marital status, household income, smoking status, and probable depression in mid-pregnancy.

History of sexual abuse

Women reporting a history of sexual abuse had 2.92 points (95% CI [1.16 to 4.68]) higher MRS total score adjusted for race/ethnicity, mainly explained by associations in the somato-vegetative (1.13 points, 95% CI [0.38 to 1.88]) and psychological (1.62 points, 95% CI [0.79 to 2.45]) scales (see Table 2). We observed similar results in Model 2 (Table 2). Moreover, reporting a history of sexual abuse was associated with higher odds of worse general health in Model 1 (OR = 1.97, 95% CI [1.03 to 3.80]) with a similar estimate in Model 2 (Table 2). History of sexual abuse explained a small amount of variation in the outcomes in fully adjusted models (R² range = 0.03 to 0.15). History of sexual abuse in childhood and adolescence was associated with greater levels of menopausal symptoms (Supplemental Table 2) while history in adolescence was associated with overall greater odds of worse general health (Supplemental Table 3).

History of financial instability

In Model 1, reporting a history of financial instability was associated with 1.92 points (95% CI [0.50 to 3.34]) higher MRS total score, reflecting associations in the somato-vegetative (0.70 points, 95% CI [0.09 to 1.32]) and urogenital (0.62 points, 95% CI [0.12 to 1.12]) scales (see Table 2). History of financial instability was also associated with higher odds of worse general health (OR = 2.29, 95% CI [1.35 to 3.88]) and greater depressive symptoms (OR = 2.64, 95% CI [1.47 to 4.74]) in midlife (see Table 2). Again, estimates were similar in Model 2. History of financial instability explained a small amount of variation in the outcomes in fully adjusted models (R² range = 0.04 to 0.12). Financial instability in childhood, adolescence or adulthood was associated with overall higher levels of menopausal symptoms and worse general health, while exposure to financial instability in childhood or adolescence was specifically associated with greater depressive symptoms (Supplemental Tables 2 and 3).

We did not find significant associations between any exposures and generalized anxiety symptoms assessed at the midlife visit (Table 2).

4. Discussion

This study examined associations of women’s history of psychosocial stressors in childhood, adolescence, and early adulthood (reported on average in their early thirties) with menopausal symptoms and well-being in midlife. We observed that a history of physical or sexual abuse, or the experience of financial instability, were associated with worse menopausal symptoms and with poorer well-being almost two decades after exposure assessment.

Our study confirms prior research observing that a history of psychosocial stressors is associated with higher levels of menopausal symptoms in midlife. For example, a study including 332 White and African American women reported that childhood abuse or neglect was associated with greater risk of reporting hot flashes and night sweats.⁹ Moreover, a study composed of 295 perimenopausal and postmenopausal women observed that childhood physical or sexual abuse was associated with greater levels of vasomotor symptoms during sleep.¹⁰ However, these studies examined exposure to childhood adverse experiences retrospectively during the midlife period and focused on history of abuse in the childhood period only. In our current study, we were able to look at history of psychosocial stressors in childhood, adolescence and early adulthood allowing us to present a better overview of the associations of abuse across multiple periods of life reported by women in average in their thirties with reports of menopausal symptoms almost two decades later.

We also reported few associations of psychosocial stressors in relation to urogenital symptoms. We observed low score variability in the urogenital scale due to relatively low levels of symptoms, which are expected given that only about 48% of women had reach menopause at the midlife visit; likely limiting our ability to detect an association. Alternatively, the levels of urogenital symptoms experienced by women might be more strongly related to alternate biological or environmental mechanisms that are less susceptible to prior history of psychosocial stressors. However, we remain cautious about specific results of MRS sub-domain given that we reported lower Cronbach alphas in the somato-vegetative and urogenital scales in comparison to the psychological scale or the MRS total score.

Our findings in the mental health domain also suggested stronger associations of psychosocial stressors with depressive symptoms, measured by the PHQ-9, in comparison to generalized anxiety symptoms, assessed by the GAD-7. The association between history of psychosocial stressors, such as physical or sexual abuse, and depressive symptoms in adulthood is well documented.^26-29 Past research also reported that history of physical or sexual abuse in childhood is related to greater risk of anxiety disorders during adulthood, including generalized anxiety disorder.^30-32 In this study, we were surprised not to observe associations between psychosocial stressors and generalized anxiety levels among women in midlife. However, in comparison to some previous studies, we did not use a clinical evaluation or interview that would have led to a formal diagnosis of anxiety disorder. Instead, we evaluated generalized anxiety levels using the GAD-7, which although brief is nevertheless a widely used and validated questionnaire.²⁰ Our sample is mostly composed of white and well-educated women with a majority reporting none or minimal anxiety and depressive symptoms. All estimates for anxiety symptoms were in the hypothesized direction and similar to those for depressive symptoms and general health, most notably in relation to prior history of sexual abuse. However, confidence intervals were wide and included the null, perhaps in part because this outcome was less frequently reported than the other two dichotomous outcomes (12.2% for anxiety symptoms versus 16.0% for depressive symptoms and 31.3% for good, fair, or poor general health).

Underlying mechanisms of these relationships are undeniably complex and multiple, and include behavioral, social, cognitive and emotional pathways.³³ Adults reporting a history of childhood abuse are susceptible to engage in harmful activities and to report negative beliefs and attitudes towards others and themselves.³³ According to Kendall-Tackett,³³ these factors may increase the risk of physical and mental health problems in these individuals over time. Moreover, adverse childhood experiences have been established as key contributors to hypothalamic–pituitary–adrenal (HPA) axis dysregulation, as indicated by abnormal cortisol dynamics levels,^34,35 which is related to psychiatric diseases and health problems.^36,37 A prior history of childhood abuse is linked to mental health problems, such as anxiety and depression^12,13 and was previously found to be related to higher odds of prenatal depression in women in this cohort.¹⁷ Additionally, people with a history of childhood abuse tend to report poorer self-rated health and more health care utilization in adulthood.^38,39 Overall, our results and those of previous studies highlighted the need to shed light on the underlying pathways that would allow a better understanding of the mechanisms for how psychosocial stressors affect menopausal symptoms and well-being in midlife, with the ultimate goal of identifying targets for intervention to prevent these adverse sequelae.

This study should be considered in the context of some strengths and limitations. Among its strengths, this study is based on almost two decades of data collection with follow-up into midlife. We used a prospective design, which may help reduce recall bias and makes the temporal relationship clearer, compared to cross-sectional or retrospective studies. Moreover, our relatively large sample size and availability of rich sociodemographic and lifestyle data allowed us to account for key covariates such as household income, and mid-pregnancy depressive symptoms. We were thoughtful about separately considering race/ethnicity, a characteristic present before potential exposures, from other characteristics such as adult income and age at pregnancy, which might be influenced by the earlier experience of a stressor, and thus at least partially serve as mediators. Nonetheless, associations were essentially identical with and without such adjustment, suggesting that these factors were neither strong confounders or mediators. We also offer a more holistic overview of the associations between prior psychosocial stressors and well-being in midlife by including information about history of both physical and sexual abuse as well as financial instability in multiple periods of life (e.g. childhood, adolescence and early adulthood) while prior studies have generally focused on the experience of abuse in childhood only.

Despite the prospective design, a limitation of the current study lies in the retrospective evaluation of psychosocial stressors, which may be influenced by factors such as mood state at the time of assessment. Experiences of psychosocial stressors may also be underreported since participants may choose not to disclose these experiences to avoid painful emotions; or they may have been forgotten if they occurred at a very young age. Second, we did not directly ascertain the women’s menopausal status with detailed menstrual histories or hormone levels; thus, menopausal symptoms examined in this study may not be specifically related to the menopause transition. However, all women were at least age 45, an age when menopause symptoms typically occur. Since our study covers a large time span, there might be selection bias in women participating in the long-term follow-up. Moreover, our sample was mostly composed of white and well-educated women, thus, our results may not generalize to populations of other ethnic backgrounds and may explain the low levels of generalized anxiety and depressive symptoms reported in this cohort. Given the number of associations examined in this study, we remain cautious about the interpretation of statistical significance for any single model. However, the fact that many significant associations were in the direction of our hypothesis and were seen across our several related exposures and outcomes brings confidence in our findings.

5. Conclusion

Among 682 women enrolled in pregnancy in the Project Viva prospective cohort, a history of psychosocial stressors from childhood up to the index pregnancy was associated with worse menopausal symptoms and poorer well-being almost two decades later. These results highlight the long-lasting influence of adverse experiences on women's physical and mental health and emphasize the importance of past history of psychosocial stressors when considering the health of women in midlife. Given the cumulative nature of life course experiences, future studies should examine how the co-occurrence of multiple stressful exposures across the life course independently and jointly affect health and well-being to identify potential targets for health promotion efforts.