Women experience a higher level of mental health morbidity as compared to men over the life course.1 A particularly vulnerable time for women occurs during the menopausal transition, which is marked by drastic fluctuations in sex hormones.2 This transition is often accompanied by the occurrence of significant life stressors and changes in personal, family, and professional responsibilities.3,4 A growing body of literature has suggested that, for some women, the perimenopause and the early postmenopausal years are associated with an increased risk of experiencing symptoms of depression and the development of an episode of major depressive disorder.5 Depression can be a disabling condition and can lead to significant personal, economic, and societal costs. Although the physiological health consequences of menopause have been relatively well-characterized in the existing literature, the specific influence on mental well-being during this time is less clear.
As the baby boom generation reaches midlife, a large number of women are now postmenopausal in Canada. It is important for health care providers to understand the window of vulnerability for depressive symptoms among mid-life women.6 In particular, the menopause transition, which occurs 2 to 3 years before the final menstrual period, and the early postmenopausal years have been associated with an increased risk for depressive symptoms as well as a twofold risk for major depressive disorder among women.7-9 Endogenous estrogens are known to modulate systems that are critical for mood regulation5,10 but the therapeutic benefit of hormone therapy (HT) has not been clearly elucidated among postmenopausal women for the treatment of depressive symptoms.7 A large systematic review of 10 studies reported that vasomotor symptoms (VMS) increase the risk for depressive symptoms during the perimenopause.11 The benefits of HT have been well-described for the treatment of VMS and for improving the quality of life among women. It is, however, unclear whether HT use is associated with a lower prevalence of depressive symptoms among postmenopausal women.
The main objective of this research was to investigate the association between menopausal status and self-reported depressive symptoms among middle-aged women in Canada. We explored the potential influences of the social determinants of health (education, income, social support) and health behaviors on current self-reported depressive symptoms. The characteristics of the menopausal transition and HT use were also examined to uncover potential associations.
Setting and study population
The Canadian Longitudinal Study on Aging (CLSA) is one of the largest and most comprehensive research platforms in the world, examining health and aging in community-dwelling participants between the ages of 45 and 85 years at the time of recruitment from the 10 provinces. People were excluded from the CLSA if they lived on federal First Nations reserves, were full-time members of the Canadian Armed Forces, lived in institutions, were unable to respond in English or French, or were cognitively impaired. Enrollment and baseline data collection from 51,338 participants across Canada were completed from 2010 to 2015. There were 21,241 participants recruited into the Tracking Cohort that provided information through telephone interviews and 30,097 participants recruited into the Comprehensive Cohort who completed in-depth physical assessments and provided blood and urine samples in addition to completing questionnaires in person. The CLSA study design has been previously described in detail elsewhere.12 Baseline data from the Tracking Cohort (Dataset version 3.1) and Comprehensive Cohort (Dataset version 2.0) of the CLSA were accessed for this research project. Before accessing and analyzing the data, ethics approval was received from the Hamilton Integrated Research Ethics Board (Protocol Number: 0204). For these analyses, cross-sectional baseline data were used. Women aged 65 and older (n = 10,749) and those missing data on the exposure or the outcome (n = 326) were excluded. We also excluded participants who had undergone a hysterectomy at baseline (n = 1,864) due to lack of complete detailed information about the procedure. Women were not asked whether they had ovarian conservation or removal, rendering the age at true menopause difficult to ascertain. The analyses included 13,216 female CLSA participants (Fig. 1).
The primary outcome of this study was depression status measured using the Center for Epidemiologic Studies Short Depression Scale (CESD-10), which contains 10 questions about items such as feelings of depression, loneliness, hopefulness for the future, and restless sleep. Four possible response options were provided to the participants for each question (all of the time, occasionally, some of the time, rarely or never) and the total score may range from 0 to 30. A total score of 10 or more is considered depressed.13,14 This self-report scale has high sensitivity (0.89) to estimate current depression with a cut-off score of 10 points, and has demonstrated both convergent and divergent validity.13,14
Self-reported menopausal status is commonly used in population-based studies. Women who reported having their menstrual periods stopped for at least 1 year without restarting, or who were over the age of 60 years were considered postmenopausal. All other women were considered premenopausal. Postmenopausal women were asked at what age menopause occurred. Based on this information, postmenopausal women were categorized as premature onset (<40 years), early onset (40-45 years), normal onset (46-55 years), and late onset (≥56 years) age at menopause. The validity of self-reported age at menopause has found to be acceptable but shown to decrease with increasing number of years since menopause.15,16 Although the option was not read to the participants, some women reported to have had a hysterectomy. These participants were excluded from the analyses.
Hormone therapy use
Women were asked if they have ever used HT. Women reporting HT use were asked at what age they started HT, if estrogen, progesterone, or both were used, and how long HT was used.
Sociodemographic and lifestyle variables
We explored sociodemographic factors previously reported to be associated with depression: highest attained education level (less than secondary, secondary, some postsecondary, postsecondary), household income (CAD$: <20,000; 20,000-49,000; 50,000-99,000; 100,000-149,000; ≥150,000), marital status (single, married/common law, widowed/divorced/ separated), and parity (0 biological children, ≥1 biological children). Lifestyle factors and health-related variables included body mass index (BMI; underweight <18.5 kg/m2; normal weight 18.5-24.9 kg/m2; overweight 25.0-29.9 kg/m2; obese ≥30.0 kg/m2); smoking status (never, former, current) and alcohol consumption (regular, occasional, infrequent); and perceived mental health and perceived general health (both categorized as poor/fair and good/very good/excellent). Social support was measured to capture whether an individual was receiving affectionate social interaction, emotional and informational support, positive social interaction, tangible social support, and whether or not they received caregiving. A social participation score was calculated based on frequency of involvement in social activities and then categorized by age-specific quintiles.
Descriptive statistics, expressed as frequencies and percentages, were calculated using chi square analyses across the different categories of depression status and menopausal status. The median and interquartile range of the CESD-10 score for each group of women was also calculated. Multivariable analyses were calculated using logistic regression (based on the presence or absence of self-reported depression on the CESD-10) and adjusted odds ratios (aORs) are presented. Age, education, and income were included in the model as a priori variables and lifestyle factors (smoking and alcohol consumption), health-related variables (BMI, perceived general health, and mental health), social support, and social participation variables were added one by one in the model to see the effect changes. Odds ratios (OR) were calculated along with the 95% confidence interval (95% CI). All hypotheses tests were two-sided and P values less than 0.05 were considered significant. The main analysis was done for women ages 45 to 64 years; however, we also performed the same analysis, stratified by age 45 to 54 and 55 to 64 years to see if any differences emerged. The stratification was completed as the literature suggests that the most vulnerable time for depressive symptoms exists during the menopause transition. All analyses were performed using sampling weights with appropriate methods for complex survey design methods. Descriptive analysis was done with sampling weights and association was measured using probability weights. Complex survey design methods were employed here to calculated estimates adjusted with strata for calculating appropriate CIs. All analyses were conducted using Stata version 14.2 software (StataCorp. 2015. Stata Statistical Software: Release 14.StataCorp LP, College Station, TX) and SAS 9.4 software (SAS Institute Inc 2013, SAS/ACCESS 9.4, Cary, NC).
Data regarding a history of physician diagnosed clinical depression and the current use of medication to treat depression were available in the Comprehensive Cohort (approximately 40% of participants), but not the Tracking Cohort. A sensitivity analyses was conducted to compare the odds of current depression based on the CESD-10 in the Comprehensive Cohort with and without these variables included.
The sociodemographic and lifestyle variables for middle-aged women, stratified by self-reported depression (estimated using CESD-10), are presented in Table 1. Using the CESD-10 score, 18.1% of the middle-aged women were identified as depressed with a median score of 12.1 (interquartile range 10.3-15.4). When stratified by age group 45 to 54 years (an estimate for the mean age of perimenopause and early menopause) versus 55 to 64 years, there was no difference in proportion who were categorized as depressed. Sociodemographic and lifestyle characteristics were different between depressed and nondepressed participants. Women who were not identified as depressed based on the CESD-10 had a higher level of education (76.8% vs 73.4%, completed at least some postsecondary), higher income (20.1% vs 10.5%, had more than $150,000 yearly household income), were married or common law (78.2% vs 65.9%), had multiple children (86.7% vs 80.9%), were former or never smokers (90.2% vs 81.7%), were normal or over weight (73.9% vs 64.4%), had higher levels of social support (affectionate social interaction: 81.6% vs 60.1%; emotional and informational support: 69.0% vs 46.7%; positive social interaction: 68.0% vs 43.1%; tangible social support: 66.5% vs 46.2%); and rated their mental and general health as good, very good, or excellent (mental health: 97.9% vs 77.1%; general health: 93.9% vs 73.6%) compared to women who met the CESD-10 criteria for depression.
Menopausal transition and HT use
The majority of women were postmenopausal and never users of HT. The average age of menopause was approximately 49 years (not depressed: 49.2 years; depressed: 48.2 years) (Table 2). There were 1,318 women currently using HT. Among women who were current HT users, the majority of women were using HT for less than 5 years, and started HT use either the same year or after menopause.
Depression status and covariates
The association of current depression status (as estimated using the CESD-10) and covariates adjusted for age, education, income, and marital status are presented in Table 3. Women who reported current smoking, regular alcohol use, obesity, and self-reported “fair” or “poor” general health had a higher risk of current depression based on the CESD-10. Moderate to high social interaction, emotional and informational support, positive social interaction, tangible social support, and social participation reduced the risk of current depression among women.
Menopausal transition, HT use, and depression status
After adjustment for age, education level, household income, marital status, BMI, smoking status, alcohol consumption, social support, perceived general health, and social participation, current users of HT had a higher odds for current self-reported depression on the CESD-10 (aOR: 1.21, 95% CI: 1.02-1.44) as did women who experienced premature menopause (aOR: 1.45, 95% CI: 1.07-1.97) (Fig. 2). Supplemental Digital Content 1, http://links.lww.com/MENO/A577 contains the unadjusted OR and aORs. The age-stratified analyses with participants aged 45 to 54 years and participants aged 55 to 64 years did not reveal any meaningful differences for the magnitude of the association of premature menopause or HT use with current self-reported depression on the CESD-10. For women aged 45 to 54 years and 55 to 64 years, respectively, the adjusted odds of depression based on the CESD-10 for HT use were 1.20 (95% CI: 0.91-1.56) and 1.24 (95% CI: 1.00-1.55) and for premature menopause the aORs were 1.19 (95% CI: 0.51-2.77) and 1.52 (1.08-2.14).
In the Comprehensive Cohort, a history of physician diagnosed clinical depression and the current use of medication to treat depression were associated with a 1.77 (95% CI: 1.44-2.14) and a 1.88 (95% CI: 1.52-2.32) greater odds of current depression based on the CESD-10 in the adjusted model (Supplemental Digital Content 2, http://links.lww.com/MENO/A578). Including a history of depression and use of medication to treat depression had a minimal impact on the magnitude of the association between any of the other variables with depression.
Overall, 18.1% of middle-aged women in the CLSA data were identified as “depressed,” as estimated by a score of 10 or more on the CESD-10. A premature menopause and current use of HT were identified as risk factors for current self-reported depression after adjustment for other important variables. We did not observe any association with duration of current HT use with current symptoms of depression. Several positive social determinants of health and health behaviors were observed as protective (eg, education level, household income, nonsmoking status, social support, and social participation).
The estimated prevalence of depressive symptoms in our sample of women (18.1%) is lower than previously reported in American, Polish, and Chinese Cohorts around the time of the menopausal transition, with rates ranging from 25% to 36%.17-19 Nevertheless, these findings from our large representative sample suggest that a large number of Canadian women are at risk. Previously identified prognostic factors for poor mental health were also demonstrated in our sample, including low social support, lower income, obesity, and cigarette smoking. Some of these factors may help health care providers to identify which patients are at the highest risk, and may be modifiable, such as cigarette smoking and obesity. In the United Kingdom, the government has recently appointed a “Minister for Loneliness” as poor social support has been identified as a significant public health challenge (https://www.gov.uk/government/news/pm-launches-governments-first-loneliness-strategy). Health care providers have begun “social prescribing,” which will potentially lessen the isolation commonly seen in older adults. Perhaps a similar program in Canada would help to improve mental health given that we found such profound differences on all of our social support and interaction items (Table 1).
A recent meta-analysis investigated the association between age at menopause and risk for depression among postmenopausal women with naturally occurring menopause.20 Among the 14 included studies, an inverse association with depression in postmenopausal women was shown for increasing age at menopause. For example, those women who experienced premature menopause (before age of 40 years) experienced a 50% increased risk for depression as compared to women with longer reproductive years. Similarly, a study using the Korean National Health and Nutrition Examination Survey (KNHANES V) reported that those women who experienced an earlier menopause (before the age of 45) experienced a 65% increased risk for depression as compared to women with longer reproductive years.21 As such, longer exposure to endogenous estrogens (having a normal or late onset of menopause, either naturally or surgically) can be protective against the occurrence of depression in later life. Considering these findings, it is of utmost importance to identify those women who experience premature menopause (from any cause), and who may be at a higher risk for depression (such as those women with previous depressive episodes). These women may benefit from psychiatric support and intervention and/or estrogen-based therapies, as clinically appropriate. The perimenopause and the early postmenopausal years represent a vulnerable time for the development of depressive symptoms as related to the dramatic shifts in endogenous estrogen levels22 and we therefore expected to find a greater proportion of self-reported depression among women closer to the mean age of the menopause transition (45-54 years) compared to women aged 55 to 64 years. This expected result was not observed, but may be in part attributable to our study measuring current symptoms of depression rather than incident depression.
The finding that current HT use was associated with an increased risk for depressive symptoms is interesting and complex. Although the models were adjusted for a range of covariates, we were unable to adjust for the presence of menopausal VMS as these data were not captured by the CLSA. VMS are the primary reason why HT is used in menopause and current use may be a proxy for more severe symptoms, a known risk factor for depressive symptoms. HT has been used to successfully treat depression in conjunction with antidepressant medications.23 Further research should explore the relationship between HT use and ongoing depressive symptoms as women experience the menopausal transition. The burden associated with symptoms or diagnoses for depression at any stage of life is considerable and undeniable. The current description of the CLSA data is from the baseline visit and therefore cross-sectional in nature. Future analyses using the longitudinal data of these women may, however, capture the changes that may occur in older age (such as loss of a partner or social isolation) and that can impact the health and well-being of women before, during, and after the menopausal transition.
A key strength of this study is the use of the CLSA Cohort data, which captures information on the menopausal transition and self-reported depressive symptoms among a large group of midlife and older women in Canada. An important limitation to these data is that both the prior physician diagnosis of clinical depression and the current depressive symptoms, as measured using the CESD-10, are self-report data which may be subject to social desirability bias. Recall bias may have affected the reporting of these variables, and the timing of the menopausal transition and details of HT use. Another limitation is that certain populations were not included in the CLSA (women living on federal First Nations reserves; full-time members of the Canadian Armed Forces; those living in institutions; those who were unable to respond in English or French; and women who were cognitively impaired), and these women are often vulnerable to mental distress, including depression. The CLSA also did not include women younger than 45 years of age; therefore, stratified analyses investigating the association of menopausal status and depression in women currently undergoing early menopause could not be evaluated. Previous studies have included a prior diagnosis of depression and the use of medication to treat depression as covariates in their models. This information was only available for Comprehensive Cohort participants (approximately 40% of participants) and therefore could not be included in our main analyses. Although both variables were associated with increased odds of being identified as depressed, as estimated using the CESD-10, they did not meaningfully alter the association between the other variables and depression when the analyses were limited to the Comprehensive Cohort. Therefore, it is unlikely that the exclusion of these variables had a meaningful impact in our main analyses.
Overall based on the CLSA Cohort, approximately 18% of women in Canada aged 45 to 64 had self-reported high levels of depressive symptoms measured using the CESD-10. Premature menopause (menopause occurring before 40 years) and current use of HT were both associated with an increased odds of having current depression after adjustment for other covariates. Other risk factors for depression included low education, low household income, living alone, being nulliparous, and low social support. Given that the menopausal transition is a vulnerable time for mental health, the identification of risk factors, including social determinants of health, age at menopause, and menopausal symptoms can help guide clinicians when assessing mental health.
This research was made possible using the data collected by the Canadian Longitudinal Study on Aging (CLSA). Baseline data from the Tracking Cohort (Dataset version 3.1) and Comprehensive Cohort (Dataset version 2.0) of the CLSA were accessed for this research project, under Application Number 150308. The CLSA is led by Drs. Parminder Raina, Christina Wolfson, and Susan Kirkland.
1. Knoll A, MacLennan R. Prevalence and correlates of depression
in Canada: findings from the Canadian Community Health Survey. Can Psychol Can
2. Soares CN. Depression
: an update on current knowledge and clinical management for this critical window. Med Clin North Am
3. Soares CN. Depression
: current knowledge and clinical recommendations for a critical window. Psychiatr Clin North Am
4. Gordon JL, Rubinow DR, Eisenlohr-Moul TA, Leserman J, Girdler SS. Estradiol variability, stressful life events and the emergence of depressive symptomatology during the menopause
5. Lokuge S, Frey BN, Foster JA, Soares CN, Steiner M. Depression
in women: windows of vulnerability and new insights into the link between estrogen and serotonin. J Clin Psychiatry
6. Soares CN. Mood disorders in midlife women: understanding the critical window and its clinical implications. Menopause
7. Maki PM, Kornstein SG, Joffe H, et al. Guidelines for the evaluation and treatment of perimenopausal depression
: summary and recommendations. Menopause
8. Bromberger JT, SChott L, Kravitz HM, Joffe H. Risk factors for major depression
during midlife among a community sample of women with and without prior major depression
: are they the same or different? Psychol Med
9. Freeman EW, Sammel MD, Lin H, Nelson DB. Associations of hormones and menopausal status with depressed mood in women with no history of depression
. Arch Gen Psychiatry
10. Chhibber A, Woody SK, Rumi MK, Soares MJ, Zhao L. Estrogen receptor B deficiency impairs BDNF-5-HT2A signaling in the hippocampus of female brain: a possible mechanism for menopausal depression
11. Worsley R, Bell R, Kulkarni J, Davis SR. The association between vasomotor symptoms and depression
during perimenopause: a systematic review. Maturitas
12. Raina PS, Wolfson C, Kirkland SA, et al. The Canadian Longitudinal Study on Aging (CLSA). Can J Aging
13. Andresen E, Malmgren J, Cater W, Patrick D. Screening for depression
in well older adults: evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression
Scale). Am J Prev Med
14. Björgvinsson T, Kertz SJ, Bigda-Peyton JS, McCoy KL, Aderka IM. Psychometric properties of the CES-D-10 in a psychiatric sample. Assessment
15. Colditz GA, Stampfer MJ, Willett WC, et al. Reproducibility and validity of self-reported menopausal status in a prospective cohort study. Am J Epidemiol
16. Den Tonkelaar I. Validity and reproducibility of self-reported age at menopause
in women participating in the DOM-project. Maturitas
17. Marsh WK, Bromberger JT, Crawford SL, et al. Lifelong estradiol exposure and risk of depressive symptoms during the transition to menopause
and postmenopause. Menopause
18. Grochans E, Szkup M, Kotwas A, Kopeć J, Karakiewicz B, Jurczak A. Analysis of sociodemographic, psychological, and genetic factors contributing to depressive symptoms in pre-, peri- and postmenopausal women. Int J Environ Res Public Health
2018; 15:pii: E712.
19. Zeng LN, Yang Y, Feng Y, et al. The prevalence of depression
in menopausal women in China: a meta-analysis of observational studies. J Affect Disord
20. Georgakis MK, Thomopoulos TP, Diamantaras AA, et al. Association of age at menopause
and duration of reproductive period with depression
: a systematic review and meta-analysis. JAMA Psychiatry
21. Jung SJ, Shin A, Kang D. Hormone-related factors and post-menopausal onset depression
: Results from KNHANES (2010-2012). J Affect Disord
22. Soares CN. Depression
in peri- and postmenopausal women: prevalence, pathophysiology and pharmacological management. Drugs Aging
23. Soares CN, Poitras JR, Prouty J, Alexander AB, Shifren JL, Cohen LS. Efficacy of citalopram as a monotherapy or as an adjunctive treatment to estrogen therapy for perimenopausal and postmenopausal women with depression
and vasomotor symptoms. J Clin Psychiatry