Evidence suggests that depressive symptoms can increase among perimenopausal and post-menopausal women,Reference Freeman1–Reference Freeman, Sammel, Lin and Nelson5 although findings are mixed regarding clinical depression.Reference Freeman1 Depressive symptoms have been found to associate with both hormonal levels and bothersome physical symptoms during perimenopause.Reference Freeman, Sammel, Liu, Gracia, Nelson and Hollander6,Reference Campbell, Gorelik, Szoeke and Dennerstein7 Few studies have investigated the relationship between generalised anxiety disorder and menopause, and results have been inconclusive overall.Reference Bryant, Judd and Hickey8 Nevertheless, one study in China found an association between increased anxiety symptoms and both peri- and post-menopause,Reference Tang, Luo, Li, Peng, Wang and Liu9 and a US study found that the perimenopausal period was associated with new-onset high anxiety levels.Reference Bromberger, Kravitz, Chang, Randolph, Avis and Gold10 There have been few UK studies to compare depression or anxiety across different menopausal stages.
The aim of this study was to investigate the odds of screening positive for possible depression and generalised anxiety disorder (GAD) in women in the menopausal stages relative to the pre-menopausal stage. Our study builds upon previous work by using a nationally representative household survey. Our hypothesis was that women in the pre- and perimenopausal stages would have a higher chance of screening positive for both outcomes, and would report higher levels of anxiety and depression symptoms.
Method
Sample
We conducted a secondary analysis of data from the Adult Psychiatric Morbidity Survey (APMS) 2014.Reference McManus, Bebbington, Jenkins, Morgan, Brown and Collinson11 The APMS is a household survey conducted once every 7 years for the purpose of understanding the prevalence of mental health problems in England. Data was collected throughout 2014. The sampling process is described in detail elsewhere.Reference McManus, Bebbington, Jenkins, Morgan, Brown and Collinson11 We restricted analyses to participants aged 40–59 years who reported female gender and answered menopause-related questions (below).
Exposure variable
Data was gathered by computer-assisted self-interviewing. Eligible participants were asked ‘Over the last 12 months, to what extent do you think you have been through or are going through the menopause?’ Participants who answered ‘Yes, I have been through menopause’ were defined as being in the post-menopause stage; ‘Yes, I am going through menopause’ as being in the perimenopause stage; and ‘No, not yet’ as being in the pre-menopause stage. We also combined the peri- and post-menopausal stages in separate analyses, for comparability with previous research.
Outcome variables
Anxiety and depression symptoms were assessed in the face to face part of the interview, using the Clinical Interview Schedule–Revised (CIS–R).Reference Lewis, Pelosi, Araya and Dunn12 The interviewers used computer-assisted self-interviewing to ask if participants experienced symptoms in the past 7 days. Participants screened positive for possible depression or GAD if they met ICD-1013 criteria for moderate or severe depression or GAD.
We also used total CIS–R score after subtracting the scores for irritability, fatigue, somatic symptoms and sleep problems, which could be inherent symptoms of menopause.
Covariates
We adjusted results for several covariates that could confound associations: continuous age in years; quintile of Index of Multiple Deprivation score;14 self-reported presence of chronic disease (asthma, cancer, epilepsy, diabetes and/or high blood pressure in the past 12 months);Reference Roa-Díaz, Raguindin, Bano, Laine, Muka and Glisic15,Reference Russell, Broderick, Taylor, Fernandes, Harvey and Barnes16 smoking status (none, ≤14 average daily number of cigarettes and ≥15 average daily number of cigarettes);Reference Parazzini17,18 Alcohol Use Disorders Identification Test score;Reference Ceylan and Özerdoğan19–Reference Saunders, Aasland, Babor, de la Fuente and Grant21 and educational level (no qualifications, foreign/other, GCSE or equivalent, A level, vocational and degree).Reference Ceylan and Özerdoğan19,Reference Bromberger and Epperson22
Statistical analyses
We used logistic regression where the outcome was binary, and linear regression for continuous CIS–R score. All analyses were conducted in Stata version 17 for Mac and Windows.Reference StataCorp23 We used original APMS survey weighting, using the svy command in Stata, to account for selection and non-response bias. We report unweighted absolute numbers with weighted percentages. We conducted all analyses as complete-case analyses, unadjusted and adjusted for putative confounding variables.
Results
Of the 7546 participants who were surveyed for the APMS 2014, 4488 reported female gender, 1510 were aged 40–59 years and 1413 (93.6% of those eligible) had answered all relevant menopause-related questions and constituted the analytic sample. Among these, 626 (45.2%) were classed as pre-menopausal; 409 (29.4%) were perimenopausal and 378 (25.5%) were post-menopausal. Further details can be seen in Table 1. The post-menopausal group tended to be older and have no qualifications compared with the other two groups.
Table 1 Characteristics of analytic sample according to menopausal stage
AUDIT, Alcohol Use Disorders Identification Test.
Association between menopause stage and depression
Participants who were post-menopausal were more likely to screen positive for moderate to severe depression (3.9%) than those who were pre-menopausal (1.7%; odds ratio 2.31, 95% CI 1.10–4.89; P = 0.028) (Supplementary Table 1 available at https://doi.org/10.1192/bjo.2023.82). Notably, the prevalence in women who were post-menopausal was similar to the overall rate in women in the 2014 APMS (3.7%), and higher than that for men (2.9%).Reference McManus, Bebbington, Jenkins and Brugha24 Following adjustment, both the size of the effect estimate and the statistical evidence of the association were increased (adjusted odds ratio 3.91, 95% CI 1.23–12.46; P = 0.027) because of negative confounding by age and alcohol dependence score. We found no statistically significant association between perimenopause and screening positive for depression before (odds ratio 1.69, 95% CI 0.73–3.92; P = 0.222) or after adjustment (adjusted odds ratio 2.17, 95% CI 0.89–5.28; P = 0.089). When combining peri- and post-menopausal groups, evidence of an association with possible depression persisted (before adjustment: odds ratio 1.98, 95% CI 1.00–3.90, P = 0.049; after adjustment: odds ratio 2.49, 95% CI 1.11–5.59, P = 0.027).
Association between menopause stage and GAD
We found no statistical evidence of an association between menopause and screening positive for GAD, either before adjustment (perimenopausal group: odds ratio 1.33, 95% CI 0.78–2.27, P = 0.299; post-menopausal group: odds ratio 1.14, 95% CI 0.66–1.96, P = 0.640) or after adjustment (perimenopausal group: adjusted odds ratio 1.57, 95% CI 0.81–3.04, P = 0.182; post-menopausal group: adjusted odds ratio 1.43, 95% CI 0.64–3.21, P = 0.383) (Supplementary Table 1). There was also no evidence of association when menopausal groups were combined (before adjustment: odds ratio 1.24, 95% CI 0.78–1.96, P = 0.360; after adjustment: odds ratio 1.54, 95% CI 0.80–2.95, P = 0.195).
Association between menopause and CIS–R scores
We found no statistical evidence for an association between total CIS–R score based on non-menopausal symptoms and menopause before adjustment (perimenopausal group: mean increase in score 0.46, 95% CI −0.28 to 1.19, P = 0.223; post-menopausal group: mean increase in score 0.55, 95% CI −0.15 to 1.25, P = 0.126) or after adjustment (perimenopausal group: mean increase in score 0.52, 95% CI −0.27 to 1.32, P = 0.197; post-menopausal group: mean increase in score 0.76, 95% CI −0.22 to 1.74, P = 0.129) (Supplementary Table 2). There was no evidence of association between CIS–R score and menopause when menopausal groups were combined (before adjustment: odds ratio 0.50, 95% CI –0.09 to 1.08, P = 0.094; after adjustment: odds ratio 0.59, 95% CI −0.17 to 1.34, P = 0.129).
Discussion
Main findings
Our results showed that women who are post-menopausal are more likely to screen positive for possible depression than women who are pre-menopausal. However, we found no significant association between perimenopause and possible depression, or between menopause and GAD, or total CIS–R score after accounting for features of menopause.
Strengths and limitations
Strengths of this study included the use of a nationally representative sample and computer-assisted self-interviewing, which may have encouraged participants to be more honest in reporting sensitive topics.
A key limitation is cross-sectional design, which means that the temporal relationship between variables cannot be elucidated. It is possible that stress associated with depression and anxiety might influence the timing of menopause onset. The numbers who screened positive for GAD and depression were small, which reduced the power to detect clinically relevant associations and precluded further subgroup analysis (e.g. by ethnicity, younger age group or severity of symptoms). Results from the CIS–R were not validated by clinical interview, and this, alongside the inclusion of inherent features of menopause in the outcome measure for some of our analyses, could have led to false positives, or conversely false negatives if symptoms of depression were dismissed as symptoms of menopause. Menopause stage was also determined only by self-report, which could have led to some misclassification. We were unable to adjust for use of hormone replacement therapy, age at menopause onset or surgical menopause. Further research could make use of electronic health records to examine such objective measures, as well as mental health diagnoses and prescriptions.
Interpretation
There is some evidence that rates of depression improve several years post-menopause, making our finding that post-menopause was the only period of elevated risk surprising.Reference Bromberger and Epperson22 One possible reason that we did not find an association between menopause stage and GAD is that menopausal anxiety might differ from diagnostic criteria for GAD, and may have an irregular temporal pattern or unpredictable onset, meaning that it could be hard to detect in cross-sectional studies.Reference Vesco, Haney, Humphrey, Fu and Nelson25 Another possibility, given that we found no association with CIS–R score after removing somatic features of menopause, is that the apparent rise in common mental disorders in this phase is driven by inherent features of menopause, such as night sweats. Future longitudinal research could focus on to what extent this is true and how the association with depression might be modified. Healthcare professionals should be aware of the increased risk of depressive symptoms after menopause.
Supplementary material
Supplementary material is available online at https://doi.org/10.1192/bjo.2023.82
Data availability
The data that support the findings of this study are available from NHS Digital. Restrictions apply to the availability of these data, which were used under licence for this study.
Author contributions
N.S. and R.R. suggested carrying out this study. A.A. planned the analyses with N.S. and R.R., carried out initial analyses and drafted the manuscript. S.M. provided feedback on the draft and suggested amendments, having originally proposed inclusion of menopause questions in the Adult Psychiatric Morbidity Survey. All authors edited the manuscript.
Funding
The authors did not receive any external funding to conduct this research. S.M. is supported by the UK Prevention Research Partnership (number MR-VO49879/1).
Declaration of interest
N.S. is a trainee editor with BJPsych Open, but was not involved in the peer-review process or decision-making process regarding publication.
Open access
eLetters
No eLetters have been published for this article.