Association of lifespan reproductive duration with depression in Swedish twins: The role of hormone replacement therapy

To examine the association between reproductive duration and postmenopausal depression (taking the use of hormone replacement therapy [HRT] into account).

Across the lifespan, women-especially at several reproductionrelated life-cycle points (such as perimenstrual changes, pregnancy, the postpartum period, and menopause)-have a higher risk of depression than men. 3 Increased vulnerability for depression in women begins with puberty and declines after menopause. 4 Although the rate of new-onset depression tends to decline after menopause, postmenopausal women are still at high risk of developing depression, even if they never experienced depression before menopause. 3,5 Across a woman's lifespan, sex steroids, especially estrogen, have been proposed to play a role in mood disorders. 4,6 Estrogen is involved in hypothalamic-pituitary-adrenal axis regulation and mood regulation. 4 In addition, fluctuations in estrogen levels have been shown to alter the brain's activation patterns, enhance the sensing of negative events, and to increase depression symptoms. 4 Lifespan reproductive duration-which reflects the duration of a woman's exposure to endogenous estrogens-has an impact on the development of postmenopausal depression. 7,8 To date, only a few studies have addressed the association between lifespan reproductive duration and depression among postmenopausal women, with inconsistent results. [9][10][11][12] Existing evidence has suggested that depression 13,14 and the onset of menarche and menopause 15,16 are influenced by genetics and early-life environment. However, it is unclear whether these factors play a role in the association between reproductive duration and depression. Moreover, hormone replacement therapy (HRT) has been proposed for the treatment of mood disorders during menopausal transition caused by estrogen withdrawal, 17 but it is unknown whether HRT affects the association between reproductive duration and postmenopausal depression.
In the present study, we aimed to (1) examine the association between reproductive duration and postmenopausal depression, (2) explore whether familial background (i.e. genetic and early-life environment factors) play a role in this association, and (3) assess the joint effect of HRT use and reproductive duration on depression using a long-term nationwide cohort of Swedish twins.

| Study population
This is a large-scale, nationwide, population-based cohort study. Out of 13 305 females with data on age of menarche or menopause, we excluded 1422 participants who underwent hysterectomy or ovariectomy before natural menopause, 214 who had outliers or extreme values for age at menarche (<9 years or >19 years) or menopause (<39 years or >60 years), and 349 who had depression before the screening. Finally, 11 320 postmenopausal women remained for the current analysis ( Figure S1).

| Assessment of reproductive duration
Ages at menarche and menopause were self-reported in the SALT telephone interview. Reproductive duration was calculated by subtracting age at menarche from age at menopause. In the current study, reproductive duration was categorized into quartiles (≤34, 35-36, 37-39, and ≥40 years). A multi-adjusted restricted cubic spline regression model showed a U-shaped trend instead of a monotonic trend for the risk of depression with increasing reproductive duration, and the inflection point was 36.75 ( Figure 1). Therefore, we combined the 35-36 years and 37-39 years reproductive duration groups as the reference group, and reproductive duration was categorized into three groups: short (≤34 years), average (35-39 years), and long (≥40 years) in the analysis.

| Characteristics of the study population
Of 11 320 postmenopausal women, the average age at recruitment was 63.57 ± 9.38 years (range 44-97 years), and the mean reproductive duration was 36.28 ± 4.04 years (range F I G U R E 1 Multivariable adjusted odds ratios (ORs) of depression in relation to reproductive duration, results from restricted cubic spline regression model. Adjusted for age, education, marital status, smoking status, alcohol consumption, physical activity, body mass index, parity, oral contraceptives, hormone replace treatment, diabetes, hypertension, cardiovascular disease, and cancer. CL, confidence limits.
Compared with women with an average reproductive duration, those with a short reproductive duration were more likely to be older, single, smokers, have fewer years of education, have lower body mass index, and report never giving birth or using HRT.
Women with a long reproductive duration were more likely to be married/cohabiting, non-smokers, have more years of education, have higher body mass index, and have a history of giving birth and using HRT. In addition, women with an average reproductive duration were more likely to use OCs and less likely to have cancer than the other groups (Table 1). There were no significant differences between the groups with respect to zygosity, alcohol consumption, physical activity, and history of diabetes, hypertension, or CVD.

| Association between reproductive duration and postmenopausal depression in unmatched case-control analysis
During the follow up, a total of 593 (5.24%) participants developed depression. In the basic-and multi-adjusted GEE models, women with a short or long reproductive duration had a higher risk of depression after menopause, compared with those with an average reproductive duration. The risk of depression was 20% higher for individuals with a short or long reproductive duration (OR 1.28, 95% CI 1.05-1.55; OR 1.25, 95% CI 1.01-1.55, respectively). Therefore, the short and long reproductive duration groups were combined into one group (i.e. non-typical reproductive duration) in the following analysis. Compared with those with an average reproductive duration, the OR for the association between nontypical reproductive duration and depression was 1.27 (95% CI 1.07-1.50) ( Table 2).
In stratified analyses, the OR of non-typical reproductive duration (versus average reproductive duration) was 1.29 (95% CI 1.07-1.55) among those who were CMD-free, but the results were not significant among those with CMDs (OR 1.15, 95% CI 0.77-1.72) (Table S2). However, there was no significant interaction between CMDs and reproductive duration on the risk of depression (P = 0.225).

| Association between reproductive duration and postmenopausal depression in co-twin matched case-control analysis
In multi-adjusted co-twin matched case-control analysis (n = 165 depression-discordant twin pairs), we observed no significant association between reproductive duration and depression (OR 0.85, 95% CI 0.51-1.44) ( Table 3).
There was no significant difference in ORs from the GEE model in unmatched case-control analysis and the conditional logistic model in co-twin control analysis (OR 1.19, 95% CI 0.87-1.63, P = 0.270).
This suggests that familial background is unlikely to play a role in the association between reproductive duration and postmenopausal depression.

| Joint effect of reproductive duration and HRT on postmenopausal depression risk
In the GEE models, women with a history of HRT use were at a higher risk of postmenopausal depression (OR 1.40, 95% CI 1.16-1.67) (Table S3). In joint effect analysis, the multi-adjusted ORs of depression were 1.19 (95% CI 0.91-1.56) for women with an average reproductive duration and a history of HRT use, 1.15 (95% CI 0.94-1.41) for those with a non-typical reproductive duration and no HRT use, and 1.82 (95% CI 1.42-2.33) for those with a non-typical reproductive duration and a history of HRT use, compared with those with an average reproductive duration and no HRT use (Table 4).
Furthermore, among women with a non-typical reproductive duration, a history of HRT use still related to the higher risk of depression by approximately 60% (OR 1.58, 95% CI 1.24-2.01) (Table S4).

| Supplementary analysis
The results were not much altered when we repeated the analysis after (1) restricting the population to only participants with no missing data on covariates (n = 7867, 69.50% of the full sample) (Table S5); and (2) additionally adjusting for baseline depression symptoms (Table S6).

| DISCUSS ION
In this large-scale, nationwide, population-based Swedish twin study, we found that women with comparatively short (≤34 years) or long (≥40 years) reproductive durations had an elevated risk of developing depression after menopause. Familial background may not account for the observed association. History of HRT use may presage a higher risk of depression in relation to reproductive duration.
Few studies have investigated the associations between reproductive characteristics and depression, particularly among postmenopausal females. Three observational studies in eastern Asia [9][10][11] have suggested that the risk of depression increases significantly with decreasing reproductive duration. However, a cohort study of 1013 French women indicated no significant association between age of menarche or reproductive duration and depression. 12 In the present study, the risk of depression was higher among women with a reproductive duration ≤34 or ≥40 years, compared with those with a reproductive duration of 35-39 years. The results suggest that there is a U-shaped, rather than a monotonic, relationship between depression risk and reproductive duration.
It is noteworthy that the women in this study had a longer mean reproductive duration than those in the aforementioned studies.  It has been proposed that HRT may relieve depressive symptoms among perimenopausal women, but due to inconsistent evidence, HRT has not been approved for this indication to date. 21,22 Several studies have failed to find a significant improvement in moods among current HRT users, [23][24][25] and women who were past HRT users have a higher risk of depression in late life. 9,12 In this study, we found that HRT was associated with higher risk of depression, and furthermore, there was a synergistic effect between non-typical reproductive duration and HRT use on depression risk among postmenopausal women. Those with a history of HRT use were more likely to experience negative events such as premature ovarian failure and severe hormonal withdrawal symptoms. 26 Despite the results found above, we would rather interpret the history of HRT as a reminder of a high-risk population but not a risk factor for postmenopausal depression.
The mechanisms involved in the association between reproductive duration and depression are incompletely understood, but may involve two distinct aspects: lifespan estrogen exposure and estrogen fluctuation. Estrogen receptors are widely distributed in the brain and closely associated with cognition, memory, neurodevelopment, and neuroplasticity function of the brain. 6 On the one hand, a short reproductive duration implies a shortened period of estrogen exposure. This could be detrimental given estrogen's role in modulating the neurotransmitter system functions (including the dopaminergic, serotonergic, and glutamatergic systems) and exerting neurotropic effects to promote neuronal plasticity. 4,6 On the other hand, longer reproductive periods may entail prolonged exposure to hormonal fluctuations, which may also increase vulnerability to depression. 27 To date, lots of the existing literature supports a role for sex steroid fluctuations in affective symptoms. 28 Sex steroid fluctuations may cause an imbalance in functional activity in the dorsal and ventral divisions of the prefrontal cortex and enhance the perception of negative events. 4 Furthermore, they can also play a role in the development of somatic symptoms, which in turn can lead to an intensification of mood disorders. 3 Therefore, we speculate that when reproductive duration extends beyond a certain duration (36.75 years in our study), the benefits of long-term estrogen exposure for mental health may be counteracted by the detrimental impact of hormone fluctuations. However, there may also be a residual confounding caused by the use of OCs or premenopausal HRT. As we know, OCs and premenopausal HRT can extend reproductive lifespan 29 and are associated with an increased risk of depression. 12,30 More evidence is needed to determine whether the increased risk of depression with longer reproductive duration is due to hormone fluctuations.
This large, nationwide, population-based twin cohort study provided us with a unique opportunity to explore the association between reproductive duration and depression while also controlling for genetic background and early-life environmental factors.
Nonetheless, some limitations of our study need to be pointed out.
First, depression cases in the SALT study were identified through medical records, and could therefore leading to an underestimation of depression cases if some people did not seek medical help for their depression. Second, participants' reproductive history was assessed TA B L E 2 Odds ratios with 95% confidence intervals of depression in relation to reproductive duration, results from generalized estimating equation models.

TA B L E 3
Odds ratios and 95% confidence intervals for the association between reproductive duration and depression in co-twin control analysis using depression-discordant twin pairs (n = 165 pairs), results from conditional logistic regression models. proposed that the duration of HRT use is not associated with depression risk. 31 Therefore, the lack of data on onset time and duration of HRT use may not have had a major impact on our results, but caution is still needed when interpreting our results. Fifth, our analysis did not reflect the changes in the impact of reproductive duration on depression over time. Finally, we were unable to separately evaluate monozygotic twins (who share 100% of their genetic background) and dizygotic twins (who share only 50% of their genetic background) to estimate the effect of genetic background due to the limited sample size.
In conclusion, our study provides evidence that both short 2017-00641.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors report no conflict of interest. Funding sources had no role in study design, data collection, data analysis, data interpretation, and writing of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.

TA B L E 4
Additive interaction between hormone replacement therapy and reproductive duration for the risk of depression.