Risk of Psychiatric Disorders Following Symptomatic Menopausal Transition

INTRODUCTION

Menopausal transition is a period late in a woman's reproductive life before the final menstrual period. Menopause is a physiological phase defined retrospectively by 12 months of amenorrhea and loss of ovarian follicular function, typically occurring in most women between the ages of 40 and 55 years. The mean duration of menopausal transition is 5 years, but symptoms during this period may begin 8 years or more before the final menstrual period.¹

Although the menopausal transition is not a disease and does not require diagnosis or treatment unless symptoms are present, for clinicians, early awareness of symptoms of menopausal transition is essential for providing appropriate counseling or treatment. In general, symptomatic menopausal transition is a clinical diagnosis. Accordingly, laboratory testing, including determining follicle-stimulating hormone and estradiol levels, is not necessary for establishing the diagnosis, and skilled clinicians can diagnose symptomatic menopausal transition based on the characteristic symptoms such as abnormal uterine bleeding, hot flashes, and night sweats, which occur as a result of declining ovarian function.² It is clear that the existence and severity of symptoms during the menopausal transition are strongly related to the woman's quality of life.³

Regardless of ethnic origin, skin color, or socio-demographic factors, women around the world suffer not only from characteristic somatic symptoms of the menopausal period,⁴ such as hot flushes, night sweats, lowered metabolism, and sexual problems, but also from psychiatric symptoms, such as changes in mood, anxiety, and sleep disturbance.⁵ However, in clinical practice, the diagnosis of symptomatic menopausal transition is based more on the somatic rather than psychiatric symptoms.^6,7 Nevertheless, experts in symptomatic menstrual transition recognize that psychiatric problems are a core issue among affected women. However, most studies focusing on the evaluation of occurrence or severity of psychiatric symptoms during menstrual transition employ rating scales or other self-reporting methods^8–13 rather than clinical diagnostic interviews to identify psychiatric disorders. Oftentimes, the self-reported depression, anxiety, and sleep disturbance can be tolerated by the patients or treated by their primary care physicians or obstetrician-gynecologists. Accordingly, the knowledge on the incidence of clinical depression, anxiety, and sleep disorders in women with symptomatic menopausal transition is currently very limited.

Several studies dedicated to menopausal transition in women with schizophrenia^14,15 and bipolar disorder^16–18 investigated changes in psychotic symptoms and mood course fluctuations. In addition, about 3 decades ago, Kukopulos et al¹⁹ and Oppenheim²⁰ provided evidence that treatment of unipolar depression during menstrual transition with antidepressants or augmentation with estrogen may increase the risk of a bipolar switch, especially in those with features of rapid cycling. However, nationwide clinical data regarding the association between menopausal transition and risk of developing subsequent newly diagnosed schizophrenia and bipolar disorders (especially in those without past history of depression) are limited.

In this study, we focused on psychiatric symptoms severe enough to require evaluation by a specialist and warrant a diagnosis of a psychiatric disorder. In particular, we designed a retrospective cohort study using a nationwide claims database to investigate the possible link between symptomatic menopausal transition and psychiatric disorders.

PATIENTS AND METHODS

RESULTS

Figure 1 displays the flowchart of enrollment. Our study sample comprised 19,028 symptomatic menopausal transition women and 19,028 control women. The basic characteristics of the 2 cohorts are shown in Table 1. The median age was 51.3 years (interquartile range, 46.4–56.9 years). Compared with the control cohort, the symptomatic menopausal transition cohort had higher percentages of comorbidities, significantly higher income, and higher degree of urbanization. During the entire follow-up period, 2164 (11.4%) symptomatic menopausal transition women and 1069 (5.6%) control women were diagnosed with psychiatric disorders (P < 0.001). The 3 most frequent types of psychiatric disorders detected in both cohorts were depressive disorders (4.6% vs 2.2%, P < 0.001), anxiety disorders (3.6% vs 1.7%, P < 0.001), and sleep disorders (2.8% vs 1.4%, P < 0.001).

As shown in Table 2, within the entire follow-up period, significantly higher IRRs of bipolar disorders (IRR = 1.71, 95% confidence interval [CI] = 1.01–2.96), depressive disorders (IRR = 2.15, 95% CI = 1.91–2.42), anxiety disorders (IRR = 2.07, 95% CI = 1.81–2.37), and sleep disorders (IRR = 2.01, 95% CI = 1.73–2.34) were observed in the symptomatic menopausal transition cohort than in the control cohort. Stratification according to follow-up durations revealed no significant differences in the incidence of schizophrenia in any group. However, IRRs of newly diagnosed depressive disorders, anxiety disorders, and sleep disorders remained significantly increased in all the follow-up duration groups (0–1, 1–5, and ≥5 year), and that of bipolar disorders increased only in the groups with a follow-up longer than 1 year (1–5 and ≥5 year).

Table 3 compares crude hazard ratios (HRs) and adjusted hazard ratios (aHRs) of the newly diagnosed psychiatric disorders for the symptomatic menopausal transition women and controls. The crude HRs obtained with the univariate analysis revealed that patients with symptomatic menopausal transition may have an increased risk of developing a subsequent bipolar disorder, depressive disorder, anxiety disorder, and sleep disorder. The results of the multivariate analysis indicated that the women with symptomatic menopausal transition had a markedly higher risk of a subsequent depressive disorder (aHR = 2.17, 95% CI = 1.93–2.45), anxiety disorder (aHR = 2.11, 95% CI = 1.84–2.41), and sleep disorder (aHR = 2.01, 95% CI = 1.73–2.34). A small but significant increase in the risk of a subsequent bipolar disorder was also observed (aHR = 1.68, 95% CI = 1.01–2.79).

DISCUSSION

This is the 1st nationwide retrospective cohort study to examine the association between symptomatic menopausal transition and the risk of subsequent new-onset psychiatric disorders. The major finding of our study is that women with symptomatic menopausal transition may have a higher risk of development of subsequent psychiatric disorders, especially depressive disorder, anxiety disorder, sleep disorder, and bipolar disorder.

Several studies have demonstrated that depression and anxiety are the most common psychiatric symptoms in women with symptomatic menopausal transition. Our study indicated that symptomatic menopausal transition might be a risk factor for subsequent clinical depressive and anxiety disorders. There may be several explanations for this link. First, the withdrawal of hormones may cause a change in the serotonin levels. Estrogens are highly lipophilic hormones that can cross the blood–brain barrier and bind to intracellular estrogen receptors distributed in many organs related to anxiety and depressive disorder, such as the hippocampus.²³ In 2009, Walf et al²⁴ found that chronic estradiol replacement in aged female rats reduced anxiety and depressive-like behaviors, suggesting an important role of estradiol in hippocampus-mediated mood regulation. Second, the role of estrogens in the development of anxiety and depressive disorders during menopausal transition may be associated with their neurotrophic and neuroprotective effects. Thus, administration of estradiol to ovariectomized rats induced an increase in the release of brain-derived neurotrophic factor and vascular endothelial growth factor, decreasing depressive-like behavior.^25,26 Third, VMS, sexual dysfunction, and obesity,^2,27 which are common symptoms in women with symptomatic menopausal transition, have been linked to the development of depressive and anxiety disorders. Last, although most arguments are in favor of a primarily biological rationale, several psychosocial stressors common in mid-life women, such as marital issues, changes in caretaking (children departing home, aging parents) or career responsibilities, issues related to aging, and limited social network have also been linked to depressive and anxiety disorders in women during the menopausal transition.^17,28,29 However, the complexity of related social, cultural, psychological, and environmental factors^30,31 has resulted in poor understanding of the role of these common psychosocial stressors in the pathogenesis of depression and anxiety in symptomatic menstrual transition.

Our results also indicate that women with symptomatic menopausal transition are at an increased risk of new-onset sleep disorders. Although the hormone changes during menopausal transition are often linked to the development of sleep disturbance, no studies have confirmed the causal relationship between the changes of serum hormone concentrations and the occurrence of sleep disturbance in women during the perimenopausal period. A recent study trying to explore the linkage between the sleep quality and the fluctuation of hormone levels in women revealed that there was no clear relationship between hormone changes and the sleep quality among women during menopausal transition, which implicated that the sleep disturbance during this period may be influenced by other factors, such as emotional disturbance.³² Another study suggested that perimenopausal women with sleep disorders often have VMS and depressive symptoms, which are associated with different patterns of the sleep disorder. For example, depression is uniquely related to difficulty falling asleep and waking up earlier than desired, whereas VMS are related to frequent awakenings during sleep.³³

Previous studies reported that progression through the female reproductive stages (premenopause, perimenopause, and postmenopause) is associated with mood course in women with bipolar disorder¹⁶ and indicated that such women are at a higher risk of unstable mood episodes during periods of intense hormonal fluctuation such as the menopausal transition. In addition, as long as 3 decades ago, Kukopulos et al¹⁹ demonstrated that menopausal transition could induce rapid cycling, whereas Oppenheim²⁰ showed that estrogen induced abruptly rapid mood cycling. These results may further emphasize the important associations between the hormonal changes and bipolar disorders, especially rapid cycling. Moreover, estrogen and progesterone can modulate neurotransmitter systems and intracellular signaling pathways known to be affected by mood-stabilizing agents. In particular, they closely interact with brain-derived neurotrophic factor, oxidative stress, and inflammation pathways, which are involved in the neurobiology of bipolar disorders.³⁴ Although our results indicate that symptomatic menstrual transition might increase the risk of a subsequent newly diagnosed bipolar disorder, no firm conclusions should be drawn at this time given the evidence of high rates of misdiagnosis and delayed treatment of bipolar disorders, especially of type II,^35,36 and lack of consistent supporting evidence.

The results of stratification for the duration of follow-up period indicated that the incidences of bipolar disorders, depressive disorders, anxiety disorders, and sleep disorders were increased in the 1st year after the diagnosis of symptomatic menopausal transition. One possible explanation for this result is surveillance bias. Women with symptomatic menopausal transition are likely to have more frequent outpatient visits than the general population, leading to earlier diagnosis of psychiatric disorders. However, as shown in Table 2, the incidences of newly diagnosed bipolar disorders, depressive disorders, anxiety disorders, and sleep disorders remained significantly increased after the 1st year of follow-up, suggesting that the observed effects were not due only to surveillance bias. The present study has several limitations related to the use of a claims database. First, although symptomatic menopausal transition is a clinical diagnosis, data on menstrual cycle as well as results of baseline and longitudinal endocrinological assessments, which could increase the validity of the diagnosis, were unavailable. Second, data on severity of symptomatic menopausal transition and use of hormone replacement therapy were limited. The effects of severity of symptomatic menopausal transition on development of subsequent psychiatric disorders require further investigation. Third, subtypes of psychiatric diagnoses were not specified (eg, bipolar I vs bipolar II, types of depressive disorders, panic vs other anxiety disorders, types of sleep disorders, eg, insomnia vs hypersomnia, delayed or advanced sleep disorder, or restless leg syndrome). Fourth, there was no access to data on personal or family history of psychiatric disorders, and seasonal influences. Fifth, the absence of records on several demographic variables such as domestic environment, lifestyle, and psychosocial stressors represents an additional limitation. Last, the prevalence of psychiatric disorders in our study may be underestimated because only subjects who sought medical evaluation were included. Thus, the risk for a new onset³⁷ or recurrence³⁸ of depressive disorders during the menopausal transition in women with a history of depression has been well documented, and such women may be more likely to attend a menopause clinic.³⁹ The higher possibility of early diagnosis in the study cohort compared to the control cohort was not fully eliminated by the subgroup analysis according to the follow-up duration between the diagnosis of symptomatic menstrual transition and depressive disorder.

In conclusion, symptomatic menopausal transition may be a risk factor for bipolar, depressive, anxiety, and sleep disorders. Although there are several limitations to the claims database study, these limitations are most likely to have a greater influence on the investigated temporal relationship between symptomatic menopausal transition and subsequent psychiatric disorders within the 1st year follow-up. Therefore, based on the results of persistently higher IRRs over longer follow-up periods (1–5 and ≥5 year), symptomatic menopausal transition may play a potential role in new-onset, but not recrudescence of, bipolar, depressive, anxiety, and sleep disorders. However, prospective studies, particularly with additional patient-level data, are warranted to confirm our findings.