Primary ovarian insufficiency (POI) is associated with increased risk for depression, episodes of which frequently occur after the onset of menstrual irregularity and before the POI diagnosis.
Abstract
Context:
A high prevalence of depressive symptoms is observed in women with primary ovarian insufficiency (POI) compared with women in whom the menopause is normally timed. Indeed, studies suggest that depression and/or its pharmacological treatment contribute to the onset of POI.
Objectives:
We characterize the prevalence of psychiatric disorders and the timing of onset of clinically significant depression relative to both the diagnosis of POI and the onset of menstrual irregularity in women with POI.
Design and Setting:
We conducted a cross-sectional clinic-based study at the National Institutes of Health Clinical Research Center.
Patients:
A total of 174 women with spontaneous 46, XX POI and 100 women with Turner syndrome participated in the study.
Main Outcome Measures:
The structured clinical interview for DSM-IV was performed.
Results:
Lifetime histories of depression in POI exceeded rates of depression reported in women with Turner syndrome and community-based samples of women (P < 0.001). The onset of depression frequently preceded the diagnosis of POI but occurred after the onset of menstrual irregularity. Analyses standardizing the periods of risk for depression showed that similar numbers of depressions occurred before and after these events.
Conclusions:
POI is associated with an increased lifetime risk for major depression. Attention to the presence of depression in POI should become an important part of the care for these women. The onset of depression frequently occurs after signs of altered ovarian function but before the diagnosis of POI. Thus, in some women the association between POI and depression suggests an overlapping pathophysiology rather than a causal relationship.
Primary ovarian insufficiency (POI) is defined by 4 or more months of oligoamenorrhea in association with plasma gonadotropin levels in the menopausal range in women before the age of 40; early menopause is defined by these features in women between the ages of 40 and 45 (1–6). There is a higher frequency of depressive-like symptoms accompanying both POI and early menopause compared with the normally timed menopause (7–11); however, in only one study was a clinically significant depressive syndrome distinguished from depressive symptoms that can be nonspecific and multidetermined (7). Specifically, Harlow et al. (7) defined depression by a self-report of receiving both a medically diagnosed depression and antidepressant medication. Interestingly, in the study by Harlow (7), because the onset of depression frequently preceded the diagnosis of POI, it also was suggested that depression and/or its treatment could contribute to the onset of POI. Depression is both underdiagnosed and undertreated. In fact, recent epidemiological studies that have employed structured psychiatric diagnostic interviews to determine the occurrence and time of onset of depression document that in the United States over 40% of major depressions do not receive any treatment (12). Additionally, these studies observed that there is a gap of 6–8 yr between the onset of the first major depression and the receipt of medical treatment (13). Thus, unless a structured diagnostic interview is employed, estimates of the prevalence and timing of depressions in women with POI based on medical attention and treatment could be inaccurate.
To date, no study has employed standardized criteria for POI and structured diagnostic interviews to determine the lifetime history of clinically significant depression in women with POI. We examined the prevalence of DSM-IV Axis I psychiatric disorders [defined as clinical disorders, including major mental disorders (14) and determined by structured diagnostic interviews] in a sample of well-characterized women with spontaneous 46, XX POI. Prevalence of psychiatric disorders in POI were compared with those reported in community- and clinic-based samples. Additionally, we investigated how frequently depression occurred prior or subsequent to the diagnosis of POI and, in those depressions that occurred before the diagnosis of POI, whether depression preceded the onset of menstrual cycle irregularity (MCI) (a proxy for the onset of ovarian insufficiency). Depressions occurring before the onset of MCI could contribute to or cause ovarian insufficiency [e.g. through proinflammatory or stress-related mechanisms (15–18)], whereas depressions occurring after the onset of MCI (and early ovarian insufficiency) are not likely to be a direct cause of ovarian insufficiency.
Subjects and Methods
Subject selection
Women with POI participated in a National Institute of Child Health and Human Development (NICHD) protocol that included an extensive medical evaluation. Participants were recruited primarily through notices on the National Institutes of Health (NIH) home page. The diagnosis of POI was based on a history of at least 4 months of noniatrogenic oligoamenorrhea occurring in women no more than 40 yr of age, at least two determinations of menopausal serum FSH levels (defined by the local laboratory), and a normal, 46, XX peripheral karyotype. All women were otherwise in good general health based on laboratory and clinical evaluations. Women with POI provided written informed consent for participation in this Institutional Review Board-approved protocol. Psychiatric evaluation was one component of the 3-d screening admission in the NIH Clinical Research Center. After a 3-month assessment period, women with POI were offered participation in a trial investigating the efficacy of testosterone augmentation of estrogen-progestin therapy on bone density.
Procedures
A total of 174 women with POI were evaluated with the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders (4th ed. DSM-IV) (SCID) (19), a structured interview that establishes the presence or absence and the date of onset of a DSM-IV Axis I diagnosis. The date on which a woman was diagnosed by her physician to have POI was obtained from medical records and confirmed by NICHD medical staff. Additionally, the date of onset of MCI was determined by patient self-report and defined as the onset of irregular menstrual cycle length (<21 or >35 d) or skipped periods relative to their established normal pattern of menses.
Operational definitions and data analysis
The percentages of women with POI who met SCID criteria for current and lifetime diagnoses of major depression, current and lifetime diagnoses of any mood disorder (i.e. combined prevalences of major and minor depression and dysthymia but not anxiety disorders), and current and lifetime diagnoses of an anxiety disorder were compared with published prevalences of these disorders from several large community-based studies (20–22) as well as published prevalences from primary care and gynecology clinic-based studies (23, 24). Comparisons were performed with one-sample exact binomial calculations using the community or clinic-based prevalences as the standard. Finally, we employed the results from a group of adult women with Turner syndrome (TS) who were studied previously with the same structured diagnostic interview (25) to compare the prevalences of Axis I conditions in women with POI to a group of women with early onset ovarian failure and hypogonadism. We used χ2 analyses to compare the lifetime prevalence of Axis I disorders in women with POI with those reported previously in this group of 100 adult women [age, 34.7 + 11.6 yr (mean ± sd)] with TS (25).
Timing of onset of depression
Women were separated into two groups based on the reported timing of the onset of their first major or minor depression compared with the date when the diagnosis of POI was established. We also separated women into two groups based on the timing of the onset of the first episodes of major and minor depression relative to each woman's report of the date of onset of MCI. The onset of MCI was used as a proxy for the first clinical signs of developing POI in those women who had previously established regular menstrual cycle function. Four women with past major depressive episodes reported primary amenorrhea and, therefore, were excluded from the analysis using MCI onset. The frequency distributions for these groups were calculated and compared to evaluate two questions: 1) whether a difference existed in the number of depressive episodes that occurred before compared with after the diagnosis of POI; and 2) whether a difference existed in the number of depressive episodes that occurred before compared with after the reported onset of MCI. The number of episodes of major depression were calculated separately from episodes of minor depression. Frequencies were compared initially using an exact binomial calculation. Due to the retrospective nature of this study, the years at risk before and after the diagnosis of POI differed in these women. The duration of time between the onset of menarche and the diagnosis of POI was on average 16.4 yr, whereas the time between the diagnosis of POI and the date when the SCID was administered was 3.1 yr. We therefore performed additional comparisons as follows: 1) Poisson rates of depression were calculated (i.e. number of women with depression per person-years in the time period) to compare the number of the depressions that occurred in women before and after the diagnosis of POI. Three sets of Poisson rates were calculated for major as well as for minor depressions that were limited to depressions occurring within 3-, 5-, and 10-yr windows surrounding the diagnosis of POI. In these calculations, women whose depression did not occur during the windows of time before or after the diagnosis of POI (i.e. 3, 5, and 10 yr, respectively) contributed person-years to the denominator. For each set, the before and after rates were compared using an exact Poisson test of homogeneity. 2) Kaplan-Meier survival analyses were performed to compare the onset of depressive episodes in the absolute number of days occurring before and after the diagnosis of POI. 3) We compared the numbers of women who met criteria for major depression in the 3 yr before and the 3 yr after the diagnosis of POI in a subsample of 58 women who were administered SCID interviews 3 yr or more after the diagnosis of POI.
Differences in the years of exposure to risk during the period before and after the onset of MCI also existed, although to a lesser extent than those surrounding the onset of POI. The average duration of time between the ages of menarche and MCI was 11.6 yr, whereas the duration of time between the onset of MCI and the age at which the SCID interview was performed was an average of 7.4 yr. Thus, we also calculated both Poisson rates and Kaplan-Meier survival analyses to compare the number of episodes of depression that occurred before and those occurring after the reported onset of MCI.
Differences in the following demographic variables in women with POI with and without a lifetime history of depression were examined by ANOVA or Fisher exact tests: age, marital status, height, body mass index (BMI), years of education, age at menarche, age at POI diagnosis, age at MCI onset, duration of MCI before the diagnosis of POI, smoking history, past or current antidepressant use, and thyroid or ovarian hormone replacement therapies.
Results
Subject characteristics (Tables 1 and 2)
Table 1.
Characteristics of women with spontaneous 46, XX POI (n = 174)
| Age (yr) | 31.6 (5.3) |
| No. married | 122 (70%) |
| No. of women with prior pregnancies | 54 (31.0%) |
| No. of women with children | 32 (18.4%) |
| Racial/ethnic composition, no. of women | |
| Asian | 8 (4.6%) |
| African-American | 16 (9.2%) |
| Caucasian | 140 (80.5%) |
| Hispanics | 10 (5.7%) |
| Height (cm) | 164.7 (7.4) |
| BMI (kg/m2) | 23.7 (3.4) |
| Years of education | 16.3 (2.1) |
| Age at menarche (yr)a | 12.7 (1.5) |
| Age at POI diagnosis (yr) | 28.6 (6.6) |
| Age at MCI onset (yr) | 24.4 (7.9) |
| Duration of MCI before POI diagnosis (yr) | 4.5 (5.4) |
| No. of smokers | 10 (5.7%) |
| No. ever treated with antidepressants | 48 (28%) |
| No. on thyroid hormone treatment | 17 (9.7%) |
| No. on hormone replacement therapy | 132 (76%) |
Reported values are expressed as mean (sd) unless otherwise indicated. To be eligible for the study, women had to meet the following inclusion criteria: 1) diagnosis of spontaneous 46, XX POI before the age of 40 yr (i.e. at least 4 months of oligomenorrhea or amenorrhea and FSH levels in the menopausal range, as defined by the local laboratory, confirmed on two separate occasions, at least 1 month apart); 2) current age between 18 and 42; and 3) no iatrogenic cause of POI or known chromosomal abnormality. An initial sample of 398 women with POI were admitted to the NIH Clinical Center and participated in the screening portion of the general protocol between July 2000 and February 2005. Of these, 209 did not qualify or declined to return to participate further in longer term integrated studies (evaluating many aspects of POI, such as bone health, metabolic effects, sexual function, cognition, mental health, and dry eye syndrome). Because recruitment for participation in the subsequent longer term stressed mainly the effects on bone health, we do not expect that there was selection bias either for or against mental health disorders in the women who did not continue in the study. After returning for further evaluation, four women declined participation, and 11 women did not complete the structured diagnostic interview because of time limitations arising from competing procedures during their admission. Thus, 174 women completed the study.
Seven women reported primary amenorrhea and could not be included in this calculation.
Table 2.
Frequency of Axis I psychiatric illness in 174 women with POI and 100 women with TS (25)
| POI | TS | |
|---|---|---|
| Current diagnosis (n = 39)a | ||
| Mood disorders | ||
| Major depression | 16 (9.2) | 5 |
| First onset | 4 (2.3) | |
| Minor depression | 11 (6.3) | 5 |
| First onset | 1 (0.6) | |
| Dysthymic disorder | 3 (1.7) | 1 |
| Anxiety disorders | ||
| Specific phobia | 3 (1.7) | 2 |
| Social phobia | 3 (1.7) | 2 |
| Generalized anxiety | 3 (1.7) | 2 |
| PTSD | 2 (1.1) | 0 |
| Agoraphobia without panic | 2 (1.1) | 0 |
| Panic disorder | 1 (0.5) | 4 |
| Past diagnosis (n = 113)b | ||
| Mood disorders | ||
| Major depression | 91 (52.3) | 33 |
| Minor depression | 20 (11.4) | 8 |
| Dysthymic disorder | 2 (1.1) | 0 |
| Anxiety disorders | ||
| Panic disorder | 3 (1.7) | 1 |
| PTSD | 7 (4.0) | 5 |
| Eating disorders | ||
| Anorexia nervosa | 2 (1.1) | 5 |
| Bulimia nervosa | 1 (0.5) | 1 |
| Substance abuse disorders | ||
| Alcohol/drug dependence | 3 (1.7) | 3 |
| Lifetime diagnosis (n = 121)c | ||
| Major depression | 95 (54.5) | 36 |
| Any mood disorder | 117 (67.2) | 47 |
| Any anxiety disorder | 24 (14.0) | 15 |
| Any Axis I disorder | 121 (69.5) | 52 |
Values in women with POI are expressed as number (percentage); otherwise all values are expressed as percentage. The numbers of women with POI meeting criteria for an Axis I disorder were tabulated. In the case of recurrent, episodic disorders like depression, the number of women with a lifetime history of depression was calculated using the results of the SCID as follows: the number of women with at least one past episode of depression plus the number of women with a current major depression that was lifetime first-onset. Thus, women with both a current and past episode of depression were not counted twice. Women who were taking antidepressant medication and whose depression was in remission would no longer meet criteria for current major depression; however, the original depressive episode for which they received treatment would be captured by the SCID and, therefore, included in the lifetime prevalences.
Met two diagnostic categories n = 5 (2.9%); current major depression with Dysthymic disorder n = 2; current major depression with PTSD n = 2; current GAD with panic disorder n = 1.
Met two or more diagnostic categories n = 14 (8.0%); major depression, PTSD n = 5; major depression, PTSD, and alcohol dependence n = 1; major depression, panic disorder and anorexia nervosa n = 1; major depression and panic disorder n = 2; major depression drug dependency n = 1; major depression and alcohol dependence n = 1; major depression and anorexia nervosa n = 1; major depression and bulimia nervosa n =1; major depression and minor depression n =1.
Both the lifetime prevalence of any mood disorder and major depression in women with POI (i.e. 67 and 54.5%, respectively) were significantly higher than those previously observed in a sample of adult women with TS (25) (i.e. POI vs. TS: any mood disorder, 67 vs. 47%; χ2 = 10.8; P = 0.001; major depression, 54.5 vs. 36%; χ2 = 8.8; P = 0.003).
The majority of Axis I diagnoses were mood or anxiety disorders. A lifetime history of a depressive disorder was defined as at least one past episode of depression or a current major depression that was lifetime first-onset. Thus, women with both a current and past episode of depression were not counted twice. A lifetime history of any Axis I disorder was present in 69.5% (n = 121) of women with POI compared with lifetime prevalences of 50% in community-based studies (21). Lifetime history of major depression occurred in 54.5% of women (n = 95), and a lifetime history of any mood disorder (i.e. includes combined prevalences of major and minor depression, and dysthymia but not anxiety disorders) in 67.2% (n = 117) of women; both rates were significantly higher than those observed in community samples (P < 0.001 for both comparisons) (21). A lifetime history of an anxiety disorder, observed in 14% (n = 24) of women with POI, was significantly lower (P < 0.001) than the reported lifetime prevalence of 36% in community studies (21). Finally, both the lifetime prevalence of any mood disorder and major depression in women with POI (i.e. 67 and 54.5%, respectively) were significantly higher than those previously observed in a sample of adult women with TS (i.e. POI vs. TS, any mood disorder—67 vs. 47%; χ2 = 10.8; P = 0.001; major depression—54.5 vs. 36%; χ2 = 8.8; P = 0.003).
With the exception of a greater number of women reporting the use of antidepressant medication (χ2 = 24.6; P < 0.001), women with POI who had a lifetime history of depression were not distinguished from those with no lifetime history of depression by any variable (Table 3) (P > 0.1 for all comparisons).
Table 3.
Characteristics of women with POI (n = 174) with and without lifetime history of major and minor depression
| Major depression (n = 95) | Minor depression (n = 22)a | No lifetime depression (n = 57) | |
|---|---|---|---|
| Age (yr) | 31.8 (4.9) | 32.6 (6.0) | 30.9 (5.8) |
| No. married | 63 (66%) | 17 (77%) | 42 (74%) |
| Height (cm) | 165.2 (6.8) | 164.4 (6.9) | 163.9 (8.4) |
| BMI (kg/m2) | 23.8 (3.5) | 23.3 (3.3) | 23.8 (3.4) |
| Years of education | 16.3 (2.0) | 16.6 (1.7) | 16.0 (2.4) |
| Plasma FSH (IU/liter) | 83.5 (36.0) | 87.5 (36.7) | 83.5 (35.2) |
| Age at menarche (yr) | 12.9 (1.6) | 12.7 (1.1) | 12.5 (1.3) |
| Age at POI diagnosis (yr) | 28.4 (6.6) | 30.0 (6.3) | 28.4 (6.8) |
| Age at MCI onset (yr) | 23.8 (7.9) | 26.1 (8.2) | 24.8 (7.8) |
| Duration of MCI before POI diagnosis (yr) | 5.2 (5.9) | 4.4 (5.2) | 4.2 (4.8) |
| No. of smokers | 3 (3.1%) | 1 (4.5%) | 6 (10.5%) |
| No. ever treated with antidepressants | 45 (47%)b | 1 (4.5%) | 2 (3.5%) |
| No. on thyroid treatment | 10 (10.5%) | 0 | 7 (12.2%) |
| No. on HT | 74 (78%) | 16 (73%) | 42 (74%) |
Data are expressed as mean (sd) unless otherwise indicated.
Two women included in this group met criteria for dysthymic disorder.
χ2 = 24.6; P < 0.001. Otherwise, all group comparisons were nonsignificant.
Twenty-two percent of women with POI (n = 39) met criteria for a current Axis I diagnosis, the majority of which were mood disorders (Table 2).
Timing of onset of depression (Fig. 1)
Fig. 1.
Dates of onset of depression, POI, and MCI were estimated to the closest month and year. In women in whom the month could not be identified, we standardized the report to January of that year for all self-reported events. A, Episodes of major depression (solid histograms) and minor depression (hatched histograms) before and after the diagnosis of POI. B, Episodes of major depression (solid histogram) and minor depression (hatched histograms) before and after the reported onset of MCI. The x-axis represents the 2-yr intervals before or after the diagnosis of POI (A) or MCI (B). The y-axis represents the first episode of major depression in each woman or, in those women who reported no episodes of major depression, the y-axis represents the first episode of minor depression. In the majority of women, whereas the onset of depression preceded the diagnosis of POI (A), depression occurred most commonly after the onset of MCI (B), a proxy for the early stages of ovarian insufficiency in these women. Of the 95 women who experienced at least one major depression, a greater number (n = 65; 68.4%) reported the onset of depression before the diagnosis of POI than after the diagnosis of POI (n = 30; 31.6%). Similarly, in those 22 women with at least a single episode of minor depression, 13 women (59.1%) reported the onset of depression before the diagnosis of POI. When the analyses were adjusted for differing years of exposure to the risk of depression before and after the diagnosis of POI, there were comparable numbers of depressive episodes occurring before and after the diagnosis of POI (Table 4). A greater number of women (n = 67; 73.6%) also reported the onset of depression after the onset of MCI (B). Sixteen of the 22 women with minor depression (72.7%) reported the onset of depression after the onset of MCI. Analyses adjusting for the years of exposure to risk demonstrated comparable numbers of women who developed depression after and before the onset of MCI (Table 4). Of the 65 women who had a major depression before the POI diagnosis, the length of time between the major depressive episode and the diagnosis of POI was 5.4 ± 5.5 yr. In the 13 women in whom minor depression occurred before the diagnosis of POI, the average length of time between minor depression and POI diagnosis was 5.4 ± 4.8 yr. For women whose depression occurred after the diagnosis of POI, the average duration between the POI diagnosis and the onset of major depression was 3.7 ± 4.3 yr, and for those with minor depression, the average duration was 2.3 ± 1.7 yr. In the women with major depression occurring after MCI (n = 67), there was an average duration between the onset of MCI and depression of 5.0 ± 4.9 yr. In those 24 women whose depression preceded MCI, there was an average of 6.1 ± 4.9 yr between the onset of depression and MCI. Observations were similar in those women with minor depression. Minor depression occurred on average 4.3 ± 3.6 yr after MCI, and in those minor depressions occurring before MCI, a difference of 5.3 ± 4.2 yr separated the two events.
Timing relative to the diagnosis of POI
As stated above, 95 women (54.5%) with POI reported at least one episode of major depression, and 22 (12.6%) reported at least one episode of minor depression (but no major depression). Of the 95 women who experienced at least one major depression, a significantly greater number reported the onset of depression before the diagnosis of POI (n = 65; 68.4%), than after the diagnosis of POI (n = 30; 31.6%) (P < 0.001). Similarly (although not statistically significant, P = 0.5), in those 22 women with at least a single episode of minor depression, 13 women (59.1%) reported the onset of depression before the diagnosis of POI, whereas nine women (40.9%) reported the onset of minor depression after the diagnosis of POI.
In contrast, in the analyses that took into account the differing years at risk before and after the diagnosis of POI, neither the Poisson rates for 3-, 5-, and 10-yr intervals [with one exception (Table 4)] nor the Kaplan-Meier analyses demonstrated a significant difference between the number of depressive episodes before and after the diagnosis of POI (for either major or minor depressions) [Kaplan-Meier: major depression, hazard ratio (after:before) = 1.5, P = 0.09; minor depression, hazard ratio (after:before) = 2.6; P = 0.054]. Nonetheless, in the 58 women who had SCID interviews performed at least 3 yr after the diagnosis of POI, a greater number of episodes of major depression (n = 9) were observed in the 3 yr before compared with the 3 yr after the diagnosis of POI (n = 5).
Table 4.
Poisson rates for major depression and minor depression
| No. of depression events | Depression years in time period | Events per year | P valuea | |
|---|---|---|---|---|
| Comparison A | ||||
| Major depression | ||||
| 3 yr | ||||
| Before POI | 30 | 134.5 | 0.22 | 0.19 |
| After POI | 17 | 50.9 | 0.33 | |
| 5 yr | ||||
| Before POI | 39 | 197.0 | 0.20 | 0.19 |
| After POI | 21 | 72.3 | 0.29 | |
| 10 yr | ||||
| Before POI | 54 | 289.9 | 0.19 | 0.16 |
| After POI | 27 | 101.5 | 0.27 | |
| Minor depression | ||||
| 3 yr | ||||
| Before POI | 4 | 31.4 | 0.13 | 0.18 |
| After POI | 6 | 17.1 | 0.35 | |
| 5 yr | ||||
| Before POI | 9 | 44.8 | 0.20 | 0.19 |
| After POI | 8 | 20.3 | 0.40 | |
| 10 yr | ||||
| Before POI | 11 | 61.8 | 0.18 | 0.04 |
| After POI | 9 | 20.5 | 0.44 | |
| Comparison B | ||||
| Major depression | ||||
| 3 yr | ||||
| Before MCI | 7 | 58.4 | 0.12 | 0.16 |
| After MCI | 33 | 144.6 | 0.22 | |
| 5 yr | ||||
| Before MCI | 10 | 87.6 | 0.11 | 0.13 |
| After MCI | 41 | 204.6 | 0.20 | |
| 10 yr | ||||
| Before MCI | 19 | 128.3 | 0.15 | 0.27 |
| After MCI | 58 | 290.0 | 0.20 | |
| Minor depression | ||||
| 3 yr | ||||
| Before MCI | 2 | 14.7 | 0.14 | 0.74 |
| After MCI | 8 | 36.2 | 0.22 | |
| 5 yr | ||||
| Before MCI | 4 | 21.0 | 0.19 | 1.0 |
| After MCI | 10 | 49.3 | 0.20 | |
| 10 yr | ||||
| Before MCI | 5 | 30.0 | 0.17 | 0.81 |
| After MCI | 14 | 66.3 | 0.21 |
Comparison A: Comparisons were performed as follows: 1) Poisson rates of depression were calculated (i.e. number of women with depression per person-years in the time period) to compare the number of the depressions that occurred in women before and after the diagnosis of POI. Three sets of Poisson rates were calculated for major as well as for minor depressions that were limited to depressions occurring within 3-, 5-, and 10-yr windows surrounding the diagnosis of POI. In these calculations, women whose depressions did not occur during the windows of time before or after the diagnosis of POI (i.e. 3, 5, and 10 yr, respectively) contributed person-years to the denominator. For each set, the before and after rates were compared using an exact Poisson test of homogeneity. 2) Kaplan-Meier survival analyses were performed to compare the onset of depressive episodes in the absolute number of days occurring before and after the diagnosis of POI. Neither the Poisson rates for 3-, 5-, and 10-yr intervals (with one exception) nor the Kaplan-Meier analyses demonstrated a significant difference between the number of depressive episodes before and after the diagnosis of POI (for either major or minor depressions) [Kaplan-Meier: major depression, hazard ratio (after:before) = 1.5; P = 0.09; minor depression, hazard ratio (after:before) = 2.6; P = 0.054].
Comparison B: Poisson rates and Kaplan-Meier survival analyses also were calculated to compare the number of episodes of depression that occurred before and after the reported onset of MCI. Neither the comparisons of the three sets of Poisson rates nor the Kaplan-Meier survival analyses identified significant differences in the number of depressive episodes (major or minor) preceding compared with those occurring after the onset of MCI [Kaplan-Meier: major depression, hazard ratio (after:before) = 1.2; P = 0.4; minor depression, hazard ratio (after:before) = 1.2; P = 0.7].
Exact Poisson test of homogeneity.
Timing relative to the onset of MCI
Of the remaining 91 women who reported at least one episode of major depression, 24 women (26.4%) reported depression before the onset of MCI, whereas 67 women (73.6%) reported the onset of depression after the onset of MCI (P < 0.001). Six of the 91 women reported the onset of both major depression and MCI to occur concurrently, and they were included in the group with depression occurring after the onset of MCI. However, we obtained similar results when we reanalyzed the data with these six women included in the group with depression occurring before the onset of MCI. Similarly, of the 22 women with minor depression, more women (n = 16; 72.7%) reported the onset of depression after the onset of MCI compared with before (n = 6; 27.3%) the onset of MCI [although of borderline statistical significance (P = 0.05)]. Neither the comparisons of the three sets of Poisson rates (Table 4) nor the Kaplan-Meier survival analyses identified significant differences in the number of depressive episodes (major or minor) preceding compared with those occurring after the onset of MCI [Kaplan-Meier: major depression—hazard ratio (after:before) = 1.2; P = 0.4; minor depression—hazard ratio (after:before) = 1.2; P = 0.7].
Discussion
This study, the first to use standardized diagnostic instruments to examine the prevalence of Axis I psychiatric disorders in a group of well-characterized women with spontaneous 46, XX POI, has two important findings. First, the evidence clearly demonstrates an increased prevalence of a lifetime history of depression (both major and minor) in POI, well beyond the rates of depression reported in several community-based samples of women. Second, in the majority of women, whereas the onset of depression preceded the diagnosis of POI, depressions occurred as commonly after as before the onset of MCI—a proxy for the early stages of ovarian insufficiency in these women. Thus, these data establish an accurate estimate of the prevalence and timing of the onset of clinically significant depression— preceding the diagnosis of POI, but in many women occurring after signs of altered ovarian function. Thus, depression is not likely to cause POI.
Although the lifetime rates of most psychiatric illnesses were comparable with other settings, the lifetime rate of mood disorders, in particular major depression, was considerably higher than those reported in either community- or clinic-based studies. In the National Comorbidity Study Replication study (NCSR) (21), rates of lifetime affective disorder were approximately 24.9%, compared with 67% in our study. Similarly, the lifetime rate of major depressive illness in women with POI was 54.5%, compared with the rate of 20% reported in the NCSR study. Conversely, the rate of lifetime anxiety disorders of 14% in women with POI is considerably lower than the 36.3% reported in the NCSR study. Thus, women with POI had a disproportionately higher lifetime risk of depression compared with other disorders and higher rates of depression relative to community- or medical clinic-based samples. Given the frequent appearance of depression before the diagnosis of POI, this selectively increased rate of mood disorders in women with POI suggests that depression and POI share common pathophysiological features. For example, the fragile-X syndrome premutation and abnormalities of immune function are both accompanied by disturbances in mood and behavior (15, 26, 27) and POI (4). The relatively lower rates of anxiety in our sample could reflect a recruitment bias in which women with more severe anxiety conditions were less likely to participate in a research study.
Our data are consistent with previous findings that in the majority of women with POI (68.4%), depression precedes the medical diagnosis of POI—albeit with the caveat that these comparisons potentially were confounded by differing lengths of time at risk before compared with after the diagnosis of POI. We observed approximately a 5.4-yr difference between the onset of depression and the diagnosis of POI, similar to the 10-yr window between the onset of depression and the diagnosis of POI in the community-based study reported by Harlow et al. (7). Nonetheless, we also observed that in the majority of women (73.6%) the onset of depression occurred after the onset of MCI in the context of the early stages of ovarian insufficiency. In these women, depression is occurring contemporaneously with the disease process (and before the final diagnosis of POI) and, therefore, is not causal. Finally, the ostensible increase in the number of episodes of depression in the 2 yr before the reported onset of MCI (Fig. 1) could conceivably reflect either a consequence of the physiological concomitants of POI or a reporting error in the exact age of onset of MCI.
These observations of an association between depression and POI suggest several possible relationships. First, depression could arise as a consequence of the diagnosis of POI. Needless to say, the implications of the diagnosis of POI for future fertility and role within the family could lead, in some women, to low self-esteem and possibly depression (2, 28). However, the high number of women in whom depression preceded the diagnosis of POI does not support this possibility as a sole explanation for the onset of depression. Second, as suggested by Harlow and Signorello (29), impaired ovarian function or an increased vulnerability for a more rapid decline in ovarian function could be a sequelae of either depression or antidepressant treatment. However, the relatively low prevalence of POI (i.e. 1% of women) is small compared with the lifetime rates of depression in 20% of women, suggesting that ovarian insufficiency is not a sequelae of depression in the majority of women. Third, several studies have identified an increased risk for major and minor depression during the normally timed menopause transition compared with the premenopause (30–33). Although the basis of this association is unknown, ovarian steroid receptors are widely distributed within the brain (34–37), and recent neuroimaging studies have confirmed the capacity of ovarian steroids to regulate activities of those brain regions and neurocircuits involved in affective regulation (38–41) in women. Thus, it is possible that either endocrine events or severe vasomotor symptoms (although clinically the latter is not a commonly reported complaint in POI) accompanying the transition to ovarian dysfunction (regardless of age) could contribute to the risk of developing depression. It is also possible that estrogen-based treatments could potentially be an effective therapeutic alternative to traditional antidepressants in depressed women with POI as they are in some older perimenopausal women (42–44). Finally, in a subgroup of women, it is possible that depression and POI reflect a common underlying pathophysiology or genetic vulnerability that increases the risk for both depression and POI. For example, the FMR1 premutation is associated not only with familial POI but also with a late-onset neurodegenerative disorder, FXTAS (fragile X associated tremor/ataxia syndrome) (4, 45, 46) as well as depression (26, 27). Recent studies suggest that depression is characterized by a loss of cellular resilience and neurodegeneration in specific brain regions (47), and it is possible that the same underlying process could lead to ovarian follicle degeneration.
In a prior study of adult women with TS (25), a condition also characterized by POI, we employed methods identical to those used in this study, including the same structured diagnostic interview. A lifetime rate of major depression was observed in only 36% of the women with TS compared with 54.5% of the women with spontaneous 46, XX POI evaluated in this study. Thus, ovarian insufficiency/failure alone is unlikely to be driving the relationship between depression and POI. Indeed, women with TS experience a longer duration of ovarian insufficiency and, therefore, should be at an increased risk for depression if our findings simply were caused by the duration of ovarian insufficiency and the presence of hypogonadism. Alternately, it is possible that the impact of ovarian failure differs when it occurs at a younger age in the context of supportive family members and physicians (as it would in many women with TS), compared with the discovery of ovarian failure in adulthood at the time a woman is attempting to conceive.
This study has several limitations. First, the differing years of exposure to risk before and after the diagnosis of POI and to a lesser extent the onset of MCI introduce a confound to our interpretation of the apparent clustering of depressive episodes around these physiological and medical events. Second, the diagnoses of both depression and MCI were determined on the basis of patient self-reports. Efforts were made to confirm dates of onset of POI, MCI, and depression. Given the premature nature of POI, there is at present no other way to determine the timing of the true onset of ovarian insufficiency. MCI is likely to be a delayed clinical manifestation of the underlying physiological changes of ovarian insufficiency that mark the onset of the disorder. Thus, our inferences about the temporal patterns in the onset of depression and MCI are at risk of misclassification. Nonetheless, earlier markers are not available, and so the cycle changes are a reasonable proxy. A prospective study could examine this possibility; however, the low prevalence rates of POI (i.e. 1% of women) would make such a study difficult to power and complete. Finally, in this clinic-based study, there is an inherent acquisition bias that could prevent women with more severe mental illness from presenting for participation. Women with more severe forms of mental illness such as bipolar illness and schizophrenia could be sufficiently disabled by their psychiatric illness to prevent their volunteering for this study. Alternately, their access to gynecological care could be limited or their physician could attribute MCI or amenorrhea to either their illness or its treatment (48–51), and not to POI. Thus, these women might not be referred for evaluation of POI. Additionally, POI is a relatively rare condition, representing only 1% of population, and our sample size of 174 women could be insufficient to identify equally rare conditions such as bipolar illness and schizophrenia with prevalence rates of 4.5 and 1%, respectively. Finally, it also is possible that the psychiatric evaluation performed in this study could attract women with more severe psychiatric illness to participate.
Future studies of this association could identify candidate genes and physiological processes that may be involved in the development of both depression and POI. Certainly a primary candidate for this risk is genomic variation including the FMR1 gene and proapoptotic genes in which pathology in both ovarian and brain tissue could occur. Finally, the high rates of depression accompanying the diagnosis of POI strongly suggest that attention to the possible presence of depression in these women should be part of the evaluation and management of all women who present with POI.
Acknowledgments
We acknowledge the collaboration with Dr. Carolyn Bondy who permitted our use of the data in women with Turner syndrome. We also acknowledge the contributions of Dr. Robert Daly (clinical support), Ms. Elinor Spoor (data management), and Dr. Robert Wesley (statistical consultation and expertise).
P.J.S. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. This work was written as part of the official duties of P.J.S. as a U.S. Government employee. The views expressed in this article do not necessarily represent the views of the National Institute of Mental Health (NIMH), National Institutes of Health, the Department of Health & Human Services, or the U.S. Government. No grants or fellowships supported the writing of this paper. This work was funded by the Intramural Research Programs of the National Institute of Child Health and Human Development and NIMH.
Disclosure Summary: None of the authors report any potential conflicts of interest relevant to the information contained within this manuscript.
Footnotes
- BMI
- Body mass index
- MCI
- menstrual cycle irregularity
- POI
- primary ovarian insufficiency.
References
- 1. Welt CK. 2008. Primary ovarian insufficiency: a more accurate term for premature ovarian failure. Clin Endocrinol (Oxf) 68:499–509 [DOI] [PubMed] [Google Scholar]
- 2. Nelson LM, Covington SN, Rebar RW. 2005. An update: spontaneous premature ovarian failure is not an early menopause. Fertil Steril 83:1327–1332 [DOI] [PubMed] [Google Scholar]
- 3. Kalantaridou SN, Davis SR, Nelson LM. 1998. Premature ovarian failure. Endocrinol Metab Clin North Am 27:989–1006 [DOI] [PubMed] [Google Scholar]
- 4. Nelson LM. 2009. Primary ovarian insufficiency. N Engl J Med 360:606–614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Santoro N. 2003. Mechanisms of premature ovarian failure. Ann Endocrinol (Paris) 64:87–92 [PubMed] [Google Scholar]
- 6. Coulam CB, Adamson SC, Annegers JF. 1986. Incidence of premature ovarian failure. Obstet Gynecol 67:604–606 [PubMed] [Google Scholar]
- 7. Harlow BL, Cramer DW, Annis KM. 1995. Association of medically treated depression and age at natural menopause. Am J Epidemiol 141:1170–1176 [DOI] [PubMed] [Google Scholar]
- 8. Liao KL, Wood N, Conway GS. 2000. Premature menopause and psychological well-being. J Psychosom Obstet Gynaecol 21:167–174 [DOI] [PubMed] [Google Scholar]
- 9. Berg JS, Moore J. 2000. Early menopause presenting with mood symptoms in a student aviator. Aviat Space Environ Med 71:251–254 [PubMed] [Google Scholar]
- 10. Schmidt PJ, Cardoso GM, Ross JL, Haq N, Rubinow DR, Bondy CA. 2006. Shyness, social anxiety, and impaired self-esteem in Turner syndrome and premature ovarian failure. JAMA 295:1374–1376 [DOI] [PubMed] [Google Scholar]
- 11. Davis M, Ventura JL, Wieners M, Covington SN, Vanderhoof VH, Ryan ME, Koziol DE, Popat VB, Nelson LM. 2010. The psychosocial transition associated with spontaneous 46,XX primary ovarian insufficiency: illness uncertainty, stigma, goal flexibility, and purpose in life as factors in emotional health. Fertil Steril 93:2321–2329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wang PS, Lane M, Olfson M, Pincus HA, Wells KB, Kessler RC. 2005. Twelve-month use of mental health services in the United States. Results from the National Comorbidity Survey Replication. Arch Gen Psychiatry 62:629–640 [DOI] [PubMed] [Google Scholar]
- 13. Wang PS, Berglund P, Olfson M, Pincus HA, Wells KB, Kessler RC. 2005. Failure and delay in initial treatment contact after first onset of mental disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 62:603–613 [DOI] [PubMed] [Google Scholar]
- 14. 1994. Diagnostic and statistical manual of mental disorders. 4th ed Washington, DC: American Psychiatric Association [Google Scholar]
- 15. Miller AH, Maletic V, Raison CL. 2009. Inflammation and its discontents: the role of cytokines in the pathophysiology of major depression. Biol Psychiatry 65:732–741 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Grambsch P, Young EA, Meller WH. 2004. Pulsatile luteinizing hormone disruption in depression. Psychoneuroendocrinology 29:825–829 [DOI] [PubMed] [Google Scholar]
- 17. Young EA, Midgley AR, Carlson NE, Brown MB. 2000. Alteration in the hypothalamic-pituitary-ovarian axis in depressed women. Arch Gen Psychiatry 57:1157–1162 [DOI] [PubMed] [Google Scholar]
- 18. Kaplan JR. 2008. Origins and health consequences of stress-induced ovarian dysfunction. Interdiscip Top Gerontol 36:162–185 [DOI] [PubMed] [Google Scholar]
- 19. First MB, Spitzer RL, Gibbon M, Williams JB. 1996. Structured clinical interview for DSM-IV axis I disorders—patient edition. New York: Biometrics Research Department, New York State Psychiatric Institute [Google Scholar]
- 20. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. 2005. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the national comorbidity survey replication. Arch Gen Psychiatry 62:593–602 [DOI] [PubMed] [Google Scholar]
- 21. Kessler RC. 2006. Appendix table 1: lifetime prevalence of DSM–IV/WMH-CIDI disorders by sex and cohort. National Comorbidity Survey. Available at: http://www.hcp.med.harvard.edu/ncs/ftpdir/table_ncsr_by_gender_and_age.pdf Accessed September 7, 2006
- 22. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, Wittchen HU, Kendler KS. 1994. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from the National Comorbidity Survey. Arch Gen Psychiatry 51:8–19 [DOI] [PubMed] [Google Scholar]
- 23. Linzer M, Spitzer R, Kroenke K, Williams JB, Hahn S, Brody D, deGruy F. 1996. Gender, quality of life, and mental disorders in primary care: results from the PRIME-MD 1000 study. Am J Med 101:526–533 [DOI] [PubMed] [Google Scholar]
- 24. Spitzer RL, Williams JB, Kroenke K, Hornyak R, McMurray J. 2000. Validity and utility of the PRIME-MD patient health questionnaire in assessment of 3000 obstetric-gynecologic patients: the PRIME-MD Patient Health Questionnaire Obstetrics-Gynecology Study. Am J Obstet Gynecol 183:759–769 [DOI] [PubMed] [Google Scholar]
- 25. Cardoso G, Daly RJ, Haq NA, Hanton L, Rubinow DR, Bondy CA, Schmidt PJ. 2004. Current and lifetime psychiatric illness in women with Turner syndrome. Gynecol Endocrinol 19:313–319 [PubMed] [Google Scholar]
- 26. Rodriguez-Revenga L, Madrigal I, Alegret M, Santos M, Milà M. 2008. Evidence of depressive symptoms in fragile-X syndrome premutated females. Psychiatr Genet 18:153–155 [DOI] [PubMed] [Google Scholar]
- 27. Bourgeois JA, Coffey SM, Rivera SM, Hessl D, Gane LW, Tassone F, Greco C, Finucane B, Nelson L, Berry-Kravis E, Grigsby J, Hagerman PJ, Hagerman RJ. 2009. A review of fragile X premutation disorders: expanding the psychiatric perspective. J Clin Psychiatry 70:852–862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Sutton EJ, McInerney-Leo A, Bondy CA, Gollust SE, King D, Biesecker B. 2005. Turner Syndrome: four challenges across the lifespan. Am J Med Genet 139A:57–66 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Harlow BL, Signorello LB. 2000. Factors associated with early menopause. Maturitas 35:3–9 [DOI] [PubMed] [Google Scholar]
- 30. Cohen LS, Soares CN, Vitonis AF, Otto MW, Harlow BL. 2006. Risk for new onset of depression during the menopausal transition. The Harvard study of moods and cycles. Arch Gen Psychiatry 63:385–390 [DOI] [PubMed] [Google Scholar]
- 31. Freeman EW, Sammel MD, Lin H, Nelson DB. 2006. Associations of hormones and menopausal status with depressed mood in women with no history of depression. Arch Gen Psychiatry 63:375–382 [DOI] [PubMed] [Google Scholar]
- 32. Bromberger JT, Assmann SF, Avis NE, Schocken M, Kravitz HM, Cordal A. 2003. Persistent mood symptoms in a multiethnic community cohort of pre- and perimenopausal women. Am J Epidemiol 158:347–356 [DOI] [PubMed] [Google Scholar]
- 33. Bromberger JT, Schott LL, Kravitz HM, Sowers M, Avis NE, Gold EB, Randolph JF, Jr, Matthews KA. 2010. Longitudinal change in reproductive hormones and depressive symptoms across the menopausal transition: results from the Study of Women's Health Across the Nation (SWAN). Arch Gen Psychiatry 67:598–607 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Kruijver FP, Balesar R, Espila AM, Unmehopa UA, Swaab DF. 2003. Estrogen-receptor-β distribution in the human hypothalamus: similarities and differences with ERα distribution. J Comp Neurol 466:251–277 [DOI] [PubMed] [Google Scholar]
- 35. Osterlund MK, Gustafsson JA, Keller E, Hurd YL. 2000. Estrogen receptor β (ERβ) messenger ribonucleic acid (mRNA) expression within the human forebrain: distinct distribution pattern to ERα mRNA. J Clin Endocrinol Metab 85:3840–3846 [DOI] [PubMed] [Google Scholar]
- 36. Brinton RD, Thompson RF, Foy MR, Baudry M, Wang J, Finch CE, Morgan TE, Pike CJ, Mack WJ, Stanczyk FZ, Nilsen J. 2008. Progesterone receptors: form and function in brain. Front Neuroendocrinol 29:313–339 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Mani S. 2008. Progestin receptor subtypes in the brain: the known and the unknown. Endocrinology 149:2750–2756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Berman KF, Schmidt PJ, Rubinow DR, Danaceau MA, Van Horn JD, Esposito G, Ostrem JL, Weinberger DR. 1997. Modulation of cognition-specific cortical activity by gonadal steroids: a positron-emission tomography study in women. Proc Natl Acad Sci USA 94:8836–8841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Dreher JC, Schmidt PJ, Kohn P, Furman D, Rubinow D, Berman KF. 2007. Menstrual cycle phase modulates reward-related neural function in women. Proc Natl Acad Sci USA 104:2465–2470 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Goldstein JM, Jerram M, Poldrack R, Ahern T, Kennedy DN, Seidman LJ, Makris N. 2005. Hormonal cycle modulates arousal circuitry in women using functional magnetic resonance imaging. J Neurosci 25:9309–9316 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Protopopescu X, Pan H, Altemus M, Tuescher O, Polanecsky M, McEwen B, Silbersweig D, Stern E. 2005. Orbitofrontal cortex activity related to emotional processing changes across the menstrual cycle. Proc Natl Acad Sci USA 102:16060–16065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Soares CN, Almeida OP, Joffe H, Cohen LS. 2001. Efficacy of estradiol for the treatment of depressive disorders in perimenopausal women: a double-blind, randomized, placebo-controlled trial. Arch Gen Psychiatry 58:529–534 [DOI] [PubMed] [Google Scholar]
- 43. Schmidt PJ, Nieman L, Danaceau MA, Tobin MB, Roca CA, Murphy JH, Rubinow DR. 2000. Estrogen replacement in perimenopause-related depression: a preliminary report. Am J Obstet Gynecol 183:414–420 [DOI] [PubMed] [Google Scholar]
- 44. Zweifel JE, O'Brien WH. 1997. A meta-analysis of the effect of hormone replacement therapy upon depressed mood. Psychoneuroendocrinology 22:189–212 [DOI] [PubMed] [Google Scholar]
- 45. Terracciano A, Chiurazzi P, Neri G. 2005. Fragile X syndrome. Am J Med Genet 137C:32–37 [DOI] [PubMed] [Google Scholar]
- 46. Marozzi A, Vegetti W, Manfredini E, Tibiletti MG, Testa G, Crosignani PG, Ginelli E, Meneveri R, Dalpra L. 2000. Association between idiopathic premature ovarian failure and fragile X premutation. Hum Reprod 15:197–202 [DOI] [PubMed] [Google Scholar]
- 47. Manji HK, Drevets WC, Charney DS. 2001. The cellular neurobiology of depression. Nature Med 7:541–547 [DOI] [PubMed] [Google Scholar]
- 48. Canuso CM, Goldstein JM, Wojcik J, Dawson R, Brandman D, Klibanski A, Schildkraut JJ, Green AI. 2002. Antipsychotic medication, prolactin elevation, and ovarian function in women with schizophrenia and schizoaffective disorder. Psychiatry Res 111:11–20 [DOI] [PubMed] [Google Scholar]
- 49. Joffe H, Hayes FJ. 2008. Menstrual cycle dysfunction associated with neurologic and psychiatric disorders: their treatment in adolescents. Ann NY Acad Sci 1135:219–229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Kenna HA, Jiang B, Rasgon NL. 2009. Reproductive and metabolic abnormalities associated with bipolar disorder and its treatment. Harv Rev Psychiatry 17:138–146 [DOI] [PubMed] [Google Scholar]
- 51. Joffe H, Kim DR, Foris JM, Baldassano CF, Gyulai L, Hwang CH, McLaughlin WL, Sachs GS, Thase ME, Harlow BL, Cohen LS. 2006. Menstrual dysfunction prior to onset of psychiatric illness is reported more commonly by women with bipolar disorder than by women with unipolar depression and healthy controls. J Clin Psychiatry 67:297–304 [DOI] [PubMed] [Google Scholar]