The risk factors and symptomatology of perimenopausal depression

Maturitas 55 (2006) 180–186

The risk factors and symptomatology of perimenopausal depression ¨ ur Ozt¨ ¨ urk a,∗ , Defne Eraslan b , Hayriye Elbi Mete c , Serdar Ozs ¨ ¸ener d Ozg¨ a

d

TAPD (Family Planning Association of Turkey) Buca Medical Center, Ko¸suyolu cad. 506 sok. No. 2, S¸irinyer/Buca, Izmir, Turkey b Ege University Medico-Social Unit, Izmir, Turkey c Ege University Faculty of Medicine Department of Psychiatry, Izmir, Turkey Ege University Faculty of Medicine Department of Obstetrics & Gynecology, Izmir, Turkey

Received 27 September 2005; received in revised form 7 February 2006; accepted 8 February 2006

Abstract Objectives: The aim of this study is to determine if the depression in perimenopausal women is symptomatologically different than depression in premenopausal women, and if these depressive women are under more risk for depression when factors like premenstrual dysphoric syndrome, socio-economical status, vasomotor symptoms and familial inclination to depression are considered. Methods: Fifty major depressive women with hormonally established perimenopause (follicle stimulating hormone (FSH) levels over 20 IU/l and estrogen levels lower than 40 IU/l) and who were not taking any psychotropic or hormone replacement therapy, were enrolled in the study. In order to investigate the characteristic clinical features of perimenopausal depression, a drug free control group consisting of 48 pre-perimenopausal women with a diagnosis of major depression was formed. Another control group was formed with 53 non-depressive perimenopausal women in order to investigate the risk factors of perimenopausal depression. Results: The depressive and non-depressive groups did not differ from each other according to level of education, marital and economical status and comorbid physical problems. No major symptomatological difference between perimenopausal depression and pre-perimenopausal depression was found. No correlation was found between the severity of vasomotor symptoms and severity of depression. Conclusions: Findings of our study suggest that vasomotor symptoms and socio-economical status do not predict the severity and existence of perimenopausal depression. Episodes of major depression are not necessarily the normal result of such vasomotor symptoms. Therefore, the evaluation and management of perimenopausal depression should be carried out as carefully as is done in episodes of depression seen in the rest of women’s life span. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Depression; Hot flushes; Night sweats; Vasomotor symptoms; Menopause

∗

Corresponding author. Tel.: +90 532 6545895; fax: +90 232 4388344. ¨ Ozt¨ ¨ urk). E-mail address: [email protected] (O.

0378-5122/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.maturitas.2006.02.001

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

1. Introduction Menopause; is a unique endocrinological phenomenon which all middle-aged women have to cope. Although the word ‘menopause’ points out the last menstrual bleeding, it is often used for the approximately 10 years long climacteric period, during which many important biological, psychological and social changes occur. The response to the transition from perimenopausal period to menopause is multifactorial and is associated with women’s individual personalities and coping styles [1]. Menopause is epidemiologically described as amenorrhea for 12 consequent months in the absence of causes such as pregnancy or lactation. The perimenopausal period is clinically described as irregular menstrual cycles or amenorrhea for 11 months or less. In the early phase of perimenopause, elevation of follicle stimulating hormone (FSH) might be the only sign [2,3]. In the perimenopausal period, FSH levels over 25 IU/l and estrogen levels under 40 pg/ml are measured in 2nd or 3rd day of menstruation, while after menopause the FSH level is over 40 IU/l and the estrogen level is under 25 pg/ml [2]. The centuries old belief that women are more prone to depression during climacteric period has lost its popularity [4,5] although the rate of depression is clearly high in perimenopausal women attending menopause clinics, sometimes reaching rates as high as 44% [6]. In community based studies, it is found that depressive symptoms are more common among perimenopausal women, especially women who have more serious vasomotor symptoms like night sweats, flushing and sexual dysfunction [7,8]. As these vasomotor symptoms diminish, affective symptoms diminish as well [3,9]. This increase in depressive symptoms may be the result of dysphoria caused by vasomotor symptoms as well as increasing psychosocial stressors within the perimenopausal period [4]. It is not clearly known if perimenopausal depression has a characteristic symptomatology. Some authors argue that, hormonal changes in a group who are more prone to affective syndromes related to reproductive events; might induce a specific type of depression [10]. It is also indicated that natural menopause has no negative mental health consequences for the majority of middle-aged healthy women [11]. However, there is no study which describes risk factors and symptomatology of perimenopausal depression

181

with women who are endocrinologically proved to be perimenopausal. The aim of this study is to determine if the depression in perimenopausal women is symptomatologically different than depression in premenopausal women, and if these depressive women are under more risk for depression when factors like premenstrual dysphoric syndrome, socio-economical status, vasomotor symptoms and familial inclination to depression are considered.

2. Method Fifty major depressive women with hormonally established perimenopause and who were not taking any psychotropic or hormone replacement therapy, were enrolled in the study. In order to investigate if perimenopausal depression has characteristic clinical features a control group consisting of 48 pre-perimenopausal women with a diagnosis of major depression and who were not under psychotropic treatment was formed. Another control group was formed with 53 non-depressive perimenopausal women who were under control in the obstetrics and gynecology department’s menopause clinic in order to investigate the risk factors of perimenopausal depression. Subjects who had follicle stimulating hormone levels over 20 IU/l and estrogen levels lower than 40 IU/l were considered as perimenopausal. Women whose FSH levels were lower and whose estrogen levels were higher were included in the pre-perimenopausal group. Subjects with FSH levels over 40 IU/l and estrogen levels lower than 25 pg/ml were excluded from the study. Psychiatric evaluation was done with SCID-I (structured clinical interview according to Diagnostic and Statistical Manual—DSM) form and depressive symptoms were evaluated by the 17-item Hamilton Depression Scale (Ham-D) and Montgomery Asberg Depression Scale (MADRS) while Greene Climacteric Scale (GCS) was used to determine the severity of menopausal symptoms. In order to form the perimenopausal study groups, all women attending the university perimenopause clinic in a period of 5 months, who were in natural perimenopause according to hormone levels and who were not taking hormone replacement therapy were interviewed by the researchers. The psychiatric exam-

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

182

ination using the SCID-I form was completed after the Ham-D, MADRS and GCS were applied. The women fulfilling the criteria for depression in the SCID-I form were included in the perimenopausal depression group, while the remaining women were included in the nondepressive perimenopausal group. For the pre-perimenopausal group, all women that came to the university psychiatric outpatient clinic during the same 5-month period, who were over the age of 40 years and were not taking any psychotropic drugs were interviewed using the SCID-I form and the Ham-D, MADRS and GCS. Blood samples were taken from those who were diagnosed as major depressive and those who were pre-perimenopausal according to their hormonal levels were included in the preperimenopausal depressive group. The history of Premenstrual Dysphoria (PMD) was questioned retrospectively by using the DSM-IV criteria for Premenstrual Dysphoria which includes irritability, anxiety, mood swings, fatigue, sleep changes, appetite changes, decreased interest and difficulties in concentration, feeling overwhelmed or out of control and some physical symptoms like bloating or breast tenderness. Minimum five of these symptoms provided that one of them is a mood symptom must exist during at least two menstrual cycles. Chi-square was used for categorical variables while student t-test and ANOVA was used for continuous variables. The differences between groups were analyzed with Pearson’s correlation test and least significant difference (LSD) test. The significance criterion was determined as p < 0.05.

3. Results Three groups were formed from a total number of 151 women according to a structured SCID-I inter-

view and hormone levels: perimenopausal without depression, perimenopausal with depression and preperimenopausal with depression. Table 1 shows some demographic features of the groups. The mean age was 50.4 ± 2.65 in perimenopausal depressive group, 51 ± 3.21 in non-depressive perimenopausal group and 44.69 ± 4.23 in pre-perimenopausal group. There was significant difference between groups according to age (p < 0.001) but test of least significant difference revealed that this difference was the effect of pre-perimenopausal depressive group while perimenopausal depressive and perimenopausal nondepressive groups did not differ from each other according to age. The duration of time after the last menstrual bleeding was 10.15 months for the nondepressive perimenopausal group and 7.5 months for the depressive perimenopausal group. There was no difference between groups according to level of education, economical and marital status and duration of time of the postmenopausal period. 2.6% (n = 4) of our subjects were illiterate and 27.2% (n = 41) had finished elementary school while 29.1% (n = 44) had graduated from high school and 29.8% (n = 45) from university. Table 2 shows the distribution of level of education according to study groups. Pearson’s correlation test revealed no significant relationship between level of education and study groups (p = 0.65). Table 3 shows the distribution of subjects according to marital status and study groups. No significant relation between marital status and groups was found as well (p = 0.45). 70.9% (n = 97) of subjects had at least one comorbid medical disease apart from depression and menopausal symptoms. However, one-way ANOVA test revealed no significant difference between study groups according to accompanying medical diseases (p = 0.98). During the last 1 year, the non-depressive perimenopausal group has used health facilities more than

Table 1 Some demographic features and mean Ham-D total scores of the groups Groups Non-depressive perimenopausal Depressive perimenopausal Depressive pre-perimenopausal Total *

1 TL = D 0.6.

n

Mean age

Duration after last menstrual bleeding (months)

Annual monthly income (Turkish Liras, TL)*

Mean HAM-D score

53 50 48

51.04 50.40 44.69

10.15 7.50 –

1200.00 1071.88 1015.63

5.38 21.52 20.02

151

48.81

1100.66

15.38

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

183

Table 2 The distribution of level of education according to study groups, no significant correlation between level of education and groups was found (Pearson’s correlation test) Level of education

Non-depressive perimenopausal

Depressive perimenopausal

Depressive pre-perimenopausal

Total

Illiterate Primary school Secondary school High school University Total

– 10 (24.4) 10 (18.9) 20 (37.7) 13 (24.5) 53 (100)

3 (6.0) 17 (34.0) 4 (8.0) 8 (16.0) 18 (36.0) 50 (100)

1 (2.1) 14 (29.2) 3 (6.3) 16 (33.3) 14 (29.2) 48 (100)

4 (2.6) 41 (27.2) 17 (11.3) 44 (29.1) 45 (29.8) 151 (100)

Values in parenthesis are in percentage. p = 0.65. Table 3 The distribution of subjects according to marital status and study groups Groups

Married Single Widow Divorced Total

Non-depressive perimenopausal n 45 Within group (%) 84.9 Depressive perimenopausal n 38 Within group (%) 76

3 5.7 2 4

Depressive pre-perimenopausal n 37 Within group (%) 77.1 Total n 120 Within group (%) 79.5

2 1.3

5 9.4

3 6

53 100.0

7 14

50 100.0

5 10.4

6 12.5

48 100.0

11 7.3

18 11.9

151 100.0

No significant relation between marital status and groups was found (Pearson’s correlation test). p = 0.45.

perimenopausal depressive and pre-perimenopausal depressive groups. The number of visits to health facilities was significantly different between groups (p < 0.05). Table 4 shows the mean number of visits to health facilities and significance according to the groups. A total number of 53 subjects had at least one or more comorbid past or present psychiatric diagnoses.

The number of comorbid psychiatric diagnoses was one in 31.8% (n = 48), two in 2.6% (n = 4) and three in 0.7% (n = 1). Past depression was the most common comorbid diagnosis (n = 30, 53%), but the three groups did not differ from each other according to their history of past depression (p = 0.39). The incidence of PMD in the non-depressive perimenopausal group was 39.6% and while it was 48% in depressive perimenopausal group, and these groups did not differ from each other in terms of history of PMD (p = 0.77). 58.3% of the pre-perimenopausal group had PMD but this group did not differ from the perimenopausal depressive group (p = 0.41) as well. 15.1% of non-depressive perimenopausal group had family history of depression while this rate was 44% for perimenopausal depressive group and 25% for pre-perimenopausal depressive group. When the existence of family history of depression was compared using Pearson’s test between groups, the difference between depressive perimenopausal patients and non-depressive perimenopausal patients was significant (p = 0.009). Table 5 shows the distribution of family history of depression according to groups. The mean Ham-D score of the non-depressive perimenopausal group was 5.38 while it was 21.52 in the perimenopausal depressive group, and 20.02 in the

Table 4 The mean number of visits to health facilities according to groups during last year (Pearson’s correlation test) Groups Non-depressive perimenopausal Depressive perimenopausal Depressive pre-perimenopausal Total *

p < 0.05.

n

Mean no. of visits to health facilities during last year (S.D.)

F

p

53 50 48

8.60 (6.11) 7.52 (5.05) 5.67 (3.57)

4.332

0.015*

151

7.32 (5.19)

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

184

Table 5 The distribution of family history of depression according to groups Groups

Without family history of depression

With family history of depression

Total

Non-depressive perimenopausal n Within group (%)

45 84.9

8 15.1**

53 100

Depressive perimenopausal n Within group (%)

28 56

22 44**

50 100

Depressive pre-perimenopausal n Within group (%)

36 75.0

12 25.0

48 100

109 72.2

42 27.8

151 100

Total n Within group (%) **

p = 0.009.

Table 6 The mean Hamilton Depression Scale total scores and standard deviation in three groups

Table 8 The scores of the vasomotor symptom subscale of Greene Climacteric scale in the two perimenopausal groups

Groups Non-depressive perimenopausal Depressive perimenopausal Depressive pre-perimenopausal Total

n

Mean (S.D.)

Groups

n

Mean (S.D.)

t

p

53 50 48

5.38 (2.60) 21.52 (7.68) 20.02 (5.65)

53

4.19 (2.14)

−0.552

0.582

50

4.62 (2.07)

151

15.38 (9.30)

Non-depressive perimenopausal Depressive perimenopausal

pre-perimenopausal depressive group. The difference between the groups was significant (p < 0.01) (Table 6). The LSD test to determine the source of this difference showed that the pre-perimenopausal depressive group and the perimenopausal depressive group were similar in terms of mean Ham-D score (p = 0.19), while the mean Ham-D score of the non-depressive perimenopausal group was significantly lower than the both depressive groups (p < 0.001), as expected (Table 7). Meanwhile, the duration after last menstrual bleeding was not related with mean Ham-D scores in perimenopausal depressive group (p = 0.613).

Mean score of non-depressive perimenopausal group in vasomotor symptom subscale of Greene Climacteric scale (GCS) was 2.92 (maximum score is 6) while it was 3.08 for depressive perimenopausal group, and the difference between groups was not significant (Table 8). Either depressive or non-depressive, both perimenopausal groups had similar severity of vasomotor symptoms. The total vasomotor symptoms scale score in GCS of depressive perimenopausal group did not overlap with total Ham-D scores according to Pearson’s correlation test (p = 0.08). There was no correlation between the “night sweats” item

Table 7 The comparison of groups according to total Hamilton Depression Scale scores (test of least significance) Group (A)

Group (B)

Mean difference (A − B) (S.D.)

p

Non-depressive perimenopausal Non-depressive perimenopausal Pre-perimenopausal depression

Depressive perimenopausal Pre-perimenopausal depression Depressive perimenopausal

−16.14** (1.12) −14.64** (1.13) −1.50 (1.14)

0.000 0.000 0.192

**

p < 0.01.

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

185

Table 9 The comparison of groups according to MARDRS emotional withdrawal item Group (A)

Group (B)

Mean difference (S.D.)

p

MADRS emotional withdrawal Non-depressive perimenopausal Non-depressive perimenopausal Depressive pre-perimenopausal

Depressive perimenopausal Depressive pre-perimenopausal Depressive perimenopausal

−2.37* (0.213) −3.01* (0.237) 0.65* (0.220)

0.000 0.000 0.004

*

p < 0.05.

of the GCS and the sleep item of the MADRS (p = 0.29). The three groups were found to be different from each other according to all sub-items of MADRS when each of items were evaluated with one-way ANOVA test (p < 0.001). But the LSD test revealed that this significance particularly arose from the difference between depressive and non-depressive perimenopausal groups. Naturally, depressive groups either premenopausal or pre-perimenopausal, had significantly higher emotional withdrawal scores compared to non-depressive perimenopausal group. Nevertheless, the perimenopausal depressive and pre-perimenopausal depressive groups also differed from each other in ‘emotional withdrawal’ sub-item of MADRS. The pre-perimenopausal depressive group had significantly higher ‘emotional withdrawal’ scores in MADRS compared to perimenopausal depressive group (p = 0.04) (Table 9).

4. Discussion In our study, no major symptomatological difference between perimenopausal depression and preperimenopausal depression was found. The preperimenopausal group had significantly higher scores in the ‘emotional withdrawal’ item of the MADRS. This finding might be related with anhedonia being a more important obstacle of daily life for young and active pre-perimenopausal women who might possibly have higher social expectations. According to the ‘domino’ theory of depression during perimenopausal period, sleep problems, dysphoria, irritability and difficulties in concentration are consequences of vasomotor symptoms such as hot flashes and night sweats [12]. In our study, the depressive and nondepressive perimenopausal women did not differ from each other according to vasomotor symptoms and items

of GCS. No correlation was found between the severity of vasomotor symptoms and severity of depression. Night sweats, which are claimed to be the main cause of depressive symptomatology via decreasing sleep quality, were not correlated with severity of depression and MADRS sleep items score either. Findings of our study suggest that vasomotor symptoms do not predict the severity and existence of depression. These results indicate that the idea of perimenopausal depression is associated with hot flushes and night sweats might not be true [13]. According to the ‘psychosocial’ theory, the affective symptoms during perimenopausal period are due to marital problems, physical illnesses and insufficient economical status in this group [14]. Our results have shown that the depressive and non-depressive groups did not differ from each other according to level of education, marital and economical status and comorbid physical problems. During the last 1 year the non-depressive perimenopausal group have used health facilities significantly more than pre-perimenopausal depressive group. As our entire study group had social insurance, this difference might be related with the nondepressive group being registered to a menopause unit and thus having more consciousness about their health problems and seeking for help. These data once more emphasizes the importance of specific menopause units in clinical follow-up of such patients. There is no significant difference between the groups according to the history of post partum depression and premenstrual dysphoric syndrome. The premenstrual data had to be collected from patients themselves retrospectively which might decrease the confidence of the data and thus causing a limitation for his study. The incidence of familial history of depression in perimenopausal depressive group was higher than pre-

186

¨ Ozt¨ ¨ urk et al. / Maturitas 55 (2006) 180–186 O.

perimenopausal depressive group, but it is interesting that perimenopausal group does not have higher incidence of past history of depression although they have a familial inclination. These data are correlated with other studies which suggest that first attack depression is more common among perimenopausal group [10]. Hormonal changes and stressful life events might induce depression in people who have familial inclination to depression. Consequently, it might be suggested that the major depression in perimenopausal period does not symptomatologically differ from depression in pre-perimenopausal period except less common seen anhedonia. It is suggested that perimenopause may represent a period of increased psychiatric vulnerability particularly for mood disorders [15]. Our findings do not confirm this idea as we could not find any symptomatological difference between pre-perimenopausal depression and perimenopausal depression and no relation between severity of vasomotor symptoms and severity of depression as well. During this period, people who face more and serious stressful life events and have familial inclination might be under higher risk for depression. According to our results, we may suggest that perimenopausal depression is not only a dysphoric state that is simply secondary to night sweats and vasomotor symptoms. It may be natural for perimenopausal women to experience minor mood fluctuations as a result of insomnia or hot flushes, but episodes of major depression are not necessarily the normal result of such vasomotor symptoms. Therefore, the evaluation and management of perimenopausal depression should be carried out as carefully as is done in episodes of depression seen in the rest of women’s life span. However, our study is not an epidemiologic study nor community based, so one must avoid generalizing these results unless they are confirmed in epidemiologic studies.

References [1] Bosworth HB, Bastian LA, Rimer BK, Siegler IC. Coping styles and personality domains related to menopausal stress. Womens Health Issues 2003;13(January–February (1)):32–8. [2] Gracia CR, Sammel MD, Freeman EW, et al. Defining menopause status: creation of a new definition to identify the early changes of the menopausal transition. Menopause 2005;12(March (2)):128–35. [3] Avis NE, Mc Kinlay SM. The Massachussets Women’s Health Study: an epidemic investigation of the menopause. J Am Med Women’s Assoc 1995;50:45–9, 63. [4] Burt VK, Altschuler L, Rasgon N. Depressive symptoms in the perimenopause: prevalence, assesment, and guidelines for treatment. Harv Rev Psychiatry 1998;6:121–32. [5] Weissmann MM. The myth of involutional melancholia. JAMA 1979;242:742–4. [6] Hay AG, Bancroft J, Johnstone EC. Affective symptoms in women attending a menopause clinic. Br J Psychiaty 1994;164:513–6. [7] Mc Kinlay SM, Jeffers M. The menopausal syndrome. Br J Preventive Soc Med 1974;28:108–15. [8] Hunter MS. Somatic experience of the menopause: a prospective study. Psychosom Med 1990;52:357–67. [9] Avis NE, Mc Kinlay SM. A longitudinal analysis of women’s attidutes towards the menopause: results from the Massachussets Women’s Health Study. Maturitas 1991;13:65–79. [10] Schmidt PJ, Roca CA, Bloch M, Rubinow DR. The perimenopause and affective disorders. Sem Reprod Endocrinol 1994;15(1):91–100. [11] Matthews KA, Wing RR, Kuller LH, et al. Influences of natural menopause on psychological characteristics and symptoms of middle-aged healthy women. J Consult Clin Psychol 1990;58(June (3)):345–51. [12] Ross LA, Alder EM, Cawood EH, Brown J, Gebbie AE. Psychological effects of hormone replacement therapy: a comparison of tibolone and a sequential estrogen therapy. J Psychosom Obstet Gynaecol 1999;20(June (2)):88–96. [13] Joffe H, Hall JE, Soares CN, et al. Vasomotor symptoms are asociated with depression in perimenopausal women seeking primary care. Menopause 2002;9(6):392–8. [14] Ballinger CB. Psychiatric aspects of the menopause. Br J Psychiatry 1990;156(June):773–87. [15] Soares CN, Cohen LS. The perimenopause,depressive disorders, and hormonal variability. Sao Paulo Med J 2001; 119(March (2)):78–83.