Archives of Sexual Behavior, VoL 19, No. 3, 1990
Changes in Women's Sexual Desire in Middle Life: The Longitudinal Study of Women in Gothenburg T o r e Hfillstr6m, M.D., Ph.D. 1 and Sverker S a m u e l s s o n , M.D., Ph.D. 1
In a general population survey, 677 urban middle-aged women were interviewed about their sexual desire at two occasions 6years apart. Data from 497 subjects, who were married at both occasions or were cohabiting with a male partner, were analyzed. Twenty-seven percent reported a decrease in sexual desire between interviews, and 10°7o experienced an increased desire. There were no clear cohort differences. A decrease in sexual desire was predicted by age, high sexual desire at first interview, lack of a confiding relationship, insufficient support from spouse, alcoholism in spouse, and major depression. Predictors of an increase o f sexual desire were weak desire at first interview, negative marital relations before first interview, and mental disorder at first interview. Although sexual desire showed considerable stability over time, a substantial proportion o f married middle-aged women experienced major changes, mostly as a decrease. Age, psychosocial factors associated with quality of marital relationship, and mental health were major contributors towards change in sexual desire. KEY WORDS: sexual dysfunction; sexual desire; women; middle life; longitudinal study; marital relations; life events; mental disorder; sociodemographic factors.
INTRODUCTION
M o s t studies in the field o f age a n d sexuality have been cross-sectional ( P f e i f f e r etal., 1972; H/illstr6m, 1973, 1979; Kivel/i etal., 1986). T h e y show t h a t the intensity o f sexual interest o r desire tends to decline with a d v a n c i n g
This study was supported by grants from the Swedish Medical Research Council (27X-4578). ~Department of Psychiatry, Sahlgren Hospital, University of Gothenburg, S-413 45 Gothenburg, Sweden. 259 0004-0002/90/0600-0259506.00/0 © 1990 Plenum Publishing Corporation
260
H~ilistr6m and Samuelsson
age. However, in these studies it is not possible to differentiate age effects from cohort differences. Another problem with most cross-sectional investigations is that marital status has not been taken into account in the analyses of age influence on sexual desire. In the analysis of longitudinal data from the first and second Duke study it was not possible to demonstrate an age effect on sexual interest among women (Pfeiffer et al., 1969; George and Weiler, 1981). Age at entry into the study was, however, associated with the degree of sexual interest, indicating a cohort effect. The present investigation is based on data from the study of women in Gothenburg (Bengtsson et al., 1973). This paper's aim is to describe changes in sexual desire in women of late middle life and to examine their associations with age, sociodemographic factors, the quality of the marital relationship, mental disorder, and life events.
METHOD Subjects
Middle-aged women were sampled systematically from the general population of women in Gothenburg, Sweden, an industrial city of 450,000 inhabitants. A representative subsample of women aged 38, 46, 50, and 54 years, born in 1930, 1922, 1918, and 1914, respectively, was selected for a psychiatric study. Of the 899 selected women, 800 (89.0O7o) participated in the first study in 1968-1969. Further details concerning the method of sampling have been provided (H~illstr6m, 1973; Bengtsson et al., 1973). There were no significant differences between the participants and nonparticipants concerning social class or contact with mental health services. The women were reexamined 6 years later (Samuelsson, 1982). Participation rates are described in Table I. Only those 497 women who lived with their spouse or other partner at both interviews were included in the analyses.
Procedure
In each cross-sectional study the subjects were interviewed by the same psychiatrist in a semistructured interview. The ratings from the first interview were not known to the interviewer in the second study. Using data from the study in 1968-1969 as background variables made the study prospective. Our hypotheses were formulated according to the findings of the first crosssectional investigation (H~illstr6m, 1973, 1979).
261
Changes in Women's Sexual Desire Table I. Participants and Participation Rates by Age First study, 1968-1969 Age (years)
Sampled n
Second study, 1974-1975, participants
Participants n
Age (years)
n
% of participants 1968-1969
sampled 1968
% of
38
124
111
44
86
77.5
69.4
46
346
309
52
264
85.4
76.3
50
323
290
56
248
85.5
76.8
54
106
90
60
79
87.8
74.5
Totals
899
800
677
84.6
75.3
The following sociodemographic variables were rated in the first interview: education, social class (1-4 according to husband's occupation), family income, age difference between spouses, work outside the home (none, part-time, full-time), and child(ren) at home (no, yes). Work outside the home was also rated in the second interview. In the interviews a sexual history was taken. In this report only the womens' sexual desire (interest) is taken into consideration. In both examinations the participants were asked about the present degree of their desire, that is, if they perceived it as strong, moderate, weak, or lacking. The emotional quality of the marital relationship was rated according to the following variables: insufficient emotional support from spouse (no, yes); development of marital relation (positive, no change, negative); lack of confiding relationship (no, yes); husband alcoholic, addictive type (no, yes). The grade of current mental disorder was rated in the interviews using the following scale: 0 = no impairment; 1 = mild symptoms, no functional impairment; 2 = moderate symptoms, clear functional impairment; 3 = marked symptoms, work capacity greatly reduced; 4 = severe symptoms, complete incapacity for work. Maximum grade of mental disorder between interviews was rated at the second interview according to the same scale. Duration of mental disorder (Grades 1-4) was rated in the second interview (none, < 4 months, < 1 year, 1-3 years, > 3 years). Major depression according to DSM-III (American Psychiatric Association, 1980) was diagnosed on data from the first interview (H~illstr6m, 1984). In the second interview the ICD-8 (World Health Organization, 1965) diagnoses were used. In this paper the three most common psychiatric diagnoses in the sample have been employed (depressive neurosis, anxiety neurosis, and neurasthenia). In the second interview, symptom score sum was calculated from current symptom ratings (reported items) according to the Comprehensive Psychopathological Rating Scale (CPRS; Asberg et al., 1978)
262
Hiillstr6m and S a m u e l s s o n
using the following variables: sadness, inner tension, hostile feelings, inability to feel, pessimistic thoughts, suicidal thoughts, hypochondria, worrying over trifles, compulsive thoughts, phobias, rituals, indecision, lassitude, fatiguability, concentration difficulties, failing memory, reduced sleep, autonomic disturbances, muscular tension, derealization/depersonalization, delusional ideas, and hallucinations. Current use of psychotropic medication was assessed at both interviews (none, up to once/month, up to once/week, less than once daily, daily). Participants were asked about the occurrence of life events during a 1-year period before the second interview. Fifty-two undesirable life events were defined before interview. Life events were weighted by a panel of 126 women sampled from the same area and in the same age groups. These were asked to rate the amount of emotional stress they would experience from each event using a visual analog scale with end points "no stress" and "worst imaginable stress." This gave a ranking and scores were then standardized in relation to the most stressful item (Samuelsson, 1982). Statistical methods used were Pitman's permutation tests for single or paired observations (Bradley, 1968). Age effects were eliminated by the method of Mantel (1963). The level of significance used was p < 0.05. Twotailed tests were used throughout. RESULTS
Overall Changes in Sexual Desire Table II shows the reported degree of sexual desire at the two investigations. Of those 15 subjects who reported a strong sexual desire at the first interview, all but one reported a lower level at the second interview. Conversely, half of those who reported initially that their sexual desire was absent regained a weak or moderately strong desire. Of those 461 women Table II. Sexual Desire at First a n d Second I n t e r v i e w S e c o n d interview First
Strong
Moderate
Weak
Absent
interview
n
o70
n
%
n
o70
n
%
Strong
1
7
11
73
2
13
1
7
Total n 15
Moderate
3
1
231
70
81
25
15
5
330 116
Weak
-
--
30
26
63
54
23
20
Absent
--
--
5
14
13
36
18
50
36
Total
4
1
277
56
159
32
57
11
497
Changes in Women's Sexual Desire
263
Table IlL Sexual Desire by Age Cohort, interview, and age in years a 1930 First 38
1930 Second 44
1922 First 46
1918 First 50
1922 Second 52
3 (4)
4 (6)
10 (5)
2 (1)
.
Moderate
51 (74)
57 (83)
139 (72)
114 (62)
120 (63)
26 (50)
81 (44)
19 (37)
Weak
12 (17)
8 (12)
37 (19)
52 (28)
59 (31)
15 (29)
74 (40)
18 (35)
Absent
3 (4)
--
6 (3)
16 (9)
13 (7)
11 (21)
29 (16)
15 (29)
Total
69
69
192
184
192
52
184
52
Sexual desire Strong
1914 First 54 .
1918 Second 56
.
1914 Second 60
.
"Percentages in parentheses.
with a strong, moderate, or weak sexual desire at first interview, 39 (8.5%) reported absence o f desire in the second interview. Sexual desire in the four birth cohorts are presented in Table III. After the age o f 50 no subject was aware of a strong sexual desire. From this age increasing proportions o f w o m e n report little or no sexual interest. N o clear cohort effects emerged. The prevalence o f hypoactive sexual desire disorder defined as absent sexual desire was 3°70 in ages 38-46, 9% at age 50, and 29070 at age 60. The perceived changes o f sexual desire in the 6 years between the interviews are shown in Table IV. Subjects w h o reported the same intesity of sexTable IV. Change of Sexual Desire by Age Age period (years) 38-44
46-52
~o
n
1
0.5
-
Moderately reduced
8
4
8
4
1
2
17
3.4
Slightly reduced
n
%
%
Total
54-60
Greatly reduced
Sexual desire
n
50-56
n
%
-
n 1
-% 0.2
5
7
42
22
50
27
18
35
115
23. t
Unchanged
50
72
127
66
110
60
26
50
313
63.0
Slightly increased
12
17
13
7
16
9
5
10
46
9.3
2
3
1
0.5
-
2
4
5
1.0
Moderately increased Greatly increased Total
69
192
184
52
497
Permutation test for paired observations
ns
p
Changes in Women's Sexual Desire in Middle Life: The Longitudinal Study of Women in Gothenburg T o r e Hfillstr6m, M.D., Ph.D. 1 and Sverker S a m u e l s s o n , M.D., Ph.D. 1
In a general population survey, 677 urban middle-aged women were interviewed about their sexual desire at two occasions 6years apart. Data from 497 subjects, who were married at both occasions or were cohabiting with a male partner, were analyzed. Twenty-seven percent reported a decrease in sexual desire between interviews, and 10°7o experienced an increased desire. There were no clear cohort differences. A decrease in sexual desire was predicted by age, high sexual desire at first interview, lack of a confiding relationship, insufficient support from spouse, alcoholism in spouse, and major depression. Predictors of an increase o f sexual desire were weak desire at first interview, negative marital relations before first interview, and mental disorder at first interview. Although sexual desire showed considerable stability over time, a substantial proportion o f married middle-aged women experienced major changes, mostly as a decrease. Age, psychosocial factors associated with quality of marital relationship, and mental health were major contributors towards change in sexual desire. KEY WORDS: sexual dysfunction; sexual desire; women; middle life; longitudinal study; marital relations; life events; mental disorder; sociodemographic factors.
INTRODUCTION
M o s t studies in the field o f age a n d sexuality have been cross-sectional ( P f e i f f e r etal., 1972; H/illstr6m, 1973, 1979; Kivel/i etal., 1986). T h e y show t h a t the intensity o f sexual interest o r desire tends to decline with a d v a n c i n g
This study was supported by grants from the Swedish Medical Research Council (27X-4578). ~Department of Psychiatry, Sahlgren Hospital, University of Gothenburg, S-413 45 Gothenburg, Sweden. 259 0004-0002/90/0600-0259506.00/0 © 1990 Plenum Publishing Corporation
260
H~ilistr6m and Samuelsson
age. However, in these studies it is not possible to differentiate age effects from cohort differences. Another problem with most cross-sectional investigations is that marital status has not been taken into account in the analyses of age influence on sexual desire. In the analysis of longitudinal data from the first and second Duke study it was not possible to demonstrate an age effect on sexual interest among women (Pfeiffer et al., 1969; George and Weiler, 1981). Age at entry into the study was, however, associated with the degree of sexual interest, indicating a cohort effect. The present investigation is based on data from the study of women in Gothenburg (Bengtsson et al., 1973). This paper's aim is to describe changes in sexual desire in women of late middle life and to examine their associations with age, sociodemographic factors, the quality of the marital relationship, mental disorder, and life events.
METHOD Subjects
Middle-aged women were sampled systematically from the general population of women in Gothenburg, Sweden, an industrial city of 450,000 inhabitants. A representative subsample of women aged 38, 46, 50, and 54 years, born in 1930, 1922, 1918, and 1914, respectively, was selected for a psychiatric study. Of the 899 selected women, 800 (89.0O7o) participated in the first study in 1968-1969. Further details concerning the method of sampling have been provided (H~illstr6m, 1973; Bengtsson et al., 1973). There were no significant differences between the participants and nonparticipants concerning social class or contact with mental health services. The women were reexamined 6 years later (Samuelsson, 1982). Participation rates are described in Table I. Only those 497 women who lived with their spouse or other partner at both interviews were included in the analyses.
Procedure
In each cross-sectional study the subjects were interviewed by the same psychiatrist in a semistructured interview. The ratings from the first interview were not known to the interviewer in the second study. Using data from the study in 1968-1969 as background variables made the study prospective. Our hypotheses were formulated according to the findings of the first crosssectional investigation (H~illstr6m, 1973, 1979).
261
Changes in Women's Sexual Desire Table I. Participants and Participation Rates by Age First study, 1968-1969 Age (years)
Sampled n
Second study, 1974-1975, participants
Participants n
Age (years)
n
% of participants 1968-1969
sampled 1968
% of
38
124
111
44
86
77.5
69.4
46
346
309
52
264
85.4
76.3
50
323
290
56
248
85.5
76.8
54
106
90
60
79
87.8
74.5
Totals
899
800
677
84.6
75.3
The following sociodemographic variables were rated in the first interview: education, social class (1-4 according to husband's occupation), family income, age difference between spouses, work outside the home (none, part-time, full-time), and child(ren) at home (no, yes). Work outside the home was also rated in the second interview. In the interviews a sexual history was taken. In this report only the womens' sexual desire (interest) is taken into consideration. In both examinations the participants were asked about the present degree of their desire, that is, if they perceived it as strong, moderate, weak, or lacking. The emotional quality of the marital relationship was rated according to the following variables: insufficient emotional support from spouse (no, yes); development of marital relation (positive, no change, negative); lack of confiding relationship (no, yes); husband alcoholic, addictive type (no, yes). The grade of current mental disorder was rated in the interviews using the following scale: 0 = no impairment; 1 = mild symptoms, no functional impairment; 2 = moderate symptoms, clear functional impairment; 3 = marked symptoms, work capacity greatly reduced; 4 = severe symptoms, complete incapacity for work. Maximum grade of mental disorder between interviews was rated at the second interview according to the same scale. Duration of mental disorder (Grades 1-4) was rated in the second interview (none, < 4 months, < 1 year, 1-3 years, > 3 years). Major depression according to DSM-III (American Psychiatric Association, 1980) was diagnosed on data from the first interview (H~illstr6m, 1984). In the second interview the ICD-8 (World Health Organization, 1965) diagnoses were used. In this paper the three most common psychiatric diagnoses in the sample have been employed (depressive neurosis, anxiety neurosis, and neurasthenia). In the second interview, symptom score sum was calculated from current symptom ratings (reported items) according to the Comprehensive Psychopathological Rating Scale (CPRS; Asberg et al., 1978)
262
Hiillstr6m and S a m u e l s s o n
using the following variables: sadness, inner tension, hostile feelings, inability to feel, pessimistic thoughts, suicidal thoughts, hypochondria, worrying over trifles, compulsive thoughts, phobias, rituals, indecision, lassitude, fatiguability, concentration difficulties, failing memory, reduced sleep, autonomic disturbances, muscular tension, derealization/depersonalization, delusional ideas, and hallucinations. Current use of psychotropic medication was assessed at both interviews (none, up to once/month, up to once/week, less than once daily, daily). Participants were asked about the occurrence of life events during a 1-year period before the second interview. Fifty-two undesirable life events were defined before interview. Life events were weighted by a panel of 126 women sampled from the same area and in the same age groups. These were asked to rate the amount of emotional stress they would experience from each event using a visual analog scale with end points "no stress" and "worst imaginable stress." This gave a ranking and scores were then standardized in relation to the most stressful item (Samuelsson, 1982). Statistical methods used were Pitman's permutation tests for single or paired observations (Bradley, 1968). Age effects were eliminated by the method of Mantel (1963). The level of significance used was p < 0.05. Twotailed tests were used throughout. RESULTS
Overall Changes in Sexual Desire Table II shows the reported degree of sexual desire at the two investigations. Of those 15 subjects who reported a strong sexual desire at the first interview, all but one reported a lower level at the second interview. Conversely, half of those who reported initially that their sexual desire was absent regained a weak or moderately strong desire. Of those 461 women Table II. Sexual Desire at First a n d Second I n t e r v i e w S e c o n d interview First
Strong
Moderate
Weak
Absent
interview
n
o70
n
%
n
o70
n
%
Strong
1
7
11
73
2
13
1
7
Total n 15
Moderate
3
1
231
70
81
25
15
5
330 116
Weak
-
--
30
26
63
54
23
20
Absent
--
--
5
14
13
36
18
50
36
Total
4
1
277
56
159
32
57
11
497
Changes in Women's Sexual Desire
263
Table IlL Sexual Desire by Age Cohort, interview, and age in years a 1930 First 38
1930 Second 44
1922 First 46
1918 First 50
1922 Second 52
3 (4)
4 (6)
10 (5)
2 (1)
.
Moderate
51 (74)
57 (83)
139 (72)
114 (62)
120 (63)
26 (50)
81 (44)
19 (37)
Weak
12 (17)
8 (12)
37 (19)
52 (28)
59 (31)
15 (29)
74 (40)
18 (35)
Absent
3 (4)
--
6 (3)
16 (9)
13 (7)
11 (21)
29 (16)
15 (29)
Total
69
69
192
184
192
52
184
52
Sexual desire Strong
1914 First 54 .
1918 Second 56
.
1914 Second 60
.
"Percentages in parentheses.
with a strong, moderate, or weak sexual desire at first interview, 39 (8.5%) reported absence o f desire in the second interview. Sexual desire in the four birth cohorts are presented in Table III. After the age o f 50 no subject was aware of a strong sexual desire. From this age increasing proportions o f w o m e n report little or no sexual interest. N o clear cohort effects emerged. The prevalence o f hypoactive sexual desire disorder defined as absent sexual desire was 3°70 in ages 38-46, 9% at age 50, and 29070 at age 60. The perceived changes o f sexual desire in the 6 years between the interviews are shown in Table IV. Subjects w h o reported the same intesity of sexTable IV. Change of Sexual Desire by Age Age period (years) 38-44
46-52
~o
n
1
0.5
-
Moderately reduced
8
4
8
4
1
2
17
3.4
Slightly reduced
n
%
%
Total
54-60
Greatly reduced
Sexual desire
n
50-56
n
%
-
n 1
-% 0.2
5
7
42
22
50
27
18
35
115
23. t
Unchanged
50
72
127
66
110
60
26
50
313
63.0
Slightly increased
12
17
13
7
16
9
5
10
46
9.3
2
3
1
0.5
-
2
4
5
1.0
Moderately increased Greatly increased Total
69
192
184
52
497
Permutation test for paired observations
ns
p
