- Email: [email protected]
Climacteric symptoms and sleep quality1
Climacteric Symptoms and Sleep Quality ¨ IVI POLO-KANTOLA, MD, RISTO ERKKOLA, MD, PhD, KERTTU IRJALA, MD, PhD, PA HANS HELENIUS, MSc, SIRKKU PULLINEN, RN, AND OLLI POLO, MD, PhD Objective: To evaluate the effect of climacteric vasomotor symptoms on sleep quality measured by self-report and polysomnography in postmenopausal women. Methods: Seventy-one healthy postmenopausal women were recruited, of whom 63 completed the study. Each subject recorded climacteric symptoms and subjective sleep quality for 14 days. Sleep quality was evaluated objectively by all-night polysomnography using the static charge– sensitive bed. Results: During polysomnography, a high frequency of climacteric vasomotor symptoms was not associated with changes in sleep latency, percentage of sleep stages, number of arousals, sleep efficiency, or total sleep time. However, a high frequency of climacteric vasomotor symptoms (range 0 – 8.9, r ⴝ .60, P < .001), somatic symptoms (range 0 –5.0, r ⴝ .25–.44, P < .05), and mental symptoms (range 0 –5.0, r ⴝ .41–.51, P < .001) was related to impaired subjective sleep quality. In stepwise regression analysis, 32% of the impairment in subjective sleep quality was explained by vasomotor symptoms (P < .001), 14% by palpitations (P < .001), and 4% by mood instability (P ⴝ .029). High body mass index predicted impaired objective sleep quality, such as prolonged latencies to stage-2 sleep (r ⴝ .27, P ⴝ .031) and slow-wave sleep (r ⴝ .51, P ⴝ .003) and decreased oxygen saturations (r ⴝ ⴚ .54, P < .001). Older women had decreased sleep efficiency (r ⴝ ⴚ .27, P ⴝ .030) and lower oxygen saturations (r ⴝ ⴚ .36, P ⴝ .004). Serum estradiol level had only a minor effect on objective sleep quality. Conclusion: Impaired subjective sleep quality associated with climacteric vasomotor symptoms did not manifest as abnormalities in polysomnographic sleep recordings. Body mass index and age appeared to have the strongest effect on objective sleep quality. (Obstet Gynecol 1999;94:219 –24. © 1999 by The American College of Obstetricians and Gynecologists.)
From the Departments of Obstetrics and Gynecology and Clinical Chemistry, Turku University Central Hospital; the Departments of Physiology and Biostatistics, University of Turku, Turku; and the Department of Pulmonary Medicine, University of Tampere, Tampere, Finland. Supported by research grants from the Medical Research Council of the Academy of Finland and the Gynecological Association of Finland. We thank Anne Kaljonen, MSc, for statistical assistance.
VOL. 94, NO. 2, AUGUST 1999
Sleep complaints, especially frequent nocturnal awakenings, increase with age.1 Elderly women suffer more often from sleep problems and use more hypnotics than elderly men.2 Somatic disorders such as chronic bronchitis, asthma, and cardiovascular symptoms affect sleep quality.3 Specific sleep disorders, including sleep apnea, cause insomnia or daytime sleepiness.4 Psychological factors also disturb sleep.5 It has been suggested that the decline of endogenous estrogen has an effect on sleep because sleep complaints increase rapidly after menopause.6 In addition to vasomotor symptoms, poor sleep quality is one of the most common symptoms of menopause. Climacteric vasomotor symptoms typically occur at night and interfere with sleep.7 We hypothesized that severe vasomotor symptoms are related to significant changes in the cortical electroencephalographic activity during sleep. However, sleep disturbances due to climacteric symptoms are poorly described in polysomnographic terms. To determine the type and the degree of specific electroencephalogram changes associated with vasomotor symptoms, we conducted all-night polysomnographic sleep studies in postmenopausal women who had a wide range of climacteric symptoms.
Materials and Methods Seventy-one healthy postmenopausal women were recruited over 15 months through a newspaper announcement, and 63 women completed the study. The age range was 47– 65 years (mean ⫾ standard deviation [SD] 56.3 ⫾ 4.4) and the minimum education was 7 years of basic school. Sixty-seven percent were married, 6% were single, and 27% were divorced or widowed. Forty-six percent were employed, 10% were unemployed, and 44% were retired. Postmenopause was defined as a serum FSH level exceeding 30 IU/L. Two women with FSH levels of 28 and 29 IU/L were accepted because of their ages (56 and 62 years). At the beginning of the study, the serum estradiol (E2) level was less than 50 pmol/L in all women except two,
0029-7844/99/$20.00 PII S0029-7844(99)00284-7
219
whose levels were 90 and 105 pmol/L, but whose serum FSH levels were postmenopausal (70 and 55 IU/L, respectively). Exclusion criteria included history of neurologic, cardiovascular, endocrinologic, or mental disease; medicated hyperlipidemia; malignancies; abuse of alcohol or medications; and heavy smoking (more than ten cigarettes per day). Latent systemic disorders possibly affecting sleep such as anemia, hypothyroidism, diabetes, and uremia were also excluded by measuring blood hemoglobin, leukocytes, sedimentation rate, serum thyrotropin, free thyroxine, vitamin B12, creatinine, glucose, and cholesterol levels. All subjects were strictly advised not to use any medications that affect the central nervous system, or any antioxidants. Incipient depression was screened using the Beck Depression Inventory.8 A complete gynecologic examination was done. Subjects were given oral and written information about the study, and they all provided written informed consent. The study was approved by the Ethics Committee of Turku University and Turku University Central Hospital. Polysomnographic sleep recording included continuous monitoring of electroencephalogram, electrooculogram, electromyogram, and electrocardiogram. A finger pulse oximeter (Biox Ohmeda 3700e; BOC Company, Louisville, CO) was used to monitor arterial oxyhemoglobin saturation. A static charge–sensitive bed (BioMatt; Biorec Oy, Helsinki, Finland)9 was used to record body movements, breathing patterns, and heart rates. Original analog signals were amplified and digitized at a frequency of 250 samples per second with 12-bit amplitude resolution and were recorded with custom software (UniPlot, Unesta Oy, Turku, Finland). Conventional criteria10 were used for sleep staging, which included stage-1, stage-2, slow-wave (including stage-3 and stage-4), and rapid eye movement sleep. Staging was carried out by the same scorer in all subjects. The scorer was masked to the presence of climacteric symptoms. Episodes of alpha arousals and body movements during sleep also were analyzed. Their frequencies were expressed as the number of episodes per hour of sleep. For all arousals, electroencephalogram alpha activity of at least 2 seconds was required. Arousal movement events were further classified with the static charge–sensitive bed as movement arousals (arousals with body movement) and electroencephalogram arousals (arousals without body movement). Highamplitude signals in the static charge–sensitive bed without simultaneous electroencephalogram alpha activity were considered body movements. This study was part of a larger survey evaluating the effect of estrogen replacement therapy (ERT) on sleep.11 In the original study design, randomization into two
220 Polo-Kantola et al
Women and Sleep
treatment groups was made in six-person blocks using random permuted blocks. Thirty-two women first received the placebo for 3 months and then, after 1 month washout with the placebo, estrogen for 3 months (group A). The remaining 31 women were treated in the reverse order (group B). Sleep studies were done twice in each subject at 4-month intervals, after placebo and after estrogen treatment. Data for the present report were collected using the placebo nights only. A statistically significant carryover effect was observed for three variables: latency to slow-wave sleep, percentage of stage-1 sleep, and electroencephalogram arousals in non–rapid eye movement sleep. For these variables, data were included only from group A. All the data were also analyzed with the 32 women in group A. The subjects self-reported daily climacteric symptoms for 14 days before the sleep study: vasomotor symptoms (hot flushes and sweating separately), sleep complaints (sleeping problems, insomnia), somatic symptoms (palpitations, numbness, muscular pain, dizziness, headache, fatigue), and mental symptoms (anxiety, depression, mood instability, memory problems, and lack of initiative). The intensity of the symptoms was evaluated on a six-step scale: 0 (“not at all”), 1 (“very little”), 2 (“to some extent”), 3 (“moderately”), 4 (“much”), 5 (“very much”), and 6 (“I cannot say”). To get an overall intensity score, the 14 daily scores were averaged. Some women express their vasomotor symptoms as hot flushes, whereas others feel incidental sweating. Therefore, the sum score of hot flushes and sweating on a scale from 0 to 10 was used to describe the severity of the vasomotor symptoms. For sleep complaints, somatic symptoms and mental symptoms scores ranged from 0 to 5. The few level-6 responses (“I cannot say”) were excluded from the analyses. Spearman correlation coefficient was used in the analysis of correlations. Because subjective symptoms were interdependent, stepwise regression analysis was used to study the connections among subjective sleep quality, climacteric vasomotor symptoms, and other climacteric symptoms. P ⬍ .05 was considered statistically significant. The power of our population to show the correlation r ⱖ .40 was greater than 90%. Statistical calculations were done using the SAS program package (SAS Institute, Cary, NC).
Results The mean (⫾ SD) body mass index (BMI) was 26.9 ⫾ 4.0 kg/m2 (range 19.8 –39.0), and the mean Beck Depression score was 4.0 ⫾ 3.8 (range 0 –15). In the morning after the sleep study, the mean serum E2 and serum FSH levels were 28.8 ⫾ 20.5 pmol/L and 70.4 ⫾ 22.7 IU/L, respectively. The subjects had not used ERT for at
Obstetrics & Gynecology
Table 1. Correlations Between Subjective Sleep Quality and Climacteric Symptoms Symptoms Vasomotor Hot flushes Sweating Sum score Somatic Palpitations Numbness Muscular pain Dizziness Headache Fatigue Mental Anxiety Depression Mood instability Memory problems Lack of initiative
Median
Range
r
P
1.0 1.3 2.6
0 – 4.3 0 – 4.7 0 – 8.9
.53 .58 .60
⬍.001 ⬍.001 ⬍.001
0.1 0.4 1.5 0.0 0.6 0.5
0 –3.8 0 – 4.0 0 –5.0 0 – 4.3 0 –3.1 0 – 4.0
.44 .30 .22 .21 .26 .25
⬍.001 .017 .079 .100 .038 .047
0.1 0.0 0.1 0.4 0.2
0 – 4.5 0 – 4.5 0 –5.0 0 – 4.0 0 – 4.0
.44 .51 .41 .44 .44
⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001
least 4 months before the sleep study. Before entering the study, 47 women (75%) had taken ERT previously, with a mean duration of 41 ⫾ 55 months (range 1–252). The mean time off ERT was 37 ⫾ 55 months (range 5–226). Each woman had had a hysterectomy for benign indications. Twenty-four (38%) had had oophorectomy, unilateral in nine (14%) and bilateral in 15 (24%). The median score for climacteric vasomotor symptoms (hot flushes plus sweating; scale 0 –10) was 2.6 ⫾ 2.5 (range 0 – 8.9), and the median score for sleep complaints (scale 0 –5) was 1.5 ⫾ 1.1 (range 0 –5.0). Women with subjectively impaired sleep experienced more climacteric vasomotor symptoms (r ⫽ .60, P ⬍ .001). They also had more climacteric somatic symptoms, such as palpitations, numbness, headache, and fatigue. More climacteric mental symptoms, such as anxiety, depression, mood instability, memory problems, or lack of initiative, were also present (Table 1). Climacteric symptoms were interdependent, so a stepwise regression analysis was performed. Thirty-two percent of the impairment in subjective sleep quality was explained by vasomotor symptoms (P ⬍ .001), 14% by palpitations (P ⬍ .001), and 4% by mood instability (P ⫽ .029). Body mass index, age, serum E2 level, or FSH level did not correlate with subjective sleep quality. Climacteric vasomotor symptoms were not related to sleep latency, latencies to sleep stages, percentages of sleep stages, frequencies of short (less than 1 minute) or long (longer than 1 minute) awakenings, number of arousals, number of body movements, number of arousals or body movements in various sleep stages, sleep efficiency (total sleep time/sleep-period time [awakening latency ⫺ sleep latency] ⫻ 100%), or total
VOL. 94, NO. 2, AUGUST 1999
Table 2. Correlations Between Vasomotor Symptoms (Sum Scores) and Sleep Variables Variable
Median
Range
r
P
Total sleep time (min) Sleep efficiency (%) Sleep latency (min) Latency (min) To stage 2 To stages 3– 4 To REM % Stage 0 1 2 3– 4 % REM % Movement time % Time out of bed SaO2 (%) Minimum (lowest) Maximum (highest) Mean
441.6 93.4 3.6
309.0 –511.2 74.4 –98.6 0.0 – 61.8
⫺.002 .11 .005
.989 .389 .969
3.6 14.4 76.8
0.0 – 67.8 4.8 –74.4 22.2–249.0
.21 ⫺.10 .13
.102 .605* .297
6.7 15.4 44.5 9.6 20.6 0.0 0.0
0.0 –25.6 7.2–28.0 25.8 – 61.2 1.2–23.7 9.2–31.5 0.0 –1.0 0.0 –1.5
⫺.07 ⫺.32 .04 .08 .10 ⫺.14 .20
.572 .072* .778 .518 .414 .286 .108
89.0 98.0 94.3
58.0 –92.0 96.0 –100.0 91.4 –96.1
.11 .14 .11
.411 .270 .407
REM ⫽ rapid eye movement; SaO2 ⫽ arterial oxyhemoglobin saturation. * A significant carryover effect in the basic study, data for calculation is drafted from the first recording night (n ⫽ 32).
sleep time (awakening latency ⫺ sleep latency ⫺ amount of stage-0 sleep). Vasomotor symptoms were not associated with the mean, minimum, or maximum arterial oxyhemoglobin saturations. (Tables 2– 4). To evaluate the relation between climacteric symptoms and the objective sleep variables, we analyzed more than 550 correlations. Just by repeating statistical tests, one would expect 29 significant correlations by chance. However, only six correlations between climacteric somatic symptoms and objective sleep variables were found, with r values from .26 to ⫺.45 (P ⬍ .05). These included more palpitations with longer rapid eye movement latency, more muscular pain with less stage-1 sleep and more rapid eye movement sleep, more dizziness with fewer movement arousals, more headache with less stage-1 sleep, and more fatigue with less stage-1 sleep. Seven correlations between climacteric
Table 3. Correlations Between Vasomotor Symptoms (Sum Scores) and Frequency of Awakenings Symptom (per hour of sleep)
Median
Range
r
P
Movement arousals No. awakenings ⬍1 min (short) No. awakenings ⬎1 min (long) EEG arousals Body movements
8.1 2.4 1.2 3.2 7.5
1.3–22.6 1.2–5.3 0.3–3.4 0.2–12.4 0.6 – 47.8
.05 ⫺.03 ⫺.04 ⫺.03 ⫺.01
.690 .802 .769 .851* .942
EEG ⫽ electroencephalogram. * A significant carryover effect in the basic study; calculation is drafted from the first recording night (n ⫽ 32).
Polo-Kantola et al
Women and Sleep
221
Table 4. Correlations Between Vasomotor Symptoms (Sum Scores) and Awakenings in Various Sleep Stages (per Hour of Sleep) Stage Stage 1 Movement arousals EEG arousals Body movements Stage 2 Movement arousals EEG arousals Body movements Stage 3– 4 Movement arousals EEG arousals Body movements REM Movement arousals EEG arousals Body movements
Median
Range
r
23.7 12.2 10.8
2.5–74.4 0.0 –36.7 0.0 – 43.7
4.2 3.1 6.3
0.0 –14.0 0.0 –20.8 0.0 –50.2
0.0 0.0 1.7
0.0 –10.6 0.0 – 4.4 0.0 – 47.0
.10 ⫺.02 ⫺.04
.432 .864 .752
5.7 1.1 8.2
0.5–28.0 0.0 –10.5 0.0 –39.0
.07 .11 ⫺.02
.589 .392 .875
.04 ⫺.10 ⫺.04 .12 .002 .002
P .753 .448 .785 .337 .986 .987
EEG ⫽ electroencephalogram; REM ⫽ rapid eye movement.
mental symptoms and objective sleep variables were observed, with r values from ⫺.26 to ⫺.56 (P ⬍ .05). These included more anxiety with less stage-1 sleep, more depression with less stage-1 sleep, more mood instability with less stage-1 sleep and with more rapid eye movement sleep, more memory problems with less stage-1 sleep and with more rapid eye movement sleep, and more lack of initiative with less stage-1 sleep. Women with higher BMI had a longer latency to stage-2 sleep (r ⫽ .27, P ⫽ .031) and to slow-wave sleep (r ⫽ .51, P ⫽ .003). High BMI was related to low mean arterial oxyhemoglobin saturation (r ⫽ ⫺.54, P ⬍ .001) and minimum arterial oxyhemoglobin saturation (r ⫽ ⫺.37, P ⫽ .003). Older women spent more time awake during the study night (r ⫽ .30, P ⫽ .019), and their sleep efficiency was decreased (r ⫽ ⫺.27, P ⫽ .030). They also had more body movements in stage-1 sleep (r ⫽ .34, P ⫽ .007), lower mean arterial oxyhemoglobin saturation (r ⫽ ⫺.36, P ⫽ .004), and lower minimum arterial oxyhemoglobin saturation values (r ⫽ ⫺.25, P ⫽ .047). Serum E2 correlated with only one objective sleep measurement: Low estrogen concentration predicted a high frequency of movement arousals in slow-wave sleep (r ⫽ ⫺.39, P ⫽ .002). Serum FSH concentrations did not correlate with objective sleep measurements. The number of arousals per hour of sleep through various sleep stages is shown in Table 4. Movement arousals (movements with alpha activity) and electroencephalogram arousals (alpha activity without movements) were most common in stage-1 sleep but were extremely rare in slow-wave sleep. The frequencies of body movements (movements without alpha activity) were similar in stage-1, stage-2, and rapid eye move-
222 Polo-Kantola et al
Women and Sleep
ment sleep, but in slow-wave sleep, the frequency of body movements was low. Women with frequent vasomotor symptoms had a sleep architecture (sleep-stage distribution) similar to those with a low frequency of symptoms. The data also were analyzed with the group of 32 women who received placebo first in the original study, with substantially the same results as for the entire 63-woman group.
Discussion The questionnaires indicated impaired subjective sleep quality in women with a high frequency of climacteric vasomotor symptoms. Somatic and mental climacteric symptoms were also more common in women with frequent sleep complaints. In contrast to the selfreported findings, polysomnography did not show any specific impairment of sleep architecture, sleep efficiency, or frequency of arousals induced by or related to climacteric vasomotor symptoms. The discrepancy between subjective and objective sleep quality was even more evident for climacteric mental symptoms, which impaired subjectively but improved objectively measured sleep. Serum E2 levels did not have a major effect. Age and obesity, rather than climacteric symptoms, appeared to have the strongest impact on polygraphic sleep variables. Sleep complaints and insomnia have been connected to menopause, especially to climacteric vasomotor symptoms. This connection is supported by a reversal of sleep complaints during ERT.11 Our results agree with these findings. Few studies have objectively measured sleep after menopause,7,12–15 and most of these studies focused on the effect of ERT on sleep quality, relating climacteric vasomotor symptoms to objective sleep quality. Erlik et al7 reported an association between hot flushes and waking episodes. Unfortunately, we did not monitor hot flushes or skin temperature during the sleep-recording nights. It is possible that the subjective scoring of climacteric symptoms reported during the 14 days before the sleep study could not predict objective vasomotor symptoms during the laboratory night. Hot flushes and sweating do arouse women, especially from light sleep (stages 1 and 2).7 Most of the arousals occurred from light sleep, but factors other than vasomotor symptoms, such as BMI, seemed to explain better the variation in arousal frequency. Regardless of climacteric vasomotor symptoms, the total sleep time and the percentages of various sleep stages were within earlier reported reference values.16 Schiff et al14 observed an increase of rapid eye movement sleep and a shortening of sleep latency with ERT, as well as an alleviation of hot flushes. However, the
Obstetrics & Gynecology
authors did not correlate the alleviation of hot flushes with objective sleep quality. We could not establish this kind of connection: Neither the percentage of rapid eye movement sleep nor the sleep latencies correlated with climacteric vasomotor symptoms. Despite the many sleep variables studied, no correlation was observed between subjective and objective sleep quality, supporting the earlier report that subjective and objective sleep quality are not always parallel.17 Women with vasomotor symptoms and subjective sleep complaints remind one of patients with insomnia with Sleep State Misperception, who have a normal electroencephalogram. Only a few somatic climacteric symptoms were correlated with objective sleep indices, and the observed correlations seemed sporadic rather than logically connected. Mental symptoms are central in the postmenopausal syndrome.18 According to the Beck Depression Inventory, none of our women were considered depressive.8 On the climacteric symptom questionnaire, some women did report climacteric mental symptoms, including depression, anxiety, mood instability, and memory problems. Sleep disturbances also occur in depression and anxiety, independent of menopause.5 Therefore, it could be argued that climacteric sleep disturbances are mediated through climacteric mental symptoms, which in the present study were more common in women with impaired subjective sleep quality. In stepwise regression analysis, however, variation in climacteric vasomotor symptoms explained the variation in subjective sleep complaints better than the variation in climacteric mental symptoms. Surprisingly, polysomnography showed that women with frequent mental symptoms had better sleep quality when assessed by less light sleep (stage 1) and by more rapid eye movement sleep. Women with higher BMI fell asleep as soon as those with low BMI, but they had difficulty entering deeper stages of sleep (stage-2 and slow-wave sleep). The prevalence of sleep-disordered breathing increases after menopause.19 High BMI predisposes to upper-airway obstruction and sleep apnea.20 Nocturnal breathing disturbances were not the focus of the present study. Some women, at least the most obese ones, could have suffered from partial upper-airway obstruction and increased respiratory resistance,9 which could have been responsible for lighter sleep. This interpretation is supported by the lower mean and minimum arterial oxyhemoglobin saturation observed in obese women. Periodic leg movements, which were not evaluated in the present study, also could have contributed to the reduced subjective sleep quality in women with vasomotor symptoms. Our study population was a selected group of healthy
VOL. 94, NO. 2, AUGUST 1999
women who responded to a newspaper announcement. Previous or latent diseases were effectively ruled out to exclude factors that could have influenced the relation between climacteric symptoms and sleep. Thus, the results of the present study cannot be directly extrapolated to clinical populations, in which menopausal symptoms often coincide with other physical and mental conditions. Moreover, the average climacteric symptoms scores in our study were quite low. By including more subjects with more severe vasomotor symptoms, we may have been able to show significant correlations with the objective markers of sleep quality. We were well aware that sleeping for the first night in a new environment (the so-called “first-night effect”) increases the arousal frequency21 and decreases sleep quality. However, it is also questionable to use data from the second night, during which the decreased sleep quality of the first night rebounds as “better than normal” sleep quality. Accordingly, only data from the third night should be accepted. However, a recent study showed no first-night effect in the distribution of sleep stages,4 so we compared the first nights in all subjects. Moreover, all women were studied in identical circumstances. Because we showed previously that placebo treatment had no effect on climacteric symptoms and sleep complaints,11 we believed that we could also study climacteric symptoms and sleep complaints during placebo treatment. The bias of giving estrogen to 31 women 4 months before the sleep study was effectively ruled out by calculating the carryover effect before the initial data analyses. In addition, the results of the 32 women who received placebo first (group A) did not differ from those obtained in the 63 women (groups A and B combined). All of our subjects had low serum E2 levels and could be considered estrogen deficient. Because of the marginal variation of serum E2 levels, serum E2 had little correlation with other characteristics. Why do climacteric vasomotor symptoms, which impair subjective sleep quality, not manifest as changes in objectively measured sleep? The two possibilities are that climacteric vasomotor symptoms do not induce marked specific changes in the polysomnogram or that polysomnography is insensitive to these changes. This study provides strong evidence that subjective and objective sleep are two distinct variables. It also calls into question whether sleep complaints can be regarded as an independent climacteric symptom or only as a consequence of the high frequency of vasomotor symptoms. Climacteric vasomotor symptoms seem to act as an alarm clock, causing arousals that women recall in the morning, but that do not have a sustained effect on sleep structure beyond the actual moments of arousal.
Polo-Kantola et al
Women and Sleep
223
References 1. Bliwise DL, King AC, Harris RB, Haskell WL. Prevalence of self-reported poor sleep in a healthy population age 50 – 65. Soc Sci Med 1992;34:49 –55. 2. Morgan K, ed. A research-based guide to sleep in later life. Baltimore: The Johns Hopkins University Press, 1987. 3. Gislason T, Almqvist M. Somatic diseases and sleep complaints. An epidemiological study of 3,201 Swedish men. Acta Med Scand 1987;221:475– 81. 4. Young T, Palta M, Dempsey J, Skatrud J, Webber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230 –5. 5. Morin C, Gramling SE. Sleep patterns and aging: Comparison of older adults with and without insomnia complaints. Psychol Aging 1989;4:290 – 4. 6. Ballinger CB. Subjective sleep disturbance at the menopause. J Psychosom Res 1976;20:509 –13. 7. Erlik Y, Tataryn IV, Meldrum DR, Lomax P, Bajorek JG, Judd HL. Association of waking episodes with menopausal hot flushes. JAMA 1981;245:1741– 4. 8. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4: 561–71. 9. Polo O. Partial upper airway obstruction during sleep. Studies with the static charge-sensitive bed (SCSB). Acta Physiol Scand Suppl 1992;145:1–118. 10. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring systems for sleep stages of human subjects. Publication no. 204. Washington DC: National Institutes of Health, 1968. 11. Polo-Kantola P, Erkkola R, Helenius H, Irjala K, Polo O. When does estrogen replacement therapy improve sleep quality? Am J Obstet Gynecol 1998;178:1002–9. 12. Shaver JL, Giblin E, Paulsen V. Sleep quality subtypes in midlife women. Sleep 1991;14:18 –23. 13. Thomson J, Oswald I. Effect of oestrogen on the sleep, mood, and anxiety of menopausal women. BMJ 1977;2:1317–9. 14. Schiff I, Regestein Q, Tulchinsky D, Ryan KJ. Effects of estrogens on
224 Polo-Kantola et al
Women and Sleep
15.
16. 17. 18.
19.
20.
21.
sleep and psychological state of hypogonadal women. JAMA 1979;242:2405–7. Purdie DW, Empson JA, Crichton C, Macdonald L. Hormone replacement therapy, sleep quality and psychological wellbeing. Br J Obstet Gynaecol 1995;102:735–9. Hauri P. The sleep disorders. 2nd ed. Kalamazoo, Michigan: Upjohn Pharmaceuticals, 1977. Bonnet MH, Arand DL. Physiological activation in patients with sleep state misperception. Psychosom Med 1997;59:533– 40. Smith RNJ, Studd JWW. Estrogens and depression in women. In: Lobo RA, ed. Treatment of the postmenopausal woman: Basic and clinical aspects. New York: Raven Press Ltd, 1994:129 –36. Block AJ, Wynne JW, Boysen PG. Sleep-disordered breathing and nocturnal oxygen desaturation in postmenopausal women. Am J Med 1980;69:75–9. Guilleminault C, Quera-Salva MA, Partinen M, Jamieson A. Women and the obstructive sleep apnea syndrome. Chest 1988;93: 104 –9. Agnew HWJ, Webb WB, Williams RL. The first night effect: An EEG study of sleep. Psychophysiology 1966;2:263– 6.
Address reprint requests to:
Pa¨ivi Polo-Kantola, MD Department of Obstetrics and Gynecology Turku University Central Hospital FIN-20520 Turku Finland E-mail: [email protected]
Received October 23, 1998. Received in revised form January 26, 1999. Accepted February 10, 1999. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Obstetrics & Gynecology
From the Departments of Obstetrics and Gynecology and Clinical Chemistry, Turku University Central Hospital; the Departments of Physiology and Biostatistics, University of Turku, Turku; and the Department of Pulmonary Medicine, University of Tampere, Tampere, Finland. Supported by research grants from the Medical Research Council of the Academy of Finland and the Gynecological Association of Finland. We thank Anne Kaljonen, MSc, for statistical assistance.
VOL. 94, NO. 2, AUGUST 1999
Sleep complaints, especially frequent nocturnal awakenings, increase with age.1 Elderly women suffer more often from sleep problems and use more hypnotics than elderly men.2 Somatic disorders such as chronic bronchitis, asthma, and cardiovascular symptoms affect sleep quality.3 Specific sleep disorders, including sleep apnea, cause insomnia or daytime sleepiness.4 Psychological factors also disturb sleep.5 It has been suggested that the decline of endogenous estrogen has an effect on sleep because sleep complaints increase rapidly after menopause.6 In addition to vasomotor symptoms, poor sleep quality is one of the most common symptoms of menopause. Climacteric vasomotor symptoms typically occur at night and interfere with sleep.7 We hypothesized that severe vasomotor symptoms are related to significant changes in the cortical electroencephalographic activity during sleep. However, sleep disturbances due to climacteric symptoms are poorly described in polysomnographic terms. To determine the type and the degree of specific electroencephalogram changes associated with vasomotor symptoms, we conducted all-night polysomnographic sleep studies in postmenopausal women who had a wide range of climacteric symptoms.
Materials and Methods Seventy-one healthy postmenopausal women were recruited over 15 months through a newspaper announcement, and 63 women completed the study. The age range was 47– 65 years (mean ⫾ standard deviation [SD] 56.3 ⫾ 4.4) and the minimum education was 7 years of basic school. Sixty-seven percent were married, 6% were single, and 27% were divorced or widowed. Forty-six percent were employed, 10% were unemployed, and 44% were retired. Postmenopause was defined as a serum FSH level exceeding 30 IU/L. Two women with FSH levels of 28 and 29 IU/L were accepted because of their ages (56 and 62 years). At the beginning of the study, the serum estradiol (E2) level was less than 50 pmol/L in all women except two,
0029-7844/99/$20.00 PII S0029-7844(99)00284-7
219
whose levels were 90 and 105 pmol/L, but whose serum FSH levels were postmenopausal (70 and 55 IU/L, respectively). Exclusion criteria included history of neurologic, cardiovascular, endocrinologic, or mental disease; medicated hyperlipidemia; malignancies; abuse of alcohol or medications; and heavy smoking (more than ten cigarettes per day). Latent systemic disorders possibly affecting sleep such as anemia, hypothyroidism, diabetes, and uremia were also excluded by measuring blood hemoglobin, leukocytes, sedimentation rate, serum thyrotropin, free thyroxine, vitamin B12, creatinine, glucose, and cholesterol levels. All subjects were strictly advised not to use any medications that affect the central nervous system, or any antioxidants. Incipient depression was screened using the Beck Depression Inventory.8 A complete gynecologic examination was done. Subjects were given oral and written information about the study, and they all provided written informed consent. The study was approved by the Ethics Committee of Turku University and Turku University Central Hospital. Polysomnographic sleep recording included continuous monitoring of electroencephalogram, electrooculogram, electromyogram, and electrocardiogram. A finger pulse oximeter (Biox Ohmeda 3700e; BOC Company, Louisville, CO) was used to monitor arterial oxyhemoglobin saturation. A static charge–sensitive bed (BioMatt; Biorec Oy, Helsinki, Finland)9 was used to record body movements, breathing patterns, and heart rates. Original analog signals were amplified and digitized at a frequency of 250 samples per second with 12-bit amplitude resolution and were recorded with custom software (UniPlot, Unesta Oy, Turku, Finland). Conventional criteria10 were used for sleep staging, which included stage-1, stage-2, slow-wave (including stage-3 and stage-4), and rapid eye movement sleep. Staging was carried out by the same scorer in all subjects. The scorer was masked to the presence of climacteric symptoms. Episodes of alpha arousals and body movements during sleep also were analyzed. Their frequencies were expressed as the number of episodes per hour of sleep. For all arousals, electroencephalogram alpha activity of at least 2 seconds was required. Arousal movement events were further classified with the static charge–sensitive bed as movement arousals (arousals with body movement) and electroencephalogram arousals (arousals without body movement). Highamplitude signals in the static charge–sensitive bed without simultaneous electroencephalogram alpha activity were considered body movements. This study was part of a larger survey evaluating the effect of estrogen replacement therapy (ERT) on sleep.11 In the original study design, randomization into two
220 Polo-Kantola et al
Women and Sleep
treatment groups was made in six-person blocks using random permuted blocks. Thirty-two women first received the placebo for 3 months and then, after 1 month washout with the placebo, estrogen for 3 months (group A). The remaining 31 women were treated in the reverse order (group B). Sleep studies were done twice in each subject at 4-month intervals, after placebo and after estrogen treatment. Data for the present report were collected using the placebo nights only. A statistically significant carryover effect was observed for three variables: latency to slow-wave sleep, percentage of stage-1 sleep, and electroencephalogram arousals in non–rapid eye movement sleep. For these variables, data were included only from group A. All the data were also analyzed with the 32 women in group A. The subjects self-reported daily climacteric symptoms for 14 days before the sleep study: vasomotor symptoms (hot flushes and sweating separately), sleep complaints (sleeping problems, insomnia), somatic symptoms (palpitations, numbness, muscular pain, dizziness, headache, fatigue), and mental symptoms (anxiety, depression, mood instability, memory problems, and lack of initiative). The intensity of the symptoms was evaluated on a six-step scale: 0 (“not at all”), 1 (“very little”), 2 (“to some extent”), 3 (“moderately”), 4 (“much”), 5 (“very much”), and 6 (“I cannot say”). To get an overall intensity score, the 14 daily scores were averaged. Some women express their vasomotor symptoms as hot flushes, whereas others feel incidental sweating. Therefore, the sum score of hot flushes and sweating on a scale from 0 to 10 was used to describe the severity of the vasomotor symptoms. For sleep complaints, somatic symptoms and mental symptoms scores ranged from 0 to 5. The few level-6 responses (“I cannot say”) were excluded from the analyses. Spearman correlation coefficient was used in the analysis of correlations. Because subjective symptoms were interdependent, stepwise regression analysis was used to study the connections among subjective sleep quality, climacteric vasomotor symptoms, and other climacteric symptoms. P ⬍ .05 was considered statistically significant. The power of our population to show the correlation r ⱖ .40 was greater than 90%. Statistical calculations were done using the SAS program package (SAS Institute, Cary, NC).
Results The mean (⫾ SD) body mass index (BMI) was 26.9 ⫾ 4.0 kg/m2 (range 19.8 –39.0), and the mean Beck Depression score was 4.0 ⫾ 3.8 (range 0 –15). In the morning after the sleep study, the mean serum E2 and serum FSH levels were 28.8 ⫾ 20.5 pmol/L and 70.4 ⫾ 22.7 IU/L, respectively. The subjects had not used ERT for at
Obstetrics & Gynecology
Table 1. Correlations Between Subjective Sleep Quality and Climacteric Symptoms Symptoms Vasomotor Hot flushes Sweating Sum score Somatic Palpitations Numbness Muscular pain Dizziness Headache Fatigue Mental Anxiety Depression Mood instability Memory problems Lack of initiative
Median
Range
r
P
1.0 1.3 2.6
0 – 4.3 0 – 4.7 0 – 8.9
.53 .58 .60
⬍.001 ⬍.001 ⬍.001
0.1 0.4 1.5 0.0 0.6 0.5
0 –3.8 0 – 4.0 0 –5.0 0 – 4.3 0 –3.1 0 – 4.0
.44 .30 .22 .21 .26 .25
⬍.001 .017 .079 .100 .038 .047
0.1 0.0 0.1 0.4 0.2
0 – 4.5 0 – 4.5 0 –5.0 0 – 4.0 0 – 4.0
.44 .51 .41 .44 .44
⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001
least 4 months before the sleep study. Before entering the study, 47 women (75%) had taken ERT previously, with a mean duration of 41 ⫾ 55 months (range 1–252). The mean time off ERT was 37 ⫾ 55 months (range 5–226). Each woman had had a hysterectomy for benign indications. Twenty-four (38%) had had oophorectomy, unilateral in nine (14%) and bilateral in 15 (24%). The median score for climacteric vasomotor symptoms (hot flushes plus sweating; scale 0 –10) was 2.6 ⫾ 2.5 (range 0 – 8.9), and the median score for sleep complaints (scale 0 –5) was 1.5 ⫾ 1.1 (range 0 –5.0). Women with subjectively impaired sleep experienced more climacteric vasomotor symptoms (r ⫽ .60, P ⬍ .001). They also had more climacteric somatic symptoms, such as palpitations, numbness, headache, and fatigue. More climacteric mental symptoms, such as anxiety, depression, mood instability, memory problems, or lack of initiative, were also present (Table 1). Climacteric symptoms were interdependent, so a stepwise regression analysis was performed. Thirty-two percent of the impairment in subjective sleep quality was explained by vasomotor symptoms (P ⬍ .001), 14% by palpitations (P ⬍ .001), and 4% by mood instability (P ⫽ .029). Body mass index, age, serum E2 level, or FSH level did not correlate with subjective sleep quality. Climacteric vasomotor symptoms were not related to sleep latency, latencies to sleep stages, percentages of sleep stages, frequencies of short (less than 1 minute) or long (longer than 1 minute) awakenings, number of arousals, number of body movements, number of arousals or body movements in various sleep stages, sleep efficiency (total sleep time/sleep-period time [awakening latency ⫺ sleep latency] ⫻ 100%), or total
VOL. 94, NO. 2, AUGUST 1999
Table 2. Correlations Between Vasomotor Symptoms (Sum Scores) and Sleep Variables Variable
Median
Range
r
P
Total sleep time (min) Sleep efficiency (%) Sleep latency (min) Latency (min) To stage 2 To stages 3– 4 To REM % Stage 0 1 2 3– 4 % REM % Movement time % Time out of bed SaO2 (%) Minimum (lowest) Maximum (highest) Mean
441.6 93.4 3.6
309.0 –511.2 74.4 –98.6 0.0 – 61.8
⫺.002 .11 .005
.989 .389 .969
3.6 14.4 76.8
0.0 – 67.8 4.8 –74.4 22.2–249.0
.21 ⫺.10 .13
.102 .605* .297
6.7 15.4 44.5 9.6 20.6 0.0 0.0
0.0 –25.6 7.2–28.0 25.8 – 61.2 1.2–23.7 9.2–31.5 0.0 –1.0 0.0 –1.5
⫺.07 ⫺.32 .04 .08 .10 ⫺.14 .20
.572 .072* .778 .518 .414 .286 .108
89.0 98.0 94.3
58.0 –92.0 96.0 –100.0 91.4 –96.1
.11 .14 .11
.411 .270 .407
REM ⫽ rapid eye movement; SaO2 ⫽ arterial oxyhemoglobin saturation. * A significant carryover effect in the basic study, data for calculation is drafted from the first recording night (n ⫽ 32).
sleep time (awakening latency ⫺ sleep latency ⫺ amount of stage-0 sleep). Vasomotor symptoms were not associated with the mean, minimum, or maximum arterial oxyhemoglobin saturations. (Tables 2– 4). To evaluate the relation between climacteric symptoms and the objective sleep variables, we analyzed more than 550 correlations. Just by repeating statistical tests, one would expect 29 significant correlations by chance. However, only six correlations between climacteric somatic symptoms and objective sleep variables were found, with r values from .26 to ⫺.45 (P ⬍ .05). These included more palpitations with longer rapid eye movement latency, more muscular pain with less stage-1 sleep and more rapid eye movement sleep, more dizziness with fewer movement arousals, more headache with less stage-1 sleep, and more fatigue with less stage-1 sleep. Seven correlations between climacteric
Table 3. Correlations Between Vasomotor Symptoms (Sum Scores) and Frequency of Awakenings Symptom (per hour of sleep)
Median
Range
r
P
Movement arousals No. awakenings ⬍1 min (short) No. awakenings ⬎1 min (long) EEG arousals Body movements
8.1 2.4 1.2 3.2 7.5
1.3–22.6 1.2–5.3 0.3–3.4 0.2–12.4 0.6 – 47.8
.05 ⫺.03 ⫺.04 ⫺.03 ⫺.01
.690 .802 .769 .851* .942
EEG ⫽ electroencephalogram. * A significant carryover effect in the basic study; calculation is drafted from the first recording night (n ⫽ 32).
Polo-Kantola et al
Women and Sleep
221
Table 4. Correlations Between Vasomotor Symptoms (Sum Scores) and Awakenings in Various Sleep Stages (per Hour of Sleep) Stage Stage 1 Movement arousals EEG arousals Body movements Stage 2 Movement arousals EEG arousals Body movements Stage 3– 4 Movement arousals EEG arousals Body movements REM Movement arousals EEG arousals Body movements
Median
Range
r
23.7 12.2 10.8
2.5–74.4 0.0 –36.7 0.0 – 43.7
4.2 3.1 6.3
0.0 –14.0 0.0 –20.8 0.0 –50.2
0.0 0.0 1.7
0.0 –10.6 0.0 – 4.4 0.0 – 47.0
.10 ⫺.02 ⫺.04
.432 .864 .752
5.7 1.1 8.2
0.5–28.0 0.0 –10.5 0.0 –39.0
.07 .11 ⫺.02
.589 .392 .875
.04 ⫺.10 ⫺.04 .12 .002 .002
P .753 .448 .785 .337 .986 .987
EEG ⫽ electroencephalogram; REM ⫽ rapid eye movement.
mental symptoms and objective sleep variables were observed, with r values from ⫺.26 to ⫺.56 (P ⬍ .05). These included more anxiety with less stage-1 sleep, more depression with less stage-1 sleep, more mood instability with less stage-1 sleep and with more rapid eye movement sleep, more memory problems with less stage-1 sleep and with more rapid eye movement sleep, and more lack of initiative with less stage-1 sleep. Women with higher BMI had a longer latency to stage-2 sleep (r ⫽ .27, P ⫽ .031) and to slow-wave sleep (r ⫽ .51, P ⫽ .003). High BMI was related to low mean arterial oxyhemoglobin saturation (r ⫽ ⫺.54, P ⬍ .001) and minimum arterial oxyhemoglobin saturation (r ⫽ ⫺.37, P ⫽ .003). Older women spent more time awake during the study night (r ⫽ .30, P ⫽ .019), and their sleep efficiency was decreased (r ⫽ ⫺.27, P ⫽ .030). They also had more body movements in stage-1 sleep (r ⫽ .34, P ⫽ .007), lower mean arterial oxyhemoglobin saturation (r ⫽ ⫺.36, P ⫽ .004), and lower minimum arterial oxyhemoglobin saturation values (r ⫽ ⫺.25, P ⫽ .047). Serum E2 correlated with only one objective sleep measurement: Low estrogen concentration predicted a high frequency of movement arousals in slow-wave sleep (r ⫽ ⫺.39, P ⫽ .002). Serum FSH concentrations did not correlate with objective sleep measurements. The number of arousals per hour of sleep through various sleep stages is shown in Table 4. Movement arousals (movements with alpha activity) and electroencephalogram arousals (alpha activity without movements) were most common in stage-1 sleep but were extremely rare in slow-wave sleep. The frequencies of body movements (movements without alpha activity) were similar in stage-1, stage-2, and rapid eye move-
222 Polo-Kantola et al
Women and Sleep
ment sleep, but in slow-wave sleep, the frequency of body movements was low. Women with frequent vasomotor symptoms had a sleep architecture (sleep-stage distribution) similar to those with a low frequency of symptoms. The data also were analyzed with the group of 32 women who received placebo first in the original study, with substantially the same results as for the entire 63-woman group.
Discussion The questionnaires indicated impaired subjective sleep quality in women with a high frequency of climacteric vasomotor symptoms. Somatic and mental climacteric symptoms were also more common in women with frequent sleep complaints. In contrast to the selfreported findings, polysomnography did not show any specific impairment of sleep architecture, sleep efficiency, or frequency of arousals induced by or related to climacteric vasomotor symptoms. The discrepancy between subjective and objective sleep quality was even more evident for climacteric mental symptoms, which impaired subjectively but improved objectively measured sleep. Serum E2 levels did not have a major effect. Age and obesity, rather than climacteric symptoms, appeared to have the strongest impact on polygraphic sleep variables. Sleep complaints and insomnia have been connected to menopause, especially to climacteric vasomotor symptoms. This connection is supported by a reversal of sleep complaints during ERT.11 Our results agree with these findings. Few studies have objectively measured sleep after menopause,7,12–15 and most of these studies focused on the effect of ERT on sleep quality, relating climacteric vasomotor symptoms to objective sleep quality. Erlik et al7 reported an association between hot flushes and waking episodes. Unfortunately, we did not monitor hot flushes or skin temperature during the sleep-recording nights. It is possible that the subjective scoring of climacteric symptoms reported during the 14 days before the sleep study could not predict objective vasomotor symptoms during the laboratory night. Hot flushes and sweating do arouse women, especially from light sleep (stages 1 and 2).7 Most of the arousals occurred from light sleep, but factors other than vasomotor symptoms, such as BMI, seemed to explain better the variation in arousal frequency. Regardless of climacteric vasomotor symptoms, the total sleep time and the percentages of various sleep stages were within earlier reported reference values.16 Schiff et al14 observed an increase of rapid eye movement sleep and a shortening of sleep latency with ERT, as well as an alleviation of hot flushes. However, the
Obstetrics & Gynecology
authors did not correlate the alleviation of hot flushes with objective sleep quality. We could not establish this kind of connection: Neither the percentage of rapid eye movement sleep nor the sleep latencies correlated with climacteric vasomotor symptoms. Despite the many sleep variables studied, no correlation was observed between subjective and objective sleep quality, supporting the earlier report that subjective and objective sleep quality are not always parallel.17 Women with vasomotor symptoms and subjective sleep complaints remind one of patients with insomnia with Sleep State Misperception, who have a normal electroencephalogram. Only a few somatic climacteric symptoms were correlated with objective sleep indices, and the observed correlations seemed sporadic rather than logically connected. Mental symptoms are central in the postmenopausal syndrome.18 According to the Beck Depression Inventory, none of our women were considered depressive.8 On the climacteric symptom questionnaire, some women did report climacteric mental symptoms, including depression, anxiety, mood instability, and memory problems. Sleep disturbances also occur in depression and anxiety, independent of menopause.5 Therefore, it could be argued that climacteric sleep disturbances are mediated through climacteric mental symptoms, which in the present study were more common in women with impaired subjective sleep quality. In stepwise regression analysis, however, variation in climacteric vasomotor symptoms explained the variation in subjective sleep complaints better than the variation in climacteric mental symptoms. Surprisingly, polysomnography showed that women with frequent mental symptoms had better sleep quality when assessed by less light sleep (stage 1) and by more rapid eye movement sleep. Women with higher BMI fell asleep as soon as those with low BMI, but they had difficulty entering deeper stages of sleep (stage-2 and slow-wave sleep). The prevalence of sleep-disordered breathing increases after menopause.19 High BMI predisposes to upper-airway obstruction and sleep apnea.20 Nocturnal breathing disturbances were not the focus of the present study. Some women, at least the most obese ones, could have suffered from partial upper-airway obstruction and increased respiratory resistance,9 which could have been responsible for lighter sleep. This interpretation is supported by the lower mean and minimum arterial oxyhemoglobin saturation observed in obese women. Periodic leg movements, which were not evaluated in the present study, also could have contributed to the reduced subjective sleep quality in women with vasomotor symptoms. Our study population was a selected group of healthy
VOL. 94, NO. 2, AUGUST 1999
women who responded to a newspaper announcement. Previous or latent diseases were effectively ruled out to exclude factors that could have influenced the relation between climacteric symptoms and sleep. Thus, the results of the present study cannot be directly extrapolated to clinical populations, in which menopausal symptoms often coincide with other physical and mental conditions. Moreover, the average climacteric symptoms scores in our study were quite low. By including more subjects with more severe vasomotor symptoms, we may have been able to show significant correlations with the objective markers of sleep quality. We were well aware that sleeping for the first night in a new environment (the so-called “first-night effect”) increases the arousal frequency21 and decreases sleep quality. However, it is also questionable to use data from the second night, during which the decreased sleep quality of the first night rebounds as “better than normal” sleep quality. Accordingly, only data from the third night should be accepted. However, a recent study showed no first-night effect in the distribution of sleep stages,4 so we compared the first nights in all subjects. Moreover, all women were studied in identical circumstances. Because we showed previously that placebo treatment had no effect on climacteric symptoms and sleep complaints,11 we believed that we could also study climacteric symptoms and sleep complaints during placebo treatment. The bias of giving estrogen to 31 women 4 months before the sleep study was effectively ruled out by calculating the carryover effect before the initial data analyses. In addition, the results of the 32 women who received placebo first (group A) did not differ from those obtained in the 63 women (groups A and B combined). All of our subjects had low serum E2 levels and could be considered estrogen deficient. Because of the marginal variation of serum E2 levels, serum E2 had little correlation with other characteristics. Why do climacteric vasomotor symptoms, which impair subjective sleep quality, not manifest as changes in objectively measured sleep? The two possibilities are that climacteric vasomotor symptoms do not induce marked specific changes in the polysomnogram or that polysomnography is insensitive to these changes. This study provides strong evidence that subjective and objective sleep are two distinct variables. It also calls into question whether sleep complaints can be regarded as an independent climacteric symptom or only as a consequence of the high frequency of vasomotor symptoms. Climacteric vasomotor symptoms seem to act as an alarm clock, causing arousals that women recall in the morning, but that do not have a sustained effect on sleep structure beyond the actual moments of arousal.
Polo-Kantola et al
Women and Sleep
223
References 1. Bliwise DL, King AC, Harris RB, Haskell WL. Prevalence of self-reported poor sleep in a healthy population age 50 – 65. Soc Sci Med 1992;34:49 –55. 2. Morgan K, ed. A research-based guide to sleep in later life. Baltimore: The Johns Hopkins University Press, 1987. 3. Gislason T, Almqvist M. Somatic diseases and sleep complaints. An epidemiological study of 3,201 Swedish men. Acta Med Scand 1987;221:475– 81. 4. Young T, Palta M, Dempsey J, Skatrud J, Webber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230 –5. 5. Morin C, Gramling SE. Sleep patterns and aging: Comparison of older adults with and without insomnia complaints. Psychol Aging 1989;4:290 – 4. 6. Ballinger CB. Subjective sleep disturbance at the menopause. J Psychosom Res 1976;20:509 –13. 7. Erlik Y, Tataryn IV, Meldrum DR, Lomax P, Bajorek JG, Judd HL. Association of waking episodes with menopausal hot flushes. JAMA 1981;245:1741– 4. 8. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4: 561–71. 9. Polo O. Partial upper airway obstruction during sleep. Studies with the static charge-sensitive bed (SCSB). Acta Physiol Scand Suppl 1992;145:1–118. 10. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring systems for sleep stages of human subjects. Publication no. 204. Washington DC: National Institutes of Health, 1968. 11. Polo-Kantola P, Erkkola R, Helenius H, Irjala K, Polo O. When does estrogen replacement therapy improve sleep quality? Am J Obstet Gynecol 1998;178:1002–9. 12. Shaver JL, Giblin E, Paulsen V. Sleep quality subtypes in midlife women. Sleep 1991;14:18 –23. 13. Thomson J, Oswald I. Effect of oestrogen on the sleep, mood, and anxiety of menopausal women. BMJ 1977;2:1317–9. 14. Schiff I, Regestein Q, Tulchinsky D, Ryan KJ. Effects of estrogens on
224 Polo-Kantola et al
Women and Sleep
15.
16. 17. 18.
19.
20.
21.
sleep and psychological state of hypogonadal women. JAMA 1979;242:2405–7. Purdie DW, Empson JA, Crichton C, Macdonald L. Hormone replacement therapy, sleep quality and psychological wellbeing. Br J Obstet Gynaecol 1995;102:735–9. Hauri P. The sleep disorders. 2nd ed. Kalamazoo, Michigan: Upjohn Pharmaceuticals, 1977. Bonnet MH, Arand DL. Physiological activation in patients with sleep state misperception. Psychosom Med 1997;59:533– 40. Smith RNJ, Studd JWW. Estrogens and depression in women. In: Lobo RA, ed. Treatment of the postmenopausal woman: Basic and clinical aspects. New York: Raven Press Ltd, 1994:129 –36. Block AJ, Wynne JW, Boysen PG. Sleep-disordered breathing and nocturnal oxygen desaturation in postmenopausal women. Am J Med 1980;69:75–9. Guilleminault C, Quera-Salva MA, Partinen M, Jamieson A. Women and the obstructive sleep apnea syndrome. Chest 1988;93: 104 –9. Agnew HWJ, Webb WB, Williams RL. The first night effect: An EEG study of sleep. Psychophysiology 1966;2:263– 6.
Address reprint requests to:
Pa¨ivi Polo-Kantola, MD Department of Obstetrics and Gynecology Turku University Central Hospital FIN-20520 Turku Finland E-mail: [email protected]
Received October 23, 1998. Received in revised form January 26, 1999. Accepted February 10, 1999. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Obstetrics & Gynecology