Discussion
Our study showed that regularly working postmenopausal women slept worse than regularly working premenopausal women during workdays. According to the sleep diary, TST and ST were shorter, sleep latencies were longer, and awakenings were more frequent in postmenopausal women during the entire 14-day period and during workdays. Our novel finding was that postmenopausal women slept nearly as well as premenopausal women during the leisure days because only sleep latency remained longer. Depressive mood had a different impact on sleep in the menopausal groups during the workdays: premenopausal women with higher BDI scores had shorter TST and ST, whereas postmenopausal women with higher scores had longer TST and ST. Furthermore, higher education level associated with longer TST and ST during the workdays in premenopausal women, whereas higher education level associated with shorter TST and ST in postmenopausal women.
To our knowledge, this is the first sleep diary study in premenopausal and postmenopausal women to deal with workdays and leisure days separately. A recent sleep diary study of 74 women, aged between 40 and 59 years, found no difference in TST, total wake time, or sleep efficiency between postmenopausal and nonmenopausal women. However, workdays and leisure days were not analyzed separately. In that study, all women had primary insomnia, whereas in our study, insomnia symptoms were not the inclusion criteria. In a cross-sectional study of 266 healthy men and women aged 20 to 50 years, the mean time in bed increased by 27 minutes during weekend versus during weeknight, and the mean sleep latency decreased by more than a minute, respectively. There were no sex differences in these variables. The effect of menopause was not evaluated, and most of the female participants were probably premenopausal considering the age range of the study cohort. In our study, in premenopausal women, the mean difference in TST was -71 minutes and the mean difference in ST was -68 minutes during workdays compared with leisure days. In postmenopausal women, the mean difference was even more, -100 and -89 minutes, respectively. These data suggest that sleep is restricted during workdays and that leisure days offer an opportunity to catch up.
In the present study, postmenopausal women reported worse sleep than premenopausal women during the entire 14-day period and during the workdays. This is in line with most of the literature on questionnaire studies, which indicate that postmenopausal women experience more sleep problems than premenopausal women. The Wisconsin Sleep Cohort Study of 589 women confirmed that perimenopausal and postmenopausal women were more dissatisfied with their sleep quality than were premenopausal women. In the Australian Longitudinal Study on Women's Health with 8,649 participants, the natural menopausal transition was associated with difficulty in sleeping. Shin et al found that, among 2,400 middle-aged Korean women, the prevalence of insomnia increases markedly in the transition from premenopause to perimenopause. However, in our previous study of 3,421 women with five age cohorts, no differences were found between groups, although the groups differed in menopause status.
Vasomotor symptoms, such as hot flashes and sweating, usually occur at night and frequently associate with sleep disturbances. Especially, the relationship between self-reported sleep problems and self-reported climacteric symptoms is well acknowledged. Ohayon found that, among 982 women aged 35 to 65 years, the prevalence of chronic insomnia symptoms increased with the severity of hot flashes, reaching more than 80% in perimenopausal and postmenopausal women who had severe hot flashes. However, in the present study, after adjustment for sleep diary results with vasomotor symptom score, TST and ST for the entire 14-day period remained shorter for postmenopausal women compared with premenopausal women; in addition, during workdays, there was a tendency for shorter TST and ST. In the postmenopausal group, there were nine women on HT, which is thought to improve sleep by alleviating vasomotor symptoms. According to the post hoc analysis, which excluded HT users, the main results remained the same, and only latency during workdays and awakenings for the 14-day period approached those in premenopausal women. This supported our earlier finding that HT use does not always prevent sleep impairment.
Increasing age associates with sleep problems, in addition to menopause. The women in the present study were originally selected by menopause status. Because of the recruitment strategy, the age of the two groups also differed by approximately 10 years. The present study is cross-sectional; thus, we cannot denote a causative role for menopause per se for the difference in subjective sleep between the premenopausal group and the postmenopausal group. Aging and more frequent comorbidities in older participants might affect sleep, especially in the postmenopausal group. However, comorbidity is unlikely to explain the difference between the two menopausal groups in our study because women with severe illnesses were excluded.
Long work hours, higher work demands, and shift work relate to the incidence of sleep disturbances. Sleep problems due to work increase with age, especially among shift workers, but probably among nonshift workers as well. In our study, this difference was evident on workdays, which might indicate that work induces sleep problems more frequently after menopause than before menopause or that menopause accelerates the effects of aging. Insufficient sleep during workdays and lack of recovery may predispose one to long-term negative health consequences and impaired quality of life. From another point of view, inadequate sleep quality on workdays may lead to decreased productivity at work, increased risk of accidents, and work absenteeism. Sleep disturbances due to work decrease during leisure days. Because sleep problems, except sleep latency, diminished during the leisure days in this study, postmenopausal women are suggested to still have quite a good capacity to recover. However, because of the unambiguous difference in subjective sleep during the workdays between premenopausal and postmenopausal women, ensuring sufficient recovery periods from work among aging employees is essential. This is likely to enable better productivity and to possibly lengthen work carriers.
Postmenopausal women scored worse in the BDI than premenopausal women. This is consistent with several previous studies, which showed that the menopausal transition is an independent risk factor for depressive symptoms in midlife women. Furthermore, mood symptoms have been shown to associate with sleep problems in perimenopausal and postmenopausal women. However, it must be noted that moderate and severe depression were excluded from our study; thus, women who scored higher in the BDI still had only mild depressive symptoms. After adjusting the results for the BDI, we found an interesting interaction between TST and ST during workdays. Higher depressive symptoms in premenopausal women were linked to shorter TST and ST, whereas higher depressive symptoms in postmenopausal women occurred with longer TST and ST during workdays. Depression can shorten or prolong sleep duration. Although women were not depressed, it was interesting to note that a slight difference in BDI score for sleep duration was distinguished between the menopausal groups.
In our study, premenopausal women had higher education level than postmenopausal women. This was expected because the standards of living and education level in Finland increased in the 1960s and 1970s. Associations between education level and sleep problems show that a lower education level predicts poorer subjective sleep quality. In our study, however, after adjusting for education, TST and ST during the 14-day period remained shorter, and latency during the 14-day period and during leisure days remained longer in postmenopausal women than in premenopausal women. During workdays, TST and ST acted differently between premenopausal and postmenopausal women. In premenopausal women, TST and ST were longer in higher-educated women than in lower-educated women, supporting the previous findings in the literature. However, a higher education level in postmenopausal women was linked with shorter TST and ST during workdays compared with postmenopausal women, who had a lower education level. This could indicate that, after menopause, work demanding higher education level may cause a decrease in subjective sleep.
Our study has several limitations. First, recruiting women by a newspaper announcement could appeal more to poor sleepers than to good sleepers. Second, our participants were relatively healthy. Because comorbidities frequently affect quality of sleep, our findings cannot be extrapolated to women with major illnesses. Third, owing to the moderate sample size, the differences in sleep measures could be overestimated or underestimated. Fourth, our study lacked objective measures of sleep, we used a sleep diary. The incongruence between subjective and objective sleep quality is well acknowledged. Because of this mismatch, it has been concluded that these measurements reflect different matters. In a study of postmenopausal women experiencing hot flashes, those who underestimated sleep duration were more likely to report low-quality sleep. However, this does not explain the difference in sleep during workdays and leisure days between premenopausal and postmenopausal women. In clinical work, questionnaires are widely used as practical tools to evaluate the severity of sleep problems. Fifth, our study was cross-sectional and cannot exclude other reasons—besides menopause—that cause sleep problems.
Our participants are likely to represent the relatively healthy, regularly working, Finnish female population around menopause. Major illnesses, moderate depression, and severe depression, as well as shift workers, unemployed persons, and retired persons, were excluded. Most premenopausal and postmenopausal participants were midrange sleepers, as they were in the recent Finnish study of more than 9,000 participants that evaluated self-reported sleep duration in the Finnish general population. Exclusion of important confounding factors, such as nocturnal breathing problems, ensured a more precise evaluation of the influence of menopause. Another strength of our design is that we studied women who worked regular hours. There have been several studies on women's sleep and shift work, but none has investigated how workload influences sleep in postmenopausal women working regular hours.