In biomedical science, it is generally surmised that symptoms associated with the transition to menopause are engendered by extant reproductive hormone fluctuations. Whether measuring hormones directly or using menopause status as a proxy, clear evidence for their sole effects is elusive, including for sleep difficulties. Using longitudinal survey data from a nationally representative Great Britain cohort study, the article by Tom et al1 in this issue of Menopause is addressed to clarifying the relationship between menopause status and sleep difficulty while accounting for age and a variety of other factors, that is, current somatic, vasomotor, and psychological symptoms; nighttime bathroom visits; and earlier measured socioeconomic status, physical and mental health, health behaviors, and the propensity to have poor sleep. The data set involved multiple repeated measures (most done annually) as reported by a large, random, social class-stratified sample of singleton birth women, born during 1 week in March 1946 and followed annually between ages 47 and 54 years. As with surveys, all data were derived from survey self-report, including sleep quality and menopause status (reported bleeding patterns). Women were classed a priori as having severe (bothered a lot by) or moderate (bothered a little by) sleep difficulties and compared with women reporting none. At each yearly assessment, women also were classed as remaining in premenopause, perimenopause, or postmenopause or transitioning from one to another stage.

Claiming importance to clinical management in the “Introduction,” the authors explicated wanting to distinguish whether “sleep difficulties in midlife women are related to the menopausal transition, other risk factors (stress of psychological distress), or other age-related changes” and whether sleep issues are newly experienced or exacerbated in menopause. Given the state of current knowledge about sleep and menopause, the fact that sleep quality will vary across a large sample chosen randomly and the results in this article, these are not “either/or” issues, and the answer is “all of the above” in some women. I comment on these issues with an emphasis on framing approaches for investigating sleep during the menopausal transition.

Age
A goal to disentangle what is caused by menopause status from what is caused by normative aging with regard to symptoms, and especially perceived sleep, is questionable. Of note in relation to sleep and normative aging in both men and women, mainly from studies not accounting for menopause status in women, the amount of physically recorded (polysomnograph) deep sleep wanes with age, or at least the electroencephalographic wave amplitudes believed to represent deep sleep do. Behaviorally, sleep is said to become lighter and less stable, so typically, older adults report waking up more often during the night. Therefore, it is not surprising nor revealing that Tom and her team found that chronological “age was a risk factor” for both the moderate and severe sleep difficulty groups of women.

Menopausal status

Most biomedical studies of outcomes associated with the menopausal transition, for example, vasomotor symptoms (VMS), sleep disruptions, sexual function, metabolic and disease risk factors, and cognitive or emotional function, are framed to show how they are attributable to ovarian hormone fluctuations or more precisely hypothalamic-ovarian (HPO) axis hormonal changes. A prevailing perspective, in part, is that these changes include progressive ovarian follicle depletion, early cycle inhibin B reductions, and increases in follicle-stimulating hormone levels that, for a period, sustain estrogen levels, perhaps rather erratically in some women. It is these among other changes that are most often seen to drive symptom and behavioral outcomes, whether any hormonal patterns are measured or not. However, supportive direct evidence is obscure, relegating hormone fluctuations to one of a number of considerations.

The HPO axis functional changes underlie the altered menstrual bleeding patterns that have been used to define the menopause stages emanating from the Stages of Reproductive Aging Workshop (STRAW, 2001). Specified was that hormonal changes (increasing follicle-stimulating hormone levels) begin in the late reproductive stage. The menopausal transition is recognized in the early phase when menstrual bleeding patterns become irregular (>7 d different from normal), progressing in the late phase to skipped menstrual cycles (two or more skipped cycles, amenorrhea for ≥60 d). Postmenopause is recognized 12 months past the final menstrual period, which, with the next 4 years, is considered early postmenopause stage and past this is considered late postmenopause stage. In this schema, high vulnerability to VMS (evidence to date shows that this is context for sleep disturbance) is said to occur in the late menopausal transition and early postmenopause stages. Because Tom and her team judged perimenopause by self-reported bleeding in the past 12 months but not in the past 3 months or decreased regularity compared with the previous year, their groups included women in the early and late menopausal transition. Similarly, the postmenopause group presumably included members in the early and late menopause. Thus, analyses according to the STRAW classifications are precluded, but of interest would be a longitudinal study that allowed classification of women according to the STRAW stages in concert with self-report measures of sleep and VMS or, better yet, subjective and objective measures of each.

Aligned with HPO axis hormonal fluctuations is the consideration of “windows of vulnerability” for experiencing symptoms like hot flashes, insomnia, or depressed mood as a function of ovarian hormonal fluctuations across reproductive stages. Although evidence is not extensive, Soares and Zitek comment on symptom vulnerability that crosses reproductive stages characterized by HPO hormone fluctuations, that is, during the menstrual cycle, pregnancy, and menopause. For example, they cite the observation that depression at menopause has been correlated with depression at premenstruum and postpartum. To reiterate, most attempts to link actual hormonal concentrations to symptom or behavioral health outcomes have failed to produce convincing evidence, implying that rapid or erratic hormone fluctuations might be more important than absolute hormone levels. Therefore, revealing dynamic temporal patterns of HPO axis hormonal changes as context for physical and cognitive/emotional changes deserves more attention.

Often, symptoms are assessed for an association with menopause status or stage as a proxy indicator for exposure to HPO axis hormone fluctuations. Indeed, sleep problems are reported by more women in the perimenopause or postmenopause stages than in the premenopause stage, and the prevalence escalates over the course of transition, suggestive of a contributing role. This was corroborated in the Tom et al study. After adjusting for age, most (moderate sleep difficulties) or all (severe sleep difficulties) of the women who were transitioning through natural menopause (as compared with premenopause) or had a hysterectomy had increased odds of self-reported sleep difficulty. Other large sample studies have revealed that rather than related to menopause status directly, perceived poor sleep quality was positively related to self-reported hot flashes, an indicator of menopause, but not to menopause status directly. Even objective sleep has not been related to menopause status, except perhaps in the context of hot flash activity. Thus, it remains equivocal as to whether sleep is disturbed by menopausal HPO hormone changes directly, indirectly, or
both.