Physical Activity and Menopausal Symptoms: Evaluating the Contribution of Obesity, Fitness and Ambient Air-Pollution Status

Abstract

Objective:

The menopausal transition is accompanied by transient symptoms that have been linked to subclinical cardiovascular disease (CVD); CVD has also been linked to air pollution. Physical activity (PA) reduces CVD, improves body composition, and can reduce menopausal symptoms. The purpose of this study was to assess the links between PA and menopausal symptoms and whether obesity, fitness, and air-pollution status play a role in this relationship.

Methods:

Women (40–60 yrs; N=243; M age=47.8, SD=5.6) from areas with high versus low air-pollution enrolled in the 4HAIE prospective cohort study completed psychological, cardiorespiratory fitness, body composition, and menopausal status screening followed by a 14-day prospective assessment of menopausal symptoms (Menopause Rating Scale) using a mobile application. Daily PA was assessed objectively across 14 days via Fitbit Charge 3 monitor. General linear mixed models were conducted and controlled for age, menopausal status, day in the study, wear-time, and neuroticism.

Results:

Peri/postmenopausal women (β=0.43, p<.001) and those residing in a high air-pollution environment (β=0.45, p<.05) reported more somato-vegetative symptoms. Hot flashes alone were associated with peri/postmenopausal status (β=0.45, p<.001), and for women residing in a high air-pollution environment, lower reporting of hot flashes was observed on days when a woman was more physically active than usual (β=−0.15, p<.001). No associations were found for cardiorespiratory fitness and visceral fat with any of the symptoms.

Conclusions:

PA may enhance resilience to hot flashes, especially when residing in high air-pollution environments where we also observed higher reporting of somato-vegetative menopausal symptoms.

INTRODUCTION

Menopause refers to the cessation of ovarian follicular activity and is manifest by the cessation of menstrual flow lasting at least 12 months. Perimenopause encompasses the time immediately before menopause (i.e., after menstrual cycles become irregular) and the first year after menopause, representing a transitional period when physiological, hormonal, and clinical changes commence, which lasts on average 3.5–4 years¹. Although there is no universal menopausal syndrome² and most women view menopause as a neutral event in their lives³, data from several longitudinal studies^4–10 have indicated that anywhere from 50 to 88% of women suffer from bothersome acute symptoms during the menopausal transition. Among the most common symptoms reported by menopausal women are hot flashes, night sweats, irritability, moodiness, tension, anxiety, low self-esteem, and emotional instability¹. However, the term “menopause management” is typically constrained to the management of vasomotor symptoms such as hot flashes and night sweats. These symptoms represent a heavy socioeconomic burden on society^11,12 and they are the primary reason for middle-aged women visiting a physician¹³. In a significant portion of women, vasomotor symptoms can last up to a decade after menopause^14,15 and they negatively impact on quality of life^16,17.

Interestingly, women with menopausal complaints display other risk factors including lower bone mineral density¹⁸, less favorable cardiovascular risk profile¹⁹, higher risk for hypercholesterolemia, hypertension²⁰, and higher body mass index²¹, pointing to a possible overlap in etiology of both cardiovascular disease (CVD) and vasomotor symptoms²². Both CVD and vasomotor symptoms have also been linked to lifestyle factors, including physical activity^23,24, although currently physical activity or exercise are not endorsed as effective nonhormone treatments for the management of vasomotor symptoms²⁴. Importantly, however, menopause has been associated with a decrease in energy expenditure and increase in total body fat and visceral adipose tissue²⁵, but maintaining or increasing participation in regular physical activity may prevent or attenuate any such increases in midlife women^26–29. Unfortunately, women who experience vasomotor symptoms spend less time on exercise per week than do those who do not have hot flashes³⁰, and only approximately half of middle-aged women engage in any leisure-time physical activity on a regular basis, with less than 25% meeting the public health recommendations of at least 30 minutes of accumulated moderate physical activity on most days of the week^31,32.

Both health status and menopause outcomes may also be influenced by contextual factors such as the environment, be it directly or indirectly via its effects on behavior. Among them, air pollution has been associated with shorter life expectancy, higher all-cause mortality, and an increased incidence of cardiovascular diseases and higher CVD mortality^33–36. Short-term exposure to air pollutants such as O₃, NO₂, SO₂ has been linked with inflammation (CRP and in TNF-α for NO₂)³⁷. Other air pollution parameters (PM_2.5, PM10, CO, NO₂, O₃, and SO₂) have been associated with BMI score and obesity^38,39 (including prenatal exposure to PAHs, PM_2.5 and NO₂)⁴⁰, and a link has been established between air pollution and increased trunk fat distribution⁴¹, which in turn has been linked to an increased odds of hot flashes⁴². Both long-term (PM_2.5, NO₂) and short-term (PM₁₀ and CO) exposure to air pollutants have been linked with an increased risk of depressive symptoms⁴³ and lower wellbeing⁴⁴. Finally, air pollution can also influence health and wellbeing indirectly, by reducing physical activity and increasing sedentary behavior⁴⁵, which can have cascading effects on different bodily processes.

The negative effects of air pollution are thought to be the result of a number of physiological mechanisms involving processes such as inflammation, oxidative damage, lipid peroxidation, carcinogenic or genotoxic effects^33,46–48. For example, PM_2.5 has been shown to contribute to the progression of subclinical atherosclerosis among women transitioning through menopause⁴⁹. Exposure to environmental pollutants has also been studied in relation to reproductive aging outcomes, showing that endocrine-disrupting chemicals can contribute to earlier onset of menopause⁵⁰, as can exposure to PM and traffic^51,52. It is unclear whether air pollution can also contribute to a more problematic menopause transition in the form of increased menopausal symptoms. While studies on possible protective role of physical activity with respect to the symptoms of menopause exist^53–56, more research is needed to better understand whether these benefits are realized to the same extent within environments with varying levels of ambient air pollution, while accounting for the effects of important covariates such as fitness level, body composition status, or neuroticism, all which have been also associated with menopausal symptoms^42,57,58. In this study, we used data from the Healthy Aging in Industrial Environment Program 4 (4HAIE) study to examine the associations between daily physical activity (Fitbit steps) and menopausal symptoms in a subsample of middle-aged women recruited from the 4HAIE cohort residing in areas of high versus low relative ambient air pollution. Specifically, we were interested in evaluating (1) whether an overall level of physical activity was associated with menopausal symptoms; (2) whether daily deviations from typical physical activity levels were associated with daily changes in symptom reporting across the 14-day monitoring period; (3) whether cardiorespiratory, body composition and air-pollution status contributed to symptom reporting, while controlling for the effects of neuroticism; and (4) whether air-pollution status modified the physical activity-symptom reporting association.

METHOD

Participants

The sample included middle-aged women (40–60 years, a period when most cases of spontaneous menopause occur⁵⁹) from the 4HAIE study. The 4HAIE study is a prospective cohort study (N=1314) on the links between air pollution, biomechanical, physiological, psychosocial, and sociodemographic variables on the incidence of running-related injuries, physical (in)activity, health outcomes and quality of life. The sample was derived using non-probabilistic quota sampling of individuals (aged 18–65) from two different regions of the Czech Republic: the Moravian-Silesian Region (MSR - an area with high relative air pollution) and the Southern-Bohemian Region (SBR - an area with relative low air pollution). Based on long-term monitoring of benzo[a]pyrene and particulate matter in the air, the MSR of the Czech Republic is considered a European hotspot⁶⁰. The SBR, due to its relatively low concentrations of air pollutants in the long term, was chosen as the control region^60–62. Concretely, for example, during (April) 2019-(August) 2022 when the recruitment and data collection took place, the average annual concentration of PM_2.5 in the key agglomerations within the two regions ranged from 20.27, 17.57, 19.45, and 18.45 for the years 2019, 2020, 2021, and 2022 in the Ostrava, Havířov, Karviná area (i.e., the main cities in the MSR) and 12.80, 10.90, 11.70, 11.50 μg/m³ in the area around České Budějovice (the main city in the SBR), respectively^63,64. In their latest recommendations from 2021, the World Health Organization (WHO) recommends annual concentrations below 5 μg/m³ (threshold that was lowered from 10 μg/m³ in the previous 2005 recommendations)⁶⁵. In both regions there is similar socioeconomic and ethnic (about 95% White) stratification and inhabitants claim Czech or Moravian nationality⁶⁶. The recruited participants comprised both active (runners) and inactive individuals. The runners were individuals who reported meeting the WHO physical activity guidelines of 150 minutes of moderate/vigorous activity per week and regular running of at least 10 km per week in the past 6 weeks of more. Inactive individuals were those reporting activity levels below the WHO recommended level.

Study Design

The details of the 4HAIE study protocol have been published elsewhere^67–69. Briefly, the 4HAIE study included baseline laboratory assessments, 12 months of prospective monitoring of physical activity and four intensive 2-week measurement bursts with ecological momentary assessment (EMA) methodology assessing contextual and psychosocial correlates of physical activity. In this investigation, data from the first 2-week intensive measurement burst are utilized to examine prospectively the day-to-day associations between physical activity and menopausal symptoms.

Measures

Background information.

The participants underwent pre-screening for eligibility criteria online (physical activity level, running history, injury status) and were subsequently screened on the telephone where menstrual cycle history and menopausal/hormonal status were ascertained based on the Stages of Reproductive Aging Workshop⁷⁰. Where it was not possible to ascertain the time of last menstrual period (e.g., when not menstruating due to using an IUD), menopausal status was determined based on age and self-reported menopausal status⁷¹. Prior to arriving to the baseline laboratory assessments, participants completed two online questionnaires. The first questionnaire was focused on the socioeconomic determinants of health and included information about age, education, socioeconomic and health status (rated subjectively on a 5-point scale as very good, good, satisfactory, poor, very poor). The second questionnaire included measures of psychological correlates of health outcomes including neuroticism which was considered as a control variable due to its established relationship with symptom reporting, including during menopause⁷². Neuroticism was measured by 10-item International Personality Item Pool (IPIP) scale⁷³. Although both depressive symptoms and anxiety have been established as correlates of menopausal symptoms, we prioritized neuroticism as a control variable for several reasons: (a) neuroticism reflects a dispositional tendency toward frequent and intense negative emotions and individuals high in neuroticism often exhibit heightened symptom reporting, including both depressive and anxiety symptoms; (b) our study focused on a sample of otherwise healthy, community-dwelling women with relatively low levels of depressive and anxiety symptoms; (c) depressive and anxiety symptoms were already included as part of the menopausal symptom assessment. Thus, prioritizing neuroticism as a control variable allowed us to avoid redundancy, enhance precision, and align with the principle of model parsimony.

Laboratory assessments.

During a 2-day laboratory visit, the participants underwent body composition assessments (height and weight assessed using Stadiometren InBody BSM 370 from Biospace, Seoul, South Korea). Percent body fat including visceral fat was assessed using dual energy absorptiometry (DXA, Hologic Discovery A, Waltham, Massachusetts). Participants also underwent maximum graded exercise testing (GXT) for the assessment of cardiorespiratory fitness (peak oxygen consumption, $\dot{V}$O_2peak). Viscreral fat and fitness status were considered as potential covariates of menopausal symptoms given previous evidence linking body composition and lower fitness levels as potential moderators of the PA-menopausal symptom relationship^42,74.

Physical activity assessment.

Daily physical activity across two weeks was assessed using Fitbit Charge 3 or 4 monitors worn on the wrist. Participants were asked to wear the monitor continuously including sleep and only remove it when there was a risk of injury (e.g., during volleyball). Total daily step count as provided by Fitbit was taken as an index of an overall level of daily physical activity. The data were used when at least 10 valid hours of “non-sleep” data were available, with a valid minute defined as a minute where either (at least one) heart rate value or non-zero step counts were recorded. This definition meant that 3% of available „person-day“ values were excluded from the analyses. In the descriptive tables, for ease of interpretation, the data on steps per day are presented as reflecting daily average steps standardized per ten valid hours. In the analyses, the actual reached daily step counts were used and models were adjusted for total daily wear time (with minutes marked by Fitbit as sleep being excluded).

Menopausal symptom assessment.

Across two weeks, while wearing the Fitbit monitor, the women in the sample completed the Czech version of the Menopause Rating Scale (MRS)^75,76 adapted for presentation in a custom-built mobile app daily as part of an evening survey (delivered between 20:00–22:00). The MRS assesses the severity of eleven symptoms in three categories (psychological, somato-vegetative, uro-genital). The Czech version of the scale adapted for daily assessments asked the respondents to rate the extent to which the symptoms applied to them today on a scale ranging from 0 (none, no difficulties) to 4 (unbearable difficulties). Psychological symptoms included four symptoms: (1) depressive mood (feeling down, sad, on the verge of tears, lack of drive, mood swings), (2) irritability (feeling nervous, inner tension, feeling aggressive), (3) anxiety (inner restlessness, feeling panicky), (4) physical and mental exhaustion (general decrease in performance, impaired memory, decrease in concentration, forgetfulness). Somato-vegetative symptoms included four symptoms: (1) hot flashes, sweating (episodes of sweating), (2) heart discomfort (unusual awareness of heart beat, heart skipping, heart racing, tightness), (3) sleep problems (difficulty in falling asleep, difficulty in sleeping through, waking up early), (4) joint and muscular discomfort (pain in the joints, rheumatoid complaints). Uro-genital symptoms included three symptoms: (1) sexual problems (change in sexual desire, in sexual activity and satisfaction), (2) bladder problems (difficulty in urinating, increased need to urinate, bladder incontinence), (3) dryness of vagina (sensation of dryness or burning in the vagina, difficulty with sexual intercourse).

Analytical Approach

We conducted linear mixed effects models using the lme4 package for R⁷⁷. Since we collected intensive longitudinal data with repeated daily assessments of both physical activity and menopausal symptoms across 14 days, the data have a hierarchical structure. Using linear mixed effects models (or multilevel modeling) accounts for the dependencies between observations within the same person over time. It also allows researchers to examine how variables change over time within individuals, as well as how they differ between individuals by pooling information from all participants and days. The model estimates the average effect of overall physical activity level on symptom reporting across all individuals (i.e., between-person effects), while also accounting for individual differences in the relationship between physical activity and symptoms (i.e., within-person effects).

Specifically, in this approach, both between- and within-person associations between PA and menopausal symptoms were evaluated. Intercept and slope (for hot flashes and somato-vegetative symptoms), respectively intercept alone (for psychological and uro-genital symptoms, as models with slope did not converge), were fitted as random effects allowing participants to have different PA values at baseline. First, we person-mean centered daily measures of PA and symptoms to derive the within-person component representing the daily deviation from within-person mean („state-like” variable). Second, we computed the between-person component by sample-mean centering the daily values. This „trait“ component represented the within-person mean across the 14 days, denoting overall level of either PA or symptoms for the individual. Separate models were run for the psychological and somato-vegetative category of symptoms plus for hot flashes, as the cardinal symptom of menopause. We did not run the model for uro-genital symptoms given the low prevalence of the symptoms in the study sample (i.e., only 11% percent of the observations represented non-zero values for this symptom subscale). The models included a number of control variables at baseline (to account for differences in these variables at the between-person level): age, visceral fat, $\dot{V}$O_2peak, neuroticism, menopausal status (premenopausal versus peri/postmenopausal status), location (MSR versus SBR), day in the study (to control for fatigue effects in reporting), Fitbit wear time. Additionally, we incorporated two interaction terms between location (MSR versus SBR) and physical activity to explore whether the between or within-person associations between physical activity and symptoms differ as a function of residing in the air-polluted region (MSR).

• Within-Person Model (Level-1): HF_di = β_0i + β_1i(Daily PA_di) + e_di

• Between-Person Model (Level-2): β_0i = γ₀₀ + γ₀₁(Overall PA_i) + γ₀₂(age_i) + γ₀₃(Viscreal Fat_i) ++ γ₀₄(VO2peak_i) + γ₀₅(neuroticism) + γ₀₆(status_i) ++ γ₀₇(location_i) + γ₀₈(day_i) + + γ09(fitbit wear timei) + γ10(Overall PAi * Locationi) + γ11(Daily PAi * Locationi) + u_0i

• β_1i = γ₁₀ + γ₁₁(Location_i) + u_1i

where γ₀₀ represents the average level of symptoms in the sample, γ₀₁, γ₀₂, …, γ₀₈ represent the influence of PA and symptoms (i.e., between-person effects), γ₁₀ represents the average effect of PA fluctuations (i.e., average within-person effects) on symptoms and u_0i are individual-level residual deviations that are uncorrelated with the day-level residuals e_di. Of particular interest for our research questions are the γ₀₁ and γ₁₀ parameters quantifying the between-person and within-person associations between PA and symptoms.

RESULTS

Sample description

Out of the 1314 4HAIE study participants, 282 women were in the 40–60 age range upon study entry (M age = 48.10, SD = 5.70). For two women, menopausal stage could not be reliably established. Additionally, two women had missing values for symptoms, six for neuroticism, and 31 did not undergo assessment of cardiorespiratory fitness due to high blood pressure prior to testing. Hence, in the end data from 243 women were used in the analyses. There were no statistically significant differences between the women included in the analyses and those that were missing some data except for age, with those excluded having higher median age (49 vs 47, p<0.05). Of the 243, 136 were from the MSR (56%) and 107 self-identified as runners during screening (44%). All women were non-smokers. Most of the women were married or lived with a partner (71.6%, n=174), with the rest reporting being single (4.1%, n=10), divorced (22.2%, n=54), or widowed (1.2%, n=3). Two women did not report their relationship status. As for education levels, university degree was reported by 48.6% (n=118), secondary education (high school or some years of higher education) was reported by 40.7% (n=99), with 9.9% (n=24) reporting education without a high school degree. The majority rated the socioeconomic situation of their family as average (81.9%, n=199), with 14.0% (n=34) reporting above average and 3.3% (n=8) below average economic situation. Two women did not report their education and two women did not report the economic status of the family. In terms of menopausal status, 58% (n=147) were categorized as premenopausal based on their menstrual history and/or self-reported menopausal status when menstrual bleeding criteria could not be applied (e.g., when using hormonal IUD). Although 50.2% self-identified as active runners, the distribution of both aerobic fitness (as measured by $\dot{V}$O_2peak) and physical activity levels (as measured by Fitbit monitor) indicated a wide range of activity levels in the sample. The sample did not differ in any of the baseline parameters between the two locations (MSR vs SBR). The descriptive statistics are summarized in Table 1.

Table 1.

Descriptive characteristics of the study sample

Variable	MSR	SBR	Total
	Number (%)	Number (%)	Number (%)
	136	107	243
Menopausal status
Premenopausal	82 (60%)	65 (61%)	147 (60%)
Peri or postmenopausal	54 (40%)	42 (39%)	96 (40%)
Hysterectomy	7 (5%)	7 (7%)	14 (6%)
Surgical menopause	3 (2%)	4 (4%)	7 (3%)
Other reason for no menstrual bleeding	2 (1%)	2 (2%)	4 (2%)
Hormonal contraception	32 (24%)	25 (23%)	57 (24%)
Intrauterine device	25 (18%)	18 (17%)	43 (18%)
	Median (Q1–Q3)	Median (Q1–Q3)	Median (Q1–Q3)
Age	47.0 (43.75–52.25)	46.0 (43.0–51.5)	47.0 (43.0–52.0)
Visceral fat	70.4 (47.4–108.7)	66.5 (40.6–96.5)	66.8 (39.9–96.3)
Neuroticism	2.2 (1.8–2.6)	2.2 (1.9–2.7)	2.2 (1.8–2.7)
CRF VO2peak	32.9 (28.1–39.1)	32.9 (28.0–38.6)	32.6 (28.0–38.1)
Education	4.0 (3.0–5.0)	3.0 (3.0–5.0)	4.0 (3.0–5.0)
Socioeconomic status	2.0 (2.0–2.0)	2.0 (2.0–2.0)	2.0 (2.0–2.0)
Health status	2.0 (2.0–2.0)	2.0 (2.0–2.0)	2.0 (2.0–2.0)
Income	14.0 (11.0–16.0)	14.0 (11.0–15.0)	14.0 (11.0–16.0)
* Fitbit steps	13921(9645–17996)	13242.36(10043.92–17265.21)	13208(10044–16961)
* Psychological	0.19(0.03–0.33)	0.17(0.07–0.42)	0.17(0.07–0.36)
* Somato-vegetative	0.23(0.12–0.48)	0.21(0.08–0.34)	0.21(0.06–0.33)
* Uro-genital	0.00(0.00–0.05)	0.00(0.00–0.05)	0.00(0.00–0.03)
* Hot flashes	0.07(0.00–0.34)	0.00(0.00–0.29)	0.07(0.00–0.32)

Note. CRF=cardiorespiratory fitness; MSR=Moravian Silesian region (air polluted region); SBR (South Bohemian Region (non-air polluted region); Q = quartile; VO2peak= peak oxygen consumption (in ml/kg/min).

^*values calculated based on 14-day person means.

Adherence to the daily physical activity and menopausal symptom assessment

The women were highly adherent both to the Fitbit and the daily surveys, completing on average 99.7% of the surveys (min=78.6%, max=100%, M=99.72, SD=1.94) and having valid Fitbit step data for 3769 out of the 3920 person-day values possible (96% adherence). After pairing the Fitbit data with daily survey data, 3608 person-day values were available for the analyses of PA-menopausal association.

Effects of physical activity on menopausal symptoms

The results of the multilevel analyses for each type of symptoms are presented in Table 2. The analysis revealed substantial variation in both daily PA and menopausal symptoms at the between- and within-person levels, respectively. The intraclass correlation coefficients reflecting between-person variation were 45.8% for PA and 44.8%, 53.1%, 59.3%, 44.3% for psychological, somato-vegetative, uro-genital, and hot flash symptoms, respectively, suggesting that 54.2% of the PA and on average 49.6% of the symptom data varied within-persons over time.

Table 2.

Key multilevel model statistics for the basic model testing effects of daily steps on daily menopausal symptoms

	Psychological	Somato-Vegetative	Hot Flashes
	Estimate (Std.Error)	Estimate (Std.Error)	Estimate (Std.Error)
Intercept, γ00	0.003 (0.082)	−0.261 (0.084)** [*]	−0.182 (0.079)* [NS]
Overall PA („trait“), γ01	−0.058 (0.075)	−0.009 (0.075)	0.032 (0.072)
Daily PA („state“), γ10	−0.047 (0.026)	0.011 (0.026)	0.079 (0.033)* [NS]
Day in the study	−0.082 (0.015)*** [***]	−0.131 (0.014)*** [***]	−0.100 (0.015)*** [***]
Location (MSR)	−0.015 (0.088)	0.228 (0.092)* [NS]	0.080 (0.086)
Age	−0.099 (0.063)	−0.037 (0.063)	−0.010 (0.060)
Visceral fat	0.008 (0.060)	0.034 (0.060)	−0.032 (0.058)
CRF VO2peak	−0.029 (0.070)	−0.099 (0.071)	−0.116 (0.068)
Neuroticism	0.261 (0.044)*** [***]	0.146 (0.044)*** [**]	0.082 (0.042)
Fitbit wear time	−0.003 (0.017)	0.025 (0.016)	−0.001 (0.017)
Peri/Post menopausal status	0.117 (0.119)	0.431 (0.119)*** [**]	0.447 (0.115)*** [**]
Daily PA × Location (MSR)	−0.041 (0.032)	−0.056 (0.033)	−0.145 (0.042)*** [**]
Overall PA × Location (MSR)	0.047 (0.087)	−0.034 (0.086)	−0.065 (0.084)
−2LL	6183.6	5779.4	6223.6
AIC	6213.6	5813.5	6257.6
R 2	0.460	0.553	0.465
Number of participants	243	243	243
Number of observations	2505	2516	2515

Note. AIC = Akaike Information Criterion (measure of the relative goodness of fit of the statistical model); CRF=cardiorespiratory fitness; 2LL = −2 Log Likelihood; MSR= Moravian Silesian Region (air polluted region); PA= physical activity; Std.Error= standard errors.; VO_2peak = peak oxygen consumption (in ml/kg/min).

Model is based on data reflecting an average of 14 occasions nested within 243 participants for a total of 3608 observations. Due to some missing data (e.g., Fitbit non-wear), the total number of observations differs between models. Estimates represent standardized estimates. Effects for Overall PA reflect between-person differences. Effects for Daily PA reflect within-person associations.

^*p < .05,

^**p < .01,

^***p < .001.

[p-value indicators in square brackets after Hochberg correction adjusting for multiple comparisons]

The multilevel models indicated that overall level of PA was not significantly associated with overall symptom burden in any symptom category (i.e., no between-person associations were found). However, at the within-person level, daily PA (i.e., doing more PA than was typical on a given day) was associated with more severe hot flashes (t_134.7 = 2.408, p < 0.05). Importantly, there was a significant daily PA by location interaction (t_110.5 = −3.466, p<0.001), such that on days when a woman (from MSR, the air-polluted region) was more physically active than was typical for her, she reported fewer hot flashes.

Effects of covariates on menopausal symptoms

Severity of psychological symptoms was associated with day in the study and neuroticism. With increasing days in the study women reported less burden from symptoms (t₂₂₉₅ = −5.302, p<0.001) and higher levels of neuroticism were associated with more symptom burden (t_240.6 = 5.976, p<0.001).

Somato-vegetative symptoms were associated with the intercept (t_239.9 = −3.106, p<0.01), days in the study t₂₃₀₀ = −9.409, p<0.001), location (t_233.7 = 2.487, p<0.05), level of neuroticim (t_235.1 = 3.356, p<0.001), and peri/postmenopausal status (t_228.2 = 3.622, p<0.001).

Similarly, hot flashes alone were associated with the intercept (t_218.8 = −2.294, p < 0.05), daily PA (t_134.7 = 2.408, p < 0.05), days in the study (t₂₂₉₀ = −6.498, p<0.001), and peri/postmenopausal status (t_219.2 = 3.893, p<0.001). No significant associations were found between any of the symptoms and cardiorespiratory fitness or visceral fat.

DISCUSSION

The findings of this study shed light on the complex relationship between physical activity, ambient air pollution status, and menopausal symptoms among middle-aged women. We followed women from the 4HAIE cohort across two weeks during which their physical activity levels were tracked using Fitbit monitors and symptoms were assessed daily on a mobile app. The first notable finding is that women residing in an area with high ambient air pollution reported more somato-vegetative symptoms. This suggests that living in a high air-polluted area may exacerbate certain menopausal symptoms, potentially due to the adverse effects of air pollutants on physiological processes or through other psychosocial mechanisms. However, the study also uncovered an association between physical activity and menopausal symptoms, specifically hot flashes. On days when women from the high-polluted area engaged in more physical activity than their typical routines (i.e., at the within-person level), they reported fewer hot flashes. This finding suggests that performing more physical activity on a given day may be associated with less severe hot flashes during menopause, despite the presence of environmental pollutants.

The results also indicated that higher level of neuroticism and peri- and post-menopausal status were significant covariates of menopausal symptoms. Neuroticism, characterized by higher levels of emotional instability and reactivity, may contribute to increased reporting of menopausal symptoms⁷². Additionally, the transition into menopause and the post-menopausal stage are periods marked by hormonal fluctuations, which can impact symptom severity. Indeed, studies show higher reporting of menopausal symptoms (vasomotor such as hot flashes, in particular) in the late perimenopausal and early postmenopausal stage^78,79. Women in this study reported only modest levels of symptom burden, which may be related to the menopausal status distribution where 60% of the sample comprised of premenopausal women and there was a smaller number of peri and postmenopausal women, which may have influenced the study results.

Interestingly, the study did not find a significant association between visceral fat or $\dot{V}$O_2peak and menopausal symptoms. This implies that body composition and cardiorespiratory fitness may not be the primary factors influencing symptom reporting in the context of the 4HAIE cohort, in spite of evidence that endurance-trained persons have better thermoregulatory control than untrained persons⁸⁰ and exhibit attenuated blood pressure and heart rate responses compared to unfit individuals and blunting of sympathetic vasomotor activation⁸¹. Other psychological, physiological, or environmental factors may play a more substantial role in the experience of menopausal symptoms among the participants. It should be noted, however, that whereas fitness levels varied substantially within the sample, the sample comprised few women in the overweight/obese category and the relative homogeneity of the sample along with Czech ethnicity may not have captured the visceral fat-menopausal symptom link. Prominent studies on menopause have shown that among correlates of menopausal symptoms are specific endocrine changes, African American race, lower educational attainment or socio-economic status, higher body mass index or body fatness, smoking, history of premenstrual complaints, stress, anxiety, depression, and sleep problems^82–84.

Regardless of visceral fat and cardiorespiratory fitness status, women residing in the highly air-polluted environment, reported higher levels of somato-vegetative symptoms. Although this finding warrants further corroboration, the underlying physiological mechanisms linking ambient air pollution to somato-vegetative symptoms during menopause could include several possible pathways. First, ambient air pollution is known to induce systemic inflammation and oxidative stress in the body^85–87. Chronic exposure to pollutants can trigger an inflammatory response and increase the production of reactive oxygen agents, leading to tissue damage and disruption of normal physiological processes⁸⁷. Inflammation and oxidative stress have been implicated in the occurrence and severity of menopausal symptoms, including hot flashes and sleep disturbances⁸⁸. Second, air pollution exposure has been linked to alterations in the autonomic nervous system (ANS) activity, specifically favoring sympathetic overactivity and/or vagal withdrawal of parasympathetic activity⁸⁹. The ANS co-regulates all physiological processes, including heart rate, blood pressure, and body temperature regulation. Disruptions in ANS balance, increased sympathetic activity as well as reduced vagal modulation of parasympathetic activity⁹⁰ have been associated with increased occurrence of somato-vegetative symptoms, including hot flashes⁹¹. Third, some air pollutants, such as polycyclic aromatic hydrocarbons (PAHs) and phthalates, have been identified as endocrine disruptors⁹². These substances can interfere with hormonal signaling pathways, including those involved in reproductive hormone regulation. Menopause is characterized by a decline in estrogen levels, and any disruption to the hormonal balance caused by pollutants may exacerbate menopausal symptoms or impact menopause onset⁵⁰.

Notably, we observed that performing more physical activity than was typical was associated with less sever hot flashes on the same day, but only in women residing in the highly air-polluted environment. There are several potential explanations for this finding. First, since the overall burden of somato-vegetative symptoms was higher in MSK (the air-polluted region), it is possible that the association transpires only at higher levels of symptom burden. Second, it is possible that the two groups (women from MSK versus SBR) differed in some parameter that we did not measure but one that was also related to hot flashes or the underlying mechanisms between physical activity and hot flashes. Third, the air pollution status variable is a crude index of air pollution exposure and does not reflect the differences in within-region exposure to air pollution, which could have influenced the study results.

The mechanisms underlying the physical activity and menopausal symptom association are not completely understood. The models in this study tested contemporaneous (i.e., same-day) associations between physical activity and menopausal symptoms, precluding determination of the direction of the observed effects. It is equally possible that on days when women experienced fewer hot flashes, they chose to be more physically active, as that more physical activity on a given day helped them deal better with their hot flashes. Still, there are several plausible mechanisms for through which physical activity can act to alleviate menopausal symptoms. Physical activity has been shown to influence the production and regulation of various hormones, including estrogen^93,94 and serotonin⁹⁵. Regular exercise can improve cardiorespiratory fitness, enhance blood circulation, and promote better heart function and so may help mitigate these symptoms through its positive effects on the cardiovascular system regulation. Physical activity is also known to have positive effects on mental health and overall well-being⁹⁶, consequently, the mitigating effects on hot flashes (especially when assessed in terms of severity or bother) could occur through psychosocial mechanisms as well. Engaging in physical activity can reduce stress levels, improve mood, and promote better sleep quality⁹⁷, consequently, the beneficial impact of exercise on mental health may indirectly alleviate the somato-vegetative symptoms experienced during menopause. Further research should examine these pathways through longitudinal studies, qualitative approaches, and interventions targeting psychosocial factors to provide a more comprehensive understanding of the psychosocial determinants of menopausal symptom burden and the role of physical activity.

Study Limitations

The study findings must be interpreted in the light of its limitations. Although this study involved 3608 observations given its intensive longitudinal design, the study sample was rather small and homogeneous. The majority of women (60%) were premenopausal according to their menstrual cycle characteristics and did not report large symptom burden overall. The peri/postmenopausal women in the sample, however, also reported relatively low symptoms, which could reflect the study’s selection criteria and recruitment process. It is possible that women with severe menopausal symptoms self-selected out of the study, leading to a sample that is skewed towards women with milder symptoms. This selection bias could underestimate the true symptom burden experienced by women in the general population. Importantly as well, menopausal symptoms can fluctuate in severity over time, and the women in the study might have been in a phase where their symptoms were relatively milder. Menopausal symptoms are known to vary in intensity and duration throughout the menopausal transition⁷⁸, and the study’s data collection period might have coincided with a period of relatively lower symptom burden for the participants. Further studies affording longitudinal assessments of the changing symptom trajectories across the menopausal transition, especially those designed to examine lagged influences and other dynamics of the physical activity and menopausal symptoms, should be pursued.

The level of air pollution in this study was characterized by area of living and not by individualized level of exposure to specific air pollutants calculated for each participant. There is substantial variation in air pollution parameters within the studied areas themselves. Additionally, the recruitment and data collection process unfolded across two years and the participants entered the study in different parts of the year (the data collection schedule by year, season and region is available in Supplemental Digital Content). Slightly more women from the MSR region (as compared to SBR) underwent data collection in the fall (43 versus 29) and winter (26 versus 14) months. Since air pollution also fluctuates across the year, it will be important for future studies to leverage the use of technology in obtaining dynamic data on personal exposure to air pollutants.

This study included a homogeneous sample (Czech ethnicity) and women of predominantly average or above average socioeconomic standing. Socio-economically disadvantaged women display more negative health behaviors (e.g., smoking, alcohol consumption, body mass index, stress, diet, physical activity)⁹⁸, all of which have been shown to impact age at menopause, menopausal symptoms, and/or health outcomes^99,100. It is possible that the relative sample homogeneity may have influenced the null findings with respect to the impact of body composition and cardiorespiratory fitness status. Importantly, 50% of the study sample identified themselves as active runners and even among the non-runners, physical activity was fairly high. It is possible that this may explain why overall level of physical activity was not associated with any of the symptoms. Endurance trained persons such as runners as well as highly active individuals display a number of physiological adaptations (enhanced fitness, better thermoregulation, lower adiposity, etc.) that have been linked with less severe menopause symptomology. It maybe that the “mitigating” effects of physical activity on symptoms appear primarily at lower-to-moderate levels of physical activity. Future studies with more diverse samples of women and women with higher symptom burden are warranted to evaluate the visceral fat, cardiorespiratory fitness, menopausal symptom association. More importantly, as there was observable heterogeneity in the PA-symptom trajectories, future studies should evaluate person-specific associations to better depict the person-specific experiences of menopausal symptoms, their dynamics and their within-person predictors¹⁰¹, paving way to personalized and context-adaptive interventions.

CONCLUSIONS

Overall, these findings highlight the importance of considering both individual and environmental factors when studying menopausal symptoms. Although the association between high ambient air pollution status and somato-vegetative symptoms should be considered preliminary, pending further corroboration, this study emphasizes the need to consider the impact of the environment and pursuing interventions aimed at reducing pollution levels or mitigating its effects. Additionally, the same-day association between higher levels of physical activity and lower severity of hot flashes suggests that promoting regular physical activity may be a valuable strategy for managing menopausal symptoms. Unlike air pollution or neuroticism, which cannot be easily changed or controlled, physical activity represents a modifiable health behavior under personal control that could help lower perceived menopausal symptom burden. Future research should further explore the underlying mechanisms behind these relationships and consider additional factors that may influence menopausal symptomatology to develop comprehensive approaches for improving women’s health during the menopausal transition.