Menopausal Hormone Therapy and Change in Physical Activity in the Women’s Health Initiative Hormone Therapy Clinical Trials

Abstract

Objective:

The menopausal transition results in a progressive decrease in circulating estrogen levels. Experimental evidence in rodents has indicated that estrogen depletion leads to a reduction of energy expenditure and physical activity. It is unclear whether treatment with estrogen therapy increases physical activity level in postmenopausal women.

Methods:

among 27,327 postmenopausal women aged 50–79 years enrolled in the Women’s Health Initiative randomized double-blind trials of menopausal hormone therapy. Self-reported leisure-time physical activity at baseline, and years 1, 3 and 6 was quantified as metabolic equivalents [MET]-hr/week. In each trial, comparison between intervention and placebo groups of changes in physical activity levels from baseline to follow-up assessment were examined using linear regression models.

Results:

In the CEE-alone trial, the increase in MET-hr/week was greater in the placebo group compared to the intervention group at years 3 (p=0.002) and 6 (p<0.001). Similar results were observed when analyses were restricted to women who maintained an adherence rate ≥80% during the trial, or who were physically active at baseline. In the CEE+MPA trial, the primary analyses did not show significant differences between groups, but the increase of MET-hr/week was greater in the placebo group compared to the intervention group at year 3 (p=0.004) among women with an adherence rate ≥80%.

Conclusions:

The results from this clinical trial do not support the hypothesis that estrogen treatment increases physical activity among postmenopausal women.

INTRODUCTION

Regular recreational physical activity (PA) is considered to be an important contributor to physical and mental health throughout the entire lifespan.¹ Physical activity helps to maintain a normal body weight and muscle strength, and is associated with a reduction in the risk of developing several chronic diseases including metabolic syndrome, type 2 diabetes, heart disease, stroke, osteoporosis, dementia, and cancer, amongst others.^2–6 The World Health Organization recommends an average weekly duration of 150–300 minutes of moderate intensity or 75–150 minutes of vigorous intensity physical exercise, or an equivalent combination for adults aged 18–64 years, and 150 minutes of moderate or 60 minutes of vigorous intensity aerobic activity, plus muscle-strengthening, flexibility, and balance exercise, for adults aged 65 years or older.^7,8

Physical activity is a complex behavioral trait influenced by multiple physical, psychological and environmental factors.⁹ In women, particular attention is given to physical activity during and after menopause,¹⁰ when there is a considerable reduction of ovarian estrogen production, and in particular, of estradiol (E2). E2 regulatory activity in the body extends beyond the reproductive system to the cardiovascular, musculo-skeletal, and nervous systems.^11,12 E2 is involved in cellular energy homeostasis,¹³ influencing resting and exercise metabolic rate, and body composition.¹⁴ During the transition from premenopausal to postmenopausal status, changes in resting metabolic rate and total energy expenditure occur.¹⁵ These metabolic changes are often accompanied by a decrease in total energy expenditure, an increase in sedentary time,¹⁶ and a decline in total and leisure time physical activity.^17,18

Estrogen treatment is associated with several health benefits including the reduction of menopausal vasomotor symptoms, such as hot flashes and night sweats, genitourinary syndrome, osteoporosis, and all-cause mortality.^19,20 However, high levels of endogenous estrogen are associated with increased risk of various cancers including those of the breast, endometrium, ovaries, and colon,²¹ similar association has been found between estrogen therapy alone or in co-administration with a progestin and cancer risk,^22–29 although, several studies have shown different results. ^29–31

The idea that estrogen might influence female physical activity levels has been reviewed in detail elsewhere.³² Experimental studies in female rodents have shown that estrogen levels play an important role in the regulation of in energy expenditure and physical activity.^33,34 However, results from previous studies that have evaluated the association between hormone therapy and physical activity in postmenopausal women have been inconsistent.^35–38 In the present study, we examined the relationship between estrogen therapy (with or without progestin) on the level of leisure time physical activity among postmenopausal women participating in the Women’s Health Initiative (WHI) menopausal hormone therapy randomized trials.

METHODS

Subjects and study design

The WHI randomized controlled trial comprises three partially overlapping components: menopausal hormone therapy (HT) (2 trials), dietary modification (DM), and calcium-vitamin D (CaD) supplementation.³⁹ The HT trials were double blind placebo-controlled trials designed to evaluate the risks and benefits of estrogen treatment, with or without progestin, in healthy postmenopausal women age 50 to 79 years at enrolment. The primary outcomes of the hormone trials were coronary heart disease and invasive breast cancer. Details regarding the design and rationale of the trials, including the randomization procedure and sample size, have been published elsewhere.³⁹ Briefly, between 1993 and 1998, postmenopausal women who were cancer free (excluding non-melanoma skin cancer), with no medical conditions with predicted survival of less than 3 years, and with no adherence concerns (e.g., alcoholism, dementia), were invited to participate to the HT trials at 40 clinical centers across the country. Women with an intact uterus (n=16,608) were randomized to receive oral conjugated equine estrogens (CEE, 0.625 mg/day) plus medroxyprogesterone acetate (MPA, 2.5 mg/day), or placebo, and were followed until trial termination after a median follow-up period of 5.6 years. Women who had had a hysterectomy (n=10,739) were randomized to CEE-alone or placebo and were followed for a median duration of 7.2 years. Some of the women enrolled in the HT trials were also enrolled in the DM trial (n=8,050), and in the calcium-vitamin D supplementation trial (n=16,069), which started subsequently. Participants in the HT trials who were currently taking HT underwent a 2-month wash-out period before randomization. No recommendation on making changes in physical activity were made in the HT trials. Adherence to the assigned regimen was evaluated every 6 months based on the weight of the pills and containers on a calibrated scale.⁴⁰ Institutional review board approval was obtained at the Fred Hutchinson Cancer Center and at each clinical center, and all study participants provided written informed consent.

Assessment of Physical Activity

Women enrolled in the clinical trials completed a questionnaire on frequency (times/week), duration (hours), and intensity (mild, moderate, vigorous) of leisure time physical activity at baseline, and at years 1, 3, and 6 of the study. The year 1 questionnaire on personal habits (e.g., physical activity) was not-administered until June 1997. Consequently, corresponding data are missing for participants randomized between 1993 and the first half of 1996. Data on physical activity were summarized as total leisure time energy expenditure score by multiplying the metabolic equivalent (MET) intensity values for each activity by the hours per week (hr/wk) of reported participation and summing over all activities (MET-hr/wk).⁴¹

Assessment of Covariates

Information on demographic, health, and reproductive characteristics was collected at baseline using self-administered questionnaires, while anthropometric measurements including weight, height, and waist circumference were obtained at the clinic visits by trained personnel and repeated at the annual follow-up visits for the duration of the trial intervention phase. Weight (kg) and height (cm) were used to calculate body mass index (BMI) (weight (kg)/height (m²)), which was categorized into groups (underweight, <18.5; normal weight, 18.5-<25.0; overweight, 25.0 –<30.0; obesity I 30.0-<35.0, obesity II & III ≥35.0 kg/m², and missing). Blood pressure was measured at baseline and at each annual visit by certified staff using standardized procedures and instruments. Incident cardiovascular disease events were identified by contacting the trial participants every 6 months, while cancer and fracture occurrence were reported annually using a health status questionnaire. Cases of cardiovascular disease, cancer, and hip fracture were centrally adjudicated based on medical records review.

Statistical Analysis

All analyses were conducted separately in the two trials based on the intention-to-treat principle. The distribution of physical activity level was highly skewed to the right, with approximately 26% of the trial participants at baseline reporting that they were inactive (MET hr/wk=0); therefore, we used quantile regression with 100 bootstrap samples to compare the low (0.26), median (0.50) and high (0.75, 0.90) percentiles of the distribution of physical activity at different times during the trials (baseline, year 1, 3 and 6) between the two arms in each trial. This statistical approach allows examination of a non-linear relationship between two variables without the need to categorize participants.⁴² Changes in physical activity levels from baseline to a specific year (1, 3, and 6) separately or in relation to the overall average throughout the intervention phase of the trial (physical activity (PA)_change=PA_average–PA_baseline), were normally distributed and compared using multivariable linear regression with adjustment for the baseline level of physical activity. Since estrogen treatment and physical activity reduce postmenopausal vasomotor symptoms, a subgroup analysis was performed by the baseline reported vasomotor symptoms. In addition, because in this cohort the effects of the menopausal hormone treatment on health outcomes varied by age and time since menopause,⁴³ we performed two separate subgroup analyses: age </≥ 60 years, and time since menopause </≥ 10 years. We performed an analysis excluding women inactive at baseline to evaluate whether changes in the level of physical activity during the trial were due to women who were inactive at baseline becoming active during follow-up, or to those who were active at baseline and changed their levels of physical activity over time. We also performed an analysis including only women who remained adherent to the intervention (i.e. using ≥80% of the study pills for the intervention group and not initiating hormone treatment in the placebo group). Sensitivity analyses were conducted including as covariates in the model variables representing the occurrence of cardiovascular disease, cancer, and hip fractures between baseline and the specific year when physical activity was reported, and body mass index as measured at the specific yearly exam. At each time point, we created a dichotomous variable indicating physical inactivity/activity (MET hr/wk = 0 />0), to evaluate if the trial intervention affected the probability of remaining physically active, and of becoming physically active over time. For the former of these analyses, we used a generalized estimating equation for longitudinal repeated outcomes to account for the inherent correlation of the observations within each individual, and for each trial arm we estimated the odds ratios and 95% confidence intervals (CI) of remaining active at each reported year compared to the baseline. For the latter analysis, we used a random-effects logistic regression model for discrete time-repeated events and we included only women who were not active at baseline (MET h/wk=0).

All statistical analyses were performed using STATA 17 statistical software. All p-values were two-sided and considered statistically significant when <0.05.

RESULTS

Baseline characteristics of the HT participants by study intervention have been previously published.⁴⁴ Of the 27,327 women enrolled in the hormone trials, 58% reported data on physical activity at year 1, 87% at year 3, and 88% at year 6. In both arms of the two trials, there was an increase in the average level of physical activity over time (Table 1). However, when the analysis was restricted to women who were active at baseline (MET hr/wk >0), we observed a reduction in the level of physical activity over time (Table 1). In the CEE-alone trial, cross-sectional analyses showed that the baseline and year 1 distribution of the level of physical activity was similar between the two arms (Table 1, Supplementary Figure 1), while, at years 3 and 6, the mid and upper percentiles of the distribution of physical activity level were significantly higher in the placebo arm compared to the intervention arm (Supplementary Figure 1). Changes in physical activity level were significantly higher in the placebo group compared to the intervention group at year 3 (p=0.002), year 6 (p<0.001), and throughout the trial intervention period overall (p<0.001). Limiting the analysis to women who were physically active at baseline showed a significantly lower reduction of physical activity in the placebo group compared to the intervention group at year 3 (p=0.011), year 6 (p=0.027), and overall (p=0.004) (Table 1). When the analysis was restricted to women who maintained an adherence rate of >80% to the CEE trial protocol the results remained significant at year 3 (p=0.043), year 6 (p=0.024), and overall (p=0.007). In separate analyses adjusting for body mass index, and for cardiovascular disease, cancer and fractures that occurred between baseline and the follow-up assessment of physical activity, we observed similar results (Supplementary Table 1). Approximately 70% of women reported vasomotor symptoms at baseline. Analyses stratified by the presence of these symptoms showed that in CEE-alone trial, changes in physical activity from baseline to year 1 and year 6 were similar in both arms of the trial, independent of symptoms, whereas at year 3 there was difference in the changes of physical activity between intervention and placebo group only in the group with vasomotor symptoms (Supplementary Table 2). Analyses stratified by age and by time since menopause showed results comparable to those obtained in the main analysis (Supplementary Table 3).

Table 1.

Median levels and average changes in physical activity during the menopausal hormone therapy trials

	Hormone therapy clinical trials
	CEE			CEE + MPA
	intervention	placebo	p-value	intervention	Placebo	p-value
baseline, n	5,305	5,427		8,503	8,092
median a	4.25 (0–12.50)	4.00 (0–12.50)	0.725	5.50 (0–15.50)	6.50 (0.75–16.33)	<0.001
year 1, n	3,046	3,054		4,871	4,868
median a	5.31 (1.00–14.25)	5.75 (1.17–14.25)	0.402	7.25 (1.67–17.25)	7.50 (1.88–17.46)	0.291
year 3, n	4,559	4,634		7,443	7,170
median a	5.25 (0.75–14.50)	5.58 (1.00–15.00)	0.486	7.50 (1.50–17.08)	7.50 (1.88–17.75)	0.999
year 6, n	4,590	4,685		7,560	7,228
median a	5.00 (0.50–13.50)	5.43 (0.75–15.00)	0.032	7.50 (1.25–17.50)	7.50 (1.88–17.50)	0.980
Changes in physical activity b
Total trial cohort
baseline – yr 1	0.43 (4.51)	0.65 (4.51)	0.290	0.42 (4.32)	0.41 (4.45)	0.687
baseline – yr 3	0.56 (5.87)	1.30 (5.81)	0.002	0.96 (5.20)	1.05 (5.34)	0.133
baseline – yr 6	0.46 (7.01)	1.34 (7.05)	<0.001	1.47 (6.44)	1.09 (6.65)	0.756
baseline –yrs 1–6, n	5,078	5,186		8,213	7,845
change	0.80 (10.83)	1.54 (11.44)	<0.001	1.39 (5.67)	1.21 (5.80)	0.591
Including only women physically active at baseline
baseline – yr 1, n	2,349	2,385		3,952	3,979
change	−0.39 (4.89)	−0.07 (4.88)	0.291	−0.25 (4.55)	−0.26 (4.65)	0.775
baseline – yr 3, n	3,402	3,460		5,740	5,703
change	−0.92 (6.12)	−0.21 (6.05)	0.011	−0.45 (5.24)	−0.05 (5.37)	0.034
baseline – yr 6, n	3,301	3,375		5,569	5,600
change	−1.47 (7.29)	−0.87 (7.33)	0.027	−0.58 (6.43)	−0.69 (6.60)	0.908
baseline – yr 1–6, n	3,660	3,725		6,062	6,065
change	−1.04 (11.31)	−0.41 (11.49)	0.004	−0.50 (5.74)	−0.35 (5.83)	0.236
Including only women with ≥ 80% adherence rate
baseline – yr 1, n	2,435	2,589		3,821	4,190
change	0.56 (4.21)	0.61 (4.34)	0.503	0.34 (4.20)	0.43 (4.33)	0.358
baseline – yr 3, n	2,831	3,008		4,681	5,026
change	0.76 (5.55)	1.26 (5.50)	0.043	0.79 (5.06)	1.24 (5.03)	0.004
baseline – yr 6, n	2,052	2,150		2,229	2,223
change	1.08 (6.87)	1.80 (6.45)	0.024	2.37 (6.68)	1.77 (6.56)	0.596
baseline – yr 1–6, n	2,102	2,193		2,283	2,268
change	1.15 (6.13)	1.86 (5.75)	0.007	2.20 (5.96)	1.94 (5.86)	0.769

^aResults represent medians (interquartile ranges) of physical activity (MET hr/wk). Equality of medians between the groups were tested using quantile regression.

^bMultivariable liner regression analysis adjusted age and physical activity (MET hr/wk) at baseline.

Abbreviations: CEE, conjugated equine estrogens; MPA, medroxyprogesterone acetate; PA, physical activity, MET, metabolic equivalent; yr, year

In the CEE+MPA trial, at baseline the median level of physical activity was significantly higher in the placebo group compared to the intervention (p<0.001), but no difference in the change of physical activity levels over time were observed between the two arms (Table 1). Comparison of the distribution of physical activity levels showed significantly higher levels in the placebo arm compared to the intervention arm at baseline and at year 3 up across most of the distribution, but no difference at year 1 and year 6. (Supplementary Figure 1). Amongst women physically active at baseline, there was a lower reduction of physical activity level in the placebo arm at year 3 (p=0.034) (Table 1); comparably, there was a significantly higher increase in physical activity at year 3 (p=0.004) in the placebo group when only women with an adherence rate of >80% were included. In analyses stratified by the presence of vasomotor symptoms as reported at baseline, results were similar to those observed in the CEE-alone trial. Specifically, changes of physical activity between baseline and year 1 and year 6 were similar in the two groups (symptoms and no-symptoms), but women with vasomotor symptoms showed a borderline-significantly higher increase in physical activity in the placebo arm compared to the intervention arm at year 3 (p=0.072), while there was no difference between the two groups among those without symptoms (p=0.455) (Supplementary Table 2). Analyses stratified by age showed that the difference in the baseline median levels of physical activity between the two arms was mainly observed amongst older women (≥60 years or ≥10 years since menopause), while average changes in physical activity levels between the arms of the trial were similar (Supplementary Table 3).

Percentages and odds of remaining physically active during the trial were similar between the two arms of each trial and showed that, compared to baseline, a higher number of women in all trial arms reported some level of physical activity during the trial, both in the total sample and in those who maintained an adherence rate ≥80% (Table 2). Over time, the odds of becoming physically active among women who were inactive at baseline were mostly similar between the intervention and placebo arms of both trials; however, in the CEE-alone, women in the placebo arm had a 25% (95% CI 1.01–1.56) greater odds of becoming active at year 6 compared to those in the intervention arm (Table 3); a similar pattern was shown among women who maintained an adherence rate ≥80% (38% increase in the placebo).

Table 2.

Odds ratios and 95% confidence intervals of staying physically active (MET hr/wk > 0) during the clinical trials

	Hormone therapy clinical trials
	CEE trial		CEE + MPA trial
	intervention	placebo	intervention	placebo
Total sample
baseline, %	71.61	71.16	73.43	77.11
OR (95% CI)	1.00	1.00	1.00	1.00
Year 1, %	78.92	80.12	82.91	83.93
OR (95% CI)	1.40 (1.28–1.53)	1.52 (1.39–1.66)	1.59 (1.48–1.71)	1.42 (1.31–1.53)
Year 3, %	78.68	79.18	81.98	83.31
OR (95% CI)	1.40 (1.30–1.51)	1.46 (1.35–1.58)	1.54 (1.45–1.63)	1.41 (1.32–1.51)
Year 6, %	75.05	76.86	80.19	82.73
OR (95% CI)	1.18 (1.09–1.27)	1.32 (1.22–1.43)	1.45 (1.36–1.54)	1.40 (1.31–1.49)
Including only women who maintained an adherance rate ≥80% during the intervention period
Year 1, %	79.47	80.57	82.86	83.74
OR (95% CI)	1.44 (1.31–1.59)	1.53 (1.39–1.68)	1.55 (1.43–1.67)	1.39 (1.28–1.50)
Year 3, %	79.76	80.44	81.48	84.36
OR (95% CI)	1.48 (1.35–1.61)	1.56 (1.42–1.71)	1.47 (1.39–1.60)	1.49 (1.39–1.61)
Year 6, %	76.40	79.33	82.02	83.45
OR (95% CI)	1.28 (1.16–1.41)	1.48 (1.34–1.63)	1.73 (1.60–1.91)	1.56 (1.41–1.72)

Generalized estimating equation for longitudinal repeated outcomes were used to make comparisons within the same group at different time point accounting for inherent correlation of the observations within each individual. The results represent odds ratios and 95 % confidence intervals of being active at each year (1, 3, and 6) compared to the intervention group at baseline.

Abbreviations: CEE, conjugated equine estrogens; MPA, medroxyprogesterone acetate; PA, physical activity, MET, metabolic equivalent; OR, odds ratio; CI, confidence interval.

Table 3.

Odds ratios (OR) and 95% confidence intervals (CI) of becoming active among women not physically active (MET hr/wk=0) at baseline

	Hormone therapy trial arms
	CEE trial		CEE + MPA trial
	intervention	placebo	intervention	placebo
Women not active at baseline, n	1,506	1,565	2,259	1,852
Year 1, % a	47.63	46.79	45.27	47.81
OR (95%CI)	1.00	0.99 (0.74–1.33)	1.00	1.05 (0.81–1.37)
Year 3, % a	57.48	56.64	58.31	57.87
OR (95%CI)	1.00	0.94 (0.75–1.18)	1.00	0.98 (0.80–1.21)
Year 6, % a	57.64	61.45	63.94	64.43
OR (95%CI)	1.00	1.25 (1.01–1.56)	1.00	1.02 (0.84–1.25)
Including only women who maintained an adhesion rate ≥ 80% during the intervention period
Year 1, % a	47.02	47.31	44.35	47.20
OR (95%CI)	1.00	1.09 (0.79–1.50)	1.00	1.05 (0.78–1.40)
Year 3, % a	58.50	57.39	57.28	59.09
OR (95%CI)	1.00	0.92 (0.70–1.22)	1.00	1.12 (0.87–1.44)
Year 6, % a	58.74	63.81	69.55	67.03
OR (95%CI)	1.00	1.38 (1.00–1.89)	1.00	0.85 (0.60–1.20)

The odds ratios and 95% confidence intervals compared placebo group to the intervention group.

^aPercentage of women who become active (MET hr/wk >0) among those inactive and with available data.

Abbreviations: CEE, conjugated equine estrogens; MPA, medroxyprogesterone acetate; PA, physical activity, MET, metabolic equivalent; OR, odds ratio; CI, confidence interval.

DISCUSSION

The present study investigated the effect of menopausal hormone therapy on physical activity levels among postmenopausal women enrolled in two randomized clinical trials, one of CEE-alone, and the other of CEE+MPA. During the trial intervention period, an increase in the level of physical activity was observed in all groups, although this increment was mainly due to the fact that some women who were inactive at baseline became active during the follow-up period. This change in reported behavior might have occurred as a response to the enrollment and engagement in the trial, as observed in other clinical trials.⁴⁵ In the CEE-alone trial, the two arms had similar levels of physical activity at baseline, but compared to women in the intervention group, those in the placebo group had a greater increase in physical activity level at years 3 and 6. When we excluded from the analysis women who were inactive at baseline, we found that the placebo group had a lower reduction of physical activity at years 3 and 6, while women in the placebo arm who were inactive at baseline had higher odds of becoming physically active at year 6 compared to the intervention arm. In the CEE+MPA trial, the baseline physical activity level was higher in the placebo group compared to the intervention group; however, the level of physical activity, and the numbers of women who become active and who remained active during the trial were similar in the two groups. We observed that the percentage of women engaging in some level of physical activity increased in all groups over time, but in both trials, it was consistently higher in the placebo group compared to the intervention group. Analyses stratified by age and, separately, by time since menopause, showed similar associations compared to those observed in the main analyses.

Physical activity has been associated with a reduced risk of developing several chronic conditions, including cancer.⁶ In postmenopausal women, substantial evidence indicates the importance of physical activity in improving health and quality of life and reducing the symptoms associated with menopause.¹⁰ Experiments in rodents have indicated that E2 interacts with neuronal circuits that regulate physical activity.^{33,34,46–48} Neuroimaging studies in humans have shown both structural and functional cerebral changes following the reduction of E2 production during and after the menopause,^49,50 and in-situ hybridization experiments in a primate model have found that specific hypothalamic E2 receptor-positive neurons, which modulate various functions including locomotion, decline during menopause,⁵¹ and that subsequent treatment with estrogen or estrogen plus progesterone cannot restore the neurons in these areas.⁵² Based on these data, we might hypothesize that, in postmenopausal women, treatment with estrogen does not substantially affect these areas due to irreversible changes that have occurred. In women, reports on physical activity have suggested a decline in physical activity during the menopausal period.⁵³ Previous epidemiological studies that have investigated the relationship between postmenopausal hormone use and physical activity had small sample sizes. A cross-sectional study conducted among 248 postmenopausal women found an increased exercise capacity in women using hormone treatment compared to those not currently using it, but no difference in the level of physical activity between the two groups.³⁵ A small intervention study compared resting energy expenditure and physical activity level in 33 postmenopausal women before and during the administration of estrogen alone or in combination with progesterone to achieve premenopausal physiological circulating levels of the hormones and found no difference in these functional parameters.³⁶ A three-year, double blind randomized trial of low-level estrogen treatment in 167 old postmenopausal women (mean age 74 years) found no difference in the rate of change of physical activity level over time between the intervention and placebo groups.³⁷ A survey among 3,479 postmenopausal women who participated to the Third National Health and Nutrition Examination Survey showed that prevalence of physical inactivity was higher among women who had never used HT compared to those who used it.³⁸ The present study is the largest that has examined the relationship between estrogen treatment and physical activity level in postmenopausal women. The study found that estrogen treatment was not associated with an increase in the level of physical activity. A previous analysis conducted in this cohort evaluated performance-based measures of physical functions including grip strength, chair stand and timed walk at baseline, year 1, 3 and 6 and found no difference in the rates of decline of these measures between intervention and placebo groups of both trials.⁵⁴

The CEE-alone trial showed differences in physical activity level between women in the two arms, with the placebo group having greater increase in physical activity quantitatively, namely, an absolute change in MET hr/wk, and qualitatively, namely, a change from physically inactive to active compared to the treated group. In the CEE+MPA trial we did not observe differences between the two arms. These findings raise the question as to whether the results may be related to the type of treatment. As previously reported, in both WHI HT trials, the intervention was associated with a significant reduction of vasomotor symptoms such as hot flashes and night sweats, and modest reduction of join stiffness.^55,56 However, in the CEE-alone trial, the prevalence of hot flashes and night sweats at baseline was higher compared to that in women of the same age who participated in the CEE+MPA trial, while within the CEE-alone trial the incidence of vasomotor symptoms measured before stopping the treatment was greater among women in the placebo group compared to those in the intervention group.⁵⁵ It is possible that women in the placebo groups, and in particular those in the CEE-alone trial who did not experience these reductions in symptoms, were more motivated to engage in physical activity or to increase their original level of activity, as physical exercise is associated with a reduction of these menopausal-related symptoms.⁵⁷ Using data on vasomotor symptoms obtained at baseline, we observed that while among women with no vasomotor symptoms there was no between-group difference in the change in physical activity by year 3, while in women who reported these symptoms the level of change in physical activity was higher in placebo group compared to the treated one. Additional health-related factors may have contributed to an engagement in physical activity. In the CEE-alone trial, the incidence of joint pain and stiffness was similar between the two trial arms at year 1, but it was lower in the placebo group (52.9%) compared to the intervention one (58.0%) by the time the study ended; ⁵⁵ this difference might have contributed in part to a higher percentage of women in the placebo arm being physically active. Information on personal motivation related to engagement in physical activity and the perceived health benefits of it could potentially provide some support for these results.

The present study was performed in a randomized controlled trial setting. Therefore, baseline demographic and health-related characteristics were similar between the two groups within each trial. The blinding to randomization arm would have reduced in part the possibility of bias in reporting personal behaviors, including physical activity. In the sensitivity analyses, we adjusted for events that occurred during the follow-up such as cancer, cardiovascular disease, fractures, as well as for body mass index, since they could have been related to both the treatment and the outcome. We also used data on intervention adherence to reduce the possibility that the results were due to a lack of treatment compliance. Leisure time physical activity was reported at multiple time points during the trial intervention phase. In addition to these strengths, the study also has some potential limitations. First, we used self-reported data on physical activity, and although this measuring tool has been validated in the WHI⁵⁸ and used previously in other cohort studies,^59,60 measurement error cannot be excluded. Second, we used data from two trials designed to evaluate different outcomes from the one examined in this study; therefore, the study power calculations were not based on evaluation of change in physical activity as the outcome. Third, both trials protocols included only one dosage of estrogen with or without a progestin.

CONCLUSIONS

The current study examined the relationship between hormone therapy and future levels of leisure time physical activity in the Women’s Health Initiative menopausal hormone therapy trials. The findings from the clinical trials suggest that in post-menopausal women, treatment with oral conjugated equine estrogens alone or in combination with medroxyprogesterone acetate does not increase the level of leisure time physical activity.