Abstract
Background:
The nature of the relationship between adiposity and hot flashes has been debated, but it has not been examined using physiological measures of hot flashes. We examined associations between body size/composition and physiologically assessed hot flashes among women with hot flashes.
Methods:
A subcohort of women in the Study of Women's Health Across the Nation (n = 52; 25 African-American and 27 non-Hispanic Caucasian; ages, 54 to 63 yr) who reported hot flashes, had their uterus and ovaries, and were not taking medications impacting hot flashes were recruited in 2008–2009. Women completed anthropometric measures [bioimpedance analysis of total percentage of body fat, body mass index (BMI), waist circumference], a blood draw (estradiol, SHBG, FSH, dehydroepiandrosterone sulfate), and 4 d of ambulatory sternal skin conductance monitoring with diary (physiological and reported hot flashes, respectively). Associations between anthropometrics and hot flashes were estimated with generalized estimating equations with covariates age, race, and anxiety.
Results:
Higher BMI (odds ratio, 0.97; 95% confidence interval, 0.94–0.99; P < 0.05) and waist circumference (odds ratio, 0.98; 95% confidence interval, 0.97–0.99; P < 0.01) were associated with fewer physiological hot flashes. Interactions by age (P < 0.05) indicated that inverse associations of body fat, BMI, and waist circumference with hot flashes were most apparent among the oldest women in the sample. Estradiol and SHBG reduced but did not eliminate age-related variations in relations between body size/composition and hot flashes.
Conclusion:
Higher adiposity was associated with fewer physiological hot flashes among older women with hot flashes. A modifying role of age must be considered in understanding the role of adiposity in hot flashes.
Most women transitioning through the menopause report hot flashes (1). For many women, hot flashes are frequent or severe (2). Despite the well-documented impairments in quality of life associated with hot flashes (3), their underlying physiology and risk factors require further clarification.
One risk factor for hot flashes is body size and composition. Body fat has traditionally been considered protective against hot flashes. This perspective, dubbed the “thin hypothesis,” is informed by an endocrine model of hot flashes in which hot flashes occur due to the pronounced reproductive hormonal changes, including declining estrogens, characteristic of the menopausal transition (4, 5). Because androgens are aromatized into estrogens in body fat, women with more body fat should have higher estrogen levels and fewer hot flashes. Thinner women, lacking this extragonadal estrogen source, should have more hot flashes. This effect would be particularly relevant for older and postmenopausal women in whom extragonadal sources represent the primary source of estrogen (6).
Recent research has called the thin hypothesis into question. Several epidemiological investigations have shown that women with higher body mass index (BMI) (1) and body fat (7, 8) are more likely to report hot flashes. These findings are consistent with a thermoregulatory model in which hot flashes represent heat dissipation events occurring in the context of the narrowed thermoneutral zone of menopausal women (9). Body fat would thereby act was an insulator, increasing the occurrence of hot flashes.
One limitation to this work is that it is largely based upon epidemiological studies using crude, questionnaire measures of hot flashes. These measures, which ask women to recall their hot flashes over weeks or months, have notable weaknesses including reporting and memory biases (10) and the influence of negative affect (11). Furthermore, highly symptomatic women have been underrepresented in these epidemiological studies, and more detailed questions about the frequency of symptoms were not included. Thus, these studies largely allow comparison of women with and without hot flashes. Whether body fat shows a dose-response relation to hot flashes among symptomatic women is unknown.
The study aim is to test relations between body size/composition (body fat percentage, waist circumference, BMI) and hot flashes (physiologically assessed and self-reported) among women with hot flashes. We examine whether these associations vary by age/menopausal stage and race/ethnicity, given that these factors may modify these relationships (6, 12–17). Finally, we examine the role of estradiol (E2), FSH, SHBG, and the free E2 index (FEI; an estimate of E2 unbound to SHBG and thereby bioavailable) in these associations given their relations to hot flashes and body size (18–20).
Subjects and Methods
Study population
The study sample was a subcohort of participants (n = 52) of the Pittsburgh site of the Study of Women's Health Across the Nation (SWAN). SWAN is a cohort study designed to characterize the menopausal transition conducted at seven sites across the United States. Details of SWAN procedures have been reported previously (21). At enrollment (1996–1997), SWAN participants (n = 3302) were aged 42–52 yr, had an intact uterus and at least one ovary, were not pregnant or breast feeding, had menstruated within 3 months, and were not using oral contraceptives or hormone therapy.
A subcohort of participants at the Pittsburgh SWAN site participated in SWAN FLASHES, an ancillary study examining relations between body size/composition and hot flashes. SWAN FLASHES assessments occurred during 2008–2009, most closely corresponding to SWAN′s 10th annual visit. By design (21), the Pittsburgh site recruited only non-Hispanic Caucasian and African-American women. Therefore, SWAN FLASHES participants described themselves as non-Hispanic Caucasian or African-American. SWAN FLASHES inclusion criteria included reporting any hot flashes or night sweats in the past 2 wk; having a uterus and both ovaries; not being pregnant; not using hormone therapy, oral contraceptives, or selective serotonin reuptake inhibitor/serotonin-noradrenaline reuptake inhibitor antidepressants within 3 months; and not currently undergoing chemotherapy for breast cancer. Furthermore, given the unequal distribution of BMI/body fat by race/ethnicity (7, 19, 22), there was effort to match participants by race/ethnicity on obesity category (<25 kg/m2, lean; 25–29.9 kg/m2, overweight; ≥30 kg/m2, obese) to examine relations between body composition and hot flashes apart from race/ethnicity. SWAN FLASHES included 52 women (25 African-American, 27 Caucasian women) in primary models. Due to missing adiposity (n = 1) and reproductive hormone (n = 4) data, 51 women were included in models incorporating body fat and 48 women in models incorporating hormones.
Design and procedures
At SWAN entry and annually thereafter, participants completed a protocol including questionnaires, blood specimens, and bioimpedance analysis (BIA) (23). During SWAN visit 10, women meeting eligibility criteria were invited to participate in SWAN FLASHES. During their SWAN FLASHES visit, height, weight, and waist circumference were measured, questionnaires were administered, and participants were equipped with physiological hot flash monitors and an electronic hot flash diary. All women underwent hot flash monitoring as they went about their daily activities for 96 h, conducted in the context of two separate 48-h sessions within approximately 4 wk. Procedures were approved by the University of Pittsburgh Institutional Review Board. Participants provided written informed consent.
Anthropometric measures
Total percentage of body fat
Adiposity was estimated from BIA (BIA-103 analyzer; RJL Systems, Mt. Clemens, MI) at the SWAN visit. BIA is based on measurement of transmission speed of an electrical pulse between electrodes attached at the feet and the knuckles of the hand. Electrical conductivity is greater in fat-free than in fat mass, and thereby resistance and reactance can be used to estimate fat and lean mass (23, 24). Sex-specific validation equations of Chumlea et al. (25) were used.
Body mass index
Height and weight were measured at the SWAN FLASHES assessment via a fixed stadiometer (Seca, Hanover, MD) and calibrated balance beam scale (Healthometer Inc., Alsip, IL), respectively. BMI was calculated (kilograms divided by meters squared).
Waist circumference
At the SWAN FLASHES assessment, waist circumference was measured at the natural waist or the narrowest part of the torso from the anterior aspect. If a waist narrowing was difficult to identify, the measure was taken at the smallest horizontal circumference between the ribs and iliac crest.
Hot flashes
Hot flash monitoring was conducted with an ambulatory sternal skin conductance monitor and electronic diary. Sternal skin conductance was recorded via the Biolog monitor (model 3991/2-SCL; UFI, Morro Bay, CA), a portable device worn in a pouch around the waist. The Biolog measures sternal skin conductance sampled at 1 Hz from the sternum via a 0.5-volt constant voltage circuit passed between two Ag/AgCl electrodes (UFI) filled with 0.05 m KCL Velvachol/glycol paste (26). Participants were instructed to avoid exercising and showering during monitoring.
Physiological hot flashes were classified via standard methods, with skin conductance rise of 2 μmho in 30 sec (27) flagged automatically by UFI software (DPSv3.6) and edited for artifact (28). Given that some women show submaximal hot flashes failing to reach the 2-μmho criterion (29, 30), all potential hot flash events were also visually inspected, and events showing the characteristic hot flash pattern yet less than 2 μmho/30 sec rise were coded as hot flashes. This coding has been shown to be reliable (κ = 0.86) (29, 30). A 20-min lockout period (during which no hot flashes were coded) was implemented after the start of the flash.
To report hot flashes, participants were instructed to: 1) complete a portable electronic diary (Palm Z22; Palm, Inc., Sunnyvale, CA) (waking hours); and 2) press event mark buttons on the hot flash monitor (waking and sleeping hours) when experiencing a hot flash.
At the SWAN FLASHES assessment before hot flash monitoring, questionnaire-assessed frequency of hot flashes (1–5, 6–8, 9–13 d, or everyday in the prior 2 wk) was also measured. Responses were dichotomized (1–8 vs. ≥9 d) due to small cell sizes.
Reproductive hormones
Concentrations of FSH, E2, and SHBG were obtained from a morning fasting blood sample during annual SWAN visit 10. Attempts were made to take samples on menstrual cycle d 2–5, achieved for 4% of the sample. A random fasting sample was taken in 96% of sample, largely reflective of the lack of regular menstrual cycling among these primarily postmenopausal women. E2 assays were conducted in duplicate and FSH, dehydroepiandrosterone sulfate (DHEAS), and SHBG in singlicate. Assays were performed using an ACS-180 automated analyzer (Bayer Diagnostics, Tarrytown, NY). E2 was measured with a modified direct ACS-180 (E2-6) immunoassay, with inter- and intraassay coefficients of variation of 10.6 and 6.4%, respectively, and a lower limit of detection (LLD) of 6.6 pg/ml (31). FSH assays were performed with a two-site chemiluminometric immunoassay, with inter- and intraassay coefficients of variation of 11.4 and 3.8%, respectively, and LLD of 1.1 mIU/ml. The DHEAS and SHBG assays were developed on site using rabbit anti-DHEAS and anti-SHBG antibodies, with LLD of 1.52 μg/dl and 1.95 nm, respectively. FEI was calculated as 100 × E2 (pg/ml)/272.11 × SHBG (nm) (32).
Covariates
Demographics, menstrual history, and health behaviors were assessed by questionnaires during the SWAN FLASHES visit. Race/ethnicity was determined in response to the question “How would you describe your primary racial or ethnic group?” Menopausal status was obtained from reported bleeding patterns, categorized as perimenopausal (bleeding in previous 3 months with a decrease in cycle predictability in past year or more than 3 but less than 12 months of amenorrhea) or postmenopausal (≥12 months of amenorrhea). Time since last menstrual period was obtained from the prior 10 yr of SWAN annual interviews. Physical activity was assessed via a modified Kaiser Permanente Health Plan Activity Survey (33), depressive symptoms via the Center for Epidemiologic Studies Depression Survey (34), and anxiety via the Spielberger State-Trait Anxiety Inventory (35).
Statistical analysis
Variables were examined for distributions, outliers, and cell sizes. Depression, anxiety, physical activity, E2, FEI, and DHEAS were log-transformed, and SHBG was square root-transformed due to skew. Differences in variables by race/ethnicity were conducted via χ2 and t tests. Associations between anthropometrics, covariates, and hot flashes were evaluated using generalized estimating equations (GEE), with an autoregressive correlation matrix and hot flashes nested within monitoring sessions and women. Each 20-min interval was classified regarding the presence/absence of a hot flash. GEE have the advantage of modeling hot flashes in the longitudinal fashion they occurred, while accounting for variations in monitoring duration and clustering of hot flashes within women. Odds ratios (OR) greater than 1.0 indicate more hot flashes. Factors associated with hot flashes (P < 0.10) were included in multivariable models. Given high correlation between state anxiety, trait anxiety, and depression, the affective variable most strongly associated with hot flashes (state anxiety) was included. Hormones were considered in a separate step, with blood draw timing (within vs. outside of cycle d 2–5) also covaried in GEE models. In models with adiposity or hormones, the time difference between SWAN FLASHES and SWAN visits (months between visits) was also covaried. Five missing adiposity values were carried forward from prior SWAN years, resulting in a median (interquartile range) of 1.5 (1.2) yr between BIA and hot flash assessments. Interactions between age and race were evaluated with cross-product terms; where significant interactions were evident, results were stratified by race or age tertile (54–56.8, 56.9–59.2, and 59.3–64 yr). Interactions by menopausal stage were not estimated, given the few perimenopausal women, although interactions by time since last menstrual period were examined. Correlations between body size/composition and hormones were estimated using Spearman correlation coefficients, partialling blood draw timing, race, age, and time difference between measures. Associations between anthropometrics and recalled hot flashes were examined using linear regression. Analyses were performed with SAS (v. 9.2;, SAS Institute, Cary, NC). Tests were two-sided, α = 0.05.
Results
Participants were on average 58 yr old (range, 54–63), overweight, and postmenopausal (Table 1). African-American women were less likely to drink alcohol, had more physiological hot flashes, and had a somewhat higher adiposity and BMI than Caucasian women.
Table 1.
Sample characteristics for the total sample and by race
| Total sample | Caucasian | African-American | |
|---|---|---|---|
| n | 52 | 27 | 25 |
| Age (yr) | 58.3 (2.3) | 58.6 (2.5) | 58.1 (2.1) |
| Education, n (%) | |||
| High school | 10 (19.2) | 8 (29.6) | 2 (8.0) |
| Some college/vocational | 27 (51.9) | 12 (44.4) | 15 (60.0) |
| College or higher | 15 (28.9) | 7 (25.9) | 8 (32.0) |
| Menopausal status, n (%) | |||
| Perimenopausal | 5 (9.6) | 3 (11.1) | 2 (8.0) |
| Postmenopausal | 47 (90.4) | 24 (88.9) | 23 (92.0) |
| Time since last menstrual period (yr) | 5.8 (3.4) | 6.1 (3.7) | 5.5 (3.2) |
| Smoking (current), n (%) | 6 (11.5) | 2 (7.4) | 4 (16.0) |
| Alcohol use (any current), n (%)c | 31 (59.6) | 21 (77.8) | 10 (40.0) |
| Physical activity score | 5.6 (1.5) | 5.6 (1.0) | 5.6 (1.3) |
| Depressive symptoms (CESD) | 7.4 (6.5) | 6.5 (6.1) | 8.4 (6.9) |
| State anxiety | 31.6 (9.4) | 31.7 (10.9) | 31.5 (7.6) |
| Trait anxiety | 33.6 (9.0) | 32.4 (8.4) | 34.8 (9.6) |
| E2 (pg/ml)a | 24.4 (21.9) | 24.1 (26.4) | 24.8 (15.6) |
| FSH (mIU/ml)a | 115.1 (63.9) | 112.2 (58.4) | 118.6 (71.2) |
| SHBG (nm)a | 45.9 (28.6) | 49.9 (33.2) | 41.2 (21.8) |
| DHEAS (μg/dl)a | 118.5 (76.2) | 122.8 (74.7) | 113.4 (79.3) |
| FEIa | 0.27 (0.29) | 0.25 (0.29) | 0.30 (0.29) |
| BMI (kg/m2)b | 29.9 (5.0) | 28.7 (4.7) | 31.1 (5.3) |
| Obesity status, n (%) | |||
| Normal | 10 (19.2) | 7 (25.9) | 3 (12.0) |
| Overweight | 18 (34.6) | 9 (33.3) | 9 (36.0) |
| Obese | 24 (46.2) | 11 (40.7) | 13 (52.0) |
| Total percentage body fatb | 40.3 (5.7) | 38.8 (5.5) | 41.8 (5.6) |
| Waist circumference | 90.8 (13.6) | 88.2 (14.7) | 93.6 (11.9) |
| No. of physiological hot flashes/24 hc | 18 (10) | 17 (7) | 21 (12) |
| Reported hot flashes/24 h | 5 (4) | 5 (3) | 6 (4) |
Data are expressed as mean (sd), unless described otherwise. CESD, Center for Epidemiological Studies Depression Survey.
n = 48.
Differences by race:
P < 0.10;
P < 0.05.
Race/ethnicity, state anxiety, depressive symptoms, DHEAS, FEI, and to a lesser extent FSH showed univariate associations with hot flashes. African-American women had more physiological hot flashes than Caucasian women [OR (95% confidence interval or CI) = 1.33 (1.03–1.71); P = 0.02]. Higher anxiety and depressive symptoms were associated with more physiological hot flashes [state anxietylog, OR (95% CI) = 2.02 (1.31–3.11), P = 0.002; trait anxietylog, OR (95% CI) = 1.78 (1.04–3.03), P = 0.04; and depressionlog, OR (95% CI) = 1.23 (1.05–1.45), P = 0.01] and self-reported hot flashes [state anxietylog, OR (95% CI) = 2.05 (1.33–3.13), P = 0.001; trait anxietylog, OR (95% CI) = 1.63 (0.99–2.72), P = 0.06; and depressionlog, OR (95% CI) = 1.32 (1.15–1.52), P < 0.0001]. Higher DHEASlog [OR (95% CI) = 0.63 (0.48–0.83); P = 0.004] and FEIlog [OR (95% CI) = 0.84 (0.73–0.96); P = 0.01] were associated with fewer physiological hot flashes. Higher FSH [OR (95% CI) = 1.001 (1.00–1.003); P = 0.06] was associated with marginally more physiological hot flashes.
In univariate models, higher BMI was marginally associated with fewer physiological hot flashes [OR (95% CI) = 0.97 (0.94–1.002); P = 0.07], and higher waist circumference was significantly associated with fewer physiological hot flashes [OR (95% CI) = 0.99 (0.98–0.99); P = 0.01]. In multivariable models adjusted for age, race, and anxiety, higher BMI and waist circumference were significantly associated with fewer physiological hot flashes (Table 2). Other factors associated with physiological hot flashes in multivariable models were African-American race/ethnicity [e.g. OR (95% CI) = 1.43 (1.14–1.78); P < 0.01, in models with BMI] and anxiety [e.g. OR (95% CI) = 2.09 (1.36–3.23); P < 0.01, in models with BMI]. The only factor associated with reported hot flashes in multivariable models was anxiety [e.g. OR (95% CI) = 2.20 (1.36–3.56); P < 0.01, in models with BMI]. Body fat percentage was not associated with physiological or reported hot flashes.
Table 2.
Association between body size/composition and hot flashes
| Physiological hot flashes | Self-report hot flashes | |
|---|---|---|
| Total percentage of body fat | 1.00 (0.97–1.02) | 1.002 (0.98–1.03) |
| BMI | 0.97 (0.94–0.99)a | 0.99 (0.97–1.02) |
| Waist circumference | 0.98 (0.97–0.99)b | 0.99 (0.98–1.002) |
Data are expressed as OR (95% CI). Covariates are: age, race/ethnicity, anxiety; time difference between body fat and hot flashes measurements are included in body fat models. Note that each anthropometric variable is considered in a separate model. Sample sizes: n = 52 for all models except body fat models (n = 51).
P < 0.05;
P < 0.01.
Interactions by age and race/ethnicity
Interactions with age were noted for body fat (P = 0.02), BMI (P < 0.0001), and waist circumference (P = 0.003) in relation to physiological hot flashes in multivariable models. Higher body fat, BMI, and waist circumference were associated with fewer physiological hot flashes only among the older women (Table 3). We also observed interactions by race/ethnicity (body fat, P = 0.004; BMI, P = 0.06), with the inverse association between adiposity/BMI and hot flashes restricted to Caucasian women [Caucasian, body fat, OR (95% CI) = 0.97 (0.95–0.99), P = 0.009; BMI, OR (95% CI) = 0.94 (0.91–0.97), P < 0.0001; African-American women, body fat, OR (95% CI) = 1.02 (1.00–1.04), P = 0.10; BMI, OR (95% CI) = 1.003 (0.96–1.05), P = 0.87]. There were no three-way interactions between age, race, and body size/adiposity (data not shown), nor were there interactions by age or race/ethnicity for reported hot flashes. There were no significant interactions between time since the last menstrual period and body size/adiposity in relation to hot flashes.
Table 3.
Associations between anthropometric variables and physiological hot flashes by age
| Low (<56.9 yr) | Medium (56.9–59.2 yr) | High (≥ 59.3 yr) | |
|---|---|---|---|
| Body fat | 1.00 (0.97–1.03) | 1.01 (0.98–1.04) | 0.96 (0.93–0.99)b |
| BMI | 0.98 (0.94–1.03) | 0.98 (0.92–1.04) | 0.89 (0.84–0.93)c |
| Waist circumference | 0.99 (0.97–1.00) | 0.98 (0.96–1.00)a | 0.96 (0.95–0.98)c |
Data are expressed as OR (95% CI). Covariates are: age, race/ethnicity, and anxiety; time difference is included in body fat models. Note that each anthropometric variable is considered in a separate model. Sample sizes: low, n = 17; medium, n = 17 (for body fat models, n = 16); high, n = 18.
P < 0.10;
P < 0.01;
P < 0.0001.
Influence of reproductive hormones
As would be expected, greater body size/adiposity was associated with higher E2, lower SHBG, and higher FEI, with associations most pronounced with increasing tertile of age (Table 4). Moreover, in examining whether these hormones accounted for age differences in relations between body size/fat and hot flashes, SHBG, FEI, and E2 each alone reduced age differences in relations between body fat (but not other anthropometric variables) and hot flashes to nonsignificance (P > 0.10). Considering race/ethnicity, greater body size/adiposity was related to lower SHBG (e.g. BMI and SHBG, Caucasian ρ = −0.66, P < 0.001; African-American ρ = 0.13, P = not significant) and higher FEI (e.g. BMI and FEI, Caucasian ρ = 0.49, P < 0.05; African-American ρ= 0.16, P = not significant) primarily among Caucasian women. Furthermore, SHBG reduced interactions between race and body fat/BMI in relation to hot flashes to nonsignificance (P > 0.10). Thus, SHBG, FEI, and E2 may play a partial role in age-related variations, and SHBG in race-related variations in body composition-hot flash relations. Neither DHEAS nor FSH was related to body/size composition, nor did they modify age and race differences in relations between body size/fat and hot flashes.
Table 4.
Spearman correlation coefficients between hormones and anthropometric factors, total sample, and by age
| E2 | SHBG | FEI | |
|---|---|---|---|
| Total percentage of body fata | 0.43g | −0.25 | 0.43g |
| Ages 54–56.8 yr | −0.02 | −0.27 | 0.33 |
| Ages 56.9–59.2 yr | 0.75g | −0.44 | 0.75g |
| Ages 59.3–64 yr | 0.47e | 0.33 | 0.45e |
| BMIc,d | 0.31f | −0.33f | 0.38f |
| Ages 54–56.8 yr | −0.22 | −0.16 | 0.03 |
| Ages 56.9–59.2 yr | 0.13 | −0.16 | 0.12 |
| Ages 59.3–64 yr | 0.56g | −0.72g | 0.75g |
| Waist circumferenceb | 0.16 | −0.45g | 0.35f |
| Ages 54–56.8 yr | −0.30 | −0.44 | 0.10 |
| Ages 56.9–59.2 yr | −0.16 | −0.24 | −0.03 |
| Ages 59.3–64 yr | 0.58g | −0.60f | 0.64g |
Data are adjusted for cycle day of blood draw, age, time difference between hormone and anthropometric measurements, and race/ethnicity for full sample models. Sample sizes: full sample, n = 48; ages 54–56.8, n = 15; ages 56.9–59.2, n = 15; ages 59.3–64, n = 18.
Differences by race:
P < 0.10 for E2;
P < 0.10,
P < 0.05 for SHBG;
P = 0.05 for FEI;
P < 0.10;
P < 0.05;
P < 0.01.
Recalled hot flashes
For comparability with other studies (1), we examined associations between anthropometric variables and questionnaire-assessed hot flashes (recalled over prior 2 wk). These findings confirmed primary models, with higher body size/adiposity associated with lower hot flash reporting [percentage body fat, OR (95% CI) = 0.88 (0.76–1.01), P = 0.07; BMI, OR (95% CI) = 0.85 (0.75–0.98), P = 0.02; and waist circumference, OR (95% CI) = 0.94 (0.89–0.99), P = 0.02; multivariable models].
Discussion
In this study, a higher percentage of body fat, BMI, and waist circumference were associated with fewer physiologically measured hot flashes among the older women in the sample. Associations between body fat and physiological hot flashes were also limited to Caucasian women. Finally, there was a suggestion that E2, SHBG, and/or FEI may play a partial role in age and racial/ethnic variations in relations between body composition and hot flashes.
The present findings are at apparent odds with epidemiological research showing positive associations between BMI/adiposity and hot flash reporting (1, 7). In fact, they are more consistent with the “thin hypothesis,” in which higher adiposity relates to reduced hot flashes, albeit modified. The inverse associations between body size/fat were only observed among the older women. There has been some prior work suggesting that the direction of the relation between adiposity/BMI and hot flashes may vary by menopausal stage or age, with positive associations between BMI/adiposity most apparent among younger women, and a null or inverse relation among older/postmenopausal women (12–16). Thus, the direction of relations between adiposity and hot flashes may shift with aging.
Emerging work demonstrates that the reproductive endocrine impact of BMI/adiposity varies by age/menopausal stage, with higher BMI related to higher E2 among older and postmenopausal women (6, 36), and lower E2 among younger women (6). Extragonadal sources such as body fat may serve as the primary site of estrogen synthesis among older women, but not in younger women who are having at least intermittent cycles. In fact, obese perimenopausal women have been shown to produce less gonadotropins, progesterone, and estrone metabolites across the menstrual cycle compared with thinner women (37), which may predispose younger high-BMI women to hot flashes. Prior work suggests that body fat may have both thermoregulatory and endocrine properties in relation to hot flashes (7). Thus, thermoregulatory vs. endocrine effects of adiposity may predominate among younger perimenopausal vs. older postmenopausal women, respectively.
We examined the role of key hormonal factors in these relations. Greater body size/adiposity was related to higher E2, lower SHBG, and greater FEI, particularly among older women. Moreover, women with higher FEI had fewer hot flashes, and FEI, E2, and SHBG accounted for age differences in relations between body fat and hot flashes. Therefore, free E2 likely played a role but did not fully account for observed age variations in body size-hot flash relations. These findings must be interpreted in light of limitations, particularly the single assessment that may not have fully captured the hormonal dynamics of these women, the lack of assessment of estrone, and an assessment schedule not precisely concurrent with hot flash measurement. However, these findings do suggest the importance of looking beyond E2 to understand age-related variations in relations between body size/fat and hot flashes. Finally, although DHEAS and FSH were related to hot flashes, they played little role in observed relations between anthropometrics and hot flashes.
Race also played an important role in observed associations. Given anthropometric differences by race/ethnicity, there has been discussion as to the degree to which adiposity and race are independently associated with hot flashes (38). African-American women here had more physiological hot flashes controlling for anthropometric factors. Thus, the higher hot flash burden among African-American women was not attributable to body size/composition. Second, inverse associations between body fat and hot flashes were apparent among Caucasian women only, consistent with prior work (17). The reason for this difference is not clear, although these racial/ethnic differences were reduced controlling for SHBG. Further work investigating racial/ethnic differences in relations between body size/adiposity and hot flashes, including a role for SHBG, is warranted.
Anxiety was related to both reported and physiological hot flashes. In fact, anxiety was the only factor related to hot flash reporting. The role of anxiety in hot flash reporting has been noted before (1, 11, 20, 29). However, in contrast to work showing anxiety related solely to reported hot flashes lacking physiological evidence (11, 29), anxious women here had more reported and physiological hot flashes. The directionality of these relations cannot be inferred, but bidirectional relations are plausible, with more hot flashes increasing anxiety and greater anxiety increasing symptom reporting.
Several differences between existing literature and this study deserve mention. First, the present study is the first to examine relations between body size and composition and hot flashes using physiological measures of hot flashes. Notably, differences between these and prior epidemiological findings cannot be explained by this methodology because similar patterns of results were observed with questionnaire hot flash measures. Furthermore, our study investigates dose-response relations between anthropometrics and hot flashes among women with hot flashes, as opposed to risk factors for any/no hot flashes. Risk factors for the presence vs. absence rather than hot flash frequency may vary, and future work should examine these relations among women with and without hot flashes.
Endocrine and thermoregulatory models of hot flashes emphasize adiposity. However, relations between BIA and hot flashes, including age differences in relations with hormones, were somewhat less robust than for other anthropometric variables. Notably, BIA was the only anthropometric measure not measured concurrently with hot flashes, thereby containing more error and possibly reducing detection of significant associations. However, future work should consider mechanisms beyond adiposity to understand relations between obesity and hot flashes.
Several limitations deserve mention. First, the sample was small. Power to detect associations, particularly interactions, may have been limited. Further investigations should be undertaken with larger samples. Second, most of the participants were postmenopausal. A different pattern of associations may be observed earlier in the transition. BIA and hormonal measures were taken at the annual SWAN visit rather than at the SWAN FLASHES visit, likely increasing error in these measures. Finally, most of the sample was overweight/obese, limiting the generalizability of findings to lean women.
This study had several strengths. It is the first to examine relations between body size/composition and hot flashes using physiological and diary-reported hot flashes, allowing a careful analysis of hot flashes and reducing biases of questionnaire measures (10, 39). This study incorporated multiple anthropometric measures, allowing comparison across measures. BIA, although less precise than resource-intensive measures such as dual-energy x-ray absorptiometry, represents a more precise measure of adiposity than the commonly used BMI. Finally, this study included approximately equal numbers of African-American and Caucasian women, allowing comparisons between groups, and with a well-characterized sample of women followed throughout the menopausal transition.
This study showed that higher adiposity, BMI, and waist circumference were associated with fewer physiologically assessed hot flashes among older postmenopausal women with hot flashes. Moreover, associations were most pronounced among Caucasian women. This study adds to the growing body of literature examining relations between body composition and hot flashes. It underscores the importance of considering how age and race may modify the relations between obesity and hot flashes.
Acknowledgments
We thank the study staff at each site and all the women who participated in SWAN.
Clinical Center: University of Pittsburgh, Pittsburgh, Pennsylvania; principal investigator (PI), Karen Matthews. National Institutes of Health (NIH) Program Office: National Institute on Aging (NIA), Bethesda, Maryland; Project Officer, Sherry Sherman, 1994 to the present; Marcia Ory, 1994–2001; National Institute of Nursing Research (NINR), Bethesda, Maryland. Central Laboratory: University of Michigan, Ann Arbor, Michigan; Daniel McConnell (Central Ligand Assay Satellite Services). Coordinating Center: University of Pittsburgh, Pittsburgh, Pennsylvania; PI, Kim Sutton-Tyrrell, 2001 to the present; New England Research Institutes, Watertown, Massachusetts; PI, Sonja McKinlay, 1995–2001. Steering Committee: Current Chair, Susan Johnson; Former Chair, Chris Gallagher.
The Study of Women's Health Across the Nation (SWAN) has grant support from the NIH, Department of Health and Human Services, through the NIA, the NINR, and the NIH Office of Research on Women's Health (ORWH) (Grants AG012546 and AG012535). This work was additionally supported by the NIH through NIA Grant AG029216 (to R.C.T.).
R.C.T. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIA, NINR, ORWH, or NIH.
Disclosure Summary: N.S. has stock options in MenogeniX. No other authors have potential conflicts to disclose.
Footnotes
- BIA
- Bioimpedance analysis
- BMI
- body mass index
- CI
- confidence interval
- DHEAS
- dehydroepiandrosterone sulfate
- E2
- estradiol
- FEI
- free E2 index
- GEE
- generalized estimating equation
- LLD
- limit of detection
- OR
- odds ratio.
References
- 1. Gold EB, Colvin A, Avis N, Bromberger J, Greendale GA, Powell L, Sternfeld B, Matthews K. 2006. Longitudinal analysis of vasomotor symptoms and race/ethnicity across the menopausal transition: Study of Women's Health Across the Nation (SWAN). Am J Public Health 96:1226–1235 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Williams RE, Kalilani L, DiBenedetti DB, Zhou X, Granger AL, Fehnel SE, Levine KB, Jordan J, Clark RV. 2008. Frequency and severity of vasomotor symptoms among peri- and postmenopausal women in the United States. Climacteric 11:32–43 [DOI] [PubMed] [Google Scholar]
- 3. Avis NE, Colvin A, Bromberger JT, Hess R, Matthews KA, Ory M, Schocken M. 2009. Change in health-related quality of life over the menopausal transition in a multiethnic cohort of middle-aged women: Study of Women's Health Across the Nation. Menopause 16:860–869 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Campagnoli C, Morra G, Belforte P, Belforte L, Prelato Tousijn L. 1981. Climacteric symptoms according to body weight in women of different socio-economic groups. Maturitas 3:279–287 [DOI] [PubMed] [Google Scholar]
- 5. Erlik Y, Meldrum DR, Judd HL. 1982. Estrogen levels in postmenopausal women with hot flashes. Obstet Gynecol 59:403–407 [PubMed] [Google Scholar]
- 6. Freeman EW, Sammel MD, Lin H, Gracia CR. 2010. Obesity and reproductive hormone levels in the transition to menopause. Menopause 17:718–726 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Thurston RC, Sowers MR, Chang Y, Sternfeld B, Gold EB, Johnston JM, Matthews KA. 2008. Adiposity and reporting of vasomotor symptoms among midlife women: the Study of Women's Health Across the Nation. Am J Epidemiol 167:78–85 [DOI] [PubMed] [Google Scholar]
- 8. Thurston RC, Sowers MR, Sutton-Tyrrell K, Everson-Rose SA, Lewis TT, Edmundowicz D, Matthews KA. 2008. Abdominal adiposity and hot flashes among midlife women. Menopause 15:429–434 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Freedman RR. 2001. Physiology of hot flashes. Am J Hum Biol 13:453–464 [DOI] [PubMed] [Google Scholar]
- 10. Erskine A, Morley S, Pearce S. 1990. Memory for pain: a review. Pain 41:255–265 [DOI] [PubMed] [Google Scholar]
- 11. Thurston RC, Blumenthal JA, Babyak MA, Sherwood A. 2005. Emotional antecedents of hot flashes during daily life. Psychosom Med 67:137–146 [DOI] [PubMed] [Google Scholar]
- 12. Whiteman MK, Staropoli CA, Langenberg PW, McCarter RJ, Kjerulff KH, Flaws JA. 2003. Smoking, body mass, and hot flashes in midlife women. Obstet Gynecol 101:264–272 [DOI] [PubMed] [Google Scholar]
- 13. den Tonkelaar I, Seidell JC, van Noord PA. 1996. Obesity and fat distribution in relation to hot flashes in Dutch women from the DOM-project. Maturitas 23:301–305 [DOI] [PubMed] [Google Scholar]
- 14. Gold EB, Sternfeld B, Kelsey JL, Brown C, Mouton C, Reame N, Salamone L, Stellato R. 2000. Relation of demographic and lifestyle factors to symptoms in a multi-racial/ethnic population of women 40–55 years of age. Am J Epidemiol 152:463–473 [DOI] [PubMed] [Google Scholar]
- 15. Hyde Riley E, Inui TS, Kleinman K, Connelly MT. 2004. Differential association of modifiable health behaviors with hot flashes in perimenopausal and postmenopausal women. J Gen Intern Med 19:740–746 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Sabia S, Fournier A, Mesrine S, Boutron-Ruault MC, Clavel-Chapelon F. 2008. Risk factors for onset of menopausal symptoms: results from a large cohort study. Maturitas 60:108–121 [DOI] [PubMed] [Google Scholar]
- 17. Thurston RC, Sowers MR, Sternfeld B, Gold EB, Bromberger J, Chang Y, Joffe H, Crandall CJ, Waetjen LE, Matthews KA. 2009. Gains in body fat and vasomotor symptom reporting over the menopausal transition: the Study of Women's Health Across the Nation. Am J Epidemiol 170:766–774 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Randolph JF, Jr, Sowers M, Bondarenko I, Gold EB, Greendale GA, Bromberger JT, Brockwell SE, Matthews KA. 2005. The relationship of longitudinal change in reproductive hormones and vasomotor symptoms during the menopausal transition. J Clin Endocrinol Metab 90:6106–6112 [DOI] [PubMed] [Google Scholar]
- 19. Randolph JF, Jr, Sowers M, Gold EB, Mohr BA, Luborsky J, Santoro N, McConnell DS, Finkelstein JS, Korenman SG, Matthews KA, Sternfeld B, Lasley BL. 2003. Reproductive hormones in the early menopausal transition: relationship to ethnicity, body size, and menopausal status. J Clin Endocrinol Metab 88:1516–1522 [DOI] [PubMed] [Google Scholar]
- 20. Freeman EW, Sammel MD, Lin H, Gracia CR, Kapoor S, Ferdousi T. 2005. The role of anxiety and hormonal changes in menopausal hot flashes. Menopause 12:258–266 [DOI] [PubMed] [Google Scholar]
- 21. Sowers M, Crawford S, Sternfeld B, Morganstein D, Gold EB, Greendale GA, Evans D, Neer R, Matthews K, Sherman S, Lo A, Weiss G, Kelsey J. 2000. SWAN: a multicenter, multiethnic, community-based cohort study of women and the menopausal transition. In: Lobo RA, Kelsey J, Marcus R, Lobo AR. eds. Menopause: biology and pathology. New York: Academic Press; 175–188 [Google Scholar]
- 22. Flegal KM, Carroll MD, Ogden CL, Curtin LR. 2010. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 303:235–241 [DOI] [PubMed] [Google Scholar]
- 23. 1996. Bioelectrical impedance analysis in body composition measurement: National Institutes of Health Technology Assessment Conference Statement. Am J Clin Nutr 64:524S–532S [DOI] [PubMed] [Google Scholar]
- 24. Boulier A, Fricker J, Thomasset AL, Apfelbaum M. 1990. Fat-free mass estimation by the two-electrode impedance method. Am J Clin Nutr 52:581–585 [DOI] [PubMed] [Google Scholar]
- 25. Chumlea WC, Guo SS, Kuczmarski RJ, Flegal KM, Johnson CL, Heymsfield SB, Lukaski HC, Friedl K, Hubbard VS. 2002. Body composition estimates from NHANES III bioelectrical impedance data. Int J Obes Relat Metab Disord 26:1596–1609 [DOI] [PubMed] [Google Scholar]
- 26. Dormire SL, Carpenter JS. 2002. An alternative to Unibase/glycol as an effective nonhydrating electrolyte medium for the measurement of electrodermal activity. Psychophysiology 39:423–426 [DOI] [PubMed] [Google Scholar]
- 27. Freedman RR. 1989. Laboratory and ambulatory monitoring of menopausal hot flashes. Psychophysiology 26:573–579 [DOI] [PubMed] [Google Scholar]
- 28. Carpenter JS, Andrykowski MA, Freedman RR, Munn R. 1999. Feasibility and psychometrics of an ambulatory hot flash monitoring device. Menopause 6:209–215 [DOI] [PubMed] [Google Scholar]
- 29. Thurston RC, Matthews KA, Hernandez J, De La Torre F. 2009. Improving the performance of physiologic hot flash measures with support vector machines. Psychophysiology 46:285–292 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Thurston RC, Hernandez J, Del Rio JM, De La Torre F. 2011. Support vector machines to improve physiologic hot flash measures: application to the ambulatory setting. Psychophysiology 48:1015–1021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. England BG, Parsons GH, Possley RM, McConnell DS, Midgley AR. 2002. Ultrasensitive semiautomated chemiluminescent immunoassay for estradiol. Clin Chem 48:1584–1586 [PubMed] [Google Scholar]
- 32. van den Beld AW, de Jong FH, Grobbee DE, Pols HA, Lamberts SW. 2000. Measures of bioavailable serum testosterone and estradiol and their relationships with muscle strength, bone density, and body composition in elderly men. J Clin Endocrinol Metab 85:3276–3282 [DOI] [PubMed] [Google Scholar]
- 33. Ainsworth BE, Sternfeld B, Richardson MT, Jackson K. 2000. Evaluation of the Kaiser physical activity survey in women. Med Sci Sports Exerc 32:1327–1338 [DOI] [PubMed] [Google Scholar]
- 34. Radloff LS. 1977. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1:385–401 [Google Scholar]
- 35. Spielberger CD. 1983. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press [Google Scholar]
- 36. Baglietto L, English DR, Hopper JL, MacInnis RJ, Morris HA, Tilley WD, Krishnan K, Giles GG. 2009. Circulating steroid hormone concentrations in postmenopausal women in relation to body size and composition. Breast Cancer Res Treat 115:171–179 [DOI] [PubMed] [Google Scholar]
- 37. Santoro N, Lasley B, McConnell D, Allsworth J, Crawford S, Gold EB, Finkelstein JS, Greendale GA, Kelsey J, Korenman S, Luborsky JL, Matthews K, Midgley R, Powell L, Sabatine J, Schocken M, Sowers MF, Weiss G. 2004. Body size and ethnicity are associated with menstrual cycle alterations in women in the early menopausal transition: The Study of Women's Health Across the Nation (SWAN) Daily Hormone Study. J Clin Endocrinol Metab 89:2622–2631 [DOI] [PubMed] [Google Scholar]
- 38. Crawford SL. 2007. The roles of biologic and nonbiologic factors in cultural differences in vasomotor symptoms measured by surveys. Menopause 14:725–733 [DOI] [PubMed] [Google Scholar]
- 39. Bradburn NM, Rips LJ, Shevell SK. 1987. Answering autobiographical questions: the impact of memory and inference on surveys. Science 236:157–161 [DOI] [PubMed] [Google Scholar]
