OBJECTIVE: Preliminary studies suggest that the menopause transition is associated with deleterious changes in body composition and abdominal fat distribution. Limitations of the methodology used in these studies, however, render their conclusions controversial. Thus, the present study used radiologic imaging techniques to examine the effect of menopausal status on body composition and abdominal fat distribution.
SUBJECTS: Fifty-three healthy, middle-aged, premenopausal women (mean±SD; 47±3 y) and 28 early-postmenopausal women (51±4 y).
MEASUREMENTS: Total and regional body composition by dual energy X-ray absorptiometry and abdominal fat distribution by computed tomography.
RESULTS: No differences in total body fat-free mass or appendicular skeletal muscle mass were noted between groups. In contrast, total body fat mass was 28% higher (23±7 vs 18±7 kg) and percentage fat 17% higher (35±6 vs 30±9%; both P<0.01) in postmenopausal women compared with premenopausal women. Postmenopausal women had a 49% greater intra-abdominal (88±32 vs 59±32 cm2; P<0.01) and a 22% greater abdominal subcutaneous fat area (277±93 vs 227±108 cm2; P<0.05) compared to premenopausal women. The menopause-related difference in intra-abdominal fat persisted (P<0.05) after statistical adjustment for age and total body fat mass, whereas no difference in abdominal subcutaneous fat was noted. A similar pattern of differences in total and abdominal adiposity was noted in sub-samples of pre- and postmenopausal women matched for age or fat mass.
CONCLUSION: Our data suggest that early-postmenopausal status is associated with a preferential increase in intra-abdominal fat that is independent of age and total body fat mass.
Preliminary studies suggest that the menopause transition is associated with deleterious changes in body composition and body fat distribution. Work from our laboratory1 showed a greater increase in fat mass and the waist-to-hip ratio and a greater loss of fat-free mass in middle-aged women who became postmenopausal compared to women who remained premenopausal during a 6 y follow-up. Björkelund and co-workers2 found a similar menopause-related increase in the waist-to-hip ratio in women who became postmenopausal compared to women who remained premenopausal. Collectively, these longitudinal studies suggest that the menopause transition is associated with an accelerated increase in total and central adiposity and a reduction in fat-free mass.
Results from these studies, however, should be interpreted with caution. First, because hydrodensitometry was used to measure body composition in our longitudinal study,1 menopause-related changes in body composition may be partially explained by changes in bone mineral density. The rapid loss of bone mineral density during the early-postmenopausal period3 may have reduced the density of fat-free mass and caused percentage body fat to be overestimated in those women who became postmenopausal.4 Overestimation of percentage fat may partially account for the menopause-related increase in adiposity and decrease in fat-free mass observed in our study.1 Second, the waist-to-hip ratio may not be an appropriate index to assess menopause-related changes in abdominal fat distribution.5,6 In fact, Björkelund and co-workers2 found that the menopause-related increase in the waist-to-hip ratio was due to a decrease in hip circumference with no change in waist circumference, suggesting that the menopause transition was not associated with changes in abdominal adiposity. In light of these methodological limitations, controversy remains regarding the effect of the menopause transition on body composition and body fat distribution.
To examine the effect of menopausal status on body composition and abdominal fat distribution, we measured body composition by dual energy X-ray absorptiometry and body fat distribution by computed tomography in healthy, middle-aged premenopausal and early-postmenopausal women. The use of dual energy X-ray absorptiometry and computed tomography represents significant methodological advances over prior longitudinal studies that used hydrodensitometry and anthropometric measurements to estimate body composition and fat distribution, respectively. Moreover, studying non-obese, middle-aged premenopausal and early-postmenopausal women minimizes the confounding effect of age and adiposity on body composition and abdominal fat distribution.
Volunteers in the present study were recruited for two on-going studies in our laboratories. Participants were recruited from Burlington, VT and surrounding areas through advertisements in local newspapers. Premenopausal volunteers were recruited to participate in the Vermont Longitudinal Study of the Menopause, a 5 y study examining changes in energy expenditure, body composition, abdominal fat distribution and metabolic function in women as they traverse the menopause. Data from the first-year evaluation of these volunteers is presented. Data from this cohort has been published previously as part of a study examining correlates of energy expenditure and substrate oxidation in premenopausal women.7 Postmenopausal women were recruited to participate in a study examining the effect of hormone replacement therapy on glucose homeostasis and abdominal fat distribution. Data from this cohort has been published previously as part of a study examining correlates on insulin sensitivity (Sites et al, in press). Baseline, pre-treatment data from this study is presented.
The inclusion criteria for premenopausal women were: (1) between 40 and 52 y of age; (2) premenopausal, as defined by the occurrence of two menses in the 3 months preceding testing, no increase in cycle irregularity in the 12 months preceding testing and a follicle-stimulating hormone level less than 30 IU/L; (3) non-smoking; (4) normal electrocardiogram at rest and during an exercise test; (5) weight stability (±2 kg) during the 6 months prior to testing; and (6) body mass index ≤30 kg/m2. Premenopausal women were excluded if they: (1) were or planned on becoming pregnant; (2) had a history or current diagnosis of diabetes, heart disease, hypertension or other chronic disease; (3) were taking hormone replacement therapy, birth control, chronic steroid therapy, neuroleptics or other medication that could affect energy expenditure or metabolic function; (4) had a history of alcohol or drug abuse; or (5) were glucose intolerant, defined as a fasting glucose level of 112 mg/dl or higher or a 2 h glucose level of greater than 140 mg/dl following a 75 g oral glucose load.
The inclusion criteria for postmenopausal women were: (1) postmenopausal, as defined by absence of menses for at least 6 months and a follicle-stimulating hormone level greater than 30 IU/L; (2) not more than 5 y postmenopausal; and (3) body mass index ≤30 kg/m2. Time since menopause (months) was defined as the time since the cessation of menses. The exclusion criteria for postmenopausal women were identical to premenopausal women except that no oral glucose tolerance test was performed. Thus, postmenopausal women with a fasting glucose level greater than 112 mg/dl were excluded. The nature, purpose and possible risks of each study were explained to each subject before she gave written consent to participate. The experimental protocols were approved by the Committee on Human Research at the University of Vermont.
Each prospective volunteer underwent an outpatient screening visit at which time medical history, physical examination, biochemical laboratory tests, treadmill test (premenopausal women only) and an oral glucose tolerance test (premenopausal women only) were performed. Volunteers who met the eligibility criteria following screening and consented to participate were admitted to the General Clinical Research Center (GCRC) for an overnight visit approximately 2 months following their screening visit. In premenopausal women, the overnight visit occurred during the follicular phase of the menstrual cycle in 42 patients and during the luteal phase in 11 patients. For 3 days prior to admission, all subjects consumed a standardized, weight-maintenance diet provided by the Metabolic Kitchen of the GCRC (60% carbohydrate, 25% fat, 15% protein). Computed tomography scans were performed the evening of admission and dual energy X-ray absorptiometry the next morning.
Body mass was measured on a metabolic scale (Scale-Tronix Inc., Wheaton, IL) with the volunteer clothed in a hospital gown. Fat mass, fat-free mass and bone mineral mass were measured by dual energy X-ray absorptiometry (DEXA) using a Lunar DPX-L densi-tometer (Lunar Co, Madison, WI). All scans were analyzed using the Lunar Version 1.3y DPX-L extended-analysis program for body composition. Appendicular skeletal muscle mass was obtained from regional fat-free mass measurements according to the model of Heymsfield et al8 as described previously.9 Briefly, specific regions of the body were defined using an on-screen image of the patient and specified anatomical cut-points.8 In this model, the fat-free mass of each limb represents skeletal muscle mass given that bone marrow, skin and associated subcutaneous tissues contribute a negligible amount to the total mass of the extremity. In our laboratory, the coefficient of variation for repeat determination in seven older women was 1% for total body fat mass, 2% for total body fat-free mass, 0.8% for fat-free tissue mass of the arms and 1% for fat-free tissue mass of the legs.
Intra-abdominal and abdominal subcutaneous adipose tissue areas were measured by computed tomography with a GE High Speed Advantage computed tomography scanner (General Electric Medical Systems, Milwaukee, WI). Subjects were examined in the supine position with both arms stretched above the head. The scan was performed between the L4–L5 vertebrae using a scout image of the body to establish the precise scanning position. Intra-abdominal adipose tissue area was quantified by delineating the intra-abdominal cavity at the innermost aspect of the abdominal and oblique muscle walls surrounding the cavity and the posterior aspect of the vertebral body. Adipose tissue was highlighted and computed using an attenuation range from −190 to −30 Hounsfield units. The subcutaneous adipose tissue area was quantified by highlighting adipose tissue located between the skin and the outermost aspect of the abdominal muscle wall. The coefficient of variation for repeated analysis of 10 subjects’ scans (i.e. reproducibility of the analysis of each scan) was less than 1%.
Differences between variables were determined by unpaired Student's t-tests. Menopausal status was assigned a dummy variable: 1 for premenopausal and 2 for postmenopausal. Analysis of covariance was used to examine differences in body composition and abdominal fat distribution after controlling for selected covariates. Fat mass was statistically adjusted for differences in age. Abdominal subcutaneous and intra-abdominal fat areas were adjusted for differences in age and total body fat mass to examine differences in abdominal fat distribution independent of menopause-related differences in adiposity. Age was closely correlated to menopausal status in our population (n=81; r=0.54; P, 0.01). Because of this colinearity, statistical adjustment for age would, in effect, partially remove the effect of menopausal status. Thus, age was adjusted for menopausal status using regression analysis, as described,10 prior to its use as a covariate. In addition to statistical adjustment, body composition and abdominal fat distribution were compared between sub-samples of pre- and postmenopausal women matched for either age (n=14; premenopausal: 49±2 vs postmenopausal: 50±3 y) or fat mass (n=40; premenopausal: 22±6 vs postmenopausal: 22±7 kg). In both cases, each pair of pre- and postmenopausal women were matched to within ±5% of the matching variable. Relationships between variables were determined by Pearson product–moment correlation coefficients. All data are expressed as mean±SD, unless otherwise specified.
Physical characteristics of pre- and postmenopausal women are shown in Table 1. Postmenopausal women were older (P<0.01), weighed more (P<0.05) and had a higher body mass index (P<0.01) than premenopausal women. Time since menopause averaged 22±15 months, indicating that postmenopausal women were studied early in the postmenopausal period. No differences were found in body composition or body fat distribution between premenopausal women studied in the luteal and follicular phase (data not shown). Moreover, in a sub-set of premenopausal women (n=43), no differences were found in total body water (isotope dilution) or the hydration of fat-free mass (percentage of fat-free mass that is water) between those studied in the luteal and follicular phase (data not shown). Thus, premenopausal women were pooled.
Total and regional body composition data are shown in Table 2. Fat mass and percentage fat were greater (both P<0.01) in postmenopausal compared to premenopausal women. No differences in total body fat-free mass, regional fat-free mass or appen-dicular skeletal muscle mass were noted between groups. Abdominal subcutaneous fat and intra-abdominal fat areas were greater (P<0.05 and P<0.01, respectively) in postmenopausal compared to premenopausal women.
Differences in fat mass and abdominal fat distribution after statistical adjustment for covariates or in sub-groups matched for selected variables are shown in Figure 1. The difference in fat mass persisted after statistical control for age (premenopausal, 18±8 vs postmenopausal, 23±8 kg; P<0.01) or when fat mass was compared within a sub-group (n=7 per group) of women matched for age (premenopausal, 16±9 vs postmenopausal, 27±9 kg; P<0.01). No difference in abdominal subcutaneous fat was noted after statistical control for age and fat mass (premenopausal, 251±50 vs postmenopausal, 233±51 kg) or in a sub-group (n=20 per group) of women matched for fat mass (premenopausal, 277±100 vs postmenopausal, 271±80 cm2). In contrast, the difference in intra-abdominal fat persisted after statistical control for age and fat mass (premenopausal, 65±22 vs postmenopausal 77±23 cm2, P<0.05). In a sub-group of women matched for fat mass, a trend toward a greater intra-abdominal fat in postmenopausal women was found (premenopausal, 71±37 vs postmenopausal, 89±32 cm2; P=0.11).
Time since menopause was not related to fat mass (r=0.22), intra-abdominal fat (r=0.10), abdominal subcutaneous fat (r=0.13) or other body composition variables (data not shown).
To our knowledge, this is the first study to examine simultaneously the effect of menopause status on both total body composition and abdominal fat distribution using radiologic imaging techniques. The main finding is that postmenopausal women are characterized by increased intra-abdominal body fat, independent of menopause-related differences in total body adiposity. The present study provides new information regarding the effect of menopause status on body composition and fat distribution because we accounted for the confounding effect of age and adiposity in our study design and statistical approach. We compared non-obese, middle-aged premenopausal and early-postmenopausal women to minimize the confounding effects of age and adiposity on body composition and fat distribution. Moreover, we compared body composition and abdominal fat distribution variables: (1) after statistically adjusting for the effects of age and adiposity; and (2) in sub-samples of pre- and postmenopausal women matched for age or adiposity. These design and analytical considerations permitted a rigorous examination of the effect of the menopause transition on body composition and fat distribution.
In a recent review, we noted considerable divergence among studies regarding the effect of the menopause transition on total body fat mass.11 In studies that measured body composition by dual energy X-ray absorptiometry in middle-aged premenopausal and early-postmenopausal women, either no difference in fat mass11 or a trend toward a greater fat mass in postmenopausal women12 was noted. In the present study, we observed a greater fat mass in postmenopausal compared to premenopausal women. The reasons for divergent results among studies are not readily apparent. Age is an unlikely contributor since the difference in age between pre- and postmenopausal women was small (∼4–5 y) and similar among studies. Moreover, differences in fat mass in the present study persisted after statistical control for age or when pre- and postmenopausal women were matched for age (Figure 1). Another possible explanation for differences among studies is the amount of time after menopause that women were evaluated. Several studies have shown a positive relationship between fat mass and time since menopause.11,12 This explanation also seems unlikely, however, given that fat mass was greater in postmenopausal women in the present study despite having a shorter average time since menopause compared to other studies finding no menopause-related difference in fat mass.12 The most likely explanation for differing results appears to be variation in the populations studied: Svendson et al studied Danish women,13 Trémollieres et al studied French women12 and we studied American women. Genetic or cultural factors, may have contributed to divergent findings. Irrespective of the reasons for differing results, when taken together with our preliminary longitudinal data,1 our findings favor the hypothesis that the accumulation of body fat is accelerated during or directly following the menopause transition.
Abdominal fat distribution
To our knowledge, only three studies have examined abdominal fat distribution in pre- and postmenopausal women using computed tomography.14,15,16 Hunter et al14 studied a large number of women throughout the age range (17–77 y) and found greater intra-ab-dominal and abdominal subcutaneous fat in postmenopausal compared to premenopausal women. Zamboni and co-workers15 tested obese women of various ages (premenopausal, 35±11 y and postmenopausal, 60±8 y) and found greater intra-abdominal fat, but lower subcutaneous fat, in postmenopausal compared to premenopausal women. Finally, Kotani et al16 examined intra-abdominal fat volume in obese women throughout the age range (10–79 y). They found that the slope of the relationship between intra-abdominal fat volume and age was 2.6-fold greater in postmenopausal compared to premenopausal women. One might conclude from the aforementioned studies that menopause is associated with an increase in intra-abdominal fat. However, because these studies examined obese women and included women throughout the lifespan, it is unclear whether these results were confounded by age, obesity or a combination of both. For example, in the study by Kotani et al,16 their finding of a greater increase in intra-abdominal fat in postmenopausal women was largely due to greater intra-abdominal fat volumes in older women (>60 y), since no apparent difference in intra-abdominal fat volume was evident between pre- and postmenopausal women of similar age (i.e. between the ages of 40 and 55 y).
Our study extends these preliminary observations by considering confounding factors such as age and adiposity in the study design (i.e. comparison of non-obese middle-aged premenopausal and early-postmen-opausal women) and analytical approach (i.e. statistically adjusting for age and adiposity and matching patients for age or adiposity). Both abdominal subcutaneous and intra-abdominal fat were greater in postmenopausal compared to premenopausal women. After statistical adjustment for age and total body fat mass, however, only intra-abdominal fat remained significantly higher in postmenopausal women. A similar pattern of differences in abdominal fat areas were found when pre- and postmenopausal women were matched for total body fat mass (see Figure 1). These findings suggest that the greater intra-abdominal fat in postmenopausal women is independent of total adiposity. Our interpretation of these results is that the menopause transition is associated with a preferential storage of fat in the intra-abdominal compartment. Changes at the cellular level, such as increased lipoprotein lipase activity or decreased lipolysis, that result from estrogen deficiency and increased androgenicity induced by the menopause transition, may explain the redistribution of fat to the intra-abdominal depot.17 Increased intra-abdominal fat is associated with a clustering of metabolic abnormalities including insulin resistance and dys-lipidemia which may predispose postmenopausal women to a greater risk for cardiovascular disease and diabetes.18,19 An understanding of the mechanisms underlying the redistribution of fat toward the abdominal compartment and the role of estrogen deficiency in this process, therefore, has important health implications for postmenopausal women.
Fat-free and skeletal muscle mass
No differences in total body fat-free mass or appendicular skeletal muscle mass were noted. These findings are in contrast to previous cross-sectional20 and longitudinal1 studies from our laboratory and others21 that suggest the menopause transition is associated with a reduction in fat-free mass and, in particular, skeletal muscle mass. The inability to detect a menopause effect on fat-free mass or appendicular skeletal muscle mass in the present study may relate to the increased body mass noted in postmenopausal women. During weight gain, approximately 25% of the excess weight gained is fat-free mass.22 Thus, assuming that the observed difference in body mass between pre- and postmenopausal women (4 kg; see Table 1) is due to weight gain, a 1 kg increase in fat-free mass in postmenopausal women would be expected (i.e. 25% of the weight difference between pre- and postmenopausal women). This increase in fat-free mass associated with weight gain may mask the menopause-related loss of fat-free mass. In fact, after statistical control for differences in body mass, fat-free mass was significantly lower in postmenopausal compared to premenopausal women (pre, 41±3 vs post, 39±3 kg; P=0.01). This difference in fat-free mass (2 kg) following adjustment for body mass is concordant with the menopause-related reduction in fat-free mass found in our previous longitudinal study (−2.5 kg). Thus, our results suggest a moderate effect of menopausal status on fat-free mass. Longitudinal studies that measure changes in fat-free mass during the menopause transition, however, are needed to clarify the effect of the menopause transition on fat-free tissue and skeletal muscle mass.
The mechanisms underlying the menopause-related loss of fat-free mass are unclear. Estrogen receptors are present in skeletal muscle,23 the largest component of fat-free mass, suggesting that estrogen may have a direct effect in maintaining various tissue components of fat-free mass. Menopause-related changes in growth factors may also contribute to changes in fat-free tissue mass. For example, prior work from our laboratory showed that insulin-like growth factor-1 levels decreased during the menopause transition.24 Finally, menopause-related reductions in physical activity1 may promote skeletal muscle atrophy and the loss of fat-free mass. Whatever the mechanism, the loss of fat-free mass, in particular skeletal muscle mass, may hasten the development of disability by reducing strength25 and endurance.26
In conclusion, results from the present study suggest that the menopause transition is associated with increased total and central adiposity. In contrast, no differences in fat-free mass were noted. Furthermore, the menopause transition appears to promote the selective accumulation of fat in the intra-abdominal compartment. Longitudinal studies that measure multiple compartments of body composition (including bone mass) and abdominal fat distribution are needed to confirm the present results.
The authors would like to thank all the participants who volunteered their time for this study. We are grateful to Denise Defalco-McGeein, Martin Brochu and Chris Potter for their skilled assistance. This work was supported by grants from the NIH (AG-13978; M01 RR-1093252; AG-151121), the United States Department of Agriculture (96-35200-3488), General Clinical Research Center (RR-00109) and the American Federation for Aging Research. Dr Tchernof is the recipient of a postdoctoral research fellowship from the Canadian Diabetes Association.