Changes in body composition during post-menopausal hormone therapy: a 2 year prospective study*

A. Arabi1, P. Garnero2, R. Porcher3, C. Pelissier4, C.L. Benhamou5 and C. Roux1,6

1 Centre d’Evaluation des Maladies Osseuses, Département de Rhumatologie, Hôpital Cochin, Université René Descartes, Paris, 2 Synarc, INSERM U403, Lyon, 3 Département de Biostatistique, Hôpital Saint-Louis, et INSERM U444, Paris, 4 Service de Gynécologie, Hôpital Hôtel-Dieu, Paris and 5 Service de Rhumatologie, Hôpital Porte Madeleine, et ERIT-M0101, Orleans, France

6 To whom correspondence should be addressed at: Hôpital Cochin, Service de Rhumatologie, Centre d’Evaluation des Maladies Osseuses, 27 rue du Faubourg Saint Jacques, 75014 Paris, France. e-mail: christian.roux{at}cch.ap-hop-paris.fr


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: Post-menopausal hormone therapy (pHT) induces changes in both body composition and bone mineral density (BMD). METHODS: In 109 post-menopausal women beginning either tibolone 2.5 mg (n = 29), tibolone 1.25 mg (n = 42) or estradiol 2 mg plus norethisterone acetate 1 mg (E2 + NETA) (n = 38), we assessed body composition, total and regional BMD by dual energy X-ray absorptiometry, and the serum bone alkaline phosphatase (BAP), osteocalcin and the urinary excretion to type I collagen C-telopeptide (CTX) at baseline and after 2 years. RESULTS: At baseline, BMD at all sites correlated negatively with age and years since menopause, and positively with lean mass and fat mass (r = 0.42, P < 0.001 and r = 0.26, P = 0.006 at the total femur). During treatment, BMD increased at all sites (P < 0.001), and serum BAP, osteocalcin, and urinary CTX decreased in all groups (P < 0.001). Lean mass increased whereas android fat and android obesity index decreased. The increase in BMD at all sites correlated positively with changes of lean mass at 2 years. CONCLUSIONS: Both fat mass and lean mass are related to BMD in post-menopausal women, the relationship being strongest with lean mass; an increase in lean mass and a change in distribution of body fat are observed during treatment with E2 + NETA and tibolone.

Key words: body composition/body mass index/fat mass/hormone replacement therapy/lean mass


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Menopause is associated with an accelerated bone loss (Hansen et al., 1995Go; Warming et al., 2002Go), an increase in body weight (Wing et al., 1991Go) with changes in body composition characterized by a decrease in lean mass, and an increase and redistribution of fat mass (Wang et al., 1994Go; Aloia et al., 1995Go; Trémollieres et al., 1996Go). The relative proportion of android fat increases in post-menopausal women; this phenomenon, although it might be beneficial for bone density (Heiss et al., 1995Go), is an independent cardiovascular risk factor (Lapidus et al., 1984Go; Larsson et al., 1984Go).

Several studies attempted to find out which body compartment is the major determinant of bone mineral density (BMD), and, beyond the mechanical loading, what are the other mechanisms by which one compartment or another affects the skeleton. These studies have yielded conflicting results. Some of them found that both fat and lean mass are independently related to BMD (Khosla et al., 1996Go; Pluijm et al., 2001Go), the influence being strongest with fat mass in some studies (Compston et al., 1992Go; Reid et al., 1992Go), whereas BMD is more closely related to lean tissue than to fat mass for others (Salamone et al., 1995Go; Chen et al., 1997Go).

Post-menopausal hormone therapy (pHT) prevents bone loss that follows menopause. Considerable attention has been focused on the effects of pHT on body weight and body composition changes occurring with menopause. Some studies found that changes are not prevented by pHT (Aloia et al., 1995Go), or that its effect on the upper body fat deposition is relatively small (Silverstein and Barret-Connor, 1996Go); other studies report that pHT blunts the increase in body weight, promotes a gynoid distribution and prevents the central shift of body fat commonly observed after the menopause (Haarbo et al., 1991Go; Reubinoff et al., 1995Go; Troisi et al., 1995Go; Perrone et al., 1999Go; Sorensen et al., 2001Go). However, a recent Cochrane review concluded that there is not enough data to determine whether pHT has a preventive effect on redistribution of body fat which is associated with the menopause (Norman et al., 2003Go).

The objectives of this study were to compare the 2 year effects of tibolone with those of a combined 17{beta}-estradiol and norethisterone acetate (E2 + NETA) treatment on BMD and body composition in post-menopausal women, and to study the relationship of the changes in each of these parameters to changes in the others.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subjects
A total of 109 healthy post-menopausal women are the basis of this study. They are part of a population of 163 women who completed a prospective randomized, double-blind study of prevention of bone loss (Roux et al., 2002Go), conducted in two centres, using two randomization lists. Subjects included in this part of the study were those referred to one centre, having body composition data available at baseline and at the end of the study. All participants were aged >=45 years with intact uterus. Criteria for entry in the study included: menopause for 1–10 years confirmed by serum levels of estrogen <=50 pg/ml and FSH >=20 IU/l. Exclusion criteria were: bilateral oophorectomy, undiagnosed vaginal bleeding, endometrial hyperplasia, inflammatory disease or any condition likely to require corticosteroid treatment, bone diseases, liver enzyme abnormalities, a body mass index (BMI) <19 or >27 kg/m2, a history of cancer, or cigarette consumption >10 cigarettes daily. Subjects were also excluded if they had used estrogens within the last 6 months, calcitonin or vitamin D treatment within the last 2 months, and drugs known to interfere with calcium metabolism (Roux et al., 2002Go).

All subjects gave written informed consent to participate in the study, which was carried out in accordance with the Helsinki declaration.

Treatments
Subjects received either tibolone 2.5 mg (n = 29), tibolone 1.25 mg (n = 42), or estradiol 2 mg plus NETA 1 mg (n = 38), daily at bedtime. All women received a daily supplementation of 500 mg of calcium with a meal. Follow-up duration was 2 years.

Assessments
Height, weight and BMI (calculated as the weight in kg per height in m2) were determined at baseline, and after 2 years of follow-up for each subject. At each time point, BMD (g/cm2) was measured at the left hip (the femoral neck and the total hip), and the lumbar spine L2–L4 in the antero-posterior scan, by dual energy X-ray absorptiometry (DEXA), using a QDR 2000 device (Hologic, USA). BMD of the whole skeleton, and the body composition, i.e. lean mass and fat mass (in grams) were determined at the same time, on the total body scans.

From body composition data, we calculated the android fat and the gynoid fat percentage, and the android obesity index for each subject. Android fat was the proportion of fat in the trunk region (area between an upper horizontal border below the chin, a lower border formed by the oblique lines passing through the hip joints and vertical borders lateral to the ribs); gynoid fat was the proportion of fat in both legs (area below the upper border formed by the oblique lines passing through the hip joints), both expressed as percentage of subtotal body fat (assuming that brain fat represents 17% of the total body fat, the head region was excluded from our analysis). Android obesity index was the ratio of android fat/gynoid fat (Trémollieres et al., 1996Go).

At the same time points, serum concentrations of bone alkaline phosphatase (using immunoradiometric assay, Ostase; Beckman– Coulter, USA), osteocalcin (using radioimmunoassay, Elsa Osteo; CisBio Internationl, France) and the urinary excretion of type I collagen C-telopeptide (CTX) corrected for creatinine, on the second-void urine sample (using enzyme linked immunosorbent assay, Crosslaps ELISA; Nordic Biosciences, Denmark) were measured in 101 women after an overnight fast.

Statistical analysis
Data are presented as mean and SD. Cross-sectional associations at inclusion between BMD at several sites, BMI, years since menopause, lean mass, fat mass and fat repartition indexes, and associations between relative variations at 2 years of these parameters were assessed using Pearson’s correlation coefficient. Multiple linear regression, equivalent to partial correlation coefficients, was used to adjust these correlations for other covariates.

Tests for changes in parameters in the pooled sample used one-sample Student’s t-tests. Comparisons of characteristics between the three groups at inclusion and relative variations of the studied parameters were performed using a one-way analysis of variance. Additionally, post-hoc two-by-two comparisons were adjusted for multiple testing using Tukey’s correction.

A multiple regression model was built to predict changes in BMD according to treatment and changes in body composition repartition indexes and biochemical markers using analysis of co-variance. The model was selected on the basis of a backward model selection procedure, where all parameters achieving a P-value of 0.20 in the univariable analysis were considered.

All tests were two-sided. For the one-way analysis of variance, type I error rate was fixed at 0.05, except for post-hoc comparisons. To account for a relative high number of tests for correlation coefficients, such tests were performed with significance level fixed at 0.01. Analyses were performed using S-Plus 2000 software (MathSoft Inc., USA).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Baseline characteristics of patients according to treatment group are shown in Table I. There was no statistically significant difference in these variables between groups.


View this table:
[in this window]
[in a new window]
 
Table I. Baseline characteristics of the study population
 
At baseline, lean mass correlated negatively with age and years since menopause (YSM), but these correlations did not reach statistical significance: P = 0.013 and P = 0.018 respectively. Fat mass correlated neither with age nor with YSM. We observed a significant increase in android fat and android obesity index with age: r = 0.27, P = 0.0047 and r = 0.25, P = 0.0085 respectively, but not with YSM: r = 0.22, P = 0.020 and r = 0.21, P = 0.027. In contrast, there was a decrease in gynoid fat, but this decrease did not reach significance with age r = –0.21, P = 0.027 or with YSM r = –0.20, P = 0.037.

The relationships between BMD and different parameters at baseline are shown in Table II. BMD at all sites was related to YSM. Similar results were observed with age (data not shown). Multiple regression analysis showed that BMD was more related to YSM than to age, whereas body composition variables were more related to age than to YSM (data not shown). A positive correlation was found between total and regional BMD and BMI. At all sites, there was a weak correlation between 0.18 and 0.27 with fat mass, and between 0.33 and 0.47 with lean mass.


View this table:
[in this window]
[in a new window]
 
Table II. Pearson’s correlation coefficients (P-value) between bone mineral density and other parameters at baseline
 
Changes observed after 2 years of treatment, expressed as a percentage change from the baseline measurements, according to treatment groups, are shown in Table III. In the whole population, BMD increased significantly at all sites. Mean gain was 2.3, 4.1 and 2.5% in the total body, lumbar spine and total hip BMD respectively. These changes were significantly different according to group (P = 0.013, P < 0.001 and P = 0.011 respectively). After adjustment for multiple testing, the increase in lumbar spine BMD was found to be higher in E2 + NETA group than in tibolone groups, while the analysis only yielded significant differences between E2 + NETA group and tibolone 1.25 mg at both other sites. Serum bone alkaline phosphatase, osteocalcin and urinary cross-links decreased by 27.4, 35.7 and 50.6%, respectively. There were no significant differences between groups.


View this table:
[in this window]
[in a new window]
 
Table III. Percentage changes at 2 years from baseline measurements with tibolone 2.5 mg, tibolone 1.25 mg, and E2 + NETA treatments
 
The mean increase in BMI was 2.5 ± 5.4, 2.6 ± 4.3 and 3.5 ± 6.4 in tibolone 2.5 mg, tibolone 1.25 mg and E2 + NETA groups respectively. A mean increase in lean mass equivalent to 7.3 ± 4.2, 3.4 ± 3.6 and 5.5 ± 4.3%, from baseline was observed in tibolone 2.5 mg, tibolone 1.25 mg and E2 + NETA groups respectively. In the whole cohort, this increase was significant (P < 0.0001), and was different according to the group (P = 0.0004), but two-by-two comparisons only yielded a significant difference between tibolone 2.5 mg and tibolone 1.25 mg groups. Total body fat decreased in the three groups, but this was globally not significant. This decline was associated with changes in fat distribution with slight but significant decrease in the percentage of android fat and the android index from baseline. These changes were more pronounced in the E2 + NETA group than tibolone 2.5 mg and tibolone 1.25 mg groups, although the only significant difference was observed between tibolone 1.25 mg and E2 + NETA groups for the percentage android fat, when adjusting for multiple testing.

The relationships between the changes in total and regional BMD with changes in body composition are shown in Table IV. There was a significant positive correlation between changes in BMD from baseline at all sites and changes in lean mass at 2 years. In a multiple regression model, the E2 + NETA treatment (P = 0.04), and the changes in lean mass (P = 0.0008) at 2 years were selected as independent predictors of changes in femoral BMD from baseline. No interactions between these covariates were found, and the multiple R2 of the model was 0.18. Similar results were obtained for BMD changes measured at other sites.


View this table:
[in this window]
[in a new window]
 
Table IV. Pearson’s correlation coefficients (P-value) between changes in total and regional bone mineral density with changes in body composition parameters
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The present study suggests a link between lean mass increase and the increase in BMD at all sites in post-menopausal women receiving post-menopausal hormone therapy.

At baseline, we observed the expected effects of hormonal deficiency: BMD at all sites was related to YSM. BMD correlated positively with BMI at all sites, as previously reported (Ribot et al., 1988Go; Harris et al., 1992Go). Both total fat mass and lean mass were significantly related to total and regional BMD, but the correlation was stronger with lean mass (~0.4, versus ~0.25 for fat mass), in accordance with others (Salamone et al., 1995Go) who suggested that a low lean mass may be considered as an osteoporotic risk factor. A BMI between 19 and 27 was an inclusion criterion for this study, and this precludes accurate analysis of the role of fat mass in obese post-menopausal women.

The mechanical loading might explain the correlation of fat and lean tissues with BMD, but does not explain the difference in these correlations, as the skeleton cannot distinguish between a pound of lean tissue and a pound of fat (Slemenda, 1995Go). The correlation of BMD with lean mass may reflect a genetic association between a higher lean tissue mass and a higher peak bone mass (Compston et al., 1992Go; Chen et al., 1997Go). Physical activity has been suggested to play a role in this relationship (Chen et al., 1997Go) as exercise results in greater muscle mass and may increase bone density. However, one may wonder whether the exercise-induced increase in BMD is related to exercise itself or to the increase in lean mass. Another suggested explanation of the link lean mass–BMD is the circulating IGF-I (Salamone et al., 1995Go), which has potent anabolic actions on both skeletal muscles and bone. In our study, lean mass increased significantly in all treatment groups, in contrast to some cross-sectional and longitudinal studies which reported either an acceleration of lean mass loss (O’Sullivan et al., 1998Go) or a failure of pHT to prevent this loss (Hassager and Christiansen, 1989Go; Haarbo et al., 1991Go). In these studies, pHT used were an estrogen plus either cyproterone acetate or a progestagen other than NETA. In one study (Sorensen et al., 2001Go), E2 + NETA were used, and results similar to ours were found. In another prospective study where oral E2 + dydrogesterone, transdermal E2 + dydrogesterone, or tibolone were used, the loss in lean mass observed with placebo was prevented in tibolone and transdermal E2 + dydrogesterone, but not with oral E2 + dydrogesterone (Hänggi et al., 1998Go). NETA is a synthetic progestagen known to have androgenic properties, and tibolone is a synthetic steroid, with a large progestogenic activity, and to a lesser extent androgenic and estrogenic properties. We did not include a placebo group in our study; thus we cannot state that the measured changes are related to treatment. However, because such changes are unexpected in untreated post-menopausal women, we speculate that the gain in lean mass may be related to an anabolic effect of NETA and tibolone on skeletal muscles.

Similarly to our findings, in a large cohort study, including 2016 early post-menopausal women, followed-up for 5 years, the change in lean mass was the best predictor of bone changes in pHT users (Jensen et al., 2003Go). The significant increase in lean mass with treatment and the positive correlation between its changes at 2 years with the changes in BMD at all sites, suggest that lean mass plays a mediating role in the effects of treatment on BMD. This effect may be direct or by the mean of another hormonal factor. Growth hormone has been reported to be a potent mitogen for osteoblasts in vitro (Hock et al., 1988Go), in-vivo IGF-I levels correlated positively with BMD at all sites in women (Sugimoto et al., 1997Go), and changes in circulating IGF-I correlated with changes in osteocalcin, PINP and bone alkaline phosphatase in women on pHT, suggesting that some of the effect of E2 on bone metabolism may be mediated by IGF-I (Garnero et al., 1999Go). However, the estrogen used in our study was an oral form, and, although one study found an increase in IGF-1 with oral estrogen therapy (Posaci et al., 2001Go), some studies (Helle et al., 1996Go; Raudaskoski et al., 1998Go; Garnero et al., 1999Go) found that oral E2 led to a decrease in these levels.

The most commonly given explanation for the correlation between BMD and fat mass was aromatization (Compston et al., 1992Go; Lindsay et al., 1992Go), as fat tissue is responsible for this phenomenon which is the principal source of estrogen in post-menopausal women. This theory was supported by the results of Jensen et al., who found that the effect of fat mass on the femoral neck and the lumbar spine BMC was markedly reduced in women receiving pHT as compared with those who were not. The positive relationship found between fat mass and BMD at baseline in this study, and the absence of correlation between fat mass changes and BMD changes under pHT, further supports this assumption. High fat mass is associated with high insulin level in obese people; in addition, insulin has anabolic effects on osteoblasts (Hickman and McElduff, 1989Go), and serum insulin levels have been reported as related to BMD in post-menopausal women (Reid et al., 1993Go), suggesting that insulin may play a role in this relationship. Although there were no significant changes in total fat mass with pHT in this study, body fat redistribution observed with age and menopause changed during treatment, in accordance with some previous reports (Haarbo et al., 1991Go; Reubinoff et al., 1995Go; Troisi et al., 1995Go; Perrone et al., 1999Go; Sorensen et al., 2001Go). Despite these individual conclusions, a substantive systematic review of randomized studies comparing pHT to placebo or no treatment, failed to find sufficient data to enable a meta-analysis on the effect of pHT on waist:hip ratio (Norman et al., 2003Go). Although android obesity has been reported as beneficial for bone mass because of its association with lower sex hormone binding globulin (SHBG), which results in higher concentrations of the free androgens and estrogens (Heiss et al., 1995Go), a high android obesity index is undesirable because it has been linked to insulin resistance and alteration in lipid profile, both associated with high risk of coronary artery disease (Lapidus et al., 1984Go; Larsson et al., 1984Go).

This study lacks information about previous pHT use, which, if it had happened, might have affected the baseline relationship between body composition and years since menopause. Other limitations to this study are related to the young age of the population and the exclusion of obese women, which cut out a large proportion of post-menopausal women, and make the impact of age and fat mass less easy to determine. Finally, we have no information on physical fitness changes during follow-up.

Despite these limitations, we conclude that bone mass in post-menopausal women is more affected by lean mass than fat mass. During E2 + NETA and tibolone treatments, a decrease in android fat and an increase in lean mass are observed.


    FOOTNOTES
 
* Data presented in part at the ASBMR 24th annual meeting at San Antonio, Texas, USA. Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Aloia, J.F., Vaswani, A., Russo, L. Sheehan, M. and Flaster, E. (1995) The influence of menopause and hormone replacement therapy on body cell mass and body fat mass. Am. J. Obstet. Gynecol., 172, 896–900.[ISI][Medline]

Chen, Z., Lohman, T.G., Stini, W.A., Ritenbaugh, C. and Aickin, M. (1997) Fat or lean tissue mass: which one is the major determinant of bone mineral mass in healthy postmenopausal women? J. Bone Miner. Res., 12, 144–151.[Medline]

Compston, J.E., Bhambhani, M., Laskey, M.A., Murphy, S. and Khaw, K.T. (1992) Body composition and bone mass in postmenopausal women. Clin. Endocrinol., 37, 426–431.[ISI][Medline]

Garnero, P., Tsouderos, Y., Marton, I., Pelissier, C., Varin, C. and Delmas, P.I. (1999) Effects of intranasal 17 beta-estradiol on bone turnover and serum insulin-like growth factor I in postmenopausal women. J. Clin. Endocrinol. Metab., 84, 2390–2397.[Abstract/Free Full Text]

Haarbo, J., Marslew, U., Gotfredsen, A. and Christiansen, C. (1991) Postmenopausal hormone replacement therapy prevents central distribution of body fat after menopause. Metabolism, 40, 1323–1326.[ISI][Medline]

Hänggi, W., Lippunert, K., Jaegert, P., Birkhauser, M.H. and Horber, F.F. (1998) Differential impact of conventional oral or transdermal hormone replacement therapy or tibolone on body composition in postmenopausal women. Clin. Endocrinol., 48, 691–699.[CrossRef][ISI][Medline]

Hansen, M.A., Overgaard, K. and Christiansen, C. (1995) Spontaneous postmenopausal bone loss in different skeletal areas followed up for 15 years. J. Bone Miner. Res., 10, 205–210.[ISI][Medline]

Harris, S., Dallal, G.E., Dawson-Hughes, B. (1992) Influence of body weight on rates of change in bone density of the spine, hip and radius in postmenopausal women. Calcf. Tissue Int., 50, 19–23.[CrossRef]

Hassager, C. and Christiansen, C. (1989) Estrogen/gestagen therapy changes soft tissue body composition in postmenopausal women. Metabolism, 38, 662–665.[ISI][Medline]

Heiss, C.J., Sanborn, C.F., Nichols, D.L., Bonnick S.L. and Alford, B.B. (1995) Associations of body fat distribution, circulating sex hormones, and bone mineral density in postmenopausal women. J. Clin. Endocrinol. Metab., 80, 1591–1596.[Abstract]

Helle, S.I., Omsjo, I.H., Hughes, S.C., Botta, L., Hulls, G., Holly, J.M. and Lonnin, P.E. (1996) Effects of oral and transdermal oestrogen replacement therapy on plasma levels of insulin-like growth factors and IGF binding proteins 1 and 3: a cross-over study. Clin. Endocrinol. (Oxf)., 45, 727–732.[ISI][Medline]

Hickman, J. and McElduff, A. (1989) Insulin promotes growth of the cultured rat osteosarcoma cell line UMR-106-01: an osteoblast-like cell. Endocrinology, 124, 701–706.[Abstract]

Hock, J.M., Centrella, M. and Cnalis, E. (1988) IGF-1 has independent effects on bone matrix formation and cell replication. Endocrinology, 122, 254–260.[Abstract]

Jensen, L.B., Vestergaard P., Hemann A.P., Gram, J., Eiken, P., Abrahamsen, B., Brot, C., Kolthoff, N., Sorensen, O.H., Beck-Nielson, H. et al. (2003) Hormone replacement therapy dissociates fat mass and bone mass, and tend to reduce weight gain in early postmenopausal women: a randomised controlled 5-year clinical trial of the Danish Osteoporosis Prevention Study. J. Bone Miner. Res., 18, 333–342.[ISI][Medline]

Khosla, S., Atkinson, E.J., Riggs, B.L. and Milton, L.J. III (1996) Relation between body composition and bone mass in women. J. Bone Miner. Res., 11, 857–863.[ISI][Medline]

Lapidus, L., Bengtsson, C., Larsson, B.D., Pennert, K., Rybo, E. and Sjostrom, I. (1984) Distribution of adipose tissue and risk of cardiovascular disease and death: a 12 year follow up of participants in the population study of women in Gothenburg, Sweden. Br. Med. J. (Clin. Res. Edn), 10, 1257–1261.

Larsson, B., Svardsudd, K., Welin, L., Wilhelmsen, L., Bjorntorp, P. and Tibbli, G. (1984) Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow up of participants in the study of men born in 1913. Br. Med. J. (Clin. Res. Edn), 288, 1401–1404.[ISI][Medline]

Lindsay, R., Cosman, F., Herrington, B.S. and Himmelstein, S. (1992) Bone mass and body composition in normal women. J. Bone Miner. Res., 7, 55–63.[ISI][Medline]

Norman, R.J., Flight, I.H.K. and Rees, M.C.P. (2003) Oestrogen and progestogen hormone replacement therapy for peri-menopausal and post-menopausal women: weight and body fat distribution (Cochrane Review). In The Cochrane Library, Issue 1, Update Software, Oxford.

O’Sullivan, A.J., Crampton, L.J., Freund, J. and Ho, K.K. (1998) The route of estrogen replacement therapy confers divergent effects on substrate oxidation and body composition in postmenopausal women. J. Clin. Invest., 102, 1035–1040.[Abstract/Free Full Text]

Perrone, G., Liu, Y., Capri, O., Critelli, C., Barillaro, F., Galoppi, P. and Zichella, L. (1999) Evaluation of the body composition and fat distribution in long-term users of hormone replacement therapy. Gynecol. Obstet. Invest., 48, 52–55.[CrossRef][Medline]

Pluijm, S.M.F., Visser, M., Smit, J.H., Popp-Snijders, S., Roos, J.C. and Lips, P. (2001) Determinants of bone mineral density in older men and women. Body composition as mediator. J. Bone Miner. Res., 16, 2142–2151.[ISI][Medline]

Posaci, C., Altunyurt, S., Islekel, H. And Onvural, A. (2001) Effects of HRT on serum levels of IGF-1 in postmenopausal women. Maturitas, 40, 69–74.[CrossRef][ISI][Medline]

Raudaskoski, T., Knip, M. and Laatikainen, T. (1998) Plasma insulin-like growth factor-I and its binding proteins 1 and 3 during continuous nonoral and oral combined hormone replacement therapy. Menopause, 5, 217–222.[ISI][Medline]

Reid, I.R., Ames, R., Evans, M.C., Sharpe, S., Gamble, G., France, J.T, Lim, T.M. and Cundy, T.F. (1992) Determinants of total body and regional bone mineral density in normal post menopausal women—a key role for fat mass. J. Clin. Endocrinol. Metab., 75, 45–51.[Abstract]

Reid, I.R., Evans, M.C., Cooper, G.J., Ames, R.W. and Stapelton, J. (1993) Circulating insulin levels are related to bone density in normal postmenopausal women. Am. J. Physiol., 265, E655–E659.

Reubinoff, B.E., Wurtman, J., Rojansky, N., Adler, D., Stein, P., Schenker, J.G. and Brzezinski, A. (1995) Effects of hormone replacement therapy on weight, fat distribution, and food intake in early postmenopausal women: a prospective study. Fertil. Steril., 64, 963–968.[ISI][Medline]

Ribot, C., Trémollieres, F., Pouilles, J.M., Bonneu, J.M., Germain, F. and Louvet, J.P. (1988) Obesity and postmenopausal bone loss: the influence of obesity on vertebral density and body turnover in postmenopausal women. Bone, 8, 327–331.[CrossRef][ISI]

Roux, C., Pelissier, C., Fechtenbaum, J., Loiseau-Peres, S. and Benhamou, C.L. (2002) Randomized, double blind, 2-year comparison of tibolone with 17 {beta} estradiol and norethindrone acetate in preventing postmenopausal bone loss. Osteoporos. Int., 13, 241–248.[CrossRef][ISI][Medline]

Salamone, L.M., Glynn, N., Black, D., Epstein, R.S., Palermo, L., Meilahn, E., Kuller, L.H. and Cauley, J.A. (1995) Body composition and bone mineral density in premenopausal and early perimenopausal women. J. Bone Miner. Res., 10, 1762–1768.[ISI][Medline]

Silverstein, D.K. and Barret-Connor, E. (1996) Long-term postmenopausal hormone use, obesity, and fat distribution in older women. J. Am. Med. Assoc., 275, 46–49.[Abstract]

Slemenda, C.W. (1995) Body composition and skeletal density—mechanical loading or something more? [Editorial.] J. Clin. Endocrinol. Metab., 80, 1761–1763. [ISI][Medline]

Sorensen, M.B., Rosenfalck, A.M., Hojgaard, L. and Ottesen, B. (2001) Obesity and sarcopenia after menopause are reversed by sex hormone replacement therapy. Obes. Res., 9, 622–626.[Abstract/Free Full Text]

Sugimoto, T., Nishiyama, K., Kuribayashi, F. and Chihara, K. (1997) Serum levels of insulin-like growth factor (IGF) I and IGF-binding protein (IGFBP)-2 and IGFBP-3 in osteoporotic patients with and without spinal fractures. J. Bone Miner. Res., 12, 1272–1279.[ISI][Medline]

Trémollieres, F.A., Pouilles, J.M. and Ribot, C.A. (1996) Relative influence of age and menopause on total and regional body composition changes in postmenopausal women. Am. J. Obstet. Gynecol., 175, 1594–1600.[ISI][Medline]

Troisi, R.J., Wolf, A.M., Mason, J.E., Klingler, K.M. and Colditz, G.A. (1995) Relation of body fat distribution to reproductive factors in pre-and postmenopausal women. Obes. Res., 3, 143–151.[Abstract]

Wang, Q., Hassager, C., Ravn, P., Wang, S. and Christiansen, C. (1994) Total and regional body-composition changes in early postmenopausal women: age-related or menopause-related? Am. J. Clin Nutr., 60, 843–848.[Abstract]

Warming, L., Hassager, C. and Christiansen, C. (2002) Changes in bone mineral density with age in men and women: a longitudinal study. Osteoporos. Int., 13, 105–112.[CrossRef][ISI][Medline]

Wing, R.R., Matthews, K.A., Kuller, L.H., Meilahn, E.N. and Platinga, P.L. (1991) Weight gain at the time of menopause. Arch. Intern. Med., 151, 97–102.[Abstract]

Submitted on November 26, 2002; resubmitted on March 13, 2003; accepted on May 6, 2003.