Oestrogen and progestogen hormone replacement therapy for peri‐menopausal and post‐menopausal women: weight and body fat distribution

Eugene J Kongnyuy; Robert J Norman; Ingrid HK Flight; Margaret C Rees

doi:10.1002/14651858.CD001018

. 1999 Jul 26;1999(3):CD001018. doi: 10.1002/14651858.CD001018

Oestrogen and progestogen hormone replacement therapy for peri‐menopausal and post‐menopausal women: weight and body fat distribution

Eugene J Kongnyuy ^1,^✉, Robert J Norman ², Ingrid HK Flight ³, Margaret C Rees ⁴

Editor: Cochrane Gynaecology and Fertility Group

PMCID: PMC13613310 PMID: 10796730

Abstract

Background

Hormone replacement therapy (HRT) is commonly prescribed to treat menopausal symptoms and to prevent post‐menopausal bone loss. However, many women are concerned about hormonal replacement therapy because they believe that such treatment will result in weight gain. The effect of HRT on body weight and fat distribution has not yet been examined in systematic reviews.

Objectives

To evaluate the effect of unopposed oestrogen or combined oestrogen and progestogen hormone replacement therapy (HRT) upon the body weight and fat distribution of peri‐menopausal and post‐menopausal women.

Search methods

We searched the Cochrane Menstrual Disorders and Subfertility Group Trials Register, MEDLINE, EMBASE, Current Contents, Biological Abstracts and CINAHL, most recently in July 2005.

Selection criteria

All randomised, placebo or no‐treatment controlled trials (RCTs) that detailed the effect of HRT on body weight or fat distribution. Previous HRT use should have ceased at least one month (in the case of patches, cream or gel) or three months (for oral preparations or subcutaneous pellets) before commencement of the study.

Data collection and analysis

Twenty‐eight RCTs with 28,559 women fulfilled the inclusion criteria for this review. Data was pooled for continuous variables and the weighted mean difference (MD) with 95% confidence intervals reported.

Main results

Outcomes were evaluated separately for unopposed oestrogen and oestrogen combined with progestogen regimens. No statistically significant difference was found in mean weight gain between those using unopposed oestrogen and non‐HRT users (0.03kg, 95% CI ‐0.61 to 0.67) and those using oestrogen with progestogen therapy and non‐HRT users (0.04 kg, 95% CI ‐0.42 to 0.50). There was no significant difference in body mass index (BMI) between women using unopposed oestrogen and non‐HRT users (‐0.14, 95% CI ‐0.40 to 0.12) or oestrogen combined with progestogen and non‐HRT users (‐0.10, 95% CI ‐0.27 to 0.07). Insufficient data were available to enable meta‐analysis of the effect of HRT on waist to hip ratio, fat mass or skinfold thickness.

Authors' conclusions

There is no evidence of an effect of unopposed oestrogen or combined oestrogen with progestogen on body weight and on the BMI increase normally experienced at the time of menopause. Insufficient evidence currently exists to enable examination of the effect of HRT on waist‐hip ratio, fat mass or skinfold thickness.

Keywords: Female, Humans, Body Mass Index, Estrogen Replacement Therapy, Estrogen Replacement Therapy/adverse effects, Estrogens, Estrogens/adverse effects, Estrogens/therapeutic use, Postmenopause, Premenopause, Progestins, Progestins/adverse effects, Progestins/therapeutic use, Weight Gain, Weight Gain/drug effects

Plain language summary

Hormone replacement therapy has no effect on body weight and cannot prevent weight gain at menopause.

Hormone replacement therapy (HRT) is used to reduce the symptoms of menopause and bone loss after menopause. Some women decline to take HRT because they believe it causes weight gain. The review of trials found no evidence that unopposed oestrogen and combined oestrogen and progestogen have an effect on body weight additional to that usually gained at the time of menopause. The review did not find any evidence that HRT prevents weight gain experienced at menopause.

Background

Description of the condition

Hormone replacement therapy (HRT) is commonly prescribed to treat menopausal symptoms and to prevent post‐menopausal bone loss (Rees and Purdie 2006). However, many women are concerned about hormonal replacement therapy because they believe that such treatment will result in weight gain (Hope 1995; Reynolds 2002).

Description of the intervention

Evidence suggests that women increase in body weight with age, and this tends to begin at or near menopause (Williams 2006). Additionally, body fat distribution changes with age, independent of weight gain, and these changes begin around the time of the menopause. There is an increase in body fat as a percentage of body weight and a redistribution of body fat with a relative increase in the proportion of abdominal fat (Tchernof 2000). This centralised abdominal (android) fat distribution is recognised as an independent predictor of (or risk factor for) cardiovascular disease (Gill 2006).

How the intervention might work

Contrary to the widely held belief of increasing body weight, HRT may in fact prevent or at least limit an increase in the body fat of menopausal women (Garcia 1990; Hassager 1991; Stevenson 1996; Chen 2005) and resist the redistribution of body fat so that a more gynaecoid, or pear‐shaped, appearance is maintained (Haarbo 1991; Stevenson 1995; Gambacciani 1997; Arabi 2003). These effects may depend upon the mode of administration of HRT (Hanggi 1998; O'Sullivan 1998), in the same manner as are apparent for symptomatic and metabolic differences (for example, in lipids and lipoproteins) (Godsland 2001; Stevenson 2004).

Why it is important to do this review

This review is an update of a previous systematic review prepared in 1999. It is an important topic since some women decline oestrogen therapy due to their concerns about resultant weight gain and thus forego its potential therapeutic benefits (Baglia 2004). Despite its benefits, HRT has been found to increase the risk of cardiovascular disease and breast cancer (MWS 2003; WHI 2004).

Objectives

1. To assess the likelihood of weight gain associated with hormone replacement therapy in peri‐menopausal and menopausal women.   2. To assess the likelihood of a redistribution of body fat from the hips and thighs (gynaecoid) to the abdomen (android) associated with hormone replacement therapy in peri‐menopausal and menopausal women.

Methods

Criteria for considering studies for this review

Types of studies

A trial was considered for inclusion if it detailed the effect of HRT on body weight or fat distribution, even if its main focus was on another outcome of HRT, for example change in the risk of osteoporosis. Only randomised controlled trials (RCTs) were included in this review.

Types of participants

Women of any ethnicity who are either peri‐menopausal or post‐menopausal.

Peri‐menopausal women are defined as those aged 45 years or more, or who have elevated follicle stimulating hormone (FSH), and who have menstruated intermittently within the last 12 months. Post‐menopausal women are defined as those aged 45 years or more and who have not menstruated for more than 12 months. Also defined as menopausal are women of any age who have had menopause artificially induced by bilateral oophorectomy (removal of both ovaries) and whose oophorectomy took place at least one month before entry into a trial.

RCTs of women with pre‐existing major disease have been excluded from consideration as are those studies where previous HRT was discontinued less than one month (in the case of patches, cream or gel) or three months (in the case or oral preparations or subcutaneous pellets) before commencement of the study.

Types of interventions

All oestrogens with and without progestogens given as peri‐ or post‐menopausal therapy versus placebo or no treatment for the control of menopausal symptoms. HRT is defined as unopposed oestrogen therapy (E) or oestrogen therapy with combined, cyclic or continuous progestogen therapy (E+P). Studies that combined HRT with another complementary therapy, such as vitamin or mineral supplementation or diet were included for consideration. The minimum period of treatment was three months with no upper limit to the period of treatment. Where answers to criteria for consideration were not readily available from a paper, first authors were contacted for further information. Other aspects of the intervention that were considered in relation to weight and body fat distribution change were dosage and length of the treatment follow‐up period.

Types of outcome measures

Primary outcomes

Body weight (kg or pounds);   Body mass index (BMI) defined as the ratio of weight to height squared W/H2;   Fat mass;

Secondary outcomes

Waist to hip ratio;   Skinfold thickness.

Search methods for identification of studies

The search strategy was kept very open and did not use search terms specifically associated with body weight parameters because it was expected that a number of trials describing other primary outcomes would include at least baseline details of body weight or fat distribution. A decision as to whether this was likely to be the case was made by IF and EJK based on information provided in abstracts.

Electronic searches

See Appendix 1.The literature search was without language restriction. All publications which described randomised controlled trials having the effect of HRT on weight or body fat distribution as a primary or secondary outcome were obtained using the search strategy for randomised trials as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005 ) by searching the Cochrane Menstrual Disorders and Subfertility Group Trials Register, MEDLINE and handsearching of specialist journals. The review authors also searched the electronic databases EMBASE, Current Contents, Biological Abstracts and CINAHL. Pharmaceutical companies were contacted by the Menstrual Disorders and Subfertility Group for information on unpublished trials but none was forthcoming.

Searching other resources

The original searches were performed in August 1998 and subsequent searches in July 2005. Trials found in this recent search are listed in the references section as studies awaiting assessment as they are still being assessed for inclusion or exclusion.

The following terms and variations of these terms were included in the search strategy:

hormone replacement therapy;   oestrogen (oestrogen) replacement therapy;   estrogens (oestrogens);   progestogens.

Data collection and analysis

Data collection and analysis was conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005).

Selection of studies

Citation lists of included trials, conference abstracts and relevant review articles were searched. Additionally, if subscribed to electronic mail alerting services for tables of contents of relevant journals.   The original searches were performed in August 1998 and subsequent searches in July 2005. Trials found in this recent search are listed in the references section as studies awaiting assessment as they are still being assessed for inclusion or exclusion.   The following terms and variations of these terms were included in the search strategy:   hormone replacement therapy;   oestrogen (oestrogen) replacement therapy;   estrogens (oestrogens);   progestogens.   The search strategy was kept very open and did not use search terms specifically associated with body weight parameters because it was expected that a number of trials describing other primary outcomes would include at least baseline details of body weight or fat distribution. A decision as to whether this was likely to be the case was made by IF and EJK based on information provided in abstracts.

Trial characteristics   1. Method of randomisation:   a) method of allocation adequately concealed (used central telephone randomisation, allocation through pharmacy, sealed opaque envelopes etc);   b) method of allocation unclear;   c) method of allocation inadequate.   2. Number of centres involved in the study.   3. Design of study.   4. Quality of blinding in relation to treatment:   a) double blind (women and investigators);   b) single blind (women or investigators);   c) no blinding;   d) unclear.   5. Number of women randomised, excluded or lost to follow up:   a) losses and withdrawals of less than 10%;   b) losses and withdrawals of 10 % or more;   c) not reported or unclear.   6. Basis for analysis:   a) intention to treat;   b) no intention to treat;   c) unclear.   7. Whether a power calculation was done.   8. Number of participants actually analysed.   9. Source of funding, if stated.   Characteristics of the study participants   1. Age and menopausal status.   2. Criteria for confirming menopausal status.   3. The location of the study and source of recruitment of participants.   4. Ethnicity.   5. Inclusion criteria.   6. Exclusion criteria.   7. Baseline equality of treatment groups:   a) groups balanced in terms of age, smoking, diet and physical activity;   b) groups not balanced in terms of age, smoking, diet and physical activity;   c) balance not reported.

Interventions used   1. Type of therapies used.   2. Mode of administration.   3. Blinding in relation to treatment:   a) containers were identical and drugs were identical in appearance;   b) HRT and placebo not identical;   c) similarity of HRT and placebo not reported.   4. Duration of trial.

Outcomes   1. Outcomes relevant to this analysis, that is body weight, BMI, waist to hip ratio, skinfold measurements and body fat mass.   2. Primary outcomes of the trial which may have differed from the outcomes relevant to this analysis.   Three of the review authors (RJN, MR and EJK) are experienced in clinical issues. Additional information on trial methodology and actual original trial data were sought from the principal author of any trials that appeared to meet the eligibility criteria, where necessary. A number of requests for additional information are currently outstanding. As most of these requests are for endpoint body weight details, it has not been possible to include the studies in the analysis at this time. Details and subsequent eligible trials will be included in updates to this review, as responses are received.   This review includes eligible trials located as at July 2005. Study selection and assessment of methodological quality took place between July and December 2005, extraction of data and analysis were performed between January and February 2006 and the update was written between March and May 2006.   Results from trials that used E and those that used E+P have been separately analysed. Those trials that had two or more treatment arms consisting of differing dosages of E or E+P respectively had their results combined for an analysis of results arising from any dosage. Dosages were also broken down for analysis into low, medium and high‐dose categories and equivalences based on moderate doses as follows.

Moderate dosages:   conjugated oestrogen: 0.625 mg to 1.25 mg;   17 beta estradiol: 1.0 mg to 2.0 mg;   micronized estradiol: 1.0 mg to 2.0 mg;   estradiol valerate: 1.0 mg to 2.0 mg;   transdermal estradiol: 50 µgm;   ethinyl estradiol: 15 µgm to 20 µgm.

Any dose below the above units was considered low and any dose above as high dosages. The breakdown of dosages was determined by RN and MR based upon their expert clinical experience, and upon Ansbacher 1994.

Data extraction and management

Two review authors independently extracted the data.

Measures of treatment effect

Statistical analysis was performed in accordance with the guidelines for statistical analysis described in the Cochrane Handbook for Reviews of Interventions as applies to continuous outcomes. The results were expressed as the absolute values of means with their standard deviations. A fixed‐effect model was used to pool the data and the mean difference (MD) with 95% confidence intervals reported.

Assessment of heterogeneity

Heterogeneity between the results of different studies was examined by inspecting the scatter in data points and the overlap in their confidence intervals and, more formally, by checking the results of the chi‐squared tests. There was no heterogeneity identified between any trials for any outcome.

Sensitivity analysis

Sensitivity analysis was performed according to the quality of allocation concealment, control group comparison (either placebo or no treatment), equality at baseline, rate of withdrawal and losses to follow up, intention‐to‐treat analysis and length of treatment period.

Results

Description of studies

Results of the search

From the studies awaiting assessment in the original review and current searches, 43 trials were identified and met initial inclusion criteria. The original review included 22 trials and excluded 10. The current update and original review identified a total of 75 trials that met initial inclusion criteria.

This update adds six trials to the 22 trials of the original review. Overall, 28 RCTs met the criteria for inclusion in this review. Aloia 1995 met the initial inclusion criteria but was unable to provide endpoint data in a form that could be used because its effects were estimated by linear regression, with an absence of means and standard deviations. This trial has not been included in the meta‐analysis although it is included in the following discussion. A further 16 trials are duplicates of the trials included. A total of 28,559 women from 28 trials were included in this review.Table

1. Meta‐analysis, subgrouping and sensitivity analysis results.

Comparison	Outcome	Analysis Type	Description	No. of Studies	MD	95% CI	Heterogeneity
Oestrogen (any dose) vs placebo or no RX	Weight	Overall meta‐analysis		9	0.03	‐0.61to 0.67	0.24
Oestrogen (any dose) vs placebo or no RX	Weight	Sensitivity analysis	method of randomization	5	‐0.04	‐0.71to 0.63	0.19
Oestrogen (any dose) vs placebo or no Rx	Weight	Sensitivity analysis	allocation concealment	5	‐0.04	‐0.71to 0.63	0.19
Oestrogen (any dose) vs placebo or no Rx	BMI	Overall meta‐analysis		2	‐0.14	‐0.40 to 0.12	0.26
Oestrogen (any dose) vs placebo or no Rx	BMI	Sensitivity analysis	method of randomization	2	‐0.14	‐0.40 to 0.12	0.26
Oestrogen (any dose) vs placebo or no Rx	BMI	Sensitivity analysis	allocation concealment	2	‐0.14	‐0.40 to 0.12	0.26
Oestrogen + progestogen (any dose) vs placebo or no Rx	Weight	Overall meta‐analysis		13	0.04	‐0.42 to 0.0.50	0.09
Oestrogen+progestogen (any dose) vs placebo or no Rx	Weight	Sensitivity analysis	Method of randomization	10	0.05	‐0.41 to 0.52	0.15
Oestrogen+progestogen (any dose) vs placebo or no Rx	Weight	Sensitivity analysis	Allocation concealment	10	0.05	‐0.41 to 0.52	0.15
Oestrogen+progestogen (any dose) vs placebo or no Rx	BMI	Overall meta‐analysis		10	‐0.10	‐0.27 to 0.07	0.57
Oestrogen+progestogen (any dose) vs placebo or no Rx	BMI	Sensitivity analysis	Method of randomisation	9	‐0.09	‐0.26 to 0.08	0.84
Oestrogen+progestogen (any dose) vs placebo or no Rx	BMI	Sensitivity analysis	Allocation concealment	8	‐0.09	‐0.26 to 0.08	0.84

Date	Event	Description
7 November 2008	Amended	Converted to new review format.
13 May 1999	New citation required and conclusions have changed	Substantive amendment

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	9	10194	Mean Difference (IV, Fixed, 95% CI)	0.03 [‐0.61, 0.67]
1.1 Subcategory	9	10194	Mean Difference (IV, Fixed, 95% CI)	0.03 [‐0.61, 0.67]
2 BMI	2	8933	Mean Difference (IV, Fixed, 95% CI)	‐0.14 [‐0.40, 0.12]
3 Waist to hip ratio	1	336	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.02, 0.02]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	13	18365	Mean Difference (IV, Fixed, 95% CI)	0.04 [‐0.42, 0.50]
2 BMI	10	17292	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.27, 0.07]
3 Waist to hip ratio	1	68	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.03, 0.03]
4 Skinfold thickness	1	22	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐66.46, 65.46]
5 Body fat mass	1	43	Mean Difference (IV, Fixed, 95% CI)	0.5 [‐3.51, 4.51]
6 Weight (kg) (nonevaluable data)			Other data	No numeric data
7 Body fat mass (nonevaluable data)			Other data	No numeric data

Weight (kg) (nonevaluable data)
Study
Aloia 1995	Pre‐trial control group mean weight 65.6 kg (SEM 1.8)   Pre‐trial treatment group mean weight 65.3 kg (1.6)   Post trial control group mean weight 65.9 kg   Post trial treatment group mean weight 67.7 kg   Body weight increase placebo group 0.1kg/year   Body weight increase treatment group 0.8 kg/year   Statistically significant increase

Body fat mass (nonevaluable data)
Study
Aloia 1995	Pre trial control group mean total fat mass 26.4 kg (SEM 2.4)   Pre trial treatment group mean total fat mass 23.9 kg (1.4   Post trial control group fat mass increase 2.9 kg/yr = 35.1 kg total fat at 3 years.   Post trial treatment group mean total fat mass 3.8 kg/yr = 35.3 total fat at 3 years.   Difference in fat increase between control and treatment group non significant

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	7	9367	Mean Difference (IV, Fixed, 95% CI)	‐0.23 [‐0.92, 0.45]
2 BMI	2	8933	Mean Difference (IV, Fixed, 95% CI)	‐0.14 [‐0.40, 0.12]
3 Wait to hip ratio	1	336	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.02, 0.02]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg)	10	17599	Mean Difference (IV, Fixed, 95% CI)	‐0.08 [‐0.55, 0.39]
2 BMI	10	17299	Mean Difference (IV, Fixed, 95% CI)	‐0.10 [‐0.27, 0.07]
3 Wait to hip ratio	1	68	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.03, 0.03]
4 Skinfold thickness	1	22	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐66.46, 65.46]
5 Weight kg (non evaluable data)			Other data	No numeric data
6 Fat mass (non evaluable data)			Other data	No numeric data

Methods	Randomisation achieved using computer‐generated random number table.   Number of centres: 1   Design: parallel, placebo‐controlled   Double‐blinded   No. of women randomised: 118   No. of women analysed: 94   No intention‐to‐treat analysis   Withdrawals and losses to follow up: n = 17. Reasons were stopping hormonal therapy (2), starting hormonal therapy (3), wished to withdraw (3), side effects of calcium augmentation (1), peptic ulcer disease (1), breast cancer (1), hyperthyroidism (1), moved out of region (2), claustrophobia (1), obesity preventing completion of neutron activation studies (1), development of rheumatoid arthritis (1). The woman with breast cancer was not in the HRT group. Seven participants are unaccounted for.   No power calculation stated.   Source of funding: Financial support from a National Institutes of Health Grant.
Participants	Postmenopausal. Natural menopause between 3 and 6 years prior to entry into study. Mean age of placebo group 53.4 yrs; calcium group 52.0 yrs, HRT group 51.2 yrs.   Location and country recruited: Recruited by community announcement from New York state, USA.   Ethnicity: Caucasian.   No biochemical confirmation of menopause status stated.   Inclusion criteria: Postmenopausal women between 6 months and 6 years after natural menopause   Exclusion criteria: any disorder known to affect bone metabolism e.g. glucocorticoid use, gastrointestinal disease, chronic illness; previous or current malignancy; undiagnosed vaginal bleeding; thrombophlebitis or thromboembolism; acute liver disease; gallbladder disease; history of liver disease; first‐degree relatives with breast cancer; hypertension; calcium urolithiasis; osteoporosis or vertebral compression fracture.   Groups stated comparable at baseline with age, years since menopause, activity and smoking.
Interventions	Rx1: conjugated equine oestrogen 0.625 mg daily + 10 mg medroxyprogesterone days 16‐25 + calcium carbonate 1700 mg daily + Vit D 400 IU daily   Rx2: calcium carbonate 1700 mg daily + Vit D 400 IU daily   Control: placebo + Vit D 400 IU daily   Oral administration.   Calcium and placebo preparations identical. Similarity of HRT not stated.   Duration of trial: 2.1 ± 1.1 (SD) years.
Outcomes	Relevant outcomes: Fat mass, Weight, BMI.   Primary outcomes of trial: to determine the efficacy of calcium supplementation in the prevention of bone loss; to examine the influence of HRT in prevention of the loss of lean body mass.
Notes	Side effects noted in the HRT treatment group were breast tenderness (2); decreased libido (1); menstrual cramps (1); eructation (1); constipation (1); headache (2); mood disorder (2); weight gain (2).   Authors contacted for additional information. Reply received. No evaluable data is available due to the way in which the analyses were performed (the effects were estimated by linear regression). The results indicated that menopause is associated with a gain in fat mass and a loss of lean body mass, but these changes in body composition are not prevented by HRT.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved using computer‐generated random number table.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double‐blinded
Incomplete outcome data addressed?   All outcomes	High risk	Withdrawals and losses to follow up: n = 17. Reasons were stopping hormonal therapy (2), starting hormonal therapy (3), wished to withdraw (3), side effects of calcium augmentation (1), peptic ulcer disease (1), breast cancer (1), hyperthyroidism (1), moved out of region (2), claustrophobia (1), obesity preventing completion of neutron activation studies (1), development of rheumatoid arthritis (1). The woman with breast cancer was not in the HRT group. Seven participants are unaccounted for.
Free of other bias?	Unclear risk	No power calculation stated.   Source of funding: Financial support from a National Institutes of Health Grant.

Methods	Block randomisation achieved using computer‐generated random number table.   Number of centres: 1   Design: parallel, placebo‐controlled,   partly double‐blinded (n = 50) and partly open (n = 66).   No. of women randomised: 116   No. of women analysed: 108   Intention‐to‐treat analysis   Withdrawals and losses to follow up: non‐attenders for measurement n = 8 (4 from E+P group and 4 from placebo group). Reasons for non‐attendance not stated.   Power calculation performed   Source of funding: Financial support from the Wishbone Trust and Special Trustees for the Nottingham Hospitals. Drugs provided free by Sandoz Pharmaceuticals and Wyeth.
Participants	Postmenopausal. Mean age of placebo group 61.3 ± 5.8 yrs; treatment groups 60.5 ± 6.3 yrs. Natural and surgical menopause.   Location and country recruited: Fracture clinics of large city hospital in Nottingham, UK.   Ethnicity: Caucasian.   Menopausal status confirmed by targeted history, examination and biochemistry. Gynaecological screen performed to ensure no contraindication to HRT; FSH and LH levels examined in women without a uterus to ensure postmenopausal status.   Inclusion criteria: Postmenopausal women who had suffered a wrist fracture within the previous 7 weeks prior to study entry.   Exclusion criteria: Overt neurological or neuromuscular condition that could impair strength, balance or mobility; treatment with drugs affecting balance.   Groups stated comparable at baseline with age and years since menopause. The placebo group was significantly heavier than the treatment groups. The treatment groups took significantly more exercise initially than the placebo group. Baseline equality of diet and smoking status not stated.
Interventions	Rx1: conjugated equine oestrogen 0.625 mg daily (Premarin) + 1000 mg/day elemental calcium.   Rx2: conjugated equine oestrogen 0.625 mg + 150 mcg norgestrel given in sequential fashion + 1000 mg/day elemental calcium.   Control: 1000 mg/day elemental calcium.   Oral administration.   HRT and placebo preparations identical.   Duration of trial: 12 months.
Outcomes	Relevant outcome: Weight   Primary outcomes of trial: to assess the effect of HRT plus calcium compared to calcium alone on balance, muscle performance and falls.
Notes	Authors contacted for additional information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Block randomisation achieved using computer‐generated random number table.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Parallel, placebo‐controlled,   partly double‐blinded (n = 50) and partly open (n = 66).
Incomplete outcome data addressed?   All outcomes	High risk	Withdrawals and losses to follow up: non‐attenders for measurement n = 8 (4 from E+P group and 4 from placebo group). Reasons for non‐attendance not stated.
Free of other bias?	High risk	Source of funding: Financial support from the Wishbone Trust and Special Trustees for the Nottingham Hospitals. Drugs provided free by Sandoz Pharmaceuticals and Wyeth.

Methods	Variable length, blocked randomisation, in which treatment assignment was stratified by clinical centre and hysterectomy status. Randomization was achieved using an on site computer system.   Number of centres: 7.   Design: parallel, placebo‐controlled.   Blinding: Triple‐blinded (women, investigators, laboratory staff).   N. of women randomised: 875 overall; 349 women randomised to groups currently included in this review.   No. of women analysed: 847 overall; 336 from groups currently included in this review.   Intention‐to‐treat analysis stated, but number analysed less than number randomised.   No withdrawals and losses to follow up stated.   Power calculation performed.   Source of funding: Financial support provided by Cooperative Agreement Research Grants, the Diabetes Research and Training Center, Indiana University School of Medicine.
Participants	Postmenopausal. From total women randomised, mean age was 56.1 ± 4.3 years. Of these women, 41% were aged 45‐54 and 59% were aged 55‐65. Natural and surgical menopause.   Location and country recruited: California, Texas, Iowa, Maryland and Washington DC, USA.   Ethnicity: In groups currently analysed:   13 African American   17 Hispanics   306 Caucasians/Asians/   Native Americans.   Menopausal status confirmed by cessation of menstruation > 1 year prior to randomisation; hysterectomy > 2 months prior to randomisation. Women with hysterectomy FSH levels equal or greater than 40 mIU/ml.   Inclusion criteria: Women of all races 45‐64 years with or without uterus and irrespective of prior hormone use.   Exclusion criteria: Natural menopause before age 44 yr, within 12 months of enrolment or > 10 years prior; hysterectomy within 2 months; BMI of 40 kg/m2 or more; medical history that might contraindicate use of HRT or curtail follow up.   Groups stated comparable at baseline for age, years since menopause, weight, smoking and physical activity. Baseline equality of diet not stated.
Interventions	Rx: conjugated equine oestrogen 0.625 mg daily (Premarin).   Control: placebo   Oral administration   HRT and placebo preparations were identical.   Duration of trial: 36 months.
Outcomes	Relevant outcomes: Body weight, Waist/hip ratio, Body Mass Index.   Primary outcomes of trial: Body weight, Waist/hip ratio, Body Mass Index, Bone density, Quality of life, Endometrial changes, Menopausal symptoms, Lipid metabolism, Blood pressure, Carbohydrate metabolism, Haemostasis.
Notes	Authors contacted for additional information. Reply received.   This trial compared 5 regimes:   Rx1: conjugated equine oestrogen 0.625 mg daily (Premarin).   Rx2: conjugated equine oestrogen 0.625 mg daily + medroxyprogesterone 10 mg daily days 1‐12 (Provera).   Rx3: conjugated equine oestrogen 0.625 mg daily + medroxyprogesterone 2.5 mg daily.   Rx4: conjugated equine oestrogen 0.625 mg daily + micronized progesterone 200 mg daily days 1‐12.   Rx5: Placebo.   Further information is required before Rx2, 3 and 4 can be analysed. This information has been requested.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Variable length, blocked randomisation, in which treatment assignment was stratified by clinical centre and hysterectomy status. Randomization was achieved using an on site computer system.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Triple‐blinded (women, investigators, laboratory staff).
Incomplete outcome data addressed?   All outcomes	Low risk	No losses
Free of other bias?	High risk	Source of funding: Financial support provided by Cooperative Agreement Research Grants, the Diabetes Research and Training Center, Indiana University School of Medicine.

Methods	Block randomisation; method of randomisation not stated.   Number of centres: single centre.   Design: parallel, placebo‐controlled.   Blinding: double‐blind for Rx1; single‐blind for Rx2.   No. of women randomised: 81 overall, 60 women randomised to groups included in this review.   No. of patients analysed: 65 overall, 45 from groups included in this review.   No intention‐to‐treat analysis.   Withdrawals overall n = 16. Withdrawals from groups analysed: n = 15 (4 from oestrogen only group ‐ 1 uncontrolled hot flashes, 3 too busy to attend; 5 from oestrogen/progesterone group ‐ 1 uncontrolled hot flashes, 2 irregular bleeding, 2 too busy to attend; 6 from placebo group ‐ 2 uncontrolled hot flashes, 3 too busy to attend, 1 breast carcinoma). No losses to follow up.   No power calculation reported.   Source of funding: Financial support provided by National Institutes of Health and Upjohn Pharmaceutical Company.
Participants	Postmenopausal (treatment group oestrogen only 53 ± 4.7 years, 3.6 ± 2.9 years since LMP; treatment group oestrogen/progesterone 53 ± 3.5 years, 4.6 ± 2.3 years since LMP; control group 50 ± 4.4 years, 3.1 ± 1.7 years since LMP).   Approximately half in each group had surgical menopause.   Location and country recruited: Population sample in the USA.   Ethnicity not stated.   Menopausal status confirmed by: Not stated. No biochemical investigation reported.   Inclusion criteria: Within 6 years of menopause; no disease or drug known to cause abnormalities of calcium metabolism; good general health; no chronic diseases.   Exclusion criteria: None stated.   Groups comparable at baseline for age, years since menopause and weight. Baseline equality of smoking, diet and physical activity not stated.
Interventions	Rx1: conjugated estrogens 0.6 mg daily (Premarin) + calcium 600‐1000 mg daily if necessary.   Rx2: conjugated estrogens 0.3 mg (Premarin) + medroxyprogesterone 10 mg (Provera) daily + calcium carbonate 600‐1000 mg daily if necessary.   Control: Placebo + calcium carbonate 600‐1000 mg daily if necessary.   Oral administration.   Rx1 and placebo identical; Rx2 and placebo not identical.   Duration of trial: 24 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: Bone mineral density.
Notes	Author contacted for additional information. Replied received.   One treatment arm, Medroxyprogesterone 20 mg (Provera) daily + calcium 600‐1000 mg daily if necessary, excluded from this review. Subjects' diets were supplemented with calcium carbonate 600 mg/day if necessary to achieve a final calcium intake of approximately 1000 mg/day. 95% of participants received some degree of supplementation.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding?   All outcomes	Low risk	Double

Methods	Randomisation achieved using an allocation procedure transparent before assignment.   Number of centres: Multicentre (number not stated).   Design: parallel, placebo‐controlled.   Blinding: double‐blinded.   No. of women randomised: 273.   No. of women analysed: 273   Intention‐to‐treat analysis.   Withdrawals n = 48 (25 from treatment groups including 8 systemic adverse experiences, 1 ovarian cyst, 1 breast enlargement); 23 from placebo group (including 3 systemic adverse experiences, 8 lack of efficacy). Reasons for remaining withdrawals not stated. No losses to follow up.   No power calculation reported.   Source of funding: TheraTech, Inc.
Participants	Postmenopausal (treatment group 50.8 ± 0.7 years, 27 <3 yrs since menopause, 61 equal or greater than 3 yrs since menopause; control group 50.7 ± 0.8 years, 27 <3 yrs since menopause, 64 equal or greater than 3 yrs since menopause). Includes natural and surgical menopause.   Location and country recruited: Population sample in United States (actual locations not stated).   Ethnicity: Treatment groups white n=158, black n = 19; other n=5.   Control group white n=81; black n=9; other n=1.   Menopausal status confirmed by: FSH equal or greater than 40 mIU/mL   serum estradiol concentration equal or greater than 20 pg/mL   Inclusion criteria: Age 21+ yrs if surgically menopausal or 45+ yrs if naturally menopausal; amenorrhoeic for 6+ months; experiencing 60+ moderate or severe flushes weekly; abstention from previous HRT for at least 6 weeks before initial screening visit.   Exclusion criteria: Use of implanted or injected hormones within previous 12 months; prior major dermatologic disease or active skin disorder; history of allergies to estradiol or transdermal medications; history of estrogen‐dependent neoplasia; history of cerebrovascular disease, thrombotic disorders or angina; history of cholecystectomy or active gall bladder disease within previous 6 months; psychological or psychiatric illness at time of entry; undiagnosed or abnormal vaginal/uterine bleeding; within previous 6 months in women with intact uterus; clinically significant abnormalities in medical history, physical examination, or mammography; use of any investigational drug within previous 60 days; evidence of pre‐malignant or malignant breast changes detected by mammography in previous 12 months; serious hepatic, renal or cardiac disease; endocrine disease; any malignancy or substance abuse within 12 months of study; + additional conditions detected at initial screening visit.   Groups comparable at baseline for age, body weight, age at menopause, time since menopause and type of menopause. Baseline equality of smoking, diet and physical activity not stated.
Interventions	Rx1: 50 µg estradiol daily.   Rx2: 100 µg estradiol daily.   Control: Placebo.   Mode of administration: patch   Intervention and placebo identical.   Duration of trial: 3 months
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: Frequency and severity of hot flushes, changes in serum concentrations of estradiol, estrone, estrone sulphate and FSH, improvements in vaginal cytology.
Notes	First author contacted for additional information. Reply received.   This study did not entirely conform to the review protocol, in that treatment with any previous oral HRT terminated 8 weeks before the study rather than 12 weeks.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomisation achieved using an allocation procedure transparent before assignment.
Allocation concealment?	High risk	C ‐ Inadequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	Low risk	Intention‐to‐treat analysis.   Withdrawals n = 48 (25 from treatment groups including 8 systemic adverse experiences, 1 ovarian cyst, 1 breast enlargement); 23 from placebo group (including 3 systemic adverse experiences, 8 lack of efficacy). Reasons for remaining withdrawals not stated. No losses to follow up.
Free of other bias?	Unclear risk	Source of funding: TheraTech, Inc.

Methods	Method of randomisation not stated.   Number of centres: single centre.   Design: parallel, placebo‐controlled design with three groups.   Blinding: double‐blind.   No of women randomised: 75   No. of women analysed: 62.   No intention‐to‐treat analysis.   Withdrawals: 13 (number of withdrawal per group as well as the reasons of withdrawal are not stated).   No power calculation reported.   Funding: Statens Sundhedsvidenskabelige Forrk‐ningsrd, Copenhagen, Denmark.
Participants	Postmenopal women aged 45 to 55 years with 6 month amenorrhoea. Women were 22.4 +/‐ 9.9 months of menopausal age in the estradiol valerate + levonorgestrel group, 20.9 +/‐ 8.4 months in the estradiol valerate + cyproterone acetate group and 17.6 +/‐ 8.6 months in the placebo group.   Includes only women with natural menopause.   Location and country recruited: Copenhagen, Denmark.   Ethnicity: Not stated (Danish menopausal women).   Menopausal status not confirmed by biochemical investigation.   Inclusion criteria: Healthy Danish postmenopausal women aged 45 to 55 years, who had passed natural menopause 6 months to 3 years previously.   Exclusion criteria: diseases and medications known to influence body composition.   Groups similar in weight and menopausal age at baseline.   Baseline age, smoking, diet and physical activity not reported.
Interventions	Rx1: 2 mg estradiol valerate daily combined sequentially with 75 uG levonorgestrel.   Rx2: 2 mg estradiol valerate combined continuously with 1 mg cyproterone acetate.   Control: placebo   Mode of administration: oral   Similarity of treatment and placebo no reported.   Duration: 2 years
Outcomes	Relevant outcome: Fat mass.   Primary outcomes of the trial: abdominal fat, bone loss, total body fat mass, total lean body mass.
Notes	First author contacted and reply awaited for additional information.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Not stated
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding?   All outcomes	Low risk	Double blinded
Incomplete outcome data addressed?   All outcomes	High risk	Withdrawals: 13 (number of withdrawal per group as well as the reasons of withdrawal are not stated).
Free of other bias?	Unclear risk	Funding: Statens Sundhedsvidenskabelige Forrk‐ningsrd, Copenhagen, Denmark.

Methods	Randomisation achieved by providing participant details by telephone to receive treatment group allocation.   Number of centres: single centre.   Parallel design, treatment versus no treatment, single blinded.   No. of women randomised: 23   No. of women analysed: 22   No intention to treat.   Withdrawals and losses to follow up: Withdrawals n = 1 (from treatment group ‐ severe bleeding).   No power calculation reported.   Source of funding: No source of funding reported.
Participants	Postmenopausal (mean age = 51.8 years). Natural menopause.   Location and country recruited: University department, Vienna, Austria.   Ethnicity not reported.   Menopausal status confirmed by: Determination of E2 FSH and LH levels.   Inclusion criteria: Spontaneous menopause >1 year prior to randomisation, healthy, no HRT up to time of first examination.   Exclusion criteria: No specific exclusion criteria reported.   Groups compatible at baseline for age, months since menopause, weight, BMI and skinfold measurements. Baseline equality of smoking, diet and physical activity not stated
Interventions	Treatment: estradiol valerate 2 mg daily + dydrogesterone 10 mg 10 days per month.   Control: No treatment.   Oral administration.   Blinding in relation to treatment: Not applicable.   Duration of trial: 10 months.
Outcomes	Relevant outcome: Body weight, Abdominal skin fold, BMI.   Primary outcomes of trial: Secretion of Growth hormone (GH) and IGF 1, Somatometric‐GH correlation patterns; stature, chest girth, waist girth, abdominal girth, body weight, skinfold triceps, skinfold subscapular and skinfold abdominal.
Notes	Author contacted for additional information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved by providing participant details by telephone to receive treatment group allocation.
Allocation concealment?	Low risk	A ‐ Adequate
Incomplete outcome data addressed?   All outcomes	High risk	Withdrawals and losses to follow up: Withdrawals n = 1 (from treatment group ‐ severe bleeding).

Methods	Randomisation achieved using assignment envelopes, sequentially numbered, sealed and opaque.   Number of centres: Single centre.   Design: parallel, placebo‐controlled.   Method of blinding not stated.   No. of women randomised: 151.   No. of women analysed: 136.   No intention‐to treat analysis.   Withdrawals n = 15 (7 from medium dose groups, 7 from high‐dose group, 1 from placebo group). Reasons were breast tenderness, vaginal bleeding, nervousness or nausea n = 6, moving away from the area n = 3, illness n = 3, not specified n = 3.   No power calculation reported.   Source of funding not stated.
Participants	Post menopausal (treatment group 52.4 (46‐56) years, 46 (29‐62) months since menopause; control group 50.7 (46‐55) years, 45 (29‐62) months since menopause).   Natural menopause.   Location and country recruited: Population sample from Copenhagen, Denmark.   Ethnicity not stated.   Menopausal status confirmed by: Spontaneous cessation of bleeding 6 months ‐ 5 years prior to study. No biochemical investigation reported.   Inclusion criteria: Healthy, natural menopause, no treatment with exogenous hormones.   Exclusion criteria: Medication known to influence the variables under study.   Groups comparable at baseline for age, years since menopause, and weight. Baseline equality of smoking, diet and physical activity not stated.
Interventions	Treatment: cyclic percutaneous estradiol (3 mg daily from day 1 to day 24) in the first year; during the second year oral micronized progesterone (200 mg) was added to the estradiol treatment for 12 days (days 13 to 24).   Control: placebo (mode of administration not stated).   Similarity of HRT and placebo not reported.   Duration: 1 year of estradiol alone vs placebo, followed by another year of estradiol + micronized progesterone vs placebo.
Outcomes	Relevant outcome: body weight.   Pimary outcome of the trial: effect on lipid profile (total serum cholesterol, low‐density lipoprotein cholesterol, high‐density lipoprotein cholesterol and serum triglycerides).
Notes	Author contacted for additional and information pending.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved using assignment envelopes, sequentially numbered, sealed and opaque.
Allocation concealment?	Low risk	A ‐ Adequate
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to treat analysis.   Withdrawals n = 15 (7 from medium dose groups, 7 from high‐dose group, 1 from placebo group). Reasons were breast tenderness, vaginal bleeding, nervousness or nausea n = 6, moving away from the area n = 3, illness n = 3, not specified n = 3.

Methods	Randomisation achieved by assignment envelopes, sequentially numbered, sealed and opaque.   Number of Centres: Single centre.   Design: Parallel design, placebo‐controlled.   Blinding: Double blinded.   No. of patients randomised: 76.   No. of patients analysed: 67.   No intention to treat analysis.   Withdrawals n = 9 (4 from placebo group ‐ 3 unacceptable climacteric symptoms, 1 lack of time; 5 from treatment group ‐ 2 lack of time, 1 gain in weight, 1 varicose veins, 1 moved away from the area).   No power calculation reported.   No source of funding reported.
Participants	Menopausal status: Postmenopausal aged 45‐54 years. Natural menopause.   Location and country recruited: Population sample from Copenhagen, Denmark.   Ethnicity not reported.   Menopausal status confirmed by: Spontaneous cessation of bleeding within previous 6 months ‐ 3 years. No biochemical investigation reported.   Inclusion criteria: Freedom from past and present diseases and medications known to influence calcium metabolism.   Exclusion criteria: None reported.   Groups compatible at baseline for age, months since menopause, weight. Baseline equality of smoking, diet and physical activity not reported.
Interventions	Treatment: Estradial valerate 2mg days 1‐21 + cyproterone acetate 1 mg days 12‐21.   Control: Placebo.   Oral administration.   Similarity of treatment and placebo not reported.   Duration of trial: 12 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: Effect on lipoprotein profiles of cyproterone acetate, when included as a progestational agent in hormone replacement therapy.
Notes	Author contacted for additional information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved by assignment envelopes, sequentially numbered, sealed and opaque.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No intention to treat analysis.   Withdrawals n = 9 (4 from placebo group ‐ 3 unacceptable climacteric symptoms, 1 lack of time; 5 from treatment group ‐ 2 lack of time, 1 gain in weight, 1 varicose veins, 1 moved away from the area).
Free of other bias?	Unclear risk	No source of funding declared

Methods	Randomisation achieved by pharmacy controlled allocation scheme.   Number of centres: single centre.   Design: crossover, placebo‐controlled.   Blinding: Triple‐blinded (women, investigators, trial statistician).   No. of women randomised: 105.   No. of women analysed: 93 at crossover.   No intention‐to‐treat analysis.   Withdrawals n = 12 patients had withdrawn at crossover ‐ this included 3 because they had felt unwell after changing medication and 1 (placebo group) because of diagnosed colonic cancer. The remaining 8 withdrawals were not accounted for at crossover, but included with the total withdrawals at the conclusion of the complete trial.   Power calculation performed.   Source of funding not stated.
Participants	Perimenopausal, aged 46.1 ± 2.9 years.   Location and country recruited: Population sample in Australia.   Ethnicity not stated.   Considered to be in menopause transition because subjects were 40‐52 years; having menstrual bleeding, albeit irregular and infrequent; experiencing minor symptoms associated with menopause. No biochemical investigations to confirm menopausal status were carried out.   Inclusion criteria: No history of hormonal treatment, including contraceptive pill, for at least one month; no evidence of malignancy; adequate protection against pregnancy; sufficient intelligence and motivation to complete the questionnaire, lived close enough to comply with the protocol; able to give informed consent.   Exclusion criteria: History of thromboembolism; chronic renal or hepatic impairment; psychiatric disorder; psychotropic drug treatment; smoking, hypersensitivity to oestrogen; focal headaches; gallbladder disease.   Groups comparable at baseline for age and smoking. The placebo group had significantly higher blood pressure and were also heavier than the treatment group (mean excess 2.6 kg, P = 0.08). Baseline equality of diet and physical activity not stated.
Interventions	Treatment: Conjugated oestrogens 0.625 mg daily (Premarin) + medroxyprogesterone acetate 10 mg daily (Provera) days 14‐27.   Control: placebo.   Oral administration.   Similarity of HRT and placebo not reported.   Duration of trial: 12 months overall; crossover at 6 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: Body weight, Blood pressure, Lipoprotein levels, Antithrombin III activity, Endometrium changes.
Notes	Author contacted for additional information. Reply received.   Most women had ceased hormone replacement therapy > 6 months before randomisation. No more than 3 subjects were taking HRT 1 month prior to randomisation, although to comply with the review protocol, those taking oral HRT should have ceased therapy > 3 months before randomisation. It is not possible to ascertain from the available data what group these 3 subjects were part of, whether they were amongst the withdrawals, or in what form the HRT was taken.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved by pharmacy controlled allocation scheme.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Triple
Incomplete outcome data addressed?   All outcomes	Unclear risk	Withdrawals n = 12 patients had withdrawn at crossover ‐ this included 3 because they had felt unwell after changing medication and 1 (placebo group) because of diagnosed colonic cancer. The remaining 8 withdrawals were not accounted for at crossover, but included with the total withdrawals at the conclusion of the complete trial.
Free of other bias?	Unclear risk	Source of funding not stated.

Methods	Block randomisation achieved by a computer‐generated random number table in blocks of 4, 8 or 12.   Number of centres: Single centre.   Design: parallel, placebo‐controlled.   Blinding: research personnel blinded to group allocation. It is not stated whether the women were blinded to their treatment, but the inference is that they were not.   No. of women randomised: 464.   No. of women analysed: 388.   No intention‐to‐treat analysis.   Withdrawals n = 73, excluded n = 3. HRT dropouts n = 55 (commonly due to hypermenorrhoea, dysmenorrhoea or uterine bleeding other than that caused by menstruation); placebo dropouts n = 21, reasons not stated.   No power calculation reported.   Financial support provided by Leiras Ltd and Schering Ltd.
Participants	Recently postmenopausal (treatment group mean age 52.9 years, mean years since LMP 1.2 years; placebo group mean age 52.6 years, mean years since LMP 1.1 years).   Location and country recruited: Population sample in Kuopio Province, Finland.   Ethnicity not stated.   Menopausal status confirmed by: 6‐24 months since last menstruation. Levels of LH and FSH.   Inclusion criteria: Non osteoporotic.   Exclusion criteria: History of breast or endometrial cancer, thromboembolic diseases and medication‐resistant hypertension.
Interventions	Rx1: estradiol valerate 2 mg days 1‐21 + cyproterone acetate 1mg days 12‐21.   Rx2: vitamin D3 (cholecalciferol 300 IU) + 93 mg Ca2+/day; no intake during June‐August.   Rx3: Rx1 and Rx2 combined.   Rx4: calcium lactate, 500 mg/day (equivalent to 93 mg Ca2+/day).   Control: calcium lactate 500 mg/day.   Oral administration.   Similarity of treatment and placebo not reported.   Duration of trial: 30 months.
Outcomes	Relevant outcome: Body weight, BMI.   Primary outcomes of trial: effect of HRT and low‐dose Vit D supplementation on bone mineral density.
Notes	Author contacted for additional information. Reply received.   Rx1 and Rx3 results combined; Rx2 and Rx4 results combined for purposes of this analysis.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Block randomisation achieved by a computer‐generated random number table in blocks of 4, 8 or 12.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Unclear risk	Blinding: research personnel blinded to group allocation. It is not stated whether the women were blinded to their treatment, but the inference is that they were not.
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to‐treat analysis.   Withdrawals n = 73, excluded n = 3. HRT dropouts n = 55 (commonly due to hypermenorrhoea, dysmenorrhoea or uterine bleeding other than that caused by menstruation); placebo dropouts n = 21, reasons not stated.
Free of other bias?	High risk	Financial support provided by Leiras Ltd and Schering Ltd.

Methods	Randomisation achieved by alternation.   Number of centres: single centre.   Design: parallel design, treatment versus no therapy.   Blinding: No blinding.   No. of women randomised: 120.   No. of women analysed: 93.   No intention‐to‐treat analysis.   Losses to follow up n = 27 (9 in treatment group; 18 in no treatment group).   No power calculation reported.   Source of funding: Rachadapiseksomphot Fund, Faculty of Medicine, Chulalongkorn University.
Participants	Postmenopausal 6 months ‐ 10 years (treatment groups 48.67 ± 5.23 years; no treatment group 48.26 ± 4.24 years). Treatment groups 32 natural and 19 surgical menopause; no treatment group 25 natural and 17 surgical menopause.   Location and country recruited: Menopause clinic, Chulalongkorn University Hospital, Bangkok.   Ethnicity not reported.   Menopausal status confirmed by: Measurement of serum gonadotropin and estradiol levels.   Inclusion criteria: Healthy, no hormone treatment within previous one year.   Exclusion criteria: Freedom from drugs known to affect lipid metabolism; evidence of any endocrinologic disorders, chronic diseases or laboratory abnormalities that could interfere with interpretation of the result of treatment.   Groups compatible at baseline for age and time since menopause. Baseline equality for diet and physical activity not reported. No difference between groups for smoking status. Groups not compatible for weight at baseline at P = 0.05 significance level
Interventions	Rx1, Natural menopause: either cyclic estradiol valerate 2 mg + norgestrel 0.5 mg, cyclic conjugated equine oestrogen 0.625 mg + medrogestone 5 mg, or continuous conjugated equine oestrogen 0.625 mg + MPA 2.5mg (numbers taking each regimen not reported).   Rx2, Surgical menopause: either cyclic conjugated equine oestrogen 0.625 mg or cyclic 17 beta‐estradiol 1.5 mg (numbers taking each regimen not reported).   Control: calcium 1,000 mg/day + parasympatholytics as needed.   Oral administration for all groups except for 1.5 mg 17 beta‐estradiol given to some in the surgical menopause group as gel.   Blinding in relation to treatment: not applicable.   Duration of trial: 12 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: effect on lipid profiles in postmenopausal women with or without hormone replacement therapy.
Notes	Author contacted for further information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	High risk	Randomisation achieved by alternation.
Allocation concealment?	High risk	C ‐ Inadequate
Blinding?   All outcomes	High risk	Unblinded
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to‐treat analysis.   Losses to follow up n = 27 (9 in treatment group; 18 in no treatment group).
Free of other bias?	Unclear risk	Source of funding: Rachadapiseksomphot Fund, Faculty of Medicine, Chulalongkorn University.

Methods	Randomisation method not stated.   Number of centres: 1.   Parallel design with no treatment control group.   No blinding of primary outcome assessment to either researchers or subjects.   No. of women randomised: 101 overall, 48 in groups analysed. No. of women analysed: 101 overall, 48 in groups analysed. Intention‐to‐treat analysis.   Withdrawals and losses to follow up: 95 women over all treatment groups completed the study. It is not stated what groups those who did not complete the study belonged to, or their reasons for non‐completion.   No power calculation reported.   Funding supported by a grant from Wyeth‐Ayerst laboratories.
Participants	Postmenopausal women aged 42‐59 years. Whether they were naturally or surgically menopausal is not stated, although subjects taking oestrogen were given progestins at the completion of the study, which would seem to indicate that the women were naturally postmenopausal.   Location and country recruited: Climacteric centre of university community in North‐Central Florida, USA.   Ethnicity not stated.   Menopausal status confirmed by absence of bleeding for at least 1 year; elevated serum FSH (>40 mIU/mL).   Inclusion criteria: Evidence of postmenopausal status; absence of significant systemic diseases; no previous hormone replacement therapy.   Exclusion criteria: Presence of cardiovascular or pulmonary disease; inability to participate in the exercise program.   Baseline equality for age, smoking, diet, physical activity not reported.
Interventions	Treatment: conjugated oestrogen 0.625 mg days 1‐25   Control: no treatment   Oral administration. No blinding in relation to treatment   Duration of trial: 6 months.
Outcomes	Relevant outcome: Body weight   Primary outcomes of trial: the effects of a moderate exercise program on levels of lipids and lipoproteins in postmenopausal women.
Notes	This trial compared 4 regimes:   Rx1: conjugated oestrogen 0.625 mg days 1‐25   Rx2: conjugated oestrogen 0.625 mg days 1‐25 plus exercise 3 times per week for 30 minutes   Rx3: exercise 3 times per week for 30 minutes   Rx4: no treatment or exercise.   Rx1 and Rx4 have been included in the analysis, pending further information concerning Rx2 and Rx3.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomisation method not stated.
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding?   All outcomes	High risk	No blinding of primary outcome assessment to either researchers or subjects.
Incomplete outcome data addressed?   All outcomes	Low risk	No. of women randomised: 101 overall, 48 in groups analysed. No. of women analysed: 101 overall, 48 in groups analysed. Intention‐to‐treat analysis.
Free of other bias?	High risk	Funding supported by a grant from Wyeth‐Ayerst laboratories.

PERMALINK

Oestrogen and progestogen hormone replacement therapy for peri‐menopausal and post‐menopausal women: weight and body fat distribution

Eugene J Kongnyuy

Robert J Norman

Ingrid HK Flight

Margaret C Rees

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Measures of treatment effect

Assessment of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1. Meta‐analysis, subgrouping and sensitivity analysis results.

Included studies

Excluded studies

Risk of bias in included studies

1.

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Discussion

Summary of main results

Overall completeness and applicability of evidence

Quality of the evidence

Potential biases in the review process

Agreements and disagreements with other studies or reviews

Authors' conclusions

Implications for practice.

Implications for research.

What's new

History

Notes

Acknowledgements

Appendices

Appendix 1. Search strategy

Data and analyses

Comparison 1. Oestrogen (any dose) versus placebo or no treatment.

1.1. Analysis.

1.2. Analysis.

1.3. Analysis.

Comparison 2. Oestrogen plus progestogen (any dose) versus placebo or no treatment.

2.1. Analysis.

2.2. Analysis.

2.3. Analysis.

2.4. Analysis.

2.5. Analysis.

2.6. Analysis.

Methods	Openly assigned to treatment or control group utilising block randomisation achieved by a computer‐generated random number table in blocks of four. Subsequent randomisation to treatment arms achieved in a double‐blind fashion.   Number of centres: Single centre.   Design: parallel design, no treatment control.   Blinding: two treatment arms were allocated in a double‐blind fashion.   No. of women randomised: 30.   No. of women analysed: 27.   Intention‐to‐treat analysis, although n = 3 were excluded from the final analysis due to lack of material at baseline as well as during the trial.   Withdrawals n = 4 (2 from treatment group ‐ 1 lack of compliance; 1 unacceptable vaginal bleeding; 2 from no treatment group ‐ 1 lack of motivation to participate, 1 HRT use for climacteric complaints).   No power calculation reported.   Funding provided by Solvay Duphar BV.
Participants	Postmenopausal 47‐59 years. Natural menopause 6 months ‐ 18 years.   Location and country recruited: Outpatient clinic of department of obstetrics and gynaecology, University Hospital, Amsterdam, The Netherlands.   Ethnicity not reported.   Menopausal status confirmed by: Serum FSH concentrations equal or greater than 21 IU/L.   Inclusion criteria: Good health, no history of renal, hepatic, endocrinologic, gastrointestinal, or cardiovascular disease   Exclusion criteria: Previous oestrogen, progestogen, or androgen therapy prior to randomization; regular consumption of B6, B12 or folate supplements; previous gonadal steroid pellet implants.   Groups compatible at baseline for age, BMI and smoking. Baseline equality of diet and physical activity not reported.
Interventions	Rx1: micronized E2 1 mg daily + dydrogesterone 5 mg days 15‐28.   Rx2: micronized E2 1 mg daily + dydrogesterone 10 mg days 15‐28   Both treatment arms changed to micronized E2 2 mg daily + dydrogesterone 10 mg days 15‐28 months 12‐15.   (The treatments were combined for analysis by the researchers as there were no significant differences found in the measured indices).   Control: no treatment.   Oral administration.   Blinding in relation to treatment: Not applicable, although containers for two treatment arms were identical.   Duration of trial: 15 months.
Outcomes	Relevant outcome: BMI.   Primary outcomes of trial: to examine the effects of oral E2 combined sequentially with dydrogesterone on fasting plasma homocysteine concentrations
Notes	Author contacted for additional information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Openly assigned to treatment or control group utilising block randomisation achieved by a computer‐generated random number table in blocks of four. Subsequent randomisation to treatment arms achieved in a double‐blind fashion.
Allocation concealment?	High risk	C ‐ Inadequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	Unclear risk	Intention‐to‐treat analysis, although n = 3 were excluded from the final analysis due to lack of material at baseline as well as during the trial.
Free of other bias?	High risk	Funding provided by Solvay Duphar BV.

Methods	Method of randomisation not stated.   Single centre study.   Design: parallel, placebo‐controlled.   Blinding: double‐blind.   Women randomised overall: n = 113.   Women randomised to groups included in this review: not stated.   No intention to treat analysis.   Number of withdrawals: n = 39 overall (17 due to vaginal bleeding, 7 due to intercurrent disease, 15 for other reasons).   Number of withdrawals from groups included in this review: not stated.   No of woman analysed: 74 overall; 35 women from groups included in this review.   No power calculation reported.   Source of funding: Velux Fonden, Velux International A/S, Denmark.
Participants	Postmenopausal women aged 70 years. Whether natural or surgical menopause not stated.   Location and country recruited: A district in Denmark with a population representative of the Danish population.   Ethnicity not stated.   Menopausal status confirmed: Not reported.   Inclusion criteria: Healthy (criteria not stated).   Exclusion criteria: None stated.   Groups comparable at baseline for parity, time after menopause, BMI, smoking habits and serum total cholesterol. Baseline equality of diet and physical activity not stated.
Interventions	Treatment: estrogen‐norethindrone combination (2 mg estradiol + 1 mg estriol days1‐12; 2 mg estradiol + 1mg estriol + 1 mg norethindrone days 13‐22; 1 mg estradiol + 0.5mg estriol days 23‐28).   Control: placebo   Oral administration.   Similarity of treatment and control treatment not reported.   Duration of trial: 12 months.
Outcomes	Relevant outcome: BMI.   Primary outcomes of trial: BMI, Total cholesterol level, High‐density lipoprotein cholesterol level, Blood pressure.
Notes	Author contacted for additional information. Letter returned 'unknown'.   This trial compared four regimes:   Rx1: Placebo   Rx2: 1,25‐dihydroxycholecalciferol 0.5 µg/daily   Rx3: Estrogen‐norethindrone combination (2mg estradiol + 1mg estriol days1‐12; 2mg estradiol + 1 mg estriol + 1 mg norethindrone days 13‐22; 1 mg estradiol + 0.5 mg estriol days 23‐28)   Rx4: 1,25‐dihydroxycholecalciferol 0.5 µg/daily + estrogen‐norethindrone combination (2mg estradiol + 1mg estriol days1‐12; 2mg estradiol + 1mg estriol + 1 mg norethindrone days 13‐22; 1 mg estradiol + 0.5 mg estriol days 23‐28).   Rx1 and Rx3 have been entered into the meta‐analysis pending a further attempt to obtain combined data from the first author.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Method of randomisation not stated.
Allocation concealment?	Unclear risk	B ‐ Unclear
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No intention to treat analysis.   Number of withdrawals: n = 39 overall (17 due to vaginal bleeding, 7 due to intercurrent disease, 15 for other reasons).   Number of withdrawals from groups included in this review: not stated.
Free of other bias?	High risk	Funding Velux Fonden, Velux International A/S, Denmark.

Methods	Randomisation was achieved by giving the next sequential participant number using computer‐generated random number. The randomisation code was prepared in blocks of 9 to allow equal randomisation to all treatment groups.   Number of centres: Multi‐centre.   Design: parallel, placebo‐controlled.   Blinding: double blinded.   No. of women randomised: 1265.   No. of women analysed: 1241.   Intention to treat: Author stated an intention‐to‐treat analysis, although the actual number analysed was less than the original number randomised.   One treatment group was terminated early due to high rate of hyperplasia. All remaining groups had rates of withdrawal 22% ‐ 30%, and 3% of the total number of women were considered non compliant. Specific reasons for withdrawal were not provided. From numbers subsequently provided by the researchers, it appears that the 24 women that withdrew comprised n = 1 from control group, n = 13 from oestrogen + progesterone treatment groups, n = 10 from oestrogen only groups).   Power calculation performed.   Source of funding Involvement of Parke Davis Pharmaceutical Research Division.
Participants	Postmenopausal (treatment group oestrogen + progesterone 52.5 ± 3.9 years, 31.2 ± 17.3 months since LMP; treatment group oestrogen only 52.0 ± 4.0 years, 32.8 ± 19.4 months since LMP; control group 51.7 ± 4.1 years, 31.5 ± 20.2 months since LMP). Natural menopause.   Location and country recruited: Outpatients at 65 centre's in the United States.   Ethnicity: 95% Caucasian, 5% other (not stated).   Menopausal status confirmed by: FSH equal or greater than 40 IU/L; estradiol equal or greater than 73 pmol/L [20 pg/mL].   Inclusion criteria: Low‐density lipoprotein cholesterol [LDL‐C] < 4.91 mmol/L; spinal trabecular bone density between 90 mg/cm3 and 150 mg/cm3; atrophic endometrium; no major illnesses.   Exclusion criteria: Baseline vaginal bleeding; baseline mammography suggestive of malignant disease, chronic use of medications that affect bone calcium metabolism; significant vasomotor symptoms requiring medical treatment.   Groups comparable at baseline for age, months since last menstrual period, bone mineral density, weight. Baseline equality of smoking, diet and physical activity not stated.
Interventions	Rx1: ethinyl estradiol 1.0 µg + norethindrone acetate 0.2 mg + calcium 1000 mg daily;   Rx2: ethinyl estradiol 2.5 µg + norethindrone acetate 0.5 mg + calcium 1000 mg daily;   Rx3: ethinyl estradiol 5 µg + norethindrone acetate 1.0 mg + calcium 1000 mg daily;   Rx4: ethinyl estradiol 10 µg + norethindrone acetate 1.0 mg + calcium 1000 mg daily;   Rx5: ethinyl estradiol 1.0 µg + calcium 1000 mg daily;   Rx6: ethinyl estradiol 2.5 µg + calcium 1000 mg daily;   Rx7: ethinyl estradiol 5.0 µg + calcium 1000 mg daily;   Rx8: ethinyl estradiol 10.0 µg + calcium 1000 mg daily;   Control: Placebo + calcium 1000 mg daily.   Oral administration.   Similarity of HRT and placebo not reported.   Duration of trial: 24 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: bone mineral density measured by quantitative computed tomography, Serum lipids,   Endometrial effects as assessed by rate of hyperplasia and proliferative status
Notes	Author contacted for additional information. Reply received.   Subjects in the 10 mg EE2 group were terminated early owing to a high rate of hyperplasia for that treatment group.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation was achieved by giving the next sequential participant number using computer‐generated random number. The randomisation code was prepared in blocks of 9 to allow equal randomisation to all treatment groups.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	Unclear risk	Intention to treat: Author stated an intention‐to‐treat analysis, although the actual number analysed was less than the original number randomised.
Free of other bias?	High risk	Source of funding Involvement of Parke Davis Pharmaceutical Research Division.

Methods	Randomisation achieved by sequential administration of pre‐numbered containers to participants, ratio of 3:2 active treatment: placebo.   Number of centres: Single centre.   Design: parallel, placebo‐controlled.   Blinding: double blinded.   No. of women randomised: 78.   No. of women analysed: 68.   No intention‐to‐treat analysis.   Withdrawals n = 10 (1 from each group within 2 days of randomisation; 1 in placebo group who wished for HRT and withdrew due to lack of bleeding; 1 in placebo group due to hospitalisation for dizziness; 6 in treatment group due to nausea, menorrhagia, feeling unwell, tablets lodged in throat, dysmenorrhoea, increase in headache). No losses to follow up.   Power calculation performed.   Financial support provided by Novo‐Nordisk Pharma AS.
Participants	Postmenopausal (treatment group 55 ± 4 years, placebo group 55 ± 4 years). Natural menopause.   Location and country recruited: Population sample from Oslo, Norway.   Ethnicity not stated.   Menopausal status confirmed by: No vaginal bleeding within 6 months prior to study, FSH level equal or greater than 26 IU l‐1, serum oestradiol levels equal or greater than 0.2 mol l‐1.   Inclusion criteria: Aged 40‐60 years with intact uterus, FSH and oestradiol levels within postmenopausal range, Total cholesterol levels in range 6.2 ‐ 10.5 mmol l‐1, Total LDL cholesterol concentration equal or greater than 4.2 mmol l‐1, Triglyceride levels equal or greater than 4.0 mmol l‐1, those women who did not require treatment of menopausal symptoms.   Exclusion criteria: BMI > 30 kg/m‐2, diastolic blood pressure > 95 mmHg, systolic blood pressure > 260 mm Hg, equal or greater than 2 mm ST depression at baseline stress electrocardiogram, use of lipid‐lowering drugs, consumption of soluble fibre or fish oil supplements within 8 weeks of the start of the study, abnormal baseline mammogram or cervical smear, cardiac disease, diabetes, contraindications to the use of oestrogen (breast or endometrial carcinoma, acute or chronic liver disease, thromboembolic disorder or cerebrovascular accident associated with previous oestrogen use, haemoglobinopathies or porphyria), evidence of secondary hypercholesterolaemia.   Groups comparable at baseline for age, years since menopause, BMI, waist‐hip ratio, smoking, diet and physical activity.
Interventions	Treatment: 17b‐oestradiol 2 mg for 22 days, 17b‐oestradiol 1 mg for 6 days + norethisterone acetate 1 mg for 10 days (Trisekvens) + dietary therapy.   Control: Placebo + dietary therapy.   Oral administration, HRT and placebo identical.   Duration of trial: 12 months.
Outcomes	Relevant outcome: BMI, waist‐hip measurement.   Primary outcomes of trial: LDL cholesterol, HDL cholesterol, HDL sub fractions, lipoprotein populations, triglycerides, apolipoproteins, fibrinogen, glucose tolerance.
Notes	All subjects had elevated LDL cholesterol levels.   Author contacted for additional information. Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved by sequential administration of pre‐numbered containers to participants, ratio of 3:2 active treatment: placebo.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to‐treat analysis.   Withdrawals n = 10 (1 from each group within 2 days of randomisation; 1 in placebo group who wished for HRT and withdrew due to lack of bleeding; 1 in placebo group due to hospitalisation for dizziness; 6 in treatment group due to nausea, menorrhagia, feeling unwell, tablets lodged in throat, dysmenorrhoea, increase in headache). No losses to follow up.
Free of other bias?	Unclear risk	Financial support provided by Novo‐Nordisk Pharma AS.

Methods	Randomisation achieved using a randomized permuted block algorithm, stratified by clinical centre site and age group.   Number of centres: 40 US clinical centres.   Design: parallel, placebo‐controlled.   Blinding: double‐blinding.   No. of women randomised: 16 608.   No of women analysed: 13 859 (the original analysis included all the 16 608 women, but the data obtained from the authors on body weight and BMI was available for only 13 859 women by the 4th year and 431 women by the 8th year).   Intention to treat analysis for the original data, but no intention‐to‐‐ treat analysis for additional data we obtained from the authors.   Unknown vital status: 307 in the treatment group and 276 in the placebo group.   Power calculation performed.   Source of funding: The National Heart, Lung, and Blood Institute funds the WHI program and Wyeth‐Ayerst Research provided the study medication.
Participants	Postmenopausal women aged 50‐79. Mean age of placebo group 63.3 +/‐ 7.1 years; treatment group 63.2 +/‐ 7.1 years.   Natural menopause.   Location and country: Population‐based sample of postmenopausal women in the United States.   Ethnicity: white (83.9%), Black (6.5%), Hispanic (5.5%), American Indian (0.3%), Asian/Pacific Islander (2.3%) and unknown (1.5%).   Menopausal status confirmed by amenorrhoea for at least 12 years; no biochemical investigations.   Inclusion criteria: Age 50‐79 years at initial screening, postmenopausal, likelihood of residence in the area for 3 years, and provision of written informed consent.   Exclusion criteria: competing risks (any medical condition likely to be associated with a predicted survival of <3 years), safety (e.g.. prior breast cancer, other prior cancer within the last 10 years except non melanoma skin cancer, low hematocrit or platelet count), and adherence and retention concerns (e.g.. alcoholism, dementia).   Groups comparable in baseline characteristics such as age, weight, smoking, but diet and physical activity not reported.
Interventions	Treatment: conjugated equine estrogens, 0.625 mg/day plus medroxyprogesterone acetate, 2.5 mg/day, in 1 tablet.   Control: matching placebo.   Oral administration.   HRT and placebo identical.   Duration: 8 years (mean 5.2 years); however we analysed data recorded in the 4th year because only 431 women had data on body weight and BMI at the end of the trial
Outcomes	Relevant outcome: body weight and BMI.   Primary outcomes: coronary heart disease and invasive breast cancer.
Notes	Authors contacted for additional information.   Reply received.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved using a randomized permuted block algorithm, stratified by clinical centre site and age group.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No. of women randomised: 16 608.   No of women analysed: 13 859 (the original analysis included all the 16 608 women, but the data obtained from the authors on body weight and BMI was available for only 13 859 women by the 4th year and 431 women by the 8th year).   Intention to treat analysis for the original data, but no intention‐to‐‐ treat analysis for additional data we obtained from the authors.
Free of other bias?	High risk	Source of funding: The National Heart, Lung, and Blood Institute funds the WHI program and Wyeth‐Ayerst Research provided the study medication.

Methods	Randomisation achieved using a permuted block algorithm, stratified by clinical centre and age group.
Participants	Postmenopausal women aged 50 to 70.   Women with prior hysterectomy.   Location and country: population‐based sample of hysterectomized postmenopausal women in the United States.   Ethnicity: White (75.3%), Blck (15.1%), Hispanic (6.1%), American Indian (0.7%), Asian/Pacific Islander (1.5%) and unknown (1.4%).   Menopausal status was based on prior hysterectomy; no biochemical investigations.   Inclusion criteria: Hysterectomized women aged 50 to70 at initial screening, had undergone hysterectomy, were likely to reside in the area for 3 years and provision of written informed consent.   Groups were comparable at baseline   Exclusion criteria: competing risks (any medical condition likely to be associated with a predicted survival of < 3years), safety (e.g. prior breast cancer), adherence and retention concerns (e.g. alcoholism) or clinical judgment of the participant's health care practitioner to continue in symptomatic or osteoporotic women.
Interventions	Treatment: conjugated equine estrogens, 0.625 mg/d   Control: matching placebo.   Oral administration.   HRT and placebo identical.   Duration: 9 years (mean 5.2 years), however we analyzed data end the end of the 4th year because of the significant number of women whose weight and BMI were not recorded after that time.
Outcomes	Relevant outcome: body weight and BMI.   Primary outcome: coronary heart disease and invasive breast cancer.
Notes	Authors contacted for additional information.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomisation achieved using a permuted block algorithm, stratified by clinical centre and age group.
Allocation concealment?	Unclear risk	D ‐ Not used

Methods	Randomisation achieved using on‐site computer system.   Number of centres: More than one (exact number not stated).   Design: parallel, placebo‐controlled.   Blinding: double‐blinded.   No. of women randomised: 58 overall, 29 from groups analysed.   No. of women analysed: 21 from groups included in this analysis.   No intention‐to‐treat analysis.   Withdrawals n = 8 from groups analysed. 3 subjects (2 treatment group; 1 placebo) withdrew due to inability to tolerate the medications; 2 participants from placebo group left due to carcinoma of the ovary (1) and cerebrovascular accident (1); and 3 were lost to follow‐up.   No power calculation reported.   Source of funding: No industry funding.
Participants	Postmenopausal (mean age over all groups analysed 52.6 ± 3.8 years, range 45‐57 years).   Location and country recruited: Metabolic bone disease clinics in UK.   Ethnicity: Ethnicity not stated.   Menopausal status confirmed by: Serum gonadotropin and 17b‐estradiol levels.   Inclusion criteria: Within 1 and 5 years of natural menopause; bone mineral density within the reference range established for normal healthy women aged between 30 and 40 years; no previous treatment with HRT or bisphosphonate's.   Exclusion criteria: oophorectomy; spinal, wrist or hip fractures; secondary causes of osteoporosis; medication known to affect calcium metabolism.   Groups comparable at baseline for age, years since menopause, BMI. Equality of smoking, diet and exercise levels not stated.
Interventions	Treatment: 17 beta estradiol 1.5 mg daily + micronized progesterone 200 mg/d for 12 days + 1.0 g/d elemental calcium.   Control: 1.0 g/d elemental calcium.   Mode of administration: Estradiol administered percutaneously in form of gel, progesterone administered orally.   Treatment and control not identical.   Duration of trial: 48 months.
Outcomes	Relevant outcome: BMI.   Primary outcomes of trial: effect of etidronate on bone mineral density.
Notes	Further information requested from author. Reply received.   This trial compared four treatment regimes:   Rx1: 17 beta estradiol 1.5 mg daily + micronized progesterone 200 mg/d for 12 days + 1.0 gm/d elemental calcium   Rx2: intermittent cyclically administered etidronate (ICE) 400 mg/d for 14 days followed by 10 weeks supplementation with 1.0 G/d elemental calcium   Rx3: combination of HRT, ICE and calcium supplementation   Rx4: 1.0 G/d elemental calcium   For the purposes of this review, Rx1 and Rx4 were analysed. Rx2 and Rx3 combined HRT treatment with a drug that is considered to be outside the framework of the review protocol.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved using on‐site computer system.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to‐treat analysis.   Withdrawals n = 8 from groups analysed. 3 subjects (2 treatment group; 1 placebo) withdrew due to inability to tolerate the medications; 2 participants from placebo group left due to carcinoma of the ovary (1) and cerebrovascular accident (1); and 3 were lost to follow‐up.
Free of other bias?	Unclear risk	Source of funding: No industry funding.

Methods	Randomisation achieved by a computer‐generated random number table to produce balanced blocks of participant numbers within each centre. A 1:1:1 treatment allocation was used.   Number of centres: 8.   Design: Parallel, placebo‐controlled.   Blinding: designed to be double‐blind, although the placebo group became unblinded due to lack of bleeding. Laboratory staff were kept blind.   No. of women randomised: 255.   No. of women analysed: 188.   No intention‐to‐treat analysis.   Withdrawals and losses to follow up: 67 discontinued the study early, mainly because of poor participant levels (n = 19) or adverse reactions (n = 19; skin irritation at patch application site (8), breast tenderness (5), abnormal bleeding (3), abdominal pain and weight gain (2), low back pain (1). Other reasons for discontinuance were unsatisfactory therapeutic effect (5), withdrawal of consent (10), and loss to follow‐up (14).   Power calculation performed.   Source of funding: No source of funding stated.
Participants	Postmenopausal; natural menopause 1‐5 years prior to study.   Location and country recruited: Outpatient clinics at participating centres in Italy.   Ethnicity not stated.   Menopausal status confirmed by: Kupperman index score >14; circulating levels of FSH >30 mlU/ml.   Inclusion criteria: BMI <26; normal histological picture at endometrial biopsy and cervical smear; normal mammography within previous 12 months.   Exclusion criteria: Previous history or evidence of endometrial hyperplasia or breast cancer; abnormal vaginal bleeding within previous 12 months; presence or history of endometriosis; active chronic skin diseases; history of thromboembolic disorders associated with previous oestrogen use; known allergy to study medications; serious illnesses such as Cushing's Disease or cancer; history of cerebral or coronary artery disease; congestive heart failure; obstructive lung disease; moderate to severe hypertension; severe liver disease; other clinically important abnormalities at laboratory screening.   Groups compatible at baseline for age, weight, time since menopause, Kupperman index. Baseline equality of smoking, diet and physical activity not reported.
Interventions	Rx 1: Cyclic regime consisting of E2 patches 0.05 mg/day (Estraderm TTS 50) days 1‐21; placebo patches days 22‐28; 10mg medroxyprogesterone acetate tablet days 14‐25; placebo tablet on other days.   Rx 2: Continuous regime consisting of E2 patches 0.05 mg/day (Estraderm TTS 50) days 1‐28; 10mg medroxyprogesterone acetate tablet days 14‐25; placebo tablet on other days.   (These groups were combined for the purposes of this meta analysis).   Control: Placebo patches and placebo tablets given over entire study period.   Mode of administration: Transdermally by patches   Similarity of treatment and placebo not reported.   Duration of trial: 12 months.
Outcomes	Relevant outcome: Body weight.   Primary outcomes of trial: The effect of two transdermal E2 regimens combined with progestin on the procoagulant and inhibiting factors of haemostatic balance in postmenopausal women.
Notes	Author contacted for additional information. Reply received
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation achieved by a computer‐generated random number table to produce balanced blocks of participant numbers within each centre. A 1:1:1 treatment allocation was used.
Allocation concealment?	Low risk	A ‐ Adequate
Blinding?   All outcomes	Low risk	Double
Incomplete outcome data addressed?   All outcomes	High risk	No intention‐to‐treat analysis.   Withdrawals and losses to follow up: 67 discontinued the study early, mainly because of poor participant levels (n = 19) or adverse reactions (n = 19; skin irritation at patch application site (8), breast tenderness (5), abnormal bleeding (3), abdominal pain and weight gain (2), low back pain (1). Other reasons for discontinuance were unsatisfactory therapeutic effect (5), withdrawal of consent (10), and loss to follow‐up (14).
Free of other bias?	Unclear risk	No source of funding stated

Study	Reason for exclusion
Bellantoni 1991	Authors contacted. Endpoint weight details not available.
Binder 1996	Author contacted and has stated that participants were not randomised to treatment and control groups. This makes the study ineligible for analysis, as per the protocol
Cagnacci 2002	Treatment was less than three months
Cherry 2002	Authors contacted. Endpoint weight details not available
Chung 1996	This is a study which has cross over at 6 months. Body weight data was not collected at this crossover time.
Davis 2000	There was no control group.
Gambacciani 1995	The treatment group consisted of oestradiol combined with cyproterone acetate.
Gambacciani 1997	Author contacted and states that participants were not randomised to treatment and control groups. Only randomised trials will be included in this analysis, as per the protocol.
Genant 1990	Author contacted. Endpoint data no longer available.
Hanggi 1997	Participants were not randomised to the untreated control group. Those women not desiring HRT chose to be in the untreated control group; participants desiring HRT were randomised to one of three treatment groups.
Hosking 1998	Treatment group was administered alendronate and control group was given placebo.
Lindsay 1984	Author contacted. Endpoint data no longer available.
Lufkin 1992	Authors contacted. They are unable to provide baseline and endpoint weight details as the original data has been destroyed.
MacLennan 1993	Participants were receiving oestrogen immediately before the trial. Such studies will be excluded, according to the protocol.
Morin‐Papunen 2001	There was no control group.
Morrison 1980	Treatment group was given medroxyprogesterone only.
Nand 1998	There was no control group.
Ongphiphadhanakul	There was no control group.
Pickar 1998	Both treatment and control groups took conjugated oestradiol.
Utian 1973	Author contacted; unable to supply endpoint weight data.