Papers

Randomised comparison of oestrogen versus oestrogen plus progestogen hormone replacement therapy in women with hysterectomy

BMJ 1996; 312 doi: http://dx.doi.org/10.1136/bmj.312.7029.473 (Published 24 February 1996) Cite this as: BMJ 1996;312:473
  1. T W Meade, Centres and participants are listed at the end of this report.a
  1. aMedical Research Council's General Practice Research Framework
  1. Correspondence to: Professor T W Meade, MRC Epidemiology and Medical Care Unit, Wolfson Institute of Preventive Medicine, St Bartholomew's and the Royal London Hospital School of Medicine and Dentistry, London EC1M 6BQ.
  • Accepted 18 November 1995

Abstract

Objective: To compare the acceptability and symptomatic and metabolic effects of two regimens of hormone replacement therapy in women with hysterectomy.

Design: Randomised, double blind comparison.

Setting: Seven group practices in the Medical Research Council's general practice research framework.

Subjects: 321 women with hysterectomy aged 35-59.

Interventions: Hormone replacement therapy with (a) conjugated equine oestrogen 625 µg daily alone or (b) conjugated equine oestrogen 625 µg daily plus the progestogen norgestrel 150 µg daily for the last 12 days of the “cycle.” Main outcome measures: Changes in blood pressure, weight, symptoms, and haemostatic and lipid values.

Results: After two years 36% (57/158) of women randomly allocated to take oestrogen alone had discontinued treatment as compared with 30% (49/163) of women allocated to take oestrogen plus progestogen. Smokers were more likely to withdraw than non-smokers. There were no clear differences between the two groups in symptoms often attributed to hormone replacement therapy or in blood pressure or weight. At one year low density lipoprotein cholesterol concentrations had fallen substantially in both groups. High density lipoprotein cholesterol concentrations rose to significantly higher values in women taking oestrogen alone compared with those taking oestrogen plus progestogen, though triglyceride concentrations and factor VII activity were also significantly higher in this group. Fibrinogen concentration tended to fall, though not significantly, in both groups, possibly more in women taking oestrogen alone.

Conclusions: Oestrogen plus progestogen was no less well tolerated than oestrogen alone. There was a fairly even balance between possibly beneficial and adverse effects of the two regimens on lipid concentrations and coagulability. Concern that the combined regimen may not have the cardioprotective effects ascribed to oestrogen alone can to some extent be allayed, with reassurance for the growing numbers of women with intact uteri using the combined regimen. Misgivings about the combined regimen in women with hysterectomy on the grounds of its acceptability and its effects on lipid values may also be unfounded.

Key messages

  • Key messages

  • In fact, in hysterectomised women norgestrel containing hormone replacement preparations may be as well tolerated as unopposed oestrogen

  • There may be little difference between norgestrel containing hormone replacement therapy and oestrogen only preparations in their lipid and haemostatic effects

  • Pending the results of clinical end point trials, anxieties about the metabolic effects of norgestrel in hormone replacement therapy can largely be allayed

Introduction

Observational studies suggest that hormone replacement therapy reduces the incidence of ischaemic heart disease and possibly also of stroke.1 2 In most of these studies hormone replacement was given predominantly as oestrogen only. When replacement therapy with oestrogen only was recognised as sometimes leading to endometrial cancer a progestogen was added in the last 10 or 12 days of the “cycle.” However, combined replacement therapy with oestrogen and progestogen may affect the lipid profile adversely and so reduce or abolish the apparent benefits due to oestrogen alone.

Around 85% of postmenopausal women have intact uteri—that is, they have not had a hysterectomy. Almost one fifth of these women are currently being given hormone replacement therapy at any one time (Medical Research Council Epidemiology and Medical Care Unit, unpublished), mostly as a combined regimen of oestrogen and progestogen. As these women account for probably well over half of all usage of hormone replacement therapy, it is necessary to establish the effects of the combined regimen. Direct comparison between replacement therapy with oestrogen alone and therapy with oestrogen plus progestogen cannot easily be made in women with intact uteri because of the increased risk of uterine cancer associated with oestrogen alone. This difficulty would not arise in women who have had a hysterectomy, and this paper describes the clinical and metabolic findings in such a comparison. When the study was initiated it was thought that most women with a hysterectomy should take hormone replacement therapy. Hence no untreated (placebo) group was included.

Patients and methods

The trial was conducted through seven group practices in the MRC's general practice research framework. A trained research nurse based in each practice identified eligible patients, arranged their recruitment between September 1991 and March 1993, and conducted follow up examinations. Results are based on follow up till September 1994—that is, for between 17 and 38 months.

Diagnostic listings were used to identify all women under 60 who had undergone hysterectomy. Of the 1018 women identified, 64 were excluded for various reasons and 950 were invited by letter to attend for an explanatory interview. Women who expressed interest in the randomised treatment comparison were given an appointment for an entry medical examination three weeks later, and 321 women entered.

Eligibility—Of the 696 women who attended for screening, 34 (4.9%) were excluded because of a family history of breast cancer in a first degree relative, 17 (2.4%) because of a past or present personal history of cancer, and 17 because of abnormal cervical smears. Various other conditions, including otosclerosis, led to the exclusion of a further 96 (13.8%) women. Eligible women aged over 50 were encouraged to have mammography through the national breast screening programme if they had not recently done so.

Treatment consisted either of oral conjugated equine oestrogen 625 µg daily (Premarin) or of oral conjugated equine oestrogen 625 µg daily plus oral progestogen (norgestrel) 150 µg for the last 12 days of each 28 day cycle (Prempak-C). In both groups oestrogen was taken for all 28 days of each cycle. Within each practice, allocation to oestrogen alone or the combined regimen was at random both for women who were not currently taking hormone replacement therapy (205) and for those who were already taking it at trial entry (116); for these women treatment was randomly reassigned. As all but 12 (10%) of these women had received oestrogen alone, the likelihood of any differential effects of oestrogen alone and oestrogen plus progestogen on reported symptoms was low and also would have waned with time. Results for clotting factor and lipid values are presented separately (table 2) for women who had not been using hormone replacement therapy at trial entry as well as for all participants. Treatment was double blind, identical placebo norgestrel tablets being included in the blister packs for women randomised to receive oestrogen alone.

Entry investigations and follow up—Smoking history was elicited. The occurrence of a range of symptoms associated with the menopause, weight, and blood pressure was recorded at each visit. Non-fasting blood samples were taken at entry and at three, six, and 12 months for serum total cholesterol and triglyceride concentrations (Ektachem 700 multilayer film methods) and for high density lipoprotein cholesterol concentrations.3 Low density lipoprotein cholesterol concentrations4 and plasma factor VII (VIIC) activity and fibrinogen concentrations5 were measured as described.

Statistical—The main comparisons were between women randomly allocated to the two treatment groups, oestrogen alone and oestrogen plus progestogen. Differences in means were tested by t tests or Mann-Whitney tests. P values for differences between the groups in clotting factor and lipid values are presented only for the results at 12 months. Because there was no untreated group, changes within each separate group must be interpreted cautiously. However, as women were not selected by entry values of any blood or other variables, changes were unlikely to be the result of regression to the mean and over the time periods studied age related changes would have been minimal. Accordingly, we comment on pronounced changes within treatment groups. The findings with respect to clotting factor and lipid values are confined to women from whom samples were obtained at each visit during the first year (the inclusion of data for those who withdrew making little difference to the results). In figure 5 the y axes are scaled to extend approximately one standard deviation above and below the zero value, thus allowing direct visual comparisons of the magnitude of the differences between the groups for different variables.

Ethical approval was obtained from local research ethics committees and from the clinical research ethics committee of the Royal College of General Practitioners.

Results

Of the 950 women invited for screening, 696 (73%) attended, of whom 677 (97%) were white. Of these, 532 (76%; 56% of the women invited) were eligible for the trial and 321 (age range 35-59 years) entered. These women (158 randomised to receive oestrogen alone, 163 randomised to receive oestrogen plus progestogen) represented 60% of those eligible, 46% of those who attended initial interview, and 34% of those invited. Among the 532 eligible women the main reasons for not entering the trial were concern about breast cancer and unwillingness to take long term therapy in the absence of symptoms.

Trial entrants—Table 1 summarises the characteristics of the two groups. There were slightly more smokers in the group allocated to the combined regimen. A few more women in the oestrogen alone group had previously taken hormone replacement therapy. These and other differences were not pronounced.

Table 1

Characteristics of trial entrants by randomised treatment

View this table:

Withdrawal from randomised treatment—Figure 1 shows that after one year 75 (23.4%) women had withdrawn from randomised treatment, 44 (28%) from the oestrogen alone group and 31 (19%) from the combined regimen. After two years the cumulative proportions of women having withdrawn were 36% (57/158) and 30% (49/163) respectively. Current smokers were more likely to withdraw than non-smokers, 31 (43%) of the 72 smokers having done so.

Blood pressure and weight remained stable throughout follow up and there were no differences between the groups (figs 2 and 3)

Fig 1
Fig 1

Percentages of women withdrawn from randomised treatment

Fig 2
Fig 2

Mean blood pressure recorded in the two treatment groups

Fig 3
Fig 3

Mean body weights recorded in the two treatment groups

Symptoms—Figure 4 shows the occurrence of symptoms at entry and at each follow up visit. There were only small differences between the treatment groups. Both groups experienced a reduction in hot flushes and night sweats. There was no clear difference in reports of breast tenderness or in answers to questions on mood (not shown). The oestrogen plus progestogen regimen may have been associated with fewer reports of vaginal dryness and bladder problems.

Fig 4
Fig 4

Percentages of women reporting symptoms since last visit

Haemostatic and lipid values—Table 2 shows the mean haemostatic and lipid values at each follow up visit and figure 5 the changes from initial values in the women not taking hormone replacement therapy at entry. Factor VIIC activity rose in women taking oestrogen alone, the difference between the two groups being significant at 12 months. Fibrinogen concentration was lower at 12 months than at entry in both groups, more so in the oestrogen alone group. However, the difference between the groups was not significant. High density lipoprotein cholesterol concentration rose in women taking oestrogen alone but was unchanged in the combined therapy group, leading to a significant difference between the groups at 12 months. Both groups showed a similar, pronounced fall in the concentration of low density lipoprotein cholesterol. As a result of these changes the concentration of total cholesterol fell in both groups but more so in the combined therapy group, so that the difference between the groups during follow up was marginally significant. Triglyceride concentrations rose in the oestrogen alone group and changed little in the combined therapy group, the difference at 12 months being significant.

Table 2

Mean values and standard deviations of haemostatic and lipid variables at each visit by randomised group

View this table:
Fig 5
Fig 5

Changes in haemostatic and lipid values from entry to three, six, and 12 months in women new to hormone replacement therapy and who provided samples at all visits (numbers in each group shown at 12 months)

Discussion

Any differences between the treatment groups were attributable to norgestrel, the progestogen in the combined regimen, which is considered to be relatively androgenic. The recent United States postmenopausal estrogen/progestin interventions (PEPI) trial did not include a preparation containing norgestrel.6 Both preparations led to a decrease in vasomotor symptoms (flushing and sweating) but reasons besides the recorded symptoms seem to have led to the somewhat higher rate of withdrawal early in the trial among women in the oestrogen alone group. It seems unlikely that the oestrogen plus progestogen regimen is less well tolerated than oestrogen alone.

Total cholesterol concentration is as strongly related to ischaemic heart disease in women as it is in men.7 The somewhat higher concentration in the oestrogen alone group was due to a large decline in low density lipoprotein cholesterol concentration in both groups and a substantial increase in high density lipoprotein cholesterol concentration in the oestrogen alone group. Triglyceride values were unchanged in the combined therapy group and rose in the oestrogen alone group. Triglyceride values may be independently related to ischaemic heart disease in women.8 Changes in triglyceride concentrations due to hormone replacement therapy might not be atherogenic9 but might still predispose to ischaemic heart disease through increased coagulability10 or impaired fibrinolysis.11 Hence the evidence for a beneficial effect of oestrogen alone compared with oestrogen plus progestogen on blood lipids as a whole is somewhat equivocal.

The higher factor VIIC activities attributable to oestrogen alone compared with oestrogen plus progestogen are clear. Raised factor VIIC activity is associated with increased mortality from ischaemic heart disease in men,5 12 13 and factor VIIC activity is correlated with indices of thrombin generation.14 Two randomised comparisons have shown that treatment with oestrogen alone leads to an increase in the prothrombin activation peptide F1.2, a marker of thrombin generation.15 16 Factor VIIC activity did not rise in the combined therapy group in our trial, and the results of a large observational study in which the predominant progestogen was medroxyprogesterone acetate also suggested that the combined regimen has less effect on factor VIIC activity than oestrogen alone.17 The apparent absence of an effect of oestrogen plus progestogen on factor VIIC activity suggests that the combined regimen is perhaps unlikely to increase F1.2, though this remains to be established.

Besides the United States trial,6 two large observational studies have found substantially lower fibrinogen concentrations in hormone replacement therapy users than in non-users17 18 with no apparent difference in this respect between opposed and unopposed hormone replacement therapy.17 Fibrinogen concentration fell in both groups in our trial, somewhat more in the oestrogen alone group than in the combined therapy group, though the difference at one year was not significant. Any difference might partly be accounted for by the combined effects of a higher withdrawal rate among women taking oestrogen alone (fig 1) and in smokers. High fibrinogen concentrations are related to ischaemic heart disease in women as well as in men.19 If the two regimens do affect fibrinogen equally, then the balance of advantage in terms of coagulability and when taking account of factor VIIC as well may favour the combined regimen. However, regimens of hormone replacement therapy may have effects that we did not study—for example, on functional pathways influencing blood flow.20 The clinical value of different regimens will depend on the balance of changes in all the pathways affected.

COMPARISON WITH UNITED STATES TRIAL

All women in the actively treated groups in the postmenopausal estrogen/progestin interventions trial6 and both groups in our trial used conjugated equine oestrogen in a dose of 625 µg. The main difference was in the progestogen. In the United States trial two groups took medroxyprogesterone acetate, one cyclically and one continuously. We used norgestrel for the combined therapy group. Reports on the effects of progestogens on lipids are confusing,21 so far being mainly dominated by extrapolations from data on oral contraceptives, observational studies, and short term trials based usually on small numbers. Despite the different progestogens used in the two trials there were several similarities between our results and the results of the United States trial. Thus in hysterectomised women oestrogen alone had the largest effect in raising the high density lipoprotein cholesterol concentration in the United States trial, and women randomised to oestrogen alone in our trial had significantly higher high density lipoprotein cholesterol concentrations at one year than those randomised to the combined regimen.

Active treatment led to a substantial reduction in low density lipoprotein cholesterol concentration in the United States trial, and (subject to the reservation that we did not include a placebo group) the magnitude of the decrease means that the same was almost certainly true in our trial. Triglyceride concentrations rose to much the same extent in all the actively treated groups in the interventions trial but only in women given oestrogen alone in our trial. A two year trial in 113 women showed that norethisterone acetate also has effects on lipids similar to those of medroxyprogesterone acetate and norgestrel.22 As in our trial, differences in fibrinogen concentrations between the actively treated groups in the interventions trial were not significant. Neither trial found that treatment affected blood pressure. In the interventions trial women given oestrogen alone gained less weight than women in other groups. We found no difference in weight. Overall rather fewer women withdrew from randomised treatment in the interventions trial than in our trial (apart from a high rate of withdrawals among non-hysterectomised women given oestrogens alone). But in both trials there was a tendency among hysterectomised women for more of those in the oestrogen only groups to stop their initially assigned treatment.

The proportion of women using hormone replacement therapy at any one time in Britain has increased considerably over the past 15 years23 24 to nearly 20% in 1993-4 (MRC Epidemiology and Medical Care Unit, unpublished). By the end of the decade some 30% or more of postmenopausal women may be taking hormone replacement therapy. Preparations containing oestrogen plus the progestogen norgestrel account for a high proportion of all usage of hormone replacement therapy.

Pending the results of end point trials we conclude, firstly, that differences between the effects of norgestrel and medroxyprogesterone acetate may be less than has often been assumed, with implications for decisions about different regimens used by women with intact uteri taking opposed hormone replacement therapy. This conclusion assumes that findings in women who have had a hysterectomy are also a guide to the effects in those who have not. This is not unreasonable given the general similarity of our results and those of the United States trial,6 in which 68% of women had not had a hysterectomy. Secondly, replacement therapy with oestrogen plus progestogen may be as cardioprotective as therapy with oestrogen alone, a possibility supported by an observational study25 and a similar finding on stroke.26 Thirdly, misgivings about therapy with oestrogen plus progestogen on the grounds of acceptability and metabolic effects in patients with hysterectomy may also be unfounded. Aside from the possible bone sparing effect of progestogen,27 oestrogen plus progestogen given as replacement therapy may often be the regimen of choice in these women as well as in those in whom the uterus is present.

Centres and participants: MRC Epidemiology and Medical Care Unit, St Bartholomew's and the Royal London Hospital School of Medicine and Dentistry, London—H C Wilkes, T W Meade, M R Vickers, W Browne, M A Walgrove, V Ruddock, Y Stirling, C Lennon, Z Islam, P J Brennan. Department of Cardiovascular Biochemistry, St Bartholomew's Hospital Medical College, London—N E Miller, M N Nanjee. Practices in general practice research framework—Dr G Walker, Mrs M Goldsborough; Dr P Jennings, Mrs B Hinton; Dr J Davies, Dr M Maggs, Mrs J Charlton, Mrs J Davies; Dr R Rolls, Mrs D Dowell, Mrs F Round, Mrs M Holton; Dr R King, Dr P Booth, Mrs P Jones, Mrs P Bennett; Dr C Bevan, Mrs S Derrick; Dr I Orpen, Mrs A Baldwin, Mrs R Heyworth.

We thank members of the MRC's working party on hormone replacement therapy (chairman: Professor G M Besser) for advice and Ruth Gold, Sara Bordoley, and Deepa Patel for administrative help.

Footnotes

  • Funding The work was funded by the MRC. Tablets were provided, initially free and later at cost, by Wyeth UK.

  • Conflict of interest None.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.