Correspondence: Ms I. Byrjalsen, Osteometer BioTech A/S, Herlev Hovedgade 207, DK-2730 Herlev, Denmark.
Objective To assess the effects on the postmenopausal endometrium of two doses of oral piperazine oestrone sulphate and interrupted norethisterone in comparison with a continuously combined regimen and placebo.
Design A prospective randomised trial.
Participants Two hundred healthy postmenopausal women.
Methods Random assignment to two years of treatment with alternating three-day cycles of 1.5 mg piperazine oestrone sulphate and 1.5 mg piperazine oestrone sulphate + 0.7 mg norethisterone (highEP), or alternating three-day cycles of 0.75 mg piperazine oestrone sulphate and 0.75 mg piperazine oestrone sulphate + 0.35 mg norethisterone (lowEP), or 2 mg 17 β-oestradiol continuously combined with 1 mg norethisterone acetate (E2+NETA), or placebo.
Main outcome measures Effect of treatment on endometrial histology, endometrial thickness, occurrence of uterine bleeding, endometrial oestrogen and progesterone receptor content, endometrial isocitrate dehydrogenase activity, and serum placental protein 14.
Results The incidence of bleeding declined with time. In the second treatment year, the women receiving lowEP reported on average 7.3 days of bleeding, highEP 16.7 days, and E2+NETA 11.2 days. Histological assessment of endometrial biopsies revealed an atrophic or slightly secretory endometrium. Serum placental protein 14 increased slightly, but was statistically highly significant, during treatment, but no cyclical variation was observed. Endometrial isocitrate dehydrogenase was low in all three hormone groups and the same low level of endometrial oestrogen receptor and progesterone receptor was found comparable to the level in the placebo group.
Conclusions Histological and biochemical assessment of the endometrium showed that interrupted hormone replacement therapy induced the same pattern in endometrial parameters as continuous combined hormone replacement therapy.
Oestrogen replacement therapy is an effective treatment for relieving menopausal symptoms, including atrophic vaginal changes in postmenopausal women. Furthermore, it prevents postmenopausal bone loss, and epidemiological studies have suggested that it also protects against cardiovascular disease1. Various prescribing schedules have been used for the addition of a progestogen to oestrogen replacement therapy for the purpose of preventing endometrial hyperplasia and protecting against the endometrial cancer associated with oestrogen monotherapy2–4. The commonest schedule is sequentially combined administration in which a progestogen is added for 10 to 13 days in each cycle of oestrogen administration. One disadvantage of this form of progestogen administration is the regular monthly uterine bleeding experienced by 80% to 90% of the women, which makes hormone replacement therapy (HRT) less acceptable.
Continuously combined administration of oestrogens and progestogens is the other commonly used prescription schedule. The scientific rationale behind this type of HRT is the antimitotic activity of the progestogen: continuous administration of the progestogen constantly opposes the oestrogen growth-stimulatory effects. As a result, the endometrium is kept atrophic or slightly secretory and proliferation is avoided. The advantage of this administration schedule is that it should not give rise to uterine bleeding episodes. However, in continuously combined HRT, the total dose of ingested progestogen is relatively high, which may decrease the oestrogen-induced increased concentration of high density lipoprotein cholesterol5,6. Therefore it may be desirable to aim at as low a progestogen intake as possible. Based on this, the concept of interrupted HRT was introduced as a modification of the continuously combined HRT7. In interrupted HRT the oestrogen is administered continuously and the progestogen administration alternates between a three-day cycle of progestogen addition, followed by a progestogen-free three-day cycle (i.e. the progestogen is administered only half of the time as compared with continuously combined HRT).
This paper reports data from a prospective, placebo-controlled study, which was intended to characterise the efficacy of two interrupted treatment regimens and compare them with a continuously combined regimen to control bleeding and to ensure endometrial safety, as assessed histologically and by measurement of the endometrial thickness in postmenopausal women. Furthermore, the endometrial effects were assessed biochemically by the measurement of the content of endometrial sex steroid receptors and secretory phase markers (i.e. the secretory endometrial protein, placental protein 14) measured in serum, and endometrial isocitrate dehydrogenase.
The study was part of a large, randomised and placebo-controlled clinical trial to evaluate and compare the effects of administration of piperazine oestrone sulphate and interrupted norethisterone on climacteric symptoms and bone and lipid metabolism8. The doses of the hormones were based on experience from earlier observations. A phase II study of 0.75 mg piperazine oestrone sulphate and interrupted 0.35 mg norethisterone showed that this regimen was effective in eliminating hot flushes in 75% of the 40 postmenopausal women entered in the study and was effective in protecting the endometrium from hyperplasia7. This indicated that a higher oestrogen dose was needed to relieve vasomotor symptoms in 100% of the women, and it was therefore chosen to investigate a dose of 1.5 mg piperazine oestrone sulphate. This dose was combined with 0.7 mg norethisterone, thus maintaining the same ratio of norethisterone to piperazine oestrone sulphate that had been shown to provide endometrial protection.
The participants were selected as a representative sample of healthy postmenopausal women, aged 45 to 65 years, who had passed a natural menopause at least one year previously8. None of the women who entered the study had diseases or were taking medications known to influence the study measurements. Endometrial biopsies were taken before treatment to ensure that none of the participants had endometrial dysplasia or malignancy. A total of 200 women entered the study. After the initial examinations, the women were assigned randomly by a simple randomisation procedure using random numbers in a double-blind manner to 30-day treatment cycles with either 1.5 mg piperazine oestrone sulphate daily, interruptedly combined with 0.7 mg norethisterone on days 4–6, 10–12, 16–18, 22–24, 28–30 (highEP, n= 50), or 0.75 mg piperazine oestrone sulphate daily, interruptedly combined with 0.35 mg norethisterone on days 4–6, 10–12, 16–18, 22–24, 28–30 (lowEP, n= 50), or in a single-blind manner to 28-day treatment cycles with 2 mg of 17 β-oestradiol daily continuously combined with 1 mg norethisterone acetate (E2+NETA, n= 50), or placebo (n= 50). The highEP, lowEP, and placebo were identical in appearance, whereas the E2+NETA was different. The medication code was available only after termination of the study. The women were examined every three months during the study.
Before treatment and (according to the interrupted combination) at three months of treatment, blood samples were taken every other day three times during a six-day tablet combination period (i.e. on days one, three, and five or alternatively on days two, four, and six). Blood sampling was repeated at two years of treatment. The thickness of the endometrium was measured by transvaginal ultrasonography and endometrial biopsies were taken on day three or four in the same cycle as the blood sampling at the two years of treatment. The participants registered daily their intake of tablets and the occurrence and intensity of uterine bleeding on a calendar.
All participants gave their written informed consent. The study was performed in accordance with Helsinki Declaration II, and approved by the local ethics committee.
Sera were stored at −18°C until assayed. Serum placental protein 14 (PP14) and serum oestradiol were measured by nonequilibrium radioimmunoassays9,10. The detection limits of the assays were, respectively, 0.3 μg/L and 0.01 nmol/L; the intra-assay and inter-assay imprecisions were, respectively, 5% and 10%, and 4% and 14%.
The endometrial tissue samples were immediately divided and placed in neutral formalin for routine histological assessment and frozen on solid carbon dioxide and stored at −85°C for the biochemical analysis. The histological assessment was performed externally (Dr Bang's Laboratory, Copenhagen, Denmark) by persons with no knowledge of the treatments.
For the biochemical analysis, the tissue was homogenised (Potter-Elvehjem) in a 10 mmol/L potassium phosphate buffer, containing 1.5 mmol/L potassium-EDTA, 10 mmol/L sodium molybdate, 10 mmol/L monothioglycerol, 10% v/v glycerol, pH 7.5. The homogenates were centrifuged at 105,000 g for 45 minutes at 4°C, and the 105,000 g supernatant was used for the measurements. Endometrial cytosolic oestrogen receptor (ERc) and progesterone receptor (PgRc) were measured by steroid ligand binding assays using dextran-coated charcoal separation11. Briefly, ERc was measured by incubation with 5 nmol/L [3H]-oestradiol (TRK 322, Amersham International, Buckinghamshire, UK) for 18–20 hours at 4°C. Nonspecific binding was assessed with parallel incubation in 500 nmol/L diethylstilbestrol. PgRc was measured by incubation with 27 nmol/L [3H]-ORG 2058 [16 α-ethyl-21-hydroxy-19-norpregn-4-ene-3, 20-dione (TRK 629, Amersham International, Buckinghamshire, UK)] for 18 hours at 4°C. Nonspecific binding was assessed with parallel incubation in 1000 nmol/L ORG 2058. Isocitrate dehydrogenase (ICDH) was measured by spectrophotometry as the rate of reduction of nicotinamide adenine dinucleotide phosphate (NADH) in the presence of isocitric acid12 by an automatic procedure described13. The cytosolic protein concentration was measured according to Bradford14. The interassay imprecisions for ERc, PgRc, ICDH and protein, were, respectively 7%, 10%, 2%, and 5%. The results of the cytosolic steroid hormone receptors and ICDH were normalised to the cytosolic protein content and expressed as fmol/mg protein and U/g protein.
The data on endometrial thickness, endometrial steroid hormone receptors, endometrial isocitrate dehydrogenase, serum oestradiol, and serum placental protein 14 were logarithmically transformed to attain normality and homogeneity of variances. These values were used for all the statistical analyses. The baseline data of the women who completed the study were compared by one-way analysis of variance. The data of the women who left the study were compared with those of the women who completed the study by two-tailed Student's t test for unpaired data within each treatment group. For each period of three tablet cycles, the bleeding index was calculated as the sum of days with bleeding weighted with the intensity of bleeding (i.e. spot bleeding has a weight factor of one, slight bleeding a factor of two, moderate bleeding a factor of three, and heavy bleeding a factor of four). The bleeding index is given per tablet cycle per woman. In the groups receiving the interrupted combined regimens, the cyclical dependency of serum PP14 was assessed by means of the General Linear Models Procedure of the SAS Institute15. A test procedure was conducted to ascertain whether the model consisting of the three main parameters: long term variation (separate levels at baseline, three months, and two years), cyclical variation (separate levels at cycle days one, three, and five), and individual response as main effects, could be reduced to a model using only the long term variation and individual response as main effects. Omission of the long term variation and effect of treatment from the model was also tested. Any difference among the treatment groups was assessed by testing whether the treatment group could be omitted from the linear model consisting of treatment group and long term variation as crossed effects. Fisher's exact test was used to assess the association between endometrial histology and treatment group. One-way analysis of variance (ANOVA) was used to compare the data on endometrial thickness and the biochemically measured parameters in the four treatment groups. The two-tailed Student's t test was used to evaluate differences in endometrial markers between two groups.
Twenty-four women in the interrupted combined high dose group (highEP) left the study, because of: uterine bleeding (n= 11), adverse effects from the medication (n= 3), reasons unrelated to treatment (n= 9), and one woman was lost to follow up. Fifteen women in the interrupted combined low dose group (lowEP) left the study because of adverse effects from the medication (n= 7) and reasons unrelated to treatment (n= 8). Nineteen women in the continuous combined group (E2+NETA) left the study because of uterine bleeding (n= 5), adverse effects from the medication (n= 6), reasons unrelated to treatment (n= 7), and one woman was lost to follow up. Twenty women in the placebo group left the study because of adverse effects from the medication (n= 3), reasons unrelated to treatment (n= 15), and two women were lost to follow up. At the end of the study one woman in the E2+NETA group and two in the placebo group did not wish to undergo gynaecological examination, and two women in the lowEP group and two in the placebo group did not wish to undergo endometrial biopsy. One woman in the lowEP group and one in the placebo group were not biopsied, owing to stenosis of the cervical canal. The amount of endometrial tissue for the biochemical analysis was insufficient for six participants in the highEP group, four in the lowEP group, five in the E2+NETA group, and five in the placebo group. Two samples were lost during the analysis, one from the E2+NETA group and one from the placebo group.
Table 1 shows the mean baseline values for the four groups of women who entered the study. The groups were comparable for age, menopausal age, and serum placental protein 14, but the group receiving lowEP had slightly higher serum oestradiol concentrations (ANOVA, P= 0.02). The baseline data of the women who left the study were not significantly different from the data of those who completed the study, except for the placebo group in which the age of the noncompleters (mean age 57.9 years) was lower than for the completers (mean age 60.1 years) (P= 0.02).
Table 1. Baseline data of the women who entered the study. Values are given as mean (SD) or geometric mean [± 1 SD]. HighEP = piperazine oestrone sulphate 1.5 mg + interrupted norethisterone 0.7 mg daily; LowEP = piperazine oestrone sulphate 0.75 mg + interrupted norethisterone 0.35 mg daily; E2+NETA = continuous 17β-oestradiol 2 mg + norethisterone acetat 1 mg daily.
E2+NETA (n= 50)
Placebo (n= 50)
Menopausal age (years)
Serum PP14 (μg/L)
Figure 1 shows the distribution (in percent) of the number of women according to the individual number of bleeding days in the first and second year of treatment with hormones. In the first year of treatment, no uterine bleeding was reported by 39% of the women in the highEP group and a total of 65% of the women experienced 12 days or less of bleeding. The number of women who had no bleeding increased to 62% in the second year of treatment and 77% had 12 days or less of bleeding. No bleeding was reported by a higher number of the women in the lowEP group. Thus 60% and 69% did not have any bleeding in the first and second year of treatment, respectively, and a total of 86% of the women reported 12 days or less with bleeding both in the first and the second year of treatment. In the women receiving E2+NETA bleeding did not occur in 48% in the first year of treatment rising to 61% in the second year of treatment. Similar to the highEP group, 65% had 12 days or less of bleeding in the first year of treatment and 81% in the second year of treatment. A high incidence of bleeding (i.e. more than 24 days per year) was reported by 31%, 6% and 32% of the women in the first year of treatment with highEP, lowEP and E2+NETA. In the second year of treatment this decreased to 19% and 13% in the highEP and E2+NETA groups, while remaining unchanged at 8% in the lowEP group.
The average number of bleeding episodes and bleeding days stratified according to the severity of bleeding in the four groups of women is given in Table 2. Bleeding episodes occurred most frequently in the highEP group with an average of 7.7 episodes per woman in the first year of treatment and a small decline to 6.4 episodes in the second year of treatment. The lowest incidence of bleeding episodes was found in the lowEP group with an average of 2.8 in the first year and 2.3 in the second year of treatment. In the E2+NETA group the number of bleeding episodes declined from an average of 4.6 episodes in the first year to 2.2 episodes in the second year, which was comparable to the incidence observed in the lowEP group. However, the number of bleeding days per bleeding episode was highest in the woman in E2+NETA group with an average of approximately five days per episode, as compared with approximately three days in the highEP and lowEP groups. The number of bleeding days declined from the first to the second year of treatment from an average of 25.9 to 16.7 days per woman per year in the highEP group, from 8.0 to 7.3 days in the lowEP group, and from 20.8 to 11.2 days in the E2+NETA group. In all three hormone groups spot bleeding accounted for most of the days with bleeding, and a gradual decrease in the number of days was reported for slight, moderate, and heavy bleeding. The decline seen in the number of bleeding days from the first to the second year of treatment in the highEP and E2+NETA groups was due to less spot bleeding. The number of days with slight, moderate, or heavy bleeding was virtually unchanged. Two of the women in the placebo group experienced one spot bleeding episode.
Table 2. Uterine bleeding (mean number of episodes and days per woman per year). Key as for Table 1.
No. of episodes
No. of days per episode
No. of days per woman
No. of episodes
No. of days per episode
No. of days per woman
Figure 2 gives the bleeding index defined as the number of days with bleeding multiplied by the weighted intensity of bleeding and calculated per woman per tablet cycle. After the first nine tablet cycles, the bleeding indexes were constant. The women in the highEP group had a bleeding index of about two, the women in the lowEP group about one; and the women in the E2+NETA had a bleeding index in between.
Figure 3 shows the mean values of serum PP14 before and during the two years of treatment in the four groups. Serum PP14 increased highly significantly from the baseline concentration of 3.3–3.5 μg/L (P < 0.001) in the hormone groups, whereas it remained unchanged in the placebo group. The difference between the three hormone groups and the placebo group was statistically significant (P < 0.01–0.001) at all time points during treatment. In the highEP group, the concentration of serum PP14 increased to 4.7 μg/L at three months of therapy, remained at the same concentration for the two years of therapy, and did not show a cyclical variation. In the lowEP group, the same pattern of serum PP14 was found. In the E2+NETA group, serum PP14 increased to 5.4 μg/L at 3 months and remained at the same concentration during the study period. The concentration of serum PP14 was significantly higher in the E2+NETA group than in the lowEP and highEP groups (P < 0.001; P < 0.01).
Table 3 classifies the endometrial biopsies according to histology. None of the endometria showed signs of malignancy or pre-malignancy. A proliferative phase endometrium was observed in one woman in the highEP group and in one woman in the E2+NETA group, and an interval phase endometrium was found in one woman in the lowEP group. Histology was comparable for all three hormone therapies (P= 0.56), with one woman in each group having a proliferative/interval phase endometrium, about 20% of the biopsied women had a secretory phase endometrium, and the rest had an atrophic endometrium. The histology produced by the three hormone therapies was significantly different from placebo (P= 0.02), in which group all the biopsied women were found to have an atrophic endometrium.
Table 3. Histological assessment of endometrial biopsies at two years of treatment. Values are given as n (%). Key as for Table 1.
HighEP (n= 26)
LowEP (n= 35)
E2+NETA (n= 31)
Placebo (n= 30)
After two years of treatment, the endometrial thickness was slightly but significantly increased in the E2+NETA group, as compared with the placebo group (P < 0.01), whereas the highEP and lowEP regimens did not induce increased thickness (Fig. 4). Endometrial thickness of > 6 mm was seen in three women in the lowEP group and two women in the E2+NETA group. Figure 5 shows the markers of the secretory phase endometrium (i.e. serum PP14 measured on the day of the endometrial biopsy and endometrial activity of ICDH) and the markers of oestrogen stimulation (i.e. endometrial concentrations of ERc and PgRc). The secretory endometrial protein PP14 was increased in all the hormone groups, compared with placebo (P < 0.001), and reflected the pattern of endometrial thickness. All three hormone treatments showed the same decreased activity of endometrial ICDH, compared with placebo (P < 0.01–0.001). Of the markers of oestrogen stimulation, the concentration of endometrial ERc was at the same low level in all three hormone treatment groups and was comparable to the concentration in the placebo group. Endometrial PgRc tended to be increased in the highEP and lowEP groups, compared with placebo (highEP: P < 0.07; lowEP: P < 0.08) which indicates a slight oestrogen stimulation.
The present study was undertaken to investigate the effect of piperazine oestrone sulphate and interrupted norethisterone on endometrial status. A number of parameters, not normally required for regulatory purposes, was introduced to clarify their usefulness in evaluating endometrial safety (i.e. assessments of serum and endometrial tissue markers).
As regards endometrial protection against hyperplasia, we found that the two hormone regimens of piperazine oestrone sulphate and interrupted norethisterone offered the same, adequate endometrial protection as did the continuous combined regimen. Histological assessment showed no development of endometrial hyperplasia during the two-year study period. Only one biopsy in the highEP group and one in the E2+NETA group was proliferative, whereas the rest were, as expected, atrophic or secretory. As measured by ultrasound, the endometrium remained low in all three hormone groups, although a slight increase in endometrial thickness was seen in the E2+NETA group.
The use of interrupted norethisterone attempts to make progestogen administration more efficient by taking advantage of the dynamics of steroid receptors. In the pre-menopausal endometrium, oestrogen and progesterone receptors are up-regulated in the oestrogen-dominated phase of the menstrual cycle and down-regulated in the progesterone-dominated phase. The same pattern of receptor regulation is observed during sequentially combined HRT16. In vitro studies have shown that maximum levels of progesterone receptors are reached after a period of three to four days of oestradiol stimulation17. Consequent on progestogen administration, the level of progesterone receptors is lowered18, thus making further addition of progestogen less effective. It was thought that a treatment regimen consisting of short alternating phases of oestrogen and progestogen stimulation would benefit from the receptor dynamics, while avoiding uterine bleeding episodes. Maximum biological effects per dose of hormones ingested could thus be expected, owing to the receptor dynamics, in contrast to continuous combined HRT where both receptors would be constantly down-regulated. In the endometrial biopsies taken at the time of maximum oestradiol stimulation, we found the same level of oestrogen receptor in all three hormone groups, which was comparable to the level found in the atrophic endometrium of the placebo group. Moreover, in a recent study of transdermal oestrogen interruptedly combined with norethisterone acetate the same immunohistological staining intensity of oestrogen and progesterone receptors was observed in the oestrogenonly phase as in the combined oestrogen-progestogen phase19, which indicates minimal, if any, regulation of the receptors. However, although not statistically significant, we found that the progesterone receptors tended to be slightly increased in the groups receiving piperazine oestrone sulphate and interrupted norethisterone. It is possible that the three-day alternating time period may be too short to allow full response of the endometrium. Studies of postmenopausal HRT have shown that maximum progestogenic effects occurred after six days of progestogen administration16.
The secretory endometrial protein, PP14, is synthesised in the glandular epithelial cells of the secretory phase endometrium20. During sequentially combined HRT, the concentration of serum PP14 varies cyclically with the highest concentration around the onset of the menstrual-like bleeding21, and it has been shown that serum PP14 is strongly correlated with the secretory activity of the endometrium in postmenopausal women receiving HRT13. In the present study we investigated whether a putative cyclical variation in the endometrium during interruptedly combined HRT was reflected in the concentration of serum PP14, but none was found. During treatment, however, the concentration of serum PP14 was slightly increased in all three hormone groups, thus reflecting the low and slightly secretory phase endometrium. The other marker of the secretory phase endometrium, isocitrate dehydrogenase, measured in endometrial biopsies stayed at the same low level in all three hormone treatment groups. The high activity of isocitrate dehydrogenase in the placebo group was unexpected, and we have at present no data to explain this observation.
Despite the same ratio of norethisterone to piperazine oestrone sulphate in the two interrupted combined regimens, the higher dose of both hormones was associated with a more frequent occurrence and longer duration of uterine bleeding episodes. In the highEP group 22% of the women left the study because of uterine bleeding episodes, compared with 10% of the women in the E2+NETA group and none in the lowEP group. Despite the gradual decline in occurrence and duration in the women who completed the study, those receiving highEP had the most bleeding episodes, whereas those receiving lowEP had the fewest. Unacceptable control of bleeding (i.e. bleeding for more than 25 days per year in the second year of treatment) occurred in 19%, 8%, and 13% in the women receiving, respectively, highEP, lowEP, and E2+NETA.
The histological and biochemical assessments of the endometrium showed that the HRT regimens of oestrogen and interrupted progestogen induced the same pattern in endometrial parameters as the continuous combined HRT. We were unable to show any cyclical variation in the endometrium during the interrupted combined administration, but a tendency towards increased levels of progesterone receptors were observed. The lowest dose of 0.75 mg piperazine oestrone sulphate interruptedly combined with 0.35 mg norethisterone was associated with acceptable control of bleeding, but more unexpected uterine bleeding episodes occurred when the hormone doses were doubled.
This study was supported by a grant from the R. W. Johnson Research Institute, Raritan, New Jersey, USA and Zürich, Switzerland.