Most women experience vasomotor symptoms during menopause, and in a large proportion of women these symptoms have a significant, negative impact on quality of life1–3. Many women request treatment to alleviate or eliminate menopausal symptoms such as hot flushes and night sweats. A recent U.S. survey of health care providers indicated that 89% initiate menopausal hormone therapy (HT) in symptomatic patients compared with only 12% for prevention.4
Estrogen is the most effective therapy for relieving menopausal symptoms, particularly hot flushes.5–7 From the perspective of both the physician and the patient, it is desirable to use the lowest estrogen dose that provides adequate symptom relief, to minimize side effects and to maximize patient acceptance and continuation of treatment. This is reflected in current guidelines issued by U.S. and European menopausal societies, which recommend prescribing menopausal HT at the lowest possible dose for the shortest possible time.5,6,8–11 If the lowest estrogen dose is not efficacious, the dose can then be increased until desired symptom relief is achieved. A number of placebo-controlled trials have shown that low-dose oral and transdermal estrogens are effective in relieving vasomotor symptoms by 60% to 70%, with minimal adverse effects.7,12–18
The lowest transdermal dose of estrogen currently available (0.014 mg 17β-estradiol [E2] per day) significantly increases bone mineral density with a side-effect profile comparable to placebo, including vaginal bleeding and breast-related symptoms,19–20 but had not been investigated with respect to its effectiveness in relieving menopausal symptoms. The purpose of this study was to estimate the efficacy of this micro-dose transdermal estrogen in relieving vasomotor symptoms. In addition to measuring the mean reduction in moderate and severe hot flushes with HT as compared with placebo, we also estimated the proportion of women with at least a 75% and 90% reduction in the frequency of moderate to severe hot flushes.
MATERIALS AND METHODS
This was a randomized, double-blind, placebo-controlled, multi-center trial that compared a 2.2 mg E2/0.69 mg levonorgestrel combination transdermal patch and a 1.0 mg E2 transdermal patch with a placebo patch in postmenopausal women at 48 centers in the United States. The study was conducted in accordance with the Declaration of Helsinki, and it complied with Good Clinical Practice guidelines. Each individual site obtained ethical review board approval of the study protocol, and each participant provided written informed consent before study enrollment.
Investigators enrolled healthy women aged 40 years or older, who were postmenopausal (amenorrhea for 12 months or more, spontaneous amenorrhea for 6 months with serum follicle-stimulating hormone [FSH] greater than 40 milli-International Units/mL, 6 weeks postsurgical bilateral oophorectomy with or without hysterectomy) and had reported at least seven moderate or severe hot flushes per day for at least 1 week (7 consecutive days) or a minimum of 50 moderate or severe hot flushes per week for at least 1 week (7 consecutive days) during the screening phase. Exclusion criteria included treatment in the preceding 2 weeks with selective serotonin reuptake inhibitors, selective norepinephrine reuptake inhibitors, monoamine oxidases, alpha-adrenergics, raloxifene, tamoxifen, or herbal preparations. Subjects who had received prior HT were required to undergo wash-out: vaginal products, 1 week; transdermal estrogen or estrogen/progestin, 4 weeks; oral estrogen and/or progestin, 8 weeks; intrauterine progestin, 8 weeks; progestin implants and injectable estrogen alone, 3 months; estrogen pellet or injectable progestin, 6 months. Other exclusion criteria included abnormal Pap test results; abnormal genital bleeding; thrombophlebitis or thromboembolic disorder, or a history thereof.
Randomization numbers were generated in blocks using the SAS RANDO (SAS Institute Inc., Cary, NC) randomization tool. Women were randomly assigned at the baseline visit to one of three treatment groups in a 1:1:1 ratio within each block. To ensure proper randomization, a central telephone randomization was performed, with investigators calling an interactive voice-response system. Study medication was assigned to each participant according to the randomization code stored in the interactive voice-response system, and participants and investigators were blinded to group assignment until database lock. The double-blind study design was guaranteed by using treatments that were identical in appearance. Each subject wore two transdermal patches, one large patch (11 cm2) and one small patch (3.25 cm2), to comply with the double-blind study design. The first group of women was randomly assigned to receive a small transdermal patch (3.25 cm2, oval-shaped, transparent patch) containing E2 and a large placebo patch; the second group received a large transdermal patch (11 cm2, oval-shaped, transparent patch) containing E2/levonorgestrel and a small placebo patch; and the placebo group received one large and one small placebo patch. The nominal delivery rate for the E2-only patch was 0.014 mg/day and for the E2/levonorgestrel patch 0.023 mg/day and 0.0075 mg/day, respectively. Women were instructed to apply patches to intact, clean, dry skin on the lower abdomen, side by side but not in contact with each other. The patches were left in place for 7 days before being replaced by two new patches. This process was repeated for 12 weeks without any patch-free intervals. Study visits occurred as follows: first screening 4 weeks before study start, second screening 1 week before study start, baseline visit (day 0), week 4, week 12 (or premature discontinuation), end of study visit (2 weeks after end of study medication). Vital signs and body weight were measured at each visit. At the second screening visit, women also underwent a bilateral mammography, an electrocardiogram, a gynecologic examination that included transvaginal ultrasound, vaginal pH, a Pap test/cervical smear, a vaginal smear, and standard blood analyses including key hormone levels (eg, FSH, E2, thyroid-stimulating hormone). At week 12, in addition to vital signs, a general physical examination was performed plus a full gynecologic examination and blood analyses. At the end of study visit, vital signs, body weight, and the outcomes of any adverse events were recorded.
The coprimary efficacy variables were the mean changes in frequency and severity of moderate and severe hot flushes from baseline to week 4 and week 12. Baseline hot flush data were calculated using all available data from the screening phase. During treatment, diary cards were used to record the daily number and severity of hot flushes, bleeding patterns, and records of women's use of the small and large patches.
The frequency of moderate and severe hot flushes for each treatment week was calculated for each woman based on the daily frequencies from the days on which information was actually recorded. The change from baseline in frequency of moderate and severe hot flushes for each woman for each week was calculated by subtracting the baseline value from the corresponding weekly frequency. The mean changes from baseline to week 4 and week 12 for a given treatment arm were calculated by averaging the changes from baseline to weeks 4 and 12, respectively, for all women in this treatment arm.
The severity of hot flushes was recorded using the following scale: 0=no symptoms, 1=mild (sensation of heat without perspiration), 2=moderate (sensation of heat with perspiration, able to continue activity), and 3=severe (sensation of heat with perspiration, causing the subject to stop activity or awaken from sleep). Severity of hot flushes was described by a severity score calculated each day:
Severity score=(2×number of moderate hot flushes)+(3×number of severe hot flushes)/total number of moderate and severe hot flushes that day
If no moderate or severe hot flush was reported on a particular day, the mean severity score was 0 for that day. The change from baseline in mean severity of hot flushes was calculated each day for each subject by subtracting the baseline value from the corresponding mean daily severity. The change from baseline in severity of hot flushes for each subject for each week was calculated by subtracting the baseline value from the corresponding weekly frequency. The mean change from baseline to week 4 and week 12 in the severity of hot flushes for a given treatment arm was calculated by averaging the changes from baseline to weeks 4 and 12, respectively, for all subjects in this treatment arm.
Secondary efficacy variables included the mean change from baseline to week 12 in vaginal pH and vaginal maturation index, the proportion of subjects with symptoms of vulvar and vaginal atrophy and urogenital symptoms at baseline and week 12, and the proportion of subjects with moderate-to-severe symptoms of vulvar and vaginal atrophy (data not shown, to be reported elsewhere).
As a result of considerable interindividual variation in the treatment response to estrogen therapy, mean values for the frequency of moderate and severe hot flushes are of limited value for the interpretation of clinical benefit. To address this issue with respect to micro-dose estrogen, we undertook a post hoc responder analysis to estimate the percentage of women with 75% or more and 90% or more reduction in the frequency of moderate and severe hot flushes from baseline to the week 12 endpoint.
Based on the findings of a previous 12-week study that evaluated relief of hot flushes, it was estimated that a sample size of 105 women per group would provide greater than 80% power to detect significant differences at the 5% level between treatments in the primary efficacy variables. Since the dose–response relationship of placebo is less than or equal to E2 which is less than or equal to E2/levonorgestrel was assumed in advance, we followed a closed-test procedure by testing hierarchically low-dose E2/levonorgestrel compared with placebo, and only if this was significant proceeded with micro-dose E2 compared with placebo. This approach does not inflate the type I error, and no type I error adjustment for multiple comparisons was necessary. Data from the full analysis set were used to evaluate the efficacy and safety variables. The full analysis set consisted of all randomized subjects who applied at least one patch of study medication.
Descriptive statistics are presented for continuous variables, and frequency distributions are presented for categorical variables. Missing values in the treatment phase were replaced using the last-observation-carried-forward approach. The terms “week 4 endpoint” and “week 12 endpoint” are used to indicate where missing data have been replaced using the last-observation-carried-forward approach for the primary analysis. For the responder analyses, the week 12 endpoint (last-observation-carried-forward) data were used, as for the primary efficacy variables. For the micro-dose E2 responder compared with nonresponder analysis and the responder analysis over time by dose, the week 12 data (completers, no last-observation-carried-forward) were applied.
An analysis of variance model (continuous variables) or a Cochran–Mantel–Haenszel test (categorical variables) was used to evaluate the differences between treatment groups at baseline. An analysis of covariance (ANCOVA) model with treatment and pooled centers as factors and the baseline measurement as a covariate was used to compare the change from baseline to the weeks 4 and 12 endpoints in moderate and severe hot flush frequency and severity between the treatment groups. Two pair-wise comparisons (E2/levonorgestrel and placebo, E2 and placebo) were performed in an ordered sequential testing procedure. If the Shapiro–Wilk test21 in the ANCOVA model for the normality of the residuals was significant at the.05 level, a nonparametric analysis, rank ANCOVA,22 was performed. Descriptive statistics are presented grouped according to treatment for both absolute values and changes from baseline values.
For the safety analysis, descriptive statistics were used to analyze vital signs, and treatment groups were compared with respect to the change from baseline to week 12 using ANCOVA. Descriptive statistics were used to analyze continuous laboratory variables (absolute values and change from baseline values).
A total of 425 women were enrolled in the study and randomly assigned to 1 of 3 treatment groups: 145 received low-dose E2/levonorgestrel; 147 women received micro-dose E2; 133 received placebo. The first visit of the first participant was in December 2004 and the last visit by the last participant was in November 2005. The disposition of women through the 12-week study is shown in Figure 1. The mean age was 52.7 years (range 40–71 years). A total of 58% of women in the study had undergone a hysterectomy, and 41% an oophorectomy (34% bilateral and 6% unilateral). There were no significant differences in the baseline demographic characteristics of the treatment groups (Table 1). Compliance to study treatment was high; mean compliance 96.2% for low-dose E2/levonorgestrel; 94.5% for micro-dose E2; 94.0% for placebo. Of the 425 women who were randomly assigned and received at least one patch (full analysis set), 67 (16%) withdrew from the study: 23 (16%) in the low-dose E2/levonorgestrel group; 28 (19%) in the micro-dose E2 group; and 16 (12%) in the placebo group. The reasons for premature discontinuation from the study are shown in Figure 1.
Of the 425 full-analysis set, eight women did not have hot flush data from their daily diary cards. Of the remaining 417 women, one in the micro-dose E2 group did not have baseline hot flush data, so change from baseline was not calculated. One woman in the placebo group did not have hot flush data up to week 4, so a total of 415 women were included in the full analysis set for the week 4 endpoint. As hot flush data were available for this woman from week 4 onward, the analysis of the week 12 endpoint included 416 women.
The mean weekly frequency of hot flushes at baseline was similar for the three treatment groups (Table 1). Since the Shapiro–Wilk test was significant at all time points, all P values were derived from a rank ANCOVA model. At the week 12 endpoint, the reductions in hot flush frequencies were: low-dose E2/levonorgestrel, −51.80 (P<.001 compared with placebo); micro-dose E2, −38.46 (P<.001 compared with placebo; covariate excluded from the model due to violation of model assumptions); placebo, −25.57 (Fig. 2). The reductions in frequency at the week 4 endpoint were: low-dose E2/levonorgestrel, −40.76 (P<.001 compared with placebo); micro-dose E2, −28.57 (P=.039 compared with placebo); placebo, −21.34 (Fig. 2).
The mean daily severity scores at baseline were similar for the three treatment groups (Table 1). Figure 3 shows the change from baseline in the mean daily severity scores over the 12-week study period in all groups. At week 12, the greatest reduction from baseline in the mean severity score was demonstrated by the low-dose E2/levonorgestrel group (−1.19; P<.001 compared with placebo) followed by the micro-dose E2 group (−0.81; P<.001 compared with placebo) then placebo (−0.35). The reductions in the mean severity score at the week 4 endpoint were: low-dose E2/levonorgestrel, −0.63 (P<.001 compared with placebo); micro-dose E2, −0.39 (P=.04 compared with placebo); placebo −0.22.
A post hoc responder analysis was undertaken to estimate more accurately the clinical benefit provided by micro-dose E2. At the week 12 endpoint (last-observation-carried-forward), the proportions of women demonstrating 75% or more and 90% or more reductions in the frequency of hot flushes from baseline were greater with low-dose E2/levonorgestrel compared with micro-dose E2 or placebo. Overall, 41.3% of women in the micro-dose E2 group demonstrated a 75% or more reduction in frequency from baseline compared with 24.2% in the placebo group (P=.003; Fig. 4). A 90% or more reduction in the frequency of hot flushes from baseline to week 12 was demonstrated by 35.0% of women in the micro-dose E2 group compared with 14.4% of placebo-treated women (P<.001; Fig. 4).
To develop a clinical recommendation for the management of nonresponders to micro-dose E2 treatment, we compared the change in hot flush frequency over time in micro-dose E2 responders (75% or more reduction in frequency from baseline at week 12) compared with nonresponders (Fig. 5). Responders started to separate at week 1. At week 4, there was a clear difference in mean values (but still some overlap of standard deviations). Responders were clearly distinct from week 8 onward. In micro-dose E2 responders, the mean reduction in hot flushes was more than 50% after 2 weeks, approximately 70% after 4 weeks, 90% after 8 weeks, and 95% after 12 weeks (Fig. 5).
Comparison of the changes from baseline in hot flush frequency only in responders in all treatment groups showed that the time course of response (onset of efficacy) was very similar between the active treatments (Fig. 6), although response rates are different. This indicates that if a woman responds to E2 she responds quickly, regardless of dose level.
A total of 233 (55%) women in the study reported at least one treatment-emergent adverse event during the study, and the proportion of women in each group was similar. The most common adverse event, as defined as those occurring in 5% or fewer of women in the low-dose E2/levonorgestrel and micro-dose E2 groups, are shown in Table 2. Breast-related adverse events like breast pain and breast tenderness were infrequent (5% or less), with no difference in frequency between the active treatment groups and placebo.
Serious adverse events were reported during the treatment phase in three (2%) women in the low-dose E2/levonorgestrel group, three (2%) subjects in the micro-dose E2 group, and three (2%) subjects in the placebo group; only one was considered possibly related to treatment (chest pain of unknown etiology in the low-dose E2/levonorgestrel group).
In women with a uterus, the mean number of bleeding/spotting days was higher in the low-dose E2/levonorgestrel (2.7) group than the micro-dose E2 (1.2), and placebo (0.5) groups (P=.018 between treatment groups). The mean number of bleeding/spotting episodes was also higher in the low-dose E2/levonorgestrel (0.8) than the micro-dose E2 (0.4) and placebo (0.3) groups (P=.023 between treatment groups). The mean length of episodes was comparable between treatment groups.
Micro-dose E2 (0.014 mg/day) was effective in relieving vasomotor symptoms in healthy postmenopausal women. As data on the efficacy of low-dose transdermal E2 for vasomotor symptom relief are currently limited to doses of at least 0.025 mg/d for skin patches17–20 and 0.0125 mg/d for transdermal gel,12 this study adds to current knowledge on HT. Results seen in the higher estrogen-dose group (E2 0.023 mg/d plus levonorgestrel 0.0075 mg/d) are also comparable with previous studies. In our study, and others, the side-effect profile of transdermal micro-dose E2 was comparable to that of placebo,19,20 and in a previous study, mammographic breast density did not increase with use of transdermal micro-dose E2 over 2 years.23
The efficacy of HTs in relieving hot flushes is usually reported as a mean reduction in frequency; however, mean values are not useful when assessing lower estrogen dosages as they do not indicate the proportion of women having a good response. Data from our responder analysis were consistent with the hypothesis that a certain threshold, which differs between individuals, has to be reached to achieve efficacy. Therefore a proportion of women will respond well to micro-dose E2, whereas others may need higher doses to reach their efficacy threshold. This hypothesis supports the concept of initiating therapy for symptom relief in all women at the lowest effective dose (ie, micro-dose) and then titrating up if necessary, as recommended by the U.S. Food and Drug Administration (FDA) or professional societies.5,6,8–11 If individual thresholds for the effect of estrogen therapy on hot flushes exist, they are likely to be age dependent, as hot flush frequency and severity tend to reduce with age. In addition, micro-dose E2 could be an appropriate choice for women wanting to restart HT if they have a recurrence of symptoms after stopping therapy.
By comparing the changes in hot flush frequency over time in micro-dose E2 responders compared with nonresponders, we showed that considerable relief is achieved by week 4, indicating that the first clinical judgment regarding efficacy is possible at this point. Responders were completely distinguishable from nonresponders by week 8; therefore, the final decision as to whether a patient is a responder (at least 70% reduction in hot flushes from baseline) or not may be made at this time. If no adequate treatment effect is achieved by week 8, an estrogen dose increase is likely to be appropriate.
In addition to the 35% complete responder rate to micro-dose E2 in this study, the onset of efficacy data was striking. Although responder rates were higher in the E2/levonorgestrel group, there was no meaningful difference in onset of efficacy between the treatment groups, indicating that onset was independent of dose. Thus, if a woman responded to micro-dose E2, she responded as rapidly as she would have to the higher estrogen dose.
Although this study population was clinically overweight, we saw no effect of weight on efficacy. In addition, micro-dose E2 was equally effective in women with and without a hysterectomy. This trial enrolled only highly symptomatic women in accordance with FDA guidelines. The majority of women have hot flushes that are mild to moderate in intensity, with only 10–15% of women having very frequent and/or severe hot flushes.24 It is therefore likely that, in clinical practice, women seeking relief from hot flushes will generally be less symptomatic than those enrolled in this trial, raising the possibility that the responder rates with micro-dose E2 will be higher in clinical practice.
It would be interesting to ascertain whether certain subpopulations of women have an even higher responder rate to micro-dose E2 (eg, depending on age, time from menopause, degree of symptoms, restarting therapy after HT discontinuation). This study did not provide sufficient power to address these issues. A potential weakness of this study is that the higher E2 dose was administered together with levonorgestrel for endometrial protection. It could be argued that this might have influenced the E2 dose-response evaluation. However, the focus of the study was the comparison of micro-dose E2 and placebo. In addition, levonorgestrel has been shown to have no influence on hot flushes in the presence of E2.25 The results obtained here with the E2/levonorgestrel patch correlate well with those seen with patches containing E2 alone.17
This study supports micro-dose transdermal E2 (0.014 mg/d) as the lowest effective dose for the relief of moderate and severe vasomotor symptoms in healthy postmenopausal women. Transdermal micro-dose E2 may therefore be a valuable therapy for many women initiating menopausal HT, combining effective symptom relief with minimal side effects and maximum patient acceptance.
1. Hickey M, Davis SR, Sturdee DW. Treatment of menopausal symptoms: what shall we do now? Lancet 2005;366:409–21.
2. Bachmann GA. Menopausal vasomotor symptoms: a review of causes, effects and evidence-based treatment options. J Reprod Med 2005;50:155–65.
3. Bachmann GA. Vasomotor flushes in menopausal women. Am J Obstet Gynecol 1999;180:S312–6.
4. Rolnick SJ, Jackson J, Kopher R, Defor TA. Provider management of menopause after the findings of the Women's Health Initiative. Menopause 2007;14:441–9.
5. National Institutes of Health State-of-the-Science Conference Statement: Management of Menopause-Related Symptoms. Ann Intern Med 2005;142:1003–13.
6. Skouby SO, Al-Azzawi F, Barlow D, Calaf-Alsina J, Ertungealp E, Gompel A, et al. Climacteric medicine: European Menopause and Andropause Society (EMAS) 2004/2005 position statements on peri- and postmenopausal hormone replacement therapy. Maturitas 2005;51:8–14.
7. MacLennan AH, Broadbent JL, Lester S, Moore V. Oral oestrogen and combined oestrogen/progestogen therapy versus placebo for hot flushes. Cochrane Database Syst Rev 2004 Oct 18:CD002978.
8. North American Menopause Society. Estrogen and progestogen use in peri- and postmenopausal women: March 2007 position statement of The North American Menopause Society. Menopause 2007;14:168–82.
9. International Menopause Society 2007. Recommendations on postmenopausal hormone therapy, February 27, 2007. [Update to the original IMS Position Statement in Climacteric 2004;7:8–11.] Available at: http://www.imsociety.org
. Retrieved April 16, 2007.
11. Center for Drug Evaluation and Research. Guidance for industry: estrogen and estrogen/progestin drug products to treat vasomotor symptoms and vulvar and vaginal atrophy symptoms–recommendations for clinical evaluation. Rockville (MD): U.S. Department of Health and Human Services, Food and Drug Administration; January 2003.
12. Simon JA, Bouchard C, Waldbaum A, Utian W, Zborowski J, Snabes MC. Low dose of transdermal estradiol gel for treatment of symptomatic postmenopausal women: a randomized controlled trial. Obstet Gynecol 2007;109:588–96.
13. Carranza-Lira S, Gooch AL, Velasco-Diaz G, Solano J, Arzola-Paniagua A. Low and ultra low-dose estrogen therapy for climacteric symptom control: preliminary report. Int J Fertil Womens Med 2006;51:171–5.
14. Utian WH, Gass ML, Pickar JH. Body mass index does not influence response to treatment, nor does body weight change with lower doses of conjugated estrogens and medroxyprogesterone acetate in early postmenopausal women. Menopause 2004;11:306–14.
15. Shulman LP, Harari D. Low-dose transdermal estradiol for symptomatic perimenopause. Menopause 2004;11:34–9.
16. Utian WH, Shoupe D, Bachmann G, Pinkerton JV, Pickar JH. Relief of vasomotor symptoms and vaginal atrophy with lower doses of conjugated equine estrogens and medroxyprogesterone acetate. Fertil Steril 2001;75:1065–79.
17. Utian WH, Burry KA, Archer DF, Gallagher JC, Boyett RL, Guy MP, et al. Efficacy and safety of low, standard, and high dosages of an estradiol transdermal system (Esclim) compared with placebo on vasomotor symptoms in highly symptomatic menopausal patients: The Esclim Study Group. Am J Obstet Gynecol 1999;181:71–9.
18. Speroff L, Whitcomb RW, Kempfert NJ, Boyd RA, Paulissen JB, Rowan JP. Efficacy and local tolerance of a low-dose, 7-day matrix estradiol transdermal system in the treatment of menopausal vasomotor symptoms. Obstet Gynecol 1996;88:587–92.
19. Ettinger B, Ensrud KE, Wallace R, Johnson KC, Cummings SR, Yankov V, et al. Effects of ultralow-dose transdermal estradiol on bone mineral density: a randomized clinical trial. Obstet Gynecol 2004;104:443–51.
20. Johnson SR, Ettinger B, Macer JL, Ensrud KE, Quan J, Grady D. Uterine and vaginal effects of unopposed ultralow-dose transdermal estradiol. Obstet Gynecol 2005;105:779–87.
21. D'Agostino RB, Stephens MA. Goodness-of-fit-techniques. New York (NY): Marcel Dekker; 1986.
22. Conover WJ, Iman RL. Rank transformations as a bridge between parametric and nonparametric statistics: The American Statistician. 1981;35:124–9.
23. Grady D, Vittinghoff E, Lin F, Hanes V, Ensrud K, Habel LA, et al. Effect of ultra-low-dose transdermal estradiol on breast density in postmenopausal women. Menopause 2007;14:391–6.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
25. Shulman LP, Yankov V, Uhl K. Safety and efficacy of a continuous once-a-week 17beta-estradiol/levonorgestrel transdermal system and its effects on vasomotor symptoms and endometrial safety in postmenopausal women: the results of two multicenter, double-blind, randomized, controlled trials. Menopause 2002;9:195–207.