Efficacy of Vaginal Estradiol or Vaginal Moisturizer vs Placebo for Treating Postmenopausal Vulvovaginal Symptoms: A Randomized Clinical Trial

Key Points

Question  Does 12-week treatment with vaginal 10-μg estradiol tablet or vaginal moisturizer improve postmenopausal vulvovaginal symptoms more than placebo?

Findings  In a randomized clinical trial of 302 postmenopausal women with moderate-to-severe vulvovaginal symptoms, vaginal 10-μg estradiol tablet plus placebo gel and vaginal moisturizer plus placebo tablet were not more efficacious than dual placebo at reducing symptom severity or improving sexual function.

Meaning  Shared decision making for treatment of postmenopausal vulvovaginal symptoms can be based on cost and patient formulation preference; vaginal estradiol tablets appear not to add benefit beyond vaginal gel or moisturizer.

Importance  Nearly half of postmenopausal women report bothersome vulvovaginal symptoms, but few data support the efficacy of 2 commonly recommended treatments.

Objective  To compare the efficacy of a low-dose vaginal estradiol tablet and a vaginal moisturizer, each vs placebo, for treatment of moderate-to-severe postmenopausal vulvovaginal symptoms.

Design, Setting, and Participants  This 12-week multicenter randomized clinical trial enrolled postmenopausal women with moderate to severe symptoms of vulvovaginal itching, pain, dryness, irritation, or pain with penetration.

Interventions  Vaginal 10-μg estradiol tablet (daily for 2 weeks, then twice weekly) plus placebo gel (3 times a week) (n = 102) vs placebo tablet plus vaginal moisturizer (n = 100) vs dual placebo (n = 100).

Main Outcomes and Measures  The main outcome was decrease in severity (0-3) of most bothersome symptom (MBS) between enrollment and 12 weeks. Additional measures included a composite vaginal symptom score, Female Sexual Function Index (FSFI) score (2-36), modified Female Sexual Distress Score–Revised item 1, treatment satisfaction and meaningful benefit, Vaginal Maturation Index, and vaginal pH.

Results  The 302 women had a mean (SD) age of 61 (4) years and were primarily white (267 [88%]), college educated (200 [66%]), and sexually active (245 [81%]). Most women (294 [97%]) provided data for the primary analysis. The most commonly reported MBS was pain with vaginal penetration (182 [60%]), followed by vulvovaginal dryness (63 [21%]). Mean baseline MBS severity was similar between treatment groups: estradiol, 2.4 (95% CI, 2.3 to 2.6); moisturizer, 2.5 (95% CI, 2.3 to 2.6); placebo, 2.5 (95% CI, 2.4 to 2.6). All treatment groups had similar mean reductions in MBS severity over 12 weeks: estradiol, −1.4 (95% CI, −1.6 to −1.2); moisturizer, −1.2 (95% CI, −1.4 to −1.0); and placebo, −1.3 (95% CI, −1.5 to −1.1). No significant differences were seen between estradiol (P = .25) or moisturizer (P = .31) compared with placebo. Mean total FSFI improvement was similar between estradiol (5.4; 95% CI, 4.0 to 6.9) and placebo (4.5; 95% CI, 2.8 to 6.1) (P = .64), and between moisturizer (3.1; 95% CI, 1.7 to 4.5) and placebo (P = .17).

Conclusions and Relevance  Our results suggest that neither prescribed vaginal estradiol tablet nor over-the-counter vaginal moisturizer provides additional benefit over placebo vaginal tablet and gel in reducing postmenopausal vulvovaginal symptoms.

Trial Registration  clinicaltrials.gov Identifier: NCT02516202

An estimated 40% to 54% of postmenopausal women report bothersome vulvovaginal symptoms,^1-3 including vaginal dryness (up to 75%) and pain with intercourse (40%).^2,4,5 In 2014, the North American Menopause Society coined the term genitourinary syndrome of menopause (GSM) to reflect the multifaceted nature of this prevalent problem, replacing genitourinary and vulvovaginal atrophy.⁶ Recent evidence shows decrements in quality of life from moderate to severe vulvovaginal symptoms in women aged 40 to 75 years comparable to those caused by other chronic conditions such as arthritis, chronic obstructive pulmonary disease, and irritable bowel syndrome.¹ Yet more than half of symptomatic women are not using any medication to treat their symptoms.^7,8

Recommendations for treatment of GSM focus primarily on vaginal products.^9,10 However, issues with recommended vaginal treatments include messiness, expense, safety concerns, and lack of symptom relief.¹¹ In postmenopausal women with vaginal dryness, itching, pain, or burning, meta-analyses of randomized trials conclude that vaginal estrogen cream use reduces symptoms in the majority of women,^12,13 but few women use them beyond 6 months.¹⁴ Four randomized clinical trials have assessed vaginal estrogen tablet efficacy for GSM,^15-18 but only 2 industry-sponsored trials have evaluated the current 10-μg product.^15,17 Although clinicians often recommend vaginal moisturizers,⁷ few studies exist to support this recommendation.^19-21

Surveys of postmenopausal women demonstrate a preference for effective, nonhormonal therapies, often due to safety concerns.²² We aimed to evaluate the efficacy of low-risk therapies: vaginal estradiol tablets and a vaginal moisturizer in women with moderate to severe vulvovaginal symptoms. We hypothesized that the vaginal estradiol tablet is more efficacious than placebo tablet and that a vaginal moisturizer is more efficacious than placebo gel, in the relief of postmenopausal vaginal symptoms.

This randomized, double-blind, placebo-controlled 12-week trial was conducted at 2 centers: Kaiser Permanente Washington Health Research Institute in Seattle and University of Minnesota in Minneapolis. We compared treatment efficacy for moderate to severe vulvovaginal symptoms between 10-μg vaginal estradiol tablets, a vaginal moisturizer, and matching placebos for each. The study was approved by institutional review boards at participating institutions. Participants provided written informed consent. Enrollment began in April 2016, targets were achieved by February 2017, and final follow-up visits occurred in April 2017.

Women were recruited through direct mailings and Facebook ads targeted to women aged 50 to 70 years within 20 miles of the clinical sites. Inclusion criteria were as follows: age 45 to 70 years, at least 2 years since last menses, report of at least 1 moderate to severe symptom of vulvovaginal itching, pain, irritation, or dryness experienced at least weekly within the past 30 days; or pain with penetration at least once monthly. Exclusion criteria included current vaginal infection, use of hormonal medication in past 2 months, use of antibiotics or vaginal moisturizer in past month, and chronic premenopausal vulvovaginal symptoms. The study protocol (Supplement 1) provides additional details of study procedures.

Randomization by permuted blocks of 9 and stratified by site was conducted via secure web-based database, and implemented by a computerized inventory system for dispensing identical-appearing tablets in bottles and gel in tubes. Participants, study personnel, and clinicians were blinded to treatment assignments.

Women were randomly assigned 1:1:1 to Vagifem 10-μg tablet + placebo vaginal gel, placebo vaginal tablet + Replens vaginal moisturizer, or placebo tablet + placebo gel. The active ingredient for tablets is 10.3 μg of estradiol hemihydrate, equivalent to 10 μg of estradiol. Placebo tablets contained inactive ingredients identical to Vagifem. The ingredients in Replens are purified water, glycerin, mineral oil, polycarbophil, carbomer homopolymer type B, hydrogenated palm oil glyceride, sorbic acid, and sodium hydroxide. The placebo was hydroxyethylcellulose gel, shown to have minimal effect on vaginal microbiota and inflammation.^23,24 Placebo gel varied slightly from Replens in viscosity, 13 800 centipoise, and pH, 4.5 (Replens, 13 000 centipoise and pH 3.0). Study medications were formulated and/or packaged by Sharp Clinical Services.

Women were instructed to use the vaginal tablet daily for 2 weeks, then twice weekly for the remaining 10 weeks, and the vaginal moisturizer every 3 days throughout the trial. During the first 2 weeks, participants were advised to use the tablet in the morning and gel in the evening. After that, participants were instructed to use products on alternate days.

Telephone contact at 1, 3, and 11 weeks after randomization assessed protocol adherence and adverse events. Follow-up visits were conducted 4 and 12 weeks after randomization. At each visit, women completed questionnaires and underwent vaginal sample collection for wet mount evaluation, pH measurement, and vaginal maturation index (VMI, a measure of vaginal mucosal cell maturation due to estrogen effects)²⁵ (at baseline and 12 weeks). At follow-up visits, women were asked to bring medications; remaining pills were counted and gel tubes weighed to provide medication adherence estimates.

The primary outcome was severity of the most bothersome symptom (MBS) defined by the participant at trial enrollment as vulvovaginal itching, pain, dryness, irritation, or pain with penetration. Severity was rated 0 to 3, signifying none, mild, moderate, or severe.²⁶ At each visit, women completed a questionnaire about presence and severity of vulvovaginal symptoms.

Prespecified secondary outcomes were composite Vaginal Symptom Index (VSI),¹⁵ satisfaction with treatment received (Likert scale: 0 = not satisfied to 10 = completely satisfied), meaningful benefit from the study medications (yes or no), Female Sexual Function Index (FSFI),²⁷ Female Sexual Distress Scale–Revised Item 1,²⁸ VMI,^10,26 and pH.²⁶ Vaginal Symptom Index was the mean severity score of the 5 vulvovaginal symptoms listed as MBS choices. Post hoc secondary outcomes were severity of pain with penetration, and vaginal dryness.²⁹ Additional questionnaires included Menopausal Quality of Life,³⁰ Generalized Anxiety Disorder 7,³¹ and Patient Health Questionnaire 8.³²

Adverse events were assessed at each visit by a questionnaire listing symptoms potentially related to active agents (increased vaginal secretions, vaginal itching, breast tenderness, vaginal bleeding, vulvovaginal skin rash). Any new complaints reported at visits or by telephone were evaluated and classified according to the Medical Dictionary for Regulatory Activities.³³

Ninety-five women per group provided 89% power to detect an effect size of 0.5 standard deviation (SD) units change from baseline to week 12 in MBS severity between intervention group and placebo,¹⁹ based on a t test with a 2-sided α = .025 to account for 2 treatment group comparisons. The planned enrollment of 318 participants allowed for 10% loss to follow-up. The modified intent-to-treat analysis included all randomized participants who provided a baseline MBS score and corresponding vulvovaginal symptom severity at week 4 or 12, regardless of adherence to treatment assignment.

The primary outcome was change in severity of the MBS between enrollment and weeks 4 and 12. Treatment group differences were assessed by repeated-measures linear regression models of the 4- and 12-week continuous outcome measures (MBS, VSI, pain with penetration, vaginal dryness, VMI as percent superficial cells, FSFI) as a function of randomization assignment, baseline value of the outcome measure, visit, and clinical site. Robust standard errors were estimated via generalized estimating equations to adjust for correlation between repeated outcome measures. To facilitate comparisons, models of MBS severity were reanalyzed among participants meeting eligibility criteria for previously published trials: baseline pH greater than 5 and VMI with no more than 5% superficial cells.^15-17 Additional analysis evaluated intervention effects in models including only women adherent to treatment, defined as using at least 80% of medication. Two variables were hypothesized a priori to modify treatment response on the primary outcome: age and years since menopause. Tests for interaction between these variables and treatment assignment were performed within the linear regression models.

Treatment group differences at week 12 in proportions of women with medication adherence, at least 2-point drop in MBS severity, at least 50% decrease in MBS severity, vaginal pH of 5 or less, VMI greater than 5% superficial cells, sexual distress, and meaningful benefit from study medication use were assessed via χ² tests. Adverse events were compared between treatment groups via Fisher exact tests. Week 12 differences in treatment satisfaction were evaluated by t tests. Analyses were conducted using SAS, version 9.4 (SAS Institute), with 2-sided P ≤ .025 considered statistically significant for the primary outcome and P ≤ .05 for secondary outcomes.

Three hundred two women were randomized to receive vaginal estradiol tablet plus placebo gel (n = 102), placebo tablet plus vaginal moisturizer (n = 100), or dual placebo (n = 100). Study retention was high: 294 of 302 (97%) women provided primary analysis data (Figure 1). Most women were between 55 and 64 years old (235 [78%]), white (267 [88%]), and married or partnered (257 [85%]). Baseline characteristics were comparable between treatment groups (Table 1).

A total of 182 (60%) women endorsed MBS as pain with vaginal penetration, 63 (21%) dryness, 20 (7%) itching, 19 (6%) irritation, and 14 (5%) pain. Baseline mean (SD) MBS severity was 2.5 (0.6). The majority of women (245 [81%]) were sexually active: 202 (67%) with a male partner, 3 (1%) with a female partner, and 136 (45%) self-stimulation. Baseline median vaginal pH was 7 (interquartile range, 6.0-7.5). Of 269 (89%) baseline VMI samples with adequate cellularity for analysis, 91% had no more than 5% superficial cells.

At completion, of participants who returned tablets for counting (263 [87%]) and gel for weighing (259 [86%]), 94% were tablet adherent and 90% were gel adherent (ie, used >80% of medication doses). Adherence did not vary significantly across treatment groups (estradiol tablets, 84 of 102 [82%] vs placebo tablets, 80 of 100 [80%]; P = .67; vaginal moisturizer, 75 of 100 [75%] vs placebo gel, 78 of 100 [78%]; P = .62).

Neither treatment reduced MBS severity between baseline and 4 or 12 weeks more than placebo (Table 2 and Figure 2). All groups had a mean 1.2- to 1.4-point decrease from baseline MBS score by 12 weeks. A decrease of 2 points signifies a clinically meaningful change from moderate to severe symptoms to mild to none. There was no difference between intervention vs placebo groups in proportion of women with a decrease of at least 2 points in MBS severity between 0 and 12 weeks (estradiol, 47 [49%] vs placebo, 43 [45%]; P = .61; moisturizer, 35 [35%] vs placebo, 43 [47%]; P = .16). Most women had a decrease of at least 50% in symptom severity (estradiol, 67 [70%] vs placebo, 62 [65%]; P = .50; moisturizer, 53 [54%] vs placebo, 62 [65%]; P = .10). Mean VSI decreased less than 1 point from a mean of 1.6 in all treatment groups.

A higher proportion of women in the estradiol group had a pH change from greater than 5 at baseline to 5 or less at week 12 compared with placebo (36 [46%] vs 10 [12%]; P < .001); no difference was observed between moisturizer and placebo (8 [9%] vs 10 [12%]; P = .60). More women in the estradiol tablet group increased VMI superficial cells from 5% or less at baseline to greater than 5% at week 12 compared with placebo (45 [57%] vs 8 [11%]; P < .001). The same proportion of women in the moisturizer and placebo groups had an increase (8 [11%]; P = .95).

Change in FSFI did not significantly vary between treatment groups, either total score or any of the 6 domains (Table 3). The FSFI domain with the greatest improvement at 12 weeks was Lubrication, increasing by a mean of 1.4 (95% CI, 1.1-1.8) points in the estradiol + placebo gel group, 1.2 (95% CI, 0.8-1.6) points in dual placebo, and 0.9 (95% CI, 0.6-1.3) points in moisturizer + placebo tablet. Most women were “frequently” or “always” distressed about sex life at enrollment (Table 1). At 12 weeks, nearly half of women in the estradiol and placebo groups endorsed “rarely” or “never” distressed (47 [47%] estradiol, 29 [29%] moisturizer, 41 [43%] placebo; estradiol vs placebo, P = .50; moisturizer vs placebo, P = .05). Mean (SD) treatment satisfaction was similar between groups: 8.6 (2.6) for estradiol tablet, 7.7 (3.2) for moisturizer, and 8.1 (3.0) for placebo. More women in the estradiol tablet group reported “meaningful benefit” from treatment than placebo (79 [80%] vs 62 [65%]; P = .02), but no difference was observed between moisturizer and placebo (57 [58%] vs 62 [65%]; P = .39).

In regression models including only medication-adherent women, changes in MBS severity and FSFI did not differ from the intent-to-treat analysis (eTables 1 and 2 in Supplement 2). In analysis limited to women meeting enrollment criteria for previous trials of pH greater than 5 and no more than 5% superficial cells on VMI (n = 205), we saw no difference in results. Neither age nor years since menopause modified response to estradiol, although women younger than 60 years demonstrated greater MBS improvement with placebo than with moisturizer (eTable 3 in Supplement 2).

Vaginal candidiasis was diagnosed by microscopy in 5 (5%) participants randomized to estradiol, 2 (2%) to moisturizer, and 2 (2%) to dual placebo. Two additional women reported a yeast infection diagnosed elsewhere. Adverse events were not different between the treatment groups (eTable 4 in Supplement 2). Three women randomized to estradiol received a diagnosis of cancer (all judged unrelated to study medication): 1 with lymphoma at 4 weeks withdrew, 1 with breast cancer at 4 weeks stopped therapy but continued other procedures, and 1 breast cancer diagnosis was made after study completion.

In this randomized clinical trial of 302 women with moderate to severe postmenopausal vulvovaginal symptoms, no treatment group differences in symptom reduction were observed for vaginal estradiol tablet plus placebo gel vs dual placebo, or vaginal moisturizer plus placebo tablet vs dual placebo. The lack of efficacy of the active treatment groups over dual placebo was similar whether women chose pain with vaginal penetration, vaginal dryness, or other symptoms as their MBS. We demonstrated similar improvement in symptoms and sexual function in all 3 treatment groups.

The North American Menopause Society recommends nonhormonal vaginal therapies as first-line treatment for GSM,⁹ while recommendations in Europe support vaginal estrogen therapy.¹⁰ Many women report substantial concerns about the long-term safety of hormonal products, and prefer to use nonhormonal products.²² Vaginal moisturizers such as Replens, containing mucoadhesives such as polycarbophil to extend benefit from intermittent dosing,³⁴ have been evaluated in small, open-label studies using the same dosing strategy as in our trial. Two studies, including 1 randomized crossover trial in breast cancer survivors, demonstrated no significant difference between placebo and moisturizer products, although both decreased symptom severity.^21,35 A recent study comparing dehydroepiandrosterone (DHEA) in a mucoadhesive base vs base alone demonstrated no difference in symptom improvement, but better sexual function in women using DHEA.²¹ We chose our placebo gel because it was shown not to alter vaginal microbiota or inflammation^23,24 and is a formulation with less mucoadhesive properties than Replens,³⁴ although of similar viscosity and pH.²³ The effectiveness of our placebo in decreasing symptom severity suggests that the mucoadhesive properties lauded by vaginal moisturizer manufacturers may not be necessary to achieve symptom relief.

Two of 3 randomized clinical trials of low-dose vaginal estrogen therapy demonstrated greater decrease in symptom severity vs placebo, although in all trials symptoms diminished in both groups.^17,36,37 Symptom reduction with estradiol tablets in our study was comparable to the existing literature for vaginal estrogen, newly approved vaginal DHEA, and oral ospemifene, a selective estrogen receptor modulator.^38,39 Our participants using vaginal estradiol tablets had a mean decrease in symptom severity from baseline of 1.4 points, similar to the 1.2 to 1.3 seen in prior studies of the vaginal tablet,^15,17 and the 1.4 points seen with oral ospemifene³⁸ and vaginal DHEA,³⁹ but slightly less than that seen with a softgel formulation (mean [SD], 1.69 [0.07]).³⁷ While 10 μg is a low dose, 2 studies comparing 10- and 25-μg doses did not show a difference in efficacy.^15,37 The differential change in VMI in our estradiol group demonstrates the biologic effect of estrogen vs placebo but was not linked to differences in symptom improvement. Overall, the largest difference between our trial and others is the magnitude of symptom improvement in our placebo group. Our placebo was quite different from placebo creams and tablets used in other trials of vaginal estrogen, and meets many of the criteria outlined in a recent review as optimal for vaginal moisturizing products.⁴⁰

The placebo effect in treatment trials for postmenopausal vaginal symptoms is substantial. In previous trials, placebo tablets were associated with a mean decrease in symptom severity of 0.8 to 0.87 points,^15,17 while placebo softgel was associated with a mean decrease of 1.28 points in dyspareunia severity.³⁷ In 2 studies of vaginal estrogen cream, dyspareunia severity decreased a mean of 0.7 to 0.9 points in placebo groups.^36,41 The 1.3-point mean symptom severity decrease in our dual placebo group is larger than other trials’ placebo effects, with the exception of the softgel study (REJOICE).³⁷ The symptoms chosen as outcomes for trials—dryness, pain with intercourse, itching, and irritation—are the most commonly reported vulvovaginal symptoms among postmenopausal women,^3-5 and severity is correlated with sexual function scores, suggesting that they are relevant patient outcomes.⁴² However, the profound placebo response seen in our trial and others, not linked to changes in pH or VMI, suggests that many factors in addition to the local vaginal environment contribute to symptoms.

Many women prescribed vaginal therapy for vulvovaginal symptoms seem unsatisfied with treatment, as continuation rates are low.⁴⁴ In previous studies, adherence to tablets was better than vaginal creams.^45-47 Other reports suggest that women prefer tablet formulation to products that increase discharge or are messy, and would be willing to pay more for a tablet formulation.⁴⁸ However, medication cost is a substantial barrier for many women, and hormone-based therapies are expensive. A 1-month supply of vaginal estrogen cream, vaginal estrogen tablets, or newer US Food and Drug Administration–approved products marketed specifically for dyspareunia (ospemifene and vaginal DHEA) can cost between $82 and $200,^49,50 while Replens costs approximately $20. Our results suggest that most women can achieve greater than 50% reduction in symptom severity with regular, consistent use of a vaginal gel with lubricant properties and do not see added symptom improvement with vaginal estradiol.

To our knowledge, this is the first randomized clinical trial evaluating short-term (12 weeks) efficacy of recommended nonhormonal and hormonal vaginal therapies for postmenopausal vulvovaginal symptoms, and the only one with a dual placebo arm. We enrolled a large cohort of women with moderate to severe symptoms, with excellent participant retention and medication adherence. While our study was not designed to compare active interventions head to head, efficacy comparisons of vaginal estradiol and moisturizer with dual placebo in the same population provide insight into the relative benefit of each. The generalizability of our trial results is limited by the relatively homogenous population, despite enrolling at 2 geographically distinct sites and using 2 recruitment strategies. In contrast to studies limiting the population to women with high vaginal pH and/or 5% or less superficial cells,^15,17,36,43 we included all postmenopausal women reporting moderate to severe vulvovaginal symptoms, making our population more consistent with women presenting to primary care settings.

Our study demonstrates that a better understanding of the underlying mechanism of GSM is needed to guide efforts to improve vaginal treatment options. Many postmenopausal women with moderate to severe vulvovaginal symptoms can be treated with a nonprescription vaginal lubricating gel. However, not all gel formulations may have the same effects, and some women may prefer nongel formulations. Treatment choice should be based on individual patient preferences regarding cost and formulation.

Accepted for Publication: January 8, 2018.

Corresponding Author: Caroline M. Mitchell, MD, Vincent Center for Reproductive Biology, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (caroline.mitchell@mgh.harvard.edu).

Published Online: March 19, 2018. doi:10.1001/jamainternmed.2018.0116

Author Contributions: Dr Guthrie had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Mitchell, Reed, Newton, Ensrud, LaCroix, Guthrie.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Mitchell, Reed, LaCroix, Guthrie.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Larson, Guthrie.

Obtained funding: Reed, Newton, Ensrud, LaCroix, Guthrie.

Administrative, technical, or material support: Mitchell, Reed, LaCroix, Guthrie.

Study supervision: Mitchell, Reed, Newton, LaCroix, Guthrie.

Conflict of Interest Disclosures: Dr Mitchell is a consultant for Symbiomix Therapeutics. Dr Reed receives grant support from Bayer Pharmaceuticals. Dr LaCroix has served on a scientific advisory board for Sermonix, Inc. No other disclosures are reported.

Funding/Support: This study was funded by the National Institutes of Health/National Institute on Aging (5R01AG048209).

Role of the Funder/Sponsor: The sponsor had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: All of the opinions and conclusions reported herein are the authors’ own, and do not reflect the official position of the National Institute on Aging.