Design of a Randomized Clinical Trial of Perioperative Vaginal Estrogen Versus Placebo with Transvaginal Native Tissue Apical Prolapse Repair (Investigation to Minimize Prolapse Recurrence Of the Vagina using Estrogen: IMPROVE)

Abstract

Objectives:

To provide the rationale and design for a randomized, double-blind clinical trial of conjugated estrogen vaginal cream (applied for at least 5 weeks preoperatively and continued twice-weekly through 12 months postoperatively) compared with placebo in postmenopausal women with symptomatic pelvic organ prolapse undergoing a standardized transvaginal native tissue apical repair.

Methods:

Study population, randomization process, study cream intervention, masking of participants and evaluators, placebo cream manufacture, standardized surgical intervention, and collection of adverse events are described. The primary outcome of surgical success is a composite of objectively no prolapse beyond the hymen and the vaginal cuff descending no more than one-third the vaginal length; subjectively no sense of vaginal pressure or bulging; and no retreatment for prolapse at 12 months. Time-to-failure postoperatively will be compared in the two groups with continued surveillance to 36 months. Secondary outcomes assessed at baseline, preoperatively (i.e., after at least 5 weeks of study cream), and postoperatively at 6 month intervals include validated condition-specific and general quality-of-life metrics, overall impression of improvement, sexual function, vaginal atrophy symptoms, and body image. Challenges unique to this study include design and manufacture of placebo and defining and measuring study drug adherence.

Results:

Recruitment of 204 women is complete with 197 randomized. There have been 174 surgeries completed with 15 more pending; 111 have completed their 12 month postoperative visit.

Conclusions:

This trial will contribute evidence-based information regarding the effect of perioperative vaginal estrogen as an adjunct therapy to standardized transvaginal native tissue prolapse surgical repair.

ClinicalTrials.gov: NCT02431897

BACKGROUND:

Reconstructive surgeries for anterior and apical vaginal prolapse are performed transvaginally using native tissue as fixation points (e.g., uterosacral or sacrospinous ligaments) or abdominally/ laparoscopically using synthetic graft material. Unfortunately, reoperation for recurrent prolapse is not rare, and rates are highest with native tissue vaginal surgery.^1,2 The abdominal/ laparoscopic approach has tradeoffs of longer operating times, increased cost, and a risk of serious complications compared to transvaginal repairs.^2,3 Therefore, there is a need to identify adjuncts to native tissue prolapse repair that may improve outcomes.

While it is difficult to separate the effects of aging in general from the specific impact of menopause and declining estrogen levels, it is clear that the pelvic organs and their surrounding muscular and connective tissue support are estrogen-responsive.^4,5 Nonetheless, the effect of adjunctive intravaginal estrogen in POP management remains uncertain. A Cochrane review of estrogens for the treatment or prevention of POP suggests that evidence is too limited and that there is a need for studies randomizing to estrogen preparations for the prevention and management of POP before and after prolapse surgery.⁶

The primary aim of the Investigation to Minimize Prolapse Recurrence Of the Vagina using Estrogen (IMPROVE) is to determine if vaginal estrogen given pre-and postoperatively with standardized native tissue transvaginal prolapse repair surgery results in improved objective and subjective prolapse recurrence and retreatment rates one year postoperatively with continued annual surveillance up to 3 years. The purpose of this report is to detail trial design, selection of primary and secondary outcomes, surgery standardization, the manufacture of placebo cream, and early study conduct.

METHODS:

The IMPROVE trial is sponsored by the National Institute on Aging (NIA). The primary clinical site (UT Southwestern Medical Center, or UTSW) also houses the data coordinating center (DCC). There were originally consortium arrangements with 3 other clinical sites (i.e., 4 total clinical sites: UTSW, University of Alabama at Birmingham—UAB, Women & Infants Hospital of Rhode Island—WIHRI, and Baylor Scott & White of Temple, Texas), which subsequently reduced to 3 current sites: UTSW, UAB, and WIHRI. A final consortium agreement was made with a private drug manufacturing partner (DavosPharma, who in turn partnered with Particle Sciences, Inc.) that coordinated the manufacture and distribution of a high quality placebo cream. Pfizer, Inc provided active study cream (conjugated estrogen cream, Premarin®) through a separate investigator-initiated award; Pfizer had no role in the study design. A steering committee comprised of the principal investigators of the 3 active clinical sites plus the DCC chair oversees daily conduct of the trial, and a program officer from the NIA and an independent Data Safety Monitoring Board (DSMB) provide at least semiannual oversite. IMPROVE is registered with Clinicaltrials.gov: NCT02431897.

Study Population

The target population was women ≥48 years and at least one year postmenopausal (or with prior bilateral oophorectomy) with vulvovaginal atrophy exposed to no oral, vaginal, or transdermal medication containing estrogens (including selective estrogen receptor modulators, or SERMs), androgens, or progestins within 4 weeks of screening. They must have had symptomatic pelvic organ prolapse (≥stage 2 anterior or apical prolapse) planning elective transvaginal native tissue apical repair. Notable exclusions include BMI >35 kg/m², current tobacco or steroid use, any other medications affecting vaginal milieu, prior vaginal apical repair surgery or surgery using mesh for prolapse, and contraindications to estrogen therapy. Table 1 further details the inclusion and exclusion criteria. Eligible women who declined enrollment were characterized consistent with CONSORT recommendations.

TABLE 1.

Inclusion and Exclusion Criteria

Inclusion criteria

1. Postmenopausal, meeting criteria a, b, c, or d:

a. Still with uterus and no menses for >1 year

b. Status-post bilateral oophorectomy (i.e., surgically menopausal)

c. Status-post hysterectomy (with or without oophorectomy) and >55 years

d. Vaginal pH ≥5 if ≤55 years and status-post hysterectomy but still with one or both ovaries

2. Minimum age: 48 years

3. Symptomatic apical and/or anterior vaginal wall prolapse, stage 2 or greater

4. No estrogen or progesterone replacement within the last month (may come off current treatment, i.e., wash out, to join the study)

5. Medically fit for elective surgery

6. Physically able to apply/insert the study drug

7. Available for clinic follow-up for minimum 1 year

Exclusion criteria

1. Concurrent use of vulvar/ local steroid treatments for other indications (e.g., lichen sclerosis)

2. BMI >35 kg/m2

3. Recent history (within last month) of vaginal infection or vaginitis

4. Contraindications to estrogen therapy (e.g., spontaneous DVT, stroke, breast or endometrial/ hormone-responsive cancer, genital bleeding of unknown cause)

5. History of connective tissue disease

6. Any estrogen, progesterone, SERM, or other medication impacting vaginal milieu, by any route (oral, transdermal, subdermal/ implantable, intravaginal, intrauterine); may come off current treatment (>1 month), i.e., wash out, to join the study

7. History of vaginal irradiation

8. Allergy to Premarin® or its constituents

9. Prior apical repair or use of mesh for prolapse repair

10. Current tobacco use

BMI, body mass index; DVT, deep vein thrombosis; SERM, selective estrogen receptor modulator

Baseline Participant Assessment

After eligibility was determined, written informed consent approved by each site’s Institutional Review Board was obtained. Paper case report forms were used to collect baseline demographic data: age, race/ethnicity, and health insurance type. A medical history was obtained to confirm nonsmoker and postmenopausal status and to collect information on pregnancies, routes of delivery, prior urinary incontinence and prolapse repairs, hysterectomy status, and recurrent urinary tract infection history. A detailed physical examination collected height, weight, and the presence/ absence of a uterus and its approximate size if enlarged. Baseline pelvic organ prolapse quantification (POPQ) measurements were collected.⁷ Pelvic floor muscle tone and strength were assessed via vaginal palpation using a modified Oxford grading system.⁸ Investigators rated their assessment of the participant’s vaginal atrophy using a non-validated metric; 4 criteria (moisture/dryness, color/pallor, presence/absence of rugae and petechiae) were each scored on a 3-point scale. Finally, the participant’s vaginal maturation index (VMI) was assessed with a plastic spatula (similar to those used for a Pap smear) sampling the upper vaginal walls for 15–30 seconds and then swirled in a ThinPrep® container, which is stable for room temperature storage.⁹

Intervention

The study intervention was conjugated estrogen (CE) cream 0.625mg/g (Premarin®) 1g or identical placebo, inserted vaginally nightly for 2 weeks, then twice weekly for a minimum of 5 weeks before surgery. This was consistent with the dosing of the pilot study preceding IMPROVE and other clinical trials being performed at the time of the pilot’s design.^10,11 CE (or placebo) cream was then resumed at hospital discharge and continued twice-weekly for one year. Participants were permitted to apply cream either with a plastic applicator or their fingertip. The first tube was dispensed at baseline/ randomization, and a second tube was dispensed at or near the time of surgery. Subsequent tubes (i.e., numbers 3, 4, and 5) were dispensed in 3-month intervals to cover 12 months of postoperative use. Additional tubes were provided if necessary (e.g., unforeseen delays from time of randomization to surgery). The planned surgery was standardized across sites, as has been previously described^12,13 (see description below).

Randomization

The DCC used SAS V9.3 to generate stratified randomization lists using a consistent block factor of size 4. Patients were stratified by site, hysterectomy status, and duration since menopause (<10 & ≥10 years). The randomization lists were distributed to each site’s research pharmacy, where the investigational drug services (IDS) pharmacists maintained the lists and oversaw dispensing of the correct randomized study cream. Surgeons, outcomes assessors, and patients were masked to assignment.

Standardization of Surgical Procedures

All participants underwent either a uterosacral ligament suspension (“USLS”) or a unilateral sacrospinous ligament fixation (“SSLF”) at their surgeon’s discretion as an inpatient surgical procedure under general anesthesia following steps similar to a prior NICHD-sponsored Pelvic Floor Disorders Network surgical trial.¹² If performing a USLS, 1 delayed absorbable monofilament suture and 1–2 permanent monofilament sutures were placed bilaterally (i.e., 4–6 total) beginning at or above the level of the ischial spine, with the delayed absorbable suture caudad and the permanent suture(s) ~1 cm cephalad. The caudad absorbable sutures ultimately became the lateral cuff angles with the permanent cephalad sutures placed more medially on the vaginal cuff. If SSLF was selected by the surgeon, 2 permanent and 2 delayed absorbable sutures (i.e., 4 total) were placed unilaterally, typically in the right sacrospinous ligament. Based on level I evidence, neither USLS nor SSLF is superior to the other for anatomic, functional, or adverse event outcomes.¹³ Participating study surgeons must have performed ≥20 of these vault suspension procedures with ≥5 in the 12 months before subject enrollment. Local site principal investigators affirmed these minimum criteria and attested to the competence of their co-investigators. Concomitant vaginal hysterectomy, anterior and posterior colporrhaphies, and midurethral slings were allowed but not transvaginal mesh for prolapse. Colporrhaphies were completed using delayed absorbable monofilament suture.

Placebo Development

Consistent with industry best-practice guidelines and US Food and Drug Administration (FDA) mandates for drug development, a placebo cream was designed to match Premarin® in color, viscosity, and pH. Formal stability testing of the placebo was completed at 3, 6, and 12 months at standard conditions of 60% relative humidity and 25°C. Additional “accelerated” stability testing was performed at 75% relative humidity and 40°C, and benzyl alcohol assays (i.e., the preservative) were also completed to project even greater stability durations. Placebo and active study drug were shipped by temperature-controlled transport using a “TempTale”, permitting IDS pharmacists to confirm they were delivered study cream that had been consistently maintained in acceptable environmental conditions. The final placebo cream design—in an effort to precisely match the active study drug—was a mixture cetyl esters wax, cetyl alcohol, white wax, methyl stearate, propylene glycol stearate, glyceryl monostearate, light mineral oil, glycerin, sodium lauryl sulfate, purified water, benzyl alcohol, and propylene glycol.

To achieve blinding, an over-labeling process was used in which Premarin® tubes were masked with a custom opaque label. Identically sized empty tubes (also sourced by Pfizer, Inc) were provided for tubing of the placebo product, and identical custom opaque labels were affixed.

Outcomes Assessed

In addition to the measures described above collected at baseline, participants provided a list of all medications related to the treatment of overactive bladder (OAB) and incontinence. Table 2 details the other outcome measures and the timing of their collection across the duration of the study.

TABLE 2.

Timing of Outcome Measures

Outcome Assessed	Baseline (≥5wk before Surgery)	Preop Visit (w/in 2wk of surgery)	Surgery/ Hospitalization	1mo Postop Visit (±2wk)	3mo Postop Call (±2wk)	6mo Postop Visit (5–7mo)	9mo Postop Call (±2wk)	12mo Postop Visit (10–14mo)	18&30mo Postop Calls (±2wk)	24&36mo Postop Visits (±2mo)
Informed Consent Form, Inclusion/Exclusion, Randomization	X
Demographics	X
Medical History	X
Exam: POP-Q, levator muscle strength assessment	X			X		X		X		X
Exam: Vaginal atrophy assessment	X	X	X	X		X		X		X
Exam: vaginal epithelium/ granulation tissue assessment	X	X		X		X		X		X
Medication Review	X	X		X	X	X	X	X	X	X
Questionnaires*	X	X			X‡	X	X‡	X	X‡	X
Vaginal Maturation Index	X		X					X
Blood Serum (E1, E2)	X							X
Blood (whole), storage for DNA/ genetics assessment (optional)	X
Vaginal Wall Biopsy (optional)			X
Study CreamTube No. dispensing/delivery	X1		X2		X3	X4	X5

POP-Q, Pelvic Organ Prolapse Quantification⁷

^*, Pelvic Floor Distress Inventory (PFDI-20²⁴), Pelvic Floor Impact Questionnaire (PFIQ-7²⁴), Pelvic Organ Prolapse/ Urinary Incontinence Sexual Function Questionnaire (PISQ-IR²⁵) plus a 6-question survey of vaginal atrophy symptomatology²⁶, a general quality-of-life instrument (SF-12), a body image questionnaire, and the patient global impression of severity (PGI-S²⁷). Postoperatively, participants completed the patient global impression of improvement (PGI-I²⁷).

^‡, 3, 9, 18 and 30-mo postoperatively: participant was queried whether she (a) had symptoms of a bulge or, “something falling out that you can see or feel in your vaginal area”; (b) had treatment for pelvic organ prolapse since the last visit date; (c) was experiencing new medical problems or complications; (d) had been hospitalized, had surgery, been to the emergency department, or visited a doctor for urologic or gynecologic conditions since the last visit date; and (e) had any change in her medications.

E1, estrone; E2, estradiol

During surgery, surgeons provided their impression of the tissue quality of the anterior and posterior vaginal walls’ epithelium and fibromuscular tissue using a visual analog scale ranging from “thin, attenuated, poor” to “thick, healthy, robust”. In all consenting participants, a small full-thickness biopsy (a triangle approximately 1.5×1 cm) was obtained from the apical/ cephalad anterior vaginal wall. This was placed in buffered solution (RNAlater®) that is stable when stored at 4°C. It was shipped overnight to UTSW on cold packs for later histologic assessment and qPCR for mRNA testing. Tissue collected locally at UTSW was initially delivered for processing to the lab in phoshate-buffered saline, allowing additional analysis by zymography (protease enzyme activity), immunoblotting, and hydroxyproline assay for cross-linked collagen content. Any abnormal endometrial findings from the surgical pathology specimens were also captured.

Primary Outcome Measure

A large simple trial (randomized, controlled) design was selected to evaluate for possible superiority of vaginal estrogen cream compared to placebo. The primary outcome matched that described in a prior trial¹² with a comparison by treatment group of time-to-failure of the prolapse repair at 1 year after surgery with “failure” defined by any one or more of the outcomes: objective prolapse of the anterior or posterior walls beyond the hymen or the apex descending >1/3 the total vaginal length, subjective sense of bulge, or retreatment. The null hypothesis for this study was that 12 months following transvaginal native tissue repair for prolapse with or without preoperative and continued postoperative local estrogen treatment, women would have the same rate of treatment success.

Statistical Design

Sample Size/ Power Calculation

The literature reports failure by the above definition occurring in approximately 40% of patients with a native tissue USLS or SSLF (i.e., 60% “success”) two years postoperatively.¹³ A 2-sided logrank test with an overall sample size of 188 subjects having surgery (94 each group) achieves 80% power at α 0.05 significance level to detect a hazard ratio of 0.5188 when the proportion of those “surviving” (i.e. “successes”) in the control group is 65% compared to 80% in the treated group, measured at 1 year; a 15% improvement was felt to be clinically meaningful. This presumes the accrual pattern across the study time period (3 years of accrual) is uniform. This size of 188 already accounts for 25% either lost to follow-up or non-adherence to study cream in the control and treatment groups. Allowing for an additional 15% attrition/loss of participants from baseline randomization to surgery, up to 222 total participants will be enrolled, halting enrollment after 188 participants have had surgery.

Data Analysis

The primary outcome will be analyzed according to the assigned treatment group (modified intention-to-treat): i.e., all participants that were randomized and underwent a surgical prolapse repair and have at least one postoperative assessment will be analyzed according to their randomized assignment group. All eligible subjects who consented to the study and were enrolled will be available for analysis of baseline characteristics, and those that were randomized and used their study drug are monitored for adverse events, but they must meet the additional qualifications above (i.e., have undergone surgery and had a postoperative assessment) to be analyzed for the primary study outcome. At baseline, continuous measures are compared between groups with Student’s t-test (or Mann-Whitney U test) and chi-square (or Fisher’s exact test) for categorical measures. Scores from repeated surveys with continuous outcomes are analyzed using a mixed effects model with repeated measures and subject as a random effect. A variety of covariance structures will be fit to the data (compound symmetry, auto regressive (1), heterogeneous auto regressive (1), Toeplitz, unstructured, etc.) and the best fitting covariance structure (using AIC or BIC criteria) will be reported. Tests for interactions will be performed between time effect and treatment effect. For categorical data (e.g., prolapse stage), analysis will be by categorical data modeling using weighted least squares and mixed logit models, with cross-sectional assessments (at specific time intervals) by chi-square. For missing continuous follow-up points, the likelihood-based mixed effects models have been reported to be less biased and more representative of the data than other methods even when the data are not missing at random.¹⁴ Time-to-event or event-free-time will be analyzed using survival methods to include the log-rank test and Cox regression. For measures that may not be assumed normal, the data will be transformed prior to analysis.

As with the primary outcome, secondary outcome measures such as continuous scores from repeated surveys will be analyzed using mixed effects models. Because midurethral slings and OAB medications were not excluded in study participants, these other interventions may impact upon incontinence rates and will be controlled for in statistical comparisons between the two groups. Gross estimates of OAB in the 2 groups will also be assessed by rates of OAB medication use. Modest changes in stress and overall urinary incontinence will be assessed from baseline to the preoperative visit (i.e., after at least 5 weeks of study cream) using the relevant items from questionnaires (Table 2) at both time points. Proportions of participants with postoperative cystitis will be compared using chi-square. Using the liberal definition of post-operative cystitis to include any patient prescribed antibiotics for cystitis symptoms, this study will have sufficient power to demonstrate an improvement from a 22% cystitis rate to 10% (80% power, α 0.05). Comparisons of VMI and atrophy symptoms will use analysis of covariance, with treatment and study center as factors and baseline values as the covariate.

Defining Study Drug Adherence

Various per-protocol analyses will be considered, including comparisons of the subset of participants in the two randomized groups who were adherent to their study medication. Therefore, defining drug adherence is critical. This is described in three ways: “per pivotal study”, “strict” adherence, and objectively using VMI. One of the 3 recommended dosage and administration schedules per Pfizer’s package insert for Premarin® cream is twice-weekly administration of 0.5g intravaginally. This was based on the 2009 pivotal study by Bachmann, et al.¹⁵ In this trial, the twice-weekly CE group of participants was defined as adherent with their drug if they self-reported, on average, at least once-weekly (and not more than three times weekly) application of the cream on their paper diary; medication tube weights were not collected. Similarly, in IMPROVE, if participants reported via their medication calendars that they applied their cream at least 50% of the times expected, this was considered adherent by this “per pivotal study” definition.

Strict adherence was defined as >80% adherent to expected dosing using the participant’s returned tubes (i.e., comparing returned tube weights to that expected with at least 80% adherence) and/or medication calendar. Should there be a discrepancy in adherence designation between tube weight and medication calendar (e.g., using tube weights, the participant appears to have been non-adherent to study protocol but her calendar suggests she was adherent with at least 80% of planned doses recorded as completed), deference will be given to the more objective tube weight.

Finally, cytomorphologically, vaginal atrophy can be defined by high numbers of parabasal and intermediate cells and very low numbers of superficial cells. In a prolonged low-estrogen state, superficial cells are no longer produced, leaving essentially only parabasal and basal cells lining the vaginal wall. The VMI served as an important objective tool to assess patient adherence with local estrogen application in those participants randomized to vaginal estrogen. This was a quantifiable measure of the proportion of vaginal epithelial superficial cells to parabasal and intermediate cells. While there was not a universally agreed-upon formula to calculate VMI or a cutoff to absolutely define “atrophic” versus “estrogenized”, several studies pointed toward the following¹⁶: 0 x % parabasal cells + 0.5 x % intermediate cells + 1.0 x % superficial cells. A cutoff of 50 or less was indicative of atrophy.¹⁷ Therefore, a score >50 was indicative of study drug adherence in those randomized to active CE cream.

Rationale for Premarin®

While adherence to prescription medications is often poor and hindered by increasing age, minimally-symptomatic conditions, polypharmacy, and labor-intensive therapy such as application of a vaginal cream¹⁸, in the setting of other rigorous clinical trials, exceptionally good adherence rates have been observed: CE cream twice-per-week discontinuation over 1 year was just 9.8% in a recent study.¹⁵ CE cream was used in the preceding pilot study and for consistency was chosen (as opposed to estradiol cream) for the IMPROVE trial.¹⁰ Theoretically, adherence could be improved over cream using an estradiol ring (Estring®), which is replaced every 3 months, but users are sometimes bothered by difficulty with ring removal, concern over vaginal infection risk, bother with partner feeling the ring, or difficulty with insertion.¹⁹ Further, a trial that randomized women undergoing reparative surgery randomized immediately postoperatively to 12 weeks of vaginal estradiol or placebo rings found increased granulation tissue attributed to a foreign body reaction/ inflammation with the placebo (12 of 21) compared to the estradiol ring (3 of 22), P<0.01.²⁰ Another alternative, the vaginal estradiol tablet (Vagifem®), was currently only available as a 10 μg tablet, which was likely not equivalent to the 1g CE cream demonstrated as effective in matrix synthesis and decreased protease activity in the pilot study referenced above.¹⁰ In addition, severely atrophic patients sometimes report failure of the tablet to dissolve and/or the tablet falling out with the emergence of the POP. Finally, there are other potential benefits stemming from the unique “Ring B” unsaturated equine estrogens in CE, including inhibition of LDL and HDL oxidation and apoptosis in neurons.²¹

Adverse Event Collection and Reporting

Beginning with the preoperative assessment (i.e., the first participant encounter after baseline randomization), adverse events were routinely queried at each encounter. All adverse events were captured and assessed for severity, relationship to study cream, and expectedness. Each event was tracked for its ultimate outcome/ resolution. The most severe adverse event encountered at the time of surgery and during hospitalization was categorized with a Dindo score.²² Serious adverse events were reported in a timely fashion (i.e., in less than 24 hours of learning of the event) to the NIA program officer, a Safety Monitor (DSMB chair), the DCC chair, and to Pfizer, Inc. The discovery of either breast or endometrial cancers were regarded as serious adverse events, deemed possibly-related to study cream, and reported accordingly.

RESULTS:

As of the time of preparation of this manuscript, 204 women had consented to participation. Owing to screen failures, 197 had been randomized. There had been 174 completed surgeries with 15 more pending. One hundred eleven women had completed their 12 month postoperative visit, and 45 had completed a 24 month postoperative visit. None had yet completed a 36 month postoperative visit.

DISCUSSION:

The goal in designing this study was to contribute robust and impactful data on the utility of vaginal estrogen as a means of augmenting transvaginal native tissue repairs by generating high-level evidence with a randomized, placebo-controlled trial. We have endeavored to partner with skilled chemists and drug manufacturers to produce the highest possible quality placebo cream, matching CE in color, viscosity, and pH that is safe and stable for application years after production. Masking with customized opaque labels has satisfactorily blinded participants, surgeons, and outcomes assessors to randomization allocation. An additional strength of the study is standardization of the allowed surgical procedures as described here in detail, which can be simply adopted by skilled pelvic surgeons. Besides determining vaginal estrogen’s impact on prolapse repair longevity, this study will contribute important information about its effect on urinary incontinence and OAB symptoms, sexual function, body image, and perioperative adverse events such as postoperative cystitis and poor wound healing (i.e., frequency of granulation tissue, persistent suture exposure, or mesh exposure in the case of those undergoing concomitant anti-incontinence slings).

There are limitations to this study. The choice of timing of the primary outcome (i.e., time-to-failure by 12 months postoperatively) was a pragmatic one; it is possible this is too soon to recognize a difference between the study groups. This is why it is planned to follow all participants for 3 years postoperatively and continue outcome assessor and participant blinding during this time period. Conversely, if there is a significant difference in failure rates between the groups noted by 12 months postoperatively (i.e., a positive study), ongoing clinical follow-up to 3 years is critical to assess the longevity of the quality of these surgical repairs and whether the initial perioperative use of estrogen cream remains clinically meaningful over time. Especially in the setting of a positive study, the comprehensive collection of data on interval hospitalizations, surgeries, emergency room and doctor visits related to urologic or gynecologic conditions will permit a meaningful cost-effectiveness analysis.

Another potential limitation is the choice of vaginal estrogen and its dosing. Premarin® cream is among the most studied of vaginal estrogen products available, but it has numerous competitors primarily consisting of various formulations of estradiol (creams, tablets, ovules, rings). The rationale for Premarin® is described above, and it seems reasonable to translate lessons learned from IMPROVE to other vaginal estrogens²³, but this remains an unproven assumption and still has limits to its generalizability to estradiol products or to estriol, which is commonly prescribed to women with genitourinary syndrome of menopause outside the US.

In conclusion, the design of a randomized, placebo-controlled trial of vaginal estrogen as an adjunct to an apical transvaginal native tissue surgical repair for prolapse in postmenopausal women is presented. Attention to high quality placebo manufacture, ongoing masking of randomization allocation, and surgery standardization should limit bias and provide robust data to assist surgeons and patients with decisions about the optimization of perioperative care to enhance prolapse repair outcomes.