Comparative estrogen exposure from compounded transdermal estradiol creams and Food and Drug Administration-approved transdermal estradiol gels and patches

This study suggests that women using compounded transdermal estradiol creams may be receiving less estrogen exposure when compared to women using FDA-approved transdermal estradiol formulations. Urinary estradiol concentrations did increase with increasing doses of compounded transdermal estradiol cream, similar to what has been seen with FDA-approved transdermal formulations.

Abstract

Objective

The aim of this study was to evaluate the amount of estrogen exposure associated with the use of compounded transdermal estradiol (E2) creams and compare it with estrogen exposure associated with the use of Food and Drug Administration (FDA)-approved transdermal E2 patches and gels.

Methods

This was a retrospective cohort study that used clinical laboratory data collected from January 1, 2016, to December 31, 2019. Participants were first divided into three groups: postmenopausal women on no menopausal hormone therapy (n = 8,720); postmenopausal women using either a transdermal E2 patch, gel, or cream (n = 1,062); and premenopausal women on no hormonal therapy (n = 16,308). The postmenopausal menopausal hormone therapy group was further subdivided by formulation (patch [n = 777], gel [n = 132], or cream [n = 153]) and dose range (low, mid, or high). The Jonckheere-Terpstra trend test was used to determine if there was a dose-dependent trend in urinary E2 with increasing dose of compounded E2 cream (dose categories for E2 cream subanalysis, <0.5 mg [n = 49], ≥0.5-≤1.0 mg [n = 50], ≥1.0-≤1.5 mg [n = 58], and >1.5-≤3.0 mg [n = 46]). Urinary E2 and other characteristics were compared across formulations (within each dose range) using Kruskal-Wallis one-way analysis of variance.

Results

A dose-dependent, ordered trend existed for urinary E2 with increasing doses of compounded E2 cream (urinary E2 medians [ng/mg-Cr], 0.80 for <0.5 mg, 0.73 for ≥0.5-≤1.0 mg, 1.39 for ≥1.0-≤1.5 mg, and 1.74 for >1.5-≤3.0 mg; Jonckheere-Terpstra trend test, P < 0.001). Significant differences in urinary E2 concentrations were observed in all three dose ranges (Kruskal-Wallis one-way analysis of variance, P = 0.013 for low dose, P < 0.001 for mid dose, P = 0.009 for high dose). Comparison of E2 concentrations of compounded creams to E2 concentrations obtained with similar doses of FDA-approved patches and gels showed that the creams had significantly lower values than the patches and gels.

Conclusions

Estrogen exposure from compounded transdermal E2 creams increases in a dose-dependent manner; however, the amount of estrogen exposure associated with compounded creams is significantly lower than estrogen exposure associated with FDA-approved transdermal E2 patches and gels. Clinicians should be aware of the direction and magnitude of these potential differences in estrogen exposure when encountering women who have either previously used or are currently using compounded E2 creams.

Estrogen therapy (ET) is considered the most effective treatment option for both vasomotor symptoms (VMS) and the genitourinary syndrome of menopause.^1,2 In the wake of the Women's Health Initiative study, the use of ET initially decreased and then began to transition from oral to transdermal formulations, presumably because of a hypothesized lower risk of adverse outcomes.^3-6 However, the risk difference between formulations has yet to be fully elucidated.^7-10 Much of the risk (and benefit) associated with the use of ET is thought to be related to the amount and duration of estrogen exposure.^11,12 Therefore, it may be prudent to focus on optimizing the amount of estrogen exposure to both achieve treatment goals and minimize risk. Quantifying these risks and risk reductions according to dose and formulation has not been thoroughly explored, potentially because no convenient tool exists to facilitate the exploration. Whereas the amount of estrogen exposure resulting from Food and Drug Administration (FDA)-approved transdermal estrogen formulations is somewhat understood, almost nothing is known about estrogen exposure resulting from the use of compounded transdermal estrogen formulations.^13,14

Compounded transdermal estrogen formulations, generally included under the umbrella of compounded bioidentical hormones, are not approved by the FDA, and many organizations recommend against their use.^1,14,15 A documented allergy to an active pharmaceutical ingredient or excipient or a clinical need for a dose unavailable in an FDA-approved formulation is the only widely accepted justification for their use.¹⁴ A study published in 2015 found that approximately 33 million compounded hormone therapy prescriptions were being filled annually.¹⁶ The majority of pharmacists surveyed in that study indicated that they expected compounded hormone therapy prescription volume to grow by 5% to 25% over the next 2 years.¹⁶ That trend has continued and we are now seeing more patients than ever using compounded estrogen products, which is one of several reasons why understanding the comparative estrogen exposure from these compounded products may be important. Understanding estrogen exposure is also important when transitioning a patient from compounded therapy to an FDA-approved formulation. This process is somewhat analogous to dose equivalency considerations when switching a patient from one statin to another.¹⁷ Given the increasing number of women using compounded products, needing to convert women to FDA-approved therapies may be a situation many providers find themselves in more often, especially if the FDA imposes sanctions on the compounding of transdermal estrogen products.

Using serum sampling to explore estrogen exposure in the context of ET use in real-world settings is fraught with difficulties. This stems from the fact that one serum sample collected at one time point does not convey a sufficient amount of information to properly answer research questions. Previous investigators seeking to answer questions about estrogen exposure or the pharmacokinetics of estrogen products have had to collect 7 to 24 serum samples over a 24-hour period.^18-20 This process is expensive, inconvenient for patients, and nearly impossible to implement in a real-world setting. In contrast, urine sampling provides a much easier, more convenient, and relatively inexpensive alternative sampling method for the purpose of answering research questions concerning estrogen exposure. In addition, to the best of our knowledge, the analytical method used in this study is the only method with published validation supporting its use as a surrogate for serum sampling.^14,21 Furthermore, our group has previously demonstrated that urinary estradiol (E2), estrone, and other estrogen metabolite concentrations display expected dose-dependent increases when patients are using FDA-approved transdermal E2 patches and gels.^22,23

The primary goal of this study was to determine if, similar to FDA-approved transdermal E2 patches and gels, urinary E2 concentrations increase in a dose-dependent manner with the administration of increasing doses of compounded E2 cream. Secondary goals were to compare E2 and additional estrogen metabolite concentrations observed in women using compounded E2 creams at varying doses to concentrations observed in women using FDA-approved transdermal E2 products at similarly varying doses.

METHODS

Data source

Data analysis for this study used a subset of data from a larger study, Precision Analytical Retrospective Data Correlation (Clinical Trials ID, NCT04305093), which comprises deidentified, retrospective, observational clinical laboratory data collected from January 1, 2016, to December 31, 2019. The data that we were interested in for the present study were age at the time of sample collection, sex, menstrual status, body mass index (BMI) calculated from self-reported height and weight, hormonal medication use, and urinary estrogen metabolite concentrations including estrone, E2, estriol, 2-hydroxyestrone, 2-hydroxyestradiol, 4-hydroxyestrone, 16-hydroxyestrone, and 2-methoxyestrone. Data on race/ethnicity were not available. The study was approved by the National University of Natural Medicine Institutional Review Board. Because all data were deidentified, the institutional review board determined that written informed consent could be waived.

Women included in the on therapy groups had an age at the time of sample collection greater than or equal to 56 and were using a transdermal formulation of E2 (patch, gel, or cream). Only women who reported applying the transdermal therapy within a predefined window (within 1 wk for an E2 patch and within 3 d for an E2 cream or gel) were included. Women were excluded from these groups if they reported any of the following: kidney disease, an adrenal disorder, use of ET other than transdermal E2 (individuals were specifically asked about the use of birth control pills, combined estrogen treatments, and oral, sublingual, pelleted, or injectable estrogen), use of diindolylmethane, use of tamoxifen, use of anastrazole, or use of testosterone. Women were also excluded if there was missing data, evidence of overly dilute urine (urinary creatinine <0.1 mg/mL), age older than 85 years, or a BMI outside the range of 16 to 60 kg/m².

Postmenopausal and premenopausal women on no therapy were included for reference. Inclusion criteria for women in the postmenopausal no-therapy group were an age at the time of sample collection of 56 years or older and no use of ET. Exclusion criteria were the same as those for the on therapy groups. Inclusion criteria for women in the premenopausal reference group were an age at collection older than 18 years, a reported menstrual status of “regular,” and a reported sample collection during the luteal phase of the menstrual cycle. Exclusion criteria were again the same as those for previous groups.

Sample collection

Study participants collected a total of four urine samples throughout the day: (1) the first urine of the day, (2) 2 hours after waking, (3) in the afternoon (approximately 5 pm), and (4) before bedtime (approximately 10 pm). Sample times varied, but the general timeline was adhered to by all participants included in the analysis. The sample collection method has been previously described elsewhere,^21,24 but briefly, samples were collected by saturating strips of filter paper (Whatman Body Fluid Collection Paper; Sigma-Aldrich, St Louis, MO) with urine either by urinating directly on the filter paper or soaking the filter paper in a clean collection cup of urine. The strips of filter paper were then dried at room temperature for 24 hours and subsequently shipped to the laboratory for analysis.

Laboratory methods

Urinary estrogen concentrations were determined using a previously described, validated method.^21,24 Assay performance characteristics were also previously published.^21,24 Briefly, urine samples were extracted from the filter paper using ammonium acetate. After extraction from the filter paper, urinary creatinine was measured using a conventional colorimetric (Jaffe) method. Creatinine concentrations from each of the four samples was used to determine the volume of each sample to include in a combined representative sample for the period over which samples were collected. Urinary estrogen concentrations from the combined representative sample were analyzed in our laboratory via gas chromatography-tandem mass spectrometry on the Agilent 7890/7000B (Agilent Technologies, Santa Clara, CA). The estrogen concentrations were normalized to creatinine to account for variation in both urine concentration and filter paper saturation.

Statistical methods

Because no variables were normally distributed, summary statistics are reported as median (interquartile range). To evaluate if increasing doses of compounded transdermal E2 creams were associated with corresponding increases in urinary E2 concentrations, the compounded E2 cream doses were grouped into four categories: (1) <0.5 mg, (2) ≥0.5 mg to <1.0 mg, (3) ≥1.0 mg to ≤1.5 mg, and (4) >1.5 mg to ≤3.0 mg. There were a total of 203 observations collected while women were using a compounded transdermal E2 cream in one of these four prespecified dose ranges. Because drug doses are by definition ordered a priori, the nonparametric Jonckheere-Terpstra trend test was used to assess for ordered differences across dose categories. The Jonckheere-Terpstra trend test is more powerful than Kruskal-Wallis one-way analysis of variance (ANOVA) for detecting differences when there is a priori ordering.^25-27 Repeated measures were removed by only keeping the first measurement collected while on therapy to meet the assumption of independence of observations.

To compare estrogen exposure between formulations, dose categories of low, mid, and high were created as follows: low, <0.5 mg for E2 cream and E2 gel, and 0.025 to 0.0375 mg for an E2 patch; mid, ≥0.5 to <1.0 mg for E2 cream and E2 gel, and 0.05 mg for an E2 patch; and high, ≥1.0 to ≤1.5 mg for E2 cream and E2 gel, and 0.075 to 0.1 mg for an E2 patch. There were a total of 417 observations in the low dose category, 380 in the mid dose category, and 265 in the high dose category. Comparisons between formulations in each dose category were made using Kruskal-Wallis one-way analysis of variance,²⁸ followed by pairwise comparisons with Dunn's test and Bonferroni correction of P values to account for multiple comparisons.²⁹ Because this was a data mining study using convenience sampling, no sample size or power calculation was performed a priori, and all comparisons should be considered exploratory and hypothesis generating in nature.

α was set at 0.05 for all statistical tests.

RESULTS

Evaluation of increasing E2 trend in compounded cream users

Summary statistics for the observations used in the evaluation of the ordered trend for E2 cream doses are shown in Table 1. A dose-dependent, ordered trend existed for urinary E2 with increasing doses of compounded E2 cream (Jonckheere-Terpstra trend test, P < 0.001) (Fig. 1). Of note, the trend was not strictly increasing but still resulted in a significant result because the Jonckheere-Terpstra test does not require all inequalities to be strict inequalities.^25-27 Summary statistics for the postmenopausal (no therapy) and premenopausal (no therapy) reference groups are shown in Table 2.

Table 1.

Characteristics of women using transdermal estradiol cream

Variable	<0.5 mg, n = 49a	≥0.5-<1.0 mg, n = 50a	≥1.0-≤1.5 mg, n = 58a	>1.5-≤3.0 mg, n = 46a
Age at collection (y)	59.0 (58.0-65.0)	60.5 (58.0-64.0)	61.5 (58.0-65.8)	63.0 (60.0-67.8)
BMI (kg/m2)	23.7 (20.8-25.7)	25.0 (22.4-25.8)	23.7 (21.3-26.4)	23.7 (21.3-25.7)
Estradiol (ng/mg-Cr)	0.80 (0.48-1.14)	0.73 (0.41-2.05)	1.39 (0.68-3.29)	1.74 (0.91-2.46)
Estrone (ng/mg-Cr)	5.60 (3.00-7.50)	5.90 (3.55-9.73)	9.05 (4.63-17.6)	10.4 (5.65-14.5)
Estriol (ng/mg-Cr)	3.30 (1.80-7.30)	4.55 (2.93-9.78)	6.05 (3.20-9.73)	7.50 (4.33-14.9)

BMI, body mass index.

^aMedian (interquartile range).

FIG. 1.

Median estradiol concentrations by estradiol cream dose range.

Table 2.

Characteristics of postmenopausal and premenopausal women on no therapy

Variable	Postmenopausal (no therapy), n = 8,720a	Premenopausal (no therapy), n = 16,308a
Age at collection (y)	61 (58-65)	38 (33-43)
BMI (kg/m2)	25.5 (22.3-28.6)	23.6 (21.2-26.6)
Estradiol (ng/mg-Cr)	0.30 (0.18-0.48)	3.09 (2.20-4.33)
Estrone (ng/mg-Cr)	2.60 (1.60-4.10)	17.8 (12.6-24.7)
Estriol (ng/mg-Cr)	1.40 (0.80-2.30)	9.70 (6.00-15.3)

BMI, body mass index.

^aMedian (interquartile range).

Comparison of formulations by dose range

Formulations were compared with each other within each of the prespecified dose ranges of low, mid, and high (Fig. 2, Tables 3-5). The number of observations from each formulation is presented in Tables 3 to 5.

FIG. 2.

Median estradiol concentrations by formulation and dose range.

Table 3.

Characteristics of low dose users of all three formulations

Variable	Cream, n = 50a	Gel, n = 11a	Patch, n = 356a	Pb
Age at collection (y)	59.0 (58.0-65.0)	61.0 (57.0-65.0)	61.0 (58.0-65.0)	0.81
BMI (kg/m2)	23.9 (20.9-25.7)	21.9 (20.6-24.9)	23.4 (21.1-25.7)	0.66
Estradiol (ng/mg-Cr)	0.80 (0.49-1.21)	1.36 (0.84-1.97)	1.10 (0.69-1.64)	0.013
Estrone (ng/mg-Cr)	5.65 (3.20-7.72)	7.60 (4.80-10.90)	6.10 (3.98-8.90)	0.29
Estriol (ng/mg-Cr)	3.30 (1.80-7.00)	2.60 (2.35-5.00)	2.85 (1.78-4.70)	0.29

BMI, body mass index.

^aMedian (interquartile range).

^bKruskal-Wallis rank sum test.

Table 5.

Characteristics of high dose users of all three formulations

Variable	Cream, n = 54a	Gel, n = 53a	Patch, n = 158a	Pb
Age at collection (y)	61.0 (58.0-65.8)	59.0 (58.0-62.0)	60.0 (58.0-64.0)	0.089
BMI (kg/m2)	23.6 (21.3-25.8)	23.5 (22.0-26.8)	23.3 (21.0-25.7)	0.40
Estradiol (ng/mg-Cr)	1.39 (0.64-3.26)	2.04 (1.10-3.07)	2.62 (1.50-4.10)	0.009
Estrone (ng/mg-Cr)	9.05 (4.62-17.6)	9.50 (7.60-19.3)	12.2 (8.00-19.1)	0.32
Estriol (ng/mg-Cr)	5.95 (3.12-9.47)	7.60 (3.70-17.1)	6.65 (3.50-10.7)	0.19

BMI, body mass index.

^aMedian (interquartile range).

^bKruskal-Wallis rank sum test.

Low dose

Summary statistics and comparisons of formulations in the low dose category are shown in Table 3. E2 was the only variable for which a significant difference between formulations was found (Kruskal-Wallis one-way ANOVA, P = 0.013). The pairwise comparisons showed that the only significant difference was between the cream and patch formulations, where the median E2 concentration for the low dose cream category was significantly lower than the median E2 concentration for the low dose patch category (Dunn's test, P = 0.02). However, the number of participants in the low dose gel group (n = 11) was relatively low.

Mid dose

Summary statistics and comparisons of formulations in the mid dose category are shown in Table 4. There was a significant difference between formulations for all three estrogen metabolites (E1, E2, and E3) in the mid dose category (Kruskal-Wallis one-way ANOVA, P = 0.001 for E1 and P < 0.001 for E2 and E3). Subsequent pairwise comparisons for E2 indicated that there was a significant difference between the cream and gel formulations (Dunn's test, P = 0.002) as well as between the cream and patch formulations (Dunn's test, P = 0.001). The median E2 concentrations of the gel and patch formulations did not differ (Dunn's test, P > 0.99). Pairwise concentrations for E1 indicated that there was a significant difference between the cream and gel formulations (Dunn's test, P < 0.001) and between the cream and patch formulations (Dunn's test, P = 0.03). As with E2, there was no significant difference between the gel and patch formulations (Dunn's test, P = 0.10). Finally, pairwise comparisons for E3 indicated that the only significant difference was between the gel and patch formulations (Dunn's test, P < 0.001). The cream formulation did not differ from either the gel (Dunn's test, P = 0.42) or the patch formulation (Dunn's test, P = 0.41).

Table 4.

Characteristics of mid dose users of all three formulations

Variable	Cream, n = 49a	Gel, n = 68a	Patch, n = 263a	Pb
Age at collection (y)	60.0 (58.0-64.0)	60.0 (58.0-62.2)	60.0 (58.0-63.0)	0.75
BMI (kg/m2)	25.0 (22.3-25.8)	23.8 (21.7-26.0)	23.6 (21.2-26.2)	0.39
Estradiol (ng/mg-Cr)	0.75 (0.44-2.20)	1.63 (1.06-2.66)	1.65 (1.10-2.37)	<0.001
Estrone (ng/mg-Cr)	5.90 (3.70-10.0)	9.95 (7.27-15.9)	8.70 (5.90-12.3)	0.001
Estriol (ng/mg-Cr)	4.70 (3.00-10.2)	6.10 (4.10-9.67)	4.10 (2.60-6.40)	<0.001

BMI, body mass index.

^aMedian (interquartile range).

^bKruskal-Wallis rank sum test.

High dose

Summary statistics and comparisons of formulations in the high dose category are shown in Table 5. Similar to the low dose category, E2 was the only variable for which a significant difference was observed between formulations (Kruskal-Wallis one-way ANOVA, P = 0.009). Pairwise comparisons revealed that the only significant difference was between the cream and patch formulations (Dunn's test, P = 0.009).

DISCUSSION

The primary goal of this study was to determine if increasing doses of compounded E2 cream corresponded to increasing urinary E2 concentrations, a proposed surrogate marker for estrogen exposure. Secondary goals were to compare E2 and other estrogen metabolite concentrations between formulations within the same dose range. To the best of our knowledge, this is the first study to explore these questions.

Although the expected, dose-dependent, increasing trend was observed with increasing doses of compounded E2 cream, the trend was not strictly increasing. This seems to suggest that the estrogen exposure from compounded creams in the mid dose range is not consistently higher than estrogen exposure from creams in the low dose range. There are multiple plausible hypotheses as to why this was observed; a lack of standardization in the compounding process across pharmacies and higher baseline E2 values in women using low dose compounded creams are two explanations that seem likely. The latter is an interesting hypothesis because, if menopausal symptoms are correlated with low endogenous E2 concentrations, it would be expected that women presenting with milder symptoms would have higher baseline E2 concentrations than women presenting with more severe symptoms. Women with milder symptoms might also be started on lower doses of ET. More work is warranted to further investigate this hypothesis.

Comparison of the E2 concentrations of compounded creams to E2 concentrations seen with similar doses of FDA-approved patches and gels showed that the actual values for compounded creams are lower in all three of the dose range categories. A statistically significant lower median E2 concentration was observed in all three dose range categories when comparing compounded E2 creams to transdermal E2 patches. The difference between compounded E2 cream and FDA-approved E2 gel was only significant in the mid dose range; however, the actual difference in medians in the other two dose range categories was similar to the actual difference observed in the mid dose range. In addition, the low number of gel observations (n = 11) in the low dose category likely limited the ability to detect a statistically significant difference. Comparison of median E1 and E3 concentrations generated results that are more nuanced than differences observed in E2. Intriguingly, the mid dose range category was the only category in which statistically significant differences in E1 and E3 were observed. Importantly, there was no statistically significant difference in BMI between the three different formulations. This suggests that the primary driver of changes in estrogen metabolite levels was related to the differing pharmacokinetic properties of the three different formulations.

Strengths of this study include (1) the use of real-world data, (2) the large number of postmenopausal women on no therapy in the comparator group, and (3) the use of gas chromatography-tandem mass spectrometry for the determination of E2 and estrogen metabolite concentrations. Some limitations must also be considered. Because the compounded E2 creams used by the women in this study were prepared at different pharmacies, we must assume some degree of variability in concentration. In addition, we do not know what base was used to compound the creams, and the choice of base has been shown to impact absorption.³⁰ However, somewhat paradoxically, this potential variability in compounding base is actually beneficial in that it represents what is likely observed in real-world clinical practice where different compounding pharmacies use differing bases. Another limitation is that some variables were self-reported including height, weight, and adherence to therapy. Finally, the relative sample sizes for women using creams and gels were small. Future studies are needed to confirm the observed differences between groups with larger sample sizes.

One of the only other studies to explore a similar research question regarding compounded ET was a randomized, controlled trial published in 2013 that examined the pharmacokinetics of compounded estrogen cream in comparison with a “standard-dose” E2 patch (0.05 mg).¹⁸ In that study, a compounded estrogen cream containing 80% E3 and 20% E2 was used. Participants had serial serum samples drawn first at baseline and then at 2, 5, 8, 12, 18, and 24 hours after administration. Similar to the results of the present study, these investigators found that serum E2 concentrations were significantly lower in the participants receiving compounded E2 cream compared with the participants receiving the 0.05 mg patch. The current findings extend those previous results in two ways: (1) by providing a comparison with FDA-approved transdermal gel products and (2) by providing comparisons to E2 patches at varying doses.

Lower E2 concentrations have been shown to be associated with lower bone mineral density and thus a higher risk for osteoporosis.³¹ Importantly, because of the significant placebo effect associated with ET,³² the estrogen exposure required to prevent osteoporosis may not be reached if only subjective symptom improvement is used to determine dosing. One implication of the present findings is that, because of apparent lower estrogen exposure, compounded E2 creams may not provide the same level of BMD preservation as FDA-approved transdermal E2 products. Historically, the postmenopausal serum E2 reference range was considered to be <30 pg/mL.³³ However, as quantification methods have improved, the true reference range has been demonstrated to be lower, likely less than 10 pg/mL.^34-37 Similarly, previous investigators reported that the serum E2 level necessary to prevent or delay bone loss was 40 to 60 pg/mL.³⁸ However, as with the postmenopausal reference range, reevaluating this range with improved quantification methods with increased sensitivity will likely demonstrate that the lower bound of the range can be decreased below 40 pg/mL. These improved quantification methods are the more recently developed liquid chromatography-tandem mass spectrometry assays, which have largely replaced the older traditional radioimmunoassays; this is, in part, due to significant cross-reactivity between estrogen metabolites with immunoassays.^39,40

Given the 20 pg/mL downward shift in the postmenopausal serum E2 range with improved quantification methods, it could be safely assumed that the range necessary to prevent or delay bone loss may be shifted by the same amount. This would place the lower end of the range at 20 pg/mL, which is two times the upper limit of the postmenopausal range (<10 pg/mL). In the present study, if we allow the interquartile range to serve as a rough estimate of the postmenopausal reference range, the surrogate upper limit of the postmenopausal E2 range is 0.48 ng/mg-Cr. Following the example of serum E2 concentrations, this would suggest that a concentration close to 0.96 ng/mg-Cr may sufficiently prevent or delay bone loss. When considering median E2 concentrations, our results suggest that compounded E2 creams at doses lower than 1 mg would fail to meet this threshold. Contrastingly, median E2 concentrations for the transdermal gel and patch formulations exceed that threshold in each dose range category, which is consistent with clinical outcomes seen in the literature. The relationship between E2 and BMD is complex, and uncoupling it from the known covariate, BMI, is necessary. In this study, the BMI of participants is lower than what would be seen in the general population, so the results are not generalizable. To properly explore research questions concerning E2 concentrations, BMD, and BMI, baseline and repeat measurements would be needed.

Similar to BMD, menopausal symptom improvement, including decreased VMS and vulvovaginal atrophy symptom improvement, has historically been thought to occur when serum E2 levels are between 40 and 80 pg/mL,⁴¹ but again, evaluation with improved quantification methods will likely shift this range lower as well. Therefore, the same interpretation applies, and consistent with the previous discussion of BMD, this suggests that compounded E2 creams at lower doses may not be providing sufficient estrogen exposure to improve menopausal symptoms.

Further research is needed to determine if there is a correlation between clinical outcomes (VMS relief, vulvovaginal atrophy symptom improvement, etc) and estrogen exposure or baseline E2 concentrations. Given the significant placebo effect observed in randomized controlled trials with treatments for VMS,³² adding the dimension of estrogen exposure to a study may shed some additional light on symptom etiology and improvement. Further validation of urinary estrogen metabolite concentrations as surrogates for estrogen exposure is also needed, ideally by comparison with serial serum measurements such as those collected in the previously discussed study conducted by Sood and colleagues.¹⁸ Combining both an investigation into correlations with clinical outcomes and serial serum measurements would provide the most useful information regarding the clinical significance of estrogen exposure measured via urinary estrogen metabolite concentrations. Because urinary sampling allows for quantification of downstream estrogen metabolite concentrations, future studies or subanalyses looking at the differences in estrogen metabolism by formulation may be warranted or prove beneficial. Finally, additional work is necessary to uncover why doses of compounded E2 creams that are similar to doses of FDA-approved transdermal E2 products result in lower estrogen exposure. This may be related to absorption differences between formulations (differing bioavailability) or some other pharmacokinetic or pharmacodynamic property difference. A similar observation has been reported for progesterone creams.⁴²

CONCLUSIONS

In conclusion, the results of this study provide some insight into estimated estrogen exposure when using compounded E2 creams when compared with what is expected when FDA-approved transdermal E2 products are used. Ideally, these results will encourage further research exploring the clinical significance of estrogen exposure and how, or if, it might be used to optimize treatment and minimize risk for menopausal women who are candidates for HT.