The impact of 3,3’-diindolylmethane on estradiol and estrogen metabolism in postmenopausal women using a transdermal estradiol patch

Abstract

Objectives:

3,3’-diindolylmethane (DIM) is a supplement, investigational drug, and the primary in vivo product of indole-3-carbinol. DIM is sometimes recommended to postmenopausal women by functional or integrative medicine providers. Some of these women may be concurrently receiving menopausal hormone therapy (MHT), and since DIM’s mechanism of action involves the alteration of estrogen metabolism, it is possible that a drug-supplement interaction exists. The objective of this study was to examine the effect of DIM on the estrogen profiles of postmenopausal women receiving MHT in the form of a transdermal estradiol (E2) patch.

Methods:

This was a retrospective, observational cohort study for which data were collected from a database containing laboratory results for patients processed between January 1, 2016 and December 9, 2019. Laboratory measurements included urinary E2 and 9 other estrogen metabolites. From this database, we identified 1,458 results that were from postmenopausal women using a transdermal E2 patch. Of these 1,458 women, 108 indicated they were concurrently taking a DIM supplement. Wilcoxon rank sum tests were used to assess differences between groups. Multivariable linear regression models were created to confirm the effect of DIM on individual estrogen metabolites.

Results:

When compared with postmenopausal women using a transdermal E2 patch alone, postmenopausal women using a patch and concurrently taking DIM had statistically significant alterations in their urinary estrogen profiles (P < 0.001 for all metabolites with differences). Multivariable linear regression models indicated that DIM had a significant effect on 6 of the 10 estrogen metabolites measured, including estrone, estriol, 2-OHE1, 2-OHE2, 4-OHE2, and 16-OHE1, as well as the 2-OHE1/16-OHE1 ratio.

Conclusions:

Postmenopausal women who are prescribed transdermal E2 patch therapy who choose to concurrently use a DIM supplement may have unexpected changes in their urinary estrogen profiles. Further research is needed to assess whether these changes are clinically significant, as there is a possibility that they may decrease the overall estrogenic impact of MHT on key clinical endpoints such as symptom improvement and bone mineral density. The presence and magnitude of these changes suggest that providers treating postmenopausal women with MHT should ask their patients if they are taking a DIM supplement and, if so, consider the potential implications of drug-supplement interactions for MHT dose management and effectiveness.

Drug-drug interactions are relatively well-defined and documented for menopausal hormone therapy (MHT).¹ Another important but underexplored source of potential interactions exists with herbs and supplements. One such supplement, 3,3’-diindolylmethane (DIM), is a compound derived from cruciferous vegetables, such as bok choy, broccoli, cabbage, cauliflower, or kale,² and is an active area of research due to its influence on estrogen metabolism pathways.^2–5 Indole-3-carbinol (I3C) is another supplement that, when given orally, is rapidly converted to DIM which is its major active product in vivo.⁶ As the vast majority of the effects of I3C come from DIM, we will use DIM in this study to refer to supplements labeled I3C or DIM.

The only herb/supplement-drug interaction mentioned in the package inserts of transdermal estradiol (E2) patches is St. John’s wort (Hypericum perforatum). The package insert text warns that since St. John’s wort is an inducer of cytochrome P450 (CYP) 3A4, coadministration “may reduce plasma concentrations of estrogens, possibly resulting in a decrease in therapeutic effects and/or changes in the uterine bleeding profile.”^7,8 There has yet to be much in the published literature or package inserts of MHT products warning about a potential interaction with the compound DIM. However, the package inserts for transdermal E2 patches do state that “inducers and/or inhibitors of CYP3A4 may affect estrogen drug metabolism.”^7,8 Although DIM is an inducer of CYP3A4, it is considered to be a stronger inducer of CYP1A1 and CYP1A2.^9,10 It is not completely clear which component(s) of DIM’s mechanism of action is responsible for the varying effects it has on estrogen metabolism; however, studies suggest that DIM has significant impacts on estrogen metabolism in premenopausal women not using MHT.^11–14 Similarly, it is not clear how increases or decreases in different estrogen metabolites correspond with therapeutic outcomes in postmenopausal women being treated with MHT, but it is clear that transdermal MHT increases all estrogen metabolites.^15–17 Considered together, the actions of DIM and MHT on estrogen metabolism allude to, at minimum, a theoretical drug-supplement interaction.

In recent years, there has been an increase in patients seeking out functional or integrative medicine providers and treatments. Simultaneously, there has also been a recent increase in the consumption of herb/supplement products and an accompanying increase in the amount of concern around unknown herb/supplement-drug interactions.^18,19 Importantly, it has been observed that when patients are taking an herb/supplement and a prescription medication simultaneously, they may stop taking the prescription medication citing a perceived lack of effect instead of the more likely issue, an interaction.¹⁸ A recent review by Hlengwa et al,¹⁹ which focused on supplement interactions with estrogen-based oral contraceptives, should serve as inspiration for menopause providers to explore similar interactions with estrogen-containing MHT.

DIM is commonly recommended by functional or integrative medicine providers with the intent to positively influence estrogen metabolism to favor preferred metabolic pathways.⁴ While some of these providers may not be aware of potential interactions with MHT, others may be aware and intentionally recommend DIM in an attempt to address a potential risk associated with a patients’ estrogen detoxification pathways, primarily breast cancer.^2,20–22 For example, if a woman’s estrogen detoxification pathway favors 4-hydoxyestone (4-OHE1) or 16-hydroxyestrone (16-OHE1), coadministration of DIM may be beneficial. Prior studies have shown that the administration of DIM positively impacts the 2-OHE1:16-OHE1 and the 2-OHE1:4-OHE1 ratios, which have been suggested to be surrogate markers for breast cancer risk.^3,23–25

To the best of our knowledge, no previous research has been published examining the potential for interaction between DIM and MHT. We have previously shown that urinary E2 and estrogen metabolites increase as expected with increasing doses of transdermal E2 patches and provide a useful approximation of estrogen exposure.¹⁶ Previous studies have also suggested that DIM decreases endogenous estrogen in premenopausal, normally cycling women.^11–14 The aim of this study was to determine if the use of DIM has similar effects on exogenous estrogen in postmenopausal women being treated with MHT in the form of a transdermal E2 patch. In addition, we wanted to determine if there is a significant difference in estrogen exposure between women using a transdermal E2 patch and DIM concurrently and women using a transdermal E2 patch alone.

METHODS

Data source

These data and analyses reported here are a substudy from a larger retrospective, observational, data mining study registered as Precision Analytical Retrospective Data Correlation (Clinical Trials ID, NCT04305093). There was no enrollment of participants, as this was a retrospective analysis of previously collected, deidentified clinical laboratory test results. In the larger database, data were available for 144,561 laboratory accessions from 129,883 individuals between January 1, 2016 and December 9, 2019. The study was approved by the National University of Natural Medicine Institutional Review Board, and since all data were deidentified, the Institutional Review Board determined that written informed consent could be waived. To arrive at the final analytic sample size, we filtered the original sample of 129,883 individuals by those who are postmenopausal and using an E2 patch. In that sample of postmenopausal women using an E2 patch, we then applied our exclusion criteria (listed later), which yielded the final sample size of 1,458 participants. Deidentified data were extracted from the database and included the following variables: age, sex, menstrual status (regular, irregular, and none), last menstrual period, body mass index (BMI) calculated from self-reported height and weight, use of DIM, comorbidities, and urinary measures of estrone (E1), E2, estriol (E3), α-pregnanediol and β-pregnanediol, 2-hydroxyestrone, 2-hydroxyestradiol, 4-hydroxyestrone, 4-hydroxyestradiol, 16-hydroxyestrone, and 2-methoxyestrone. Data on race/ethnicity were not available. Exclusion criteria for this substudy included the following: a BMI <16 or >60 kg/m², pregnancy, any reported kidney disease, reported adrenal insufficiency or excess, suspected or diagnosed polycystic ovary syndrome, missing data for urinary estrogens or estrogen metabolites (with the exception of 4-hydroxyestradiol and 2-methoxyestradiol), use of any MHT other than a transdermal E2 patch, use of β-human chorionic gonadotropin, use of an aromatase inhibitor, use of a selective estrogen receptor modulator, or evidence of overly dilute urine (any spot urine creatinine <0.1 mg/mL). All of the abovementioned relevant information came from patient requisition forms, which were submitted at the same time as the urine samples.

Sample collection

The urine sample collection and assay methods used to produce the data analyzed in this retrospective study, including coefficients of variation for quality control samples and other assay performance characteristics, have been described previously in analytical validation studies.^26,27 Participants were instructed to collect a total of 4 urine samples throughout the day according to the following timeline: (1) the first urine of the day, (2) 2 hours after waking, (3) in the afternoon (∼5 pm), and (4) before bedtime (∼10 pm). This general timeline was adhered to by all participants included in the analysis. The sample collection method has been previously described.^26,27 Briefly, samples were collected by saturating strips of filter paper (Whatman Body Fluid Collection Paper; Sigma-Aldrich) with urine either by urinating directly on the filter paper or soaking the filter paper in a clean collection cup. The strips of filter paper were allowed to dry at room temperature for 24 hours and then shipped to the laboratory for analysis.

Laboratory methods

Urinary estrogen concentrations were determined using a validated method that has been previously described in detail elsewhere.^21,24 Briefly, urine samples were extracted from the filter paper using ammonium acetate, and the 4 samples were combined into a single representative sample. This combination was performed by using creatinine concentrations from each of the 4 samples to determine the sample volume that should be added to the combined sample to best represent measured concentrations over a 24-hour period. The combined sample was then analyzed through gas chromatography-tandem mass spectrometry on an Agilent 7890/7000B (Agilent Technologies). Estrogen concentrations were normalized to creatinine to account for variation in urine concentration and filter paper saturation.

Statistical analysis

Comparisons of estrogen metabolite concentrations (and the 2-OHE1:16-OHE1 ratio) between women using DIM and women not using DIM were done with the Wilcoxon rank sum test. Multivariable linear regression models were created to confirm the impact of DIM on estrogen metabolite concentrations (and the 2-OHE1:16-OHE1 ratio) after accounting for covariates including age, BMI, transdermal E2 patch dose, and urinary creatinine. Estrogen metabolite concentrations were log-transformed to meet the assumptions for linear regression. Alpha was set at 0.05. As the study design is retrospective and observational, all analyses were exploratory and hypothesis-generating in nature. All analyses were conducted using R version 4.3.1 (R Software for Statistical Computing).

RESULTS

Characteristics (age and BMI) and E2 patch doses for the 1,458 postmenopausal women included in the analysis are shown in Table 1. Urinary estrogen metabolite concentrations for this sample of postmenopausal women on transdermal E2 patches with or without DIM are shown in Table 2. All variables are presented as median (interquartile range) as none were normally distributed. A comparison of distributions (E2 patch only vs E2 patch + DIM) of select estrogen metabolites is shown by a dose of E2 patch in Figure 1 for E1 concentrations, Figure 2 for E3 concentrations, Figure 3 for 2-OHE1, Figure 4 for 16-OHE1, and Figure 5 for total urinary estrogen concentrations. When compared with women using a transdermal E2 patch alone, women who were concurrently taking DIM had alterations in their urinary estrogen profile that included lower E1, lower E3, higher 2-OHE1, higher 4-OHE1, lower 16-OHE1, higher 2-OHE2, and lower total estrogens (P values are shown in Table 2). To confirm these effects of DIM on estrogen metabolite concentrations, multivariable linear regression models were used that accounted for age, BMI, E2 patch dose, and urinary creatinine. These results indicated that DIM had a significant effect on the concentrations of E1 (P = 0.004), E3 (P < 0.001), 2-OHE1 (P = 0.004), 2-OHE2 (P = 0.004), 4-OHE2 (P = 0.01), 16-OHE1 (P = 0.02), and the 2-OHE1/16-OHE1 ratio (P < 0.001); however, in the models for E2 (P = 0.19) and 4-OHE1 (P = 0.10), the effect of DIM did not reach the threshold for significance. Beta estimates and 95% CIs for the associations of DIM with each estrogen metabolite are shown in Table 3.

TABLE 1.

Characteristics of postmenopausal women using transdermal E2 patches with or without DIM

	Therapy received
Variable	Patch only, N = 1,350a	Patch + DIM, N = 108a
Age at collection	57 (53, 61)	58 (54, 64)
BMI	23.8 (21.5, 26.3)	22.8 (20.6, 25.9)
Estradiol patch dose (mg)
0.025	261 (19)	23 (21)
0.0375	300 (22)	16 (15)
0.05	475 (35)	36 (33)
0.075	164 (12)	19 (18)
0.1	150 (11)	14 (13)

BMI, body mass index; DIM, 3,3’-diindolylmethane.

^a Median (interquartile range); n (%).

TABLE 2.

Urinary estrogen metabolite concentrations in postmenopausal women using transdermal E2 patches with or without DIM

	Therapy received
Urinary estrogen metabolite (or metabolite ratio) (ng/mg-Cr)	Patch only, N=1,350a	Patch + DIM, N=108a	P b
Estrone	8.30 (5.10, 12.60)	6.40 (4.20, 9.90)	0.002
Estradiol	1.58 (0.91, 2.60)	1.32 (0.87, 2.44)	0.2
Estriol	3.80 (2.20, 6.50)	2.60 (1.48, 4.45)	<0.001
2-hydroxyestrone	3.16 (1.73, 5.10)	3.92 (2.32, 7.01)	0.002
4-hydroxyestrone	0.40 (0.23, 0.64)	0.48 (0.29, 0.85)	0.039
16-hydroxyestrone	0.40 (0.21, 0.70)	0.28 (0.16, 0.58)	<0.001
2-methoxyestrone	1.45 (0.89, 2.30)	1.39 (0.80, 2.29)	0.5
2-hydroxyestradiol	0.28 (0.14, 0.48)	0.40 (0.17, 0.68)	0.003
4-hydroxyestradiol	0.10 (0.10, 0.20)	0.10 (0.10, 0.20)	0.051
Missing values	386	28	–
2-methoxyestradiol	0.30 (0.20, 0.50)	0.30 (0.10, 0.40)	0.3
Missing values	382	26	–
Total estrogens	20.90 (14.00, 31.30)	17.95 (13.50, 27.40)	0.044
2-OHE1:16-OHE1 ratio	8.24 (4.26, 14.45)	16.57 (4.71, 28.08)	<0.001
Missing values	2	1	–

DIM, 3,3’-diindolylmethane.

^a Median (interquartile range).

^b Wilcoxon rank sum test.

FIG. 1.

Urinary estrone concentration distributions in postmenopausal women with or without DIM by dose of transdermal estradiol patch. DIM, 3,3’-diindolylmethane.

FIG. 2.

Urinary estriol concentration distributions in postmenopausal women with or without DIM by dose of transdermal estradiol patch. DIM, 3,3’-diindolylmethane.

FIG. 3.

Urinary 2-hydroxyestrone concentration distributions in postmenopausal women with or without DIM by dose of transdermal estradiol patch. DIM, 3,3’-diindolylmethane.

FIG. 4.

Urinary 16-hydroxyestrone concentration distributions in postmenopausal women with or without DIM by dose of transdermal estradiol patch. DIM, 3,3’-diindolylmethane.

FIG. 5.

Total urinary estrogen concentration distributions in postmenopausal women with or without DIM by dose of transdermal estradiol patch. DIM, 3,3’-diindolylmethane.

TABLE 3.

Beta estimates and 95% CIs for the associations of DIM with estrogen metabolites and the 2/16 ratio

Metabolite or metabolite ratio (log-transformed)	Beta estimate	95% CI lower	95% CI upper
Estrone	−0.081513972	−0.14693177	−0.01609617
Estradiol	−0.053596181	−0.12658149	0.01938913
Estriol	−0.186321820	−0.27372969	−0.09891395
2-hydroxyestrone	0.119350068	0.04069780	0.19800233
4-hydroxyestrone	0.060032550	−0.01834412	0.13840922
16-hydroxyestrone	−0.124262098	−0.21989829	−0.02862591
2-methoxyestrone	−0.023159314	−0.09930479	0.05298616
2-hydroxyestradiol	0.103523101	0.01136707	0.19567913
4-hydroxyestradiol	0.103621294	0.04270567	0.16453692
2-methoxyestradiol	0.007379363	−0.07577119	0.09052992
2-OHE1:16-OHE1 ratio	0.243612165	0.13153929	0.35568504

Adjusted for age, body mass index, E2 patch dose, and urinary creatinine.

DIM, 3,3’-diindolylmethane.

DISCUSSION

Some postmenopausal women receiving MHT concurrently use DIM, both of which significantly impact estrogen metabolism. To our knowledge, this is the first study to investigate the possibility of a drug-supplement interaction between MHT and DIM, with potential clinical implications for MHT dose management and effectiveness. In the current study, we found that women who took a DIM supplement while using a transdermal E2 patch had significant differences in their urinary estrogen profile compared with women using a transdermal E2 patch alone. Overall, the women using DIM concurrently had lower total estrogen exposure; however, there was some nuance regarding which metabolites were higher or lower relative to women using the patch alone. Interestingly, for E1, E3, and total urinary estrogen, the difference in distributions between the patch alone and patch-plus-DIM groups seemed to get larger as the E2 patch dose increased. This could point to a phenomenon in which DIM’s effect on estrogen metabolism is enhanced when larger amounts of exogenous E2 are present. Further, more targeted research is needed to explore that hypothesis. Nevertheless, these findings suggest that a drug-supplement interaction is present, which is in line with the warning in the prescribing information of transdermal E2 patch products concerning CYP3A4 inducers and inhibitors.^7,8 The impact of DIM on CYP3A4 can be seen in the decrease in 16-OHE1 in the women who were using DIM and an E2 patch concurrently (Fig. 4). Similarly, DIM’s impact on CYP1A1 can be seen in the increase in 2-OHE1 (Fig. 3). Given the observed impact related to CYP1A1 activity, it may be prudent to consider adding a warning about CYP1A1 inducers and inhibitors to the prescribing information of E2 patch products. This finding is timely, as other investigators are also suggesting that clinicians need to consider drug interactions beyond the inducers and inhibitors of the CYP3A family of enzymes for 17α-ethinylestradiol-containing oral contraceptives.²⁸ Our findings suggest that similar scrutiny is needed for E2 formulations used in MHT.

The major clinical implication that can be derived from the results of this study is that menopause providers should consider asking their patients what non-prescription herbs or supplements they are taking and specifically ask about DIM or I3C supplements. Depending on clinical observations and patient preferences, a dose adjustment or therapy change may be warranted. It is important for clinicians to realize that the solution may not be as simple as directing the patient to discontinue the DIM supplement. For a patient whose genetic profile causes their estrogen metabolism to prefer the 4-OHE1 pathway, DIM supplementation may have a positive impact, increasing the metabolism of exogenous estrogen through a detoxification pathway with more stable intermediates or, alternately, a pathway with less reactive quinone intermediates.^3,4 Clinicians must approach this scenario with compassion, cultural competence, and as a partner in the patient’s menopause journey. This approach may require the clinician to consider that the patient may not be willing to discontinue DIM. The goal in such a scenario would be to optimize therapeutic outcomes, such as symptom improvement while considering the impact of DIM on estrogen exposure.

Because there is not yet an established correlation between symptom improvement and estrogen exposure, there is no definition for what represents a clinically significant alteration in a woman’s estrogen profile. In regard to the goals of therapy for DIM, multiple studies have been published in the literature that suggest the use of the 2-OHE1/16-OHE1 ratio as a biomarker for cancer risk, with the goal of DIM use being to increase this ratio.^{2,4,21,23,25,29} However, this scenario also does not have an established target level for the ratio that is correlated with avoiding an adverse outcome. The values of this ratio obtained in our lab are different than the values seen in the literature; much of the reason for this is related to methodological advancements and comparing liquid chromatography-tandem mass spectrometry to gas chromatography-tandem mass spectrometry. More work and increased laboratory method standardization are needed before clinically useful parameters can be established in these 2 scenarios.

One limitation of this study was that we were unable to collect information on the dose or formulation of the DIM/I3C supplements used by patients. Although there is a limited range of doses available, there are some purported differences in bioavailability between different formulations.²⁹ Novel methods have previously been used to quantify urinary metabolites of DIM,³⁰ which would be useful for measuring the level of DIM exposure in studies where dose-related information is not available. However, for this study, we were unable to measure urinary DIM levels following consumption of DIM supplements in women reporting the use of the E2 patch, as this was a retrospective analysis. Since the impact of DIM is likely dose-dependent, the nonstatistically significant drop in E2 may become clinically significant in regard to estrogen therapy dosing if the dose of DIM is relatively high and/or if the individual patient’s response to DIM is particularly strong. Related to this, we also did not have data on the frequency of administration or adherence to either DIM or the transdermal E2 patch, the brands/types of E2 patches, or concurrent use of progesterone. We did not collect data on smoking status, race, or ethnicity, all of which may affect 2-hydroxylation polymorphisms.^31,32 In addition, we did not have any genetic data available, and single nucleotide polymorphisms for CYP1A1 increase 2-hydroxylation, and, in combination with DIM, the alteration of estrogen metabolism would potentially be more extreme than what would be seen with either alone. While we did attempt to collect data on all medications and supplements women were taking, the data were self-reported, so it is possible that some were not captured. In functional medicine practice, other “estrogen lowering” supplements, such as calcium-d-glucarate, are sometimes given concurrently with DIM. If any of those types of supplements were taken by the women included in this study, it is possible that some of the changes seen in the urinary estrogen profile may be a result of synergy between DIM and the other supplements. While the likelihood of a functional medicine provider recommending concurrent calcium-d-glucarate to a premenopausal woman with abnormally elevated E2 is relatively high, it is far less likely that it would be recommended to a postmenopausal woman prescribed estrogen therapy. Finally, all analyses in this study are considered exploratory; therefore, the observed results will need to be confirmed in subsequent studies. However, our results do provide strong evidence that should be incorporated into current clinical practice.

Despite these limitations, the results of this study show that the impact of exogenous estrogen can be decreased with the use of DIM. In future studies, it may be useful to explore DIM’s effect on the estrogen profile and symptom management when patients are being treated with nonhormone menopause therapies (such as fezolinetant). It would also be beneficial to explore if any uterine bleeding changes are seen with the use of DIM as this is listed as a possible effect of CYP3A4 inducers taken concurrently with E2 patches.^7,8 Inadvertently inducing what would be considered abnormal uterine bleeding in postmenopausal women could represent a significant cost to the health care system if referrals to a gynecologist are initiated. An additional issue worth raising is interindividual variability in the absorption of E2 from transdermal patches. Some postmenopausal women on transdermal E2 patches have been found to be poor absorbers of E2,³³ and consideration of this subgroup of E2 patch users in future analyses of DIM’s effects may be warranted. Since all women included in this study were postmenopausal, another unanswered question relates to the potential effect of DIM in perimenopausal women. Given that the perimenopausal period is characterized by frequent and rapid fluctuations in estrogen levels,³⁴ does DIM provide any meaningful benefit by potentially stabilizing estrogen concentrations? Finally, some dietary interventions or diet choices, such as veganism or vegetarianism, likely result in the consumption of high levels of DIM through cruciferous vegetables. Another future research question to explore is what the impact of diet is on symptom presence or symptom severity, both with or without MHT and in both perimenopausal and postmenopausal women.

It is important to acknowledge that the differences observed in this study are associated with changes to estrogen metabolism and may have little to no impact on the efficacy of a transdermal E2 patch. With respect to a potential impact of changes to estrogen metabolism on menopausal symptoms, some related evidence suggests that vasomotor symptom severity is negatively associated with levels of certain urinary estrogen metabolites³⁵ and that genetic variation in estrogen metabolism pathways may contribute to interindividual variation in vasomotor symptom prevalence.^36,37 Determining whether or not altering estrogen metabolism in postmenopausal women using an E2 patch has an impact on vasomotor symptom improvement or any other desired therapeutic outcome would need to be evaluated in a future study.

CONCLUSION

In summary, this study has uncovered a potentially significant drug-supplement interaction between transdermal E2 patch therapy and DIM. We posit that this interaction is likely also present when women are using other transdermal forms of E2 such as E2 gels. However, given that oral E2 undergoes significant first-pass metabolism in the liver, the impact of DIM would potentially be different if taken concurrently in that scenario. To both reinforce these findings and elucidate more nuanced insights concerning the mechanism of the interaction, larger and more intentionally designed studies are warranted. However, until then, clinicians should be aware of the potential for this interaction and appropriately screen and educate patients who are prescribed transdermal MHT.