Volume 17, Issue 3 e13721
ARTICLE
Open Access

Bidirectional pharmacokinetics of doravirine, tenofovir, and feminizing hormones in transgender women (IDentify): A randomized crossover trial

Kevin Lam

Kevin Lam

Department of Pharmacology, Physiology, and Cancer Biology, Thomas Jefferson University, Philadelphia, Pennsylvania, USA

Search for more papers by this author
Walter K. Kraft

Corresponding Author

Walter K. Kraft

Department of Pharmacology, Physiology, and Cancer Biology, Thomas Jefferson University, Philadelphia, Pennsylvania, USA

Correspondence

Walter K. Kraft, Department of Pharmacology, Physiology, and Cancer Biology, Thomas Jefferson University, 132 South 10th Street, 1170 Main Building, Philadelphia, PA, USA.

Email: [email protected]

Search for more papers by this author
Tingting Zhan

Tingting Zhan

Department of Pharmacology, Physiology, and Cancer Biology, Thomas Jefferson University, Philadelphia, Pennsylvania, USA

Search for more papers by this author
Edwin Lam

Edwin Lam

Clinical Pharmacokinetics Research Lab, National Institutes of Health, Bethesda, Maryland, USA

Search for more papers by this author
First published: 29 February 2024

Abstract

Transgender women may have concerns of drug interactions between feminizing hormone therapy (FHT) and antiretrovirals, leading to nonadherence. This randomized, three-period crossover, open-label, phase I trial assessed the effects of doravirine (DOR) and tenofovir disoproxil fumarate (TDF) on the pharmacokinetics (PKs) of estradiol, spironolactone, and total testosterone and vice versa in healthy transgender women. Volunteers were randomized 1:1 into two sequences containing three treatment groups (DOR, lamivudine [3TC], and TDF alone; estradiol, spironolactone, and placebo; and DOR/3TC/TDF, estradiol, and spironolactone). Eight subjects enrolled in the study and six had completed all study periods. The geometric mean ratios for DOR area under the concentration-time curve from zero to last measured concentration (AUC0-last), maximum concentration (Cmax), and concentration at 24 h (C24) were similar. However, tenofovir (TFV) AUC0-last, Cmax, and C24 moderately increased by 14%–38%. Last, estradiol AUC0-last, Cmax, and C24 were increased by 10%–13%. Whereas most 90% confidence intervals did not meet the bioequivalence bounds of 80%–125%, the point estimates fell within the intervals. Log-transformed DOR, TFV, and estradiol PK parameters computed with and without co-administration were not statistically different (p > 0.05). There were no serious adverse events. There is not a clinically significant impact of FHT on DOR/TFV PKs. Similarly, there is no observed impact on estradiol PKs and total testosterone following use of DOR/3TC/TDF.

Study Highlights

  • WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Transgender women are disproportionately affected by human immunodeficiency virus-1 (HIV-1), and many have concerns of drug–drug interactions between antiretrovirals and feminizing hormone therapy (FHT). Doravirine (DOR) is a non-nucleoside reverse transcriptase inhibitor available as a single fixed-dose tablet with lamivudine (3TC) and tenofovir disoproxil fumarate (TDF). Despite increasing data demonstrating no clinically significant interaction between tenofovir (TFV)-containing pre-exposure prophylaxis regimens, there are limited studies assessing antiretrovirals for initial treatment such as DOR in the presence of FHT and vice versa.

  • WHAT QUESTION DID THIS STUDY ADDRESS?

This study evaluated the effects of concomitant estradiol and spironolactone on the pharmacokinetics (PKs) of DOR and TFV and vice versa. Additionally, effects on total testosterone as an indirect assessment of FHT were evaluated.

  • WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

Co-administration of estradiol and spironolactone decreased area under the concentration-time curve from zero to last measured concentration (AUC0-last), maximum concentration (Cmax), and concentration at 24 h (C24) of DOR and TFV by 2%–7% and 14%–38%, respectively. Co-administration of DOR/3TC/TDF increased AUC0-last, Cmax, and C24 by 10%–13%. Combined with the lack of breakthrough masculinization, PK changes are unlikely to be clinically relevant overall.

  • HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

Observed PK changes in DOR, TFV, estradiol, and total testosterone are not expected to be clinically significant. These data suggest the combination is unlikely to affect viral suppression, feminization, and cause intolerability or breakthrough masculinization. These data can be used to inform patients and healthcare providers.

INTRODUCTION

The global prevalence of human immunodeficiency virus type 1 (HIV-1) infection is estimated to be 19% among people who identify as a transgender woman (TGW).1 TGW have ~49-fold higher risk for HIV-1 infection compared to the general population and a higher prevalence than transgender men.2-4 Antiretroviral (ARV) therapy is recommended to those living with HIV-1 to prevent progression to acquired immunodeficiency syndrome, transmission, and premature death.5 TGW living with HIV-1 infection express concerns over hypothetical drug–drug interactions with feminizing hormone therapy (FHT). In a cross-sectional study, 57% of TGW participants with HIV-1 have reported this concern to their healthcare provider with 40% citing this as a reason to not use ARV therapies, FHT, or both.6 FHT is typically comprised of an estrogen (e.g., 17-β estradiol) and androgen blocker (e.g., spironolactone and finasteride) with or without a progestogen.7, 8

Doravirine (DOR) is a non-nucleoside reverse transcriptase inhibitor (NNRTI) approved in 2018 for the treatment of HIV-1 infection in combination with other ARVs in adults. It is available as a complete, fixed-dose combination tablet containing DOR 100 mg, lamivudine (3TC) 300 mg, and tenofovir disoproxil fumarate (TDF) 300 mg, making it a convenient therapeutic option for ARV-naïve patients. Although reserved as an initial option under certain clinical scenarios in the United States, DOR/3TC/TDF is a recommended first-line therapy in Europe.5, 9 Despite the low usage of DOR- and NNRTI-based regimens in 2020 (1% and 11% of persons living with HIV are prescribed ARV therapy, respectively), the metabolic profile of DOR is favorable to patients with dyslipidemias, especially for TGW, as they are at an increased risk for cardiovascular events due to HIV-1 infection and hormone therapy.10-15 The effects of chronic estradiol in TGW on ARV clearance mechanisms are not well understood. Chronic estradiol elevations have been shown to upregulate expression of transcription factors, resulting in the induction of CYP enzymes.16-18 Furthermore, spironolactone has been shown to induce drug metabolizing enzymes and transporters through similar mechanisms.19-21 These findings suggest the possibility of drug–drug interactions between DOR and hormones used by TGW. Although not significant, a 12% reduction in tenofovir (TFV) area under the curve from time zero to 24 h (AUC0-24) with concomitant estradiol valerate and cyproterone acetate with no change in hormone levels has been reported.22 Whereas most studies assessing the effects of hormones in TGW involved TFV-containing pre-exposure prophylaxis (PrEP), limited data exist on FHT effects on a contemporary ARV regimen for treatment. There are also limited data regarding the effects of an ARV regimen on the PKs of FHT.

The primary objective of this study was to evaluate the effects of administering DOR/3TC/TDF on the PKs of estradiol/spironolactone in TGW and vice versa. The secondary objective was to assess the safety and tolerability of co-administered DOR/3TC/TDF, estradiol, and spironolactone. Last, the exploratory objective was to evaluate the change in total testosterone concentrations as a measure of the effects of FHT following concomitant DOR/3TC/TDF administration.

METHODS

Study design and participants

This was a phase I, single-center, randomized, placebo-controlled, open-label, three-period crossover, two-way drug interaction trial conducted at the Clinical Research Unit at Thomas Jefferson University. Participants were 18–45 years old, HIV-1 negative, and self-identified as a TGW (male-to-female) at screening. Recruitment was achieved through social media groups such as Grindr and Facebook as well as partnerships with the LGTBQ+ community centers and healthcare providers. Participants must have no history of an orchiectomy and have received estradiol and/or spironolactone (oral and/or injectable) for at least 3 months prior to beginning the study with a self-reported adherence of at least 90% to prescribed doses. All participants agreed to briefly interrupt hormonal therapy as well as discontinue any PrEP therapy containing tenofovir alafenamide or TDF at least 14 days (i.e., 4-5 half-lives) prior to and during the study. Exclusion criteria included current use of any ARV (other than for PrEP as mentioned) that has not been discontinued within 30 days of screening, use of any hormonal replacement therapy other than estradiol and/or spironolactone, estimated creatinine clearance of 60 mL/min or less by the Cockcroft-Gault equation (using female as the sex modifier), and receipt of an investigational drug from another study within 4 weeks or five half-lives (whichever came first) before the first anticipated dose of study drug in this study.

This study was approved by the Thomas Jefferson University Institutional Review Board. The study was conducted in accordance with Good Clinical Practice standards and applicable federal and/or local regulatory requirements. All participants provided written informed consent prior to beginning the study. This trial is registered on clinicaltrials.gov (NCT04283656).

Randomization and masking

Eligible participants were assigned (1:1) one of two sequences (E or F) containing three treatment-specific groups via simple randomization. Participants received a treatment-specific group for each period (5 days each) with a 14-day washout in between periods I and II as well as II and III. Participants and staff were not masked to treatment allocation.

Procedures

Each participant received their treatment per their specific groups (treatment A, B, and C) during each period with a 14-day washout in between periods (Figure S1). Treatment A was a single oral fixed-dose of DOR/3TC/TDF 100 mg/300 mg/300 mg combination tablet alone. Treatment B was a standardized hormone therapy, consisting of estradiol 4 mg twice daily, spironolactone 200 mg twice daily, and placebo given orally for 1 day. Treatment C was the co-administration of a single dose of DOR/3TC/TDF 100 mg/300 mg/300 mg, estradiol 4 mg twice daily, and spironolactone 200 mg twice daily for 1 day. All participants discontinued their hormone therapy and were randomized to either sequence E or F 14 days prior to starting the study. Urine drug screen, urinalysis, physical examinations, and blood draw for estradiol level were done for all participants 1 day prior to beginning each period. Participants who were randomized to sequence E received one treatment group per period (i.e., treatment A, B, or C) as follows. On the first study day (period I), participants received treatment A with 8 ounces of water following an 8-h overnight fast. Blood draws for DOR/TFV and testosterone PKs were done 0.5-, 1-, 2-, 6-, 12-, 24-, 48-, 72-, and 96 h postdose. Following a 14-day washout period, participants received treatment B with 8 ounces of water and had estradiol PK blood draws at the same timepoints for period II. Participants received treatment C for period III, which involved all study activities performed in the previous two periods. Participants who were randomized to sequence F received their treatment groups in reverse order (i.e., treatment C, B, and A for periods I, II, and III, respectively). Research nurses and/or pharmacist directly observed the administration of study doses during designated days participants domiciled in the Clinical Research Unit.

Plasma DOR and TFV concentrations were determined using an ultra-performance liquid chromatography tandem mass spectrometry (Syneos Health Clinique). The lower limit of quantification was 1 ng/mL over a calibration range of 1–1000 ng/mL and 0.5 ng/mL of a range of 0.5–500 ng/mL for DOR and TFV, respectively. Plasma estradiol and total testosterone concentrations were determined using Roche electrochemiluminescence immunoassays (Roche Diagnostics) at the Jefferson University Hospital Clinical Chemistry Laboratory. The lower limit of quantification was 5 pg/mL and 12.0 ng/dL for estradiol and testosterone, respectively, with no upper limit. There was no interference noted in the DOR and TFV assays when all analytes were spiked in test samples. Plasma 3TC was not assessed in this study.

Safety and tolerability were assessed throughout the study, including monitoring for adverse events (AEs), physical examinations, vital signs, electrocardiograms, complete blood count with differential, and complete metabolic panel. AEs were coded using MedDRA Dictionary version 26.0 (MedDRA MSSO) and summarized using descriptive statistics.

Outcomes

The primary end point was the change in PK parameters for DOR/TFV and estradiol alone and during co-administration, measured as the AUC from time zero to last measured concentration (AUC0-last), maximal concentration (Cmax), and the concentration at 24 h (C24). The tolerability to co-administered drugs and proportion of participants with treatment-emergent AEs by type and severity were secondary end points. The exploratory end point was the change in total testosterone concentration before and after co-administration.

Statistical analysis

A sample size of 10 participants was determined to achieve 80% power to reject the null hypothesis (the mean ratios for DOR/TFV and estradiol AUC and Cmax will fall outside of the no-effect boundaries of 80%–125%) with a two-sided significance level of 0.05 assuming a mean ratio of 1 for no effect. Sample size calculations accounted for intrasubject variability in DOR, TFV, and estradiol AUC0-last, AUC extrapolated to infinity (AUC0-inf), and Cmax based on prior PK studies.23-25 Accounting for 20% dropout, 12 participants was identified as the target sample size.

Plasma concentrations of DOR and TFV with and without estradiol and spironolactone were utilized to calculate PK parameters for each participant via a noncompartmental analysis using PKanalix version 2023R1 (Lixoft). Similarly, plasma estradiol concentrations with and without DOR/3TC/TDF were used to calculate PK parameters for each participant. PK parameters included AUC0-last, AUC0-inf, C24, Cmax, time to maximum concentration (Tmax), apparent oral clearance (CL/F), and apparent half-life (t1/2). AUC0-last was computed using the linear-up, log-down trapezoidal integration method, whereas other parameters were estimated with a log-linear regression via a least-squares method. R version 4.2.2 (R Foundation) was used to compute and compare log-transformed DOR/TFV PK parameters and 90% confidence intervals (CIs) computed in the absence and presence of estradiol and spironolactone through two one-sided t-tests. These were compared to the log-transformed no-effect boundaries of log(0.8) and log(1.25). Additionally, Wilcoxon signed rank test of the paired differences (between treatment and reference) in Tmax was performed. For each PK parameter, a p value of less than 0.05 was considered statistically significant. Geometric mean ratios (GMRs) with 90% CIs were constructed for DOR, TFV, and estradiol with and without co-administration (test and reference, respectively). The mean total testosterone concentration-time curve with and without co-administration was constructed using R.

RESULTS

Between February 14, 2022, and December 30, 2022, eight participants were enrolled and six had completed the study; one participant withdrew due to an AE and one was lost to follow-up, as shown in Figure 1. The overlap with study start and the coronavirus disease 2019 (COVID-19) pandemic led to less than planned for subject enrollment. Subject characteristics are presented in Table 1. Six participants were included in the PK analyses, but all eight participants were included in the safety population. Due to the reduced sample size, a post hoc power analysis determined that a sample size of six subjects would achieve 80% power to detect an effect size of 1.43 or greater, using two-sided paired t-test on the (log-transformed) PK parameters at 5% significance level.

Details are in the caption following the image
Study flow.
TABLE 1. Baseline characteristics at enrollment.
Overall (n = 8) Sequence E (n = 4) Sequence F (n  = 4)
Age (years) 25.5 (24–38) 25.5 (25–38) 27 (24–31)
Weight (kg) 90.6 (56.9–132.2) 70.1 (56.9–83.8) 111.5 (97.4–132.2)
Height (cm) 181.25 (165.5–190.5) 173.3 (165.5–185) 181.5 (181–190.5)
Body mass index (kg/m2) 28.25 (20.8–39.9) 22.2 (20.8–28.3) 32.6 (28.2–39.9)
Race; n (%)
Non-Hispanic White 8 (100) 4 (100) 4 (100)
Non-Hispanic Black 0 (0) 0 (0) 0 (0)
Asian 0 (0) 0 (0) 0 (0)
Mixed or other 0 (0) 0 (0) 0 (0)
Latino 0 (0) 0 (0) 0 (0)
  • Note: All data are expressed as median (range) except race, which is expressed as median (SD).

Geometric mean plasma concentrations (±SD) for DOR, TFV, and estradiol were plotted against time, as seen in Figures S2–S4. Co-administration of estradiol and spironolactone resulted in small decreases in DOR AUC0-inf, AUC0-last, C24, Cmax, and t1/2 by 3.3%, 2.5%, 3.1%, 6.8%, and 3.2%, respectively (Table 2). The CL/F slightly increased by 3.5% and Tmax was similar with and without co-administration. The co-administration of estradiol and spironolactone resulted in larger increases in TFV AUC0-inf, AUC0-last, C24, Cmax, by 16.7%, 19.4%, 13.8%, and 37.6%, respectively, whereas CL/F and t1/2 decreased by 14.3% and 10.2%, respectively (Table 3). Tmax was the same before and after co-administration. The co-administration of DOR/3TC/TDF led to moderate decreases in estradiol AUC0-inf and t1/2 by 3.17% and 22.5% respectively (Table 4). Estradiol AUC0-last, C24, and Cmax moderately increased by 10.2%, 13.1%, and 12.3%, respectively, but Tmax was shorter after co-administration. Log-transformed DOR, TFV, and estradiol PK parameters compared to log-transformed no-effect boundaries are also presented in Figures S5–S7. The mean total testosterone (±SD) concentrations were plotted against time, as shown in Figure 2. Upon visual inspection, co-administration of DOR/3TC/TDF with estradiol and spironolactone produced lower mean total testosterone concentrations than with estradiol and spironolactone alone. However, both curves were generally similar to each other. All calculated 90% CIs fell outside of the predefined criteria, but the differences in log-transformed PK parameters were not statistically significant and the point estimates fell within the bounds. The changes observed in DOR, TFV, and estradiol Tmax were also not found to be statistically significant. One subject was not included due to having total testosterone levels below the lower limit of quantification of the assay.

TABLE 2. Pharmacokinetic parameters of doravirine following concomitant estradiol and spironolactone administration.
Parameter Doravirine
GMR (90% CI) Log-paired difference (90% CI) Paired t-Test p-value
AUC0-inf, h·ng/mL 0.97 (0.76, 1.24) −0.03 (−0.25, 0.18) 0.759
AUC0-last, h·ng/mL 0.98 (0.76, 1.25) −0.02 (−0.23, 0.18) 0.820
CL/F, mL/h 1.03 (0.81, 1.32) 0.03 (−0.18, 0.25) 0.759
C24, ng/mL 0.97 (0.76, 1.24) −0.03 (−0.24, 0.18) 0.774
Cmax, ng/mL 0.93 (0.66, 1.32) −0.07 (−0.36, 0.22) 0.650
T1/2, h 0.97 (0.81, 1.16) NA
Tmax, ha 6.00 (1, 6)b 6.00 (0.5, 12)c 0.789
  • Abbreviations: AUC0-inf, area under the concentration-time curve extrapolated to infinity; AUC0-last, from zero to last measured concentration; C24, concentration at 24 h; CI, confidence interval; CL/F, apparent oral clearance; Cmax, maximum plasma concentration; GMR, geometric mean ratio; NA, not applicable; T1/2, apparent half-life; Tmax, time to Cmax.
  • a Presented as median (range) computed with Wilcoxon signed rank test.
  • b Data represents co-administration of DOR/3TC/TDF, estradiol, and spironolactone.
  • c Data represents administration of DOR/3TC/TDF alone.
TABLE 3. Pharmacokinetic parameters of tenofovir following concomitant estradiol and spironolactone administration.
Parameter Tenofovir
GMR (90% CI) Log-paired difference (90% CI) Paired t-Test p-value
AUC0-inf, h·ng/mL 1.17 (0.91, 1.50) 0.15 (−0.06, 0.37) 0.208
AUC0-last, h·ng/mL 1.19 (0.91, 1.57) 0.18 (−0.06, 0.42) 0.199
CL/F, mL/h 0.86 (0.67, 1.10) −0.15 (−0.37, 0.06) 0.208
C24, ng/mL 1.14 (0.93, 1.40) 0.13 (−0.05, 0.31) 0.200
Cmax, ng/mL 1.38 (0.73, 2.60) 0.32 (−0.22, 0.86) 0.287
T1/2, h 0.90 (0.77, 1.04) NA
Tmax, ha 2.00 (1, 6)b 4.00 (1, 6)c 0.414
  • Abbreviations: AUC0-inf, area under the concentration-time curve extrapolated to infinity; AUC0-last, from zero to last measured concentration; C24, concentration at 24 h; CI, confidence interval; CL/F, apparent oral clearance; Cmax, maximum plasma concentration; GMR, geometric mean ratio; NA, not applicable; T1/2, apparent half-life; Tmax, time to Cmax.
  • a Presented as median (range) computed with Wilcoxon signed rank test.
  • b Data represents co-administration of DOR/3TC/TDF, estradiol, and spironolactone.
  • c Data represents administration of DOR/3TC/TDF alone.
TABLE 4. Pharmacokinetic parameters of estradiol following concomitant administration of a single dose of doravirine, lamivudine, and tenofovir disoproxil fumarate.
Parameter Estradiol
GMR (90% CI) Log-paired difference (90% CI) Paired t-Test p-value
AUC0-inf, h·pg/mL 0.97 (0.70, 1.35) −0.03 (−0.32, 0.26) 0.830
AUC0-last, h·pg/mL 1.10 (0.92, 1.32) 0.10 (−0.07, 0.26) 0.297
C24, pg/mL 1.13 (0.99, 1.29) 0.12 (0.00, 0.25) 0.097
Cmax, pg/mL 1.12 (0.92, 1.36) 0.12 (−0.08, 0.31) 0.281
T1/2, h 0.78 (0.53, 1.13) NA
Tmax, ha 4.00 (0.5, 12)b 4.00 (1, 12)c 1.000
  • Abbreviations: AUC0-inf, area under the concentration-time curve extrapolated to infinity; AUC0-last, from zero to last measured concentration; C24, concentration at 24 h; CI, confidence interval; Cmax, maximum plasma concentration; GMR, geometric mean ratio; NA, not applicable; T1/2, apparent half-life; Tmax, time to Cmax.
  • a Presented as median (range) computed with Wilcoxon signed rank test.
  • b Data represents co-administration of DOR/3TC/TDF, estradiol, and spironolactone.
  • c Data represents administration of estradiol (with spironolactone) alone.
Details are in the caption following the image
Total testosterone concentration-time profiles with and without co-administration of doravirine, lamivudine, and tenofovir disoproxil fumarate. Data plotted represents arithmetic mean (±SD).

All eight subjects reported at least one AE during the study (Table 5). All AEs were mild in intensity and resolved by the end of the study without additional action. None of the AEs were related to the study interventions given in this study. Three of eight subjects (37.5%) had AEs related to the discontinuation or washout of prior estradiol and spironolactone (Table S1).

TABLE 5. Summary of adverse events reported in the safety population.
System organ class All Subjects (n = 8) Sequence E (n = 4) Sequence F (n = 4)
Subjects with any adverse event 8 (100.0) 4 (100.0) 4 (100.0)
Subjects with any adverse event related to study medication 0 (0.0) 0 (0.0) 0 (0.0)
Gastrointestinal disorders 2 (25.0) 2 (50.0) 0 (0.0)
General disorders and administration site conditions 4 (50.0) 1 (25.0) 3 (75.0)
Infections and infestations 2 (25.0) 2 (50.0) 0 (0.0)
Investigations 1 (12.5) 1 (25.0) 0 (0.0)
Musculoskeletal and connective tissue disorders 1 (12.5) 0 (0.0) 1 (25.0)
Nervous system disorders 4 (50.0) 3 (75.0) 1 (25.0)
Psychiatric disorders 3 (37.5) 2 (50.0) 1 (25.0)
Skin and subcutaneous tissue disorders 3 (37.5) 2 (50.0) 1 (25.0)
  • Note: Data are presented as n (%). Multiple occurrences of an adverse event in an individual subject are counted only once.

DISCUSSION

Co-administration of estradiol and spironolactone slightly decreased DOR exposure (i.e., AUC0-last, C24, and Cmax), but moderately increased TFV exposure. Furthermore, co-administration of DOR/3TC/TDF resulted in moderately increased estradiol exposure. Moreover, the changes seen in total testosterone following co-administration were not expected to be clinically significant. Overall, these data suggest no clinically significant changes in the PKs of DOR, TFV, and estradiol as well as total testosterone with co-administration. All reported AEs were mild and not related to the study interventions administered in this study despite three participants reporting AEs associated with discontinuing prior hormone therapy. This was anticipated given the study design as participants were asked to temporarily discontinue FHT when indicated.

The TGW are at a disproportionately high risk for acquiring HIV-1 infection compared to the general population. One of the commonly voiced concerns among this population is drug interactions with FHT. This is especially concerning for TGW living with HIV, who have an increased risk for cardiovascular diseases and potentially decreased bone mineral density.7, 10, 12-14, 26 DOR has demonstrated a favorable lipid profile compared to an efavirenz-based regimen, which is ideal for this population.15 Additionally, DOR is available either as a standalone drug or a single complete regimen with 3TC and TDF, the latter being advantageous for patient convenience. Being a CYP3A4 substrate, DOR is susceptible to drug–drug interactions compared to 3TC or TDF. Estradiol and spironolactone have been implicated to affect transcription factors that can have downstream effects on CYP450 activity.16-21 Moreover, several ARVs, such as bictegravir, are metabolized through the same metabolic pathway where altered concentrations in ARV and/or FHT can heighten the risk for AEs, impact viral load, and disrupt desired feminization. Prior studies examining the interaction between concomitant ARV and FHT in TGW have primarily involved PrEP regimens.27 Additional studies are available assessing older NNRTIs, protease inhibitors, and oral contraceptives in cisgender women. However, these existing PK data require extrapolation to a population with diverse identities and under-representation in research. Second, the majority of these studies also used ethinyl estradiol, which is not recommended for FHT due to an increased risk of venous thromboembolism and no demonstrated superiority in feminization over other estrogen preparations.8 Ethinyl estradiol also has differences in metabolic pathways compared to 17β-estradiol, which makes extrapolation to TGW inappropriate.28 There is even less conclusive data explaining the potential interaction between ARVs and anti-androgens. Regardless, the data generated in this study align with prior data suggesting no clinically relevant interaction between ARVs and FHT.26-30

Although likely due to variability, observed changes in DOR Cmax, C24, and AUC0-last may be explained by the inductive effects of estradiol and spironolactone on CYP3A. However, the underlying mechanism for TFV and FHT interactions is uncertain, especially with observed increases in TFV AUC0-last, C24, Cmax, decreased clearance, and prolonged Tmax. Prior studies reported modest decreases in plasma TFV exposure and/or rectal tenofovir diphosphate concentrations with concomitant FHT.22, 31-35 Additionally, TFV and 3TC are renally cleared and not metabolized by CYP450. All participants were healthy and had a baseline creatinine clearance of at least 60 mL/min, the latter being a possible reason for conflicting data as creatinine clearance was not assessed throughout the study in addition to the small sample size. It has been hypothesized that altered renal clearance may affect TFV concentrations, but this mechanism has yet to be validated.22 FHT has been speculated to delay gastric emptying time based on hormonal changes during menstrual cycles in cisgender women, which could explain the delayed Tmax seen in this study.30 As TDF is a substrate for P-glycoprotein (P-gp), an increase in TFV exposure was reported following co-administration with an inhibitor, such as cobicistat. Cisgender women have been reported to have reduced P-gp expression than cisgender men, suggesting that FHT reduces P-gp expression, resulting in higher TFV exposures as seen in this study.30 TFV PK has generally been associated with high variability and may be influenced by other covariates such as race and body mass index. Nonetheless, the observed changes in plasma TFV were not expected to be clinically relevant. There are conflicting data on the effects of ARVs on the PKs of FHT. Three studies observed a decrease in estradiol exposures whereas others reported increased estradiol concentrations.22, 31, 36, 37 The mechanism is not well understood, but the differences seen here could be attributed to the estradiol formulations studied, age group, and presence of HIV-1 infection. Estradiol is primarily metabolized by CYP3A4, which is not expected to be affected by concomitant DOR and TDF co-administration.30 Similarly, the observed increases in estradiol levels were not expected to be clinically significant. Although total testosterone was an exploratory end point and evaluated as a measure of FHT, observed concentrations following co-administration of DOR/3TC/TDF were lower compared to FHT alone. These data align with prior studies demonstrating reduced testosterone and spironolactone exposures.36, 38 Spironolactone is extensively metabolized by hepatic esterases, and is not expected to be affected by CYP inducers or inhibitors. On the other hand, testosterone is subject to CYP3A4/5 metabolism which can alter exposures in TGW.30 The lack of meaningful impact seen in plasma estradiol concentrations is anticipated to show minimal changes in total testosterone concentrations. The magnitude of these changes, in addition to safety data, suggest no changes in breakthrough masculinization. As this was an exploratory analysis, no criteria to evaluate the magnitude of observed differences were established a priori. Regardless, total testosterone was utilized as an efficacy and safety measure of FHT. Although not required for participation nor maintained throughout the study, five of eight participants had achieved target total testosterone levels (<50–55 ng/dL) as recommended by University of California San Francisco and the Endocrine Society.7, 8 This is particularly of interest to TGW as some may choose to prioritize feminization over ARV therapy, the former being a common primary motivator for TGW to seek medical care. The totality of these results demonstrated no clinically significant interaction among DOR/3TC/TDF, estradiol, and spironolactone. However, larger studies are warranted to elucidate the interaction between TFV and FHT as well as downstream effects on plasma estradiol and total testosterone.

There were several limitations of this study, including the low number of participants that enrolled and completed the study. Enrollment stopped before reaching the target sample size due to logistical disruptions caused by the COVID-19 pandemic, and two participants either withdrew or were lost to follow-up. The objectives of this trial were primarily PK, so the reduced power may still be acceptable. The 3TC PKs were not evaluated due to budget constraints, but 3TC is not metabolized by CYP450 but rather renally cleared. Based on the results seen in this study as well as prior studies, the effects of FHT on 3TC PKs are not expected to be clinically significant. Similarly, 3TC is not expected to have clinically meaningful effects on estradiol exposures and total testosterone. Although selected doses for estradiol and spironolactone do not necessarily represent typical doses for FHT, the maximum daily dose was used to maximize the probability of observing a drug interaction with DOR/3TC/TDF. Rather than dosing participants to steady-state, single doses of DOR/3TC/TDF were used. The accumulation factor for DOR and TFV once daily were expected to be reasonably small. Thus, single doses were determined to be appropriate to predict steady-state parameters, which are confirmed by the half-lives calculated in this study. Although estradiol clearance was not calculated due to software limitations, the study primarily focused on changes in serum concentration and exposure as a marker of efficacy. Last, nearly all reported GMRs, 90% CIs, and log-transformed PK parameters did not meet strict a priori bioequivalence parameters partly due to small sample size. The no-effect boundaries of 80%–125% are notoriously conservative, which may not be clinically relevant in this study as they are used for generic product evaluation. However, this is mitigated by the locations of the point estimates. Moreover, it is likely that clinically meaningful CIs are closer to wider intervals, such as 67%–150% or 50%–200%. The current study was not powered for bioequivalence, but rather on a paired t-test. However, the data shows that the assumption of the effect size being at least 1.43 for post hoc power calculations was incorrect. As such, the 80% and 125% bounds were plotted in Figures S5–S7 for illustrative purposes only. Regardless, it is evident that no dose modifications are required with the co-administration of DOR/3TC/TDF and feminizing hormones as the bidirectional interaction is not clinically significant.

In conclusion, clinically significant interactions between DOR/3TC/TDF and FHT should not be expected. Concomitant estradiol and spironolactone slightly reduced DOR Cmax, C24, and AUC0-last whereas moderate increases were seen for TFV, which is not expected to impact clinical efficacy. DOR/3TC/TDF is not anticipated to significantly impact plasma estradiol and total testosterone levels. Co-administration of DOR/3TC/TDF, estradiol, and spironolactone appeared to be safe and well-tolerated with no reported breakthrough masculinization.

AUTHOR CONTRIBUTIONS

K.L. wrote the manuscript. E.L. designed the research. K.L., W.K.K., and E.L. performed the research. K.L. and T.Z. analyzed the data. T.Z. contributed new reagents/analytical tools.

ACKNOWLEDGMENTS

The authors thank the volunteers, nurses, and staff of the Clinical Research Unit at Thomas Jefferson University, Douglas F. Stickle, Lauren Furst, Michael S. Willett, and Matthieu Place for their contributions to this research.

    FUNDING INFORMATION

    This work was supported by an Investigator Studies Program grant provided by Merck & Co., Inc. (MISP59198). K.L. and E.L. were supported by the National Institutes of Health institutional training grant (T32GM008562). The funders had no role in study design, data collection and analysis, or manuscript writing.

    CONFLICT OF INTEREST STATEMENT

    During the conduct of the study, E.L. was a T32 fellow at Thomas Jefferson University and subsequently at the NIH. His current employer is Janssen Research & Development, LLC. All other authors declared no competing interests for this work.