Estradiol and Spironolactone Plasma Pharmacokinetics Among Brazilian Transgender Women Using HIV Pre-Exposure Prophylaxis: Analysis of Potential Interactions

This was a drug-drug interaction (DDI) study evaluating FHT and daily oral PrEP with TDF/FTC nested in the trans-specific PrEP demonstration project, PrEParadas, conducted in Rio de Janeiro, Brazil, from August 2017 to January 2020. PrEParadas study procedures and the results of DDI interactions of FHT on TDF/FTC PK are described elsewhere [7, 27]. PrEParadas inclusion criteria were: TGW aged ≥ 18 years, living in Rio de Janeiro or its metropolitan area, HIV negative status at screening and enrollment (baseline visit), and reporting engagement in at least one of the following: condomless anal sex in the last 6 months, sexually transmitted infection diagnosis in the last 12 months, transactional sex in the last 6 months, current sexual partner living with HIV regardless of HIV viral load. Participants who had history of pathological bone fracture, creatinine clearance (CL_CR) < 60 mL/min (estimated using the Cockcroft–Gault equation, using assigned sex at birth) [29], use of any medication known to interact with at least one of the study drugs, and any previous transfeminine bottom surgery (orchiectomy and/or vaginoplasty) were not enrolled.

The Evandro Chagas National Institute of Infectious Diseases-Fiocruz Institutional Review Board approved the study. PrEParadas study is registered with clinicaltrials.gov (NCT03220152). All participants signed informed consent forms before any study procedure. Participants included in the DDI study were off FHT for at least 15 (oral regimens) or 45 (injectable regimens) days before screening (Fig. 1). The standardized study FHT (estradiol valerate 2–6 mg plus spironolactone 100–200 mg) was initiated at the screening visit. Throughout the 12 weeks of follow-up (enrollment and Weeks 4, 7, and 9), the participants’ estradiol trough plasma concentrations (C_trough) were available to the study endocrinologist, who could adjust FHT dosage based on clinical evaluation and participant’s self-satisfaction, as recommended by available guidelines of transgender health care [30, 31]. The main criteria for estradiol dose adjustment were participant’s goals and self-satisfaction. Physiological female levels (100–200 pg/mL) served as a safety parameter for estradiol levels [10] to avoid levels above 200 pg/mL. Fifteen days after the FHT initiation, participants had the first intensive PK (PK1, only FHT) evaluation to assess the FHT PK and then initiated PrEP. A second intensive PK evaluation (PK2, FHT plus PrEP) was performed at the Week 12 visit to assess possible DDI of PrEP drugs on FHT PK.

We evaluated participants’ age (years), CL_CR (mL/min), and race (Black, Pardo, White and other) at baseline; weight (kg), body mass index (BMI; kg/m²), alanine aminotransferase (ALT; U/L), and aspartate aminotransferase (AST; U/L) at baseline and PK2. Estradiol levels were measured at screening, enrollment (PK1), follow-up visits (Weeks 4, 7, and 9), and Week 12 (PK2) at pre-dose sampling (C_trough).

Blood samples were collected prior and after (0, 0.5, 1, 2, 4, 6, 8, and 24 hours) the directly observed dosing administration in fasted state for both PK1 and PK2 evaluations. Thirty minutes after the drug intake, we offered a standard breakfast. One week before PK1 and PK2, all participants received reminders to adhere to FHT and PrEP. Adherence to FHT was evaluated by self-report at each PK visit. Pharmacokinetic visits were rescheduled if the participant reported missing any dose 7 days before each visit. Pre-exposure prophylaxis adherence was also evaluated by dried-blood spots (DBS) levels of tenofovir-diphosphate (tenofovir-DP) and emtricitabine-triphosphate (emtricitabine-TP) at Week 12; adherence was stratified into: low (less than 350 fmol per punch, suggestive of < 2 doses of PrEP per week), medium (350–699 fmol per punch, suggestive of two to three doses of PrEP per week), and high (700 fmol per punch or greater, suggestive of 4+ doses of PrEP per week) [32].

Blood samples were immediately centrifuged after collection, and plasma was stored at − 80 °C in cryotubes. Estradiol and metabolites (estrone and estrone sulfate), spironolactone and one metabolite (canrenone) were determined in plasma samples by validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods at the Fiocruz Pharmacokinetics Service (Rio de Janeiro, Brazil). Estradiol/estrone and chlorthalidone, as internal standard (IS), were extracted from plasma samples with methyl tert-butyl ether. After evaporation to dryness, the residue was reconstituted in an acetonitrile:water solution (70:30, v/v). We used a C18 column and water:acetonitrile (68:32, v/v) as mobile phase. The transitions of m/z 271.11 → 145.0 and m/z 269.114 → 145.005 were monitored for estradiol and estrone, respectively. The m/z 336.87 → 190.16 transition was monitored for the IS. Estradiol and estrone concentrations were linear in the range of 25–500 pg/mL and 25–1000 pg/mL, respectively. The inaccuracy and imprecision were lower than 15%. Both lower limit of quantification (LLQ) imprecision and inaccuracy were below 20%.

Estrone sulfate was determined after a simple acetonitrile protein precipitation, with chlorthalidone as IS, C18 column and water:acetonitrile (65:35, v/v) as mobile phase. The transitions of m/z 349.037 → 269.200 and m/z 336.936 → 189.932 were monitored for the quantification of estrone sulfate and for the IS, respectively. Estrone sulfate concentrations were linear in the range of 0.1–50 ng/mL. The inaccuracy was lower than 6.5%. Both LLQ imprecision and inaccuracy were below 17%. Spironolactone, canrenone, and diazepam (IS) were extracted from plasma samples with methyl tert-butyl ether and the supernatant was evaporated to dryness. The residue was reconstituted in a 65:35 (v/v) solution of methanol:formic acid (0.1% in water). We used a C8 column and ultrapure water with formic acid 0.1%:methanol (35:65, v/v) as mobile phase. The transitions of m/z 341.169 → 107.169 and m/z 341.152 → 107.041 were monitored for the quantification of spironolactone and canrenone, while the transition of m/z 285.320 → 193.087 was monitored for the IS. Spironolactone and canrenone concentrations were linear in the range of 1–200 ng/mL and 1–250 ng/mL, respectively. The inaccuracy and imprecision were lower than 10%. Both LLQ imprecision and inaccuracy were below 12%. All methods presented accuracy and precision according to the established in the Brazilian Health Regulatory Agency (ANVISA) guideline [33].

We used LC-MS/MS for tenofovir-DP and emtricitabine-TP quantification of DBS samples, as previously described [32, 34].

Non-compartmental PK parameters, i.e., AUC_τ, maximum concentration at steady-state (C_max,ss), t_max, minimum concentration (C_min), apparent total body clearance (CL/F), apparent volume of distribution at steady-state (V_ss/F), and t_½ were estimated for estradiol and spironolactone. Estrone, estrone sulfate, and canrenone had AUC_τ, C_max,ss, t_max, and C_min estimated (Monolix Software^® Suite 2021R2, Lixoft^®, Antony, France). We excluded participants: (1) with low PrEP adherence (tenofovir-DP suggestive of < 2 doses/week or undetectable levels of emtricitabine-TP), (2) who did not attend successive study visits, (3) who had taken medication prohibited by the study protocol (i.e., medications that could interact with PrEP or FHT), (4) who took their FHT pills prior the pre-dose PK sampling, and (5) who had blood collection difficulties. In the descriptive analyses, we used medians and interquartile range (IQR) and absolute and relative frequencies, respectively, for continuous numerical variables and for nominal variables. Non-compartmental PK parameters were summarized as geometric means and compared between PK1 and PK2 as geometric mean ratios (GMRs, 90% CI) using a paired t-test after log transformation. Our sample size provided 80% power to detect a difference of at least 23 and 23.5% on estradiol and spironolactone plasma geometric mean AUC_τ, respectively, at a significance level of 0.05. Since our participants were using different doses of FHT, we presented AUC_τ and C_max,ss of spironolactone, estradiol, and their metabolites normalized to a dose of 100 mg of spironolactone or 2 mg of estradiol valerate (equivalent to 1.53 mg of estradiol when accounting for the molecular weights), respectively. This means we used the ratio of the individual parameter (AUC_τ and C_max,ss) by the multiple of the lower dose (2 mg of estradiol valerate or 100 mg of spironolactone) according to the following equations: where: AUC_τ and C_max,ss: individual parameter normalized, AUC_{τ_obs} and C_{max,ss_obs}: individual parameter calculated from individual plasma concentration-time curve, E2V dose: estradiol valerate dose administered, SPR dose: spironolactone dose administered.

A linear regression model with random effect for individuals was used to evaluate the association between estradiol C_trough and time (study week) of PrEP use (enrollment [PK1], weeks 4, 7, 9, and 12 [PK2]). We considered p < 0.05 as statistically significant. All statistical analyses were performed in R v.4.0.5 software, utilizing the ‘nlme’ library to develop statistical models.

From August 2017 to January 2020, 33 participants were enrolled and underwent PK1 evaluations (Fig. 2). Of these, 27 underwent PK2 evaluations, with 8 participants excluded afterwards due to use of prohibited medication (n = 1), FHT intake before directly observed dosing (n = 2), and low adherence (self-reported or based on tenofovir-DP DBS levels) (n = 5). As such, PK2 assessment analysis included 19 participants and a total of 304 observations.

All participants presented high PrEP adherence. At baseline, median age was 26 years (23–27.5), and BMI was 22.4 kg/m² (20.2–27.4) (Table 1). Estradiol valerate doses ranged from 2 to 4 mg at PK1 and from 2 to 6 mg at PK2. Most participants were on daily estradiol valerate 2 mg throughout the 12 weeks of follow-up: 16/19 (84%) participants at PK1 and 10/19 (53%) at PK2 (Table 2). Spironolactone doses ranged from 100 to 200 mg during the study; 18/19 (95%) and 10/19 (53%) participants were taking 100 mg at PK1 and PK2, respectively.

Concentration-time plasma profiles of estradiol, estrone and estrone sulfate are presented in Fig. 3 and Supplementary Figure 1 (one participant on 6 mg of estradiol valerate at PK2). No differences were observed between estradiol PK parameters at PK1 (only FHT) and PK2 (FHT plus PrEP) evaluations (Table 3).

Considering estradiol metabolites, estrone AUC_τ and C_max,ss were 20% and 16% lower at PK2 (FHT plus PrEP) than PK1 (only FHT) (GMR [90%CI] AUC_τ: 0.80 [0.71–0.91]; C_max,ss: 0.84 [0.73–0.95]). No other differences on estradiol metabolites PK parameters were detected. Furthermore, there were no differences in estradiol C_trough values after FHT initiation (p = 0.54) throughout the 12 weeks of follow-up (Suppl. Fig. 2).

Spironolactone and canrenone plasma concentration versus time at PK1 and PK2 are presented in Fig. 4. Spironolactone AUC_τ was 24% lower at PK2 (FHT plus PrEP) than PK1 (only FHT) (GMR [90% CI]: 0.76 [0.65–0.89]) while CL/F was 32% higher (1.32 [1.13–1.54]) (Table 4). Similarly, canrenone AUC_τ and C_max,ss were lower at PK2 (0.85 [0.78–0.94] and 0.82 [0.74–0.91], respectively).