Pharmacokinetics, Safety, and Tolerability of Once-Daily Darunavir with Cobicistat and Weekly Isoniazid/Rifapentine

Abstract

Background:

Once-weekly isoniazid with rifapentine (HP) for three months is a recommended treatment for latent tuberculosis infection (LTBI) in persons with HIV. HP reduces exposures of certain antiretroviral medications, resulting in limited options for the concomitant use of these therapies. Here we examined the pharmacokinetics (PK), safety, and tolerability of darunavir/cobicistat with HP.

Methods:

This was an open-label, fixed sequence, two-period crossover study in persons without HIV. Participants received darunavir 800 mg/cobicistat 150 mg once-daily alone for four days, then continued darunavir/cobicistat once-daily for Days 5-19 with HP coadministration on Days 5, 12, and 19. Intensive PK assessments were performed on Days 4, 14, and 19. PK parameters were determined using noncompartmental methods. Geometric mean ratios (GMR) with 90% confidence intervals (CIs) were calculated and compared between phases using mixed effects models.

Results:

Thirteen participants were enrolled. Two withdrew after Day 4 and one withdrew after Day 14. Of the three withdrawals, two were attributed to drug-related adverse events. Darunavir area under the concentration-time curve (AUC), maximum concentrations (C_max), and concentrations at 24 hours post-dose (C_24h) were reduced by 71%, 41%, and 96% ~48-72 hours after HP administration (Day 14), respectively, and 46%, 17%, and 89% with simultaneous HP administration (Day 19), respectively. On Day 14, 45% of the pre-dose and 73% of C_24h concentrations were below the darunavir EC₅₀ (0.055 ug/mL).

Conclusions:

Darunavir exposures were significantly decreased with HP coadministration. Temporal relationships between HP coadministration and the extent of induction or mixed inhibition/induction of darunavir metabolism were apparent. Coadministration of darunavir/cobicistat with 3HP should be avoided.

INTRODUCTION

Tuberculosis is a leading cause of death among persons with HIV (PWH). Diagnosis and treatment of latent tuberculosis infection (LTBI) and initiation of antiretroviral therapy (ART) are critical to preventing progression to active tuberculosis. Historically, daily isoniazid preventive therapy (IPT) for 6 to 9 months has been a cornerstone of LTBI treatment in PWH,^1,2 but adherence and completion rates have been suboptimal given the long treatment duration and tolerability issues and may have compromised the public health impact of IPT.^3–5

More recently, multiple shortened LTBI treatment regimens consisting of rifamycins with or without isoniazid have shown higher rates of treatment completion and improved tolerability in comparison to IPT.^3–6 Once-weekly isoniazid with rifapentine for 3 months (3HP) is one such recommended treatment in persons with HIV (PWH).^1,2 3HP is associated with higher completion rates in comparison to IPT for 6 to 9 months, and lesser toxicity. However, drug-drug interactions between 3HP and antiretroviral medications are of concern. Rifapentine is a potent inducer of various drug-metabolizing enzymes and transporters, notably cytochrome P450 3A4 (CYP3A4) and P-glycoprotein.^7,8 This may result in subtherapeutic exposures of antiretroviral medications that are substrates for these enzymes and transporters, which would confer a higher risk of virologic failure and antiretroviral resistance in PWH.

Available data supports the use of dolutegravir, efavirenz, and raltegravir with 3HP,² but there are currently no data with boosted protease inhibitors. To better inform what antiretroviral options may be considered in PWH and LTBI, prospective drug-drug interaction studies with more contemporary HIV treatments are needed. Darunavir boosted with cobicistat is a component of treatment options that may be considered in certain clinical scenarios in PWH within the United States due to its high barrier to resistance.¹ Darunavir is a substrate for CYP3A4 and cobicistat is a substrate for CYP3A4 and CYP2D6,⁹ and thus rifapentine will likely decrease exposure of darunavir and cobicistat. However, isoniazid can also inhibit CYP450 enzymes,^10–13 which may counteract some of the inductive effects by rifapentine, and thus HP may have variable effects on the pharmacokinetics of darunavir/cobicistat. Here we examined the pharmacokinetics, safety and tolerability of darunavir with cobicistat when co-administered with HP in healthy adult volunteers.

METHODS

Study Population

Persons without HIV between the age of 18 to 65 years, weight between 45 to 120 kg, and BMI 18.0 to 30.0 kg/m² were eligible to participate. Participants were required to be healthy based on their physical exam, medical history, vital signs, and clinical laboratory screening tests; have no evidence of HIV, active or latent TB, or active hepatitis A, B, or C; and abstain from alcohol use during the study. Female participants of childbearing potential were required to have a negative serum or urine pregnancy test at baseline and throughout the study, and be willing to use non-hormonal contraceptive methods. Key exclusion criteria included known hypersensitivity reactions to study medications, related rifamycin analogues, or sulfonamides (type 1); current or previous use of investigational drugs within 30 days of entry; prescription, over-the-counter, herbal, or holistic medications within five half-lives of receiving study medication, except for short courses of therapy reviewed on a case-by-case basis for drug-drug interactions; and any other health conditions that may interfere with interpretation of study results or cause harm to study participants.

Study Design

This was the second arm of an open-label, fixed sequence, two-period crossover study in persons without HIV (ClinicalTrials.gov NCT02771249). The first arm evaluating the combination of dolutegravir with HP was terminated early due to severe toxicities of flu-like symptoms and increase in serum transaminases in healthy volunteers.¹⁴ This study was approved by the National Institute of Allergy and Infectious Diseases Institutional Review Board and was overseen by an independent safety monitoring committee. All study procedures were conducted in accordance with the ethical standards of the Declaration of Helsinki, as revised in 2000, and all participants provided written informed consent.

Participants were screened within 89 days of study start. Baseline safety laboratory assessments were collected within one week of initiating study drug (between Day -6 to Day 0). Participants received darunavir 800 mg/cobicistat 150 mg (Prezcobix^®, Janssen, Inc.) orally once-daily alone for four days, then continued darunavir/cobicistat once-daily for an additional 14 days in combination with weekly isoniazid (15 mg/kg per dose, maximum dose of 900 mg), rifapentine (900 mg dose if weight ≥50 kg), and pyridoxine 50 mg, all given orally on Days 5, 12, and 19. Blood samples were drawn from participants at time 0 (pre-dose), 1, 2, 3, 4, 5, 6, 8, 10, and 24 hours post-dose on days 4 (darunavir/cobicistat alone), 14 (~48-72 hours after the second isoniazid-rifapentine dose), and 19 (simultaneously with the third isoniazid-rifapentine dose). On these intensive PK days, the time (0) pre-dose sample was drawn after a fasting period of 8 hours, after which participants received a standardized 500 calorie moderate fat-containing breakfast and were administered study medication with 240 mL of water. Blood samples were centrifuged within 1 hour of collection at 3200 rpm for 10 minutes at 4°C, and plasma was then isolated and stored at −80°C until sample analysis.

Safety Assessments

Participant safety and tolerability was monitored throughout the study period which included physical exam, self-reported symptoms, vital signs, and clinical safety laboratory assessments. Safety assessments were performed at all study visits and graded according to the Division of AIDS Adverse Events Grading table (November 2014, v2.0),¹⁵ except for total bilirubin which was graded according to the Toxicity Grading Scale for Healthy Adult and Adolescent Volunteers Enrolled in Preventive Vaccine Trials¹⁶ due to transient, asymptomatic increases in this measure with the use of rifamycins.

Analytical Methods

Darunavir concentrations were analyzed using validated LC-MS-MS methods at the Infectious Disease Pharmacokinetics Laboratory at the University of Florida. Briefly, plasma samples were analyzed using a liquid chromatography-tandem mass spectrometry assay, performed on a Thermo Scientific TSQ Vantage instrument. The linear calibration curve range was from 0.05 to 10.00 mg/liter. The ranges for the coefficient of variation of validation quality control sample precisions were 4.6-6.9%. The ranges for the accuracy of validation quality control samples were 98-110%.

Sample Size Considerations, Pharmacokinetic & Statistical Analyses

Sample size calculations were performed defining a 25% or greater decrease in AUC_0-24h as clinically significant with an intrasubject variability of 30% for darunavir/cobicistat.^17–20 Using a two-sided 90% confidence interval, a sample size of 15 provided at least 80% power to detect this change. Post hoc sample size calculations demonstrated 10 participants would yield at least 80% power to detect a 30% or greater decrease.

All participants who completed at least one intensive pharmacokinetic visit were included in the PK analysis. PK parameters were determined using noncompartmental methods (Phoenix WinNonlin^© v8.3, Princeton, NJ). Pre-dose (time 0) concentrations (C_0h), maximum plasma concentration (C_max), and time to reach C_max (T_max) were obtained directly by visual inspection of darunavir plasma concentration versus time profiles. Concentrations at 24 hours post-dose (C_24h) were based on observed data for samples collected exactly at 24 hours post-dose or predicted concentrations at 24 hours based on partial area under the concentration-time curve (AUC) estimates. The apparent elimination rate constant (λ_Z) was determined through calculation of the absolute value of the slope of the log–linear regression of at least three declining points in the terminal phase. The elimination half–life (t½) was calculated as 0.693/λ_Z. The steady-state area under the concentration vs. time curve from time zero through 24 hours post-dose (AUC_0–24h) was calculated using linear up-log down trapezoidal rule. Apparent oral clearance at steady-state (CL_ss/F) was calculated as 800 mg / AUC_0-24h, and apparent volume of distribution (V_ss/F) was calculated as CL/F / λ_Z.

Pharmacokinetic data were log-transformed prior to analyses. Least square (LS) means of the log-transformed values were calculated and compared between different phases using mixed effects models with participants as random effects and study day as a fixed effect (Phoenix WinNonlin© v8.3, Princeton, NJ). Comparisons of C_0h and C_24h against the half maximal effective concentration (EC₅₀) for darunavir (0.055 ug/mL) and changes in inhibitory quotients (calculated as the ratio of the observed concentration to the EC₅₀) were also made.²¹ Log-transformed LS mean differences between phases and 90% confidence intervals (CIs) were then back-transformed to generate geometric mean ratios (GMR) with 90% CIs on the original scale. P-values were calculated based on the least square mean difference between study days. P<0.05 was considered statistically significant.

RESULTS

Study Population

A total of 13 participants were enrolled (10 males, 3 females; 8 white, 3 black, 1 multiple race, 1 unknown; 2 Hispanic/Latino) between August 2017 and July 2019. The median (range) age and weight were 25 (21-45) years and 70.7 (62-85.1) kg, respectively. Two participants withdrew from the study after the Day 4 intensive PK assessment after only receiving darunavir/cobicistat. One experienced heparin lock flush intolerance and opted to withdraw from the study; and the other was removed due to elevated C-reactive protein, total and direct bilirubin which were attributed to a viral illness but deemed possibly related to darunavir/cobicistat. A third participant was removed from study following the Day 14 PK assessment due to a low white blood cell count after receiving darunavir/cobicistat and two weekly HP with pyridoxine doses. This AE was deemed possibly related darunavir/cobicistat, isoniazid, or rifapentine, and later resolved after discontinuing study drugs.

Pharmacokinetic Results

Concentration-time profiles for darunavir showed marked reductions in drug exposures after initiating HP (Figure 1). Relative to darunavir/cobicistat alone, darunavir AUC_0-24h and C_max were decreased by 71% (90% CI -65%, -76%; p<0.0001) and 41% (90% CI -31%, -50%; p<0.0001) on Day 14, respectively, and 46% (90% CI -22%, -47%; p=0.0007) and 17% (90% CI -2%, -30%) on Day 19, respectively (Table 1). More marked decreases were apparent on Day 14 (~48-72 hours after RPT administration) in comparison to Day 19 (simultaneous administration). Apparent oral clearance at steady-state (CL/F) of darunavir was increased by 244% (90% CI 187%, 312%; p<0.0001) and 55% (90% CI 29%, 88%; p=0.0007) on Days 14 and 19, respectively, in comparison to Day 4. The net effect on the apparent volume of distribution at steady-state (V_SS/F) differed depending on the timing of HP administration, with a 54% increase (90% CI 23%, 93%; p=0.003) on Day 14 and 47% decrease (90% CI -33%, -58%) on Day 19.

Figure 1. Median darunavir concentrations by nominal time and study day.

Concentration time profiles presented on linear scale and log scale (insert).

Table 1.

PK Parameters for Darunavir Boosted with Cobicistat (DRV/c) Alone and in Combination with Weekly Isoniazid-Rifapentine (HP)

PK Parameter	DRV/c alone (Day 4) (n=13)	DRV/c + HP (Day 14) (n=11)	DRV/c + HP (Day 19) (n=10)	Day 14 vs. Day 4	Day 19 vs. Day 4	Day 14 vs. Day 19
AUC0-24h (hr*ug/mL)	82.6 (59.9%)	23.1 (56.8%)	49.4 (37.6%)	0.29 (0.24, 0.35) p<0.0001	0.64 (0.53, 0.78) p=0.0007	0.45 (0.37, 0.55) p<0.0001
Cmax (ug/mL)	7.86 (42.2%)	4.53 (45.7%)	6.22 (39.3%)	0.59 (0.50, 0.69) p<0.0001	0.83 (0.70, 0.98) p=0.06	0.71 (0.60, 0.84) p=0.002
Tmax (hr)	4.0 (2.7-4.0)	3.0 (2.0-4.0)	3.4 (2.0-4.3)	−0.33 (−1.13 to 1.0) p=0.79	−0.13 (−0.83 to 0.82) n/a	0.0 (−0.75, 0.0) n/a
C0h (ug/mL)	1.75 (112%)	0.08 (82.0%)	0.33 (90.1%)	0.04 (0.03, 0.07) p<0.0001	0.19 (0.11, 0.34) p=0.0001	0.22 (0.13, 0.40) p=0.0005
C24h (ug/mL)	1.24 (122%)	0.04 (122.0%)	0.12 (66.2%)	0.04 (0.02, 0.06) p<0.0001	0.11 (0.07, 0.18) p<0.0001	0.33 (0.19, 0.54) p=0.002
t1/2 (hr)	10.8 (85.0%)	4.44 (31.2%)	3.37 (11.3%)	0.45 (0.37, 0.56) p<0.0001	0.35 (0.28, 0.43) p<0.0001	1.31 (1.05, 1.64) p=0.052
CLss/F (L/hr)	9.68 (59.9%)	34.6 (56.8%)	16.2 (37.6%)	3.44 (2.87, 4.12) p<0.0001	1.55 (1.29, 1.88) p=0.0007	2.21 (1.83, 2.68) p<0.0001
V/F (L)	151.3 (53.1%)	221.8 (78.9%)	78.8 (34.8%)	1.54 (1.23, 1.93) p=0.003	0.53 (0.42, 0.67) p=0.0001	2.89 (2.29, 3.65) p<0.0001

Key: AUC_0-24h = area under the concentration time curve from time 0 through 24 hours post-dose; C_max = maximum concentration; C_0h = time 0 (pre-dose) concentration; C_24h = concentration at 24 hours post-dose; t_1/2 = half-life; CL_ss/F = apparent oral clearance at steady-state; V/F = apparent volume of distribution.

Data for each phase presented as geometric mean (geometric CV%) except for T_max (reported as median [IQR]); phase comparisons presented as GMR (90% CI) except for Tmax (reported as median (IQR) difference). P-values were calculated based on the least square mean difference between phases.

C_0h and C_24h were significantly reduced with concomitant HP administration, by 96% at both time points on Day 14 (C_0h 90% CI -93%, -97%; p<0.0001; C_24h 90% CI -94%, -98%; p<0.0001), and 81% (C_0h 90% CI -66%, -89%; p=0.0001) and 89% (90% CI -82%, -93%; p<0.0001) on Day 19. Several C_0h (5/11 [45.4%]) and C_24h (8/11 [72.7%]) values were below the protein-adjusted darunavir EC₅₀ for wild-type virus (0.055 ug/mL) on Day 14 (Figure 2). Geometric mean IQs had declined from 31.8 and 22.5 on Day 4 to 1.4 and 0.8 at these same time points on Day 14, respectively. By Day 19, these effects had attenuated as the geometric mean pre-dose concentration had increased and all were above the EC₅₀. Following simultaneous administration of HP on Day 19, the geometric mean C_24h decreased again with one concentration falling below the darunavir EC₅₀ and two more falling just above this threshold (0.068 and 0.069 ug/mL). The corresponding geometric IQs on Day 19 at the pre-dose and 24-hour post-dose time points were 6.0 and 2.2, respectively.

Figure 2. Darunavir concentrations at time 0 (pre-dose) and 24 hours post-dose by study day.

Data presented as geometric mean (geometric standard deviation). In comparison to Day 4, C_0h and C_24h were decreased by 96% (90% CI -93%, -97%) and 96% (90% CI -94%, -98%) on Day 14, respectively; and C_0h and C_24h were decreased by 81% (90% CI -66%, -89%) and 89% (90% CI -93%, -82%) on Day 19, respectively. In comparison to Day 19, C_0h and C_24h were 78% lower (90% CI -60%, -87%) and 67% lower (90% CI -46%, -81%) on Day 14, respectively. P-values were calculated based on the least square mean difference between phases.

Safety & Tolerability Results

A total of 69 adverse events occurred among 13 participants, of which 51 events among 11 participants were deemed possibly or probably related to study medication(s). Nearly all related adverse events were mild or moderate in severity, and the most common involved the central nervous system (headache, dizziness, fatigue; n=12 events) or gastrointestinal (nausea, diarrhea; n=8 events) systems. One grade 3 asymptomatic direct bilirubin increase (0.4; reference range, 0-0.3 mg/dL) occurred ~24 hours after the second HP dose and was deemed possibly related to darunavir/cobicistat and isoniazid and probably related to rifapentine. This increase was transient and resolved by the next day, the participant was asymptomatic, and AST and ALT were normal throughout this period.

DISCUSSION

In this healthy volunteer PK study, darunavir exposures were significantly decreased following the initiation of HP. Several trough concentrations fell below the darunavir EC₅₀, suggesting that there may be a higher risk of therapeutic failure in persons with HIV receiving darunavir/cobicistat with 3HP, particularly in the clinical context of prolonged subtherapeutic darunavir exposures over the full 3HP treatment period. Temporal relationships between HP coadministration and the extent of induction or mixed inhibition/induction of darunavir metabolism were apparent. Coadministration of once daily darunavir/cobicistat with 3HP should be avoided.

Rifapentine is a potent inducer of CYP450, UGT, and efflux transporters.^22,23 Though reduced darunavir exposures were expected in this study, the extent of reduction and whether simultaneous administration with isoniazid would counteract induction by rifapentine were unclear. Several C_0h and C_24h concentrations fell below the EC₅₀ of darunavir on Day 14.²¹ Though these effects were attenuated by Day 19, darunavir exposure-response relationships have demonstrated an increased risk of virologic failure as IQs decrease, and the IQs on Day 19 were still markedly lower in comparison to Day 4.²⁴ Temporal effects of HP on darunavir concentrations were apparent given the higher relative exposures of darunavir with simultaneous administration on day 19 in comparison to ~48-72 hours after 3HP on day 14. This is consistent with the pregnane X receptor (PXR) turnover of approximately 1-3 days^25,26 and other drug-drug interaction studies between weekly rifapentine with isoniazid and dolutegravir,^14,27 doravirine,²⁸ efavirenz,²⁹ raltegravir,³⁰ and tenofovir alafenamide (TAF).³¹ The differential effects on CL_ss/F and V/F in this and prior studies are likely due to a combination of enzyme and/or transporter inhibition by isoniazid with residual induction by rifapentine when co-administered simultaneously,^12,32 followed by induction by rifapentine as inhibitory effects diminish and PXR is activated. The relative effects on CL_ss or V relative to the effects on bioavailability cannot be clearly differentiated from these data. Although the PK data herein cannot be extrapolated to differing rifapentine-containing LTBI regimens (i.e., one month of daily isoniazid and rifapentine (1HP)), coadministration of once daily darunavir/cobicistat with 1HP should also be avoided in the absence of supporting data as daily rifapentine would likely result in more pronounced decreases in darunavir exposure.

It is unknown if the magnitude of decrease in darunavir exposure with HP seen in this study will apply to darunavir boosted with ritonavir (as opposed to cobicistat) or to other boosted protease inhibitors. Additionally, this study did not evaluate whether the decreased exposure could be overcome with twice daily dosing, though the programmatic feasibility of this approach on a global level may be limited. Standard doses of boosted protease inhibitors are contraindicated with rifampin due to markedly reduced drug exposures.¹ Double dosing of protease inhibitors and pharmacoenhancers has been evaluated in more recent studies with mixed results on PK and safety depending on the specific combinations assessed.^33,34 Boosted proteases inhibitors can be used with rifabutin, another rifamycin,^17,35–38 but are implicated in two-way interactions with rifabutin owing to its CYP3A4 metabolism, which has caused tolerability issues in some studies.³⁵ Rifapentine is metabolized by esterases, thus the two-way CYP3A4 interaction as seen with rifabutin is not a concern. However, protease inhibitors may inhibit esterases in the intestinal tract, thus the implications of higher protease inhibitor dosing and the subsequent impact on safety and tolerability of rifapentine are unclear.³⁹ This combination was well-tolerated overall in our study with no observed trends in adverse events, which stood in contrast to our previous study of once-daily dolutegravir with HP in healthy volunteers¹⁴ and other studies examining coadministration of rifamycins with boosted protease inhibitors.³⁵ The underlying mechanism behind the previous findings in healthy volunteers receiving HP with dolutegravir is unknown, but a separate study in PWH demonstrated that this combination was well-tolerated and trough concentrations measured with once-daily dosing exceeded the protein-adjusted IC₉₀.²⁷

There were limitations in this study. This study was performed in healthy volunteers and not in PWH. Assessing this combination in PWH was not possible due to the concerns over reduced ARV exposures and potential for virologic failure with subtherapeutic exposures. However, it is possible that PK or tolerability differences may exist between populations. We were also unable to fully recruit to the minimum of 15 targeted based on initial sample size calculations. However, the magnitude of decrease observed exceeded the 25% threshold, and post hoc sample size calculations demonstrated that a change of 30% with paired data from 10 participants would yield at least 80% power to detect this magnitude of difference. Cobicistat concentrations also were not quantified in this study, but likely were also markedly reduced given that cobicistat is primarily metabolized by CYP3A4 similarly to darunavir.⁹

Based on the results of our study, once daily darunavir exposures when boosted with cobicistat were significantly decreased following the 3HP initiation and should be avoided. 3HP remains an important LTBI treatment option for PWH on antiretroviral therapy. As such, alternative management strategies should be considered for those receiving a darunavir/cobicistat-containing regimen, such as switching to a dolutegravir-containing regimen in those eligible or avoiding 3HP in cases where darunavir/cobicistat must be used. Consultation with an infectious diseases specialist and HIV pharmacist is recommended to construct an optimized medication regimen that addresses treatment for both latent TB and HIV.